Novon a journal of botanical nomenclature from the Missouri Botanical Garden.

Volume 9
Number 1

1999

NOVON

A New Species of Breutelia (Bartramiaceae) from Central America,
with a Key to the Species of Breutelia in Central America

Bruce Allen

Missouri Botanical Garden, P.O. Box 299, St. Ix>uis, Missouri 63166-0299, U.S.A.

Dana Griffin III

Department of Natural Sciences, Florida Museum of Natural History, University of Florida,

Gainesville, Florida 32611-02035, U.S.A.

ABSTRACT. Breutelia pilifera differs from all other
species of Breutelia by its non-sheathing leaves that
have costae excurrent into long, naked awns and
reddish orange basal leaf cells across the leal in¬
sertion. A key to the 14 species of Breutelia in
Central America is given.

Breutelia is a robust genus found mostly at high
elevations in the Neotropics. The genus has some
distinctive, easy to recognize species, but most are
taxonomically difficult. Two characters of critical
taxonomic importance are leal stance and the de¬
gree or type of alar region differentiation. Unfor¬
tunately, proper character-state evaluation of these
features often requires previous familiarity with the
genus. There are two basic leaf stances: spreading
from the insertion and sheathing at the base. This
leaf stance distinction is used in the beginning of
many keys to the species of Breutelia (Bartram,
1949; Griffin, 1994; Churchill & Linares, 1995). It
works tolerably well, but for some species it is dif¬
ficult to evaluate because the two forms grade into
one another and sometimes both can be found with¬
in single collections. Likewise, the degree or type
of alar region differentiation, which often involves
subtle distinctions, can be difficult to evaluate be¬
cause of variation both within and among collec¬
tions. Proper evaluation of this last feature requires
the examination of many leaves.

Breutelia has an odd endostome structure that

helps to bind the Bartramiaceae together. In those
members of the family with a well-developed en¬
dostome the segments are broad and when exam¬
ined with a hand lens appear to be positioned op¬
posite to rather than alternating with the exostome
teeth. Closer examination reveals each segment is
split along the median line nearly to the basal
membrane, and each half segment diverges outward
toward the cilia where it often meets the diverging
half segment of the next segment over, thus framing
the cilia (see Fig. IB). This endostome strucuture
is sometimes inaccurately noted in the literature as
“cilia cohering” (Zales, 1973); however, Shaw and
Hohrer (1984) correctly evaluated the structure.
Within the Bartramiaceae this type of endostome is
found in Bartramia, Breutelia, Philonotis, and Pla-
giopus. Curiously, similar endostomial modification
is found in the Bryaceae ( Acidodontium and
Brachymenium columbicum (De Notaris) Brother-
us).

Breutelia is closely related to Philonotis, and al¬
though the two genera are generally easy to distin¬
guish due to the larger size of Breutelia, technically
they are difficult to separate. Sporophytically, the
genera are identical; gametophytically, only the
larger size, generally narrower and longer leaf cells,
consistently plicate leaf bases, and differentiated
alar cells of Breutelia separate the two genera.
However, Philonotis approximates Breutelia in all
of these features; even their size distinction is

Novon 9: l^U 1999.

2

Novon

blurred, since large plants of Philonotis sphaericar-
pa (Hedwig) Bridel and small plants of Breutelia
jamaicensis (Mitten) A. Jaeger overlap in size.

There are 14 species of Breutelia in Central
America, one of which is described below as new
to science, in preparation for the next volume of
the Moss Flora of Central America (Allen, 1994).

Breutelia pilifera B. H. Allen & D. G. Griffin, sp.
nov. TYPE: Costa Rica. Cartago: Tapantf For¬
est R eserve ca. 25 km SE of Cartago, Griffin
A: Eakin 196 (holotype, MO-4410818; iso¬
types, F, FLAS, NY, US). Figure 1A-H.

Species foliis habentibus aristis longis (0.6— 0.8 mm)
midis a congeneribus differt.

Plants medium- to large-sized, yellowish green
above, glossy, in loose tufts to 70 mm high. Stems
red, in cross section hyalodermis present, central
strand well developed; densely tomentose below,
rhizoids reddish brown, smooth. Leaves 3—4 mm
long, widely spaced, spreading from base, stiffly
erect-incurved above, apices not twisted when dry,
erect-spreading when wet, straight, occasionally
falcate, ovate-lanceolate, long, slenderly acuminate,
not or weakly plicate at base, not decurrent; mar¬
gins plane to narrowly recurved at midleaf, serru¬
late; costa long excurrent into a long, naked awn,
0.6— 0.8 mm long; leaf cells straight-walled, strongly
papillose from the upper end, upper cells linear-
elongate, firm-walled, 30—50 X 2.5— 3.7 p,m, lower
cells linear-elongate, 55—75 X 2.5— 3.7 |xm, firm-
walled; leaf base entirely orange to reddish yellow,
alar region with small cluster (8—15) of enlarged,
thin-walled cells at basal angle. Dioicous. Perigonia
discoid. Setae flexuose, 25—30 mm long; capsules
inclined, asymmetric, 2.5—3 mm long, subglobose,
furrowed when dry, striate when wet, exothecial
cells firm- to thick-walled, stomata immersed, nu¬
merous at base; operculum plano-convex; exostome
teeth red, triangular, finely papillose below, coarse¬
ly papillose near tips, dorsal trabeculae faint, ven¬
tral trabeculae strongly thickened; endostome seg¬
ments broad, yellow, papillose, segments split along
the median line and each half diverging toward the
cilia, cilia rudimentary 0—2. Spores reniform, 17—
23 pun, faintly papillose, pale yellow.

Habitat. Lower montane rainforest, epilithic on
wet rocky roadside embankment; 1000—1800 m.

Breutelia pilifera is a medium- to large-sized
species known only from two collections, both made
in the Tapantf Forest Reserve of Costa Hica. ft has
densely tomentose stems, a shiny, yellow-green col¬
or with widely spaced leaves that spread from the

base and are stiffly erect-incurved above, which
give the plants a “worm-like” aspect not seen in
any other Central American species of Breutelia.
Its most distinctive features are found in its costa,
which is long excurrent into a naked awn, and its
leaf base, which is orange or reddish yellow
throughout. No other species of Breutelia has this
combination of features. In Central America no oth¬
er species has a costa as strongly excurrent as that
of B. pilifera. The only other Central American spe¬
cies of Breutelia with an orange to reddish yellow
leaf base is B. reclinata Brotherus. That species
differs from B. pilifera in having much larger leaves
(8—9 mm long) that are tightly sheathing at base,
and a percurrent to shortly excurrent costa. The leaf
cells of B. pilifera are straight-walled and very
strongly papillose by projecting upper cell end-
walls.

Breutelia pilifera differs from most members of
Breutelia in having faintly papillose rather than
warty-tuberculose spores. Faintly papillose spores
are common in Philonotis, and the same spore or¬
namentation is found in Breutelia jamaicensis,
which otherwise is distinctly philonotoid in aspect.
Interestingly, however, B. affinis (Hooker) Mitten,
another species with a philonotoid aspect, has warty
spores.

The type of B. pilifera was originally distributed
by the Moss Exchange of the American Bryological
and Lichenological Society as B. jamaicensis. These
specimens do not bear a collection number; in¬
stead, they have the herbarium number 019961.
The only type specimens we examined that have
the collection number 196 are the holotype (MO)
and one isotype (FLAS).

Paratype. COSTA HICA. Cartago: Reserva de Ta¬
pantf, Gdmez 18861 (CR. MO).

Key to the Species of Breutelia in Central. America

1. Leaf base tightly sheathing . 2

I '. Leaf loosely appressed or erect at base to wide-

spreading from the insertion . 6

2(1). Leaves to 9 mm long; orange at base; alar
cells undifferentiated, consisting of elon¬
gate, porose cells ... B. reclinata Brotherus
2'. Leaves to 7 mm long; hyaline at base, red¬
dish orange only at insertion line; alar cells
strongly differentiated or with at least 2-4
enlarged cells along the basal margin and
a marginal row of enlarged rectangular

cells . 3

3(2). Alar cells greatly enlarged, thin-walled and
bulging in 4—7 rows, extending nearly to leaf
shoulders; leaf margins broadly revolute at
shoulders; leaves without groups of short lax

cells at top of basal plications .

. B. inclinata (Hampe & Lorentz) A. Jaeger

Volume 9, Number 1
1999

Allen & Griffin

Breutelia pilifera from Central America

3

Figure 1. Breutelia pilifera B. H. Allen & D. G. Griffin. — A. Habit. — B. Part of the endostome ventral (inner)
surface. — C. Leaf. — D. Capsule and operculum. — E. Upper leaf cells at margin. — F. Basal leaf cells in alar region.
— G. Median leaf cells at margin. — H. Inner basal leaf cells near costa. Scale in mm: bar = 0.05 (E— H); bar = 0.1
(B); bar = 0.55 (C); bar = 1.3 (D); bar = 1.78 (A). All figures from Griffin & Bakin 196 (MO).

4

No von

3'. Alar cells somewhat enlarged, thin-walled to
firm-walled; leal margins plane or narrowly rev¬
olute; leaves often w ith groups of short lax cells

at top of basal plications . 4

4(3). Leaves erect-appressed to erect with di¬
varicate tips when dry .

... B. polygastrica (Carl Muller) Brotherus
4'. Leaves squarrose, deflexed, or wide-

spreading, never erect-appressed . 5

5(4). Alar region with 4 -6 rows of enlarged, lax, hy¬
aline cells extending well up the leal margin

. B. austro-arcuata (Carl M tiller) Paris

5'. Alar region with 2-4 enlarged cells along the
hasal margin and a marginal row of enlarged

subrectangular to rectangular cells .

. B. brittoniae Renauld & Cardot

6(1). Costa excurrent into a long, naked, awn
to I mm long; entire leaf base orange to

reddish yellow .

... B. pilifera B. H. Allen & I). C. Criffin
6'. Costa pereurrent or short excurrent; only
the alar region and the area near the leaf

insertion orange or reddish . 7

7(6). Leaves erect-appressed; marginal basal cells
quadrate in 5—10 rows, inner basal cells short-

rectangular to quadrate .

. B. jamaicensis (Mitten) A. Jaeger

7'. Leaves spreading or erect at base, falcate or
squarrose; marginal basal cells enlarged and
lax, or undifferentiated, inner basal cells elon¬
gate . 8

8(7). Leaves long deeurrent; all basal leal cells
and decurrent cells lax and inflated; leaf

margins plane .

. B. auriculala E. B. Bartram

8'. leaves not decurrent, inner basal leaf cells
long-rectangular to linear, firm to thick-
walled, basal cells at margins lax and en¬
larged in 2-4 rows or firm to thick-walled

and small; leaf margins recurved . 9

9(8). Plants robust, leaves to 6—8 mm .

. B. karsteniana (Carl Muller) A. Jaeger

9'. Plants medium- to large-sized, leaves 3—5 mm

long . 10

10(9). Leaves ovate-lanceolate, smooth, grad¬
ually and evenly narrowed above to tip;

caspnles furrowed . 11

10'. Leaves ovate to broadly lanceolate,
smooth or rugose, abruptly narrowed
above to tip; capsules rugulose to nearly
smooth . 12

1 1(10). Branch and stem leaves closely spaced, erect
and overlapping at base, stems not evident;
leaves with small group of enlarged cells at
basal angles, and with enlarged cells extend¬
ing up the basal margin .

. B. chrysea (Carl Muller) A. Jaeger

11'. Branch leaves (occasionally stem leaves)
well spaced, spreading from the insertion,
dark-red stems conspicuous; leaves with I —

2 enlarged cells at the extreme basal angles,
otherwise alar cells undifferentiated . . .

.... B. tomenlosa (Swartz ex Bridel) A. Jaeger
12(10). Alar cells weakly differentiated, I —

3 inflated cells at extreme basal an¬
gles and a few enlarged supra-alar

cells . B. squarrosa A. Jaeger

12’. Alar cells well differentiated, 4—9
rows of enlarged, firm-walled or in¬
flated cells at basal angles, often ex¬
tending up the basal margin in some

leaves . 13

13(12). Intramarginal alar cells non-porose, length to

width 1-2:1 .

. B. subarcuata (Carl Muller) Schimper

13'. Intramarginal alar cells porose, length to

width often more than 2:1 .

. B. rhythidioides Herzog

Acknowledgments. This work was supported by
National Science Foundation Grant DEB-9400996
(to the Missouri Botanical Garden).

I iterature Cited

Allen. B. II. 1994. Moss f lora of Central America. Part
1. Sphagnaceae— Calymperaceae. Monogr. Syst. Bot.
Missouri Bot. Card. 49; 1—242.

Bartram. L. B. 1949. Mosses of Guatemala, fieldiana. Bot.
25: 1-442.

Churchill, S. P. & E. Linares C. 1995. Prodromes Bryol-
ogiae Novo-Granatensis. Bibliot. Jose Jeronimo Triana
12: 1—453 + i-xxvi.

Griffin, G., 111. 1994. Breutelia. Pp. 539—550 in A. J.
Sharp, H. Cnmi & P. M. Eckel (editors). The Moss f lora
of Mexico. Mem. New York. Bot. Gard. 69.

Shaw, A. J. & J. R. Rohrer. 1984. Endostomial architec¬
ture in diplolepideous mosses. J. Hattori Bot. Gib. 57:
41-61.

/.ales, W. M. 1973. A Taxonomic Revision of the Genus
Bhilonotis for North America, North of Mexico. Ph.l).
Dissertation. University of British Columbia, Vancouver.

Lepidium beckii (Brassicaceae), a New Species from Bolivia

Ihsan A. Al-Shehbaz

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

ABSTRACT. Lepidium beckii, a shrubby new spe¬
cies from Depto. La Paz (Bolivia), is described and
illustrated. Its relationship to L. philippianum and
L quitense is discussed.

Of the world’s ca. 180 species of Lepidium L.,
some 52 species are restricted to South America
(Al-Shehbaz, 1993). All except two of the South
American species are herbaceous. A third woody
species, L. beckii, is described below, and the char¬
acters that distinguish it from the other two woody
members, L. quitense Turczaninow and L. philip¬
pianum (Kuntze) Thellung, are discussed.

Lepidium beckii Al-Shehbaz, sp. nov. TYPE: Bo¬
livia. La Paz: Prov. Jose Roma de Loayza, Ban¬
os Termales de Urmiri, matorral, 17‘’09'S,
68°05'W, 3500 m, 26 Jan. 1996, St. G. Beck
21944 (holotype, MO; isotype. LPB). Figure 1.

Suffrutex omnino puberulus. Folia sessilia, anguste li-
nearia, (0.5— )1— 1.5(— 2.5) cm longa. 0.5— 1.2 mm lata. Flos
sepalis oblongis, 1.7—2 mm longis, 0.9— 1.1 mm latis, an¬
guste membranaceo-marginatis; petalis albis, anguste ob-
lanceolato-spathulatis, 2.5-3 mm longis, 0.5— 0.7 mm la¬
tis; staminibus 4. Silicula oblonga vel ovato-oblonga, 4—5
mm longa, 3—3.5 mm lata, puberula; stylo persistente 0.5—
0.7 mm longo.

Subshrubs, all parts except petals and filaments
horizontally puberulent with trichomes 0.05-0.1
mm long. Stems 10—15 cm tall, woody along entire
growth of previous years, 1—2 mm wide, slightly
striate. Leaves all cauline, not forming rosettes,
narrowly linear, (0.5— )1— 1.5(— 2.5) cm long, 0.5— 1.2
mm wide, sessile, somewhat fleshy, entire, widest
at or near base, acute at apex, often strongly con¬
cave adaxially, ascending to subappressed on stem,
puberulent along both surfaces. Inflorescences few-
flowered, ebracteate corymbose racemes, elongated
in fruit. Fruiting pedicels flattened, ascending to
divaricate, 3.5—5 mm long. Sepals oblong, ascend¬
ing, 1.7-2 mm long, 0.9— 1.1 mm wide, not saccate,
membranous margin and apex 0.2— 0.4 mm wide,
apex obscurely denticulate. Petals white, narrowly
oblanceolate-spatulate, 2.5—3 mm long, 0.5— 0.7
mm wide, not clawed, gradually attenuate to base,
apex obtuse. Stamens 4; filaments white, distinctly
dilated at base, 2—2.6 mm long; anthers ovate, 0.5—

0.6 mm long. Silicles oblong to ovate-oblong, 4—5
mm long, 3—3.5 mm wide; valves puberulent,
keeled, not veined; apical notch 0. 1-0.2 mm deep;
style exserted from apical notch, 0. 5-0.7 mm long,
sparsely puberulent basally; stigma entire. Seeds
reddish brown, oblong, ca. 2X1 mm; cotyledons
accumbent.

Lepidium beckii, which is named in honor of
Stephan G. Beck who made the type collection, is
most closely related to the Ecuadorian L. quitense
Turczaninow and the Chilean L. philippianum
(Kuntze) Thellung, both of which have woody low¬
er parts, four stamens, horizontally puberulent
parts, and sessile, usually entire uppermost cau¬
line leaves. The remaining South American spe¬
cies are herbs with two or rarely six stamens. From
Lepidium beckii, L. quitense is readily distin¬
guished by being a taller subshrub (20—50 cm)
with many-flowered inflorescences, sepals ca. 1
mm long, petals shorter than sepals, glabrous
fruits 2.5— 3(— 3.5) mm long, and shorter styles
(0.1— 0.3 mm long) included in the apical notch of
the fruit. In contrast, L. beckii is a shorter plant
(10—15 cm) with few-flowered inflorescences, se¬
pals 1.7—2 mm long, petals longer than sepals,
puberulent fruits 4—5 mm long, and longer styles
(0.5— 0.7 mm long) exserted from the apical notch
of the fruit. Lepidium philippianum differs from L.
beckii in having well-developed rosettes at the
branches of the woody caudex, long petiolate basal
leaves, apically toothed lower leaf blades (2— )3— 5
mm wide, and glabrous fruits. Lepidium beckii
does not produce basal rosettes, and it has sessile,
entire leaves 0.5-1. 2 mm wide and puberulent
fruits. Furthermore, L. philippianum has a woody,
much branched caudex, rather than being typical
subshrubs as are L. beckii and L. quitense.

Hitchcock (1945) suggested that Lepidium phi¬
lippianum is functionally dioecious, but I have not
seen sufficient material to prove that, and no de¬
tailed field studies have been done on the species.
I agree with Hitchcock that the division by Thel¬
lung (1906) of both L. philippianum and L. quitense
into several varieties is unwarranted.

Novon 9: 5-7. 1999.

6

Novon

figure I. Ijepidium beckii Al-Shehbaz (holotype). — A. Part of plant. — B. Cauline leaf. —
E. Stamen. — F. Fruit and pedicel. Scales A = 1 cm; B = 5 mm; C— F = 1 mm.

-C. Sepal. — D. Petal.

Volume 9, Number 1
1999

Al-Shehbaz

Lepidium beckii from Bolivia

7

Acknoivledgments. I am grateful to Stephan G.
Beck lor sending his Brassicaceae collections lor
my study and to Roy Gereau for correcting the Lat¬
in.

Literature Cited

Al-Shehbaz, I. A. 1993. Lepidium tayioriae (Brassica¬
ceae), a new species from Chile. Novon 3: 93—95.

Hitchcock, C. L. 1945. The South American species of
Lepidium. Liloa 11: 75—134.

Thellung, A. 1906. Die Gattung Lepidium (L.) R. Br. Line
Monographische Studie. Neue Denkschr. Allg. Schweiz.
Naturf. Ges. 41(1): 1-340.

A New Species of Hemilophia (Brassicaceae) from China

Ihsan A. Al-Shehbaz

Missouri Botanical Garden, P.O. Box 209, St. Louis, Missouri 63166-0299, U.S.A.

Katsutoshi Arai

Karuizawa Botanical Garden, Hocchi 1166, Karuizawa-machi, Kitasaku-gun,

Nagano Pref., Japan

Hideaki Ohba

Department of Botany, University Museum, University of Tokyo, 7-3-1 Hongo,

Tokyo, 113 Japan

Abstract. Hemilophia sessilifolia is described
and illustrated. Its relationship to the other species
of Hemilophia is discussed.

The genus Hemilophia Franchet consists of three
species, //. pulchella Franchet, //. rockii 0. E.
Schulz, and the new species //. sessilifolia Al-Sheh-
haz, Arai & H. Ohba, all endemic to southwestern
Sichuan and adjacent Yunnan, China.

Hemilophia is readily distinguished from the re¬
lated Chinese Dipoma Franchet by the appendaged
and strongly inflated bases of the median staminal
filaments, slender replum, entire leaves, and
straight fruiting pedicels, and by lacking the fruit
septum. Dipoma lias slender filaments, strongly
flattened replum, apically dentate leaves, strongly
curved and loop-forming fruiting pedicels, and
complete septum. Hemilophia is separated from the
Himalayan and central Asian Dilophia Thomson by
having slender rhizomes, appendaged and strongly
inflated bases of the median staminal filaments,
slender replum, and large (2-3 mm) seeds, I or 2
per fruit. Dilophia is a non-rhizomatous perennial
with slender filaments, broadly winged replum, and
small (ca. 1 mm) seeds, several per fruit.

The new species, Hemilophia sessilifolia , which
is known only from the type collection, was col¬
lected during a recent Sino-Japanese expedition to
Dauxue Shan, Yunnan.

lltMiiilopliia sessilifolia Al-Shehbaz, Arai & H.
Old ta, sp. nov. TYPE: China. Yunnan: Deqe,
around Dauxue Shan, 28°34'N, 99°48'E,
4300—4550 m, 28 Aug. 1996, Wu Sugong, Hi¬
roshi Ike da, Michio Wakabayashi, Futoshi Mi¬
yamoto, Yang Yongping & Takao Kikuchi 953
(holotype, TI; isotypes, KUN, MO). Figure 1.

Species ab //. pulchella Franchet et //. rockii (). E.
Schulz foliis caulinis sessilibus, sepalis membranaceis in-
tegris (acile cadentibus, petalis persistentibus, stylo gla-
bro. et trichomatibus simplicibus obtecta bene differt.

Herbs perennial, rhizomatous. Rhizomes slender,
not scaly. Stems 1 to several from basal rosette or
solitary from rhizomes, 3—9 cm tall, simple or few-
branched, with spreading to appressed simple tri-
chomes 0.06—0.25 mm long. Basal leaves oblan-
ceolate, 6—10 X 1.5— 2.5 mm, dry at anthesis,
densely beset with straight simple trichomes 0.8-
1.3 mm long. Cauline leaves oblong to oblong-ob-
lanceolate, 2.5— 5(— 8) X l-2(— 2.5) mm, sessile, en¬
tire, obtuse at apex, glabrous or minutely Hairy as
on stem. Inflorescences corymbose racemes, not
elongated in fruit, bracteate throughout; bracts sim¬
ilar to cauline leaves. Sepals obovate, membranous,
early caducous and leaving 4 receptaeular teeth,
not saccate, 1.8— 2.7 X 1.2— 1.5 mm, entire, round¬
ed at apex. Petals persistent to fruit maturity,
broadly obovate, 2-lobed, abruptly narrowed to
claw, 6—7 X 4—5.5 mm, creamy white with dark
green veins, pale to light brown at mouth, apical
notch to 2 mm deep; claws 1.5— 2.5 mm long. Fil¬
aments white; lateral filaments slender, unappen-
daged, 1. 8-2.2 mm long; median filaments 1.9-2. 5
mm long, basal Half strongly inflated to form an
oblong appendage 1.1— 1.3 mm long; anthers green.
Nectar glands surrounding base of lateral stamens,
subtending base of median filaments. Fruiting ped¬
icels slender, straight, divaricate, 5—6 mm long, pi¬
lose. Fruit oblong, dehiscent, glabrous; valves thin
papery, navicular, 3-4 X 1.5-2 mm, with 3 rows
of crests, 1 row on midvein and on either margin;
replum slender, not flattened; septum absent; gy-
nophore obsolete to 0.1 mm long; style glabrous,
conical, 1—1.5 mm long; stigma entire. Seeds 1 or

Novon 9: 8—10. 1999.

Volume 9, Number 1
1999

Al-Shehbaz et al.

Hemilophia sessilifolia from China

9

Figure 1. Hemilophia sessilifolia Al-Shehbaz, Arai & H. Ohba. — A. Plant. — B. Sepal. — C. Petal. — I). Median
filament. — E. Lateral filament. — F. Fruit and persistent petals. — G. Fruit, dorsal view. — H. Fruit after removal of
valves and seeds. — I. Seed. Scale bars: A = 1 cm; B-E, G-I = I mm; F — 5 mm. Drawn from the isotype al MO.

10

Novon

2 per fruit, oblong, 2.5—3 X 1.1-1. 3 mm, smooth;
cotyledons obliquely accumbent.

Hemilophia sessilifolia is readily distinguished
from the two other species of the genus, //. pul-
chella and H. rockii, by the sessile cauline leaves,
membranous, readily caducous entire sepals, per¬
sistent petals, glabrous conical style, and exclu¬
sively simple trichomes. Both //. pulchella and //.
rockii have petiolate cauline leaves, green sepals

that remain until fruit develops, caducous petals, a
minutely papillose cylindric style, and 2-forked tri¬
chomes on the stem. Hemilophia pulchella is sep¬
arated from //. rockii by having purplish petals 3-
4 mm long instead of white petals 5-7 mm long,
stem trichomes with long and straight instead of
short and crisped rays, and an entire instead of
ciliate sepal margin.

Acknowledgment. We thank Neil Harriman for
his review of the manuscript.

Juncus anthelatus (Juncaceae, Juncus subg. Poiophylli), a New
Status for a North American Taxon

Ralph E. Brooks

Black & Veatch, 4004 Kruse Way Place, Suite 200, Lake Oswego, Oregon 97035, U.S.A.

Alan T. Whittemore

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

ABSTRACT. The taxon long known as Juncus ten¬
uis Willdenow var. anthelatus Wiegand differs from
J. tenuis in morphology (it is more robust than J.
tenuis, with a more diffuse inflorescence and slight¬
ly smaller capsules), habitat, flowering time, and
isozyme profile, and should be treated as a distinct
species, Juncus anthelatus (Wiegand) R. E. Brooks,
stat. nov.

Studies in Juncus subg. Poiophylli Buchenau in
eastern North America (Brooks, 1989; Brooks &
Clemants, in press) revealed that the entity formerly
known as Juncus tenuis Willdenow var. anthelatus
Wiegand is genetically distinct from J. tenuis, and
should be considered a distinct species, Juncus an¬
thelatus.

Juncus (subg. Poiophylli ) anthelatus (Wiegand) R.
E. Brooks, stat. nov. Basionym: Juncus tenuis
Willdenow var. anthelatus Wiegand, Bull. Tor-
rey Bot. Club 27: 523—524. 1900. Juncus mac-
er Gray var. anthelatus (Wiegand) Fernald, J.
Bot. 68: 367. 1930. TYPE: not designated.

Juncus macer Gray f. discretiflorus I . J. Hermann, Rhodora
40: 82. 1938. Juncus tenuis Willdenow f. discretiflo-
rus (F. J. Hermann) Fernald, Rhodora 47: 123. 1945.
TYPE: U.S.A. Indiana: Harrison County, 13 July
1935, Deam 56381 (holotype, GH; isotype, INI)).

Caespitose perennial herb from a branching rhi¬
zome. Culms erect, to (3-)7-9 dm tall. Cataphylls
1-2 per culm. Leaves basal, 2-3(— 5) per culm; au¬
ricles 2—3.5 mm long, acutish, membranous; blades
flat, ( 10 — )20 — 30 cm X 0.5— 2.3 mm. Inflorescences
10— 100-flowered, diffuse, internodes of monochasia
greater than 6 mm; primary bract usually exceeding
inflorescence. Flowers: pedicels 0.1— 0.3(— 3) mm;
bracteoles 2; tepals green, lanceolate, 3.2-4. 5 X
0.7-1 mm, inner and outer series subequal, in fruit
proximal part enfolding the capsule, apex erect;
stamens 6, filaments 0.8— 1.1 mm, anthers 0.3-0. 7
mm; style ca. 0.2 mm. Capsules tan, 1-locular,

widely ellipsoid to obovoid, 2-3.2 X 1. 1-1.6 mm.
Seeds tan, ellipsoid, 0.33—0.55 mm, untailed. Chro¬
mosome number: 2 n = 80.

Flowering and fruiting spring. Exposed or par¬
tially shaded sites in moist or seasonally wet, sandy
or clay soils. Widespread in the eastern United
States (Alabama, Arkansas, Connecticut, Georgia,
Illinois, Indiana, Iowa, Kansas, Kentucky, Louisi¬
ana, Maine, Michigan, Minnesota, Mississippi, Mis¬
souri, New Hampshire, New York, North Carolina,
Ohio, Oklahoma, Pennsylvania, Tennessee, Virgin¬
ia, Vermont, Wisconsin).

Representative specimens examined. U.S.A. Arkansas:
Logan Co., 8 Apr. 1985, Teale s.n. (KANU). Georgia:
Pike Co., 7 Aug. 1985, Brooks 17593 & Kuhn (KANU).
Illinois: Clark Co., 22 Aug. 1986, Brooks 18547 & Kuhn
(KANU). Kansas: Leavenworth Co., 7 June 1985, Brooks
17473a A: Kuhn (KANU). Maine: Penobscot Co., 6 Oct.
1932, Ogden 599 (MAINE). Massachusetts: Hampshire
Co., 28 Aug. 1976, Ahles 82384 (VT). Mississippi: Hinds
Co., 8 Aug. 1985, Brooks 17627 & Kuhn (KANU). Mis¬
souri: Camden Co., 26 May 1985, Brooks 18135 & Kuhn
(KANU). Ohio: Lucas Co., 15 Oct. 1979, Stuckey 10005
(OS). Pennsylvania: Adams Co., 8 Aug. 1986, Brooks
18390 (KANU). Tennessee: Knox Co., Kearney, Jr. s.n.
(OS). West Virginia: Fayette Co., 17 Aug. 1985, Brooks
17962 & Kuhn (KANU).

Juncus anthelatus tends to be more robust than
J. tenuis, with a more diffuse inflorescence and
slightly smaller capsules, and it often grows in wet¬
ter habitats than J. tenuis, though the two species
are sometimes found growing in the same area. Al¬
though the two species can be difficult to distin¬
guish morphologically, J. anthelatus flowers one to
two weeks earlier than nearby populations of J. ten¬
uis, both in the field and in the greenhouse, and
the two species show consistent isozyme differenc¬
es; J. anthelatus has a distinctive PGI-1 profile,
which does not vary within or between populations
across the range of the species (Brooks, 1989).
Morphologically, the two species are best distin¬
guished as follows:

Novon 9: 1 1-12. 1999.

12

Novon

la. Capsules 2—3.2 mm long, less than 34 as long as
tepals, borne widely spared along the usually
diffuse branches of the inflorescence; culms

(3— )7— 9 dm tall . Junius anthelatus

lb. Capsules (3.3— )3. 8— 4.7 mm long, 34 or more as
long as tepals, congested or the branch inter¬
nodes about as long as the tepals; culms 1.5—5

dm tall . Juncus tenuis

I .iterature Cited

Brooks, K. E. 1989. A revision of Junius subgenus l‘oio-
phylli (Juncaceae) in the eastern United States. Univ.
Microfilms International, Ann Arbor, Michigan.

- & S. E. Clemants (In press). J uncus. In: Flora of

North America Editorial Committee. Flora of North
America, vol. 22. Oxford Univ. Press, New York.

Two New Species of Hoffmannia (Rubiaceae) from Panama, with
Remarks on Circumscription of Mesoamerican
Species of the Genus

William C. Burger

Botany Department, The Field Museum, Chicago, Illinois 60605-2496, U.S.A.

Abstract. The unusual degree of variation with¬
in Hoffmannia populations and the difficulty in
clearly defining some Mesoamerican species are
discussed. Two new species, Hoffmannia divyeri
and H. fortunensis , both endemic to the Chiriquf
highlands, are described.

Having worked with Hoffmannia for the Flora
Costaricensis series (Burger & Taylor, 1993), it
seemed appropriate to expand that effort and treat
the genus for the Flora Mesoamericana project.
Work with the Costa Rican material had indicated
that most species in the genus were very variable
morphologically, and that making specific distinc¬
tions in some groups was extremely difficult or ar¬
bitrary. Unfortunately, studying over 3000 speci¬
mens of Hoffmannia in the region covered by the
Flora Mesoamericana has made these difficulties
abundantly clear.

While individuals of most plant species vary
considerably, there are usually concordant suites of
consistent morphological traits that allow us to
identify and distinguish species from their close
congeners. Floral dimensions and morphology are
usually sufficiently constrained so that they are
useful in characterizing species. Leaf dimensions
may vary greatly, but leaf form and venation pat¬
terns are usually also helpful in identifying and dis¬
tinguishing closely related species. Many species
of Hoffmannia , however, exhibit a wider-than-nor-
mal range of variation in both floral dimensions,
vesture, plant size, and leaf size and shape. This
wider range of variation is rarely exhibited by any
one plant; it becomes apparent ordy when a larger
suite of collections is gathered together.

We use larger suites of collections to assess pop¬
ulation variation, characterize species, determine
species differences, delineate ranges, and look for
local or regional differentiation. If we are able to
effectively separate congeneric material into two
closely related species, and then find that they con¬
sistently grow in different areas or slightly different
habitats, we are reassured that they are indeed dif¬
ferent species. If we find a few collections that are

intermediate in a number of characteristics that
distinguish two closely related species, both rep¬
resented by large suites of collections, we may be
confident that we have identified hybrids. If there
are many intermediates between two distinctive
morphological populations, we may be dealing with
a cline or subspecific differentiation. In both these
instances, however, we must first describe and
characterize two different suites of character states,
delimiting our species or subspecies. The problem
in Hoffmannia is that the variation is often so great,
and there is so little correlation between different
characters, that delimiting two closely related “spe¬
cies” among a large suite of similar collections can
be quite arbitrary. The only effective way to tackle
this kind of problem is to separate out the most
distinctive entities first, then try to tease out other
distinctive taxa, before finally dealing with a large
complex of similar specimens.

Clearly, one of the most easily distinguished spe¬
cies in the genus is Hoffmannia vesiculifera Stand-
ley, which has two inflated elongate vesicles along
the lateral sides of the usually short petiole. Hoff¬
mannia kirkbridei Dwyer also has such vesicles, but
that species has very different petioles and leaf
form, is known from only two collections, and may
prove to be a variant of //. vesiculifera. Setting aside
the two collections tentatively assigned to //. kirk¬
bridei, and using the vesicles as a defining aut-
apomorphy for //. vesiculifera, we can bring together
a number of specimens that display a wide range
of variation in both leaf and inflorescence charac¬
teristics. The plants themselves grow from 0.1 to
1.5 m in height when flowering, and range from
central Costa Rica to central Panama. The inflores¬
cences vary from sessile congested fascicles to pe¬
dunculate and distally branched paniculate forms.
Leaves range from 10 to 30 cm in length and vary
from elliptic to oblong or obovate. Calyx lobes may
be 4—20 mm long; the corolla varies from about 10
to 20 mm in length. These extremes are never
found on the same individual, and they do not char¬
acterize particular habitats, localities, or differing
elevations. Rather, the extreme individuals appear

Novon 9: 13-17. 1999.

14

Novon

to he nothing more than unusual variants. Some of
the more distinctive variants have been the basis
for erecting new species, but all can be linked with
the earliest type through intermediate collections.
In the eastern part of this species’ range, inflores¬
cences are often borne on peduncles with small
leaves at their apex (modified axillary branches). In
the same area, there are plants with flowers densely
crowded in the leaf axils, as in the type of H. morii
Dwyer, based on Mori & Kallunki 5915 (MO). At
first, the two forms appear to be different species,
but a larger series of specimens will include some
with intermediate inflorescence characteristics. In
short, //. vesiculifera appears to be a species easily
identifiable by the petiolar vesicles, larger calyx
lobes, generally short habit, and pubescence; but it
is a species that varies rather dramatically in vir¬
tually all of its morphological features.

If we think of //. vesiculifera as being typical of
Hojfmannia species, wide patterns of variation in
other species of the genus become easier to under¬
stand. What gives us problems is the fact that these
other species usually do not possess a similarly
unique identifying trait. Hojfmannia vesiculifera
and the poorly known 11. kirkbridei Dwyer are the
only species of the genus to have developed petiolar
pouches large enough to house small ants. A num¬
ber of species, such as the two new species being
described below, do have a suite of distinctive traits
and limited geographic range that make them rel¬
atively easy to recognize, but by the time we sep¬
arate out such species we are left with wide-ranging
taxa that may be catchalls for large numbers of sim¬
ilar-looking material.

Hojfmannia regalis (J. I). Hooker) Hemsley was
based on horticultural material grown in Europe,
and the name has been used for some wild collec¬
tions from southern Mexico. The geographic origin
of the original cultivated material was not known,
but similar material of Hojfmannia is found only in
Chiapas and Guatemala. The large leaves are oc¬
casionally broad and brilliantly colored with shades
of purple beneath and a rich dark green above, the
obvious reason for its introduction as a horticultural
novelty. As in the case of 11. vesiculifera , there is
great variation in leaf size, shape, and coloring.
Leaf bases, petioles, and inflorescences are also
highly variable. The result of this extensive within-
population variation is that there is no way to clear¬
ly separate those specimens that have in the past
been identified as Hojfmannia lineolata J. Donn.
Smith and 11. riparia Standley. This latter material
has been distinguished by its more often elliptic
leaves with well-defined petioles and a lamina base
that is not so long-decurrent. A large suite of col¬

lections makes it clear that these differences are
bridged by many intermediates at all elevations.
While the large, broad, brilliantly colored leaves
are only found at elevations above 1200 m, there
is sufficient variation at all elevations to make rec¬
ognition of more than one species dubious.

Hojfmannia subauriculata Standley is quite dis¬
tinctive because of its long-tapering lamina base
and winged petioles, but it also exhibits consider¬
able variation. The winged petioles are usually
rounded and subauriculate at the stem, but not al¬
ways. The plants are often short in stature with in¬
florescences usually borne at the lower leafless
nodes, even at rooting nodes. The species ranges
from Costa Rica to Colombia. Specimens from Da¬
rien, Panama, are distinctive because their consis¬
tently oblanceolate leaves tend to dry grayish, and
their winged petioles are not aurieulate at the base.
Also the stems are often horizontal and rooting
along the proximal nodes in eastern Panama. The
name 11. eliasii Dwyer would be appropriate if these
plants of eastern Panama were to be separated from
//. subauriculata as a distinct species, but it seems
best to consider them part of a single variable spe¬
cies.

An extremely variable complex is found in the
Chiriquf highlands of Panama, for which the name
H. pittieri Standley has often been used. These pop¬
ulations may intergrade with material referable to
the earlier named 11. arborescens J. Donn. Smith of
Costa Rica’s Talamanca mountains. I have separat¬
ed these collections by annotating those with usu¬
ally axillary branched or cymose inflorescences as
H. arborescens, and those in which the flowering
inflorescences are simpler and largely confined to
the lower leafless nodes as H. pittieri, consistent
with the types. Unfortunately, there are intermedi¬
ate collections both as regards their inflorescences
and in other characteristics. It may be that this
complex of forms should be further divided accord¬
ing to specific morphological traits, but I suspect
that this is unrealistic. A closer examination of the
populations in Chiriquf, and especially their cytol¬
ogy, might help unravel this problem. There is a
report of within-species polyploidy in Hojfmannia
(C. M. Taylor, pers. comm.), and this may be a fac¬
tor in the variability found within the genus.

The most commonly collected species of Hojf¬
mannia in Mexico and Central America is 11. ni-
cotianifolia (M. Martens & Caleoti) L. 0. Williams.
Based on material collected in Veracruz, Mexico,
the name suggests plants with larger leaves. Among
Veracruz collections there are large-leaved collec¬
tions, a greater number of smaller-leaved collec¬
tions, and some intermediates. Based on this vari-

Volume 9, Number 1
1999

Burger

Hoffmannia from Panama

15

ation in the region in which the earliest type was
collected, I believe it is best to interpret //. nico-
tianifolia very broadly. However, adopting a broad¬
er inteq>retation can cause difficulty in dealing with
distinctive variants that may be common in some
geographical regions such as Honduras or Nicara¬
gua.

A large percentage of Hoffmannia specimens
from Nicaragua dry with a characteristic orange-
brown or olive-green coloration and many have
larger obovate to oblanceolate leaves. Their small
flowers and short pedunculate inflorescences with
thin pedicels clearly ally them with H. nicotiani-
folia. Because they are quite distinctive and char¬
acteristic of so much material from Nicaragua, they
could be recognized as a separate species, first de¬
scribed by L. 0. Williams as //. oreophila. However,
with much variation and many intermediate collec¬
tions, it seems best to include H. oreophila in a
broad circumscription of //. nicotianifolia. This is
consistent with also treating H. angustifolia Stand-
ley of Guatemala and Honduras as a synonym of
H. nicotianifolia. After a review of many collec¬
tions, I believe a broader, more inclusive species
concept is justified, making //. nicotianifolia a spe¬
cies that ranges from northeastern Mexico to east¬
ern Panama. However, larger circumscription does
not eliminate the problem of having collections that
appear to be intermediate with other closely related
species. In the case of H. nicotianifolia there are
collections that appear to be intermediate with H.
cryptoneura Standley, //. excelsa (Kunth) K. Schu¬
mann, H. honduremis Standley, H. laxa Standley,
//. oreophila, H. pittieri, H. psychotriifolia (Ben-
tham) Grisebach, H. regalis, and //. steyermarkii
Standley. Whether any of the “intermediate” col¬
lections are true hybrid-like products of interspe¬
cific gene flow or simply variants with character¬
istics of one of the other species is difficult to
determine.

Hoffmannia psychotriifolia is quite similar to H.
nicotianifolia but differs from the latter by the usu¬
ally longer corolla and by the consistently sessile
inflorescences. The two are sometimes collected to¬
gether. The names Hoffmannia rotundata Standley
(holotype: Purpus 7268, MO) and H. chuipensis
Standley (holotype: Purpus 7268, US) were based
on a collection that Standley recognized as mixed.
All the herbarium sheets seen have some elements
of the two similar species. Having reviewed the de¬
scriptions and examined sheets from F, US, and
MO, it appears that H. rotundata is a synonym of
H. nicotianifolia in a wide sense, while H. chiapen-
sis is a synonym of //. psychotriifolia. The type

sheets and their respective parts have been anno¬
tated accordingly.

In central Panama there are specimens of Hoff¬
mannia nicotianifolia with more ovate leaves that
have been identified as H. cercidifolia Dwyer by
Dwyer. However, the type of H. cercidifolia (Duke
15597 (2), MO) was collected in Darien and ap¬
pears to represent a different species. Unfortunate¬
ly, the type has only a few subsessile immature
flowers and provides insufficient information to
come to any firm conclusion regarding its place¬
ment. In addition, there are no similar collections
from Darien to give us a better idea of the popu¬
lation that this collection represents. It seems likely
that the type of H. cercidifolia may prove to be an
unusual collection of an earlier named South Amer¬
ican species, but it will require better sampling of
the Darien populations before that likelihood can
be verified.

At present a draft treatment of Hoffmannia for
the Flora Mesoamericana includes 49 species.
Having accepted many broader species concepts, I
have also tended to tentatively place unusual col¬
lections within a related described species rather
than erect new names. However, two small groups
of collections from the Chiriquf highlands appear
to represent two new and distinctive species.

Hoffmannia dwyeri W. Burger, sp. nov. TYPE:
Panama. Chiriquf: vicinity of Boquete, Cerro
Pate de Macho, SW slope, 8°46'N, 82°25'W,
1800-1950 m, 19 June 1987, Croat 66403
(holotype, MO; isotype, F). Figure 1.

Species habitu cum Hoffmannia valerii Standley uptime
congruens, sed differt coroll is longioribus, foliis basi cu-
neatis, petiolis alatis, stigmatibus latis, habitatione alti-
tudo alta (1700—2 1(K) m).

Herbs 0.2-0. 8 m tall, leafy stems terete, 2-6 mm
thick, subglabrous to villous with reddish brown
hairs 0.5—2 mm long; stipules 1—3 mm long, thin,
deciduous. Leaves opposite, subglabrous to villous,
lamina obovate to elliptic-obovate or narrowly ob¬
ovate, 7—23 X 3—9 cm, base cuneate and decurrent
to form a wing along the petiole, attenuate or
abruptly rounded and subauriculate at the base,
apex acute to acuminate, 2° veins 6—12 pairs; pet¬
ioles 2—12 mm long. Inflorescences axillary, 2-6
per node, 2-5 cm long, flowers in a close cymose
grouping or paniculate, peduncle 10-27 mm long,
villous to subglabrous. Flowers on pedicels 1—4 mm
long, hypanthium 2-4 mm long, villous with hairs
1—2 mm long, calyx lobes 2-A mm long, narrowly
triangular; corolla funnelform, 8—16 mm long, white
with pink tips to pinkish red, with few distal tri-

16

Novon

figure 1. Hoffmann ia dwyeri W . Burger. — A. Flowering stern. — B. Flower. A based on Hummel 7412 (F) and Sytsma
el dl. IH.'ifr ( \i < )); B from Hummel 7412 (MO).

chomes, lube 4—5 mm long; anthers ca. 3.2 mm
long, subsessile; style ca. 9 mm long, stigma ca. 3
mm long, broadly ovate. Fruits oblong, 8—9 mm
long, 5—6 mm thick, pink to red-magenta, spongy,
seeds blue.

Hoffmannia dwyeri has an unusual characteris¬
tic: the leal base is cuneate-decurrent and the
winged margins are often revolute near the base,
forming a partly enclosed space abaxially. Its short
herbaceous habit, long reddish brown multicellular
hairs (when present), short densely flowered inflo¬
rescences, prominent calyx lobes, relatively large

corolla, and broad stigmatic lobes are additional
distinguishing features. Pubescence and floral mor¬
phology are similar to //. valerii Standley, but that
species differs in the form of the lamina base and
petiole, and its lower elevation habitat. The new
species has only been collected around the Cerro
de Pate Macho and upper headwaters of Rfo Palo
Alto, between 1700 and 2100 m elevation. It flow¬
ers in March-June.

John Dwyer, who studied Hoffmannia intensively
over more than 20 years, recognized material of this
species as new in two of his annotations but did
not publish either of the names. Because Dwyer’s

Volume 9, Number 1
1999

Burger

Hoffmannia from Panama

17

(1969, 1980) publications and many identifications
have been so helpful to all those working with this
very difficult genus, it seems especially fitting that
a distinctive species should bear his name.

Paratypes. PANAMA. Chiriqui: vicinity of Boquete;
Cerro Pate de Macho, SW slope, 1800-1950 m. 19 June
1987, Croat 66403 (MO); Rfo Palo Alto mad to Chiriqui
border with Bocas del Toro Province near peak of Cerro
Pate Macho, 6200 ft., 20 Nov. 1978. Hammel 5814 (MO);
end of road past Palo Alto to Bocas, 6200 ft.. 18 Mar.
1979. Hammel 6510 (MO); Palo Alto, 4.5 mi. NE of Bo¬
quete, forest along western branch of headwaters of Rio
Palo Alto, 6300 ft., 24 May 1979. Hammel 7412 (MO); S
slopes of Cerro Pate Macho along Rfo Palo Alto, 1300—
1800 m, 11 Nov. 1981, Knapp et al. 2083 (MO); trail to
Cerro Pate Macho, above Palo Alto, 1700-2 KM) m, 15
Mar. 1982. Knapp et al. 4249 (MO); SE slopes and summit
of Cerro Pate Macho, 4 km NE of Boquete, 1700-2KM)
m, 26 May 1981. Systma et al. 4836 (MO).

HofTiiiannia fortunensis Dwyer ex W. Burger, sp.
nov. TYPE: Panama. Chiriqui: S.O. del cam-
pamento de Fortuna (Hornito), sitio de presa
subiendo hasta la finca Pittf, 1000-1200 m,
14 ago. 1976, M. D. Correa, R. Dressier & C.
Garibaldi 2415 (holotype, PMA not seen; iso¬
types, F, MO).

Suffrutices ad 2 m alti, caulibus teretibus, glabris. Folia
ad 23 cm longa, lamina anguste lanceolata, glabra, in sic-
co discolor, venis lateralibus 9—15 paribus. Inflorescentiae
brevi. Flores glabri, lobis calycis brevibus, corollis 4—7
mm longis. Fructus maturi non visi.

Slender shrubs 0.5—2 m tall, leafy stems 2—7 mm
thick, terete, glabrous. Leaves opposite, blades 11—
23 cm long, 1.5—8 cm wide, linear-lanceolate to
narrowly ovate-elliptic or elliptic-oblong, eharta-
ceous, dark above, grayish beneath, glabrous, base
cuneate and decurrent, apex acute to acuminate, 2°
veins 9-15 pairs, loop-connected along the margin,
petioles 15—65 mm long. Inflorescences 1—3 per
axil, at lower or leafless nodes, racemose, 3—9 cm
long, glabrous, drying dark, peduncles 2—4 cm,
0.3— 0.4 mm thick. Flowers glabrous, pedicels to 10
mm long, hypanthium 2—3 mm long, calyx lobes
0.2— 0.5 mm long, corolla 4—7 mm long, pale green

or reddish green, corolla tube 1.5—2 mm long; an¬
thers 2.5 mm long. Mature fruits not seen.

The usually long petioles and very narrow leaves
with clearly demarked loop-connected secondary
veins close to the lamina margins are distinctive
features. In addition, the leaves often dry very dark
above and grayish beneath. Slender racemose inflo¬
rescences with a single raehis drying dark, small
flowers, and restriction to the Fortuna area of the
Chiriqui highlands between 1000 and 1500 m el¬
evation are additional distinctions. This species ap¬
pears to be related to Hoffmannia laxa Standley,
with which it shares the slender glabrous inflores¬
cences and small flowers. There may be lower-el¬
evation collections that have a few characteristics
of both species, such as von Wedel 2363 (MO).

Paratypes. PANAMA. Chiriqui: vicinity of Gualca 1 1
mi. from Planes de Hornito, La Fortuna on road to dam
site, elev. 4(XM) ft.. 9 July 1980. Antonio 5050 (MO); N.O.
del Campamento Fortuna (Hornito), 1 000— 1 2(K) m. 16
ago. 1976, Correa et al. 2515 (MO); Fortuna dam site,
forest area. 1400-1600 m, 15 Sep. 1977, Folsom el al.
5568 (MO); near site of dam, lower slopes of Cerro For¬
tuna, 1150 m. 18 June 1982. Knapp & Vodicka 5584
(MO).

Acknowledgments. David Lorence kindly pro¬
vided me with information relating to Hoffmannia
from his draft manuscript “Nomenclator of Mexican
and Central American Rubiaceae,” which was very
helpful. Mireya Correa verified that her collection
2415, the designated holotype for H. fortunensis,
was present at PMA. David Lorence and Charlotte
Taylor provided helpful suggestions for revision of
the manuscript. I am grateful to the curators of the
following herbaria for making their collections
available for study: CAS, F, MEXU, MO, US, WIS.

Literature Cited

Burger, W. C. & C. M. Taylor. 1993. Flora Costaricensis.

Family #202 Rubiaceae. Fieldiana Rot. n.s. 33: 1-333.
Dwyer, J. D. 1969. The genus Hoffmannia (Rubiaceae) in

Panama. Ann. Missouri Bot. Card. 56: 269-286.
- . 1980. Rubiaceae. In R. E. Woodson & R. W.

Schery (editors), Flora of Panama. Ann. Missouri Bot.

Card. 67: 1-522.

Beilschmiedia manantlanensis (Lauraceae), una Nueva Especie de

Jalisco, Mexico

Ramdn Cuevas G.

Institulo Manantlan de Ecologfa y Conservaeion de la Biodiversidad, Universidad de Guada¬
lajara, Av. Independencia Nacional 151, Autlan de Navarro, Jalisco, C.P. 48000 Mexico

Theodore S. Cochrane

Dept, of Botany, University of Wisconsin-Madison, Birge Hall, 430 Lincoln Dr., Madison,

Wisconsin 53706-1381, U.S.A.

RESUMEN. Se propone a Beilschmiedia manant¬
lanensis, una nueva especie de Lauraceae conocida
unieamente de la vertiente sur de la Reserva de la
Biosfera Sierra de Manantlan, la cual crece en las
canadas que bajan hacia Cuzalapa y Casimiro Cas¬
tillo. La especie se describe e ilustra, y se compara
morfoldgicamente con otras especies.

Abstract. Beilschmiedia manantlanensis is a
new species from the rugged western portion of the
Sierra de Manantlan Biosphere Reserve in Jalisco,
Mexico, an area known to be rich in endemic and
disjunct species. The new species is described and
illustrated, and its distinguishing characters from
similar species are summarized.

A traves de la continuacion de los proyectos de
inventario de la flora vascular de la Sierra de Ma¬
nantlan y Flora Arbdrea de la Estacidn Cientifica
Las Joyas, Municipio de Autlan, Jalisco, Mexico,
que se llevan a cabo entre personal de la Univer¬
sidad de Guadalajara y la Universidad de Wiscon¬
sin, se siguen encontrando especies nuevas para la
ciencia, como la perteneciente al genero Beilsch¬
miedia, la cual es propuesta a continuacion.

Beilsehmiedia iiianantlanensis Cuevas & Coch¬
rane, sp. nov. TIPO: Mexico. Jalisco: Mpio.
Autldn, Canada del Alentrisco, 1—2 km al S
de Corralitos, 19°36'19"N, 104°18'20"W, 1900
m, en bosque inesdfilo de montafia, 26 mar.
1998 (fl), L. Guzman & F. J. Santana 1616
(holotipo, ZEA; isotipos, BM, CAS, CHAPA,
ENCB, F, G, IBUG, IEB, MEXU, MICH, MO.
NY, TEX, UCR, US, WIS, XAL). Figuras 1. 2.

Arbor 15-35 m alta, 0.5-1. 2 m diametro. Folia alter-
nata vel subvert icil lata, laminis ellipticis vel oblongis vel
oblongo-oblanceolatis, 10—17 cm longis et 3-7 cm latis,
a[)ice acutis, obtuse subacuminatis vel rotundatis, basi cu-
neatis vel obtusis. penuinerviis, uervis primariis utroque
costae latere 10—16 subpatentibus (sub angulo 50—75°

abeuntibus), venulae terminis interdum liberis et reticu¬
lum venularum ergo incompletum. luflorescentia axillaris
aspectu terminalis racemoso-paniculata, 2-4 cm longa, fe-
rrugineo-tomentella glabrescentes glauca. Flores parvi bi-
sexuales, perianthio cupulatis, lobis ovato-triangularibus
ad suborbiculares 1—2 mm longis, tubo receptaculi 1.5-3
mm longo; androeeium verticillastris 4 staminum 3, sta-
miuibus fertilibus 9, bilocularibus, seriebus 2 exterioribus
eglandulosis, serie 3 conspicue biglandulosis. antheris se-
rierum 2 exteriorum introrsis, serierum interiorum |>oris 2
lateralibus dehiscentibus, serierum 4 staminodiis 0.6—
0.75 mm longis; gynoecio 2-3 mm longo, ovario glabro.
Bacca ellipsoidea 3.5— 4.5 cm longa et 1 .8-2.5 cm dia¬
metro.

Arbol perennifolio, de 15—35 m de altura, 0.5—
1.2 m de didmetro, corteza escamosa, rojiza; ra-
millas de 3—4 mm de diametro, rojizas, glabras a
glabrescentes, generalmente pruinosas, exfoliantes
con la edad; hojas alternas o mas comunmente
agrupadas en los apices de las ramillas, limbo elfp-
tico a oblongo u oblongo oblanceolado, algunas ve-
ces obovado, de 10—17 cm de largo, 3-7 cm de
ancho, glabros, cuando jovenes esparcido puberu-
lent os en ambas caras, dpice agudo, redondeado o
oscuramente aeuminado, marge n entero, base agu-
da, algunas veces obtusa, nervat ion prominente-
mente reticulada, algunas de las venillas de los ul-
timos ordenes terminando libres, y entonces forman
una reticula incompleta, con 10-16 pares de ner-
vios primarios que forman dngulos de 50—75° con
el nervio medio, el nervio medio y primarios de-
presos en el haz y salientes en el enves; pecfolos
de 5-14 mm de largo, acanalados adaxialmente,
glabros a esparcidamente pubescentes, algunas ve¬
ces exfoliantes; inflorescencias axilares (con apa-
riencia de terminales o subterminales), paniculifor-
mes, de 22—10 mm de largo, ferrugfneo tomentosas,
el tomento caedizo, y entonces glaucas; pedunculo
de 5-12 mm de largo; brdcteas ovadas a obovadas,
de 2.5-4 mm de largo, dpice agudo, aeuminado u
obtuso, ferrugfneo tomentosas, con los margenes ci-
liados; pedicelos de 3-4 mm de largo, glabros; flo-

Novon 9: 18-21. 1999.

Volume 9, Number 1
1999

Cuevas & Cochrane

Beilschmiedia manantlanensis de Mexico

19

Figura 1. Beilschmiedia manantlanensis Cuevas & Cochrane. — A. Rama y un detalle de la venacion de la lamina.
— B. Hoja. — C. Flor con bracteola, vista lateral. — D. Fruto. — E. Corte longitudinal del fruto con pericarpio delgado.
- — F. Semilla. — G. Corte transversal de la semilla. (A, B y D— G de Cuevas et al. 5083 ; C de Cuevas et al. 4076.)

res hermafroditas, copuliformes, de 3-4 mm de lar¬
go, tepalos 6, imbricados, ovado triangulares a
suborbiculares, de 1-2 mm de largo, 1-1.5 mm de
ancho, glabros o con tomento disperso en el dorso,
internamente pubescentes, apice obtuso a redondo,
margen entero a crenado, tubo receptaeular de 1.5—
3 mm de largo; androceo de 4 verticilos, de 3 es-
tambres cada uno; el primero y segundo verticilo
epitepalos, con anteras de 0.75—1 mm de largo, di-
tecas, cada teca abriendo por un poro ventral, gla-
bras, Hlamentos m4s cortos que las anteras, de 0.25
mm de largo, con pubescencia ferrugfnea esparci-
da, estambres de la tercera serie con anteras y fi-

lamentos de 0.5—0.75 mm de largo, las anteras di-
tecas, cada teca abriendo por un poro lateral,
filamentos ferrugmeo tomentosos, cada estambre de
la tercera serie con un par de glandulas en su base,
6stas de 0.5—0.75 mm de largo, bilobadas a tetra-
lobadas, glabras, la cuarta serie la lorman los es-
taminodios, los cuales son ovado triangulares, la-
minares, de 0.6—0.75 mm de largo, con el apice
acuminado a agudo, ferrugmeo tomentosos; gineceo
de 2—3 mm de largo, glabro, el ovario ovoide c6-
nieo; frutos elipsoides, de 3. 5-4. 5 cm de largo,
1.8— 2.5 cm de ancho, glabros; semilla elipsoide, de
2.3— 3.3 cm de largo, rojiza.

20

Novon

D

E

E

I fl
N
©

rv

Figura 2. Beilschmiedia manantlanensis Cuevas & Cochrane. — A. I Malle <le un estambre de la primera y segunda
serie, vista abaxial (izquierda). vista adaxial (derecha). - li. Detalle de un estambre de la tercera serie, vista abaxial
(izquierda), vista adaxial (derecha). — C. Detalle de la glandula basal (izquierda) v del hlamento y anlera (derecha) de
un estambre de la tercera serie. mostrando los poros laterales. — D. Fstaminodio de la cuarta serie. (De Cuevas el al.
4076. )

Etimologia. Esta especie recibe el nombre por
el lugar donde crece, la Sierra de Manantlan, cerca
de la costa Pacifica en el estado de Jalisco, Mexico.

Distribut ion, ecologfa y fenologia. Hasta donde
se sabe, es endemiea a la vertiente sur de la Sierra
de Manantlan, en los munieipios tie Autlan de Na¬
varro, Cuautitlan y Casimiro Castillo en el estado
de Jalisco, Mexico. Es do esperarse, sin embargo,
quo se encuentre en otras localidades de la Sierra
Madre del Sur. Florece de marzo a abril, v fructifica
en junio y julio.

Beilschmiedia manantlanensis crece en el bosque
mesofilo de montana, en altitudes de 1000 a 1900
in, en las canadas que bajan bacia Cuzalapa y Ca¬
simiro Castillo; convive con Fraxinus uhdei (Wen-
zig) Lingelsheim, Clusia salvinii J. D. Smith, Ju-
glans olanchana Standley & L. O. Williams,
Calliandra I arris Rose, Gymnanthes riparia
Klotzsch, Rondeletia amoena (Planchon) Hemsley,
Cedrela odorata L., Guarea glabra Vahl, Sideroxy-
lon portoricense Urban subsp. minutiflorum (Pittier)
T. D. Pennington, Prunus cortapico Kerber ex
Koehne, Sebastiana hintonii Lundell y Photinia
parviftora L. O. Williams. Aun cuando el bosque
mesofilo de montana, en el cual vive R. manant¬
lanensis, no oeupa una amplia extension superficial
en las partes altas de la Sierra de Manantlan, con¬
done una gran diversidad Horfstica, asf como un
buen numero de especies endemicas al occidente
de Mexico como Magnolia iltisiana Vazquez, Pe-
riptera ctenotricha Fryxell, Rondeletia manantla¬
nensis Lorenee, Satureja jaliscana MeVaugh & R.
Schmid y Trophis sp. nov., y algunas disjuntas como
Acer skutchii Rehder, Calatola laevigata Standley

y Forchhammeria sessilifolia Standley (Vazquez et
al., 1995).

Estado de conservacion, La especie aunque no
es muy abundante en los sitios que se le ha en-
contrado, no parece presentar n ingun problema de
sobrevivencia, piles se han observado individuos de
distintas edades estableciendose el bosque mesofilo
de montana. Por otro lado se ha visto bastante re-
generacion, la cual parece ser facilitada por la dis¬
persion de los frutos por aves grandes como “chon-
cho” ( Penelope purpurascens Wagler).

Usos. Hasta la fecha no contamos con infor-
macion acerca de los usos de la especie, pero por
las caracterfsticas que present an los fustes es de
esperarse que represente un importante recurso
maderable.

Relaciones taxondmicas. El genero Beilschmie¬
dia, al igual que otros de la familia Lauraceae, es
poco conocido. Se estima que tlebe tener entre 10—
20 especies en el Neotropieo (van tier Werff, 1991).
Kostermans (1938), en su revision de Lauraceas de
America, consignb 15 especies, y solo menciono a
B. mexicana (Mez) Kostermans, para Mexico; mien-
tras que Allen (1945) hizo referencia a oeho es¬
pecies para Mexico y Centro America, senalando
unicamente a B. mexicana para el pat's. Finalmen-
te, Lorea Hernandez (1995) reconocirt ocho espe¬
cies para Mexico, ademas de [tresentar una clave
para su identificacion. Aun cuando el numero de
especies de Beilschmiedia, registradas para Mexico,
ha aumentado desde 1945, la especie aqui' publi-
cada representa una clara novedad para la botAni-
ca.

Beilschmiedia manantlanensis pertenece al gru-

Volume 9, Number 1
1999

Cuevas & Cochrane

Beilschmiedia manantlanensis de Mexico

21

po de especies que presentan venacion prominente,
laxa e incompletamente reticulada en amhas su¬
perficies. Es diferente al resto de especies ueotro-
picales por tener hojas concentradas hacia al apice
de las ramillas, largos pelos en las inflorescencias
y forma y tamano de las flores. Se diferencia de B.
hondurensis Kostermans, por presentar las hojas
con los apices agudos a redondeados o obscura-
mente aeurninados, la venacion incompletamente
reticulada y los frutos mas grandes. Un mayor nu-
mero de pares de nervios primarios en las hojas e
inflorescencias mas compactas, con mayor numero
de flores y tomento ferrugfneo o glaucas cuando el
tomento cae, son caracteristicas que permiten di-
ferenciarla de B. pendula (Swartz) Hemsley (inclu-
yendo B. costaricensis (Mez & Pittier) C. K. Allen
y B. brenesii C. K. Allen; ver Burger & van der
Werff, 1990).

Eardtipos. MEXICO. Jalisco: Sierra de Manantlan,
Canada del Alentrisco, 600 m al S de Corralilos, 18(H)-
1000 m, en liosque mesofilo de montana, 10 jun. 1006.
E. I. Sanchez & R. Cuevas 112 (ZEA); Canada del Alen-
triseo, 1—2 km al S de Corralilos. I9 36'19"N,
104°18'20"W. 1000 m. en bosque mesofilo de montana, 3
mar. 1001. R. Cuevas , L. Guzmdn & E. V. Sanchez 4076
(MO, WIS, ZEA); Canada del Alentrisco, 1—2 kin al S de
Corral itos. 19°36'19"N, 104°18'20"W, 1600-1850 m, en
bosque mesofilo de montana, 27 nov. 1096, R. Cuevas, L.
Guzmdn & E. I. Sanchez 508.1 (IBUG. MO, WIS, ZEA);
Canada de La Vaca. 3 km al SW de El Zarzamoro, en
bosque mesofilo de montana, 6 sep. 1995, R. Cuevas, E.
Sanchez A- E. J. Santana 5012 (ZEA); Arroyo de La Cas-
cada, 2.5—3 km al NNF de La Pareja. 1000—1200 m, en
bosque mesofilo de montana. 15 mayo 1097, R. Cuevas,
E. V. Sanchez, L. Lopez, M. Olvera & B. Figueroa 5505
(ZEA).

Agradecimientos. Al Henk van der Werff del
Jardfn Botanico de Missouri, San Louis, por confir-
mar nuestras sospechas de que la Beilschmiedia de
la Sierra de Manantlan era una taxa indescrito y
por la revision del manuscrito; a Aaron Rodriguez
por sus eomentarios al manuscrito; al James C.
McKeown por la revision del texto en Latin; a En¬
rique V. Sanchez por la elaborat ion del dihujo, asf
como al personal de la Estacion Cientffica Las Jo-
yas y del Laboratorio de Botanica por el apoyo en
el trabajo de campo. Trabajo realizado con apoyo
de la Comision Nacional para el Conocimiento y
Uso de la Biodiversidad y la Universidad de Gua¬
dalajara.

Literatura Citada

Allen, C. K. 1945. Studies in the Lauraceae VI. Prelim¬
inary survey of the Mexican and Central American spe¬
cies. J. Arnold Arbor. 26: 280-434.

Burger, W. C. & 11. van der Werff. 1900. Family #80 Lau¬
raceae. In: W. C. Burger (editor). Flora Costaricensis.
Fieldiana, Bot. n.s. 23: 1-138.

Kostermans, A. .1. C. H. 1038. Revision of the Lauraceae
V. A monograph of the genera: Anaueria, Beilschmiedia
(American species) and Aniba. Recueil Trav. Bot. Need.
35: 834-931.

Lorea I lemdndez, F. C. 1095. A new species of Beilsch¬
miedia (Lauraceae) from Guerrero, Mexico. Novon: 5:
45-47.

Vazquez G., J. A., R. Cuevas G., T. S. Cochrane, U. H.
litis, F. J. Santana M. <X L. Guzmdn II. 1995. Flora de
Manantlan. Plantas vaseulares de la Reserva de la Bios-
fera Sierra de Manantlan, Jalisco-Colima, Mexico. Sida,
Bot. Misc. 13: 1-312.

Werff, H. van der. 1991. A key to the genera of Lauraceae
in the New World. Ann. Missouri Bot. Card. 78: 377—
387.

Studies on Geocalyeaceae (Hepaticae). XI. Supraspecific New
Taxa and New Combinations in Chiloscyphus Corda for Australasia

John J. Engel

Department of Botany, The Field Museum, Chicago, Illinois 60605-2496, U.S.A.

Abstract. Chiloscyphus sect. Novae-zeelandii,
sect. Semiteres, sect. Spiniferi, and sect. Hemispini
are described as new. Chiloscyphus subg. Connati,
subg. Microlophocolea, and subg. Cyanolophocolea
are new combinations.

The following new taxa and new combinations
are tin* result of a systematic study of the genus
Chiloscyphus Corda in Australasia. The names are
here published separately to make them immedi¬
ately available for use.

1. Chiloscyphus sect. Novae-zeelandii Engel,

sect. nov. TYPE: Jungermannia novaezeelan-
diae Lehmann & Lindenberg, in Lehmann,
Nov. Min. Cogn. Stirp. Pug. 6: 33. 1834.

I’lantae dioicae, spicis masculis elongatis, bracteis
quam foliis brevioribus, abler velut in sect. Heterophylli.

Plants dioecious; branching ol Frullania- and
lateral-intercalary types or strictly lateral-interca¬
lary; leaves with apex polymorphic, undivided and
rounded to retuse to 1-lobed to shallowly bilobed,
at times all or a combination of these variants on
1 shoot, the leal margins entire or armed with 1 or
more teeth; underleaves bifid to 0.4—0.85, the lam¬
ina margins 1—3 dentate-ciliate; androecia forming
elongate spikes, the bracts smaller than leaves; gy-
noecia on main shoots or long leafy branches, never
on short intercalary branches that lack leaves, nev¬
er on terminal branches.

The section belongs in subgenus Lophocolea
(Dumortier) Engel & Schuster and is restricted to
Australasia.

2. Chiloscyphus sect. Semiteres Engel, sect.

nov. TYPE: Jungermannia semiteres Lehmann,
Linnaea 4: 363. 1829.

Sectio Novae-zeelandii similis, gynoeciis interdum in
ramis brevibus aphyllis latero-intercalaribus vel ramis ter-
minalibus vel surculis principalibus, seta cellulis magis
numerosis constructis, atque pariele interno capsulae non
aequaliter incrassato differt.

Plants with branching terminal-lateral and lat¬
eral-intercalary or strictly lateral-intercalary; leaves

undivided, the margins entire or dentate; under¬
leaves 2— 4-lobed at least to 0.5, at times divided
nearly to the base, the lamina margins on each side
with a tooth or lobule. Gynoecia often on abbrevi¬
ated lateral-intercalary branches lacking normal
leaves, sometimes on terminal branches or main
shoots.

The section belongs in subgenus Lophocolea, and
is pan-south temperate in distribution.

3. Chiloscyphus subg. Connati (Lindenberg) En¬

gel, comb. nov. Basionym: Plagiochila sect. V.
Connatae Lindenberg, Species Hepathiearum
XXIX. 1839. Lophocolea subg. Connatae (Lin¬
denberg) Piippo, Ann. Bot. Fenn. 131: 165.
1985. TYPE: (see Piippo, 1985: 165): Chilos¬
cyphus ciliolatus (Nees) Gottsche.

4. Cliiloscyphus sect. Spiniferi Engel, sect. nov.

TYPE: Chiloscyphus spiniferus (Hooker f. &
Taylor) Engel & Schuster.

Plantae dioicae. Folia ad apieem subaeque bilobata.
Celhdae foliorum amphigastriorum atque perianthiorum
laeves sine protuberationibus, eaedem foliorum trigonis
minutis. Amphigastria admodum magna, asteroidea, (4)6—
lobata, angusle utriuque connata.

Plants dioecious; leaves subequally bilobed; leaf,
underleaf, and perianth cells smooth, without pro¬
tuberances, the leal cells with small trigones; un¬
derleaves exceedingly large, (4— )6-lobed, narrowly
connate on both sides.

The section belongs in subgenus Connati and is
restricted to New Zealand.

5. Chiloscyphus subg. Microlophocolea

(Spruce) Engel, comb, et stat nov. Basionym:
Lophocolea sect. Microlophocolea Spruce,
Trans. & Proc. Bot. Soc. Edinburgh 15: 426.
1885. Lophocolea sect. Microlophocolea
(Spruce) Schiffner, in Engler & Prantl, Natiirl.
Pflanzenfam. 1 (3): 92. 1893. Chiloscyphus
sect. Microlophocolea (Spruce) Engel & Schus¬
ter, Nova Hedwigia 39: 410. 1985 (1984).
TYPE: Lophocolea liebmanniana Gottsche
(lectotype, fide Grolle (1976)).

No VON 9; 22-24. 1999.

Volume 9, Number 1
1999

Engel

Studies on Geocalycaceae XI

23

6. Chiloscyphus sect. Heinispini Engel, sect.

nov. TYPE: Chiloscyphus parvispinus Engel.

Plantae dioeciae, rami plerumque latero-intercalares in-
terdum terminalo-laterales nunquam ventro-intercalares,
folia pagina dorsali hispida ventrali laevi, lobi amphigas-
triorum integri vel parce (1—2) dentati.

Plants dioecious; branches mostly of lateral-in¬
tercalary type, less often of Frullania- type; leaves
with dorsal surface hispid, the ventral surface of
leaf uniformly smooth; underleaf lobes entire or at
most with 1-2 teeth.

The section belongs to subgenus Microlophocolea
and is restricted to Australasia.

7. Chiloscyphus subg. Cyanolophocolea

(Schuster) Engel, comb, et stat nov. Basionym:
Lophocolea sect. Cyanolophocolea Schuster,
Hep. Anthoc. N. Amer. 4: 238. 1980. Chilos¬
cyphus sect. Cyanolophocolea (Schuster) Engel
& Schuster, Nova Hedwigia 39: 409. 1985
(1984). TYPE: Lophocolea echinella Linden-
berg & Gottsche.

Subgenera and sections of Australasian Chilos¬
cyphus may be distinguished by the following key.

Synoptic Key to Australasian Subgenera and
Sections of Chiloscyphus

la. Dorsal (and sometimes ventral) leaf surfaces
armed with conspicuous sharp-pointed laminar
and marginal processes formed of 1— several
cells; perianths armed on external (and often in¬
ternal) surfaces with spinose cellular processes;
antheridial stalk 1-seriate.

2a. Leaf insertion extended to stem midline dor-
sally; gynoecia variable in position: acrogyn-
ous or cladogynous (the cladogynous lateral-
or ventral-intercalary, often or usually bearing

vegetative leaves) . subg. Microlophocolea

3a. Ventral surface of leaf (suboptimal
plants aside) with teeth; underleaf seg¬
ments with opposing teeth or cilia, the
abaxial surface of both segments and
lamina often with scattered teeth; ven¬
tral-intercalary branches at least some¬
times present . sect. Microlophocolea

3b. Ventral surface of leaf uniformly
smooth; underleaf segments entire or at
most with 1—2 teeth, the teeth never
regularly opposing, the abaxial surface
of both segments and lamina smooth;
ventral-intercalary branches lacking . .

. sect. Hemispini

2b. Leaf insertion not extending to stem midline
dorsally and delimiting a broad leaf free gut¬
ter 6—10 cells wide; gynoecia usually on ab¬
breviated ventral-intercalary branches de¬
void of vegetative leaves .

. subg. Cyanolophocolea

lb. Leaf surfaces smooth or armed with thick-walled
tubercles; perianths smooth, never armed with
spinous processes; antheridial stalk 1- or 2-se-
riate.

4a. Aspect heteroscyphoid: subopposed leaves
consistently and distinctly connate with ad¬
jacent, large, conspicuous and broad under¬
leaves, thus each gyre hearing a single, con¬
tinuous, leaf-complex; innermost capsule
wall cells with semiannular hands often fur¬
cate and anastomosing to delimit fenestrae;
antheridial stalks biseriate .... subg. Connati
5a. Leaf, underleaf, and perianth cells armed
on both surfaces with prominent, central,
dome-like, thick-walled tuberculae one

each per cell . sect. Leucophylli

5b. Leaf and underleaf cells smooth, with¬
out protuberances of any sort.

6a. Underleaves 2-4-lobed or -lobu-
late, very broadly connate with lat¬
eral leaves, without conspicuous
basal teeth; leaf cells with nodose

trigones . sect. Connati

6b. Underleaves (4— )6-lobed and with
basal accessory teeth, narrowly
connate with lateral leaves; leaf

cells with small trigones .

. sect. Spiniferi

4b. Aspect chiloscyphoid: leaves usually clearly
alternate, never conspicuously united with
underleaves to form solitary' ring-like units,
the underleaves free or narrowly connate
with leaves on one side (or if consistently
narrowly connate on both sides, then con-
nation obscure); innermost capsule wall
cells with complete to incomplete to spur¬
like tangential thickenings that fail to fork
or fork only sporadically and do not delimit
fenestrae; antheridial stalks 1-seriate.

7a. Underleaves undivided and entire or at
most bifid to 0.3, the lamina margins at
most with 1 tooth on each side; ventral-
intercalary branching common; leaves
oriented subtransverse to subsuccu-
bous, very rarely strongly succubous,
the insertion strongly recurved at ven¬
tral end, distinctly inverted J-shaped

. subg. Notholophocolea

7b. Underleaves bifid to 0.4 or more, the
lamina margins often with 2—3 teeth or
a laciniiform to lobuliform process on
each side; ventral-intercalary branching
rare or lacking; leaf orientation strongly
succubous, the insertion not or hardly
recurved at ventral end, at most weakly

inverted J-shaped . subg. Lophocolea

8a. Leaves fundamentally undivided,
the apices on a single shoot at
times variable and then rounded to
1- to 2-lobed, but never consistent¬
ly bilobed.

9a. Gynoecial position variable,
some to many on short, lateral-
intercalary branches lacking
normal leaves (others on lead¬
ing leafy shoots), at times on
terminal branches; leaf apices

24

Novon

broadly rounded to retuse or
occasionally short bifid, not
polymorphic, the shoots with¬
out a combination of undivid¬
ed, I- and 2-lobed leaves;
perianth wings sporadic or
lacking; inner layer of capsule
wall cells with radial walls de¬
void of continuous sheets of
pigmented thickening; seta 8—

1 1 cells diam., w ith 26-30 epi¬
dermal cell rows; gemmae

lacking . sect. Semiteres

9b. Gynoecia on main shoots or
long leafy branches, never on
short intercalary branches that
lack vegetative leaves, never on
terminal branches; leaf apices
often polymorphic, the shoots
with a combination of undivid¬
ed, I - and 2-lobed leaves; peri¬

anth wings rather common; in¬
ner layer of capsule wall cells
with continuous sheets of pig¬
mented thickening on vertical
radial walls; seta (where known)

6-8 cells diam., with 17-21
epidermal cell rows; gemmae
frequent . . . sect. Novae-zeelandii
8b. Leaves fundamentally bilobed . . .

. sect. Lophocolea

Acknowledgments. I thank Patricia Eckel and

Gary Merrill for providing the Latin diagnoses.

l iterature Cited

Grolle, R. 1976. Verzeichnis der Lebermoose Europas und
benachbarter Gebiete. Feddes Repert. 87: 171—279.

Piippo, S. 1985. Bryophyte flora of the Huon Peninsula,
Papua Newr Guinea. XII. Geocalycaceae (Hepaticae).
Ann. Rot. Fenn. 131: 129—167, f. 1—19.

Austral Hepaticae. 26. The Identity, Taxonomic Position, and
Ecology of Trichocolea julacea Hatcher (Trichocoleaceae)

John J. Engel

Department of Botany, The Field Museum, Chicago, Illinois 60605-2496, U.S.A.

ABSTRACT. A new species, Leiomitra julacea, is
proposed for the New Zealand Trichocolea julacea,
which is invalid according to the International Code
of Botanical Nomenclature. The species is de¬
scribed and illustrated.

The genus Trichocolea Dumortier is well repre¬
sented in tropical montane areas and to a lesser
extent the south temperate zone. Only a single spe¬
cies, T. tomentella (Ehrhart) Dumortier, occurs in
temperate areas in the Northern Hemisphere, where
it is disjunctly widespread in range. Trichocolea to¬
mentella is the type of the genus.

Gottsche (1864: 132) subdivided Trichocolea into
two sections, sect. Hirtiflora Gottsche (= sect. Tri¬
chocolea) and sect. Laeviflora Gottsche. He char¬
acterized section Hirtiflora as having a hairy “in-
volucrum,” i.e., the calyptra, which is fused with
and surrounded by bracts and bracteoles. Three
species were assigned to this section: T. tomentella,
T. mollissima (Hooker f. & Taylor) Gottsche, and T.
lanata (Hooker) Nees. The other section, Laeviflora,
was defined as having a smooth calyptra, devoid of
bracts, and contained one species, T. tomentosa
(Swartz) Gottsche, which is therefore the type of the
section. Lindberg (1875) created a new genus,
Leiomitra Lindberg, which included T. tomentosa
and a new species, L. capillata Lindberg. He dis¬
tinguished the genus from Trichocolea in part by
the lack of regularly pinnate branching, and on the
characters cited by Gottsche for Trichocolea sect.
Laeviflora. However, Gottsches section was not cit¬
ed as a synonym.

Leiomitra has been variously treated by subse¬
quent authors. Spruce (1884—1885) recognized
Leiomitra as a genus and included four neotropical
species: L. tomentosa (Swartz) Lindberg, L. flaccida
Spruce, L. sphagnoides Spruce, and L. paraphyllina
Spruce. Stephani (1888) and Bescherelle (1893)
also accepted the genus. Schiffner (1893—1895)
treated Trichocolea as a genus and recognized
Gottsche’s sections as subgenera; Leiomitra Lind¬
berg was cited as a synonym of subgenus Laeviflora
(Gottsche) Schiffner. Schiffner’s description of sub¬
genus Hirtiflora (Gottsche) Schiffner emphasized

succubously oriented leaves, repeatedly pinnate
branching, and the presence of a fleshy, wooly “ca¬
lyptra” having both bracts and sterile archegonia
inserted on it; the subgenus included six species.
By contrast, subgenus Laeviflora had incubously
oriented leaves, less regularly pinnate branching,
and a “calyptra” that was both thin and smooth
throughout, or at least in the upper portion. Schiff¬
ner included seven primarily neotropical species in
subgenus Laeviflora, including T. tomentosa. Spruce
(1895) was the first to treat Leiomitra as a subge¬
nus, subg. Leiomitra (Lindberg) Spruce. Subgenus
Laeviflora (Gottsche) Schiffner has less than a
three-week priority over subgenus Leiomitra (Lind¬
berg) Spruce (15 January vs. 5 February fide Sta-
fleu & Cowan, 1985).

Leiomitra was included as a synonym of Tricho¬
colea, without subgeneric status by, e.g., Stephani
(1898—1924), who regarded the two genera as the
same on developmental grounds; Verdoom (1932);
Evans (1939); Muller (1951-1958); Hatcher
(1957); Fulford (1963); Gradstein (1989); Fulford
and Sharp (1990); Gradstein and Florschiitz-de
Waard (1990); and Schuster (1966). Little (1949:
10) lectotypified Leiomitra Lindberg with L. tomen¬
tosa (while including the genus in the synonymy of
Trichocolea ). Leiomitra was recognized as a distinct
genus by Schuster (1980, 1984) and by Grolle
(1983). Schuster and Grolle placed both Trichocolea
and Leiomitra in the Trichocoleaceae.

Schuster (1963) added a third genus to the Tri¬
chocoleaceae, Eotrichocolea Schuster, based on Tri¬
chocolea polyacantha (Hooker f. & Taylor) Gottsche
et al. of New Zealand. Schuster (1980) discussed
interconnections of the Trichocoleaceae with a
group of genera that includes Temnoma in the
Pseudolepicoleaceae Fulford & J. Taylor, and
merged the two families into one family, Trichoco¬
leaceae, with four subfamilies: Trichocoleoideae,
Temnomoideae Schuster, Blepharostomatoideae
Grolle, and Chaetocoleoideae Schuster (Schuster,
1980, 1984).

Hatcher (1958) described Trichocolea julacea
Hatcher, based on a plant from Stewart Island, New
Zealand. Known only from the protologue, this spe-

Novon 9: 25-28. 1999.

26

Novon

cies has remained somewhat of a mystery plant.
Examination of the type, and material recently col¬
lected in New Zealand, reveals that the plant pos¬
sesses characters that place it in Leiomitra, and a
new combination is therefore required, as follows:

Leiomitra julacea Hatcher ex Engel, sp. nov. Tri-
chocolea julacea Hatcher, Trans. Roy. Soc.
New Zealand 85: 245. f. 30-36. 1958, nom.
inval. (Art. 37). TYPE: New Zealand. Stewart
Is.: Port Pegasus, Sawmiller’s Arm, 1949, Mar¬
tin 621 (holotype, F; isotype, TWM). Figure 1.

Plants spongy, distinctly julaceous, with all sides
appearing as a mass of interwoven cilia, loosely
creeping, chocolate brown (the shoot tips light
green), the shoots to 2 mm wide. Branching irreg¬
ularly sympodial to pseudodichotomous, the lead¬
ing axis soon losing its dominance, the main shoot
and branches alike in vigor. Stem lacking para-
phyllia, hidden on all sides by interwoven leaf cilia.
Rhizoids occasional, in tight bundles, from stem at
immediate base of underleaves, the tips often
branched. Leaves widely spreading but with lobes
arching toward shoot apex, rather closely imbricate,
distinctly succubously (almost longitudinally) in¬
serted, the lamina twisted: ventral sector ol leaf ori¬
ented parallel with substrate, the dorsal sector of
leal antically assurgent and the dorsal margin seen
on edge, the leaves obtrapezoidal to subreniform,
asymmetrically 4— 6-lobed, the ventral pair of lobes
largest, the dorsal lobe smallest. Lobes caudate, not
3-fid, the ventral pair of lobes 4-5 cells wide at
base (tier immediately above sinus base), then 2—
3 tiers of laterally juxtaposed cells basal to the uni-
seriate row of 6—10 cells; cells ol uniseriate row ±
cylindrical, with dilated septa, 13—25 /am wide,
62-98 /am long, the cells toward base of uniseriate
row rather thick-walled, the cells becoming pro¬
gressively smaller and less thick-walled toward
lobe apices, the cuticle of penultimate and terminal
cells striolate-papillose, that of the remaining cells
finely striolate; terminal cell of uniseriate row
somewhat tapered, 10-12 /am wide, 68—90 /am
long, slightly thick-walled in the tip; margins of
lobes copiously armed with pairs of opposing cilia,
the cilia of lobes and sinus bases repeatedly di-
chotomously branched, the branches of each di¬
chotomy stiffly diverging, but the armature primar¬
ily directed adaxially and oriented at right angles
to the leal plane, the leaf thus appearing to have a
stiffly ciliate adaxial leaf surface, the cilia ulti¬
mately forming a densely interwoven dendroid
crown to the leal; marginal cilia of lobe uniformly
uniseriate throughout, the cells elongate, rather

thick-walled, the cuticle of cells toward base of cil-
ium striolate, the cuticle of cells toward and in¬
cluding terminal cell striolate-papillose; sinus ba¬
ses plane and not reflexed. Disc slightly convex,
asymmetric, 3-4 cells high at dorsal sinus, 5-6
cells high at ventral sinus (from base to sinuses);
margins of disc with dichotomously branched cilia
similar to those of lobes, the cilia often adaxially
displaced and lying at right angles to the disc
plane. Underleaves somewhat smaller than leaves,
narrowly connate on both sides, cuneate, quadrifid,
the marginal armature similar to leaves and like¬
wise forming a mass of interwoven cilia; disc 3-4
cells high. Cells of disc thin-walled, strongly elon¬
gated, aligned in somewhat irregular tiers, in lam¬
ina middle 22—26 /am wide X 84—108 /am long;
cuticle markedly long striolate, with only a few pa¬
pillae. Asexual reproduction lacking. Otherwise un¬
known.

Nomenclature. The name Trichocolea julacea is
invalid according to the ICBN, Art. 37 (Greuter et
al., 1994), which requires that after 1 January 1958
a holotype for new species must be indicated.
Hatcher (1958) made no reference to a nomenela-
tural type for the species (see also Art. 7), but stat¬
ed that “portions ol the original collection,” gath¬
ered by Martin, are in four herbaria, including that
of Martin followed by the number 621 in parenthe¬
ses. The specimen retained by Hatcher (now at F)
bears the locality that matches the protologue, but
also includes the number 621 and “TYPE” on the
label. That specimen is herein designated the ho¬
lotype. The above description and accompanying
illustration (Fig. 1) are based on this collection.

Distribution and ecology. Known only from the
type collection and the Cascade ultramafic moraine
in southern Westland Province. There the species
occurs at ca. 135 m in an area of ultramafic rocks
and outcrops with rather open vegetation consisting
mainly of Gleichenia, Lycopodium, Juncus, the li¬
chen Cladina, and scattered Leptospermum. The
species forms thick, nearly pure mats on well-
drained soil of slopes and vertical banks of drain¬
age channels. Hatcher (1958) mentioned that the
type was “epiphytic; on forest trees.”

Additional specimens examined. NEW ZEALAND.
South Island: Westland Prov., Cascade Road, Cascade
ultramafic moraine, W of Martyr Saddle, SSW of Jackson
Bay, ca. 135 m, Engel 21776. 23001 (F).

Taxonomy. Only juvenile gynoecia are known
for the species (the type), but the irregularly sym¬
podial to pseudodichotomous branching pattern
(with the leading axis soon losing its dominance)
and the distinctly succubously to nearly longitudi-

Volume 9, Number 1
1999

Engel

Austral Hepaticae. 26

27

Figure 1. I^iomitra julacea Hatcher ex Engel. — 1. Ventral sector of leaf, adaxial aspect (some portions of marginal
armature not shown for clarity). — 2. Shoot, dorsal view. — 3. Median cells of lamina. — 4. Leaf, adaxial view. — 5.
Leaf, abaxial view. — 6. Cross section of shoot showing mass of interwoven, branched cilia (UL = underleal) of 1 gyre.
— 7. Underleaf (some portions of marginal armature not shown for clarity). — 8. Leaf lobe, distal sector. — 9. Lateral
armature of leaf lobe, distal sector. (All from holotype.)

28

Novon

nally inserted leaves are characters that place this
species in Leiomitra. Trichocolea s. str. has regu¬
larly 2— 3(4)-pinnate branching, with the main axis
retaining its dominance and the leaf insertion vary¬
ing from weakly succubous to weakly ineubous.

The leaves of this species are remarkable for
their development of a densely interwoven “basket-
work" of cilia (Fig. 1: 4—6), which causes the axis
and branches to appear spongy and julaceous (Fig.
1: 2); the cilia are repeatedly dichotomously
branched and widely divergent (Fig. 1: 1), the ul¬
timate divisions of each leaf forming a densely in¬
terwoven dendroid “crown.” Living plants, as well
as the type, are brownish. Branching is irregularly
sympodial to pseudodichotomous, as in Leiomitra
lanata (Hooker) Schuster, but the form of the leaves
will immediately distinguish this species. The spe¬
cies is readily distinguished in the field by the ju¬
laceous, wiry aspect ol the plants coupled with the
brownish color.

Leiomitra julacea and L. lanata, also of New Zea¬
land, are the oidy south temperate members of the
genus; all other species of Trichocolea reported
from the south temperate belong to that genus (s.
str). The two New Zealand species of 'Leiomitra may
be distinguished by the following key.

kt \ to Nt-.vt Zkai.am) Sri i i i s or Leiomitra

la. Leaf lobes equally 3-fid. consisting of an adaxi-

ally divergent segment and a pair of abaxially
divergent segments, each segment usually 1(2)
times again 3-fid, ending in a pair of ± equally
divergent, uniseriate cilia; leal disc 6—10 cells
high; plants grass green to yellowish green; com¬
mon. North and South Islands . L. lanata

lb. Leaf lobes caudate, not 3- fid, the margins armed

with pairs of opposing cilia, the cilia repeatedly
dichotomously branched and forming a densely
interwoven dendroid crown; leaf disc 5— 6 cells
high; plants brown; rare, Stewart Is. and southern
Westland Prov . I,, julacea

Acknowledgments. I thank Marlene Donnelly
lor preparing the habit figures for tbe illustration,
and Gary Merrill for valued advice on the manu¬
script.

I .ilerature Cited

Beseherelle, E. 1893. Enumeration lies hepatiques con¬
iines jusqu’iei aux Antilles franyaises (Guadeloupe et

Martinique). J. Bol. (Moral) 7; 174-180, 183-194.
Evans, A. W. 1939. The classification of the Hepatieae.

Bot. Rev. (Lancaster) 5: 49-90.

Kul ford, M. 1963. Manual of the Leafy Hepatieae of Latin
America. Part I. Mem. New York Bot. Gard. I 1: I — I 72.
pi. 1-37 (unnumbered).

- & A. .1 Sharp. 1990. The leafy Hepatieae of Mex¬

ico: One hundred and twenty-seven years after C. M.
Gottsche. Mem. New York Bot. Gard. 63: 1— 80.

Gottsehe, C. M. 1864. Hepatieae. In: Triana & Planchon,
Prodromes florae novo-granatensis. Ann. Sri. Nat. Bot.,
Ser. 5, 1: 95-198, pi. 17-20.

Gradslein, S. R. 1989. A key to the Hepatieae and An-
thocerotae of Puerto Rico and tbe Virgin Islands. Bry-
ologist 92: 329-348.

- X J. Horschtitz de Waard. 1990. Results of a bo¬
tanical expedition to Mount Roraima, Guyana. I. Bryo-
phytes. Tropical Bryology 1: 25—54.

G renter, W., F. R. Barrie, H. M. Runlet. W. G. Chaloner,
V. Demoulin. 1). L. Hawksworth, P. M. Jprgensen. I). 11.
Nicolson. P. C. Silva. P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Grolle. R. 1983. Nomina generica Hepatirarum: Refer¬
ences. types and synonymies. Acta Bot. Fenn. 121: I —
62.

Hatcher, R. E. 1957. The genus Trichocolea in North.
Central and South America (Hepatieae). Lloydia 20:
139-185.

- . 1958. The genus Trichocolea in New Zealand.

Trans. Roy. Soc. New Zealand 85: 237-246, figs. 1-3.

Lindberg, S. 0. 1875. Hepatieae in Hibernia mense Jnlii
1873 lectae. Acta Soc. Sci. Fenn. 10: 467-559.

Little, E. L. 1949. Nomina conservanda proposals in lle-
paticae. Bryologist 52: 1-22.

Midler, k. 1951 — 1958. Die Lebermoose Europas. In: Ra-
benhorst’s Kryptogamen-Klora, ed. 3. 6. pp. 1-1365.
figs. 1-514.

Schiffner. V. 1893-1895. Hepatieae (Lebermoose). Pp. 3-
141. f. 1—73 in: A. Engler & K. Prantl (editors). Die
natilrlii hen Pflanzenfamilien, Ted I (Abt. 3. I llalfte):
1 (3). Leipzig. | Pp. 1-96, 10 Oct. 1893; pp. 97-144,
15 Jan. 1895.]

Schuster, R. M. 1963. Studies on antipodal Hepatieae. I.
Annotated keys to the genera of antipodal Hepatieae
with special reference to New Zealand and Tasmania.
J. Hattori Bot. Lab. 26: 185-309.

- . 1966. The Hepatieae and Anthocerotae of North

America east of the hundredth meridian. Vol. I. Colum¬
bia Univ. Press, New York.

- . 1980. New combinations and taxa of Hepatieae,

I. Phytologia 45: 415—437.

- . 1984. Evolution, phvlogeny and classification of

the Hepatieae. Pp. 892—1070. f. 36-100 in: R. M.
Schuster (editor). New Manual of Bryology, Vol. 2. Hat-
tori Botanical Laboratory. Nichinan.

Spruce, R. 1884—1885. Hepatieae Amazonicae et Andi-
nae. Trans. & Proe. Bot. Soc. Edinburgh 15: 1—308, pi.
1-4. 1884 (part I); i— xi. 309-589. pi. 5—22. 1885 (part
II).

- . 1895. Hepatieae elliottianae, insulis Antillanis

Sti Vincentii et Dominica a elar. W. R. Elliott, annis
1891—92, lectae, Ricardo Spruce determinatae. J. Linn.
Soc., Bot. 30: 331-372. pi. 20-30.

Stafleu, F A. & R. S. Cowan. 1985. Taxonomic Literature.
A selective guide to botanical publications and collec¬
tions with dates, commentaries and types. Vol. Ill: Sal-
Ste. Regnum Veg. 112: 1-1066.

Stephani, K 1888. Westindisehe Hepatieae. Hedwigia 27:
276—299.

- . 1898-1924. Species Hepaticamm. Geneve &

Bale, 6 vols.

Verdoorn, F. 1932. Classification of hepaties. Pp. 413-432
in F. Verdoorn (editor). Manual of Bryology. The Hague.

Austral Hepaticae. 28. Plagiochila bazzanioides Engel & Merrill, a
Remarkable New Species of Plagiochilaceae from New Zealand

John J. Engel and G. L. Smith Merrill

Department of Botany, The Field Museum, Chicago, Illinois 60605-2496, U.S.A.

ABSTRACT. Plagiochila bazzanioides, a new spe¬
cies from New Zealand, is described and illustrat¬
ed.

Plagiochila is the largest genus of hepatics, with
perhaps as many as 1600 described species (Bon¬
ner, 1962). As in other species-rich groups, the ac¬
tual number ot species will probably prove to be
considerably less. According to Schuster (1980:
348), “The total number ol valid species ot Plagi¬
ochila may prove to be below 500.” The genus is
widely distributed and essentially cosmopolitan.
Centers ot taxic diversity are chiefly in the Neo¬
tropics and southeast Asia, and the genus is rather
poorly represented in temperate areas. For exam¬
ple, Schuster (1980) included 20 species for North
America, and Inoue and Schuster (1971) included
25 species from temperate Australasia. This paper
describes an additional distinctive species occur¬
ring in New Zealand, as follows.

Plafjimdula bazzanioides Engel & Merrill, sp.
nov. TYPE: New Zealand. North Island: North
Auckland Province, Rangitoto Island, imme¬
diately E of Auckland, 50—160 m, 4 Feb.
1995, Engel 20769 (holotype, F; isotype,
CHR). Figure 1.

Species insignis foliis falcatis canalioulatis decurvatis,
marginis foliorum peelinatis dentibus gracilibus uncinatis,
ramis plurimis flagelliformibus, androeciis interealaribus,
paraphylliis caulis dentatis lamelliformibus a species nob¬
is notis optime distincta.

Plants bazzanioid in habit, ascending, without
differentiation ot a horizontal leafless axis, rich,
deep olive green; leafy shoots medium, to 5 mm
wide. Stems with paraphyllia inconspicuous, typi¬
cally consisting of a single, short, sharply toothed
lamella adjacent to dorsal leaf base; stem in cross
section with cortical cells in (2)3 layers of very
thick-walled cells, the medullary cell walls thin,
colorless. Branching exclusively lateral-intercalary,
sometimes with geotropic, microphyllous, rhizoi-
dous, flagelliform, lateral-intercalary branches.
Rhizoids rather dense on leafy shoots. Leaves hor¬
izontally spreading to ventrally secund, often con¬

spicuously falcate, spreading at ± right angles to
stem, closely imbricate, frequently canaliculate-
subtubular to the tips, the dorsal margins narrowly
recurved and forming a distinct convex fold, lend¬
ing the shoot a tightly pleated aspect in dorsal view,
in ventral aspect the ventral margin aligned at right
angles to the stem except for the basal sector, which
lies parallel to the stem, in ventral view the ventral
margins pectinate-toothed, the teeth often stiffly re¬
flexed and hooked dorsally, the ventral surface of
stem moderately or totally hidden; leaf insertion
elongated, distinctly straight and steeply oblique
for most of its length, abruptly recurved at ventral
end; leaves asymmetrically narrowly elongate-trap-
ezoidal, tapering to a markedly narrow, almost par¬
allel-sided distal 0.2 to 0.3; leaf apex narrowly
rounded to truncate, tending to become bidentate
by 2 conspicuous, often diverging apical teeth that
are often larger than adjacent teeth; dorsal margin
nearly straight to broadly curved inward, with ±
regularly spaced, slender, recurved teeth to the
base (best seen in ventral aspect); ventral margin
± straight from a subauriculate base, conspicuous¬
ly ciliate-dentate particularly in basal portion, the
teeth stiffly reflexed (hooked dorsally), the margin
often entire or subentire in the distal 0.5, decur¬
rent, the decurrent strip dentate. Marginal teeth
slenderly acuminate, 1—2 cells wide at the base,
consisting almost entirely of a uniseriate row of 4—
6 distinctly thick-walled cells, the terminal cell
elongate, to 4.2:1. Underleaves consisting of an
amorphous clump of cells, in part short-filamen¬
tous, at end of decurrent ventral leaf base. Cells in
median portion of leaf thin-walled and with trigones
bulging to knotlike; median cells 20—29 /am wide
X 30—36 /am long; cells of leal bases scarcely elon¬
gated and not forming a basal field; cuticle smooth.
Asexual reproduction lacking.

6 Plants ± smaller than the 9 ones, mostly sim¬
ple; androecia on main shoot and leading lateral-
intercalary branches, becoming intercalary in posi¬
tion (often with several successive androecia
occurring on a stem or branch); bracts many, in up
to 15 pairs, strongly saccate in dorsal half of base;
apex broadly rounded, spatulate, entire; ventral mar-

Novon 9: 29-31. 1999.

200 n

30

Novon

200 n

Volume 9, Number 1
1999

Engel & Smith Merrill
Austral Hepaticae. 28

31

gin reflexed, entire or with a few spinose teeth near
base; lobule large, the free margin very short, in-
flexed, slightly to moderately dilated, with a few mi¬
nute teeth; antheridia 1 per bract, the stalk 2-seriate.
Gynoecia at apices of main or leading shoots, usually
with 1—2 innovations in absence of fertilization, the
gynoecia with irregular, ciliiform-margined para-
phyllia in dorsal sector of axil of innermost bracts;
bracts of innermost series similar to leaves except
somewhat larger and broader, narrowly ovate, with
margins armed as in leaves. Perianth slightly inflated
in basal sector but strongly bilaterally compressed
above, the dorsal and ventral keels wingless; mouth
truncate, with crowded, irregular, spinose teeth and
cilia, the teeth 2-3 cells wide at the base and with
a uniseriate row of up to 8 cells.

Sporophyte not seen.

Taxonomy. The dorsal surface of the stem in
Plagiochila bazzanioides bears sharply toothed, la-
melliform paraphyllia (Fig. 1: 1) as in FI circum-
dentata Stephani, and is assigned to Plagiochila
sect. Annotinae Carl, together with that species, P.
annotina Menzies ex Lindenberg, and P. baylisii
Inoue & Schuster, all of New Zealand (see Inoue
& Schuster, 1971). The latter two species, however,
lack stem paraphyllia. Plagiochila circumdentata
has broadly ovate to deltoid-ovate leaves with a
rounded, non-canaliculate apex and a strongly am-
pliate ventral margin. This species also differs in
the position of androecia, often occurring on clus¬
ters of terminal, Frullania- type branches, whereas
in P. bazzanioides the androecia are confined to the
main shoot and leading lateral-intercalary branch¬
es, often becoming intercalary in position (Fig. 1:
13), with several successive androecia produced on
a single shoot or branch.

The marginal teeth of the leaves of Plagiochila
bazzanioides are unique among New Zealand spe¬
cies of the genus in being slender, stiffly hooked.

and claw-like (Fig. 1: 1-5), consisting almost en¬
tirely of a uniseriate row of up to 6 cells in length
(Fig. 1: 7). The distinctly pectinate dorsal leaf mar¬
gins resemble the teeth of a garden rake, and are
apparent even under the dissecting microscope.

Plagiochila bazzanioides is notable for the re¬
markable Bazzania- like aspect of the plants, a com¬
bination of the ventrally decurved, falcate, narrowly
channeled leaves with a truncate, few-toothed apex,
and the frequent production of geotropic, flagelli-
form branches up to 2.5 cm long. Otherwise, in leaf
shape and orientation P. bazzanioides resembles a
slender P. obscura Colenso: both species have more
or less attenuate, narrowly trapezoidal leaves. As in
P obscura, the leaf insertion is straight and steeply
oblique for most of its length (Fig. 1: 1), and the
ventral margin is straight and not ampliate, except
for the short, truncate portion near the base, which
in situ lies parallel with the stem (Fig. 1: 2). Also,
as in P. obscura, the distal portion of the ventral
margin is often entire, with the basal portion con¬
spicuously armed with regularly spaced, slender,
hooked teeth (Fig. 1: 2).

Distribution and ecology. Known only from
Rangitoto Island (just east of Auckland) in forest
primarily of Metrosideros that has developed over a
lava flow. The species occurs at 50—160 m on thick
soil over shaded rock, as well as festooning the
sides of very protected, moist niches.

Paratype. NEW ZEALAND. North Island: same lo¬
cality as type, Engel 20756— c. 6 (F).

Literature Cited

Bonner, C. E. B. 1962. Index Hepaticarum. Pars 1. J. Cra¬
mer, Germany.

Inoue, H. & R. M. Schuster. 1971. A monograph of the
New Zealand and Tasmanian Plagiochilaceae. J. Hattori
Bot. Lab. 34: 1-225, f. 1-77.

Schuster, R. M. 1980. The Hepaticae and Anthocerotae
of North America East of the Hundredth Meridian, Vol.
4. Columbia Univ. Press, New York.

Figure 1. Plagiochila bazzanioides Engel & Merrill. —1. Sector of main shoot, dorsal view (note stem paraphyllia).
— 2. Sector of main shoot, ventral view (note rudimentary underleaves). — 3—5. Leaves drawn to same scale (3 = dorsal
aspect; 4, 5, ventral aspect). — 6. Teeth of dorsal margin of leaf. — 7. Teeth of ventral margin of leaf. — 8. Leaf apex.
- — 9. Median leaf cells. — 10. Innermost 9 bract (ventral aspect). — 11. Perianth. —12. Portion of perianth mouth. —
13. Sector of shoot with androecium. (Figs. 1-12, from holotype; Fig. 13, from Engel 20756, New Zealand, North
Auckland Prov., Rangitoto Is., immediately east of Auckland.)

Taxonomic Notes on Neotropical Maprounea Aublet
(Euphorbiaceae)

Hans- Joachim Esser

Herbarium Division, University of Utrecht, Postbus 80102, 3508 TC Utrecht, Netherlands.
Current address: Institut fur Allgenieine Botanik, Universitiit Hamburg,
Ohnhorststrasse 18, D-22609 Hamburg, Germany

Abstract. A key to and some remarks on Ma¬
prounea in the Neotropics are presented. Maprou¬
nea guianensis var. obtusata is discussed and raised
to specific rank with the new name Maprounea
amazonica. Maprounea brasiliensis is accepted as
distinct from M. guianensis. The taxa are separated
mainly by characters of the leaves and by fruit size.
Additionally, Maprounia glauca is established as a
new synonym of Mabea taquari, and a lectotype is
designated for the latter.

Maprounea Aublet (Euphorbiaceae, Hippoma-
neae) is a small genus of shrubs and trees occurring
in the Neotropics and in Africa. Although the South
American taxa have been treated twice quite re¬
cently, their taxonomy is still not fully resolved.

Pax and Hoffmann (1912) accepted two neotrop¬
ical species, Maprounea guianensis Aublet and M.
brasiliensis A. Saint-Hilaire, both without infraspe¬
cific taxa. Allem (1976), however, considered these
two species to be identical and united them under
the older name, M. guianensis. Senna (1984) dis¬
agreed with him and accepted both species as well
as a number of varieties of M. guianensis, with the
exception of M. guianensis var. undulata Muller Ar-
goviensis, which she did not consider to be distinct
from the typical variety.

Allem’s (1976) study was based mainly on plants
from central Brazil and concentrated on floral char¬
acters, neglecting the varieties described from other
regions. The revision of Senna (1984) was uncriti¬
cal, although it included very detailed illustrations
of all taxa. From these illustrations it is quite ob¬
vious that the leaves of M. guianensis var. obtusata
(Muller Argoviensis) Muller Argoviensis are very
different from those of all other neotropical taxa of
the genus.

On further examination of leaf and fruit charac¬
ters in particular I now conclude that this variety
represents a distinct species, and also that M. bras¬
iliensis should be kept separate from M. guianensis.

Key to the Neotropical Species oe Marrovnea

la. Leaves apically obtuse to rounded to emarginate,

abaxially eglandular at base (Fig. 1 A); 0-1 pis¬
tillate flower per inflorescence; staminate inflo¬
rescence at least as wide as long; fruits 8—10 mm
long . M. amazonica

lb. Leaves apically acute to mucronate to acuminate,
rarely obtuse, abaxially often with basal glands,
rarely without; (0— )2— 3 pistillate flowers per in¬
florescence; staminate inflorescence longer than
wide; fruits 4—10 mm long.

2a. Tree up to 25 m tall; leaves distinctly longer
than wide, apically acute to acuminate,
abaxially with 0-l(-2) pair(s) of base glands
(Fig. IF), often with additional submarginal
glands in upper half; fruits 4— 6(— 8) mm long

. M. guianensis

2b. Shrub up to 1.5 m tall; leaves hardly longer
than wide, apically mucronate, rarely ob¬
tuse, abaxially usually with 2-4 pairs of bas¬
al glands (Fig. 1G), eglandular in upper half;
fruits 8-10 mm long . M. brasiliensis

Maprounea amazonica Esser, nom. et stat. nov.
Replaced name: Maprounea guianensis Aublet
var. guianensis forma obtusata Muller Argo¬
viensis, Linnaea 32: 115. 1863. Maprounea
guianensis Aublet var. obtusata (Muller Argo¬
viensis) Muller Argoviensis, in Martius, FI.
Bras. 11.2: 543. 1874. TYPE: Brazil. Amazon¬
as: Barra (= Manaus), igapo of an igarape.
May 1851, Spruce 1475 (lectotype, selected
here, G-DC not seen, microfiche; isolectotypes,
BM, BR, E, GH. K, NY, OXF, P, TCD). Figure
1 A-E.

The name Maprounea obtusa Pax already exists
for an African species. A new combination based
on the neotropical variety obtusata would be very
similar to the African name, and because of the risk
of confusion a new name, M. amazonica, was cho¬
sen according to Article 53.3 of the International
Code of Botanical Nomenclature (Greuter et al.,
1994).

Shrub or tree up to 12 m tall and 15 cm stem
diam. Bark smooth. Totally glabrous. Monoecious.
Stipules broadly elliptic, ca. 0. 7-0.9 X 0.5-0.9

Novon 9: 32-35. 1999.

Volume 9, Number 1
1999

Esser

Neotropical Maprounea

33

Figure 1. A— E. Maprounea amazonica Esser. — A. Leal, showing glands on lower surface. — If. Pistillate flower. —
C. Seed with large caruncle and foveolate surface. — If. Mericarp of fruit with irregularly caducous septum and single
apical vascular strand. — E. Remaining columella of fruit without septal remnants. — F. Maprounea guianensis Aublet.
Leaf, showing glands on lower surface. — G. Maprounea brasiliensis A. Saint-Hilaire. Leaf, showing glands on lower
surface. Scale bar = 5 mm (A, E— G) or 2 mm (R— D). A, R drawn from Prance et al. 11752 (L); C— E from Prance et
al. 16174 (U); F from Spruce 2271 (TCD); G from Hatschbach 26979 (HBG).

mm, entire, eglandular; petioles 0.9— 1.8 cm long;
leaf blades membranous to chartaceous, ovate to
orbicular, 2.5—5 X 1.5— 3.5 cm, apex obtuse to
rounded to emarginate, margin entire, base slightly
emarginate to rounded to slightly obtuse, distinctly
whitish below and with (0— )1— 6 glands on each side
of midvein, glands 0.2— 0.4 mm diam., submarginal,
often associated with secondary veins, strictly
eglandular at base, secondary veins in 8—13 pairs
and irregularly brochidodromous. Inflorescences
greenish to yellowish (staminate flowers sometimes
reddish), usually on short lateral branches, con¬
sisting of a peduncle 5-8 mm long and a head of
staminate flowers 2-2.5 mm long and 2-3 mm
diam. and 0—1 pistillate flower inserted near base

of peduncle. Staminate flowers nearly sessile; calyx
ca. 0.75 mm long; stamens 2, filaments up to 1 mm
long, partly fused, anthers ca. 0.2 mm long. Pistil¬
late flowers: pedicel 2—9 mm long; calyx with 3 free
sepals, when flowering nearly as long as ovary; style
nearly absent, stigma ca. 1.5 mm long. Fruits
greenish, pedicel 7-17 mm long; schizocarp glob¬
ular and slightly sulcate, smooth, dry, opening first
loculicidally and soon afterward septicidally; mer-
icarps 8-10 mm long, pericarp ca. 1.2 mm thick,
septa very thin, fragile, early and irregularly ca¬
ducous, remaining columella terete, not winged
(Fig. IE). Seeds 4—4.5 X 4 mm, with a large car¬
uncle and distinctly foveolate surface (Fig. 1C).

The invariable absence of a distinct leaf apex

34

Novon

and of basal glands (Fig. 1A) are unique in neo¬
tropical Maprounea. Additionally, the style is very
short, almost absent (Fig. I B), whereas it is usually
distinct and 0.5—2 mm long in the other species.
The staminate flowers are similar in all taxa, just
as they are in closely related species throughout
the Hippomaneae. Maprounea amazonica grows
along rivers, particularly in igapo forests and pe¬
riodically inundated white sand campinas, whereas
both other species occur in non-inundated savan¬
nas, and secondary and evergreen forests.

In the Neotropics, Maprounea is represented by
one widespread species, M. guianensis, which oc¬
curs from Panama to Bolivia, and two species with
a more restricted distribution, M. amazonica and
M. brasiliensis.

Maprounea amazonica appears to be confined to
northern and central Amazonia and is most com¬
mon in the Rio Negro basin. A single record of M.
amazonica for eastern Amazonia (Para), cited by
Senna (1984), unfortunately could not be examined
for this study. In the Guianas of northeastern South
America (French Guiana, Guyana, Suriname) M.
guianensis is the only species that occurs there.
One hundred thirty-two collections I studied for a
recent revision of the genus for the Flora of the
Guianas (Esser, in prep.) belong to this species.

Maprounea brasiliensis seems to be restricted to
the cerrado of central Brazil, in Goias, Minas Ger¬
ais and the Distrito Federal. Senna (1984) also cit¬
ed collections from Bahia, but these were not avail¬
able for examination. The distinction between M.
brasiliensis and M. guianensis is less clear-cut than
those between either of them and M. amazonica.
Among the specimens of M. guianensis studied,
only one (Williams 6470 from Peru) showed an ex¬
ceptional fruit length of 6—8 mm; in all other cases
the fruit size was sufficient for distinction from M.
brasiliensis.

The Alrican M. africana M tiller Argoviensis, of
which the abovementioned M. obtusa Pax is a syn¬
onym, has leaves very similar to those of M. ama¬
zonica. It differs sufficiently by inflorescences with
(l-)2— 3 pistillate flowers, long styles, staminate in¬
florescences distinctly longer than wide, and 6-8-
mm-long seeds with a completely smooth surface.

Additional specimens examined. BRAZIL. Amazon¬
as: Rio Negro between Parana da Floresta and Rio Bran¬
co, 26 June 1979 (fl). Alencar 42 ((ill): Rio Negro, Tam-
ma, 26 Apr. 1911 (11), Ducke MG 11574 (BM); Manaus,
Igarape da Caehoeira Grande, 28 Mar. 1937 (II), Ducke
435 (A); Rio Taruma-mirim, Comunidade Nossa Senhora
de Fatima, 16 May 1992 (II). Ferreira 230 (k). If) May
1992 (fr), Ferreira 244 (k); Rio Negro. Rio Marie, 12 June
1947 (II), Frdes 22394 (U); Bio Negro between lllia do
Japo and Bareelos, 27 June 1979 (fr). Poole 1987 (k): Bio

Negro opposite Manaus 3—6 km upstream, 8 Apr. 1971
(lemale II), Prance. Coelho et id. 11752 (k. U); Rio Negro.
Bio Cuieras just below mouth of Rio Brancinho, 26 Sep.
1971 (If). Prance, Coelho A Monteiro 14878 (LI); Rio
Uneiuxi, 5 km above mouth. 8 Nov. 1971 (fr). Prance,
Maas et id. 19174 (k. U); Rio Negro 30 km above Ma¬
naus. 28 Sep. 1974 (fl). Prance 22793 (U); Rio Araea near
Serrinha, 80 m, 25 July 1985 (fr), Prance, Mota A Silva
29799 (k): without locality, without date (II). Regnell HI-
939 (BM. P); Caehoeira Grande bei Manaus, May 1910
(If), l lie 8898 (k. L). COLOMBIA. Vichada: Parque Na-
cional Natural "Kl Tuparro,” S edge of Laguna Guaipe, 80
m. 27 Feb. 1985 (fr). Zarucchi A Barbosa 3490 { k). VEN¬
EZUELA. Amazonas: Bana, 4.5 km SSW from mouth of
R fo Temi, 100 m, 6 Mar. 1996 (fr). Berry, Romero A Brake
9 035 (TCI)); bank of Orinoco River near Samariapo, 90
m. 17 Nov. 1965 (fr). Breteler 4714 (k); Cano Caname,
100 m, 3 May 1979 (fl), l)a mdse, Huber A Tillett 17151
(k); Cano Perm de Agua, 100 m. Nov.-Dee. 1978 (fr),
Huber A Tillett 2751 (k); Cano Bulron, afluenle del Cano
Ueata, 70 m, 21 Oct. 1989 (fr). Romero & Melgueiro 2 1 5 1
(CM); Samariapo, 124 m. 2 July 1942 (fl). Williams 15957
(A); Cano San Miguel just above Limoncito, 15 km from
Rio Cuaima, 120 m, 28 June 1959 (fl). Wurdack A Ad-
derley 43218 ((ill. k).

Hamilton (1825) described Maprounia glauca.
This name has never been used subsequently, and
its interpretation has remained uncertain. It was
usually placed in Maprounea, and Maprounia was
treated as an orthographic variant of it (see Farr et
al„ 1979; Howard et al., 1981). However, Senna
(1984) excluded the species from Maprounea, stat¬
ing that it may belong to Mabea Aublet.

Study of the type at Paris (P) revealed that it does
indeed belong to Mabea taquari Aublet. This syn¬
onymy was recognized by M. Emmerich (1970 in
sehed.), but remained unpublished. I therefore pro¬
pose this new synonymy:

Mabea taquari Aublet, Hist. PI. Guiane 2: 870,
4: pi. 334, flg. 2. 1775. TYPE: Guyane Fran¬
chise. Cayenne, Aublet s.n. (lectotype, desig¬
nated here, BM; isolectotypes, G, P, W).

Maprounia glauca W. Hamilton, Prodr. PI. Ind. Occid.:
54. 1825. Syn. nov. TYPE: Guyana. Without locality.
Herb. Desvaux s.n. (holotype, P).

Be cause Hamilton mentioned only this one spe¬
cies, and no other species was ever described under
Maprounia, Maprounia W. Hamilton becomes a
synonym of Mabea,

Acknowledgments. Specimens from the follow¬
ing herbaria were used in this study: A, BM, BR.
E, G, GH, HBG, K. L, NY, OXF, P, TCD, U, W. I
am indebted to the directors and curators of these
institutions who made their holdings available for
study. The work was financed by a grant (B 85—
303) from the Netherlands Organization for Scien¬
tific Reserarch (NWO). 1 am also indebted to M.

Volume 9, Number 1
1999

Esser

Neotropical Maprounea

35

Jebb, J. Parnell, and two anonymous reviewers who
commented on the manuscript. Drawings are by the
author.

I .iterature Cited

Allem, A. C. 1976. Uma especie linica d e Maprounea (Eu-
phorhiaceae) na America do Sul. Acta Amazon. 6: 4 1 7 —
422.

Earr, E. 1C. J. A. Leussink & F. A. Stafleu. 1979. Index
Nominum Genericorum. Regnum Veg. 100.

Greuter, W.. E. R. Barrie. H. M. Burdet, W. G. Ghaloner,
V. Demoulin, 1). L. Hawksworth, P. M. Jprgensen, D. H.
Nicolson, P. C. Silva. P. Trehane & J. McNeill. 1994.

International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 1.31.

Hamilton. W. 1825. I ’rodromus plantarum Indiae occiden-
talis. Treuttel & Wiirtz, London, Paris & Strasbourg.

Howard, R. A., K. S. Clausen & W. T. Gillis, Jr. 1981.
William Hamilton (1783-1856) and the Prodromus
plantarum Indiae occidentalis (1825). J. Arnold Arbor.
62: 211-242.

Pax, F. & K. Hoffmann. 1912. Euphorbiaceae— Hippoma-
neae. Pp. 1-319 in A. Engler (editor), Das Pflanzen-
reich IV. 147. V (Heft 52). Engelmann, Leipzig.

Senna, L. Mendonya de. 1984. Maprounea (Euphorbi¬
aceae). Considerayoes taxinomicas e anatomicas das es-
pecies sul-amerieanas. Rodriguesia 36(61): 51-78.

A New Triraphis (Poaceae: Eragrostideae) from Brazil: First
Record of a Native Species in the New World

Tarciso S. Filgueiras

Reserva Ecologica do IBGE, Caixa Postal 08770, 70312-970 Brasilia, DF, Brazil

Fernando 0. Zuloaga

Instituto de Botanica Darwinion, Casilla de Correo 22, San Isidro 1642, Argentina

ABSTRACT. Triraphis devia Filgueiras & Zuloaga
from central Brazil is described as a new species,
illustrated, and compared with morphologically
similar species in the genus. Anatomical and eco¬
logical data are presented together with a biogeo-
graphical discussion. A tentative key to all the spe¬
cies of Triraphis is provided. This is the first record
of a native Triraphis in Brazil and in South Amer¬
ica.

During field excursions undertaken by the senior
author and collaborators for the Biogeography of the
Cerrado Biome project (Felfili et al., 1994) several
grass species were collected that could not be iden¬
tified. One of these grasses is herein described as
a new species of Triraphis R. Brown.

Triraphis comprises seven species, six of which
occur in Africa and one in Australia, always in sa¬
vanna habitats, often in sandy or stony soils (Clay¬
ton & Renvoize, 1986; Watson & Dallwitz, 1992).

Triraphis devia Filgueiras & Zuloaga, sp. nov.
TYPE; Brazil. Goias: Chapada dos Veadeiros,
municfpio de Alto Parafso, ea. 17 km da ci-
dade, em dire^ao a Colinas de Goiifs,
14°09'30"S, 47°39'54"W, ea. 1200 m, ereto,
eerca 60 cm de altura, campo limpo, 23 Nov.
1994, T. S. Filgueiras & D. Alvarenga 316H
(holotype, IBGE; isotypes, F, ICN, K. MEXU,
MO, SI, SP, US). Figures 1, 2.

T. andropogonoidi (Steudel) E. Phillips ex Africa similis
autem culmis 12-49 cm longis, laminis 1.8-11.2 cm loti-
gis et 1 .2-2.8 cm latis, panicula 2-6 cm longa, spiculis
4— 8 flosculis absimilis.

Densely caespitose perennials. Rhizomes small,
knotty. Culms 12—49 cm tall, erect, cylindrical, un¬
branched in the vegetative portion, delicate, fragile,
with 1—2 elongated intemodes; internodes 6.5—27.5
cm long, solid, filled with pith, glabrous, striate,
stramineous to purplish, nodes darkish, glabrous.
Leaves mostly basal, a few cauline; basal leaves

strongly imbricate; basal sheaths 3. 5-5. 5 cm long,
strongly keeled, glabrous, both margins hyaline,
glabrous; cauline sheaths 3.5-5 cm long, keeled to
round at the back, striate, margins hyaline, gla¬
brous to sparsely pilose toward the apex; ligule a
dense fringe of hairs, the hairs ca. 0.5 mm long,
pale; the ligular hairs longer and denser at the ex¬
tremities of the ligule; collar undifferentiated; au¬
ricles not present; blade 1.8-11.2 cm long and 1.2-
2.8 mm wide, keeled to flat, linear, glabrous on both
surfaces or with a few scattered hairs, strongly stri¬
ate, surface rough, margins denticulate, apex na-
viculate to subnaviculate, the central nerve occa¬
sionally projected beyond the blade into an
awn-like structure. Inflorescence an open, terminal
panicle, 2—6 cm long and 1.6-4 cm wide (inclusive
of awns), with 3—7 spikelets. Pedicels 5—18 mm
long, scabrid. Spikelets 6-12 mm long and 5-9 mm
wide (inclusive of awns), solitary or paired, laterally
compressed, dark to purplish, 4— 8-flowered, the
apical floret reduced to a sterile lemma; glumes 2,
persistent on the pedicel; lower glume 4—4.5 mm
long, 1 -nerved, rough along the nerve, apex acute,
mucronate or aristate; mucro or awn 0.5-2. 5 mm
long; upper glume 4-4'. 5 mm long, 1 -nerved, thin¬
ner than the lower glume, the apex 2-notched, the
central nerve projected beyond the body of the
glume into a straight mucro or awn, the mucro or
awn 0. 5-1.5 mm long; rachilla 0. 5-0.8 mm long,
well developed between the florets, pilose at the
base; lemma 4-4.3 mm long, pilose at the base, 3-
nerved, with 3 awns, two lateral and one terminal;
lateral awns 4—5 mm long, originating from the lat¬
eral nerves, retrorsely barbed, straight, divergent;
central nerve projecting into a straight awn. the awn
3.8-4. 2 mm long, retrorsely barbed; apex of lemma
deeply 2-notched, the lateral expanses acute to ir¬
regularly eroded at the apex; palea 4—4.5 mm long,
0.8-1. 1 mm wide, strongly 2-keeled, hyaline, apex
2-notched; lodicules ca. Vi as long as the ovary, 2,
fleshy, glabrous; ovary obconic, stipitate, pale;
styles 3; lateral styles 0.5— 0.8 mm long; the central

Novon 9: 36-41. 1999.

Volume 9, Number 1
1999

Filgueiras & Zuloaga
Triraphis devia from Brazil

37

Figure 1. Triraphis devia Filgueiras & Zuloaga. — A. Habit. — B. Detail of apex of the blade. — C. Portion of the leaf
in the ligular area. — I). Spikelet, lateral view. — E. Lower glume, dorsal view. — F. Upper glume, dorsal view. — G.
Lemma, dorsal view. — H. Lemma and palea, ventral view. — I. Lemma, ventral view. — J. Palea with lodicules, ovary,
and stamens. K, L. Caryopsis. — K. Embryo view. — L. Hilum view. ( Filgueiras & Alvarenga 3168.)

38

Novon

A

Figure 2. Leaf blade anatomy of Triraphis devia. — A. Blade outline showing V-shaped blade, with the central median
bundle structurally distinguishable from the lateral first-order bundles; sclerenchyma conspicuous and associated with
l’vbs toward the margins. — B. Detail of radiate chlorenchyma, l'vb and 2’vb with specialized chloroplasts in the outer
bundle sheath. ( Filgueiras & Alvarenga 3168.)

style 0. 1-0.5 mm long; stigmas 2, at the apex of
the lateral styles, purple; stamens 3, filament mi¬
nute, pale, anthers 1.8—2 mm long, yellowish to tan.
Apical floret reduced to a sterile lemma, occasion¬
ally the sterile lemma containing a minute, rudi¬
mentary sterile lemma inside. Caryopsis 1.8—3 mm
long, ellipsoid, tan, with the 3 persistent stylar ba¬
ses; embryo ca. Vi as long as the fruit; hilum V4-V3

as long as the caryopsis, dark, punctate-oblong,
basal.

Paratypes. BRAZIL Goias: Mun. Alto Parafso,
Barque Nacional Chapada dos Veadeiros, campo limpo, 28
Sep. 1995, T. S. Filgueiras & F. C. A. Oliveira 3282 (B,
BM. IBGE, P); Chapada dos Veadeiros, 17 km de Alto
Paratso/Colinas, Erva ca. 0.40 m de altura. campo limpo
arenoso, 23 Nov. 1994, M. Aparecida da Silva & F. C. A.
Oliveira 2382 (IBGE, SP).

Volume 9, Number 1
1999

Filgueiras & Zuloaga
Triraphis devia from Brazil

39

Leaf Anatomy

The standardized terminology of Ellis (1976,
1979) was used to describe the anatomical struc¬
ture of the leaf blades. The following abbreviations
are used in the anatomical description: vbs: vas¬
cular bundles; Evb(s): first-order vascular bun¬
dle^); 2'vb(s): second-order vascular bundle(s);
obs: outer bundle sheaths.

LEAF BLADE IN TRANSVERSE SECTION (FIG. 2)

Outline: V-shaped; two halves of lamina sym¬
metrical about the median vascular bundle; leaf
blade section includes between 18 and 20 vbs; ad-
axial and abaxial surface without ribs and furrows.
Midrib : a definite keel; median vascular bundle
structurally distinguishable from other l’vbs. Vas¬
cular bundle arrangement: 3 2’vbs between consec¬
utive lateral l'vbs; all vbs positioned at the same
level. Vascular bundle description: l'vbs elliptical
to subcircular in outline, circular metaxylem ves¬
sels narrow with diameters less than half those of
the obs cells; 2'vbs elliptical in outline, with phlo¬
em and xylem distinguishable. Vascular bundle
sheaths: outer Kranz parenchyma sheath conspic¬
uous, regular in outline, continuous in 2’vbs, inter¬
rupted in l’vbs by sclerenchyma girders toward
both surfaces or only toward the abaxial surface;
specialized chloroplasts present, of centrifugal po¬
sition (?); adaxial and abaxial bundle sheath exten¬
sions absent. Inner mestome sheath entire, of small
cells with uniformly thickened walls.

Sclerenchyma: small, inconspicuous scleren-
chyma strands associated with all vbs, both adaxi-
ally and abaxially, except toward the margins where
they are conspicuous and associated with first-order
vascular bundles.

Mesophyll: chlorenchyma radiate, interrupted
toward both surfaces by sclerenchyma girders, com¬
pactly arranged with a few intercellular air spaces;
chlorenchyma cells tabular; 2—3 chlorenchyma
cells between consecutive vbs. Arm cells absent.
Fusoid cells absent.

Adaxial epidermal cells: bulliform cells con¬
spicuous and associated with the median vascular
bundle, fan-shaped and covering V2—V4 the width of
the transverse section; also in small groups in ad¬
axial furrows and between the vascular bundles.
Epidermal cells small, regular in size, the cuticle
thick and continuous; papillae and prickles pre¬
sent. Abaxial epidermal cells: bulliform cells ab¬
sent; epidermal cells small; papillae absent; hooks
present.

Discussion

This new species shares the following generic
characters with other taxa included in Triraphis: in¬
florescence an open panicle, with spikelets laterally
compressed, several flowers per spikelet, lower and
upper glume 1-nerved, shorter than the spikelet,
lemmas with three awns, and a similar leaf anatomy
and Kranz syndrome. Triraphis andropogonoides,
the closest relative of T. devia, can be distinguished
by being rhizomatous plants, with long creeping
rhizomes, the culms 120 cm tall, the leaves 20—40
cm long, inflorescences 12—30 cm long, and spike-
lets 5— 15-flowered.

The following is a tentative key to distinguish the
species in Triraphis (based on Gibbs Russell et ah,
1990):

I . Plants annual . 2

I '. Plants perennial . 3

2(1). Spikelets 2-4 mm long; anthers 0.2— 0.4 mm . . . .

. T. pumilio R. Brown

2'. Spikelets 6-10 mm long; anthers 1.2—2 mm . . . .

. T. purpurea Hackel

3(1). Culms profusely branched .

. T. rarnosissima Hackel

3’. Culms not branched . 4

4(3). Panicle 2-6 cm long .

. T. devia Filgueiras & Zuloaga

4'. Panicle 12—40 cm long . 5

5(4). Plants up to 60 cm tall; panicle contracted; spikelets

4 mm long . T. mollis R. Brown

5'. Plants up to 140 cm tall; panicle open; spikelets 6—

1 1 mm long . 6

6(5). Panicle sparse; central awn of lemma longer than

the lemma . T. schinzii Hackel

6'. Panicle dense; central awn of lemma shorter than

the lemma .

. T. andropogonoides (Steudel) E. Phillips

The arrangement of mesophyll tissue, together
with the presence of specialized chloroplasts in the
outer bundle sheath cells, indicates that this new
species has a Kranz, PS (=XyMs + ) anatomy (ter¬
minology of Hattersley & Watson, 1976). Also, two
or three chlorenchymatous mesophyll cells are
found between successive vascular bundles, which
further indicates a C+ anatomical organization (Hat¬
tersley & Watson, 1975). This agrees with the anat¬
omy described for Triraphis (Watson & Dallwitz,
1992). The position of the specialized chloroplasts,
in the outer bundle sheath, was difficult to ascertain
due to the absence of fresh material to carry on the
anatomical observations. It is noteworthy to men¬
tion the presence of conspicuous sclerenchyma tis¬
sue toward both margins, combined with the pres¬
ence of well-developed bulliform cells toward the
keel of the blade; consequently, it is possible to
conclude that the blades can be easily folded under
water stress conditions.

40

Novon

Two populations of Triraphis devia were located
during field excursions, one inside the Parque Na-
cional Chapada dos Veadeiros and the other along
an unpaved road, between the town of Alto Parafso
and the village of Sao Jorge. The latter population
is obviously under great threat because the area is
used as a native pasture. In addition, erosion is
beginning to eliminate the vegetation along the un¬
paved road, including plants of this new species.
The population inside the Parque is legally pro¬
tected. In both cases, the plants grew in stony soil,
apparently extremely poor in nutrients. The vege¬
tation where the species was found is classified as
a campo limpo de cerrado (i.e., tropical seasonal
shortgrass field of cerrado [Eiten, 1983]).

Almost all the plants seen had some leaves
clipped. Cattle, horses, wild animals, and insects
seem responsible for the leaf predation of Triraphis
devia. These plants are evidently fire-resistant,
since all populations examined displayed signs of
fire damage.

The discovery of this new species of Triraphis in
the high elevations (ca. 1200 m) of central Brazil
is exciting and puzzling: exciting because no native
species of this genus has ever been reported for
South America. The only known record of a species
of Triraphis in South America is that of T. andro-
pogonoides, taken in Rio de Janeiro, Brazil, in 1833
(Filgueiras & Burman, in press). That record ob¬
viously represents a case of a furtive introduction
that never became successfully established. Trira¬
phis andropogonoides, a native of South Africa
(Gibbs Russell et al., 1990), has never been re¬
ported or collected again in Rio de Janeiro or else¬
where in Brazil or South America (Nicora & Rii-
gulo, 1987; Watson & Dallwitz, 1992). This is also
puzzling because it is the sole known species of a
predominantly African genus in the New World,
with a single species in Australia.

Regarding the plant and animal relationships be¬
tween Africa, America, and Australia, two different
points of view developed in the last decades.
Thorne (1973), Smith (1973), and other authors em¬
phasized differences between the continents and
explained present disjunct distributions as exam¬
ples of long-distance dispersal between Africa and
America; consequently, the relationships between
these continents are, according to these authors,
derived mostly from immigration, supplemented by
a few subsequent transatlantic long-distance dis¬
persal events (Gentry, 1993). On the other hand.
Raven and Axelrod (1974) showed, in their synthe¬
sis of angiosperm biogeography, that there are many
examples, within angiosperms, of a vicariant distri¬
bution between Africa, South America, and Austra¬

lia: e.g., the Proteaceae, Restionaceae, Podocarpa-
ceae, Winteraceae, and other groups. A similar
biogeographical pattern was summarized in the An-
nonaceae by Schatz and Le Thomas (1993), and in
the Asteraceae by Bremer (1993). Zoological evi¬
dence seems to follow the same pattern. For ex¬
ample, the Lung-fishes of the group Dipnoi are rep¬
resented today by otdy three genera, one from
Australia, one from Africa, and another from South
America (Greenwood, 1975).

Within the Poaceae, Clifford and Simon (1981)
and Simon and Jacobs (1990) emphasized that Tri¬
raphis is a unique chloridoid genus with a disjunct
distribution between Australia and Africa. They in¬
cluded a number of chloridoid genera as Gondwan-
an elements, such as Chloris Swartz, Diplachne
Beauvois, Enneapogon Beauvois, and Eragrostis
Wolf. Simon (1989) stressed that major groups were
well dispersed throughout the world before the final
breakup of the Gondwanaland supercontinent. Con¬
sequently, the present distribution of Triraphis , with
one species in Australia, another in South America,
and six species in Africa can be explained by a
vicariant event, in which the genus evolved in arid
areas after the breakup of the supercontinent Gond-
wana; Raven and Axelrod (1974) stated that “many
of the endemic families and taxa of the two conti¬
nents may have survived in, or later evolved in
such edaphic deserts.”

The specific epithet alludes to the fact that the
plants of this new species are geographically de¬
viated ( devius , a, urn) from other species in the ge¬
nus.

Acknowledgments. We gratefully thank O. Mor-
rone (SI), who kindly helped us with the anatomical
data, and Z. E. Rugolo (SI) for critical comments
regarding the taxonomic position of the new spe¬
cies. G. Davidse (MO) reviewed the final version of
the manuscript. Vladimiro S. Dudds, San Isidro,
provided the excellent illustration. The senior au¬
thor (T.S.F.) thanks the Conselho Nacional de De-
senvolviinento Cientffico e Tecnologico (CNPq) for
a scholarship (Proe. # 301 190/86-0).

Literature Cited

Bremer, K. 1993. Intercontinental relationships of African
and South American Asteraceae: A cladistic biogeo¬
graphic analysis. Pp. 105-135 in P. Goldblatt (editor).
Biological Relationships between Africa and South
America. Yale Univ. Press, New Haven and London.
Clayton, VL I). A S. A. Renvoize. 1992. A classification
system for the grasses. Pp. 328-389 in G. P. Chapman
(editor), Grass Evolution and Domestication. Cambridge
Univ. Press, Cambridge.

Clifford. H. I. <X R. K. Simon. 1981. The biogeography of

Volume 9, Number 1
1999

Filgueiras & Zuloaga
Triraphis devia from Brazil

41

Australian grasses. Iri A. Keast (editor). Ecological Bio¬
geography of Australia. W. Junk, The Hague.

Eiten, G. 1983. Classifica^ao da vegetagao do Brasil.
CNPq, Brasilia.

Ellis, R. P. 1976. A procedure for standardizing compar¬
ative leaf blade anatomy in the Poaceae. I. The leaf
blade as viewed in transverse section. Bothalia 12: 65—
109.

- . 1979. A procedure for standardizing comparative

leaf blade anatomy in the Poaceae. II. The epidermis
as seen in surface view. Bothalia 12: 641-672.

Felfili, J. M.. T. S. Eilgueiras, M. Haridasan, M. C. Silva
Junior, R. C. Mendonga & A. V. Rezende. 1994. Projeto
biogeografia do bioma cerrado: vegeta^ao e solos. Cad-
ernos de Geociencias 12: 75-166.

Filgueiras. T. S. & A. G. Burman. In press. Classification
of the Poaceae genera in Brazil. Revista Brasil. Bot.

Gentry, A. H. 1993. Diversity and floristic composition of
lowland tropical forest in Africa and South America. Pp.
500—547 in P. Goldblatt (editor). Biological Relation¬
ships between Africa and South America. Yale Univ.
Press, New Haven and London.

Gibbs Russell, G. E., L. Watson, M. Koekemoroer, L.
Smook, N. P. Barker, H. M. Anderson & M. J. Dallwitz.
1990. Grasses of South Africa. Bot. Surv. S. Africa 58:
1-437.

Greenwood, P. H. 1975. Norman’s “A History of Fishes,’"
3rd ed. Ernest Benn, London.

Hattersley, P. W. & L. Watson. 1975. Anatomical param¬
eters for predicting photosynthetic pathways of grass
leaves: The “maximum lateral cell count” and the
“maximum cells distant count.” Phytomorphology 25:
325-333.

- & - . 1976. Cj grasses: An anatomical cri¬
terion for distinguishing between NADP-Malic enzyme
species and PCK or NAD-malic enzyme species. Aus¬
tral. J. Bot. 24: 297-308.

Nicora, E. G. & Z. E. Rugolo de Agrasar. 1987. Los ge-
neros de grammeas de America Austral. Hemisferio
Sur. Buenos Aires.

Raven, P. H. & D. I. Axelrod. 1974. Angiosperm bioge¬
ography and past continental movements. Ann. Missouri
Bot. Card. 61: 539-673.

Schatz, G. E. & A. Le Thomas. 1993. Annonaceae: A
primitive dicot family with an ancient center in Africa-
South America. Pp. 86—104 in P. Goldblatt (editor). Bio¬
logical Relationships between Africa and South Amer¬
ica. Yale Univ. Press, New Haven and London.

Simon, B. K. 1989. The biogeography of tropical Austra¬
lian grasses. Proc. Ecol. Soc. Australia 15: 267-269.

- & S. W. L. Jacobs. 1990. Gondwanan grasses in

the Australian llora. Austrobaileya 3: 239—260.

Smith. A. C. 1973. Angiosperm evolution and the rela¬
tionship of the floras of Africa and America. Pp. 49—62
in B. Meggers, E. Ayensu & W. Duckworth (editors).
Tropical Forest Ecosystems in Africa and South Amer¬
ica: A Comparative Review. Smithsonian Institution
Press, Washington, D.C.

Thome, R. F. 1973. Floristic relationships between trop¬
ical Africa and tropical America. Pp. 27—48 in B. Meg¬
gers, E. Ayensu & W. Duckworth (editors). Tropical For¬
est Ecosystems in Africa and South America: A
Comparative Review. Smithsonian Institution Press,
Washington, D.C.

Watson. J. & M. J. Dallwitz. 1992. The Grass Genera of
the World. C. A. B. International, Wallingford.

Una Especie Nueva de Furcraea (Agavaceae) de Chiapas, Mexico

A b is at Gn rc(a - Men doza

Jardfn Botanieo, Instituto de Biologfa, UNAM, A.P. 70— 614, Del. Coyoacan, 04510 Mexico,

D.F., Mexico

RESUMKN. Se describe e ilustra por primera vez
a Furcraea niquivilensis, del estado de Chiapas,
Mexico. Esta especie tiene afinidad con F. mac-
dougallii pero difiere de ella por la forma de la
hoja, la panfcula piramidal, el tamano mayor de las
flores y forma y tamano de los bulbilos.

Abstract. Furcraea niquivilensis from Chiapas,
Mexico, is described and illustrated for the first
time. It is related to F. macdougallii, from which it
differs I >y the shape of the leaves, the pyramidal
panicle, as well as by the flower size, and shape
and size of the bulbils.

Como parte de la revision taxonomica del genero
Furcraea que se estd realizando para Mexico, se
visitaron localidades del estado de Chiapas cerca-
nas a la frontera eon Guatemala, de donde prove-
man dos ejemplares de herbario esteriles, colecta-
dos por Eizi Matuda en los anos 1973-1974. Estos
espeefmenes fueron identificados tentativamente
por Garcfa- Mendoza y Lott (1994) como F. mac-
dougallii ; sin embargo, con la obtencion de mate¬
rial fertil en el cainpo, se encontro que, se trata de
uiia especie nueva que se describe a continuacion.

Furcraea niquivilensis Matuda ex Garcfa-Men-
doza, sp. nov. TIPO: Mexico. Chiapas: Muni-
cipio de Motozintla, Barrio Tuchamen, 10 km
al S de Niquivil camino a Pavincul, alt. 2347
m, 15°14'02"N, 92°12'43"W, 2 abr. 1997 (H),
,4. Garcia-Mendoza, L. de la Rosa & A. Cas¬
taneda 6411 (holotipo, MEXU; isotipos,
ENCB, K, MO). Figura 1.

Flanta monocarpiea arborescens; folia lanceolata
(170— )190— 210 cm longa, 12-14 cm lata, marginis den-
tibus 5— 6(-8) mm longis, decurrentibus, supra infraque
muricata, apice acuminata. Panicula pyramidalis, ramis
primariis 2-2.3 m longis, secundariis 30-60 cm longis.
Flores (7— )7.f> — 8 cm longi; bulbili (4— )5.5— 6.5 cm longi,
(3— )4.5— 6(— 6.5) cm lati, sphaeroidei vel late conici.

Plantas monocarpieas, arborescentes, troncos 1—
3 m de alto, 30—40 cm de aneho, simples, con una
roseta de hojas en el apice, diametro de la roseta
en la parte media de 4—5 m; 80-150 hojas por
roseta. Hojas (1 70— )190— 210 X 12—14 cm, lanceo-

ladas, ereetas, verdes, las secas persistentes a lo
largo del tronco, fibrosas, eoriaeeas, superficie del
haz y enves escabrosa o muricada, base de la hoja
angostdndose hasta 7—8.5 cm, en corte transversal
depreso-obovada, de 5—5.5 cm de grosor, apice eon
mucron de 1-4 mm, castafio oscuro; margen recto,
dentado, dientes de 5— 6(— 8) mm de largo, 3-4 mm
de aneho en la base, decurrentes, por lo que el
margen llega a ser corneo, sobre pequenas mami-
las, antrorsos o rectos en la base, retrorsos en la
parte media y antrorsos en la parte superior, cas-
tanos con la base amarillenta, separados entre sf
por ( 1 — )2 — 4 cm a la mitad de la hoja y por 0.6-
1.5(— 2) cm en la base. Panfculas de 6-9 m de alto,
bu lbilff eras, piramidales; pedunculos de 1.5 m de
longitud, verdes, puberulentos; ramas primarias de
2-2.3 m de largo, puberulentas, ramas secundarias
de 30-60 cm de largo, puberulentas. Flores
(7— )7.5— 8 cm de largo, en fascfculos de 1—3; pe-
dieelos 0.5—1 cm, puberulentos; tepalos externos
(3— )4— 4.5 X 1.1— 1.3 cm, oblongos, glabreseentes,
los internos 4-4.5 X 1.2—14 cm, oblongos, gla¬
breseentes sobre la prominente vena media, papi-
losos en el area de sobreposicion, ambos de color
verde-blanquecino con tintes rojizos en la parte ex¬
terna, blanquecinos por dentro; estambres 2—2.5
cm de largo, 5-8 mm en su parte mas engrosada,
papilosos en la base, de color verde-atnarillento;
anteras 4—6 mm de largo, 2—3 mm de aneho, oblon-
gas, amarillas; ovario 3.6-3.8 X 0. 4— 0.6 cm, eilhi-
drico, puberulento, verde; estilo 2. 5-2.8 cm de lar¬
go, 0.6-0. 8 mm en la base, trisulcado, papiloso,
verde-amarillento, estigma entero, en ocasiones ei-
liado. Bulbilos (4— )5.5 — 6.5 cm de alto, (3 — )4.5 — 6( —
6.5) cm de aneho, esferoides a ampliamente coni-
cos, cubiertos por 4—6 bracteas ampliamente ova-
das, pardas, cartaceas, persistentes. Fmtos y se-
millas deseonoeidos.

Etimologia y usos. El epfteto espeeffieo alude
a la region principal donde crece la especie, po-
blada por el grupo etnieo Maine, quienes la cono-
cen con el nombre generieo de maguey y la plantan
como cerco para delimitar sus terrenos y evitar la
erosion del suelo; asf mismo, ulilizan las libras de
las hojas para hat er eanastos y cuerdas, actividad

Novon 9: 42-45. 1999.

Volume 9, Number 1
1999

Garci'a-Mendoza

Furcraea niquivilensis de Mexico

43

Figure ] . Furcraea niquivilensis Matuda ex Garcfa-Mendoza. —a. Planta completa. — b. Hoja y sus secciones trans-
versales. — c. Corte de la hoja, mostrando el env6s muricado. — d. Mucron. — e. Flor con tepalos abiertos. — f.
Fstambre. — g. Estilo. — h. Rama con bulbilos. Ilustracidn basada en el especimen Garcia-Mendoza et al. 6441.

44

Novon

que se realiza en muy l>aja escala en la actualidad,
pero que fue sobresaliente en el pasado.

Distribution y habitat. Furcraea niquivilensis
se ha encontrado liasta el momento unicamente
como planta cultivada cerea de las poblaciones; sin
embargo, es posible que aun crezea dentro de los
bosques de Pinus-Quercus o bosques niesofilos de
montafia de la region circundante, tanto de Chiapas
como del vecino Departamento ile San Marcos en
Guatemala. En la actualidad estos bosques se en-
cuentran muy perturbados y no se encontraron
ejemplares de esta especie creciendo en forma sil-
vestre. Las plantas cultivadas crecen en laderas
hasta de 50° de pendiente, sobre suelos arenosos
de color pardo y en altitudes que oscilan entre los
1800 y 2650 m sum.

Relaciones taxondmicas. Furcraea niquivilensis
y F. macdougallii Matuda, tienen caracterfsticas
morfolbgieas similares, como son, la presencia de
tallo, superficie de la hoja escabrosa o muricada
por el haz y el enves; superficie puberulenta en
ramas primarias, secundarias, pedicelos de las flo-
res y ovarios. Sin embargo, difieren en varios atri-
butos importantes que se senalan a continuation.

Las hojas de F. niquivilensis son lanceoladas y
presentan una relacion entre el largo y el ancho de
13—15:1, mientras ijue, Furcraea macdougallii tie-
ne hojas lineares, con una relacion largo-aneho de
19—21:1. Los dientes del margen de la hoja en F.
niquivilensis miden 5— 6(— 8) mm y el apice es acu-
minado, mientras que, en F. macdougallii los dien¬
tes del margen miden 2-4 mm y el apice es lar-
gamente acuminado. La panfcula en F. niquivilensis
tiene forma piramidal, con las ramas primarias de

2— 2.3 m y las ramas secundarias de 30—60 cm de
largo; en contraste, la panfcula de F. macdougallii
es romboidal, las ramas primarias miden 1-1.5 m
de largo y las ramas secundarias 15—26 cm. Tam-
bi£n difieren en el tamano de las flores, ya que, en
F. niquivilensis 6stas alcanzan tamanos de 7.5-8 cm
de largo y en F. macdougallii son de 3—3 . 5( — 4) cm.
Por ultimo, una diferencia conspicua entre ambas
especies se observa en el tamano y forma de los
bulbilos; en F. niquivilensis son esferoides o am-
pliamente conicos y miden de (4— )5.5 — 6.5 cm de
alto por (3— )4.5 — 6(— 6.5) cm de ancho, en tanto que,
los bulbilos de F. macdougallii son conicos y miden

3— 5( — 6) cm de alto por 1-1.5 cm de ancho. De
hecho, F. niquivilensis posee, los bulbilos mas gran-
des conocidos para cualquier especie de Furcraea
en Mexico y Centroamerica y quizd sean los bul¬
bilos mayores en tamano de todo el genero.

Furcraea macdougallii , es una especie restrin-
gida a la Mixteca Baja en los lfmites de Puebla y
Oaxaca y posiblemente al Istmo de Tehuantepec,

area de donde fue descrita originalmente, pero don-
de no se le ha visto mds. Crece sobre suelos are¬
nosos de origen calizo, en altitudes entre los 750 y
1200 m snm. Se le conoce solo de ambientes se-
eundarios derivados de la selva baja caducifolia es-
pinosa, donde convive con Agave marmorata Boezl.
Ceiba parvifolia Rose, Haematoxylum brasiletto
Karst, Hechtia glomerata Zuccarini, Myrtillocactus
geometrizans (C. Martius) Console, Neobuxbaumia
tetetzo (F. A. C. Weber) Baekeberg, Plumeria rubra
L. y Tabebuia guayacan (Seemann) Hemsley.

En el estado de Chiapas crecen cuatro especies
de Furcraea adicionales. F. quicheensis Trelease, se
reconoce por sus hojas glaucas, con margen fina-
inente denticulado, flores glabras, verdes y eapsu-
las oblongas. Furcraea samalana Trelease, se ca-
raeteriza porque sus hojas no tienen dientes en la
mitad o % superiores, la inflorescencia es laxa con
flores blanquecinas y bulbilos foliosos. Ambas es¬
pecies son endemicas de Chiapas y Guatemala.
Trelease (1915) describio F. guatemalensis y aun-
que Garcfa-Mendoza y Lott (1994) mencionan que
crece en Chiapas, hasta el momento no se le ha
hallado; los especfmenes de herbario identificados
con este nombre posiblemente correspondan a una
especie nueva. Existe un taxon mas, cuya cireuns-
cripcion taxonomica y nomenclatural esta aun bajo
estudio.

Paratipos. MEXICO. Chiapas: Municipio de Moto-
zintla, Zaragoza, 36 km al SE <le Motozintla y 10 km al
S de Niquivil, alt. 2720 m, 15°I4'16"N. 924 3'5rW, 2
alir. 1997, .4. Garcia- Mendoza, /.. de la Rosa & A. Cas¬
taneda 6439 (MEXU); Municipio de Motozintla, Barrio
de luehamen. 10 km al S de Niquivil camino a Pavilion!,
alt. 2347 m. 15°14'02"N, 92°12'43"W. 2 abr. 1997, A.
Garcia- Mendoza, L de la Rosa A. Castaneda 6440
(CHIP, ENCB. MEXU, TEX), 6442 (DES, ENCB.
MEXU, XAE), 6443 (ENCB, MEXU. MO, UAMIZ, US),
6444 (ENCB. MEXU). 6445 (MEXU), 6446 (MEXU):
Barrio Tonincanaque, 8 km al S de Niquivil camino a
Pavincul. 2650 m, I5°14'55"N, 92°13'28"W, 2 abr.
1997, 4. Garcia- Mendoza, I,, de la Rosa & A. Castaneda
6448 (MEXU); Allende, 12 km al NO de Niquivil camino
a Motozintla, 2137 m, 15°19'07"N, 92°14'11"W, 2 abr.
1997, 4. Garcia- Mendoza, /.. de la Rosa & A. Castaneda
6455 (MEXU); Aserradero viejo, entre Toliman a Niqui¬
vil, alt. 1800 m, 24 ene. 1973, E. Matuda 38517
(MEXU); Orilla de arroyo rocoso en bosque de pinos. 13
abr. 1974, E. Matuda 38625 (MEXU).

Agradecimientos. Este trabajo se hizo con apo-
yo financiero de la Comisibn Nacional para el Co-
nocimiento y Uso de la Biodiversidad (CON A BIO)
a traves del proyecto Hill. Agradezco a Patricia
Davila, Raquel Galvan, Fernando Chiang y dos re-
visores anonimos sus comentarios al manuscrito.
Fernando Chiang elaboro la descripcion latina.
Gracias a Alejandro Castaneda y Laura de la Rosa

Volume 9, Number 1
1999

Garci'a-Mendoza

Furcraea niquivilensis de Mexico

45

por su apoyo durante las colectas en Chiapas. La
exeelente ilustracion es obra de Elvia Esparza.

I .iteratura Citada

Garcfa-Mendoza, A. & E. J. Lott. 1994. Furcraea Vent.
Pp. 45-47 en: G. Davidse, M. Sousa & 0. Chater (edi-

tores), Flora Mesoamericana, Vol. 6. Alismataceae a
Cyperaceae. Universidad Autonoma de Mexico, Insti-
tuto de Biologia, Mexico 1)4.; Missouri Botanical Gar¬
den, St. Louis; The Natural History Museum, London.

Trelease, W. 1915. The Agaveae of Guatemala. Trans.
Acad. Sci. St. Louis 23(3): 129—152. t. 6—35.

A New Species of Chinese Adenophora (Campanulaceae)

Ge Song and Hong De-yuan

Laboratory of Systematic and Evolutionary Botany, Institute of Botany, Chinese Academy of
Sciences, Beijing 100093, People’s Republic of China

Abstract. Adenophora ningxianica is described,
and its relationships to A. wawreana Zahlbruckner
and A. gmelinii (Sprengel) Fischer are discussed.

The name Adenophora ningxianica Hong, which
first appeared in llong (1983) and subsequently in
many local Chinese floras, was not validly pub¬
lished because it was not accompanied by a Latin
description or diagnosis. It is herein validated, and
a detailed description in English is provided.

Adenophora ningxianica Hong ex Ge & Hong,
sp. nov. TYPE: China. Ningxia: Ala (Helan)
Mountains, Taszekow, 1600 m, 28 Aug. 1933,
Y. Y. Pei 120 (holotype, PE).

Species sirnilis A. wawreanae et A. gmelinii, sed ab ilia
differt foliis sessilibus glabris, eapsula tenui cylindrica ca.
8 mm longa, seminibus majoribus; ab hae differt corolla
parva et angusta, stylo exserto, calycis lobis angustis par¬
vis et nigro-callosis.

Perennial herbs. Stems 30-50 cm tall, several
from base, unbranched, erect, glabrous, rarely pu¬
bescent. Basal leaves cordate or obovate; cauline
leaves alternate, sessile or short petiolate; leaf
blade 2-7 cm long, 0.2-1. 8 cm wide, lanceolate to
linear-lanceolate or rarely linear, glabrous on both
surfaces, base cuneate, margin serrate. Inflores¬
cences few-flowered racemes or panicles. Pedicel
0.5— 1.5 cm long, slender. Calyx tubes obovoid, gla¬
brous; calyx lobes 2-6 mm long, ca. I mm wide,
subulate or subulate-lanceolate, margin mostly with
a pair of small verrucate teeth or rarely entire. Co¬
rolla 1.4—1. 6 cm long, narrowly campanulate, blue
or purplish blue; lobes ca. 3.5 mm long, ovate-tri¬
angular. Disc 2-2.5 mm long, tubular, glabrous.

Style 1.6-1. 8 cm long, slightly exserted. Capsule
ca. 8 mm long, ca. 3 mm diam., long ellipsoid.
Seeds ca. 2 mm long, yellow, oblong, with a wing¬
like ridge. Flowering July-August, fruiting Septem¬
ber— October.

Adenophora ningxianica is most closely related
to A. wawreana Zahlbruckner and A. gmelinii.
From these, it differs in having a narrower corolla,
a slightly exserted style, sessile or short-petiolate
and glabrous cauline leaves, smaller calyx lobes
with toothed margins, slender capsules, and larger
seeds. Adenophora wawreana has a broadly cam¬
panulate corolla, a distinctly exserted style, pubes¬
cent cauline leaves with petioles ca. 2.5 cm long,
4— 10-mm-long calyx lobes, broader capsules, and
seeds 0.8-1. 2 mm long. Adenophora gmelinii dif¬
fers from A. ningxianica in having a broadly cam¬
panulate corolla, a slightly included style, entire
and wider calyx lobes, and broader capsules 4—7
mm in diameter.

Paratypes. CHINA. Ningxia: Helan Mt., Suyukou, V!
(A He 73 IS (PE); Helan Mt.. Wudaotai, Yellow Hirer Ex-
ped. 3950 (PE); Helan Mt., Cilawu Vellay, Group lie 2SS
(PE). Gansu: Lanzhou, Xinlong Mt., Y. (A lie 5300 (PE).

Acknowledgments. We are grateful to Yang Qin-
er for revising the Latin diagnosis and Ihsan Al-
Shehbaz for his help with the manuscript. This
work was supported by the National Natural Sci¬
ence Foundation (Youth) of China (grant
#39600009).

I .iterature Cited

Hong De-yuan. 1983. Adenophora. In: Hong De-yuan (ed¬
itor), Campanulaceae. FI. Reipubl. Popularis Sin. 73(2):

92-140.

No von 9: 46. 1999.

The Identity of Taraxacum altune D. T. Zhai & Z. X. An

Ge Xue-jun

South China Institute of Botany, Chinese Academy of Sciences, Guangzhou, Guangdong 510650,

People’s Republic of China

Zhai Da- tong

Department of Geography, Normal College of Shanxi University, Taiyuan, Shanxi 030012,

People’s Republic of China

Abstract. Taraxacum altune D. T. Zhai & Z. X.
An is reduced to synonymy ot Crepis minuta Kita-
mura.

The recently described Taraxacum altune (Zhai
& An. 1995) was based on a flowering specimen
collected by Y. H. Wu in 1988. During the revision
of Taraxacum lor Flora Reipuhlicae Popularis Sin-
icae, we examined additional specimens of the
same species, including fruiting material. The
aehenes of this material are blackish, cylindric,
narrowed toward the apex, beakless, smooth, and
ribbed, and t lit* scapes have a small leaf. In Tarax¬
acum the leaves are all basal, and the aehenes are
fusiform to oblanceolate, often spinulose near the
apex, usually slender-beaked, and clearly with a
more or less swollen region between the beak and
body. Therefore, plants of T. altune represent a spe¬
cies of Crepis h. Critical comparison of the type

material of T. altune with Chinese species of Crepis
(Tang, 1985) reveals that T. altune is conspecific
with Crepis minuta Kitamura. The type of C. minuta
was collected by Kawaguchi (SM 96895, now TNS)
from Xizang (Tibet).

Crepis minuta Kitamura, Acta Phytotax. Geobot.
XV: 70. 1953.

Taraxacum altune I). T. Zhai & Z. X. An, J. Aug. 1st Agri.
College. 18(3): 1. 1995. Syn. nov. TV PK: China. Xin¬
jiang: Qiemo, Y. //. Wu 2644 (HNWP).

Additional material examined. CHINA. Xizang: Na-
garze. Qinghai-Xizang Exped. 74—2061 (KUN).

Literature Cited

Tang Y. C. 1985. Crepis L. In: Wu Cheng-yih (editor). FI.

Xizang 4: 958—961. Science Press. Beijing.

Zhai D. T. & An Z. X. 1995. New Species of Taraxacum
from Xinjiang. .1. Aug. isl Agri. College. 18(3): 1-5.

Novon 9: 47. 1999.

Four New Species of Memora (Bignoniaceae) from South America1

Warren 1). Hauk

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, L.S.A. Current
address: Department of Biology, Denison University, Granville, Ohio 43023, U.S.A.

ABSTRACT. Four new South American species of
Memora (M. cidii, M. contractu , M. sastrei, and M.
velutina) are described and three are illustrated.
Memora cidii has flattened inflorescence axes, cam-
panulate calyces, and bracteoles extending beyond
the calyx tip. Glands at the base of the corolla lobes
distinguish M. contractu , and M. sastrei is unique
in having conspicuous yellow-dendroid pubescence
on its inflorescence axes and calyces. Memora ve¬
lutina is distinguished by pilose pubescence
throughout. Relationships of these species to other
Memora species are discussed.

Memora Miers (Bignoniaceae) is a genus of 32
species of lianas, scandent shrubs, or treelets found
in tropical South America east of the Andes (Gen¬
try, 1977, 1982, 1997). Memora is distinguished
from other genera of tribe Bignonieae by the com¬
bination ot terete branchlets and pinnately or bi-
pinnately compound leaves (Gentry, 1978). Memora
species produce showy, yellow to red, tubular-fun-
nelform corollas, linear-oblong to oblong-orbicular
dehiscent capsules, and hialate to corky seeds
(Gentry, 1977, 1997). Gentry and Tomb (1980) sug¬
gested that Adenocalymna , Memora, and two spe¬
cies of Tanaecium may compose a “basically nat¬
ural group" because most of these species share
medium-textured yellow corollas, simple tendrils,
and primarily biseriate ovules.

During preparation of a review of Memora, spec¬
imens annotated by A. H. Gentry were discovered
that were neither described fully nor published va¬
lidly. Although Gentry did not publish them before
his untimely death, these taxa represent evolution¬
ary units suitable for recognition as species. In a
draft of his Flora de Colombia treatment of Big¬
noniaceae, Gentry described Memora sastrei but
did not provide a Latin description or diagnosis.
The other three Memora species presented here, M.
cidii, M. contractu, and M. velutina, had no written
descriptions and few notes pertaining to their cir¬
cumscription.

1 . Memora cidii A. H. Gentry ex Hauk, sp. nov.
TYPE: Brazil. Amazonas: Mun. of Novo Ari-
puana, BR 230, Rod. Transamazonica, 400 km
from Humaitd, 07°15'S, 60°00'W, 4 May 1985,
6’. Ferreira 6013 (holotype, MO!; isotypes,
NY!, US!). Figure 1.

Frutex scandens, ramulis teretibus glabris sine conso-
cibus glandularibus in nodis interpetiolaribus. Folia op-
posita imparipinnata compositis foliolis 1-3 jugatis ovatis-
ellipticis obtusis; pseudostipulis ineonspicuis glandulosis.
Infforescentia paniculata axillaris; bracteolis cupulatis.
Flos calyce tmneato campanulato extus glanduloso apice
pubemlo duabus bracteolis subtentus; corolla tubuloso-
infundibuliformi tubo glabro; thecis antherarum divarica-
tis; ovario oblongo lepidoto; disco pulvinato. A speciebus
aliis rhachidibus et pedunculis complanatis olivaceis, in-
florescentiae bracteolis triangufaribus caducis differt.

Lianas-, branches drying tan to brown, not striate,
glabrate, the surface rough, with interpet iolar glan¬
dular fields lacking and interpetiolar transverse
ridge inconspicuous or absent, lenticels not evi¬
dent; pseudostipules persistent, subfoliaceous,
ovate-oblong, 4-5 X 2-3 mm, glandular, glabrate.
Leaves opposite, estipulate, petiolate, 50-55 cm
long, 2— 3-pinnate with a pair of opposite pinnae
and the terminal pinna often modified into a tendril,
each pinna imparipinnate with 1-3 sets of opposite
simple or compound pinnules, the terminal segment
somewhat larger than the lateral segments; petioles
4-6 cm long, terete, glabrate; petiolules 1—3 cm
long, inconspicuously sulcate, glabrate; ultimate
segments entire, ovate to ovate-elliptic, 16-22 X
7-11 cm, equilateral, plane, chartaceous, glabrate,
apieally obtuse, basally oblique to broadly acumi¬
nate, the venation brochidodromous with 8—10
principal vein pairs, marginally plane; joints of the
compound leaf conspicuously enlarged. Inflores¬
cences elongate, axillary racemes to 15 cm long,
branched, several- to many-flowered; raehis and
peduncles flattened, glabrate, and bracteate, the
bracts triangular, 4-6 X 1-2 mm, caducous, mi¬
nutely glandular, glabrate with ciliate margins; ped¬
icels 1—2 mm long, glabrate; bracteoles ovate-ellip-

1 I bis paper is number 3 of the GENTRA INVITATION SERIES, in acknowledgment of the contributions to the
study of the Bignoniaceae made by Alwyn II. Gentry.

Novon 9: 48-54. 1999.

Volume 9, Number 1
1999

Hauk

Memora from South America

49

tic and cupular, 8-11 X 4—5 mm, glandular,
glabrate with ciliate margins, extending beyond the
calyx tip. Flowers ovoid in bud; calyx campanulate,
5-8 X 5—6 mm, eglandular, costate, the outer calyx
surface glabrate, the inner surface gland-tipped pu¬
bescent, the margin ciliate and intact, apically
truncate except for 5 minute teeth; corolla zygo-
morphic, tubular-funnellorm, yellow, exserted ca.
45 mm beyond the level ol the calyx lip (50—55
mm total length), 2-3 mm wide at the calyx mouth,
with an inner ring of flattened gland-tipped pubes¬

cence 8—10 mm from the corolla tube base, other¬
wise glabrate; corolla lobes 5 (2 upper and 3 lower),
short-orbicular, 12 X 12 mm, the inner surface
glandular-lepidote, the outer surface glabrate, no
glandular fields present at the base of each lobe;
fertile stamens didynamous, ca. 15 or 25 mm long,
inserted near the level of the calyx lip, the single
staminode inserted near the level of the fertile sta¬
mens; fertile anthers with two spreading thecae, in¬
cluded, glabrous; disc present, 1 mm tall; ovary 5
mm long, cylindrical, glabrate to minutely lepidote;

50

Novon

style 25 mm long; stigma bipartite and included.
Fruit unknown.

Distribution. Known only from Amazonas, Bra¬
zil. Elevation unknown. Collected in terra firine
soil.

Phenology. The sole collection was in flower in
May.

Memora cidii is known only from the type col¬
lection. An unusual character of M. cidii is that the
inflorescence axes are flattened, as is found (to a
lesser extent) in M. bracteosa (DC.) Bureau & K.
Schumann, M. mollis A. H. Gentry, and M. race-
mosa A. H. Gentry. Memora cidii has a more highly
branched inflorescence than do M. bracteosa and
M. racemosa. Memora cidii , M. mollis , and M. ra-
cemosa share campanulate, ribbed calyces, but M.
cidii differs from M. racemosa and M. mollis in hav¬
ing bracteoles that extend beyond the calyx. Me¬
mora cidii has much smaller bracteoles than M.
bracteosa. The inflorescence bracts of M. cidii and
M. racemosa are generally caducous, whereas those
of M. bracteosa and M. mollis are persistent. Fur¬
ther insights into the affinities of M. cidii await
study of fruit, seed, and pollen characters.

2. Memora contracts A. 11. Gentry ex Hank, sp.
nov. TYPE: Brazil. Rondonia: Municfpio de
Ariquemes, Mineraryao Mibrasa, Setor Alto
Candeias, km 128. 16 May 1982, L. (). A. Teix-
eira el al. 502 (holotype, MO). Figure 2.

frutex scandens, rainulis teretibus glabris sine conso-
cibus glandularibus in nodis interpetiolaribus. Folia op¬
posite imparipinnata foliolis 1-2 jugatis ellipticis acutis;
pseudostipulis foliaceis ellipticis glandulosis. Inflorescen-
tia racemosa-paniculata axillaris; bracteolis eupulatis.
Flos calyee subspathaceo extns glanduloso pnbendo dua-
bus bracteis puberulis snbtentus; corolla tubnloso-infun-
dibuliformi tubo glabro; thecis antherarum divaricatis;
ovario oblongo lepidoto; disco pulvinato. A speciebus aliis
lobnlis corollinis glandulis elevatis subtentis, pseudostip¬
ulis foliaceis ellipticis differ!.

Lianas', branches brown to gray, not striate, gla-
brate to uncinate, with interpetiolar glandular fields
lacking and interpetiolar transverse ridge present,
lenticels evident; pseudostipules persistent, folia-
ceous, ovate-elliptic, 1-3 X 0.3-1. 5 cm, glandular,
glabrate to uncinate-puberulent, with prominent
veins converging at the base. Leaves opposite, es-
tipulate, petiolate, 27—50 cm long, imparipinnate
with 2 —A pairs of opposite pinnae, tendrils not ob¬
served; petioles 4—8 cm long, terete to slightly sul-
cate, minutely uncinate-puberulent; petiolules 0.3—
2.0 cm long, sulcate, uncinate-puberulent; ultimate
segments entire, 14—22 X 3.5-10 cm, mostly ellip¬
tic to ovate, equilateral, plane, chartaceous, unci¬

nate-puberulent, apically acuminate to acute-acu¬
minate, basally acute to acuminate, the venation
brochidodromous with 5—7 principal vein pairs,
marginally slightly undulate; joints of the com¬
pound leal conspicuously enlarged. Inflorescences
compact, axillary racemes to 10 cm long, un-
branched, several-flowered; rachis and peduncles
terete, minutely puberulent and bracteate, the
bracts linear-elliptic, 5—10 X 1-3 mm, glandular,
glabrate to minutely puberulent; pedicels 1—3 mm
long, minutely puberulent; bracteoles elliptic, 5-12
X 4—6 mm, glandular, minutely puberulent, ex¬
tending Va—Ya the length of the calyx. Flowers ovoid
in bud; calyx bilabiate-subspathaceous, 14—18 X
5-6 mm, glandular, ecostate, the outer calyx sur¬
face glabrate to minutely puberulent, the inner sur¬
face glabrate, the margin splitting irregularly V4-V3
the length of the tube, the valves apically acute
with 5 minute teeth; corolla zygomorphic, tubular-
funnelform, yellow, exserted ca. 55—60 mm beyond
the level of the calyx lip (65 mm total length), 3-
5 mm wide at the calyx mouth, with an inner ring
of pilose pubescence 10-20 mm from the corolla
tube base; corolla lobes 5 (2 upper and 3 lower),
irregularly orbicular, 15 X 15 mm. the inner sur-
taee lepidote-glandular, the outer surface glabrate,
raised glandular fields at the base of each lobe;
fertile stamens didynamous, 15 or 22 mm long, in¬
serted 10 mm beyond the level of the calyx tip. the
single staminode ca. 5 mm long, inserted at the
same level as the fertile stamens; fertile anthers
with two spreading thecae, included, glabrous; disc
1 mm tall; ovary 4 mm long, cylindrical, sparsely
lepidote; style ca. 40 mm long; stigma bipartite and
included. Fruit linear-oblong, 9-20 X 2.5— 3.0 cm,
the valves flattened but not conspicuously thick¬
ened, < 4 mm thick, the margins not obviously
serially constricted, drying dark brown, the midline
evident but inconspicuous, the surface smooth to
verrucate, many-seeded; seeds flattened, 2.5— 3.0 X
2.0 cm, 2-4 mm thick, irregularly quadrate, wing¬
less, the body not distinct.

Distribution. Known from Brazil and Venezue¬
la. Tbe Brazilian coll ections were made at 650 m,
in terra firme forests.

Phenology. Ten fertile collections of M. con-
tracta are known. Single flowering collections are
from May, August, and September, and two collec¬
tions each are from October and November. All
three fruiting collections are from November.

Three of the collections examined were identi¬
fied initially as M. schomburgkii , to which M. con¬
tractu is closely related. However, the presence of
glandular fields at the base of the corolla lobes in

Volume 9, Number 1
1999

Hauk

Memora from South America

51

Figure 2. Memora contracta A. H. Gentry ex Hauk. — A. Inflorescence and stem. — B. Leaves and inflorescence. —
C. Fruit. — I). Seed, proximal and distal sides. (A, It from Vieira el al. 469 , MO; C, I) from das Santas el al. 1 69,
MO.)

M. contracta clearly differentiates the two species
(Fig. 2). As in M. schomburgkii, the calyx of M.
contracta (Fig. 2) is bilabiate to subspathaceous,
but it is generally smaller than that of M. schom¬
burgkii. The fruits of M. schomburgkii are longer
than those of M. contracta , and its seeds thicker.
The pseudostipules of M. contracta are unique in
Memora for their size, elliptic shape, and con¬
spicuous venation. Vegetatively, M. contracta re¬
sembles M. aspericarpa A. H. Gentry, but the lat¬
ter has less conspicuous bracts and bracteoles,
and thicker fruit and seeds.

Paratypes. VKNEZUELA. Bolivar: Dtto. Sifontes,
16-17 Sep. 1989, 1/. Calella el al. 1447 (NY. MO!).
BRAZIL. Amazonas: Manaus— Porto- Vel ho highway, km
124, 25 Mar. 1974. I). G. Campbell et al. P209I0 (MO!);
Rio Taruma, 26 Nov. 1974, A. Gentry & ./. Ramos 12364
(MO); Minera§ao, Campo Novo, 16 Oct. 1979, C. Vieira
et al. 469 (MO!). Para: 18 km from Tucurui, I Nov.
1981, I). C. Daly et al. 1171 (MO!); Parauapebas, Re-
serva Biologiea da Serra dos Carajas, 20 Nov. 1991. G.
dos Santos et al. 1 69 (MO!); Parauapebas, Reserva Biol-
dgica da Serra dos Carajas, 21 Nov. 1991. G. dos Santos
et al. 191 (MO!), 192 (MO!), 193 (MO!). Rondonia: Por¬
to Velho— Cuiaba highway, 10 km S of Ariquemes, 14
Aug. 1968, F. Farera A B. /,. Wrigley 7037 (INPA, MG,

52

Novon

MO!); Mindanao Taboca at Massangana, 1 I Oct. 1979,
J. L. Zarticchi et al. 2674 (JBSI), MO!).

3. M« ‘inora sastrei A. H. Gentry & Hauk, sp.
nov. TYPE; Colombia. Amazonas: Rio Igara-
Parana, La Chorrera, reste de fbret primaire,
22 Sep. 1973, Sastre 229.1 (holotype, COL;
isotypes, MO, P).

Frutex scandens. ramulis teretibus glabris sine conso-
cibus glandularibus in midis interpetiolaribus. Folia op¬
posite, pinnate foliolis 3—5 ellipticis vel elliptico-ovatis
scabre pubeseentia; pseudostipulis carentibus. Inflores-
centia raeemosa eompacta axillaris dense pubeseentia
dendroideis flavidis; bracteolis deciduis. Flos calyce tub-
nloso-carnpanulato irregulariter bilabiato apiee glanduloso
dense pubeseente triehomatibus dendroideis flavidis dua-
bus bracteolis glandulosis sieut calyce inflorescentiaque
pubescentibus subtenlus: corolla tubuloso-inlnndibulifor-
mi. A speciebns aliis pubeseentia dendroidea flavida, ca¬
lyce apiee glanduloso differt.

Arbolito 5—8 m <le alto, las ramitas subteretes o
un poeo angulosas, sin seudoestfpulas. Hojas pin-
nadas eon 3—5 loliolos elfpticos a elfptico-ovados,
subagudos a subtruncados en la base (apiee des-
conoeido), 13—30 cm ile largo, 6-14 de ancho, cor-
idceos, bulados, las nervaduras primarias, secun-
darias, y tereiarias conspicuamente impresas por
eneima y prominente por debajo; glabras en la haz,
escabro no-pubescente en el enves, apareeiendo
blanqueeino por la densa pubesceneia, con trico¬
mas largos, simples y ramificados, y trieomas eor-
tos, dendroides. Inflorescencia en racimo corto, ax-
ilar, densamente pubeseente con tricomas
dendroides canela-amarillentos, las flores subten-
didas por braeteolas 5—7 nun de largo, temprana-
mente deeiduas. Caliz tubular-eampanulado, 20—26
mm de largo, 8-15 de ancho, irregularmente bila-
biado, densamente pubeseente con trieomas den¬
droides de color canela y glandulas con forma de
plato cerea del apiee; eorola amarilla, tubular-in-
lundibuliforme o tubular-campanulada, 8-1 1 cm de
largo, 15—25 mm de aneho en la boea del tubo, los
lobos easi 2 cm de largo. Cdpsula desconocida.

Esta espeeie esta muy emparentada con M. cla-
dotricha, pero difiere notablemente por la pubes¬
ceneia mucho mas gruesa, especialmente en el eal-
iz y la superficie inferior de las hojas, las braeteolas
mas pequenas y los loliolos mas bulados. Memora
cladotricha puede tener el nervio medio y los se-
cundarios impresos por la haz y prominente por el
enves, pero la venat ion mds fina es uniformemente
promfnula por debajo y promfnula a subplana por
eneima. El tipo tie M. cladotricha es mas pubes-
cente que la mayorfa de los otros ejemplares del
taxon y, por entle, intermedio entre M. sastrei y la

mayoria de las colecciones ubicadas en M. clado¬
tricha.

The description and discussion of Memora sastrei
provided here were originally written in English by
A. H. Gentry, and then translated into Spanish. Un¬
fortunately, the original English description could
not be located. Gentry visited COL and P during
the period that he wrote the Flora de Colombia
treatment, and he annotated the holotype and pre¬
sumably examined all the isotypes. The duplicates
at MO are only fragmentary and could not be the
basis for the description presented here. Gentry did
not provide a Latin description or diagnosis of M.
sastrei in the Flora de Colombia manuscript; the
Latin description was provided by the author.

Paratypes. COLOMBIA. Amazonas: Rio lgara-Para-
na. La Chorrera, layon conduisant il la Sabana, km 1, 18
Sep. 1973, Sastre 2246 (MO, !’)•

4. Memora veliitinu A. H. Gentry ex Hauk, sp.
nov. TYPE: Brazil. Para: BR 230, Transama-
zonian highway, 188 km W of Altamira, 30
Nov. 1977, 6. T. Prance , .4. S. Silva , M. J.
Balick, A. J. Henderson , B. IL Nelson A R. P.
Bahia P25H91 (holotype, MO; isotypes, NY,
US). Figure 3.

Frutex scandens. ramulis teretibus dense pilosis sine
consocibus glandularibus in nodis interpetiolaribus. Folia
opposita bipinnata foliolis ovato-elliptieis obtusis; pseu¬
dostipulis foliaceis orbiculato-eordatis sparsim pilosis
glandulosis. Inflorescentia raeemosa pendula axillaris;
bracteolis cupulatis. Flos calyce campanulato apiee trun¬
cate minute 5-denticulato extus glanduloso duabus brac¬
teolis foliaceis glandulosis subtenlus; corolla tubuloso-in-
fundibuliformi tubo glabro; tliecis antherarum divaricatis;
ovario oblongo lepidoto; disco pulvinato. A speciebns aliis
ramulis dense pilosis differt.

Lianas ; branches drying yellowish green, not
striate, conspicuously pilose, with interpetiolar
glandular fields lacking and interpetiolar transverse
ridge present, lenticels inconspicuous or absent;
pseudostipules persistent, foliaceous to subfolia-
ceous, elliptic to cordate-orbicular, 4—20 X 3—20
mm, glandular, sparsely pilose with eiliate margins.
Leaves opposite, estipulate, petiolate, 8-17 cm long,
bipinnate with a pair of opposite pinnae and the
terminal pinna modified into a simple tendril, each
pinna imparipinnate with 1—2 sets of opposite sim¬
ple ultimate segments; petioles 2.5— 3.0 cm long,
inconspicuously sulcate, pilose; petiolules 3-7 mm
long, inconspicuously sulcate, pilose; ultimate seg¬
ments entire, 4—9 X 1.5— 4.5 cm, elliptic to ovate,
equilateral, plane, chartaceous, glabrate above ex¬
cept short-pilose along the midrib and proximal
portions of the secondary veins, pilose beneath.

Volume 9, Number 1
1999

Hauk

Memora from South America

53

I' if? ure 3. Memora velutina A. H. Gentry ex Hauk. — A. Leaves and inflorescences. — B. Pseudostipules. — C. Leaflet
pubescence, adaxial and abaxial views. (From Prance et al. P25891 , MO, NY, US.)

apically acute to slightly acuminate, basally obtuse
to rounded, the venation broehidodromous with 4—
6 principal vein pairs, marginally plane to slightly
revolute; joints of the compound leaf conspicuously
enlarged. Inflorescences elongate, axillary racemes
to 10 cm long, unbranched, few- to several-flow¬
ered; rachis and peduncles terete, pilose, and brac-
teate, the bracts linear-triangular, 0.5— 1.0 X 0.5
mm, caducous, eglandular, pilose; pedicels 2—1 mm

long, glabrate to sparsely pilose; bracteoles ovate-
elliptic and cupular, 10—13 X 5—7 mm, persistent,
glandular, glabrate to sparsely pilose with pilose
margins, extending beyond the tip of the calyx.
Flowers ovoid in bud; calyx campanulate, 10-12 X
5-6 mm, eglandular, ecostate, the outer and inner
calyx surfaces glabrate, the margin intact and
flared, apically truncate except for 5 minute teeth;
corolla zygomorphic, tubular-funnelform, yellow,

54

Novon

exserted 35—40 nun beyond the level of the calyx
lip (45 inm total length), 2—3 mm wide at the calyx
mouth, with an inner ring of pilose pubescence 7—
10 mm beyond the corolla tube base; corolla lobes
5 (2 upper and 3 lower), 15 X 12—15 mm, short-
orbicular, the inner and outer surfaces glandular-
lepidote, no glandular fields present at the base of
each lobe; fertile stamens didynamous, ca. 15 or 20
mm long, inserted near the level of the calyx tip,
the single staminode inserted near the level of the
fertile stamens; fertile anthers with two spreading
thecae, glabrous, included; disc 1 mm tall; ovary
4—5 mm long, cylindrical, lepidote; style ca. 25 mm
long; stigma bipartite and included. Fruit unknown.

Distribution. The type and only known collec¬
tion was made on a roadside along the Transama-
zonian highway, L88 km west of Altamira. No in¬
formation is available on elevation or habitat.

Phenology . The type collection was in flower in
November. No collections of fruit are known at this
time.

Memora velutina (Fig. 3) has several unique or
unusual characters, the most conspicuous of which
is its pilose pubescence. The density of trichomes
seen in M. velutina is unusual in Memora , and the
pilose pubescence is unlike the pubescence found
on other Memora species. The combination of a
campanulate calyx and bracteoles exceeding the
calyx is found only in M. velutina and M. longilinea
A. Sampiao, but the species are not similar other¬
wise. Memora mollis and M. racemosa have cam¬
panulate calyces, but both have small bracteoles.
Memora velutina is also distinctive because of its
few-flowered, somewhat delicate inflorescences.
The large, foliaceous, cordate-orbicular pseudostip¬

ules contrast with the elliptic leaflet shape. The
only other Memora species with pseudostipules of
this shape is M. adenophora Sandwith. Affinities to
existing Memora species are not clear, but probably
lie with i M. racemosa , M. longilinea, and M. mollis.
Investigations of fruit and seed type, wood anatomy,
and pollen morphology are needed to further assess
the taxonomic affinities of this species.

Acknowledgments. I thank the Missouri Botan¬
ical Garden and Peter Raven for making A1 Gen¬
try’s collections and databases available. Loans of
material from the curators of US and NY are grate¬
fully acknowledged. I thank William G. D’Arcy and
Roy Gereau for assistance with the Latin descrip¬
tions, Rosa Ortiz-Gentry for helpful comments on
the Spanish text, Kerry Barringer for general com¬
ments on the text, and Jose L. Fern&ndez- Alonso
and Porter P. Lowry for information pertaining to
type specimens. Susan A. Moore provided the il¬
lustrations. This work was supported by the Na¬
tional Science Foundation (grant DEB-9509270).

I .iterature Cited

Gentry, A. II. 1977. ham. 173. Bignoniaceae. Pp. 3— 173 in
G. Hading & It. Sparre, Flora of Ecuador. Opera Bot. 7.

- . 1978. Bignoniaceae. Pp. 245—283 in B. Maguire,

The Botany of the Guyana Highland. Mem. New York
Bot. Gard. 29.

- . 1982. Bignoniaceae. Pp. 7—433 in /. Luces de

Febres & J. Steyermark. Flora de Venezuela. Instituto
Nacional de Parques, Direccidn de Investigaciones
Biologieas, Vol. 8 (4).

- . 1997. Bignoniaceae. Pp. 403-491 in J. A. Stey¬
ermark, I’. F. Berry A B. k. Holst (general editors).
Flora of the Venezuelan Guayana, Vol. 3. Missouri Bo¬
tanical Garden, St. Louis.

- A A. S. Tomb. 1979 1 1980). Taxonomic implica¬
tions of Bignoniaceae palynology. Ann. Missouri Bot.
Gard. 66: 756— 777.

New Species of Xylopia and Uvaria (Annonaceae) from Tanzania

David M. Johnson

Dept, of Botany-Microbiology, Ohio Wesleyan University, Delaware, Ohio 43015, U.S.A.

Leonard B. Mwasumbi and Frank M. Mbago
Dept, of Botany, University of Dar es Salaam, P.0. Box 35060, Dar es Salaam, Tanzania

Abstract. Two new species of Annonaceae from
the threatened lowland coastal forests of Tanzania
are described and illustrated. Xylopia mwasumbii
D. M. Johnson has distinctive flattened monocarps,
unique among its African congeners. Uvaria pug-
uensis D. M. Johnson is a small-flowered species
belonging to a complex including, in East Africa,
U. angolensis, U. lucida, and U. tanzaniae. A key
to the taxa of the U. angolensis group in East Africa
is provided.

Annonaceae are especially diverse in the low¬
land coastal forests of Tanzania and Kenya, with six
genera and 35 species endemic to these forests
(Verdeourt, 1971, 1986; Verdcourt & Mwasumbi,
1988; Vollesen, 1980). At the same time that the
biological diversity of these coastal forests is being
recognized, this rare forest formation is disappear¬
ing due to pressure from land development and
charcoal production (Burgess et ah, 1992; Mwa¬
sumbi et ah, 1994).

Pugu Forest Reserve near Dar es Salaam, Tan¬
zania, is the best-studied coastal forest in Tanzania
(Burgess et ah, 1992), yet we have just recently
gathered sufficient material to describe two new
species of Annonaceae, one in Xylopia and one in
Uvaria, from this small reserve and nearby areas.
Both spec ies occur in forested uplands with a dis¬
tinctive suite of associated species, including Ba-
phia puguensis Brummitt, Hugonia castaneifolia
Engler, Hymenaea verrucosa Gaertner, Hymenocar-
dia ulmoides Oliver, Landolphia kirkii Dyer, La-
siodiscus holtzii Engler, Manilkara sulcata (Engler)
Dubard, Monanthotaxis fornicata (Baillon) Verd¬
court, Nesogordonia holtzii (Engler) Capuron, Sco-
rodophloeus fischeri (Taubert) J. Leon, Suregada
zanzibariensis Baillon, and Uvaria pandensis Verd¬
court, as well as species of Croton, Haplocoelopsis,
Ochna, Megalochlamys, Rinorea, and Tessmannia.

Verdcourt (1971), in his treatment of Xylopia for
the Flora of Tropical East Africa, first drew atten¬
tion to the specimen Semsei 3704 from the Pugu
Forest Reserve, identifying it as Xylopia “Species

B" and commenting, “The material is scarcely ad¬
equate for description and only one of the two spec¬
imens seen bears flowers; both bear globose arti¬
choke-like galls covered with conical tubercles
which appear to be derived from the flowers.” The
name Xylopia “Species B " has, in the meantime,
crept into East African floristic literature (e.g.,
Hawthorne, 1993), but it is now possible to name
and describe this species in detail.

Xylopia mwasumbii D. M. Johnson, sp. nov.
TYPE: Tanzania. Coast Region: Kisarawe Dis¬
trict, Pugu Forest Reserve, along N road 0.5
km E of brick factory, 6°52'S, 39°06'E, 200
m, 16 Feb. 1996, Johnson & Ndangalasi 1834
(holotype, OWU; isotypes, DSM, K, MO). Fig¬
ure 1A-H.

Inter africanas Xylopia e species petalis lanceolatis oli-
vaceis, staminibus non-capitatis, staminodiis nullis, car-
pellis duobus vel tribus, stigmatibus glabris oblongo-fal-
catis, toro leviter tantum concavo, el monocarpiis
foliiformibus falcatis venosis compressis distincta.

Tree 4-9 m tall, DBH up to 8 cm, often with
multiple trunks, the principal trunk often arching
rather than erect; bark white with gray and orange
mottling, exfoliating in places; slash brown on yel¬
low (ex Hawthorne). Nodes occasionally with two or
more axillary branches. Twigs 0.7— 2.1 mm diam.,
gray, longitudinally wrinkled, with epidermis soon
exfoliating, sparsely lenticellate, sparsely ap-
pressed-pubescent to glabrate, the simple hairs
0.2-0.4 mm long. Lamina of larger leaves 4.6— 7.7
cm long, 2.4— 4.3 cm wide, paler abaxially, subco-
riaceous, broadly elliptic to elliptic, occasionally
ovate, elliptic-ovate, or obovate-elliptic; base cu-
neate, decurrent on petiole; apex 4—8 mm long,
blunt-acuminate, or occasionally obtuse and emar-
ginate; lamina with adaxial surface glabrous, ab-
axial surface sparsely sericeous; midrib impressed
to plane adaxially, abaxially raised and keeled; sec¬
ondary veins 9-1 I per side, departing at 50—60°
from the midrib, weakly brochidodromous, anasto¬
mosing 1—2 mm from margin; secondary and high-

Novon 9: 55-60. 1999.

56

No von

Figure I . Xylopia mwasumbii. — A. Flower at anthesis. face view. — B. Flower at anthesis, lateral view. — C. Mono-
carp, lateral view. — D. Habit, with flower bud. — E. Inner petal, adaxial view. — F. Outer petal, adaxial view. — G.
Stamen, abaxial view. — H. Longitudinal section of flower, with petals and stamens removed, showing two carpels seated
in concavity of torus. C, I) from Johnson & Mdangalasi 1884 (OWU); A-B. E-H from Johnson I928A (OWU). Note:
A-C and E-H were drawn from fluid-preserved material and their dimensions are consequently ca. 20% greater than
those given in the species description.

Volume 9, Number 1
1999

Johnson et al.

New Annonaceae from Tanzania

57

er-order veins indistinct, raised adaxially, slightly
raised abaxially. Petiole 4.5—8 mm long, 0. 9-1.2
mm wide, semiterete, flattened to canaliculate
adaxially, glabrous to sparsely appressed-pubes-
cent. Inflorescences of 1—3 flowers, the pedicels
arising either from the axils of leaves or from the
axils of bracts on a common peduncle 3-4 mm
long; pedicels articulate at base, 2. 5-4. 6 mm long,
0.8— 0.9 mm thick, bibracteate, sparsely pubescent
to glabrate; bracts 0. 8-1.1 mm long, quadrate,
semicircular, or deltate, occasionally tearing down
the center as the inflorescence enlarges. Buds lin¬
ear-lanceolate, apex obtuse. Calyx 1.4— 2.2 mm
long, 2.5— 3.3 mm diam., cup-shaped, coriaceous,
sparsely sericeous abaxially; lobes 0.6— 1.5 mm
long, 1.5— 2.2 mm wide, broadly triangular, obtuse
to acute at apex. Corolla olive-yellow in vivo, co¬
riaceous to slightly fleshy; outer petals spreading at
anthesis, 8—10 mm long, 2-3 mm wide, linear-lan¬
ceolate to narrowly triangular, acute at apex, eu-
neate at base, appressed-pubescent adaxially, se¬
riceous abaxially; inner petals erect al anthesis,
5.7-7. 5 mm long, 1.8— 2.2 mm wide, linear-lanceolate,
basal Vi concave adaxially, keeled abaxially, gla¬
brous except for a few hairs along abaxial keel.
Stamens pale orange-yellow in vivo, ca. 40, 1.5— 1.8
mm long, narrowly oblong, glabrous; anthers sep¬
tate at anthesis; apex of connective rudimentary,
not exceeding anther thecae, coriaceous to fleshy;
filament ca. 0.7 mm long; staminodes absent. Car¬
pels 2—3; ovaries 1.5— 1.6 mm long, narrowly ob¬
long, sericeous; ovules 4 in a single row; stigmas
white in vivo, 0.7-1. 1 mm long, lanceolate-falcate,
acute, longitudinally grooved ventrally, glabrous.
Torus 1.5 mm diam., elevated above the calyx with
a shallow central concavity in which the ovaries are
seated, glabrous. Fruit of 1—3 monocarps borne on
a pedicel 4.5—6 mm long, 1.1— 1.4 mm thick,
sparsely pubescent to glabrate, with bracts and se¬
pals persistent; torus of fruit ca. 2 mm diam.; mon¬
ocarps near maturity yellow-green with green stipe
and venation in vivo, probably dehiscent, 2—2.7 cm
long, 1.1 cm wide, oblong, compressed laterally, ve¬
nation of pericarp raised and distinct, finely ver-
rucose, sparsely pubescent to glabrate; stipe 5—6.5
mm long, 1.5— 1.6 mm wide; apex truncate, with an
oblique rnucro 1.7—2 mm long; wall ca. 0.5 mm
thick. Seeds 2-4, at 60—90° to long axis of mono¬
carp, 9 mm long, 6 mm wide, ellipsoid, elliptic in
cross section, smooth; mieropyle circular, flat, 2.2-
2.4 mm diam.; aril cupular and forming a complete
ring around mieropyle, 5 mm diam., 2.2 mm deep,
white, fleshy, readily detached from seed.

Distribution and phenology. Known only from
dry evergreen forest on several small ridgetops and

plateaus near the coast of east-central Tanzania, at
elevations of 0—300 m. Specimens with flowers have
been collected from February through May, and
those with fruits from February through June. No
definable peak of flowering was observed, but rather
a few flowers at anthesis were present at any one
time on a given tree. Similarly, only a very few fruits
could be found on an individual tree at any one time.
Vegetative growth seems to have been stimulated by
the short rains of November— December, with pro¬
duction of flower buds following shortly thereafter.

Vernacular names and uses. Mgwaza, mgwaza
dume (Zararno). Used locally for tool handles and
building poles.

I am pleased to name this distinctive species for
my co-author, Leonard Mwasumbi, authority on the
flora of Tanzania and herbarium superintendent at
the University of Dar es Salaam, who taught me
much about the flora of East Africa and guided me
to localities for this species.

In Verdcourt’s (1971) key to East African species
of Xylopia, X. mwasumbii will key to Xylopia odor-
atissima, a species of the dry interior with pro¬
nounced pubescence on twigs, leaves, pedicels, and
fruits, longer and more narrow petals, and mono¬
carps that are irregularly oblong and not com¬
pressed. Xylopia mwasumbii does not, however,
closely resemble this or any other East African spe¬
cies of Xylopia, with its subcoriaceous elliptic
leaves, olive-yellow flowers, absence of prolonged
anther connectives, lack of staminodes, small num¬
ber of carpels, and flattened monocarps. It is sym-
patric with two other species of Xylopia, X. parvi-
fiora (A. Richard) Bentham and X. arenaria Engler.
The former diflers in being a large tree, the flowers
of which have linear petals that flare into a saccate
base. Xylopia arenaria is a small tree, more similar
in habit, but its flowers are pale orange-yellow and
the leaves are rounded at the base and chartaceous.
Petal orientation at anthesis is distinctive in each
of the three species: in X. mwasumbii the inner
petals are erect and the outer petals are wide-
spreading, in X. parviflora all six petals curve out¬
ward from the base and then inward at the apices,
and in X. arenaria both series of petals are also
erect, but the apex of each squarrose inner petal
emerges through a gap between two outer petals.

The dry evergreen coastal forests where X. mwa¬
sumbii grows are both poorly known and very lim¬
ited in extent. Within this rare forest formation the
species may be commoner than is currently rec¬
ognized: its green-tinted flowers and fruits are in¬
conspicuous, and its principal flowering and fruit¬
ing occur during the long wet season. Also, it has
a strong resemblance to species of Diospyros and.

58

Novon

in herbarium material, to other Annonaceae genera
such as Sphaerocoryne and may thus he misiden-
tified in collections. The flower-galls described by
Verdcourt seem to be a frequent feature of the trees,
and similar galls are found in other African species
of Xylopia as well.

Paratypes. TANZANIA. Coast Region: I’ande. Haw¬
thorne 1360 (K), Hawthorne 1714 (K, fule B. Verdcourt);
Pande, W edge, Hawthorne 1469 (K); Kisiju. by the sea.
Hawthorne 1790 (DSM); Kisarawe District. Engn forest
Reserve, ridges along N road between Pngu railway station
and brick works, 6°52'30"S, 39°06'E, Johnson <V Mwa-
sumbi 1899 (DSM, OWU); Kisarawe District. Engn Forest
Reserve, along N road 0.5 km E of brick factory, 6°52’S.
39°06'E, Johnson 1920 (DSM, OW U), Johnson 1928A
(OWU, spirit collection only); Kisarawe District. Pugn
Forest Reserve, ridgetop over road tunnel, 6°52'30"S,
39°05'30"E, Johnson & Mwasumhi 1936 (DSM. OW U);
Kisarawe District, Pngu f orest Reserve, S of Dar— Kisar¬
awe road, ridgetop near Mwakanga railway station, 6°55'S.
39°06'E, Johnson & Ndangalasi 1948 (DSM. OWU); Ba-
gamoyo District, Zaraninge Forest Reserve, 54 km E of
llwy., 6°04— 13'S, 38°35— 42'E, Johnson & Mbago 1963
(DSM, OWU); Kisarawe District, Pngu forest Reserve,
bus roundabout area ca. 4 km E of Kisarawe, 06°53'30"
S, 39°06'E, Johnson 19648 (DSM. OWU); Kisarawe Dis¬
trict, Kazimzumbwi Ruvu South [sic], Magogo 6 18 (NUT.
FED); Kisarawe District, Banda Forest Reserve near
Mfyoza village, Ruffo 301 (EA, K, NUT. TED); Kisarawe
District, Pugu Forest Reserve, Semsei 3704 (EA, K. TED);
Kisarawe District, Banda forest Reserve, Shabani 471
(EA, K, TED); Pande Forest Reserve, 25 km WNW of Dar
|es Salaam], Wingfield 33 1 1 (DSM. EA, K).

Uvaria L. is the largest genus of Annonaceae in
the tropical East African flora, and many of its spe¬
cies groups are taxonomically difficult. Still, it is
clear that there are undescribed species among the
taxa represented. One of these came to the notice
of L. Mwasumhi in the course of collaborative work
with the research group studying Annonaceae sec¬
ondary compounds in the Dept, of Chemistry at the
University of Dar es Salaam under the leadership
of M. H. H. Nkunya. Almost all of the collections
of this species come from the Pugu Forest Reserve,
where the plant is fairly common in the same hab¬
itats in which Xylopia mwasumbii occurs, hut the
plant has also been reported from the Pande and
Kiono Forest Reserves as “ Uvaria sp. nov.” (Bur¬
gess et ah, 1992; Mwasumhi et ah, 1994).

Uvaria puguensis D. M. Johnson, sp. nov. TYPE:
Tanzania. Coast Region: Kisarawe District,
Pugu Forest Reserve, ridge trail from N road,
6°52'S, 39°06'F, 200 m, 24 Apr. 1996, John¬
son 1928 (holotype, OWU; isotypes, DSM,
MO). Figures 2A, C-E, G, I, J.

Uvariae lucidae affinis, sed lamina basi distincte sub-
cordata, pedicellis 3—5 mm longis, ealyce 3-4 mm longo.

petalis ovatis, 5—6 mm longis, flavo-viridibus vel pallide
aurantiaco-flavis, staminibus 2.2— 2.7 mm longis diversa.

Woody liana climbing to 6 m. Twigs 0.9—2 mm
diam., longitudinally wrinkled, gray to black, at
first ferruginous-pubescent, the stellate hairs 0.1—
0.6 mm diam. with the tips of the stellae erect or
oppressed, at length glabrate. Lamina of larger
leaves 8.8—12.4 cm long, 4.3— 6.6 cm wide, shiny
and gray-green adaxially, dull and greenish yellow
abaxially, subcoriaceous, elliptic, obovate, or ob-
lanceolate; base distinctly subcordate, occasionally
rounded; apex 6—10 mm long, acuminate; expanded
lamina with adaxial surface glabrous except lor a
few hairs on the midrib, abaxial surface with scat¬
tered stellate hairs; midrib impressed adaxially,
raised abaxially; secondary veins 9—1 1 per side,
departing at 50-70° from the midrib, weakly bro-
chidodromous, slightly impressed adaxially, slightly
raised abaxially; higher-order veins indistinct on
both surfaces. Petiole 6-10 mm long, 1.5-2 mm
wide, terete, pubescent. Inflorescences internodal,
usually closer to the distal nodes, 1— 2-flowered. In¬
florescence axis 3—5 mm long, 1.3— 1.8 mm thick,
covered by yellowish rusty stellate pubescence,
with two clasping bracts ca. 2 mm long; buds glo¬
bose. Calyx connate, tearing longitudinally as flow¬
er opens, 3-4 mm long, 5 mm diam., cup-shaped,
coriaceous, pubescent abaxially. Corolla slightly
fleshy; petals light green in vivo and erect in pis¬
tillate phase of anthesis, yellow-green to pale or¬
ange-yellow in vivo and strongly recurved in sta-
minate phase; inner and outer petals subequal, 5—
6 mm long, 3.6-4 mm wide, ovate, glabrous anil
somewhat verrucose adaxially, pubescent abaxially
and on margins. Stamens mustard-yellow at anthe¬
sis in vivo , 30—40, 2.2— 2.7 mm long, narrowly ob¬
long, glabrous except for the puberulous connective
apex, which is 0.3— 0.5 mm long and quadrate to
ovate; anthers latrorse; filament short or lacking.
Carpels 6—8; ovaries 3.4— 4.1 mm long, oblong, lon¬
gitudinally sulcate on ventral surface, pubescent;
ovules 8—9 in a single row; stigmas 0.7— 0.8 mm
long, obconic to quadrate, pubescent on sides. To¬
rus 3 mm diam., Hat, umbonate in center, with a
few scattered hairs. Fruit of 5—8 monocarps borne
on a pedicel 5 mm long, 1.9 mm thick, covered
with dense appressed yellowish gray stellate pu¬
bescence; bracts persistent; torus of fruit 3 mm
diam.; monocarps (immature) 1.7 cm long, 0.7— 0.8
cm wide, oblong, slightly constricted between the
seeds, subsessile, pubescent; apex rounded, with a
small beak 1.2 mm long; wall ca. 0.5 mm thick.
Seeds 1-4, at 90° to long axis of monocarp.

Distribution and phenology. Known from ridge-
tops and slopes in several areas of dry evergreen

Volume 9, Number 1
1999

Johnson et al.

New Annonaceae from Tanzania

59

anthesis, lateral view. — G. Stamen, abaxial view. — I. Carpel. — J. Monocarp, lateral view. It. F, H: Uvaria luriila. —
B. Detail of leaf lamina base. — F. Flower in staminate phase of anthesis, face view. — II. Stamen, abaxial view. I) &
E from Johnson & N dun galas i 1949 (OWU), A, C. G, I, & .1 from Johnson 1928 (OWU), B, F, & II all from Johnson
& Murray 1887 (OWU). Note: D— F and G— I were drawn from fluid-preserved material and their dimensions are
consequently ca. 20% greater than those given in the species description.

60

Novon

coastal forest in east-central Tanzania, at elevations
up to 300 m, where it has been found in flower
from February through June, and fruiting in April.

Uvaria puguensis belongs to a group of species
that includes, in East Africa, 11. angolensis, U. lu¬
cida, and U. tanzaniae. The group is characterized
by connate sepals, which form a cup that is torn
open by the developing corolla, and by petals that
become completely revolute during anthesis. Dis¬
tinctions among these taxa are presented in the key
below. The only one of these species known to be
sympatric with U. puguensis is U. lucida.

Kky to ink Fast Akrican Taxa ok thk Uvaria a\go-
u:\sis Groki* (adaitki) kkom Vkrdcoijrt, 1971, 1986;

THK AI.PHANUMKRIC COOKS KOI.I.OWINO KACH SPKCIKS
NAMK REFKK TO THK Fl.ORA OK TROPICAL K AST AKR1CA
KTORISTIC l>IVISIOI\S. \S CITKI) IN Vk.RDCOI HI. 1971)

la. Petals 12-2.7 mm long, uniformly tomentose on
both surfaces; carpels/monocarps typically 20—

35.

2a. Leaf bases distinctly cordate; leaf strongly

hirsute beneath .

. U. tanzaniae Verdeourt (T3, TO)

2b. Leaf bases broadly cuneate, rounded, or
slightly cordate; leaf sparsely pubescent be¬
neath .

. . IJ. angolensis Oliver (L I— 4; Tl, 4. 7; Sierra
Leone to Cameroon. Congo, Sudan. Zambia.

Angola)

lb. Petals 5—13 mm long, glabrous or with hairs
sparse toward the apex adaxially, pubescent
abaxially; carpels/monocarps 5—20.

3a. Leaf base cuneate or rounded (Fig. 2B),
rarely subcordate; inflorescence axis 5—10
mm long; petals 7— 13(— 20) mm long, green
or grayish brown, oblong-ovate or rounded-

obovate; stamens 2.5-4 mm long .

. U. lucida Bentham

4a. Leaves mostly relatively broad, obtuse
to shortly acuminate at the apex, drying
pale green; indumentum of calyx and
monocarps mostly pale brown or orange-
brown; calyx, petals, stamens, and mon¬
ocarps with dimensions toward ihe low¬
er of the limits given .

11. lucida Bentham subsp. lucida (K4, 7;

T3, 6, 8; /,;? P)

4b. Leaves relatively narrower, obtuse to
acuminate (acuminate in Fast African
material), drying a darker greet); indu¬
mentum of calyx and monocarps mostly
dark brown; calyx, petals, stamens, and
monocarps with dimensions toward the

upper of the limits given .

. . II. lucida Bentham subsp. circus (N. F.
Brown) Verdeourt (T6. Malawi. Mozambique.

Zimbabwe, South Africa)
3b. Leal base distinctly subcordate; inflores¬
cence axis 3—5 mm long; petals 5—6 mm
long, yellow-green to pale orange-yellow.

ovate; stamens 2.2— 2.7 mm long .

. U. puguensis I). M. Johnson (T6)

This species is a locally common component of
the forest flora in the Pugu Hills, more common
there than its close relative 11. lucida. In our ex¬
perience, the flower hurls frequently showed insect
feeding damage, so that few llowers reached anthe¬
sis; fruits were consequently scarce.

Para types. TANZANIA. Coast Region: kisarawe Dis¬
trict. Pugu Forest deserve, ridges along N road between
Pugu Railway Station and brick works, 6°52'30"S,
39°06'30"F, Johnson & Mwasumbi 1898 (DSM. OWL);
Kisarawe District. Pugu Forest Reserve, along N road be¬
tween Pugu railway station and brick works, 6°52'30"S.
39°()0'P. Johnson 1916 (DSM. OWL)); Kisarawe District.
Pugu Forest Reserve, S of I )ar-Kisarawe road, ridgetop near
Mwakanga railway station, 6°55'S, 39°06'F, Johnson &
Ndangalasi 1949 (DSM. OWL); Pande Forest Reserve,
Mwasumbi 14705 (DSM); Pugu Forest Reserve. Mwasumbi
12582 (DSM).

Acknowledgments. Financial and/or logistical as¬
sistance for the research reported here was received
from the following institutions and individuals: Bot¬
any, Chemistry, & Zoology Departments, University
of Dar es Salaam; Tanzania Commission on Science
& Technology; Fulbright African Regional Research
Program. DMJ also thanks Ohio Wesleyan University
lor providing the sabbatical leave during which this
study was made. Particular thanks are given to H.
J. Ndangalasi, who served as field assistant and
guide in the Pugu Forest Reserve area. Comments
by L. Chatrou, N. Murray, B. Verdeourt. R. Bameby,
and an anonymous reviewer improved the manu¬
script. The following herbaria made collections
available for study: DSM, EA, K. MO, NHT, OWU,
TFD; we particularly thank W. Mziray at NHT and
M. Kisena Mabula at TFD lor their assistance.

I .itcrature Cited

Burgess, N. I).. L. B. Mwasumbi, W. J. Hawthorne, A.
Dickinson & R. A. Doggett. 1992. Preliminary assess¬
ment of the distribution, status and biological impor¬
tance ol coastal forests in Tanzania. Biological Conser¬
vation 62: 20.5-218.

Hawthorne. \\. I). 1993. Fast African coastal forest bot¬
any. Pp. 57-99 in J. C. Lovett & S. K. Wasser (editors).
Biogeography and Ecology of the Rain Forests of Fast¬
en) Africa. Cambridge Lniv. Press, Cambridge, L.k.
Mwasumbi. L. B„ N. I). Burgess & G. P. Clarke. 1994.
Vegetation of Pande and Kiono Coastal forests. Tanza¬
nia. Vegetatio 113: 71-81.

Verdeourt, B. 1971. Annonaceae. In F. Milne-Redhead &
R. M. Polhill (editors). Flora of Tropical Fast Africa.
Crown Agents for Oversea Governments and Adminis¬
trations, London.

- . 1986. New taxa of Fast African Annonaceae.

Kew Bull. 41: 287-297.

- & L. B. Mwasumbi. 1988. A new species of Uvar¬
ia (Annonaceae) from Tanzania. Kew Bull. 43: 99-101.
Vollesen. K. 1980. Notes on Annonaceae from Tanzania.
But. Not. 1.33: 53—62.

Calathea singularis and Strom ant he palustris, Two New Species of

Neotropical Marantaceae

Helen Kennedy

Herbarium, Botany Department, University of British Columbia, Vancouver, B. C. V6T 1Z4,

Canada

ABSTRACT. Calathea singularis from southeastern
Brazil and Stromanthe palustris from Costa Rica are
described as new. Calathea singularis , vegetatively
unique within the Marantaceae, has a pulvinus of two
separate parts rather than the pulvinar cells contin¬
uous and encircling the uppermost portion of the pet¬
iole. Stromanthe palustris, endemic to the Atlantic
lowlands of Costa Rica, is described for inclusion in
A Manual to the Plants of Costa Rica treatment.

Calathea singularis H. Kennedy, sp. nov. TYPE:
Brazil. Espirito Santo: Munic. de Sao Mateus,
Reserva Biologica de Sooretama, Lagoa do
Macucu, umbrofila crescendo em local urnido,
ca. 40 m, 16 Mar. 1972, D. Sucre 8698 (ho-
lotype, RB 157123; isotypes, K, RB). Figure 1.

Species pulvino baud integro autem bi-partito a con-
generibus et insupra confamiliis diversa.

Rosulate, shortly caulescent herb, 30—70 cm
high, bearing 4—15 leaves, 3-6 basal and 1 cauline
on main shoot, lateral shoots develop in axils of
basal leaves bearing (0)1—3 leaves at the base and
1 ( — 3) cauline. Plants deciduous, dying back to the
rhizome during the dry season; the roots bearing
terminal, tuber-like swellings. Leaves homotropic.
Leal blade lightly pleated, herbaceous, elliptic to
broadly elliptic, apex rounded with short acumen,
base rounded to obtuse; 23-47 X 13-29 cm. Leal
blade above with stomates present, grass-green,
glabrous, midrib likewise; leaf surface below glau¬
cous silvery green or faintly tinged purple, gla¬
brous, midrib pale green, V-shaped, glabrous. Pul¬
vinar region narrowly triangular in cross section,
pulvinar cells present only in a small area between
the base of the blade and t he sheath and as a nar¬
row band on the very back shortly extending along
the midrib, green with pulvinar cells brownish, gla¬
brous, region between blade and sheath 0.2-0.4
cm, total length of area of pulvinar cells 0.9-1. 7
cm, depth 0.3-0. 5 cm. Petiole lacking. Leaf sheath
auriculate, the central portion triangular in cross
section, the sides broadly spreading above to
strongly reflexed at base, herbaceous, light green.

often reddish purple at very base of subtending
leal, 14-23 cm long in subtending leaf, others 23-
34 cm long. Stem green, internode below subtend¬
ing leal 6-20 cm long. Inflorescence terminal on
main shoot, imbricate, ovoid, apex acute (in side
view), 8—14 X 3.5—5 cm, additional inflorescences
terminating the shoots produced in the axils of the
basal leaves. Peduncle green, occasionally tinged
pink at apex, minutely pilose (14X), 3.5— 7(— 9) cm
long. Bracts 36^44, spirally arranged, herbaceous,
lowermost bract reniform, obtuse with short acu¬
men, bract shape changing from depressed ovate to
depressed elliptic in lower bracts to broadly ob-
ovate in upper ones, apex obtuse to 90°, occasion¬
ally with slight acumen, 3. 1 — 4.2 X 3.5— 5.8 cm,
bracts 10 and higher less than 4.3 cm wide; each
bract subtends 2-4 flower pairs. Outer surface of
bracts green, whitish at base in upper ones, gla¬
brous; inner surface lighter green, glabrous. Biear-
inate prophyll membranaceous, broadly ovate, ob¬
tuse, translucent faint green to transparent,
glabrous, 1.8— 2.2 X 1.4— 2.1 cm, 0.7—1 cm wide,
carina to carina. Secondary bract membranaceous,
ovate, acute, translucent taint green, glabrous, 1.5—
1.8 X 0.8—1 cm. Bracteoles 1 per flower pair, mem¬
branous medial, carinate, transparent, glabrous,
0.3— 0.9 X 0.2-0. 3 cm long, occasionally reduced
or absent in first flower pair. Sepals membrana¬
ceous, ovate-elliptic, obtuse, translucent white, gla¬
brous, 5—6 X 3.5^1 mm. Corolla 43^47 mm long,
tube cream above, white below, glabrous, 27-30 X
2.5 mm with slit to 10 mm long; corolla lobes sub¬
equal, narrowly elliptic to oblong, obtuse to acute,
pale cream-yellow, glabrous, 12-16 X 4-7 mm.
Outer staminode broadly obovate to elliptic, emar-
ginate, pale cream-yellow, 11-15 X 12-13 mm.
Callose staminode oblong, apex acute, the apical 2
mm petaloid, the rest callose, pale cream-yellow,
9—10 X ca. 4 mm. Cucullate staminode pale cream-
yellow, ca. 6 X 4 mm. Stamen with lateral petaloid
appendage to 1.5 mm wide. Style and stigma cream,
the back ol the style turns brown after tripping.
Ovary smooth, pink, glabrous, ca. 2 X 2 mm. Cap¬
sule unknown.

Novon 9: 61-65. 1999.

62

Novon

Calathea singularis belongs to Calathea sect.
Breviscapus Bentham, by virtue of its several basal
leaves, the inflorescence borne on the leafy shoot,
and spirally arranged bracts. It is distinguished

from all other species ol Marantaceae by the unique
pulvinus structure of two distinct regions ol pulvi-
nar cells rather than a single continuous cylindric
band of pulvinar cells surrounding the uppermost

Volume 9, Number 1
1999

Kennedy

Neotropical Marantaceae

63

Figure 2. Stromanthe paluslris H. Kennedy. — A. Habit. — B. Inflorescence. — C. Flower on upper leaf surface. —
I). Flower, face view.

portion of the petiole. It is most closely related to
Calathea albo-vaginata (K. Koch) K. Schumann,
sharing the deciduous habit, reflexed sheath mar¬
gins, and lack of a proper petiole. Calathea sin-
gularis differs from C. albo-vaginata in the green
vs. white margin of the leaf sheath and the bracts
erect, not reflexed at the apex.

Ihiratypes. Bit A/ll.. Espfrito Santo: Mimic. I. inhar¬
es, Reserva Florestal da Cia. Vale do Rio Doce, Est. Gav-
ea. aril. \-2. km 20.100, erva a beira da estrada, D. A.
Colli 1026 (CY RD). Cultivated. Jardim Botanico do Rio

de Janeiro, 1954, O'. Barroso s.n. (RR 88428): Jardim Bo-
tanico do Rio de Janeiro, without collector or date (RB
971 10).

Stromanthe paluslris II. Kennedy, sp. nov.
TYPE: Costa Rica. Heredia: Finca La Selva,
the OTS Field Station on the Rfo Puerto Viejo
just E of its junction with the Rfo Sarapiquf,
along Loop Trail on S side of Quebrada El Sal¬
to, ea. 6 m from trail, ca. 100 m, Aug. 1978,
B. Kirchoff 6578 (holotype, CR; isotype,
DUKE). Figure 2.

64

Novon

Aliis speeiebus Americae centralis combinatione floruni
purpureorum, bractearum viridium et foliorum basalium
praesentium bene distincta.

Caulescent, branched herb, 1.3-2. 1 m high,
bearing 3-5 basal leaves and 2-3 cauline on main
stem, additional shoots often produced in the axils
of the cauline leaves. Cataphylls stiff, coriaceous,
narrowly ovate, apiculate, dark green tinged with
purple; tomentose to subvillous basally, the hairs
borne on a raised, wart-like, cushion of cells lack¬
ing purple pigmentation; up to 53 cm long. Leaves
antitropic. Leal blade herbaceous, ovate-oblong,
apex rounded with eccentric acumen, base un¬
equal, obtuse, often very shortly, abruptly, attenu¬
ate; 15-19 X 7-10.5 cm in subtending leal, 26-
62 X 11—20 cm in others. Leal blade above shiny
dark green, glabrous, midrib dark olive-green, pi¬
lose along each side at junction with blade; leal
surface below light green, glabrous except pilose
along margin at apex, midrib olive-green, occasion¬
ally tinged with purple, pilose along sides, more
densely so at the base. Pulvinus broadly elliptic in
cross section, dark olive-green, tomentose along the
front, the rest glabrous, articulate, ca. 2X wider
than petiole, 0.9-4. 8 cm long. Petiole dark green,
tomentose to villous toward base, somewhat rough¬
ened with minute irregular warts, bearing a slight
groove, (0)20—77 cm long, usually absent in up¬
permost cauline leaves. Leaf sheath with stiff, pa¬
pery margins, not auriculate, deep green, lighter at
margin, roughened, villous, hairs golden brown to
3 mm long; sheath ol uppermost leaf 5—7 cm long,
others 15—80 cm long. Base of leaf sheath swollen,
whitish yellow. Stem green, slightly roughened,
subvillous, hairs falling in age, the intemode be¬
tween the basal leaves and lowermost cauline leal,
60-110 cm long. Inflorescences terminal and lat¬
eral, several per shoot, imbricate initially, spread¬
ing with age, ovate, laterally compressed, 5—8.7 X
2.5— 3.3 cm. Peduncle light green to yellow-green,
sparsely pilose, 1.5—13 cm long, longer in the axial
inflorescences than the terminal ones. Bracts de¬
ciduous, distichous, 8-19 (generally only 6-12
present at any one time during the main flowering
period) dorsiventrally oriented, conduplicately fold¬
ed, very broadly elliptic to subround, apex rounded,
occasionally with a very short acumen, 1.4— 2.1 X
1.1— 1.7 cm. Bracts herbaceous, chartreuse, gla¬
brous except for a tuft of hairs at apex; inner sur¬
face shiny light green, each subtending up to 2 or
more flower pairs. Bicarinate prophyll membrana¬
ceous, elliptic, apex obtuse, translucent light char¬
treuse, carina sparsely pilose at apex, the rest gla¬
brous, ca. 0.7 X 1 cm. Sepals herbaceous, obovate
to elliptic, obtuse, translucent chartreuse, glabrous.

13.5—16 X 5—6 mm. Corolla tube white, glabrous,
ca. 3 mm long; corolla lobes subequal, elliptic, ob¬
tuse, white, pilose at apex, 7—8 X 4-5 mm. Outer
staminodes oblong, rounded with minute acumen,
basal % white, deflexed apical Vs purple, ca. 5 mm
wide. Callose staminode petaloid apically, rounded,
reddish purple at apex, white basally, ca. 9 X 6.5
mm. Cueullate staminode white, ca. 5 mm long.
Anther light brown, 2 mm long. Style and stigma
white. Ovary white, densely sericeous, hairs golden
brown to 2 mm long, 2.5 X 2 mm. Capsule un¬
known.

Stromanthe palustris and S. tonchat (Aublet) Eich-
ler are the only species of Stromanthe in Central
America with purple flowers. Stromanthe palustris is
readily distinguished from S. tonchat by the several
basal and cauline leaves vs. strictly cauline leaves,
the densely villous vs. sparsely appressed pilose leaf
sheaths (hairs obvious to the naked eye vs. hairs only
visible with lens or microscope), and the longer pul¬
vinus (0.9—1. 5 cm vs. 0.15-0.5 cm). Stromanthe
hjalnuirssonii (Komicke) 0. G. Petersen from Nica¬
ragua, which has basal leaves and villous leaf
sheaths, has cream-colored flowers and only one or
two cauline leaves and does not have the additional
leafy shoots in the axils of the lower cauline leaves
found in S. palustris. The specific epithet, palustris,
refers to the habitat of the species, growing in
mucky, somewhat swampy, soil.

The description given under Ctenanthe sp. nov.
in Hammers (1986: 240) publication of Maranta-
ceae for La Selva is of Stromanthe palustris. ‘'‘’Cte¬
nanthe sp. nov." referred to the populations on the
Osa Peninsula that had purple bracts and distinctly
villous leaf sheaths and petioles. These populations
are now recognized as merely one end of the (clinal)
variation within Ctenanthe dasycarpa (Donnell-
Smith) K. Schumann, not a distinct species. Cte¬
nanthe dasycarpa has basal leaves and may also
have villous sheaths, but has persistent rather than
deciduous bracts.

Ctenanthe and Stromanthe are closely related,
both having antitropic leaves. While the majority
of species are easily classified as belonging to one
or the other of these genera, a few are problematic
having some characters found in both. Although the
longer sepal length (13 mm) is characteristic of
Ctenanthe, I feel the aspect of deciduous rather
than persistent bracts is more significant for its
placement in Stromanthe. Unfortunately, the seed
and capsule, which might have provided additional
characters, were not found in any of the specimens
examined.

Paratypes. COSTA MCA. Heredia: Finca I .a Selva.

Volume 9, Number 1
1999

Kennedy

Neotropical Marantaceae

65

the OTS Field Station on the Rfo Puerto Viejo just E of
its junction with the Rfo Sarapiquf, along Quebrada El
Saltito, back of Loop Trail. 900 m E. ca. 1(H) m, 29 Aug.
1980. H. Hammel 9627 (DUKE). Linion: Uitoy Cerere
Reserve and vicinity in Valle La Estrella S of Finca Con¬
cepcion, Los Jabillos, near station, 140 m, 09°42'N,
83°02'W. 2 Aug. 1985, li. Hanuncl & 47. Grayum 14352
(CR); Hitoy Cerere Reserve and vicinity in Valle La Es¬
trella. 09°40'00"N, 83°02'05"W, 28 Oct. 1990, B. Hammel
et al. 17948 (CR).

Acknowledgments. I am especially grateful to
the late Roberto Burle Marx of Brazil for the op¬
portunity to work with his extensive Marantaceae
collection. I thank Fatima de Gomes for help at
Sitio Burle Marx, Dimitri Sucre for sharing his live
and herbarium collections, and Dorothy Dunn de
Araujo for assistance and housing in Brazil. Field¬
work in Costa Rica was supported by NSF BSR-

8700068 to Michael Grayum and Barry Hammel for
“A Manual to the Plants of Costa Rica*’ project. I
thank Barry Hammel and Michael Grayum for help
and consultation during my fieldwork in Costa Rica
and for the use of their field vehicle; Josiane Le
Corff for help in taking the measurements; and the
staff at INBIO for their assistance while based at
the INBIO research facility. I thank the curators of
CR. CVRD, DUKE, F, K, MO, RB, and UBC for
use of their specimens and/or herbarium facilities.
I thank Fred Ganders for support and Tom Bau¬
mann and Janet Dick for use of their DOS-com¬
patible computers.

Literature Cited

Hammel, R. E. 1986. The vascular flora of La Selva Bi¬
ological Station, Marantaceae. Selbyana 9: 234—242.

Trois Especes Nouvelles d’Oleaceae et Note sur la Presence
c VO lea capensis dans P Arch i pel des Cornores

Jean- Noel hi hat

Museum national d’Histoire naturelle, Laboratoire de Phanerogamic, 16 rue Button, 75005

Paris, France, labat@nmhn.fr

Mare Pignal

Museum national d’Histoire naturelle, Laboratoire de Phanerogamic, 16 rue Button, 75005

Paris, France, pignal@mnhn.fr

Olivier Pascal

Direction de V Agriculture et de la Foret, Service de l’Environnement et de la Foret,

B.P. 103, 97600, Mainoudzou, Mayotte

Rfisi M f'. L’etude des caraeteres morphologiques
permet la description de trois especes nouvelles
d’Oleaceae de l’Archipel des Comores: Noronhia
cochleata Labat, M. Pignal & 0. Pascal, Chionan-
thus cordifolius Labat, M. Pignal & 0. Pascal et C.
insularis Labat, M. Pignal & 0. Pascal. Les oliviers
sauvages de l’Archipel sont pour la premiere fois
rattaches a Ole a capensis L. s.l.

Abstract. Morphological characters support the
description of three new species of Oleaceae from the
Comoro Archipelago: Noronhia cochleata Labat, M.
Pignal & 0. Pascal, Chionanthus cordifolius Labat. M.
Pignal & O. Pascal, and C. insularis Labat, M. Pignal
& 0. Pascal. For the first time the Archipelago “Wild
Olive Tree” is identified as Olea capensis L. s.l.

Les inventaires floristiques recents des plantes
indigenes ou naturalisees de Mayotte menes par le
Service de l’Environnement et de la Foret en col¬
laboration avec le Museum national d’Histoire na¬
turelle de Paris ont permis d’approfondir la con-
naissance de cette Hore insulaire. Les recoltes
systematiques realisees lors de ce projet ont nota-
blement enrichi les collections herbariologiques
peu abondantes et souvent anciennes. L’etude de
ce nouveau materiel et la revision des specimens
de l’Archipel des Comores presents dans l’herbier
du Museum (P) ont permis d’identifier plusieurs
taxons nouveaux. Une espece de Sapotaceae a deja
ete deerite (Labat et at, 1997). Pour la famille des
Oleaceae, ce travail revele l’existence de trois es-
pfeces nouvelles. II conduit egalement dans cette
region it clarifier la position systematique des po-
pulations appartenant an genre Olea.

Note. L’herbier cite “MAYOTTE” ne possfede
pas d’acronyme officiel; il s'agit de l’herbier du Ser¬
vice de l’Environnement et de la Foret de Mayotte.

Noronhia cochleata Labat, M. Pignal & 0. Pas¬
cal, sp. nov. TYPE: Mayotte. Mlima Combani,
28 dec. 1995 (fl), 0. Pascal 281 (holotype,
P(P75025); isotypes, K, MAYOTTE, MO,
P(P75026, P75027)). Figure 1.

A Noronhia boivinii Dubard, domatiis conspieuis. inflo-
rescentiis majoribus compositisque, fm<‘tibus ellipsoideis
apiculatisque, differl.

Petit arbre jusqu’a 7 m de hauteur et atteignant
rarement 10 cm de DBH, ou grand arbuste, 6corce
des petits rameaux grisatre parfois tres clair, por-
tant des lenticelles blanchatres ou homochromes
avec le reste des rameaux, glabre. Feuilles oppo-
sees, vertes dessus, glauques dessous, limbe ob-
long-elliptique, 7.5—13.5 X 2-5 cm, non coriace,
glabre, presentant sur la face inferieure & l'aisselle
de la plupart des nervures secondaires des doma-
ties craterilormes velues bien visibles a l’oeil nu,
ponctuations brun clair presentes, mais souvent
non visibles, meme a la loupe, nervation brochi-
dodrome, nervure primaire deprimee dessus, sail-
lante et generalement brun rougeatre dessous, 7 it
10 paires de nervures secondaires alternant avec
une nervure inter-secondaire et se divisant avant
d’atteindre Pare reliant les nervures secondaires,
nervures inter-secondaires et tertiaires habituelle-
ment plus visibles a la face superieure de la feuille,
seules les nervures secondaires sont bien marquees
a la face inferieure, sommet longuement cuspide.

Novon 9: 66-72. 1999.

Volume 9, Number 1
1999

Labat et al.

Oleaceae des Comores

67

cuspide longue de 0.5-1. 7 cm a sommet arrondi et
emargine, parfois a peine distinctement, marge tres
legerement et irregulierement sinueuse et crispee,
legerement repliee sur la face inferieure, base ai-
gue, attenuee sur le petiole; petiole epais, rougeatre
et lisse a l'etat jeune devenant grisatre et fissure
par la suite, canalicule a la face superieure, de 6—
8 mm de longueur et de ca. 1.5 mm de diametre,
glabre. Inflorescence axillaire en cymes triflores ou
en pleiochazes longues de 1.5— 2.5 cm environ, pu-
bescentes; pedoncule non renfie long de 1—8 mm;
ramifications vertes a noiratres sur le sec, greles,
longues de 3-8 mm; pedieelle de la fleur terminale
de 2 mm ou parfois sub-sessile, jusqu’a 4 mm de
longueur pour les fleurs laterales; bractees ob-
longues triangulaires, longues de 1-3 mm, pubes-
centes; bracteoles triangulaires de 0.8-1 mm de

longueur et 0.5 mm de largeur, sub-glabres, inais
terminees par une petite touffe de poils blanes
longue de 0.12 mm au maximum. Fleurs petites;
calice gamosepale a tube court, 4 lobes triangu¬
laires verts de 1 X 1 mm, glabres et surmontes des
memes petits poils blanes; corolle gamopetale,
blanche devenant jaune puis orange en vieillissant,
4 lobes sub-orbiculaires, coehleaires, de 3.5 mm
de largeur et 3 mm de hauteur, soudes sur 1 13 in-
ferieur, tube de la corolle tres court de 0.5 mm
environ, coronule tres petite, formant un bourrelet
circulaire de 0.6 mm de hauteur et 0.5 mm de lar¬
geur; androcee a 2 etamines soudees au tube de la
corolle par leur filet, filet 0.4 mm de hauteur et 0.3
mm de largeur, antheres introrses, basifixes, gros-
sierement rectangulaires, 1 mm de hauteur, 0.8 mm
de largeur et 0.5 mm d’epaisseur, pubescentes ex-

68

Novon

terieurement, dehiscentes par 2 fentes lat^rales; gy-
necee supcTe, ovaire biloculaire, pluriovule, 0.6-
0.8 mm de longueur et 0.6 mm de largeur; style
conique de 0.5 mm de long et 0.4 mm de large a
la base, surmont6 par 2 stigmates tubereuleux et
papilleux de 0.3 mm. Infrutescence grisatre, ealice
persistant. Fruits immatures en drupes ellipsoides,
1.5 X 1 cm, apiculees.

Noronhia cochleata appartient au groupe Genui-
nae defini par Perrier de la Bathie (1952); elle dif-
lere de N. boivinii Dubard, espece allopatrique des
forets orientales malgaches, par la presence de
domaties, ses inflorescences plus longues et com-
pos^es, ses fruits ellipsoides et apicules. N. coch¬
leata et N. comorensis S. Moore presentent une dis¬
tribution syinpatrique, mais cette derniere espece
en differe principalement par des fleurs fascicules
et des corolles campanulas il lobes reduits 4 de
larges ondulations arrondies. Un specimen, Jac-
quemin 86 0-J, lie de Moheli, piste du Badjo au
smnmet du M'Ze, Koukoule, 12 nov. 1970 (P), pa-
rait tres proche de ce taxon, mais les fleurs presque
entierement detruites par des insectes ne permet-
tent pas une identification certaine.

Distribution. Endemique de Pile de Mayotte.

Ecologie. Petit arbre de sous-bois, relativement
commun dans les massifs forestiers humides de Pile,
absent des vestiges de forets seches (zones pour les-
quelles la saison seche est superieure a 6 inois).

Phenologie. Sempervirent, floraison de novem-
bre & janvier, fructification en mars-avril.

Nom vernaculaire. Saritsileytri (en chibuchi,
dialecte malgache de Mayotte).

Paratypes. \1 AYOTIE. Majimbini, 6 nov. 1995 (ff).
Pascal lot) (MAYOTTE, P); Mlima Choungui, 16 nov.
1995 (fl). Pascal 205 (G, K. MAYOTTE, MO. I*. WAG);
Sohoa, 30 nov. 1995 (fl), Pascal 257 (k. MAYOITE, MO.
I’); Sohoa, 14 mars 1996 (fr), Pascal 418 (k. MAYOITE,
MO. P); Dapani, 7 jan. 1997 (fl) Pascal 857 (k, MA-
YOTTE, MO. P. WAG).

Ghionanthus eordifolius Labat, M. Pignal & O.
Pascal, sp. nov. TYPE: Mayotte. Mlima Choun¬
gui, 29 dec. 1995 (fl), 0. Pascal 288 (holotype,
P(P75021); isotypes, K, MAYOTTE). Figure 2.

A Chionantho mildbraedii (Gilg & Schellenberg)
Steam, foliis parvioribus, ovatis, domatiis absentibus, iri-
florescentiis brevioribus, tliffert.

Petit arbre de 6 m de hauteur, ecorce des rameaux
brun grisatre lonce, glabre, avec quelques lenticelles
homochromes avec1 le reste des rameaux. Feuilles
opposees, face superieure brune, presque luisante,
face inferior re vert brunatre, limbe ove & [)lus rar-
ement oblong-ove, (4— )6.5-10 X (1.4— )3-4.3 cm.

legerement coriace, glabre, sans domaties a Paisselle
des nervures secondaires, densement couvert de
ponctuations brun elair a noires, seulement visibles
a la loupe, nervation brochidodrome, nervure pri-
maire deprimee dessus, saillante et generalement
brun rougeatre dessous, 6 a 8 paires de nervures
secondaires alternant avec une nervure inter-secon-
daire se divisant avant d'atteindre Parc reliant les
nervures secondaires, apex obtus a aigu, parfois
courtement acumine, base tronquee, legerement cor-
dee; petiole epais, presentant la tneme couleur et
texture que les rameaux, souvent legerement eoude,
de 3—4 mm de longueur et de 1.5—2 mm de dia-
metre, glabre. Inflorescences en pleiochazes laches,
axillaires, de 3 cm environ, glabres, a pedoncule
grisatre, de 3—7 mm de longueur et a ramifications
noires sur le sec, greles, de 7—15 mm de longueur;
pedicel les floraux identiques aux ramifications, mais
plus courts, (0-)2 X 5 mm; bractees de 1 mm en¬
viron, jamais foliacees, pubescentes; bractdoles
identicjues aux bractees, mais deux fois plus petites,
inserees a 1.5—2 mm en dessous du calice. Calice
gamosepale, 4 lobes verts triangulaires a extremite
arrondie, de 1 mm de hauteur et 1-2 mm de largeur,
soudes sur leur tiers inferieur, formant une collerette
etalee, meme dans les fleurs en boutons; corolle & 4
lobes jaunes, epais, oves, de 4 mm de longueur et
3 mm de largeur, soudes & la base, pas de coronule;
androcee a 2 etamines soudees a la corolle, sub-
sessiles, antheres cordiformes tres epaisses de 2 mm
de largeur et 0.8—1 mm d'epaisseur, basifixes, in-
trorses a dehiscence par des fentes presque laterales;
gynecee supere, pyriforme, de 1.2 mm de hauteur
en ineluant le style, et 1 mm de large & la base,
ovaire compose de 2 loges & 2 ovules, style peu dis¬
tinct de Povaire, 2 stigmates. [Fruit inconnu.]

Chionanthus eordifolius est proche de C. mild¬
braedii, . mais ses feuilles, plus petites, sans doma¬
ties, sont ovees et tres larges dans le tiers inferieur
avec une base tronquee et legerement eordee sur
un petiole court et trapu donnant un aspect presque
sessile a la feuille. Get aspect est caraoteristique
de cette espece nouvelle, par rapport a toutes les
autres especes africaines et malgaches connues.
Les inflorescences sont egalement laches, mais
beaucoup plus courtes que ehez C. mildbraedii.

Distribution. Endemique de Pile de Mayotte.

Ecologie. Petit arbre de sous-bois rare, unique-
merit rencontre vers 500 m d’altitude en foret hu-
mide. Cette espece semble affectionner les cretes
dans la zone des brouillards matinaux. Ajiparem-
ment absent des vestiges de forets seches.

Phenologie. Sempervirent, floraison connue
uniquement en decembre.

Volume 9, Number 1
1999

Labat et al.

Oleaceae des Comores

69

F igure 2. Chionanthus cordifolius Labat, M. Fignal & (). Pascal. — A. Aspect general. — H. Detail de la face inferieure
de la feu i lie. — C. Fleur vue de prof i 1 . — D. Lobe de la corolle. — F. Corolle et etamines vue de dessus. — F. Ftamine.
— G. Ovaire. (A— G dessine d'apres Pascal 288.)

Ghionaiitliiis insularis Labat, M. Pignal & (). Pas¬
cal, sp. nov. TYPE: Mayotte. Benara, 350 m,
11 oct. 1996 (fl), (). Pascal 718 (holotype, P
(P75020); isotypes, B, G, K, MA, MAYOTTE,
MO, NY, P(P75017, P75018, P75019 alcool),
WAG). Figure 3.

A Chionantho cordifolio Labat. M. Pignal & (). Pascal,
foliis majoribus, oblongis ellipticis, basi baud cordata, in-
florescentiis majoribus, staminibus apiculatis, differ!.

Arbre de 15 m de hauteur (jusqu’a 30 m scion
Humblot?) et de 40—45 cm de diametre, ecorce des
rameaux grise, glabre, portant d’assez nombreuses
lenticelles legerement plus claires. Feuilles oppo-
sees, face superieure vert brunatre, face inferieure
plus claire, litnbe oblong-elliptique, (8 — ) 1 1 — 20 X
(3— )4— 7 cm, peu coriace, glabre, domaties crateri-
formes, velues a subglabres, rarement absentes, en
general 1—5 situees a Paisselle de quelques ner-
vures secondaires entre le quart et la moitie infe-
rieur du limbe, rarement plus haul, densement cou-

verte de ponctuations brun clair a noires,
settlement visibles a la loupe, nervation brochido-
drome, nervure primaire deprimee dessus, saillante
et generalement brun rougeatre dessous, 8 a 10 paires
de nervures secondaires alternant avec une nervure
inter-secondaire se divisant avant d’atteindre Tare
reliant les nervures secondaires, apex acumine a
cuspide, cuspide longue de 1—1.5 cm, non echan-
cree a Textremite mais souvent necrosee, base ai-
gue a attenuee; petiole epais, presentant la meme
couleur et texture que les rameaux, souvent legere¬
ment coude, de 7—10 mm de longueur et de ca. 2
mm de diametre, glabre. Inflorescences en pleio-
chazes axillaires laches, pouvant parfois parattre
terminales, jusqu’a 7.5 cm de longueur, glabres-
centes, a pedoncule grisatre, court, 2—6 mm de lon¬
gueur et a ramifications noires sur le sec, longues
de 15—35 mm; pedicelle souvent a section carree,
long de 3 — 8( — 10) mm pour la fleur terminale et de
5— 10(— 14) mm pour les Heurs laterales; bractees
parfois triangulaires et longues de 3 mm ou parfois

70

Novon

Figure 3. Chionanthus insularis Labat, M. Pignal & (). Pascal. — A. Aspect general. — 15. Detail <le la lace inferieure
<le la feuille avec domatie. — C. Bouton floral. — I). Corolle et etanunes vue de dessus. — K. Ktaniine et base de la
corolle. — F. Ovaire. — (5, H. Fruit. (A-F dessine d'apres Pascal 713 et G-H d'apres Pascal ()I7.)

foliaeees et tie mm, acuminees; 2 bracteoles

opposes, triangulaires, longues de 1 mm, ciliees.
Calice gamos£pale, 4 lobes verts de 2.3 mm de
largeur et 1.5 mm de hauteur, soudes a leurs bases,
cilies; corolle gamopetale, rotacee, ii 4 lobes jaune
vil tr6s epais, elliptiques-oves, soudes a la base,

4.5 mm de longueur et 3 mm de largeur, extremite
interieure sillonnee, pas de coronule; androcee a 2
^famines soudees a la corolle, opposees aux se-
pales, filet aplati de 0.8 mm de large et 1.2 mm de
hauteur, antheres basifixes, sub-orbiculaires, ca.

1.6 mm de diainetre, introrses, a fentes laterales,
connectif elargi en une eourte pointe terminale; gy-

nec6e supere, ovaire pyriforme, large de 1 mm k la
base, biloeulaire, style peu distinct de l'ovaire, de
0.3 mm tie hauteur, 2 stigmates. Infrutescences ri-
gides, grisatres, ne ptirtant habituellement qu’un
seul fruit develojipe. Fruits en drupe ellipsoidale,
4.5 X 2 cm, a fines cretes saillantes et presentant
un rostre court au sommet.

Chionanthus insularis differe de C. cordifolius
par ses feuilles plus grandes, oblongues-elliptiques,
non eordees a la base, ses inflorescences plus grandes
et ses etamines ajiiculees. Cette nouvelle espeee est
egalement assez proche par la structure ties inflo-

Volume 9, Number 1
1999

Labat et al.

Oleaceae des Comores

71

rescences et la morphologie florale de C. mildbrae-
dii, mais elle en differe en particulier par ses eta-
inines apiculees, ainsi que par la forme el la taille
de ses fruits. II existe une espece, Linociera may -
ottensis H. Perrier (Perrier de la Bathie, 1949,
1952), decrite sur un specimen recolte sur Pile de
Mayotte: Boivin 3196. Steam (1980) a volontaire-
ment exclu cette espece de sa revision des Chio-
nanthus d’Afrique, considerant que sa position ta-
xonomique etait incertaine. En effet, le specimen
type est tres incomplet et en mauvais etat. Cepen-
dant, l’etude des caracteres observables permet de
considerer qu'il s’agit d’une espece differente: les
debris de fleurs presents ont montre que chez L.
mayottensis l’etamine est plus allongee et que
l’apicule est plus long et bn, de plus les inflores¬
cences n’ont pas de pedoncule grisatre difference,
sont pubescentes, plus courtes et plus compactes
que chez C. insularis. Le specimen fructifere, Loup
71, de C. insularis, recolte sur la Grande Comore,
porte des fruits tres immatures lisses, mais il est
probable que les cretes apparaissent plus tardive-
ment. Elies sont effectivement presentes sur les
fruits, legerement plus developpes, mais encore im¬
matures du specimen Humblot 1407.

Distribution, Endemique des lies de Mayotte et
de la Grande Comore.

Ecologie. Espece rare a Mayotte, connue par
un seul individu rencontre vers 500 m d’altitude
lors des inventaires realises par le Service de
l’Environnement et de la Foret de Mayotte. Sa pre¬
sence dans la foret de Nioumbadjou, dans la region
occidentale de la Grande Comore, sensiblement
plus humide (3 a 4 m d’eau par an) et possedant
le seul climat axerique de Pile (sans saison seehe),
peut expliquer en partie sa rarete a Mayotte ou les
conditions climatiques ne seraient pas favorables a
son developpement. Gependant, L absence de don-
nees concernant son abondance et sa repartition a
la Grande Comore, ou dans les autres lies de
l’Archipel, ne permet pas de conclure sur les pref¬
erences ecologiques de cette espece.

Phenologie. Sempervirent, floraison observee
en aout et octobre et fructifleation en aout et oc-
tobre— novembre.

Nom vernaculaire. N’driali mendje; N’drihali
(Grande Comore).

Paratypes. MAYOTTE. Foret de Combani, 13 nov.
1884 (jfr). Humblot 1407 (P); Benara. I aout 1996 (fr),
Pascal 617 (O. K. MAYOTTE. MO. P). GRANDE CO¬
MORE. Nioumbadjou, 4 aout 1952 (ft). Service des Eaux
et Forets de Madagascar 5432-SF (P); village de Nioum¬
badjou, chemin du capitaine Dubois en direction de
N’Kourani, 9 oct. 1984 (jfr). Ixnip 71 (P|.

Su k i.a Presence o'Olka capensis L. aux
Comores

Perrier de la Bathie (1952), dans son traitement
pour la flore de Madagascar et des Comores, avait
examine des recoltes provenant de Mayotte (Boivin
3197) et d’Anjouan (Boivin s.n.). Ce materiel
n’avait ete integre dans la flore que sous la forme
d’une note: “Boivin a recolte sur Pile de Mayotte
(Comores), des specimens ( Boivin 3197), trop in-
complets pour etre decrits, mais appartenant tres
probablement au genre Olea .” Perrier de la Bathie
ne rapproche pas non plus ces plantes des especes
africaines puisqu’il note sur le specimen de Ma¬
yotte, qu’il s’agit d’une “espece probablement nou-
velle, mais trop pauvrement representee (. . .),” et
sur celui d’Anjouan, il mentionne qu’il s’agit sans
doute de la meme espece. Les recoltes plus com¬
pletes que nous avons etudiees, en particulier celles
de 0. Pascal de Mayotte, ne nous permettent pas a
notre tour de rattacher ces plantes aux especes
malgaches. La systematique des Olea de Madagas¬
car doit etre entitlement revue a la lumiere de
l’abondant materiel recolte, essentiellement par le
Service Forestier de Madagascar, depuis le traite¬
ment de Perrier de la Bathie. Cependant, l’espece
la plus affine est 0. madagascariensis Boivin ex H.
Perrier. Les specimens des Comores s’en differen-
cient principalement par leurs inflorescences plus
greles, leurs lobes du calice moins prononces, leurs
fruits j)lus larges et arrondis au sommet.

Par ailleurs, des caracteres morphologiques, es-
sentiellement la presence d’inflorescences termina-
les, les differencient trbs clairement d 'Olea europea
L. subsp. africana (Mdler) P. S. Green et les rap-
prochent au contraire d 'Olea capensis L. Coates
Palgrave (1996) note chez O. capensis une ecorce
gris fonce, avec un rhytidome fissure verticalement
et la formation earacteristique d un exsudat noir
dans l’ecorce entaillee. Ces caracteres ont ete ob¬
serves pour les individus de Mayotte.

Olea capensis est une espece africaine largement
distribute dans les forets sem{)ervirentes et semi-
decidues d’Afrique tropicale, sa repartition s’etend
jusqu’en Afrique du Sud. L’extreme variabilite de
O. capensis a conduit Verdoorn (1956) a regrouper
dans l’espbce deux autres taxons d’Afrique du Sud
en deflnissant deux unites infra-speeifiques: 0. ca¬
pensis subsp. enervis (Harvey) I. Verdoorn et 0. ca¬
pensis subsp. macrocarpa (C. H. Wright) I. Ver¬
doorn. En 1983, Kupicha souscrit a une dehnition
encore plus large de l’espece en mettant en synony-
mie de nombreuses especes d’Afrique et en ne re-
connaissant aucune sous-espece. Par la suite, Friis
et Green (1986), tout en admettant la complexity

72

Novon

de ce groupe, reviennent i la conception de Ver-
doorn pour rAfri(|ue du Sud et reconnaissent en
plus deux sous-especes pour TAfrique tropicale
(subsp. hochstetteri (Baker) Friis & P. S. Green et
subsp. welwitschii (Knoblauch) Friis & P. S. Green).

Les caracteres foliaires des specimens des Co-
mores sont intermediaires entre ceux des deux
sous-especes d’Afrique tropicale, alors que le fruit,
une drupe ovoi'de de 1.5 X 1 cm, est proche de
ceux observes ehez la sous-espece macrocarpa du
Transvaal, Natal et Province du Gap. Les inflores¬
cences sont plus laches et greles que dans les re-
coltes africaines en general. II est done difficile
d’assimiler les plantes des Comores a l une des
sous-especes existantes.

Les delimitations taxonomiques de ee groupe sont
loin d'etre claires. En reprenant a notre compte les
propos de Friis et Green (1986): "There is no doubt
that a full and careful analysis of 0. capensis sensu
tat. throughout its range is needed but unfortunately
neither of us is in a position to undertake this at
present. However, for practical purposes names are
required, to enable reference to be made to the en¬
tities in north-east tropical Africa”; il nous semble
inopportun de decrire un nouveau taxon infra-spe-
cifique. De plus, 0. madagascariensis, qui appartient
egalement a ce complexe, devra etre examinee au
tneme titre que les populations continentales. En at¬
tendant une etude plus approfondie, nous conside-
rons que les plantes examinees appartiennent it
l’esp^ce 0. capensis sensu lata , signalee ici pour la
premifere fois dans l’archipel des Comores (Grande
Comore, Mayotte, Anjouan, Moheli).

Notes tcologiques et physionomiques. A Ma¬
yotte, cette espf'ce est commune dans toutes les
reliques de foret naturelle humide de File, entre
150 m (limite inferieure des reliquats) et 600 m
d’altitude. C'est un arbre massif, muni de contre-
lorts puissants largement Stales, le tronc principal
depasse rarement 6—7 m de hauteur, 3 ou 4 branches
mattresses dressees prennent en general le relais
et forment un houppier en entonnoir largement eva-
s£ au sommet. Get allure peu forest iere est habi-
tuellement plus earaeteristique des milieux ouverts.
Les grands sujets (jusqu’il 25 m de hauteur el 80
cm de diametre) sont connus dans les massifs les
plus humides (Benara et M'Tsapere) sur pente et
haut de pente. Sur les cretes de ces massifs, comme
dans des zones moins arrosees (plateau de Sohoa
et Ouchoungui sur la cote ouest. Mont Choungui au
sud de File) leur port est plus rabougri et les in-
dividus rejettent abondamment a la base des souches.
La periode de floraison parait relativement breve,
au plus fort de la saison des pluies, de mi-janvier

a fin fevrier; la fructification s’etale de mars it sep-
tembre et jusqu'en novembre.

Noms vernaculaires. M’Rimeou, Mkarare
(Grande Comore), Chivoumze, Sheivindze (An¬
jouan), Antsagniry (dialecte chibuchi), M'chel6l6
(dialecte chimahorais) (Mayotte).

Materiel etudu s. GRANDE COMORE. Foret cfe la
Grille, 09 fev. 1957 (fl). Service des Eaux et Forets de
Madagascar 16524 SF (I’); foret du Karthala. Convales¬
cence. route Capitaine Dubois, 19 fev. 1957 (fl). Service
des Faux et Forets de Madagascar 16598 SF (I’). MOHE¬
LI. Foret Bambao, Fomboni, 16 mars 1957 (fr), Service des
Faux et Forets de Madagascar 1671 6 SF (I*). ANJOUAN.
Sin. loc., mai 1850 (fr). Boivin s.n. (P(P80210)); Koni
Djodjo. 1 nov. 1984 (fr). Ducatillion 447 (I’): Koni Djodjo,
I nov. 1984 (fl). Ducatillion 448 (I*): face sud-est du mont
Tingui, Hindi. Ouani. 12 mars 1957 (fl), Service des Faux
et Forets de Madagascar 16674 SF (K. P). MAYOTTE. Au
pied du Chongui, au dessus de Dapani. juil. 1848 (fr),
Boivin 8197 (P); Choungui, 6 sep. 1995 (Ir). Pascal 4
(MAYOTTE, P); Dapani. 17 jan. 1996 (11). Pascal 335 (B.
HR, G. K. MA, MAYOTTE. MO. NY. P. PRE. WAG); Da¬
pani. 23 jan. 1996, (fl), Pascal 355 (K. MAYOTTE, P);
Tchaourembo. 8 lev. 1996 (fr). Pascal 384 (MAYOTTE,
P); Sohoa. 14 mars 1996 (fr), Pascal 419 (MAYOTTE, P);
Dapani, 1 I avr. 1996 (fr). Pascal 459 (K. M AYOTTE. P);
Sohoa. 25 sep. 1995 (fr). Pascal et Ali andi 67 (MA-
^ OTTE, P); foret d'Ouchoungui, 25 mars 1957 (fl). Service
des Faux et Forets de Madagascar 16775 SF (P); foret
d'Ouchoungui. 25 mars 1957 (fl. fr). Service des Faux et
Forets de Madagascar 16776 SF (P).

Renter dements. Nous remercions J. Lemeux et
D. Storez qui ont realise les illustrations et J.-J.
Florel qui a traduit les diagnoses latines. Cette re¬
cherche a ete linancee par une convention entre le
Museum national d'Histoire naturelle et la collec-
tivite territoriale de Mayotte (convention d'etude du
21 fevrier 1996: etudes sur la diversity vegetale
naturelle de Mayotte).

Bibliographic

Coates Palgrave, k. 1996. Trees of Southern Africa. Struik
Publishers, Cape Town.

friis. I. & P. S. Green. 1986. Olea capensis (Oleaceae) in
north-east and east tropical Africa. Kew Bull. 41: 36.
Kupicha. E. k. 1983. Oleaceae. 108. Pp. 300-327 in E.
Launert (editor). Flora Zambesiaca, vol. 7. f lora Zam-
besiaca Managing Committee, London.

1 .a bat, J.-N.. M. Pignal & (). Pascal. 1997. Une nouvelle
espeee de Libramia (Sapotaeeae) de Tile de Mayotte
dans 1’Archipel des Comores. Adansonia, ser. 3, 19(2):
213-216.

I\ ;rrier de la Bathie, H. 1949. Revision des Oleaeees de
Madagascar et des Comores. Mem. Inst. Sci. Madagasc.,
ser. B. 2: 275-310.

- . 1952. Oleaeees. In II. Humbert (6diteur), Flore

de Madagascar et des Comores, fam. 166.

Steam, W. T. 1980. African species of Chionanthus L.
(Oleaceae) hitherto included in Linociera Swartz. J. Bot.
Linn. Soc. 80: 191-206.

Verdoom. I. C. 1956. The Oleaceae ol southern Africa.
Bothalia 6: 549-601.

Nomenclatural Consequences of the Synonymization of Hypsela
reniformis (Campanulaceae: Lobelioideae)

Thomas G. Lammers

Department of Botany, Field Museum of Natural History, Chicago, Illinois 60605-2496, U.S.A.

Iammers@fmnh.org

ABSTRACT. The recent synonymization of Hypsela
reniformis with Pratia repens has necessitated sev¬
eral nomenclatural changes. The enlarged circum¬
scription of P. repens dictates the validation of a
new combination to cover the species when it is
considered a member of Lobelia. Because //. reni¬
formis was the type of its genus, its three remaining
congeners can no longer employ that generic name.
They are transferred to Isotoma, from which they
differ solely in the possession of indehiscent (vs.
capsular) fruit.

Recently, Chiapella (1996) demonstrated statis¬
tically that two long-recognized South American
species of Lobelioideae (Campanulaceae), Hypsela
reniformis (Kunth) C. Presl (including Pratia lon-
giflora J. D. Hooker) and Pratia repens Gaudi-
chaud, could not be distinguished in any meaning¬
ful way, i.e., that the two names pertained to a
single species. This conclusion has nomenclatural
consequences that extend beyond South America
and a simple taxonomic merger. The purpose of this
paper is to effect changes necessitated by Chiapel-
la’s (1996) findings, in preparation for a forthcom¬
ing checklist of the species of Campanulaceae (cf .
Lammers, 1995, 1998a).

Generic Considerations

Coincidentally, each species is the type of its
respective genus. As a result, their merger into one
species means that these two genera must likewise
be merged. Pratia Gaudichaud, the name under
which far more species have been described, has
11 years priority over Hypsela C. Presl, and was the
name adopted by Chiapella (1996), as well as by
Baillon (1885), who treated Hypsela as a section of
Pratia.

However, there has been a growing trend over
the past 40 years to not recognize Pratia as a dis¬
tinct genus. The primary character used to distin¬
guish it from Lobelia L. has been its indehiscent
(vs. capsular) fruit (Bentham, 1876; McVaugh,
1943; Wimmer, 1943, 1953). However, various

workers have called attention to intraspecific vari¬
ation in fruit type (Moeliono & Tuyn, 1960), a lack
of correlation between fruit type and other char¬
acters (Albrecht, 1994), and the greater overall re¬
semblance of various species of Pratia to species
of Lobelia than to congeners (Adams, 1972; Wilbur,
1991; Murata, 1995), and subsumed Pratia into Lo-
belia. As a result, Hypsela should also be regarded
as a synonym of lobelia, as indicated in the partial
synonymy below:

Lobelia L., Sp. PL 929. 1753; Gen. PL (ed. 5) 401.
1754. TYPE: Lobelia cardinal is L. (lectotype,
designated by Hitchcock & Green (1929:
184)).

Pratia Gaudichaud, Ann. Sci. Nat. 5: 103. 1825. hibelia
[no rank indicated] Pratia (Gaudichaud) Heynhold.
Nom. Hot. Holt. 1: 473. 1840. Lobelia sect. Pratia
(Gaudichaud) J. Murata, J. l ac. Sci. Univ. Tokyo,
Sect. 3, Hot. 15: 356. 1995. TV PK: Pratia repens
Gaudichaud.

Hypsela C. Presl. Prodr. Monogr. Lobel. 45. 1836. Syn.
nov. Lysipomia sect. Hypsela (C. Presl) A. DC., in
l)<].. Prodr. 7: 350. 1839. Pratia sect. Hypsela (C.
Presl) Baillon. Hist. PI. 8: 366. 1885. TYPE: Hypsela
reniformis (Kunth) C. Presl.

The Correct Name eor Pratia repens s.e. in
Lobelia

Chiapella (1996) used the name Pratia repens for
the more broadly circumscribed species resulting
from the merger. The basionym of H. reniformis,
Lysipomia reniformis Kunth, had been published six
years earlier, but its use in Pratia was preempted
by P. reniformis (Chamisso) Kanitz. The situation
becomes more complicated when the species of
Pratia are transferred to Ijrbelia.

Transferring the epithet of P. repens to Lobelia is
precluded by L. repens Thunberg. Consequently, I
recently (Lammers, 1998a) validated the name L.
pratiana Gaudichaud ex Lammers, originally sug¬
gested by Murata (1995) as a nomen novum. This
is the correct name in Lobelia for Pratia repens s.
str. However, it is not the correct name for the spe¬
cies in its expanded circumscription.

Novon 9: 73-76. 1999.

74

Novon

The epithets ol Hypsela reniformis and Pratia
longiflora likewise cannot he transferred to Lobelia ,
due to the existence of L. reniformis Chamisso and
L. longiflora L., respectively. However, the synon¬
ymy of Hypsela reniformis (Wimmer, 1943; Jeppe-
sen, 1981; Lammers et al., 1993) includes four ad¬
ditional heterotypic names, all of which antedate
Lobelia pratiana. Two of these have equal priority,
and there is no impediment to taking up either ep¬
ithet. The one that seems more appropriate mor¬
phologically is transferred to Lobelia, to provide a
name in that genus for Pratia reperus s.l. It would
also he the correct name for //. reniformis if that
species were to be maintained as distinct from P.
repens hut likewise assigned to Lobelia. For clarity,
complete synonymy is provided here for the spe¬
cies, as circumscribed by Chiapella (1996).

Lobelia oligophylla (Weddell) Lammers, comb,
nov. Basionym: Pratia oligophylla Weddell,
Chlor. And. 2: 10. 1857. Hypsela oligophylla
(Weddell) Bentham & J. D. Hooker ex Zahl-
bruekner. Bull. Torrey Bot. Club 24: 387.
1897. TY PE: Peru. Cordilliere de Tacora, no-
tamment au voisinage de la metairie de Chun-
lunquaiani, pelouses un peu mareeageuses des
punas , 4000 m, Weddell s.n. (holotype, P not
seen).

Lysipomia reniformis Kunth, in H.B.k., Nov. Gen. Sp. 3:
320. 1819; non htbelia reniformis Chamisso, Lin-
naea 8: 210. 1833. Hypsela reniformis (kunth) C.
Presl, Prodr. Monogr. Lobel. 45. 1836. TYPE: Ec¬
uador. Propter speluneam Antisanae, 15,000 ft.,
Humboldt A lionpland s.n. (holotype, P not seen).
Pratia repens Gaudichaud, Ann. Sci. Nat. 5: 103. 1825;
non Lobelia repens Thunberg, Prodr. FI. Cap. 40.
1794. htbelia pratiana Gaudichaud ex Lammers,
Novon 8: 34. 1998. TYPE: Falkland Islands. Gau¬
dichaud s.n. (holotype, P not seen).

Pratia longiflora .1. I). Hooker, FI. Antaret. 325. 1846;
non htbelia longiflora L., Sp. PI. 930. 1753. Hypsela
longiflora (J. I). Hooker) F. Philippi, Cat. PI. Vase.
Chil. 232. 1881. TYPE: Chile. Straits of Magellan,
Cape Negro, Darwin s.n. (holotype, k not seen).
Pratia subsessilis Weddell, Chlor. And. 2: 10. 1857. Hyp¬
sela sulisessilis (Weddell) Bentham & J. I). Hooker-
ex Zahlbruekner. Bull. Torrey Bot. Club 24: 387.
1897. TYPE: Bolivia. Province de Cinti. pelouses un
peu mareeageuses el bonis des ruisseaux, dans les
punas, 3500 m, Weddell s.n. (holotype, P not seen).
Pratia ataeamensis B. A. Philippi, FI. Atacam. 208. 1860.
Hypsela ataeamensis (B. A. Philippi) F. Philippi, Cat.
PI. Vase. Chil. 232. 1881. TYPE: Chile. Frequens
ad fontes deserti Ataeamensis, Jan. -Feb. 1854, Phi¬
lippi s.n. (holotype, SCO).

Pratia pencana B. A. Philippi, Anal. Univ. Chile 18: 53.
1861. TYPE: Chile. Coquimho: lluanta. 1860, Vohk-
mann s.n. (holotype, SGO).

Disposition of Orphaned Hypsela Species

There remains the question ol what to do with
the remaining species of Hypsela. In the most re¬
cent monograph (Wimmer, 1943), three species
were recognized in addition to the type. None occur
in South America, but rather are endemic to Aus¬
tralasia: H. rivalis F. E. Wimmer to New Zealand,
and //. sessiliflora F. E. Wimmer and //. tridens F.
E. Wimmer to Australia. These species obviously
can no longer go by the name Hypsela.

It might be logical to suggest that they follow the
generic type into Lobelia, However, this would not
be an optimal disposition. Even prior to Chiapella’s
(1996) research, I had suspected that the South
American and Australasian elements of Hypsela
were not congeneric. In the three Australasian spe¬
cies, the staminal column is adnate to the corolla
lor at least hall its length, while in L. oligophylla
it is lree or only adnate to the corolla at its base.
Such pronounced adnation is unusual in the suh-
family; most genera, including htbelia, are char¬
acterized by having the staminal column free from
the corolla or only adnate to the corolla at its base
(Bentham, 1876; Schbnland, 1889; McVaugh,
1943; Wimmer, 1953). Similarly, the entire corolla
tube that characterizes these species is an unusual
feature in htbelia, occurring in very lew species,
none ol them Australasian.

It might be reasonable to erect a new genus for
tht ^se three orphans, as no existing genus of Lo-
belioideae combines their herbaceous habit, creep¬
ing prostrate stems, small flowers solitary in leal
axils, corollas with a cylindric entire tube and sim ill
subequal lobes, filament tubes adnate to the middle
ol the corolla tube, ventral anthers with a long flat
triangular trichome and tufts of minute hairs at the
apex ol each, and indehiscent Iruits. However,
there is one genus in which all these features save
one may be found.

Isotoma (R. Brown) Lindley (excluding Hippo-
broma G. Don) comprises 11 species endemic to
Australia (Wimmer, 1953; Melville, 1960; Elliot &
Jones, 1990). The genus is characterized by its her¬
baceous habit, corollas with a cylindric entire tube
and small subequal lobes, filament tubes adnate to
the middle of the corolla tube, ventral anthers with
a pair ol long setae and tufts of minute hairs at
their apex, and apieally loculicidal capsules. All
but one species, I. hypocrateriformis (R. Brown)
Druce, have solitary axillary flowers, and two of
tin ^se, I. fluviatilis (R. Brown) F. Mueller ex Ben¬
tham and /. luticola Carolin, have creeping pros¬
trate stems. As noted by Wimmer (1943), //. rivalis
bears a striking resemblance to 1. fluviatilis and has

Volume 9, Number 1
1999

Lammers

Nomenclature in Lobelioideae

75

often been mistaken for it (cf. Allan, 1961; Beu-
zenberg & Hair, 1959, 1983).

The only feature that distinguishes Isotoma from
the three orphan species are its capsular fruits.
Given current thinking on the inappropriateness of
indehiscence as a generic character in Campanu-
laceae (Moeliono & Tuyn, 1960; Adams, 1972; Wil¬
bur, 1991; Albrecht, 1994; Murata, 1995; Lam¬
mers, 1992, 1993, 1998a, 1998b; Morris &
Lammers, 1997; Pepper et al., 1997), establishment
of a new genus on this basis would be inadvisable.
Just as Lobelia can accommodate the species of
Pratia, Isotoma is judged to be the best home for
the three species orphaned by the synonymization
of Hypsela reniformis. The requisite combinations
are established here.

Isotoma (R. Brown) Lindley, Bot. Reg. 12: pi. 964.
1826. Basionym: Lobelia [no rank indicated]
Isotoma R. Brown, Prodr. 565. 1810. Laurentia
[no rank indicated] Isotoma (R. Brown) End-
licher. Gen. PL 512. 1838. Laurentia subg. Is¬
otoma (R. Brown) Petermann, Pflanzenr. 444.
1845. Laurentia sect. Isotoma (R. Brown) F. E.
Wimmer, Ann. Naturhist. Mus. Wien 56: 335.
1948. TYPE: Lobelia hypocrateriformis R.
Brown.

Isotoma rivalis (F. E. Wimmer) Lammers, comb,
nov. Basionym: Hypsela rivalis F. E. Wimmer,
Pflanzenr. IV.276b: 121. 1943. TYPE: New
Zealand. South Island: Broken River, Kirk 464
(holotype, CGE not seen).

Isotoma sessiliflora (F. E. Wimmer) Lammers,
comb. nov. Basionym: Hypsela sessiliflora E.
Wimmer, Pflanzenr. IV.276b: 121. 1943.
TYPE: Australia. New South Wales: Port Jack-
son, Bauer s.n. (holotype, W not seen).

Isotoma tridens (F. E. Wimmer) Lammers, comb,
nov. Basionym: Hypsela tridens F. E. Wimmer,
Pflanzenr. IV.276b: 121. 1943. TYPE: Austra¬
lia. Sine loc., Cunningham s.n. (holotype, W
not seen).

Acknowledgment. The manuscript was re¬
viewed by Fred Barrie (MO) [trior to submission.

Literature Cited

Adams, C. I). 1972. Campanulaceae. Pp. 734—737 in
Flowering Plants of Jamaica. University of the West In¬
dies. Mona.

Albrecht, I). 1994. A new combination in Lobelia L. Tel-
opea 5: 791—792.

Allan. H. II. 1961. Lobeliaeeae. Pp. 796—802 in Flora of
New Zealand, vol. 1. R. F. Owen. Wellington.

Baillon, II. F. 1885. Fobelieae. Pp. 362—367 in Uistoire
des Plantes, vol. 8. Hatchette, Paris.

Bentham, G. 1876. Campanulaceae. Pp. 541—564 in C.
Bentham & J. I). Hooker, Genera Plantarum, vol. 2.
Reeve, London.

Beuzenberg, F. J. & J. B. Hair. 1959. Contributions to a
chromosome atlas of the New Zealand llora — 3. Mis¬
cellaneous families. New Zealand J. Sci. 2: 531—538.

- & - . 1983. Contributions to a chromosome

atlas of the New Zealand flora — 25. Miscellaneous spe¬
cies. New Zealand J. Bot. 21: 13—20.

Chiapella. J. 1996. Nota sobre la identidad de Pratia re¬
pens (Campanulaceae: Lobelioideae). Bol. Soc. Argent.
Bot. 32: 123-135.

Elliot, W. B. & I). I.. Jones. 1990. Isotoma. Pp. 450—453
m Encyclopedia of Australian Plants Suitable for Cul¬
tivation, vol. 5. Lothian, Melbourne.

Hitchcock, A. S. & M. L. Green. 1929. Standard-species
of Linnean genera of Phanerogamae (1753—54). Pp.
1 11—199 m Anonymous (editor), International Botanical
Congress, Cambridge (England), 1930, Nomenclature,
Proposals by British Botanists. His Majesty's Stationery
Office, London.

Jeppesen, S. 1981. Lobeliaeeae. Pp. 9-170 in G. Harling
& B. Sparre (editors), Flora of Ecuador, no. 14. Swedish
Natural Science Research Council, Stockholm.

Lammers, T. G. 1992. New combinations for Asian Cam¬
panulaceae. Bot. Bull. Acad. Sin. 33: 285—287.

- . 1993. Chromosome numbers of Campanulaceae.

111. Review and integration of data for subfamily Fo-
belioideae. Amer. J. Bot. 80: 660-675.

- . 1995. Transfer of the southern African species of

Lightfootia, nom. i I leg., to Wahlenbergia (Campanula¬
ceae, Campanuloideae). Taxon 44: 333—339.

- . 1998a. New names and new combinations in

Campanulaceae. Novon 8: 31—35.

- . 1998b. Review of the neotropical endemics Hur-

meistera, Centropogon, and Siphocampylus (Campanu¬
laceae: Lobelioideae), with description of eighteen new
species and a new section. Brittonia 50: 233—262.

- . I). L. Kama & N. B. Morin. 1993. Campanula¬
ceae. Pp. 310-326 in F. Brako & J. F. Zaruechi (edi¬
tors), Catalogue of the Flowering Plants and Gymno-
sperms of Peru. Monogr. Syst. Bot. Missouri Bot. Gard.
45: i— xlii. 1 — 1286.

McVaugh, IF 1943. Campanulaceae (Lobelioideae). N.
Amer. FI. 32A: 1-134.

Melville. IF 1960. Contributions to the flora of Australia:
VI. The pollination mechanism of Isotoma axillaris
Lindl. and the generic status of Isotoma Lindl. Kew
Bull. 14: 277-279.

Moeliono, B. & P. Tuyn. 1960. Campanulaceae. Pp. 107-
141 in C. G. G. J. van Steenis (editor). Flora Malesiana
(ser. I) vol. 6, part 1. Noordhoff-Kolff, Djakarta.

Morris, K. E. & T. G. Lammers. 1997. Circumscription of
Codonopsis and the allied genera Campanumoea and
Ijeptocodon (Campanulaceae: Campanuloideae). I. Pal-
ynological data. Bot. Bull. Acad. Sin. 38: 277—284.

Murata, J. 1995. A revision of infrageneric classification
of Ijobelia (Campanulaceae— Lobelioideae) with special
reference to seed coat morphology. J. Fac. Sci. Univ.
Tokyo, Sect. 3. Bot. 15: 349—371.

Pepper, A. S.-IF, M. II. G. Gustafsson A V. A. Albert. 1997.
Molecular systematics of neotropical Lobelioideae

76

Novon

(Campanulaceae), with emphasis on Burmeistera, Centro-
ftogon, and Siphocampylus, and the utility of fruit and
lloral characters in lohelioid classification. Amer. J. Bot.
B4 (6, Suppl.): 222. [Abstract.]

Sehonland, S. 1889. Campanulaceae. Pp. 40—70 in A. En-
gler & k. Prantl, Die natitrlichen Pflanzenfamilien. En-
gelmann, Leipzig.

Wilbur, R. L. 1991. Synopsis of the Mexican and Central

American representatives of Lobelia section Tylomium
(Campanulaceae: Lobelioideae). Sida 14: 555—567.

Wimmer, F. E. 1943. Campanulaceae— Lobelioideae I. Teil.
Pp. i— viii, 1—260 in B. Mansfeld (editor), Das Pflanzen-
reich lV.276b. Wilhelm Engelmann, Leipzig.

- . 1953. Campanulaceae— Lobelioideae II. Teil. Pp.

i-viii, 2t>l— 814 in R. Mansfeld (editor). Das Pflanzen-
reich IV.2761). Akademie-Verlag, Berlin.

A New Combination in Nageia (Podocarpaceae)

Robert R. Mill

Royal Botanic Garden Edinburgh, 20A Inverleith Row, Edinburgh EH3 5LR, Scotland,

United Kingdom

Abstract. The new combination Nageia nan-
koensis (Hayata) R. R. Mill is proposed, and the
alfinities of the species with N. formosensis and N.
nagi are discussed briefly. Nageia nankoensis is en¬
demic to northern Taiwan, N. formosensis only to
the south tip ol the island; the latter is not native
in northern Taiwan as mis-stated by some authors.
It is shown that this error may have arisen because
of a wrong citation of the type of N. formosensis by
Li, which has been followed in later literature.

The revision of the gymnosperms for the Flora
of China is in progress. Following the views of
many authors, Nageia Gaertner (Podocarpaceae) is
here regarded as a genus separate from Podocarpus
L'Heritier ex Persoon. Indeed, Nageia was recently
raised to family rank, as Nageiaceae, by Fu (1992).
This ranking is not accepted here, pending further
work in progress on Podocarpaceae sensu lato at
the Royal Botanic Garden Edinburgh. However,
Fu’s generic limits of Nageia agree with those
adopted by me for the Flora oj China treatment.
De Laubenlels (1969, 1987) took a much broader
view of the genus and included various species
from Africa, South America, and the Australasia-
Pacific region, which Page (1989) segregated as the
genera Afrocarpus C. N. Page and Retrophyllum C.
N. Page.

In the Flora oj China, it is proposed to recognize
the taxon Podocarpus nankoensis Hayata at species
rank within Nageia. Most other necessary combi¬
nations in Nageia have been made by de Lauben-
fels (1987) and Page (1989). However, Podocarpus
nankoensis was regarded as a synonym of N. nagi
(including N. formosensis) by de Laubenfels, and of
N. formosensis by Page.

Nageia nankoensis (Hayata) R. R. Mill, comb,
nov. Basionym: Podocarpus nankoensis Hayata,
Icon. PI. Formos. 7: 39. 1918. TYPE: [Taiwan].
“Hab. Nanko, leg. B. Hayata Mai 1916'’ (ho-
lotype, TI not seen).

Nageia nankoensis and N. formosensis (Dumrner)
C. N. Page are both endemic to Taiwan. The latter
was based on Podocarpus formosensis Hummer, the

type of which ( Henry 1357, K) was collected by
Schmiiser at the extreme south tip of the island.
Unfortunately, Li and Keng (1954: 43) and Li
(1963: 40) have wrongly stated that the type of N.
formosensis is Henry 1446, which was collected at
Tamsui (Tanshui) at the north end of the island.
This has led to considerable confusion between N.
formosensis and N. nankoensis in some later liter¬
ature, including the sinking of the latter as a syn¬
onym of the former by Page (1989) and the citation
of N. formosensis as occurring in northern Taiwan
by Li (1963). The two plants are in fact separable
by leaf and branch characters and are here recog¬
nized as separate species, as was done under Po¬
docarpus by Hu (1964: 33, 35). This has necessi¬
tated a new combination in Nageia for Podocarpus
nankoensis. Hu’s distribution information for both
species is correct.

Nageia formosensis is distinguishable from N.
nankoensis and N. nagi (Thunberg) Kuntze, by its
leaf morphology: many leaves are obovate or ob-
ovate-elliptic rather than lanceolate or ovate-ellip¬
tic, and those that are obovate have very broad,
truncate tips, which are frequently blackened
(Dummer likened this character to them having
been touched by a hot iron). Obovate leaves have
not been seen in any specimens of N. nagi or N.
nankoensis, although leaf shape in both is variable,
particularly in length: width ratio. (This has led to
the recognition of varieties within N. nagi as rep¬
resented in Japan). Nageia formosensis, which only
occurs in southern Taiwan, has been recently re¬
duced to varietal rank, as N. nagi var. formosensis
(Dummer) J. Silba (Silba, 1990: 38), but this seems
too low a rank; if a broad concept of N. nagi is
preferred, N. formosensis should be recognized as a
subspecies of it.

Nageia nankoensis is more similar to N. nagi
than it is to N. formosensis but tends to have nar¬
rower, more distinctly lanceolate leaves. It is en¬
demic to northern Taiwan. Specimens from Hainan
that have been identified as N. nankoensis (e.g.,
McClure 8131 (E, K) from Nodoa; C. I. Lei 745 (K)
from Ku Tung Village; W. T. Tsang 1 7430 (E, K)
from Paak Shek Shan) all have broader, much more

No von 9: 77-78. 1999.

78

Novon

broadly obtuse leaves than any plants of N. nan-
koensis from Taiwan. They have been compared
with specimens of N. nagi from Japan and China
and showed no significant differences. Therefore,
these Hainan plants are here considered to belong
to N. nagi. This accords with the identifications of
Hu (1964: 35—36), who came to the same conclu¬
sion.

Both Nageia fornwsensis and N. nankoensis were
included in the long synonymy of N. nagi (as De-
cussocarpus nagi (Thunberg) de Laubenfels) by de
Laubenfels (1969: 357). His concept of N. nagi,
and of other species of Nageia, was very broad.

Acknowledgments. The Royal Botanic Garden
Edinburgh is supported by The Scottish Office Ag¬
riculture, Environment and Fisheries Department.
1 thank the directors of BM and K lor allowing me

to examine material, including the type of Nageia

formosensis.

Literature Cited

De Laubenfels, I). J. 1969. A revision of the Malesian and
Pacific rainforest conifers, 1. Podocarpaceae, in part. J.
Arnold Arbor. 50: 274—369.

- . 1987. Revision of the genus Nageia (Podocar¬
paceae). Rlumea 32: 209—21 1.

In. D. /.. 1992. Nageiaceae — A new gymnosperm family.
Acta Phytotax. Sin. 30: 515—528.

Hu, S. Y. 1964. Notes on the flora of China IV. Taiwania
10: 13-62.

Li, H. L. 1963. Woody Flora of Taiwan. Livingston Pub¬
lishing. Narberth, Pennsylvania.

- & H. Keng. 1954. leones Gymnospermum For-

niosanarum. Taiwania 5: 25—83.

Page, C. N. 1989. New and maintained genera in (be co¬
nifer families Podocarpaceae and Pinaceae. Notes Roy.
Rot. Gard. Edinburgh 45: 377—395.

Silba, J. 1990. A supplement to the International Census
of the Coniferae II. Phytologia 68: 7—78.

A New Species of Larix (Pinaceae) from Southeast Tibet and Other
Nomenolatural Notes on Chinese Larix

Robert R. Mill

Royal Botanic Garden Edinburgh, 20A Inverleith Row, Edinburgh, EH3 SLR, Scotland, United

Kingdom

Abstract. Larix kongbuensis, a new species from
southeastern Xizang (China) near its borders with
Arunaohal Pradesh (India), is described. Previously
misidentified as L. mastersiana, its alfinities with
that Sichuan endemic and with L. griffithii (eastern
Himalaya) and L. speciosa (Yunnan. Myanmar) are
discussed. Within L. potaninii Batalin, variety aus¬
tralis A. Henry ex Handel-Mazzetti has been found
to be an earlier name for variety macrocarpa Y. W.
Law. The nomenclature of the “Sikkim Larch” is
discussed, and it is concluded that Lirix griffithii
J. D. Hooker, not L. griffithiana Carriere, is the
correct name for this species.

In the course of revising the genus Larix for the
Flora of China , it was discovered that three spec¬
imens from southeastern Xizang named by their
original collector as L. mastersiana were not that
species. Because they do not match any other spe¬
cies in the group with exserted, reflexed bract
scales, they are described below as L. kongboensis.

Larix kongboensis R. R. Mill. sp. nov. TYPE:
China. Xizang: Tamnyen La, 29°18'N,
94°45'E, alt. 11,000 ft. [3353 ml, 21 July
1938, F. Ludlow, G. Sherriff & G. Taylor 4914
(holotype, E; isotype, BM; female cones and
mature leaves of current year).

Affinis /,. griffithii J. I). Hooker bractearum apicibus
sursum eurvatis set! foliis niulto minoribus, aestate tanturn
9—22 mm (non 30—55 mm) fascieulis densioribus foliorum
40—50 (non 28—36), brachyblaslis fertilibns loliatis, stro-
bilis maseulis multo brevioribus (6—8 mm, non 10—16
mm) rubescentibus (hand flavescentibus), strobilis femi-
neis maturis 2-plo (non 2.5— 3-plo) longioribus quam la-
tioribus differt. A txirice mastersiana Rehder & Wilson
apicibus brachyblastorum glabris (non densiuscule flavi-
dopilosis). brachyblastis brevioribus et angustioribus, epi-
dermide foliorum epapillosa, foliis in fascieulis magis nu-
merosis (40—50, non 25—10). et apicibus bractearum
sursum eurvatis recedit.

Trees, 9—25 m tall. Bark and branching habit
unknown. Long shoots of first year not seen, of sec¬
ond year pale reddish brown, glabrous, of subse¬
quent years pale gray. Short shoots 2-1.5 X 4.5-6

mm, broadly obconical; leal cushions glabrous.
Leaves in fascicles of 40—50, linear-oblong, straight
or slightly falcate, 6—12 mm in spring, finally 9—22
X 0.6— 1.1 mm; midrib raised adaxially only prox-
iinally; stomatal lines 3—5 on each side of midrib
abaxially, none adaxially; epidermis not papillose.
Cone-bearing short shoots with leaves; cones short
pedunculate (peduncle not more than 4 mm). Pol¬
len cones erecto-patent at 45°, reddish, broadly
conical, 6—8 X 5.5— 6.5 mm. Seed cones at flow¬
ering stage not seen. Cone deflexed outward from
main shoot axis, gray-brown when ripe, oblong-el¬
lipsoid, 4.5—5 X 2.2— 2.5 cm; seed scales broadly
obovate-reniform, median ones ca. 8 X 10 mm,
apex broadly rounded, entire or shallowly retuse,
abaxial surface minutely whitish pubescent; bract
scales exserted, lanceolate, exposed part ca. 8 X
5.5 mm, strongly reflexed, tapered gradually then
finally more abruptly into a cusp; cusp 1.5—1 mm,
turned upward at tip. Seeds not seen. Flowering
April. Fruiting late summer— autumn.

The three specimens cited and described as the
new species Larix kongboensis were all originally
named as L. mastersiana Rehder & Wilson, and one
of them is also mentioned under that name in a
quotation from Ludlow and Sherriff’s journal
(Fletcher, 1975: 88), where the trees concerned are
described as “a very fine larch.” The area of south¬
eastern Tibet (Xizang, China) in which the speci¬
mens were collected lies in Kongbo, from which the
epithet is derived. This area, close to the border
with Arunaohal Pradesh (northeastern India), has
yielded numerous new endemic species, primarily
as a result of the collecting expeditions by Ludlow
et al. and by F. Kingdon-Ward.

Larix mastersiana Rehder & Wilson is an en¬
dangered species, endemic to a small area of west¬
ern Sichuan. It differs from L. kongboensis in sev¬
eral characters, notably the pale yellowish-pilose
leaf “cushions,” longer and relatively narrower
short shoots, fewer (25—40) leaves per fascicle
which have a papillose epidermis, and the bract
scales whose apical cusp is not turned upward at
the tip.

Novon 9: 79-82. 1999.

80

Novon

birix kongboensis is closer to L. griffithii in its
characters and distribution, and the latter occurs in
Bhutan, Nepal, and Sikkim, and has been recorded
from Xizang (Law et al., 1978: 171-173). It is pos¬
sible that at least some of the Xizang records of L.
griffithii actually belong to L. kongboensis. birix
griffithii differs by its fewer, much longer leaves (3-
5.5 cm), its usually leafless cone-bearing short
shoots, much larger, yellowish pollen cones, and
seed cones twice as long and 2.5— 3 X as long as
broad (not 2X as long as broad). Characters L.
kongboensis shares with L. griffithii but not L. mas-
tersiana include the upturned apex ol the bract
scales, the non-papillose leaf epidermis, and the
glabrous leaf cushions.

The only other Chinese larch with reflexed bract
scales is Ixirix speciosa W. C. Cheng & Y. W. Law.
Specimens of this have been seen from Yunnan as
well as northern Myanmar; again, it has been re¬
corded from Xizang but verification is necessary. It
is easily separable from L. kongboensis by its stout,
often very long short shoots (up to 3 cm), longer
leaves (2.5— 5.5 cm), pollen cones much larger (1.6-
2.2 cm), chocolate-colored mature cones, and bract
scales whose apical cusp is not turned upward at
the tip.

Paralypes. CHINA. Southeastern Xizang: I.uriang,
Kong Chu, Kongbo, 29°44'N, 94°48'K, 10,500 ft. [32(H)
rn], fairly common at the upper limit of spruce and pine,
12 Apr. 1947. F. Ludlow , G. Sheriff & G. Taylor 12.176
(KM. K; male cones and old female cone of previous year);
Natranipa-I.ung, Cliavul (dm. 11.5(H) ft. [3505 m], tree
30-80 ft., rocky slopes and rhododendron and bamboo
jungle, transition zone, 27 Apr. 1936. F. Ludlow & G.
Sherriff 1267 (K, KM; spring leaves).

Larix potaninii Batalin var. australis A. Henry
ex Handel-Mazzetti, Symb. Sin. 7(1): 14. 1929.
TYPE: China. NW Yunnan: “prope fines Ti-
beto-Birmanicas inter Huvios Lu-djiang (Sal¬
ween) et Djiou-djiang (Irrawadi or. sup.), in
jugi Tschiangschel, 27°52'[N], lateris utri-
usque regione temperata in silvis mixtis,
substr. micoschistaceo,” 2850-3500 m, 2 X 5
July 1916, //. F. von Handel-Mazzetti 9176
(Diar. Nr. 1754,[17]70) (holotype, W not seen;
isotype, E). (The fourth character of the Diar.
Nr. is difficult to decipher but is assumed to
be a 4.)

birix potaninii Katalin var. maerocarpa V. W. Law, in W.
C. Cheng, I.. k. Fu X C. Y. Cheng, Acta Phytotax.
Sin. 13(4): 84. 1975. Syn. nov. TV PK: China. Central
Yunnan: 28(H)— 36(H) m. Feng Guomei 9347 (holo¬
type, PE not seen).

Augustin Henry's neglected varietal epithet ap¬

pears as a very brief note in Handel-Mazzetti’s ac¬
count of birix (Handel-Mazzetti, 1929), but as the
minuscule description mentions that the cones are
7 cm long, it is sufficient to equate the name with
Y. W. Law’s variety maerocarpa based on material
from Midi, southwestern Sichuan. The Edinburgh
isotype (Handel-Mazzetti 91 76) consists of two
packets mounted on a herbarium sheet, one con¬
taining only leaves and bearing a label “ birix Po¬
taninii Bat. var. nova australis Henry det. A. Hen¬
ry,” the other containing only cones and unlabeled,
but with a loose ticket within labeled “ad No. 9176
birix Potaninii var. australis."' This ticket seems to
be a discarded label (for Handel-Mazzetti 8323 )
that has been torn in half and re-used. Three of the
four cones measure 4. 8-5. 5 cm; the fourth is slight¬
ly incomplete. They are thus at the minimum of the
range allowed lor variety maerocarpa by Y. W. Law,
and at the maximum found in variety potaninii.
They are also smaller than variety australis as de¬
scribed by Handel-Mazzetti.

Larix griffithii J. D. Hooker, Himal. Journal 1:
255-256 (as “Saar,” without Latin name), 2:
44 & 481 (Jan. 1854). SYNTYPES: [Nepal]
Nango Mt., J. D. Hooker (K); [Sikkim] Lachen,
J. D. Hooker & Thomson (K).

Abies griffithiana J. I). Hooker ex Lindley & Gordon, J.
Roy. Ilort. Soc. London 5: 214. 1850, nom. subnud.
No type designated.

larix griffithiana hort. ex Carrie re, Traile (Jen. Conif. 278.
Jim. 1855.

The name of the Sikkim larch has vacillated be¬
tween birix griffithiana and L. griffithii almost
since the species was first named, due in part to
the two epithets appearing within one month of
each other. There is disagreement even in the most
recent treatments: Grierson and Long (1983) used
L. griffithiana Carriere, Earjon (1990) L. griffithi¬
ana (Lindley & Gordon) Carriere, and Silba (1986)
L. griffithii Hook. f. A similar confusion exists when
several popular books on trees are compared.

Carriere 's epithet griffithiana originally appeared
as Abies griffithiana J. D. Hooker ex Lindley &
Gordon (1850). However, Lindley and Gordon’s de¬
scription merely reads “A tree 40 to 60 feet high
(The Sikkim Larch)”; the distribution, “Eastern Ne¬
pal, Sikkim,” is also given. This description con¬
tains no characters that diagnose this larch against
any other species of birix, or indeed against any
other tree. It is not considered enough to validate
Li nc n< >y and Gordon’s name, and the epithet griffi¬
thiana should be considered to date from Carriere ’s
first adequate description, published in June 1855.

Hooker’s epithet griffithii is generally considered

Volume 9, Number 1
1999

Mill

Larix kongboensis from Tibet

81

to date from July 1855, when it appeared in Hook¬
er’s Illustrations oj Himalayan Plants (Hooker,
1855) as Larix griffithii J. D. Hooker & Thomson;
Hooker himself (1888) c ited this reference as the
place of publication, with L. griffithiana Gordon,
Abies griffithiana Lindley & Gordon, and Pinus
griffithii Parlatore listed as the only synonyms. It
would thus be antedated, by one month, by L. grif¬
fithiana Carriere, of whose publication Hooker (and
also Bean, 1908) seems to have been unaware;
hence the preponderance of use of griffithiana in
most, but not all. recent literature, including Flora
Reipublicae Popularis Sinicae.

However, Hooker also published the name in his
Himalayan Journals (Hooker, 1854), both volumes
of which were published in January 1854, 18
months before Larix griffithiana Carriere. In vol¬
ume 1, pp. 255—256, there is a detailed description
of a larch, with two vernacular names, but no Latin
epithet. Hooker had discovered the tree at Nango
Mountain, eastern Nepal (Bean, 1908). The de¬
scription reads as follows: ‘'‘'December 5 [1848].—
. . .We descended at first through rhododendron and
juniper, then through black silver-fir ( Abies Webbi-
ana), and below that, near the river, we came to
the Himalayan larch; a tree quite unknown, except
from a notice in the journals of Mr. Griffith, who
found it in Bhutan. It is a small tree, twenty to forty
feet high, perfectly similar in general characters to
a European larch, but with larger cones, which are
erect upon the very long, pensile, whip-like branch¬
es; its leaves,— now red— were falling, and covering
the rocky ground on which it grew, scattered
amongst other trees. It is called ‘Saar’ by the Lep-
chas and Cis-Himalayan Tibetans, and ‘Boarga-sel-
la’ by the Nepalese, who say it is found as far west
as the Cosi river; it does not inhabit Central or West
Nepal, nor the North-West Himalaya.”

The first actual mention of the name Ixirix grif¬
fithii is in volume 2, p. 44, which is the usual ci¬
tation found in the literature, generally qualified by
the statement “nomen” or “nom. nud.” (e.g.. Bean,
1908). Hooker was here describing the conifers of
a steep hill in the Lachen valley, near a point where
the river divides, and where some stacks of differ¬
ent conifer woods were stored for export to Tibet.
Discussing this wood pile. Hooker wrote, “Of these
the larch (Larix Griffithii, ‘Sail’) splits well, and is
the most durable of any; but the planks are small,
soft and white.”

The vernacular name “Sah” is presumably a dif¬
ferent rendering of “Saar" used in volume 1, p.
255, and could be used to argue that he was refer¬
ring in both places to the same taxon. However,
what is more significant is that it is (dear from the

index to the whole work that the plant supplied with
a full description but no Latin name in the first
volume belongs to the same species as the one giv¬
en a name but only a description of its sawn planks
in the second. The index entries on p. 481 of vol¬
ume 2 read:

Larch, Himalayan, i. 255; sketch of, ii. 55.

Larix Griffithii, i. 255; ii. 44.

From these it is obvious that Hooker knew that
his description of “Saar” in the first volume was of
Larix griffithii, which should thus not be consid¬
ered a nomen nudum in this publication but vali¬
dated by an indirect reference (Art. 32.5 of ICBN,
Greuter et al., 1994), via the index, to the descrip¬
tion in volume 1. Since the two volumes were si¬
multaneously published, the name can be taken to
date from January 1854. Larix griffithii Hooker is
thus the correct name for the Sikkim Larch. “Sik¬
kim Larch” (as used by Mitchell, 1974) is a better
English name for this species than “Himalayan
Larch,” which is sometimes used (as for example
by Hooker), because of the existence of another
Himalayan larch, L. himalaica W. C. Cheng & L.
K. Fu.

As Hooker (1854) mentioned incidentally, Larix
griffithii was in fact first discovered by William
Griffith, on 10 May 1838, in Bhutan, “above Woo-
lookha, 9,600 ft.” Griffith (1848: 189, no. 1010)
gave a short Latin description: “Arbor parva, ra-
mulis elongatis pendulis, foliis plurimis complan-
atis, conis ascendentibus, squamis rubro fuscis,
longe apiculatis, apiculo viridi deflexo.” However,
his entry is merely labeled “Larix sp.” The sup¬
porting specimen is at K.

Acknowledgments. The Royal Botanic Garden
Edinburgh is supported by The Scottish Office Ag¬
riculture, Environment and Fisheries Department.

I .iterature Cited

Mean. W. J. 1908. larix Griffithii. Bot. Mag. 134: t. 8181.
Karjon, A. 1990. Pinaceae. Drawings and Descriptions of
the Genera Abies, Cedrus, Pseudolarix, Keteleeria, No-
tholsuga, Tsuga, Cathaya, Pseudotsuga, Ixirix and Pi-
cea. Ilegnum Veg. 121.

Fletcher, II. 1975. A Quest of Flowers. Edinburgh Univ.
Press. Edinburgh.

Greuter, W., F. R. Barrie, II. M. Burdet, W. G. Chaloner,
V. Deinoulin, I). F. Hawksworth. P. M. J0rgensen. I). II.
Nicolson. P. C. Silva. P Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Grierson, A. J. C. & 1). G. Long. 1983. Flora of Bhutan,
Vol. I. Royal Botanic Garden, Edinburgh.

Griffith, W. 1848. Posthumous Papers Bequeathed to the
Honourable, the East India Company, and Printed by
Order of the Government of Bengal. Vol. 11. Itinerary
Notes of Plants Collected in the Khasyah and Bootan

82

No von

Mountains, 1837—38, in Afghanistan and Neighbouring
Countries, 1839 to 1841. By the late William Griffith.
Esq., E.L.S. Arranged by John McClelland, F.L.S. Bish¬
op’s College Press, Calcutta.

Handel-Mazzetti, H. F. von. 1929. Symbolae Sinicae Bo-
tanische Ergebnisse der Expedition der Akademie der
Wissenschaften in Wien nach Siidwest-China. 1914/
1918. Part 7(1). Julius Springer, Wien.

Hooker, J. I). 1854. Himalayan Journals; or. Notes of a
Naturalist in Bengal, the Sikkim and Nepal Himalayas,
the Khasia Mountains, &c. John Murray, London.

- . 1855. Illustrations of Himalayan Plants chiefly

selected from drawings made for J. F. Cathcart Esq. of
the Bengal Civil Service. Privately published, London.

(Frequently attributed, wrongly, to Cathcart in the lit¬
erature.]

- . 1888. Coniferae. In: J. I). Hooker, Flora of British

India 5: 643—655.

Law. Y.-W., W.-C. Cheng & L. K. Fu. 1978. hirix. In:
Cheng W.-C. & L.-k. Eu (editors), FI. Reipubl. Popu-
laris Sin. 7: 168-196.

Lindley, J. & G. Gordon. 1850. A catalogue of coniferous
plants, with their synonyms. J. Hort. Soc. London 5:
193-228.

Mitchell, A. 1974. A Field Guide to the Trees of Britain
and Northern Europe. Collins, London.

Silba, J. 1986. Encyclopaedia Coniferae. Phytologia Mem¬
oirs VIII. Corvallis, Oregon.

Hylaea (Apocynaceae-Apocynoideae),
a New Genus from South America

./. Francisco Morales

Institute Nacional de Biodiversidad (INBio), Apto. 22-3100, Santo Domingo de Heredia,

Costa Rica

Abstract. The genus Hylaea, restricted to Ve¬
nezuelan— Brazilian Amazonia, is described here.
Two new combinations, Hylaea arborescens and //.
leptoloba, are proposed.

As a result of recent study of material of Pres-
tonia (Apocynaceae) from throughout its range and
during preparation of the accounts of the Apocy¬
naceae for the Flora Mesoamericana and the Man¬
ual to the Plants oj Costa Rica, it became evident
that the delimitation of Prestonia needed critical
study and that several nomenclatural adjustments
were appropriate. Recently, some effort toward res¬
olution of these nomenclatural problems in Presto¬
nia has been made (Morales, 1996, 1997a, 1997b).

Prestonia leptoloba and P arborescens, described
and included in Prestonia by Monaehino (1957,
1961), possess a different corolla structure (sta¬
mens completely exserted, absence of the annular
corona in the orifice of the corolla tube) that does
not allow inclusion of these taxa in Prestonia.
Therefore, a new genus is here created with two
new necessary combinations.

Hylaea J. F. Morales, gen. nov. TYPE: Hylaea lep¬
toloba (Monaehino) J. F. Morales.

Herba volubilis vel arbor parva, Prestoniae et Lauber-
tiae affinis; folia opposita petiolata ad coriacea eglandu-
lifera; infforescentia subcorymbosis lateralis pedunculata
multiflora; corolla regularis 5-partita infundibuliformis vel
hypoerateriformis appendieibus epistaminalibus exserta
inconspieue stamina exserta puberulus annulo faucium
absens; fructus ignotus.

Suffruticose lianas or small trees. Stems subter-
ete to somewhat flattened, inconspicuously puber-
ulent to glabrate, inconspicuously lenticellate, latex
unknown. Leaves opposite, coriaceous, glabrous,
colleters absent at the base of the leaf blade, petiole
slightly fused at the base, mostly with colleters in
the axils. Inflorescence a subcorymbose or subco-
rymbose-umbelliform cyme, axillary, minutely pu-
berulent to glabrous or glabrate, many-flowered,
usually long-pedunculate, bracts foliaceous or scar-
ious. Flowers: sepals scarcely imbricate at the base.

bearing a solitary episepalous colleter within, sub¬
entire to inconspicuously erose, corolla infundibu-
liform or salverform, glabrous without, free corona
appendages, completely exserted, inserted in the
orifice of the corolla mouth, the limb 5-parted, ae-
tinomorphic, dextrorsely convolute, stamens 5, to¬
tally exserted, infrastaminal indument within, an¬
thers connivent and agglutinated to the pistil head,
densely puberulent, consisting of 2 parallel, uni¬
formly fertile thecae borne adaxially near the apex
of an enlarged, sagittate, 2-auriculate, sterile part
of the anther, auricles short, acute to short-acumi¬
nate, filaments short, puberulent to pilose, carpels
2, united at the apex by a common stylar shaft sur¬
mounted by the fusiform style head, ovary glabrous,
ovules numerous, multiseriate, nectaries 5, sepa¬
rated. Follicles unknown.

Hylaea comprises two species restricted to Ve¬
nezuelan— Brazilian Amazonia. The generic name
recalls the area where the plants were collected.

Kky to the Species ok IIyiaka

la. Tree; leaf blades 6— 10. 5(— 12) X 2.4 — 4( — 5) cm;
inflorescence bracts 0. 1-0.2 cm, scarious ....

. //. arborescens

lb. Liana; leaf blades (13-)26-.'l() X (6-)12-13.5

cm; inflorescence bracts 0.7—1 cm. foliaceous,
early caducous . H. leptoloba

Hylaea arborescens (Monaehino) J. F. Morales,
comb. nov. Basionym: Prestonia arborescens
Monaehino, Mem. New York Bot. Card. 10:
63—64, fig. 43E-M. 1961. TYPE: Venezuela.
Amazonas: Rio Guainfa, occasional along
Cano San Miguel just above Limoncito, 15 km
from Rfo Guainfa, 100-400 m, 28 June 1959
(fl), Wurdack & Adderley 43232 (holotype, NY;
isotypes, F, 1NB, MO, VEN; photograph, INB
ex F).

Tree 4—5 m tall; branchlets inconspicuously pu¬
berulent to glabrous or glabrate, inconspicuously
lenticellate. Leaves: petiole 0.4— 0.8(— 1) cm; blade
6— 10. 5(— 12) X 2.4-4(— 5) cm, coriaceous, elliptic
to oblanceolate, obtuse to acute at the apex, acute

Novon 9: 83-85. 1999.

84

Novon

basally, glabrous; secondary veins slightly im¬
pressed. veiidets inconspicuous. Inflorescense sub-
corymbose-umbelliform, axillary, ± agglomerate,
many-flowered, minutely puberulent to glabrous or
glabrate; peduncle 4.5-8.5(-10) cm; pedicels 2.5-
5( — 6) cm; bracts ca. 0.1— 0.2 cm, searious. Flowers:
sepals 2—3 X 1.5—3 mm, ovate, acute to obtuse,
subcoriaceous, minutely puberulent to glabrate;
colleters ca. 0.5 mm long, subentire to inconspic¬
uously erose; corolla narrowly inlundibuliform,
white and pink-tinged, glabrous; tube 8-10 X 3.5—
4 mm; corona lobes less than 1 mm, annular corona
absent; lobes 0.9-1. 2(-l. 5) X 0.6-0.9(-1.2) cm,
obovate; anthers 6 mm, densely puberulent, totally
exserted; ovary 1—1.5 mm long; style head ca. 1.5
mm long; nectaries about equaling the ovary, sep¬
arate, entire. Follicles unknown.

This species, known only from the type collec¬
tion, is restricted to Venezuelan Amazonia. It is
characterized by its arborescent habit, coriaceous
leaves, free corona appendages, lack ol an annular
corona, and completely exserted anthers. The most
important characters on which the segregation of
Hylaea from Prestonia are based are the absence
of an annular corona and the totally exserted an¬
thers. The close affinity of Hylaea with Prestonia is
exemplified by the free corona lobes and corymbose
to subcorymbose inflorescences, present in both
genera.

Although Monachino (1961) and Morillo (1978,
1995) included t his species within Prestonia , care¬
ful examination of the type collection showed sig¬
nificant differences with respect to the main char¬
acters of that genus. Both authors mentioned the
presence of a papillose faucal annulus; however,
careful examination of the flowers revealed that it
is absent.

In Prestonia the annular corona is always situ¬
ated in the corolla mouth and below this, within the
tube, are the free corona lobes (when present) or
the callus ridges. In the flowers ol Hylaea arbores-
cens, the appendages are superimposed over the
purported annular corona, which is really composed
only of a thickening of the adpressed pubescence
of the corolla tube within: a true annular corona is
not present.

Following Morales (1997a), the presence of an
annular corona is one of the most important features
that separate Prestonia and Laubertia from other
closely related genera. Both genera are easily dis¬
tinguished by this character, which is present in all
the species. Other species (e.g., P. woodsoniana
(Monachino) A. H. Gentry, P. caudata Woodson,
and P agglutinata (Jacquin) Woodson) previously

included in Prestonia but which lack an annular
corona were recently transferred to Echites (Mo¬
rales, 1997b).

In some Mesoameriean species of the section 7o-
mentosae (e.g., P clandestina J. F. Morales, P. mexi-
cana A. DC., P. speciosa Donnell Smith), the an¬
nular corona is conspicuously five-lobed and
somewhat reduced, but is always present. In some
South American species (e.g., P. solanifolia (Muller
Argoviensis) Woodson) the annular corona is very
reduced but always visible. Therefore, following the
generic concepts put forward by Woodson (1933,
1936) and Piehon (1950), and in conjunction with
the foregoing rationale, Prestonia arborescens Mon¬
achino should not be included within Prestonia be¬
cause other closely related genera such as Lau¬
bertia would then also be drawn in to Prestonia.

The genus Hylaea can be separated from most
closely related genera by the following key:

la. Anthers totally exserted; annular corona absent

. // ylaea

lb. Anthers included or with the apices partially ex¬
serted, never totally exserted; annular corona
present, conspicuous or inconspicuous.

2a. Sepals with colleters within .

. Prestonia (including Rhodocalyx )

2h. Sepals without colleters within .... Ixiubertia

During the course of an examination of Prestonia
types from South America, another species of Hy¬
laea was found. Previously included in Prestonia
by Monachino (1957), it is characterized by its
scandent habit and larger leaves. Therefore, a new
combination is necessary and is proposed here.

Hylaea leptoloba (Monachino) J. F. Morales,
comb. nov. Basionym: Prestonia leptoloba
Monachino, Phytologia 6: 12. 1957. TYPE:
Brazil. Amazonas: Iraruca, I^ana, 16 Nov.
1945 (H), Erdes 21d9I (holotype, NY; isotypes,
IAN, INB; photocopy, INB ex IAN).

Liana; branehlets inconspicuously puberulent to
glabrate, inconspicuously lenticellate. Leaves: pet¬
iole 0. 1-0.2 cm; blade (13— )26— 30 X (6-)12-13.5
cm, coriaceous, elliptic to broadly elliptic, sub-
abruptly acuminate at the apex, obtuse basally, gla¬
brous; secondary veins slightly impressed, veinlets
inconspicuous. Inflorescence subcorymbose, axil¬
lary, ± agglomerate, many-flowered, glabrous to
glabrate; peduncle 4.5-6 cm; pedicels 2-5.5 cm;
bracts 0.7—1 cm, foliaceous, early caducous. Flow¬
ers: sepals 2-3 X 1.5—2 mm, ovate, acute, subco¬
riaceous, minutely puberulent to glabrate; colleters
ca. 0.5 mm long, subentire; corolla salverform,
white and purple-tinged, glabrous; tube 9-10 X 3-
4 mm; corona lobes less than 1 mm; annular corona

Volume 9, Number 1
1999

Morales

Hylaea from South America

85

absent; lobes 0.9—1 X 0.5-0. 6 cm, narrowly ob-
ovate; anthers ca. 6 mm, densely puherulent; ovary
less than 1 mm; style head ca. 1.5 mm; nectaries
surpassing the ovary, separate, entire. Follicles un¬
known.

Known only from the type collection. Hylaea lep-
toloba is closely related to //. arborescens, but dif¬
fers by its scandent habit, larger leaves, and con¬
spicuously foliaceous inflorescence bracts.

Acknowledgments. I diank the curators and di¬
rectors of A, AAU, ALCB, B, BHCB, BM, BR. C,
CAS, CEN, COAH, COL, DS, DUKE, ECON,
ESAL, F. G, GH, GOET, K, HB. HBG, HRB. HUA,
IAN, INPA, LZ, M, MG, MO, NY, 0, QCA, QCNE,
TRIN, R, RB, S, U, UB, UC, UPS, US, USF, USJ,
W, VIC, and Z for the loan of material. I thank
Michael Grayuin for critical comments on the man¬
uscript.

Literature Cited

Monachino, J. 1957. A new species of Prestonia from Bra¬
zil. Phytologia 6: 12—13.

- . 1961. Botany of the Guyana Highlands. Mem.

New York Bot. Card. 10(4): 63-4)4.

Morales, J. F. 1996. Novelties in Prestonia (Apoeynaceae).
Novon 6: 285—287.

- . 1997a. A synopsis of the genus Prestonia sect.

Tomentosae in Mesoamerica. Novon 7: 59-66.

- . 1997b. A re-evaluation of the genera Echites and

Prestonia sect. Coalitae (Apoeynaceae). Britton ia 49:
328-336.

Morillo, G. 1978. Estudio preliminar de las espeeies ve-
nezolanas de Prestonia (Apoeynaceae). Mem. Soc. Ci.
Nat. La Salle 110(38): 195-226.

- . 1995. Prestonia. Pp. 548—552 in J. Steyermark,

P. Berry & B. Holst (general editors). Flora of the Ve¬
nezuelan Guyana, vol. 2. Timber Press, Portland,
Oregon, and Missouri Botanical Garden, St. Louis.
Piehon, M. 1950. Classification des Apocynacees XXV.
Echitoidees. Mem. Mus. Natl. Hist. Nat. Set. B, Bot.
1(1): 1-144.

Woodson, R. E. 1933. Studies in the Apoeynaceae IV. The
American genera of Echitoideae. Ann. Missouri Bot.
Card. 20: 606-790.

- . 1936. Studies in the Apoeynaceae IV. The Amer¬
ican genera of Echitoideae XXVI. Prestonia. Ann. Mis¬
souri Bot. Gard. 23: 276—367.

A New Species of Macoubea (Apocynaceae) from Mesoamerica

J. Francisco Morales

lnstituto Nacional de Biodiversidad, Apto. 22-3100, Santo Domingo de Heredia, Costa Rica

Abstract. A new species of Macoubea from Cos¬
ta Rica and Panama is described and illustrated.
Its relationships with the two South American spe¬
cies of the genus are discussed.

Macoubea has been considered to comprise just
two primarily South American species, M. guianen-
sis and M. sprucei (Allorge, 1985), with some spe¬
cific varieties. Nevertheless, recent work on the
Apocynaceae for “The Manual to the Plants of Cos¬
ta Rica” and the Flora Mesoamericana demonstrate
the existence of a third species, endemic to Me¬
soamerica, which is described here.

M acoubea mesoamericana J. F. Morales, sp. nov.
TYPE: Costa Rica. Limon: Canton de Tala-
manca, Amubri, camino lastreado al SO del
pueblo, 24 June 1989, Hammel et al. 17516
(holotype, INB; isotypes, CR, F, MO, NY).

[Specimens have been distributed as Macoubea
sprucei (Muller Argoviensis) Markgraf.]

Arbor parva. Kolia elliptica, 1 2— 19( — 21.5) X 5-12.5
(-14) cm, glabra, subcoriacea. Inflorescentiae terminales,
sepala 1.5—2 mm, eorollae tubo 3—5 mm, lobis 7-9 X 4
mm. Fruetus 4.5—7 X 2.fi — 3.6 cm, seminibus 1.3— 1.5 X
0.4— 0.5 cm.

Trees 10—50 in high; branchlets terete to subter-
ete, glabrous to glabrate. Leaves petiolate; petiole
glabrous, 1—3 cm long; blade 12— 19(— 21.5) X 5—
12. 5(— 14) cm, firmly membranaceous to subeoria-
ceous, broadly elliptic; acuminate to acute at the
apex, cuneate basally; glabrous on both sides; sec¬
ondary veins brochidodromous; veinlets conspicu¬
ously reticulated. Inflorescence corymbose, many-
flowered, agglomerate, terminal, rarely subterminal,
glabrous to glabrate; peduncle 0.8— 3.4 cm long;
pedicels 0.2— 0.6 cm long; bracts ovate, ca. 1 mm
long, fimbriate, conspicuous. Flowers greenish yel¬
low, yellowish to white; sepals broadly ovate, 1.5—
2 mm, obtuse to rounded, imbricate, glabrous to
glabrate, marginally ciliate, bearing many colleters
inside; corolla salverform, tube 3—5 mm long, ca. 2
mm wide, sparsely puberulent to glabrate without;
lobes obliquely obovate to narrowly obovate, 7—9 X
4 mm, spreading to somewhat reflexed; stamens in¬
serted near the base of the corolla tube; filaments

sparsely puberulent; anthers narrowly oblong, 4—
4.5 mm, glabrous, acuminate, sagittate at tbe base;
ovary ca. 2 mm, puberulent, 2-carpellate; pistil
head ca. 1 mm. Fruits 4.5-7 X 2.8— 3.6 cm, hemi-
syncarpous, moderately to obscurely bilobed, the
locules rounded; seeds 1.3— 1.5 X 0.4— 0.5 cm, ob¬
long, finely rugose, minutely pitted and naked.

Macoubea mesoamericana is known only from
southeastern Costa Rica, on the Atlantic slopes of
the Cordillera de Talamanca, and Panama, near El
Llano— Cartf road, at 200—500 m. Flowering Sep¬
tember— March. Fruiting November— April.

Although this species was previously considered

4 cm

Figure 1. Fruits of Macoubea mesoamericana J. F. Mo¬
rales. — A. Herrera 3343. — B. Chacdn 1 35.

Novon 9: 86-418. 1999.

Volume 9, Number 1
1999

Morales

Macoubea mesoamericana

87

4cm

B.

Figure 2. Fruits of Macoubea sprucei (Muller Argoviensis) Markgraf. — A. Revilla 3747. — B. Schultes 26072.

as Macoubea sprucei (Allorge, 1985; Gentry, 1974),
there are great differences between the two species.
In the treatment of Allorge (1985), plate 76 of M.
sprucei based on the Panamanian specimens rep¬
resents M. mesoamericana. An explanation is in or¬
der.

The fruits of Macoubea mesoamericana are al¬
ways hemisyncarpous, moderately to obscurely bi-
lobed (Fig. 1). In M. sprucei (Fig. 2) the fruits are
reniform and formed by two fused carpels, despite
the fact that the fruits are sometimes reduced to
only one locule by abortion (e.g., the type of M.
witotorum, a synonym of M. sprucei). Of the ap¬
proximately 40 fruiting collections examined from
South America none had fruits like those of the
Mesoamerican collections. In addition, the fruits of
M. mesoamericana are always conspicuously short¬
er than M. sprucei.

Another difference is found in the seed texture

(Fig. 3). In Macoubea mesoamericana the seeds are
minutely and finely rugose, inconspicuously and
minutely pitted, features consistent in all the col¬
lections available. The seeds of M. sprucei and M.
guianensis are coarsely rugose, but with conspicu¬
ous and irregular depressions along the testa. In
addition, Macoubea sprucei and M. guianensis are
restricted to Amazonia, while M. mesoamericana is
known only from Costa Rica and Panama. The three
species have similar flowers and leaves; sometimes
it may be difficult to separate Macoubea sprucei
from M. guianensis. Also the texture and pubes¬
cence of the leaves are highly variable (Monachino,
1945). Many species and intraspecific varieties
were recognized previously by Monachino (1945)
and Allorge (1985). However, the recognition of va¬
rieties based only on these characters seems un¬
warranted.

Paratypes. COSTA RICA. Limon: Talamanca, Bratsi,

88

Novon

i cm

3**

t _ I

1 cm

Figure 3. Seeds of Macoubea. — A. \1. sprucei (Mtiller Argoviensis) Markgraf ( Revilla 3747). — If. M. mesoamericana
J. F. Morales ( Cliuron 185).

Alto Ixiri. Aguilar et al. 938 (INB. MO); Barque Interna-
cional l.a Amistad, Crorina, Chacon 185 ( 1 N li. MO); Brat-
si. Talamanca, Herrera 3343 (COL, CIL F. INB. MO, NV.
USJ ). PANAMA. Panama: along El Llano— Cartf road.
Correa el al. 1847 (DUKE, MO, PMA), tie Severs el al.
4845 (MO), de Severs et al. 6219 (MO), Gentry el al.
14198 (MO). Kennedy 241 1 (MO), Kennedy 2518 (MO),
Liesner 1145a (MO), McPhi ‘rson et al. 8201 (MO), Mc¬
Pherson el al. 8216 (MO). Mori el al. 5106 (MO). San
Bias: El Llano— Cartf road. Herrera el al. 1046 (MO), de
Severs et al. 4347 (MO), de Severs el al. 5005 (MO), de
Severs el al. 6193 (MO).

Kky to thk Species ok Macih kka

la. Corolla lobes 1.1— 1.6 cm long; fruits 1-carpellate
by abortion, globose, subglobose, or subspherical

. M. guianensis Aublet

lb. Corolla lobes 0.7-1 cm long; fruits hemisyncar-
pous, composed by two locules fuseil basally,
locules somewhat rounded to reniform, some¬
times reduced to only one locule by abortion but
never subspherical or globose.

2a. Fruits 9—13 X 6-7.5 cm; seeds coarsely ru¬

gose. with conspicuous and irregular de¬
pressions along the testa; Amazonia . . .

.... M. sprucei (Muller Argoviensis) Markgraf
2b. Fruits 4.5— 6(— 7) X 2.8— 3.6 cm; seeds mi¬
nutely and finely rugose, inconspicuously
and minutely pitted; Costa Rica and Panama
. M. mesoamericana J. F. Morales

Acknowledgments. 1 thank the curators and di¬
rectors of COL, CR. F, GH, MO, NY, US. and USJ
lor the loan of material. I also thank Barry Hammel
(MO) for comments on the manuscript.

Literature Cited

Allorge, L. 1985. Monographie des \ pocynarees Taber-
naemontanoidees Americaines. Mem. Mus. Natl. Hist.
Nat., Ser. B. Bot. 30: 170-176.

Gentry, A. II. 1974. Notes on Panamanian Apocynaceae.

Ann. Missouri Bot. Card. 61: 89I-<XH).

Monachino, J. 1945. A revision of Macoubea and the
American species of Landolphia (Apocynaceae). Lloy-
dia 8: 291-317.

Rhodocalyx (Apocynaceae), a New Synonym of Prestonia

J. Francisco Morales

Institute) Nacional de Biodiversidad (INBio), Apto. 22-3100, Santo Domingo de Heredia,

Costa R iea

ABSTRACT. The genus Rhodocalyx is reduced to
synonymy under Prestonia. A new combination, P.
erecta, is proposed here.

Prestonia R. Brown (Apocynaceae, Apocyno-
ideae) is a neotropical genus of about 55 species.
These lianas (rarely suffrutescent herbs) are char¬
acterized by having the corolla tube usually with a
corona within (sometimes the free corona lobes are
replaced by callus ridges), with a conspicuous an¬
nular corona (entire to five-lobed), clear latex, se¬
pals with a single episepalous colleter (sometimes
deeply lacerated), and truncate seeds (Woodson,
1936a; Morales, 1997a, 1997b). The genus was di¬
vided into four sections by Woodson (1936a), ac¬
cording to the presence or absence of the free co¬
rona lobes in the corolla tube, shape and length of
the sepals, and the nature of the corolla indument.
As explained by Woodson (1936a) and Morales
(1997a), section Tomentosae, characterized by pu¬
bescent corollas, is taxonomically the most difficult
section in the genus.

During studies toward a clarification of some no-
menclatural problems in Prestonia (Morales,
1997a), it became evident that a re-evaluation of
the characters traditionally used to delimit the ge¬
nus, as well as an analysis ot its relationships with
some closely related genera, was needed in order
to make the generic delimitation of the genus more
natural and consistent. Recently, Morales (1997b)
evaluated Prestonia sect. Coal it ae and transferred
three species to Echites, because they lacked both
an annular corona and corona lobes within the co¬
rolla, the principal characters of Prestonia , thus
maintaining both genera as natural groups.

According to the concepts of Woodson (1933,
1936a) and Pichon (1950), Prestonia is easily dis¬
tinguished by having the corolla tube with a con¬
spicuous annular corona, usually with five free ep-
istaminal corona lobes within (but these sometimes
absent or reduced to callus ridges), and a solitary
episepalous colleter.

The two genera most closely related to Prestonia
are Laubertia and Rhodocalyx, both of which also
have a thickened annular corona in the mouth.
Laubertia A. DC. is easily distinguished by the

eglandular sepals lacking colleters within; careful
study of this genus shows that Laubertia could sim¬
ply be a Prestonia that lost its calycine colleters,
but here it is considered to stand as a good genus.
This character is very important to separate some
genera of Apocynaceae (e.g., Rhabdadenia, Elytro-
pus, Cycladenia ) and therefore, despite the pres¬
ence of a conspicuous faucal annulus, the eglan¬
dular sepals easily separate Laubertia.

Rhodocalyx Muller Argoviensis (1860) has tra¬
ditionally been separated from Prestonia by its
erect habit, terminal inflorescences, and corolla
tube with a conspicuous annular corona in the
mouth but lacking free corona lobes (Woodson,
1933; Pichon, 1950; Ezcurra, 1992). The need to
evaluate these characters when considering the re¬
duction of Rhodocalyx to synonymy of Prestonia
was previously mentioned by Ezcurra (1992), who
revealed the close affinity of these genera and the
similarity of R. rotundifolius Muller Argoviensis to
P. riedelii (Muller Argoviensis) Markgraf.

Recently, Stranghetti and Sumiko (1996) re-eval¬
uated Rhodocalyx in a comparative morphological
study with two species of Mandevilla and two spe¬
cies of Prestonia that are vegetatively and superfi¬
cially (fide Stranghetti & Sumiko, 1996) somewhat
similar to Rhodocalyx, which reflects basically the
same work as Stranghetti (1992). Stranghetti and
Sumiko concluded that Rhodocalyx must be con¬
served as a monotypic genus. However, in the pro¬
cess of preparing a synopsis of Prestonia and Man¬
devilla, and alter the examination and evaluation of
the critical morphological characters (habit, indu¬
ment, leaves, inflorescences, sepals, corollas, and
gynoecia) in most of the species of both genera, 1
have found important evidence which indicates that
Rhodocalyx should be included in the synonymy of
Prestonia. Explanation is in order.

I disagree with the relationships of Rhodocalyx
with respect to Mandevilla suggested by Stranghetti
and Sumiko (1996). As was shown previously, the
most important characters separating Prestonia and
the closely related Laubertia from the rest of the
neotropical genera of Apocynaceae are the shape
of the style head, attachment of the anthers to the

Novqn 9: 89-91. 1999.

90

Novon

style head, and the annular corona in the orifice of
the corolla tube. Mandevilla never has an annular
corona; instead, one of the most important morpho¬
logical features shared by Mandevilla and the
closely related Allomarkgrafia, Forsteronia (subg.
Forsteronia only), Macrosiphonia, Mesechites, Quio-
tania, and Tintinnabularia is the presence of col-
leters on the leaf midvein adaxially (Hansen, 1985;
Morales, 1996, 1997c, 1997d; Zarucchi, 1991): the
leaves of Rhodocalyx lack colleters adaxially on the
leaf blade.

Rhodocalyx is restricted to the “campos” and sa¬
vannas of central and southern Brazil and the Si¬
erra de Amambay in northern Paraguay. Commonly,
species of Mandevilla from this phytogeographic
area have terminal to subterminal inflorescences
and an erect habit, in contrast to the typically ax¬
illary inflorescences and scandent habit of the ge¬
nus. Despite the anomalous morphologies of these
species, all the rest of the characters fall well with¬
in the range of Mandevilla. This suggests that those
character states (erect habit and terminal inflores¬
cence) traditionally used to separate Rhodocalyx
from Prestonia , and emphasized by Stranghetti and
Sumiko (1996), are insignificant.

Terminal inflorescences, a character supposedly
restricted to Rhodocalyx , are found in some species
of Prestonia. At the time of Woodson’s (1936a)
monograph, all species of Prestonia then known
were characterized by axillary inflorescences. Since
then, additional species in section Tomentosae have
been found with both axillary or terminal to sub¬
terminal inflorescences. For example, Steyermark et
al. 111114 (F, VEN, Z), Prestonia parvifolia Ben-
tham from Venezuela, lias terminal or subterminal
inflorescences. In addition, some recently de¬
scribed species (e.g., P. tysonii A. H. Gentry) clear¬
ly have terminal inflorescences. Therefore, the dis¬
tinction of the two genera based on this character
seems unwarranted.

In addition to the annular corona of the corolla
tube, another important feature on which Rhodo¬
calyx was segregated from Prestonia is the corolla
tube without free corona lobes, a character suppos¬
edly restricted to Prestonia. Although the presence
of free corona lobes is characteristic of Prestonia
and Cycladenia in the New World, in some species
of Prestonia the corolla tube lacks free corona lobes
behind the anthers or only has inconspicuous callus
ridges. This is obvious in P. dusenii (Malme) Wood-
son, P. riverae J. F. Morales, P mexicana A. DC.,
P. hammelii J. F. Morales, P. speciosa Donnell
Smith, P. clandestina J. F. Morales, and P. riedelii
(Muller Argoviensis) Markgraf. This is the situation

in Rhodocalyx, which has conspicuous callus ridges
within the corolla tube.

As to other characters used by Stranghetti and
Sumiko (1996), leaf pubescence is highly variable
among all the species of Prestonia sect. Tomentosae
and Rhodocalyx. Therefore, this feature has little
taxonomic value to separate Prestonia from Rho¬
docalyx. The differences in the corolla color are not
always constant in Prestonia: in P. mollis Kunth
corolla color varies from cream to cream-purple to
greenish purple.

This inconsistency and the independent variation
of the characters used to separate these genera sug¬
gest that it is necessary to relegate Rhodocalyx to
the synonymy of Prestonia. This brings a measure
of consistency to the generic classification, without
which other well-defined genera, such as Mande¬
villa, could be subdivided based on the same kind
of weak characters. Rhodocalyx rotundifolius M till¬
er Argoviensis is thus transferred below to Presto¬
nia. A new combination is proposed here.

Prestonia R. Brown, Mem. Wern. Soc. 1: 69. 1811.
TYPE: Prestonia tomentosa R. Brown.

Haemadictyon Lindley, Trans. Hort. Soc. London 6: 70.
1826. TYPE: Haemadictyon venosum Lindley | =
Prestonia quinquangularis (Jacquin) Sprengel],
Rhodocalyx Miiller Argoviensis, in Mart. FI. liras. 6 (1):
172. I860. Syn. nov. TYPE: Rhodocalyx rotundifo¬
lius Muller Argoviensis.

Belandra S. E. Blake, Contr. Gray Herb. 52: 78. 1917.
TYPE: Belandra coneolor S. F. Blake [= Prestonia
longifolia (Sesse & Vloyino) J. F. Morales].

Prestonia erecta (Malme) J. F. Morales, comb,
nov. Basionym: Rhodocalyx erectus Malme,
Ark. Bot. 21 A (6): 15. 1927, nom. nov. for
Echites erecta A. DC. Echites erecta A. DC.,
Prodr. 8: 469. 1844, non Thunberg, Gen. Echi-
tis Observ.: 7. 1819; non Vellozo, FI. Flumin.
113. 1829. Rhodocalyx rotundifolius Muller
Argoviensis, in Mart. FI. Bras. 6 (1): 173, pi.
51. 1860, nom. nov. for Echites erecta A. DC.,
non Prestonia rotundifolia K. Schumman ex
Woodson, Ann. Missouri Bot. Gard. 23: 318.
1936. TYPE: Brazil: data lacking, Claussen
343 (holotype, G-DC not seen; isotype, G; pho¬
to of holotype at INB).

Prestonia erecta flowers from late August to Jan¬
uary, and fruits are borne from January to April.
This species is commonly found in savannas and
open areas. For excluded species, see Woodson
(1936b).

Prestonia erecta J. F. Morales is closely related
to P. riedelii (Muller Argoviensis) Markgraf but can

Volume 9, Number 1
1999

Morales

Rhodocalyx

91

be distinguished by the erect habit, short-petiolate
leaves, and terminal inflorescences of the former
species.

Acknowledgements. I thank the curators and di¬
rectors of F, G, G-DC, VEN, and Z for the loan of
material. I thank Michael Grayum for critical com¬
ments on the manuscript.

Literature Cited

Ezcurra, C. E. 1992. Rhodocalyx. In: Spichiger & Ramella
(editors). Flora of Paraguay, Vol. 17: 103—105. Missouri
llotanieal Garden, St. Louis, and Conservatoire et Jar-
din botaniques de la Ville de Geneve, Geneve, Swit¬
zerland.

Hansen, B. E. 1985. A Monographic Revision of Forstero-
nia (Apocynaceae). Ph.D. Dissertation, University of
South Florida, Tampa. [University Microfilms Interna¬
tional, Ann Arbor. |

Morales, J. F. 1996. Una nueva especie de Tintinnabularia
(Apocynaceae). Novon 6: 392-394.

- . 1997a. A synopsis of the genus Prestonia sect.

Tomentosae in Mesoamerica. Novon 7: 59—63.

- . 1997b. A re-evaluation of the genera Echites and

Prestonia sect. Coalitae (Apocynaceae). Brittonia 49:
328-336.

- . 1997c. A synopsis of the genus Allomarkgrafia

(Apocynaceae). Brittonia 49: 337-345.

- . 1997d. A synopsis of the genus Mandevilla (Apo¬
cynaceae) in Mexico and Central America. Brittonia 50:
214-232.

Muller Argoviensis, J. 1860. Apocynaceae. Rhodocalyx.

hi: C. F. P. Martins, Flora Brasiliensis 6(1): 172—173.
Pichon, M. 1950. Classification des Apocynaeees XXV.
Echitoidces. Mem. Mus. Natl. Hist. Nat. Ser. B. Bot. 1:
1-144.

Stranghetti, V. 1992. Reavaliayao do genera monotfpico
Rhodocalyx Miill. Arg. (Apocynaceae). Tese do mestra-
do, Universidadc Estadual de Campinas, Campinas, Sao
Paulo, Brazil.

- & L. Sumiko. 1996. Reavaliayao do genera mon¬
otfpico Rhodocalyx Miill, Arg. (Apocynaceae). Revista
Bras. Bot. 19: 133-144.

Woodson, R. E. 1933. Studies in the Apocynaceae IV. The
American genera of Echitoideae. Ann. Missouri Bot.
Card. 20: 606-790.

- . 1936a. Studies in the Apocynaceae IV. The

American genera of Echitoideae XXVI. Prestonia. Ann.
Missouri Bot. Card. 23: 276—283.

- . 1936b. Studies in the Apocynaceae IV. The

American genera of Echitoideae XXVII. Rhodocalyx.
Ann. Missouri Bot. Card. 23: 367—370.

Zarucchi, J. I,. 1991. Quiotania: A new genus of Apocy¬
naceae— Apocynoideae from northern Colombia. Novon
1: 33-36.

A New Species of Lilaeopsis (Apiaceae) from Mauritius

Gitte Petersen

Botanical Institute, University of Copenhagen, Gothersgade 140, DK-1123 Copenhagen K,

Denmark

James Aff alter

The State Botanical Garden of Georgia, The University of Georgia,

2450 South Milledge Avenue, Athens, Georgia 30605, U.S.A.

ABSTRAC T. A new species of Lilaeopsis, L. mauri-
tiana , is described and illustrated. The species is
endemic to Mauritius (southwest Indian Ocean),
where it has been found at a single locality in Le
Val Nature Park. The predominantly temperate and
alpine genus Lilaeopsis has not previously been rec¬
ognized from Mauritius.

Lilaeopsis E. L. Greene (Apiaceae) consists of
small, perennial, rhizomatous herbs occupying
moist or truly aquatic habitats. Five species occur
in North America, six in South America (including
two North American disjuncts), two in Australia,
and two in New Zealand. An unidentified species
has been reported from Kerguelen Island in the
South Indian Ocean, but a possible occurrence of
Lilaeopsis on Madagascar (Raynal, 1977) is consid¬
ered dubious (Affolter, 1985: 34). In the treatment
of the Apiaceae lor the Flore des Mascareignes
(Scott, 1990), Lilaeopsis is not included.

The species ol Lilaeopsis are morphologically dif¬
ficult to distinguish due to their much reduced veg¬
etative habit and relatively simple, uniform inflo¬
rescences. The leaves are linear to spathulate,
hollow, and transversely septate. They are assumed
to consist of only the axis of the compound leaves
common to most Apiaceae species. Mainly morpho¬
logical and anatomical characters of the fruit have
been applied to distinguish species (Affolter, 1985).

Lilaeopsis inaiiritiana G. Petersen & Affolter, sp.
nov. TYPE: Gult. from Mauritius, Le Val Na¬
ture Park near Le Val, 20°21'S, 57°37'E, 5
Mar. 1992, II. Windelftv s.n. (holotype, G; iso¬
types, GA, UC). Figures 1, 2.

Note. All types are cultivated specimens de¬
rived through vegetative propagation of the original
collection; 5 Mar. 1992 is the collection date.

Differt a aliis speciebus Lilaeopsidis fructibus eostis
conspieuis, tenuibus, obtusis. sed sine cellulis spongiosis
et sine cellulis pusillis. incrassatis, lignosis, ornatis.

Glabrous, perennial herbs with creeping rhi¬
zomes. Rhizomes 0.5-1. 2 mm diain. Leaves arise
individually or in 2-3(^l)-leaved clusters directly
from horizontal rhizomes, hollow and terete to el¬
liptical in cross section proximally, becoming flat¬
tened and solid in cross section distally. linear to
subulate, 2.8-13.0 cm long, 0.4— 1.3 mm broad, 4—
9 septate, attenuate at apex, expanded at base into
a scarious sheath, 0.1— 0.5 cm long. Peduncles 6—
26 mm long, borne directly at nodes on the hori¬
zontal rhizome, subtended by a short bract 0.9— 1.2
mm long, sometimes opposite a leaf. Involueral
bracts 0.4— 2.0 mm long. Umbels 3-6-fIowered;
pedicels 0.9— 6.0 mm long. Petals greenish white.
Fruits globose to ellipsoid or obovoid, 1.9-2. 5 mm
long, 1. 7-2.0 mm broad, spongy cells absent from
all ribs; dorsal and intermediate ribs prominent and
narrowly rounded, lateral ribs broadly rounded; vit-
tae 6—8, 2-3 on the commissure. Chromosome
number 2 n — 22. Figures 1, 2.

The phylogeny of Lilaeopsis, which included 13
species in the monograph by Affolter (1985), is
completely unknown. Hence, it is difficult to de¬
termine whether the closest relatives to L. mauri-
tiana should be found among the New World or
Australasian species.

Two important characters for separating many
species of Lilaeopsis are the abundance and distri¬
bution of “spongy cells" in the fruit. These are stor¬
age tracheids, nearly isodiametric in shape, that
lack a living protoplast at maturity. In cross sec¬
tions of fresh, green fruits they are visible to the
naked eye as bright white patches; under micro¬
scopic examination the spongy cells display char¬
acteristic secondary wall thickenings. They appear
to enhance dispersal in Lilaeopsis by increasing the
buoyancy of the fruits.

Spongy cells are completely absent from the
fruits of L. mauritiana (Fig. 2). This character has
been observed in only three other species of the

Nov on 9: 92-94. 1999.

Volume 9, Number 1
1999

Petersen & Affolter
Lilaeopsis mauritiana (Apiaceae)

93

Figure 1. Lilaeopsis mauritiana G. Petersen & Affolter. — A. Habit. — B. Fruit. Scale bars: A = I cm. B = I mm.
(Based on H. Win del 0r s.n.)

genus. Fruits of the widespread and highly variable
L. macloviana (Gandoger) A. W. Hill from South
America sometimes lack spongy cells, hut unlike
L. mauritiana the lateral ribs are then low and ob¬
scure rather than broadly rounded. Another vari¬
able species, L. polyantha (Gandoger) H. J. Eichler
from Australia, also occasionally produces fruits
without spongy cells, hut in that case the fruits pos¬
sess distinctive thick-walled lignified cells that are
not present in the Iruits of L. mauritiana. Finally,
fruits of L. ruthiana Affolter, from New Zealand,
often lack spongy cells, hut all the fruit ribs are
low and obscure — much less prominent than those
of L. mauritiana.

Further studies are clearly needed to produce a
phylogenetic hypothesis for Lilaeopsis. As morpho¬

logical characters are scarce and variable, molec¬
ular data may prove most helpful.

Distribution and habitat. Lilaeopsis mauritiana
has ordy been collected at a single locality in Le
Val Nature Park near Le Val in the southeast part
of the island of Mauritius. The habitat, located at
an altitude of approximately 300—500 in, is a mod¬
erately flowing, clear-watered stream, along which
the plants were found growing both fully submerged
and on the drier part of the bank.

The flora of Mauritius is characterized both by a
large number of introduced species and by a large
number of endemics (Baker. 1877; Vaughan, 1937;
Vidal, 1988). The locality occupied by L. mauriti¬
ana is a popular recreation area, and in the stream
it was found growing together with Aponogeton

94

Novon

Figure 2. Lilaeopsis mauritiana (H. Windelpv s.n.).
Transverse section of mature fruit embedded in GMA ami
stained in PAS-ABB. Scale bar = 0.5 mm.

madagascariensis H. Bruggen. Aponogeton mada-
gascariensis is usually recognized as a Madagascan
endemic introduced to Le Val, Mauritius, in the mid
1850s (H. W. E. van Bruggen, pers. comm.;
Vaughan, 1937, under the synonym A. fenestralis J.
D. Hooker). The possibility exists that Lilaeopsis
mauritiana has been introduced to Mauritius, but
as the species is clearly distinct from all other de¬
scribed species it must be considered endemic to
the island until proven otherwise. With the discov¬
ery of Lilaeopsis mauritiana on Mauritius, the pos¬

tulated occurrence of a species of Lilaeopsis on
Madagascar (Raynal, 1977) merits further attention.

Acknowledgments. We thank Kate Jensen, Lis-
betli Knudsen, Jimmy Olsen, and Flemming Sarup
for skillful technical assistance. Jan Petersen made
the line drawings, and Peter Wagner kindly pre¬
pared the Latin diagnosis. We owe special thanks
to Claus Christensen for placing the new species at
our disposal.

Literature Cited

Affolter, J. M. 1985. A monograph of the genus Lilaeopsis
(Umbelliferae). Syst. Bot. Monogr. (r. 1—140.

Baker, J. G. 1877. Flora of Mauritius and the Seychelles.
L. Beeve, London.

Baynal, .1. 1977. Le genre Lilaeopsis (( )mhel liferes) h Mad¬
agascar. Adansonia 17: 151—154.

Scott, A. J. 1990. Ombelliferes. In: .1. Bosser, T. Cadet, J.
Cueho, 11. B. Julien & W. Marais (editors), F lore des
Mascareignes: La Reunion, Maurice, Rodrigues. 90.
Bhizophorac^es it 106. Araliac6es. Mauritius Sugar In¬
dustry Besearch Institute, OBSTOM, and Royal Botanic
Gardens, Kew. Port Louis, Mauritius.

Vaughan, R. E. 1937. Contributions to the flora of Maur¬
itius. An account of the naturalized flowering plants re¬
corded from Mauritius since the publication of Baker's
“Flora of Mauritius and the Seychelles” (1877). J. Linn.
Soc., Bot. 51: 285—308.

Vidal, J. E. 1988. Quelques aspects de la flore et de la
vegetation des lies de 1’Ocean Indien (Seychelles, Re¬
union, Maurice). Bull. Soc. Bot. France 135, litres
Bot. 1988: 361-368.

Maytenus rupestris (Celastraceae), a New Species from Minas
Gerais, Southeastern Brazil

Jose Rubens Pirani

Instituto de Biociencias, Universidade de Sao Paulo, Caixa Postal 11461, 05422-970, Sao

Paulo, SP, Brazil

Rita Maria de Carvalho-Okano

Departamento de Biologia Vegetal, Universidade Federal de Vigosa, 36570-000, Vigosa, MG,

Brazil

ABSTRACT. Maytenus rupestris Pirani & Carvalho-
Okano, a now species of Celastraceae endemic to
the Serra do Cipo, Minas Gerais, Brazil, is de¬
scribed and illustrated. Included in section May¬
tenus, it is distinct from its congeners (especially
from M. boaria Molina) by its ascending, very nar¬
row leaves with revolute margins that are entire or
sparsely denticulate on the distal half, and by its
few-flowered cymes.

Maytenus Molina emend. Molina is one of the
largest genera in the Celastraceae, with around 200
pantropical species, mostly South American (Born-
stein, 1989). According to Carvalho-Okano (1992),
77 species occur in Brazil, from Amazonia to the
south of the country. In the state of Minas Gerais,
14 species occur, 10 belonging to section Maytenus
(M. evonymoidis Reissek, M. floribunda Reissek, M.
glazioviana Loesener, M. gonoclada Martius, M.
imbricata Reissek. M. ligustrina Reissek. M. radl-
koferiana Loesener, M. robusta Reissek, M. salici-
jolia Reissek, and M. urbaniana Loesener), and 4
to section Oxyphylla Loesener (M. acanthophylla
Reissek, M. aquifolia Martius, M. comocladiaefor-
mis Reissek, and M. rigida Martius).

While performing floristic surveys along the
mountains of the Espinhago Range, in Minas Ger¬
ais and Bahia, a new species was found in river
margins in the Serra do Cipo and is presented here.
(Note: the acronym CFSC — Colegao Flora da Serra
do Cipo.)

Maytenus rupestris Pirani & Carvalho-Okano, sp.
nov. TYPE: Brazil. Minas Gerais: Municfpio de
Santana do Riacho, Serra do Cipo, Ribeirao
Indequice, Cachoeira do Cornelio, perto da
Estrada da Usina, 5 Oct. 1981 (fl), A. Furlan,
/. Cordeiro, M. L. Kawasaki & J. R. Pirani
CFSC 7506 (holotype, SPF; isotypes, K, NY.
RB, SP. VIC). Figure 1.

Ab omnibus congeneribus foliis anguste oblongo-ellip-
ticis margine revoluto integerrimo vel ultra medium re¬
mote denticulato, cymis brevibus paueifloribus differt.

Shrub or small tree, 1-3 m tall, glabrous. Twigs
grayish, 4-angled. Leaves alternate, ascending, nar¬
rowly oblong-elliptic, apex obtuse or sometimes
acute or truncate, often mucronulate, base attenu¬
ate and decurrent to the short petiole, margin entire
or rarely with 2 very small teeth on the distal part,
revolute, in vivo dark green and lustrous on the
adaxial face, light green and dull on the abaxial
face, in sicco grayish to glaucous on both faces or
slightly ochraceous on the abaxial face, 30—80 mm
long, 4—8 mm wide; primary vein prominent on
both faces at least toward base, venation brochi-
dodromous, secondary veins straight to slightly
curved, evident on the abaxial face, obscure on the
adaxial face. Stipules 7-10 mm long, triangular,
blackish. Inflorescence a reduced cyme with 2— 3(— 4)
flowers, axillary, pedunculate, ramified, 6-14 mm
long; bracts triangular, 5-7 mm long, apex acute to
acuminate, margin irregularly serrulate. Flowers 5-
merous, short-pedicellate, cream to greenish col¬
ored; sepals ovate, ca. 1 mm long; petals free, ob¬
long, apex rounded, ca. 3 mm long, margin thin and
slightly undulate, reflexed at anthesis; stamens 5,
alternate with the petals, ca. 2 mm long, filaments
flattened, attenuate to the apex, broadened toward
base, inserted on the base of the intrastaminal disc,
reflexed at anthesis, anthers ovoid, yellow; intras¬
taminal disc fleshy, 5-lobed, ca. 2 mm diam., mar¬
gin undulate; gynoecium of 2 fused carpels, the
base of the ovary slightly fused to the disc, stigma
2-lobed on a short style. Fruit a loculicidal capsule,
ellipsoid to obovoid, apiculate, ca. 10 mm long, ca.
7 mm diam., yellow to orange when ripe, opening
in 2 reflexed valves consisting of 2 halves of ad¬
joining carpels; seed 1, erect, completely surround¬
ed by a white, soft aril.

Novon 9; 95-97. 1999.

96

Novon

Figure 1. Maytenus rupestris Pirani & Carvalho-Okano. — A. Flowering shool; note 4-angled twigs. — B. Cyme with
4 flowers. — C. Flower at anthesis, before dehiscence of the anthers. — I). Flower with the stamens reflexed and anthers
opened. — E. Longitudinal section of a flower. — F. Petal, adaxial view. — G. Cross section of an ovary. — H. Fruit
before dehiscence. — I. Capsule opened showing seed surrounded by an aril. — J. Capsule after dropping of the seed,
with reflexed valves. A, C— G. drawn from Furlan el al. CFSC 7506: B. drawn from Joly CFSC 3691 ; 11— J. drawn from
Pirani CFSC 1 1449.

Volume 9, Number 1
1999

Pirani & Carvalho-Okano
Maytenus rupestris from Brazil

97

Maytenus rupestris is distinct from the other spe¬
cies of the genus by its ascending, very narrow
leaves, with revolute margins, commonly entire but
sometimes with small teeth on the distal part, and
also by its few-flowered (2—3, seldom 4), reduced
cymes. Besides this combination of characters, M.
rupestris is easily recognized in the field by the col¬
or of its leaves, which are dark green and shiny
above and light green and dull below, turning gray¬
ish to glaucous on both faces when dried, probably
because of the presence of an epicuticular wax lay¬
er.

Maytenus rupestris is included in section May¬
tenus, along with the bulk of the taxa in the genus,
because of its entire (vs. spiny) leaf margins. May¬
tenus boaria Molina, from southern Brazil. Argen¬
tina, Chile, and Peru, also has very narrow leaves,
but it seems not to be closely related to the new
species, since its leaves have glandular-serrate, not
revolute margins, its stipules are linear, fimbriate,
and its inflorescences are multi-flowered fascicles.

Maytenus rupestris is known so far only from the
Serra do Cipo, in the southern part of the Espin-
ha<jo mountain range, in the state of Minas Gerais.
It is a shrub or much-branched tree up to 3 m that
grows as a heliophyte in open vegetation among
sandstone rocks (“campo rupestre”), by the river
and sides of gullies of that region. It occurs some¬
times close to riparian forests, but not in the shad¬
ow of the tree canopies. Although it is probably
endemic to the Serra do Cipo, since it has not been
found elsewhere so far, locally it is relatively com¬

mon. The flora of the “serras” of the Espinha§o
Range shows a high number of endemic species,
and even endemic genera, in several families, but
mostly referring to taxa from open formations (Giu-
lietti et al., 1997). However, in many instances,
even essentially forest genera (like the case of May¬
tenus) happen to show one or a couple of endemic
“campo rupestre” members in the Espinha^o.

Paratypes. BRAZIL. Minas Gerais: Santana do Ria-
cho, Serra do Cipo. Rodovia Belo Horizonte-Concei^So do
Mato Dentro, km 128. 4 Nov. 1972 (11), A. B. Joly CFSC
3691 (SP. SPF, VIC); km 129, 20 Oct. 1973 (fl). A. B. Joly
et al. CFSC 4607 (FCAB. SP); km 1 17, 4 Sep. 1973 (fl).
./. Semir et al. CFSC 4381 (SP); km 117, 26 July 1973
(fr), /V. L. Menezes CFSC 4293 (SP); Estrada da Usina Dr.
Pacffico Mascarenhas, 14 Nov. 1983 (fl. fr). J. R. Pirani
& /.. Rossi CFSC 9215 (SPF, VIC); 2 km acima da entrada
para a Estrada da Usina. 6 Sep. 1987 (fl). R. Simao CFSC
10427 (SPF); km 111, Vale do Corrego Vitalino, 20 May
1989 (fr), J. R. Pirani CFSC 11449 (MBM, SP, SPF, VIC).

Literature Cited

Bornstein, A. J. 1989. Celastraceae. Flora of the Lesser
Antilles Leeward and Windward Islands, 5: I 13-125.
Carvalho-Okano, R. M. de. 1992. Estudos taxonomicos do
genero Maytenus Mol. emend. Mol. (Celastraceae) do
Brasil Extra-Amazonico. Unpublished Doctoral Thesis,
instituto de Biologia, Universidade de Campinas, Cam¬
pinas.

Giulietti, A. M„ J. R. Pirani & R. M. Harley. 1997. Es-
pinhago Range Region, eastern Brazil. Pp. 397-404 in
S. I). Davis, V. H. Heywood, (). Herrera-MacBryde. J.
Villa-Lobos & A. C. Hamilton (editors). Centres of Plant
Diversity. A Guide and Strategy for Their Conservation,
Vol. 3. The Americas. IUCN Publications Unity, Cam¬
bridge.

A New Species of Miconia (Melastomataceae) from Serra da
Canastra National Park, Minas Gerais, Brazil

Rosana Romero arid Renato Goldenberg

Departamento <le Botanica, Uni vers idade Estadual de Campinas, Caixa Postal 6109, CEP

13083-970 — Campinas, SP, Brazil

ABSTRACT. A new species of the genus Miconia
sect. Cremanium from highland “campo rupestre”
vegetation of the Serra da Canastra National Park,
Sao Roque de Minas, is described and illustrated.
It is distinguished from other species in this section
by the 4-celled ovary with glabrous apex and leaves
lanate below.

This new species of Miconia was discovered dur¬
ing a floristic survey carried out by the Herbarium
Uberlandense (HUFU) from the Universidade Fed¬
eral of Uberlandia in the Serra da Canastra Nation¬
al Park, southwestern Minas Gerais state.

Miconia angelana R. Romero & R. Goldenberg,
sp. nov. TYPE: Brazil. Minas Gerais: Sao
Roque de Minas, Parque Nacional da Serra da
Canastra, vale da nascente do rio Sao Fran¬
cisco, lado das matas, beira do cdrrego, 1100
m, 46°15'— 47°00'W, 20°00'-20°30'S, 20 Nov.
1996, R. Romero <£- J. N. Nahajima 3773 (ho-
lotype, HUFU; isotypes, k, MO, RB, UEC,
US). Figures 1—5.

Arbor pan a ea. 2 m. Lamina oblongo-lanceolata, apice
acuto vel acuminato, basi rotundata, margine remote un-
dulato-denticulato, subtus dense lanato. Panicula elonga-
ta: tfores 5-meri, sessiles, ramis inflorescentiae in glome-
mlo denso verticillato-congestis. Antherae oblongae vel
cuneatae, ad apicem truncatae, 2-porosae. Ovarium 4-lo-
eulare, apice glabro.

Small tree, ca. 2 m tall, the stem, branches, pet¬
ioles, lower leaf surface, inflorescence, and hypan-
thium densely covered with dendritic hairs with
long and slender arms, canescent to yellowish
brown; young branches slightly flattened, older
branches terete and striate. Petioles 1.2— 2.2 cm
long, striate; leaf blade 5.5—14.0 X 1.5— 4.0 cm,
oblong-lanceolate, apex acute to shortly acuminate,
base narrowly rounded, margin slightly and re¬
motely denticulate, hyaline, above covered with
dendritic hairs, soon becoming glabrous, 3-nerved,
with an additional tenuous marginal pair, below
with the primaries, secondaries, and transverse
veins prominent, impressed above. Panicle 7.5—

13.5 cm long, the sessile, 5-merous flowers capi¬
tate-congested on very short lateral branchlets;
bracts 0.6— 1.0 cm long, linear, bracteoles 3.0-3. 5
mm long, linear, persistent. Hypanthium 2.3— 3.0 X
2. 5-3. 3 mm, campanulate. Calyx lobes 0.8— 1.4 mm
long, triangular, apex acute to rounded, external
teeth inconspicuous and hidden by the indument.
Petals 2.0— 2.5 X 1.8— 2.0 mm, obovate, apex retuse
and asymmetrical, margins and abaxial side papil¬
lose. Stamens 10, subisomorphic; filaments 2. 0-2. 3
mm long, geniculate, glabrous; connective dorsally
thickened and prolonged 0.3-0.6 mm below the
thecae, with two ventral short lobes and one dorsal
short spur; anthers 1.3-1. 6 mm long, oblong to cu-
neate with truncate apex, biporose. Ovary 1.2 mm
long, lower half adherent to the hypanthium, 4-loc-
ular, apex elongated, papillose, glabrous, 2-3
ovules per locule; style 3.5— 4.0 mm long, filiform,
glabrous, stigma truncate. Berry 2.5— 3.0 X 3.0— 3.5
mm, globose, blackish, surface smooth and gla-
brescent. Seeds 1.2— 1.7 mm long, 10-15 per fruit,
narrowly to broadly ovoid, surface smooth.

This species grows on the margins of small rivers
in rocky soil of the highlands of the Rio Sao Fran¬
cisco headwaters; it was collected with flowers in
November, with fruits in December.

The section Cremanium Bentham & Hooker,
which has eight species in Brazil, occurs mainly in
the southeastern region. The new species is placed
in this section based on the presence of cuneate
and biporose anthers (Cogniaux, 1891). The 4-
celled ovary with glabrous apex, seeds 1-1.7 mm
long, and leaves lanate below distinguish M. an-
gelana from M. hyemalis A. St. Hilaire & Naudin
ex Naudin, to which it is certainly most closely re¬
lated. The latter differs in having a 3-celled ovary
with a stellate puberulent apex, longer seeds (2—
2.6 mm), and leaves stellulate-tomentose below.
Miconia angelana also resembles M. lymanii Wur-
daek, endemic to Santa Catarina (Wurdack, 1962),
which has a 3-celled ovary with puberulous apex
and leaves stellulate-furfuraceous below.

The specific epithet was chosen in honor of An-

Novon 9: 98-100. 1999.

Volume 9, Number 1
1999

Romero & Goldenberg
Miconia angelana from Brazil

99

Figures 1—5. Miconia angelana R. Romero & R. Goldenberg. — 1. Flowering branch. — 2. Petal. — 3. Frontal and
lateral views of stamen. — 4. Rods. — 5. Longitudinal section of flower to show ovary and the insertion of style, petals
removed. Drawn from the holotype.

100

Novon

gela Borges Martins, for her important work on Bra¬
zilian Melastomataceae.

Paratypes. BRAZIL. Minas Gerais: Sao Roque tie
Minas, Barque Nacional da Serra da Canastra, nascente
do Rio Sao Francisco, borda tlo corrego, 19 Nov. 1995, J.
N. Nakajima et al. 1509 (BHCB, HUFU, MO. UEC); nas-
cente do rio Sao Francisco, beira de uma vertente, 20 Nov.
1996, R. Romero & J. A. Nakajima 5705 (HUFU, VIC);
vale do rio Sao Francisco, campo umido dois meses apos
queimada, 7 Dec. 1994, J. N. Nakajima & R. Romero 69 4
(HUFU, SP, US).

Acknowledgments. The first author is grateful to
John j. Wurtlaek for helpful comments about this
species, to the Smithsonian Institution lor a re¬

search grant to study the collections in the United
States National Herbarium, the Universidade Fed¬
eral de Uberlandia for the facilities of the Herbar¬
ium Uberlandense (HUFU), the Instituto Brasileiro
do Meio Ambiente (IBAMA) for the permission to
collect at Serra da Canastra National Park, and
CNPq for a research grant. The authors thank Ed¬
uardo Kickhofel for preparing the illustration.

Literature Cited

Cogniaux, A. 1891. Melastomataceae. Pp. 1—1256 in A.
De Candolle & C. De Candolle, Monographiae phaner-
ogamarum 7. G. Masson, Paris.

Wurdack, J. J. 1962. Melastomataceae of Santa Catarina.
Sellovvia 14: 10*1—217.

Pycnostachys recurvata (Lamiaceae), a New Species from Ethiopia

Olof Ryding

The Botanical Museum, Gothersgade 130, DK-1123 Copenhagen K, Denmark

ABSTRACT. Pycnostachys recurvata, a new species
from Ethiopia, is described and illustrated. Com¬
parisons are made to P. pseudospeciosa and P. che-
valieri, which it resembles by having recurved
leaves, and to P niamniamensis, P. reticulata, and
P. deflexifolia, which it resembles in calyx shape.

Pycnostachys is a distinct genus belonging to
Lamiaceae subfamily Nepetoideae tribe Ocimeae
subtribe Plectranthinae. It is characterized by hav¬
ing dense, head- to spike-like inflorescences and
subequally spiny-lobed calyces with scales between
the lobes. According to Bruce (1939), who revised
the genus, there are 37 species, all in tropical to
southern Africa and on Madagascar. According to
Cufodontis (1963), the east African P. deflexifolia
Baker occurs in Ethiopia. However, the Ethiopian
material distributed under this name and under P.
reticulata (E. Meyer) Bentham has been found to
represent a distinct taxon, which here is described
as a new species. Material from the herbaria C,
ETH. K, EPS, and WAG has been studied.

Pycnostachys recurvata Ryding, sp. nov. TYPE:
Ethiopia. Shewa Region: about 60 km W of
Ambo along “Lekemti” (Nekemte) road, ca.
1800 m, 10 Nov. 1965, W. J. J. (). de Wilde &
B. E. E. de Wilde-Duyjjes 8754 (holotype, C;
isotypes, ETH, K, UPS, WAG). Figure 1A-C.

Differt ah aliis speciebus cognatis Pycnostachyis com¬
bi natione constanti ex herba perenni, foliis sessilibus re-
curvis basi rotundato, tubo calycis fructicantis sursum
curvato, leviter constricto, lobis calycis moderate pubes-
centibus, 3—4.5 mm longis.

Perennial herb, 0.6— 1.6 m tall. Stems hexangular
or quadrangular, ± furrowed, rather densely pu¬
bescent with up to 0.2— 0.5-mm-long recurved to
spreading eglandular hairs and with sessile glands,
in older parts more sparsely pubescent. Leaves ter-
nate and/or sometimes decussate, mostly smaller on
the flowering shoots than on the main stem, sessile,
gradually recurved, lanceolate, ovate or elliptic, up
to 20-60 X 8 — 28 mm, 1 .6 — 5 times as long as
broad, rather densely pubescent with short eglan¬
dular hairs below on the veins, otherwise sparsely
pubescent and ± densely glandular with sessile
glands, apex subacute or acute, base rounded, mar¬

gin serrate. Inflorescence terminal on branches,
dense, initially head-like, becoming elongate and
spike-like in the fruiting stage, sometimes with a
few spaced three-flowered whorls at the base, up to
30—70 mm long; bracts narrowly lanceolate-linear,
up to 5—7 X 0.6— 1.0 mm, recurved. Flowers sessile.
Calyx enlarging up to a length of 6.5—8 mm in the
fruit, with 5 erect triangular-acute scales alternat¬
ing with and forming a whorl near inside the lobes;
tube bent upward in the proximal half, slightly nar¬
rowed along the distal half, mostly slightly recurved
near the mouth, with 10 slightly prominent longi¬
tudinal veins and a thicker vein at the throat, with
sessile glands and hairy at least in the apical part,
moderately hairy with eglandular hairs at the throat;
lobes 5, narrowly linear-lanceolate, with a thick
mid-rib terminating as a subulate apex, ± deflexed,

3— 4.5 mm long in fruiting stage, moderately pu¬
bescent with short hairs. Corolla 2-lipped, 10-17
mm long, light to dark blue; tube narrow and tu¬
bular at the base, strongly bent near the middle,
abruptly widening from the bend up to the throat,

4— 8 mm long, ventral side 1—2 mm longer than
dorsal side; upper lip 4-lobed, side-lobes narrower
than mid-lobes; lower lip much longer than upper
lip, 4—8 mm long, deeply boat-shaped, with the
apex arched and pointing upward. Stamens 4, all
connate along 1.5—2 mm; posterior pair shorter than
the lower lip of the corolla, strongly bent upward;
anterior pair as long as or slightly longer than lower
lip of corolla; anthers rather deeply cup-shaped,
0.6— 0.8 mm long; pollen-mass yellow. Style strongly
upcurved, becoming slightly longer than the lower
lip of the corolla, minutely bifid at apex. Disc with
a ca. 1-mm-long abaxial lobe that is longer than
the ovary and hides the two abaxial ovary lobes.
Nutlets 1.4— 1.7 X 1.0— 1.2 mm, smooth, slightly
glossy, brown, apparently not mucilaginous.

Distribution and habitat. Wet grasslands and
swamps, in areas with an annual rainfall of 800—
1600 mm; altitude 1200—2000 m. Flowering mostly
from October to December (or March). Southwest¬
ern and southern Ethiopia (Fig. ID).

The sessile, rather narrow leaves and the peren¬
nial habit suggest that P. recurvata belongs to the

Novon 9: 101-103. 1999.

102

Novon

Figure 1. A-C. Pycnoslachys recuruata Ryding. — A. Branch with inflorescence. — B. Flower. — C. Fruiting calyx. — D.
Map of hthiopia showing the known distribution ol P. recurvata (dots) and the divergent Flhiopian form of P. niamniamensis
(star). A drawn from Mooney 8469 (K); B drawn from de Wilde A* de Wilde- Duyjjes 8754 (C); C drawn from Mooney 8611
(K).

Volume 9, Number 1
1999

Ryding

Pycnostachys recurvata from Ethiopia

103

first, third, or fourth group in Bruce’s (1939) infra¬
generic division of Pycnostachys.

Pycnostachys recurvata differs from most of the
other species of Pycnostachys hy having recurved
leaves. Pycnostachys pseudospeciosa Buscalioni &
Muschler from Zambia and Zaire usually resembles
P. recurvata by having more or less recurved leaves
(Buscalioni & Muschler, 1913: fig. on p. 259).
However, P. pseudospeciosa differs from P. recurvata
in some other respects, mainly in the shape of the
fruiting calyx. Unlike that of P recurvata , the fruit¬
ing calyx of P pseudospeciosa is gradually and only
slightly upcurved, gradually broadened toward the
throat, very broad at the throat, and strongly hairy
at the base of the lobes. Pycnostachys chevaileri
Briquet from Chad also has recurved leaves, but
this species differs from P. recurvata by having a
cuneate leal-base, and narrower, ca. 7-mm-long ca¬
lyx lobes, and tending to have smaller (ca. 10 mm
long) flowers. Several Pycnostachys species, such as
P. niamniamensis Giirke, P. reticulata (E. Meyer)
Bentham, and P. deflexifolia Baker, resemble P. re¬
curvata in the shape of the fruiting calyx. In her¬
baria, the latter two names have often been used
for material of P. recurvata. Pycnostachys niamnia¬
mensis from Sudan and east Africa, including a
rather divergent form from western Ethiopia (rep¬
resented by the collection Mesfin & Kagnew 2269),
differs from P. recurvata by having erect and
straight leaves, narrower, 5— 8-mm-long calyx lobes,
and more spreading calyx scales, and by tending to
get longer spikes in the fruiting stage. Pycnostachys
reticulata from southern Africa and south tropical
Africa differs from P. recurvata by having straight
and erect leaves that are cuneate rather than round¬

ed at the base. Its calyx lobes tend to be more
subulate than those of P. recurvata. Pycnostachys
deflexifolia from east Africa has straight but de-
flexed leaves that are cuneate at the base. This spe¬
cies also differs from P. recurvata by having mostly
smaller (ca. 8 mm long) flowers and longer (4—10
mm long) calyx lobes.

The two collections from Shewa differ from other
material of P. recurvata by having slightly larger
flowers.

Paratypes. ETHIOPIA. Welega Region: Barri
(9°2'N, 36°10'E), 26 Mar. 1958, Smeds 1275 (K). Shewa
Region: 185 km from Addis Abeba on Lekemti road
(8°58'N, 37°30'E), 3 Nov. 1962, Mooney 100.35 (El’ll).
Ilubabor Region: Buno Bedele Awraja. ca. 6 km E of
Chora town, 21 Nov. 1981, Mesfin T. & Kagnew G. Y. 2444
(ETH, K). Kefa Region: near Sombo (7°33'N, 36°37'E),
13 Nov. 1960, Mooney 8611 (ETH, k); about 10 km SA
of Jimma, 18 Dec. 1965, W. J. de Wilde & de Wilde-Du-
yfles 9270 (C, ETH. k, WAG). Sidamo Region: Zimbaba
"(5°50'N, 39°10'E), 14 Oct. 1962. Mooney 9707 (ETH.
WAG). Bale Region: Kumbi (6°25'N. 39°25'E), 29 Dec.
1959, Mooney 8469 (ETH, k).

Acknowledgments. Material has been obtained
on loan from ETH, K, and UPS, and WAG has pro¬
vided xerox copies of their herbarium specimens. I
thank the directors and curators of these herbaria
and Alan Paton at K for critical comments on my
manuscript.

Literature Cited

Bruce, E. A. 1939. The genus Pycnostachys. Bull. Misc.
Inform. 1939: 563-593.

Buscalioni, L. & R. Muschler. 1913. In M. Piscicelli, Nel-
la Region dei Laghi Equatoriali. Libreria Luigi Pierro.
Napoli.

Cufodontis, G. 1963. Enumeratio Plantarum Aethiopiae
Spermatophyta (sequentia), Labiatae (p.p.). Bull. Jard.
Bot. Etat 33, suppl.: 829-854.

Jacquemontia revoluta (Convolvulaceae), a New Species from

Minas Gerais, Brazil

Rosangela Simao-Bianchini

Instituto de Botanica de Sao Paulo, Caixa Postal 4005, Sao Paulo, SP, Brazil,
CEP 01061-970. RBianchini@smtp-gw.ibot.sp.gov.br

ABSTRACT. A new species of Convolvulaceae,
Jacquemontia revoluta Simao-Bianchini, is de¬
scribed and illustrated. This species is known only
from the rocky slopes of Serra do Cipo (Minas Ger¬
ais, Brazil), where it is associated with ant nests.
The taxon is distinct in the genus by the linear
leaves with revolute margins and 5-stellate (almost
scale-like) trichomes. In addition to the description
and illustration, comments on its probable relation-
shif )s, habitat, and distribution are presented. Jac¬
quemontia revoluta is probably related to J. linoides
(Choisy) Meissner; both species are rare.

Jacquemontia Choisy is a fairly large genus with
174 taxa described and about 126 species accepted
at present. The remaining 48 names have been
placed in synonymy of other species of Jacquemon¬
tia or transferred to Ipomoea, Iseia, or Odonellia.
The genus occurs primarily in the American trop¬
ics. The latest revisionary treatment of this genus
is the one by Robertson (1971), which considers 54
names in 27 species; this study did not include the
species from South America, only mentioning 40-
50 species in the region. Meissner (1869) referred
to 33 Brazilian species, and other treatments of the
genus are related to local floras (0’Donell, 1960a,
1960b; Austin, 1975, 1982a, 1982b).

During floristic surveys of the Convolvulaceae of
the Espinha^-o range in the state of Minas Gerais,
Brazil (Simao-Bianchini & Pirani, 1997), a new
species of Jacquemontia was found. Jacquemontia
revoluta Simao-Bianchini, here presented, is known
only from the Serra do Cipo, a region rich in en¬
demic taxa, many of them recently described.

Jacqucinontia revoluta Simao-Bianchini, sp. nov.
TYPE: Brazil. Minas Gerais: Santana do Ria-
cho, Serra do Cipo, Rodovia Belo Horizonte—
Conceig3o do Mato Dentro km 108, 7 May 1987
(fl, fr), R. Simao & V. C. Souza CFSC 10090
(holotype, SPF; isotypes, K, SP). Figure 1.

Suffutex, ramis erectis vel adscendentibus, 20-40 cm
longis, pubescentia adpresso-stellata, internodiis 0.5—
1 .5(— 2.5) cm longis. Folia attenuata linearia, acuta, mar-

gine revoluta, pubescentia stellata, parca, 2-5 cm longa,
0.5-1 mm lata, brevissime petiolata. Cymae 1-3 flora, pe-
dunculis axillaribus, 3-7 mm, bracteis exiguis, pedicellis
( 1 — )3 — 5 mm; sepala ovata vel ovalia-oblonga acutiuseula,
subaequalia, glabra, 2—1 mm longa, 1-3 mm lata; corolla
campanulata pallide coerulea, 7-10 mm longa; stigmata
2, ovalia, divaricata. Capsula bilocularis, 4-valvis; semina
glabra.

Perennial undershrubs, stems erect or ascending,
20-40 cm tall, glabrescent, trichomes appressed-
stellate, short-stalked, scale-like, internodes 0.5—
1.5(-2.5) cm long. Leaves simple, entire, attenuate-
linear, base and apex acute, margins revolute,
glabrescent, sparsely stellate trichomes with 5 ap-
pressed, equal branches or one bigger and patent,
blade 2-5 cm long, 0.5-1 mm wide, petiole 0-2
mm long. Inflorescence cymose 1— 3-flowered, ax¬
illary, peduncles 3-7 mm long, bracts linear, 1 mm
long, pedicels (1— )3— 5 mm long; 5 sepals subequal,
ovate to ovate-oblong, acute, glabrous, 2-4 mm
long, 1—3 mm wide; corolla campanulate, pale blue,
7-10 mm long, anthers ovate, with cordate bases
and obtuse apices. Ovary glabrous, 2-locular, the
locules 2-ovulate, stigmas 2, ellipsoid. Fruit cap¬
sular, 4-valvate, seeds 4, glabrous.

Superficially, J. revoluta resembles an Evolvulus
because of the erect habit and small flowers, but
its stigmas are characteristic of the genus Jacque¬
montia. This species is clearly distinct from the
others in Jacquemontia by the linear leaves and the
indument of appressed-stellate trichomes, some¬
what hyaline, with 5 equal branches (scale-like), or
4 small and 1 long and patent (Fig. IB).

Jacquemontia revoluta is similar to J. linoides
(Choisy) Meissner, but this species is entirely gla¬
brous, bearing leaves that are larger and without
revolute margins, and the inflorescence has a long
peduncle up to 7 cm, with 6 flowers. Both of them
are rare: Jacquemontia linoides is known only from
Bahia and Maranhao (the type was collected at Ser-
tao, maybe from Bahia), while J. revoluta is restrict¬
ed to stony slopes on the Serra do Cipo, where it
has always been found associated with ant nests.

Paratypes. BRAZIL. Minas Gerais: Santana do Ria-

Novon 9: 104-106. 1999.

Volume 9, Number 1
1999

Simao-Bianchini

Jacquemontia revoluta from Brazil

105

Figure 1. Jacquemontia revoluta Simao-Bianchini. — A. Flowering branch. — B. Pari of the leaf blade showing the
stellate trichomes and the revolute margins. — C. Detail of flower. — 1). Sepal. — E. Corolla split longitudinally to show
gynoeeium and stamens, the calyx having been removed. — F. Ovary transversally sectioned showing 2 locules with 2
ovules in each one.

106

Novon

cho, Serra do Cipo, Rodovia Belo Horizonte— Concei^ao do
Mato Dentro km 1 10, 24 Mar. I 986 (fl), I). C. Zuppi it-
Kameyama CFSC 9643 (SPF), 14 Apr. 1087 (fl, fr), V. C.
Souza CFSC 10075 (SI’, SPF); Chapeu de Sol. Dec. 1958
(fl), A. R Duarte 4544 (B, MB, SI).

Acknowledgments. The author thanks J. R. Pir-
ani for several useful comments on the original
manuscript and for help with the Latin version of
the description, and M. L. Kawasaki for correcting
the English and for helpful suggestions.

Literature Cited

Austin, D. F. 1975. Convolvulaceae. In: R. E. Woodson et
al.. Flora ol Panama. Ann. Missouri Bot. Card. 62: 157—
224.

- . 1982a. Convolvulaceae. In: G. Harling & B.

Sparre, Flora of Ecuador 1 5: 1 —98.

- . 1982b. Convolvulaceae. In: Z. Luces de Febres

& J. A. Steyermark (editors), Flora de Venezuela 8(3):
15-226.

Meissner, C. F. 1869. Convolvulaceae. In: C. F. P. Martius
(editor). Flora Brasiliensis 7: 199—370.

O'Donell, C. A. 1960a. Convolvulaceas argentinas II. Lil-
loa 30: 5—38.

- . 1960b. Las especies de Jacquernontia de Peru.

I.illoa 30: 7 1 -89.

Robertson. K. R. 1971. A Revision of the Genus Jacque¬
rnontia (Convolvulaceae) in North and Central America
and the West Indies. Ph. 14. Dissertation, Washington
University, St. Louis, Missouri.

Simao-Bianchini, R. & J. R. Pirani. 1997. Flora da Serra
do Cipo. Minas Gerais: Convolvulaceae. Bob Bot. Univ.
Sao Paulo 16: 125—149.

A New Species, Notes on Subgeneric Taxa, and New Synonyms in
Hessea (Amaryllidaceae: Amaryllideae) from South Africa

D. A. Snijman

Compton Herbarium, National Botanical Institute, Private Bag X7, Claremont 7735,

South Africa

ABSTRACT. Hessea tenuipedicellata, a new granite-
loving species from Namaqualand, shares slender,
pliant, upwardly curved pedicels, white flowers,
and channeled, narrow, tepals with II. stenosiphon
(Snijman) D. & U. Miiller-Doblies. Hessea subg.
Hessea is amplified to include //. tenuipedicellata
and H. stenosiphon , and H. subg. Kamiesbergia
(Snijman) Snijman is a new synonym of H. subg.
Hessea. In II. subg. Namaquanula (D. & U. Miiller-
Doblies) Snijman, Namaquanula etesionamibensis
D. & U. Miiller-Doblies is a new synonym of Hessea
bruce-bayeri (D. & U. Miiller-Doblies) Snijman.

Hessea Herbert, a genus of small autumn-flow¬
ering plants, is endemic to the winter and autumn-
rainfall regions of the Nama Karoo, Succulent Ka¬
roo, and Fynbos biornes in southern Africa.
Cladistic analysis of tribe Amaryllideae using mor¬
phological data showed that Hessea resolves within
subtribe Amaryllidinae in a terminal elade that in¬
cludes its aetinomorphic-flowered allies, Strumaria
Jacquin and Carpolyza Salisbury. The subtribe is
recognized by filaments that are connate at the base
but are secondarily free in Carpolyza and some spe¬
cies of Strumaria ; by seeds that have chlorophyll
in the integument; and by stomata on the testa
(Snijman & Linder, 1996). Other representatives of
the sub tribe that are basal to Hessea, Strumaria,
and Carpolyza are Amaryllis L., which has unique
pink to white seeds, Nerine Herbert, Brunsvigia
I leister, and Crossyne Salisbury. Hessea sensu Snij¬
man (1994) is divided into three subgenera that
share derived floral characters. The perigone in
Hessea is actinomorphic, and this ultimately turns
brown and remains open with age, unlike the per¬
igone that becomes deeply pigmented and finally
collapses in all other aetinomorphic-flowered
Amaryllidinae.

New Species

Hessea tenuipedicellata was first collected in leaf
on the southernmost boundary of the Namaqualand
rocky hills in 1992. The hysteranthous-leaved
bulbs flowered in cultivation one year later and pro¬

duced distinctive, delicate, white flowers, which
confirmed that the plant was undescribed. Like all
representatives of subgenus Hessea sensu Snijman
(1994), the species has distinctly centrifixed an¬
thers, a condition that results from the sheath in
the anther connective having equally long dorsal
and ventral walls (Miiller-Doblies & Muller- Dob-
lies, 1985). In addition, the plant has basally fused
stamens and two foliage leaves surrounded by a
sheathing cataphyll: characters that conform with
the narrowly defined Hessea sensu Miiller-Doblies
and Miiller-Doblies (1985, 1992).

Hessea tenuipedicellata Snijman, sp. nov. TYPE:
South Africa. Northern Cape Province: 3018
(Kamiesberg) farm Uilklip, N boundary of
Knersvlakte, in pockets of loamy soil on
granite domes (DC), 19 Apr. 1994, Snijman
1437 (holotype, NBC; isotypes, MO, PRE).
Figure 1 .

Species nova Hesseae stenosiphoni affinis, cujus pedi-
cellos tenues flexibiles sursum curvos, flores albos et te-
pala angusta canaliculata habet. Differt perigonii tubo
brevissimo et filamentis baud bis articulatis.

Delicate, deciduous bulbous herb. Bulb solitary,
deep seated, subglobose, ca. 15 mm across, covered
with brown, cobwebby tunics; producing annually
a non-sheathing prophyll, a sheathing cataphyll, a
sheathing foliage leaf, and a non-sheathing foliage
leaf; neck slender, up to 8 cm long. Foliage leaves
2, absent at anthesis, linear, opposite, spreading
slightly, up to 16 cm long, ca. 1 mm wide, glabrous.
Inflorescence umbel-like, lax, 13 cm across; scape
slightly curved, rigid, dry, slender, 7.5—13.0 cm
long, 1 mm thick, putty-colored, reddened basally;
spathe valves 2, linear-lanceolate, 1.5— 2.1 cm long,
1.5 mm across proximally, transparent with pinkish
veins; bracteoles not visible; pedicels pliant, slen¬
der, upwardly curved, 5-10 cm long, ca. 1 mm
across, green. Flowers 5—12, opening sequentially
at 1- to 4-day intervals, suberect to spreading, sub-
stellate, up to 15 mm across, glistening white, often
with a lemon-green center, sometimes with pale
green toward the base of the outer surface, cream

Novon 9: 107-1 10. 1999.

108

Novon

Figure 1. Hessen tenuipedicellata Snijman. — A. Inflorescence. — B. Mature leaves and bulb. — C. Flower from above.
— D. Flower, lateral view. — E. Flower with two tepals and stamens removed. — F. Anther, dorsal view. — (i. Anther,
ventral view. — H. Anther, lateral view. Scale bars: A, B = 10 mm; C— E = 5 mm; F— H = 1 mm. Drawn by John C.
Manning. Based on Snijman 1437.

Volume 9, Number 1
1999

Snijman

Hessea from South Africa

109

to pale lemon in bud, ± scented, remaining open
and turning light brown with age. Tepals 6, scarcely
connate but adnate to filament tube for ca. 1 mm,
recurved in the distal half, 5—9 mm long, 2.5 mm
wide, with a thickened midrib in the proximal quar¬
ter, channeled, sometimes with strongly undulate
edges. Stamens 6, 5—6 mm long, slightly shorter
than the tepals; filaments connate into a tube 1.5
mm long; anthers centrifixed, 1.5 mm long before
dehiscence, wine-red; pollen cream-colored. Ovary
with 3 ovules per locule. Style slender, up to 5 mm
long; stigma trifid; nectar cenlrally pooled around
style base. Fruit not seen.

Phenology. The main flowering period in the
wild is confined to a few weeks between mid-April
and early May. The leaves appear shortly after flow¬
ering and die back with the onset of the summer
drought, usually in September.

Distribution, habitat, and ecology. Hessea tenu-
ipedicellata is known from only one locality on the
eastern slopes of the Uilklipberg, an isolated gran¬
ite mountain surrounded by the Hat, sandy plains
of the northern Knersvlakte in Namaqualand, South
Africa. The population of a few hundred plants is
confined to pockets of soil on large granite domes.
Several patches showed disturbance by porcupines,
but the effect on the bulbs could not be gauged.
The permanent vegetation cover consists of sparse,
succulent perennials, of which the asteraceous stem
succulent, Senecio junceus ffarvey, is most abun¬
dant. The closely related //. stenosiphon (Snijman)
D. & U. Miiller-Doblies occurs in similar edaphic
sites approximately 65 km to the northwest, on ex¬
posed granite rocks of the Kamiesberg and its high-
lying foothills (Fig. 2). In addition to the geographic
distance that separates the species, several impor¬
tant differences in floral structure probably effect
ecological isolation. The flowers in //. tenuipedicel-
lata are substellate and short-tubed (1 mm long),
with widely spreading filaments and a well-exserted
style. In contrast, the hypocrateriform flowers of H.
stenosiphon have a style hidden within a narrow, 8—
12-mm-long perigone tube, and outer anthers that
occlude the perigone throat. Even if these species
attract the same insect visitors, it is nevertheless
likely that their flowers deposit pollen in places
where the stigma of the other species in the pair is
unable to contact it.

Most species of Hessea have straight, rigid ped¬
icels at anthesis, and flowers that vary from pink
to white with dark pink to red markings. In con¬
trast, H. tenuipedicellata is easily recognized by its
long, slender, pliant, upwardly curved pedicels, and
glistening white flowers that lack pink markings.

Furthermore, the substellate flowers have chan¬
neled, narrow tepals, and each flower opens at one-
to four-day intervals so that few reach anthesis si¬
multaneously (Fig. 1). The only other known Hessea
species with similar slender green pedicels, a se¬
quential flowering pattern, and channeled, narrow,
whitish tepals is //. stenosiphon. According to Snij¬
man (1994), this species belongs to the monotypic
subgenus Kamiesbergia (Snijman) Snijman, which
was defined by several unique characters: hypocra¬
teriform flowers; dimorphic stamens; and anthers in
which the pocket that encloses the filament t ip is
extremely short and is located near the base of the
connective.

Paratypes. SOUTH AFRICA. Northern Cape Prov¬
ince: 3018 (kamiesberg) farm Uilklip, SW of Kliprand
(DC), 16 Aug. 1992, Bruyns 5312 (k. ARC. RRE), 6 Apr.
1994, Snijman 1422 (NBG).

New Synonyms

Hessea subg. Hessea. TYPE: Hessea stellaris (Jac-
quin) Herbert, typ. cons.

Kamiesbergia Snijman, Bothalia 21: 125. 1991. Hessea

110

Novon

subg. Kamiesbergia (Snijrnan) Snijman, Contr. Bolus
Herb. 16: 71 . 1 994. Syn. nov. TYPE: Kamiesbergia
stenosiphon Snijman (= Hessea stenosiphon (Snijman)
I). & U. Miiller-Doblies).

Tht* morphology of Hessea tenuipedicellata,
which is intermediate between subgenus Hessea
and subgenus Kamiesbergia , presents evidence ad¬
ditional to that of Miiller-Doblies and Miiller-Dob-
lies (1992) that two of the three subgenera recog¬
nized by Snijman (1994) should be combined.
He ssea subg. Hessea is amplified to include H. tenu¬
ipedicellata and //. stenosiphon , and 11. subg. Ka¬
miesbergia is placed into synonymy under subgenus
Hessea. With this new delimitation, subgenus Hes¬
sea matches genus Hessea sensu Miiller-Doblies
and Miiller-Doblies (1992), except in rank and the
number of species recognized in each. The taxon is
characterized by derived eentrifixed anthers of
which the more or less basifixed state in //. steno¬
siphon (Snijman, 1994) is inferred to result from a
reduction in the length of the sheath in the anther
connective (Miiller-Doblies & Miiller-Doblies,
1992).

Hessea bruee-bayeri (D. & U. Miiller-Doblies)
Snijman, Contr. Bolus Herb. 16: 75. 1994. Na-
maquanula bruee-bayeri D. & U. Miiller-Dob¬
lies, Bot. Jahrb. Syst. 107: 20. 1985. TYPE:
South Africa. Northern Cape Province: 2816
(Oranjemund) Perdemonde, 1.5 km N of Ku-
boos/Oranjemund road on road to Koeskop
(BC), 10 Mar. 1979, U. & D. Miiller-Doblies
79112a (holotype, PRE).

Namaquanula etesionamibensis I). & LJ. Miiller-Doblies,
Feddes Repert. 105: 333. 1994. Syn. nov. TYPE:
Namibia. 2B17 (Vioolsdrif) Dabimub River, 27 km E
of Bosh Pinah waterworks, gravel terraces along
Orange Biver (AA). 8 Aug. I9BB. U. & I). Miiller-
Doblies 88070c (holotype, WIND; putative isotypes,
BOT, PRE, STE. not yet deposited, not seen).

With the discovery of H. tenuipedicellata, the ge¬
nus Hessea sensu Snijman (1994) is considered to
have 14 species. Described from a few collections
that flowered poorly in cultivation, Namaquanula
etesionamibensis D. & U. Miiller-Doblies was de¬
limited from Namaquanula bruee-bayeri D. & U.
Miiller-Doblies (= Hessea bruee-bayeri (D. & U.
Miiller-Doblies) Snijman) by small quantitative
characters, primarily of the flower: the perigone is
10-12 mm, not 5—8 mm long; the tepals are widely

as opposed to slightly spreading; and trichomes are
absent above the hook borne near the base of each
filament (Miiller-Doblies & Miiller-Doblies, 1994).
Additional specimens from the alluvial gravel
plains along the Orange River, which D. and U.
Miiller-Doblies did not examine, are florally inter¬
mediate between N. etesionamibensis and N. bruee-
ba yeri. Williamson 3405 (Snijman, 1994: fig. 34) in
particular has a perigone 5—6 mm long, spreading
tepals, and trichomes on and below the filament
hooks. To encompass the continuous, clinal varia¬
tion now evident in the populations located north
and south of the Orange River, Namaquanula ete¬
sionamibensis is placed into synonymy under H.
bruee-bayeri in //. subg. Namaquanula sensu Snij¬
man (1994). Thus the perigone of//, bruee-bayeri
varies from approximately as long as the filaments
in specimens from the Aus district, southern Na¬
mibia, to half as long as the filaments in those from
the Riehtersveld, South Africa.

Addit ional specimens examined. NAMIBIA. 2716
(Witputz), 27°50'S, 16°36'W (DC), 30 Mar. 1986, Van
lierkel 549 (NBC, PBE); 2810 (Oranjemund), 1 mi. N of
Orange Biver at Seedlings Drift (BB), Mar. 1960, Hall
1901 (NBC). SOUTH AFBICA. North* *rn Caj>e Prov¬
ince: 2816 (Oranjemund) Perdemonde, E of Arrisdrif
(BC), 1(1 Mar. 1977, Bayer 350 (NBC); 4 km NE of Bees-
bank (BC). 10 Mar. 1985, Williamson 3405 (k. NBC,
PBE).

Acknowledgments. 1 thank Danny Geldenhuys
for bringing this new species to my attention; Peter
Bruyns, Jasper and Bertha Coetzee, Callie and Alta
Meyer, and Colin Paterson- Jones for assistance dur¬
ing held studies; Peter Goldblatt for advice on the
manuscript; and John Manning for the illustration.

Literature Cited

Miiller-Doblies, I). & U. Miiller-Doblies. 1985. De Lili-
ifloris notulae 2. De taxonomia subtribus Strumariinae
(Amaryllidaceae). Bot. Jahrb. Syst. 107: 17-47.

- & - . 1992. De Liliilloris notulae 4. Notes

on Hessea (including Kamiesbergia: Amaryllidaceae of
Southern Africa). Gleditschia 20: 15—20.

- & - . 1994. De Liliifloris notulae 5. Some

new taxa and combinations in the Amaryllidaceae tribe
Amaryllideae from arid Southern Africa. Feddes Be-
pert. 105: 331—363.

Snijman, I). A. 1994. Systematics of Hessea, Strumaria
and Carpolyza (Amaryllideae: Amaryllidaceae). Contr.
Bolus Herb. 16: 1 — 1 62 .

- & H. P. Linder. 1996. Phylogenetic relationhips,

seed characters, and dispersal system evolution in
Amaryllideae (Amaryllidaceae). Ann. Missouri Bot.
Card. 83: 362-386.

A New Rhadamanthus Species (Hyacinthaceae) from the
Northwestern Cape, South Africa

D. A. Snijman and John C. Manning

Compton Herbarium, National Botanical Institute, Private Bag X7, Claremont 7735,

South Africa

Peter Goldblatt

B. A. Krukoff Curator of African Botany, Missouri Botanical Garden, P.0. Box 299, St. Louis,

Missouri 63166, U.S.A.

Abstract. Rhadamanthus involutus is a rare,
hysteranthous-leaved species from the Bokkeveld
Escarpment, west of Nieuwoudtville, Northern Cape
Province, South Africa. Most closely related to R.
alhiflorus B. Nordenstam, R. involutus is distin¬
guished by apically recurved outer tepals that are
conspicuously spotted with green at the base, and
suberect inner tepals that have involute margins
which overlap to form a tube. The bulbs grow on
exposed sandstone sheets between patches of dry
fynbos and flower in midsummer.

Rhadamanthus Salisbury is a small, poorly un¬
derstood genus of Hyacinthaceae that is endemic
to southern Africa. Nordenstam (1970) and Ober-
meyer (1980) recognized 11 species, mostly from
semi-arid habitats in the Nama Karoo, Succulent
Karoo, and Fynbos biomes (see Rutherford. f997).
The genus is distinguished from its close ally, Ur-
ginea Steinheil, solely by its derived anther dehis¬
cence. Unlike the unspecialized, longitudinally de¬
hiscent anthers of Urginea, the anthers in
Rhadamanthus dehisce either apically, or if by lon¬
gitudinal slits then these are initiated from the apex
and extend below the midline but never reach the
base of the thecae. Despite this variation, dehis¬
cence in Rhadamanthus is never complete and the
anther opening always remains somewhat pore-like,
having its greatest width near the top of the anther
(Nordenstam, 1970). Although subtle, Nordenstam
(1970) chose to maintain this distinction, awaiting
the resolution of the difficult and complex relation¬
ships between the remaining genera in tribe Scil-
leae (sensu Hutchinson, 1959). Jessop (1977) sup¬
ported this view in his re-evaluation of Drimia
Jacquin ex Willdenow, Urginea, and related genera.

Plants of Rhadamanthus are easily overlooked in
the field. The leaves are consistently hysteranthous,
the flowers are inconspicuous, and the bulbs flower
during the unfavorable summer drought. Thus the

geographic records of many Rhadamanthus species
in Nordenstam’s (1970) treatment of the genus are
regarded as incomplete, and the fruit and seeds of
several species still remain unknown.

Rhadamanthus involutus was first discovered in
1993 in flower in sparse, dry fynbos on the Bok¬
keveld Escarpment west of Nieuwoudtville, north¬
western Cape. The flowers differ uniquely in the
form, disposition, and markings of the outer and
inner tepals. The tepals are unequal and biseriate,
with the outer whorl distinctly overlapping the in¬
ner whorl at the base. The outer tepals recurve api¬
cally at anthesis and each bears a distinctive, basal,
olive green spot. In contrast the inner tepals are
unmarked, suberect, and involute so that each has
a tapering tubular form. The non-nodding flowers
are also unusual in the genus. Anther dehiscence
in R. involutus is tardy and incomplete, and the
longitudinal slits which proceed from the apex of
the thecae do not extend to the base. Obermeyer
(1980) recognized four other species of Rhadaman¬
thus (R. fasciatus B. Nordenstam, R. alhiflorus B.
Nordenstam, R. namibensis Obermeyer, and R. ka-
rooicus Obermeyer) with this type of anther dehis¬
cence and described subgenus Rhadamanthopsis
Obermeyer to accommodate them. Rhadamanthus
involutus is the fifth known species of this subge¬
nus. Although its affinities in the genus are not well
understood, R. involutus has white flowers and ex¬
tremely short filaments relative to the anthers.
These characters also occur in R. alhiflorus and
suggest an alliance with this southwestern Cape
species.

Field observations on the bees (family Antho-
phoridae) that visit the flowers of R. involutus in¬
dicate that the anther dehiscence in Rhadamanthus
is associated with pollination by bee vibration. Ad¬
ditional characters that Rhadamanthus species
share with other known buzz-pollinated taxa (Er-

Novon 9: 111-113. 1999.

112

Novon

Figure 1. Rhadamanthus involulus J. C. Manning & Snijman. — A. Inflorescence. — B. Bulb and mature leaves. — C.
h lower, view from above. — I). Flower, lateral view. — E. Anther after dehiscence. — F. Gynoecium. — G. Mature capsule.
— II. Seed. Scale bars: A, B = 1 cm: C— H = I mm. Drawn by John C. Manning. Based on Snijman & Manning 1525.

Volume 9, Number 1
1999

Snijman et al. 113

Rhadamanthus involutus from South Africa

ickson & Buchmann, 1983) are the stigma, which
is inconspicuous and dry, and the stamens, which
converge around the pistil. Whether the pollen
morphology of Rhadamanthus is consistent with
that of other buzz-pollinated taxa (see Erickson &
Buchmann, 1983) still requires confirmation, but it
is certainly dry.

Rliadainanthus involutus J. C. Manning & Snij¬
man, sp. nov. TYPE: South Africa. Northern
Cape Province: 3119 (Calvinia) Bokkeveld Es¬
carpment, farm Arendskraal, W of Nieuwoudt-
ville, in humus-rich soil amongst exposed
sandstone sheets (AC), 20 Dec. 1995, Snijman
& Manning 1525 (holotype, NBC; isotypes, K,
MO. PRE). Figure 1.

Bulbus ovoideus. Folia 3—6 filiformia suberecta glabra.
Seapus erectus minute puberulus prope basin. Bracteae
inferiores breviter calcaratae. Perigonium campanulatum,
tepalis biseriatis albidis basi connatis. Tepala exteriora
recurvata versus apicalia, conspicue maculata basi. Tepala
interiora marginibus involutis et imbricatis. Filamenta
brevissima. Antherae connivenes, thecis desuper tandem
supra basin dehiscentibus.

Deciduous bulbous herb. Bulb solitary or rarely
clumped, ovoid, 2-3 cm wide, extended into a neck
up to 1.5 cm long; outer coat brown, papery; inner
scales tightly overlapping, white, fleshy. Leaves hys-
teranthous, up to 6, suberect; base non-amplexi-
caul; blade filiform, up to 20 cm long, hemiterete,
glabrous, dark green with a pale apricot-colored
base, dying back from the tip. Inflorescences 2 per
shoot, usually maturing one at a time within the
same season; scape erect, straight, up to 16 cm
long, ca. 1 mm diam. at the middle, with minute
papillae arranged in vertical lines toward the base;
raceme ca. 3—7 cm long, straight, with 16—20 flow¬
ers arranged in a loose spiral; bracts 1. 5-2.0 mm
long, oblong to square, apiculate, the lower spurred;
spur 0.5— 1.0 mm long; pedicels patent, 3-4 mm
long, glabrous. Flowers spreading, campanulate, 2-
5 open at a time. Tepals biseriate, with the outer
whorl overlapping the inner whorl at the base, ca.
4.0 X 2.5 mm, glistening white with a greenish
brown midrib, basally fused for ca. 1 mm; outer
tepals ovate, shallowly cup-shaped in the proximal
half, bearing a conspicuous olive green eye out¬
lined with reddish brown at the base, recurved dis-
tally, with a minutely penicillate apex; inner tepals
shortly clawed and closely adpressed to the sta¬

mens basally, with involute margins in the upper
two-thirds forming a distally tapering tube, glabrous
at the apex. Stamens adnate to the perigone base
for ca. 0.5 mm; filaments smooth, flat, 0.5 mm long;
anthers basifixed, introrse, 1.5 mm long, brownish
orange, arching inward and covering the ovary, de¬
hiscing tardily by longitudinal slits from the apex
to just short of the base, with slits remaining widest
in the upper half; thecae rounded and glabrous ba¬
sally; pollen yellow. Ovary ovoid, 1.5 mm long,
brownish orange; style columnar, 1.5 mm long,
white, apically truncate with a shallowly three-sul-
cate stigma. Capsule ca. 3 X 3 mm, ovoid, erect on
curved erecto-patent pedicels, glossy, golden green.
Seeds compressed, elliptic-oblong to reniform, ca.
2 mm long, shiny black, irregularly folded and
wrinkled, finely reticulate.

Phenology. Peak flowering is limited to a few
weeks in late November and early December. The
leaves appear in autumn and die back with the ad¬
vent of the summer drought.

Distribution and habitat. Rhadamanthus invo¬
lutus is known from two populations on the Bok¬
keveld Escarpment, west of Nieuwoudtville in the
Northern Cape Province. The populations, compris¬
ing a few hundred plants, are found in seasonally
moist, open patches of humus-rich soil between ex¬
posed sandstone sheets. The surrounding vegeta¬
tion is sparse, dry fynbos.

Paratype. SOUTH AFRICA. Northern Cape Prov¬
ince: 31 19(Calvinia) Nieuwoudtville, top of Vanrhynspas
(AC). 30 Nov. 1993, Manning 2098 (NBG).

Literature Cited

Erickson, E. H. & S. L. Buchmann. 1983. Electrostatics
and pollination. Pp. 173—184 in C. E. Jones & R. J.
Little (editors). Handbook of Experimental Pollination
Biology. Scientific & Academic Editions, New York.
Hutchinson, J. 1959. The Families of Flowering Plants 2.

2nd ed. Clarendon Press, London.

Jessop, J. P. 1977. Studies in the bulbous Liliaeeae in
South Africa: 7. The taxonomy of Drimia and certain
allied genera. .1. S. African Bot. 43: 265—319.
Nordenstam, B. 1970. Studies in South African Liliaeeae
3: The genus Rhadamanthus. Bot. Not. 123: 155-182.
Obermeyer, A. A. 1980. A new subgenus Rhadarnant hop-
sis and two new species of Rhadamanthus. Bothalia 13:

137-139.

Rutherford, M. C. 1997. Categorization of biomes. Pp. 91 —
98 in R. M. Cowling, I). M. Richardson & S. M. Pierce
(editors), Vegetation of Southern Africa. Cambridge
Univ. Press, Cambridge.

Valeriana tajuvensis (Valerianaceae),
a New Species from Southern Brazil

Marcos Sobral

Faculdade de Farm&cia UFRGS, Ipiranga 2752, 90610-010, Porto Alegre, RS, Brazil.

sobral@vortex.ufrgs.br

Abstract. Valeriana tajuvensis, a species from
the Serra Geral massif in Rio Grande do Sul and
Santa Catarina, Brazil, is described anti compared
with related Brazilian species of the genus.

Resumen. Valeriana tajuvensis, especie de las
montafias de Serra Geral en Rio Grande do Sul y
Santa Catarina, Brasil, es descrita y comparada con
las especies brasilenas afines del g6nero.

The last specific study of the genus Valeriana in
Brazil (Borsini, 1962) listed 15 species, all of them
consisting of herbs restricted to the southern and
southeastern states of the country and generally
collected in montane or submontane habitats.

During recent fieldwork on a montane massif in
the southernmost Brazilian state of Rio Grande do
Sul, a strikingly distinct, arborescent, and some¬
what uncommon species of Valeriana was collected;
it is described here.

Valeriana tajuvensis Sobral, sp. nov. TYPE: Bra¬
zil. Rio Grande do Sul: Tajuva, mun. Morrin-
hos do Sul (29°21'S, 49°58'W), 19 Aug. 1995
(hermaphrodite fl), J. A. Jarenkow 2708 & M.
Sobral (holotype, ICN; isotypes, MBM, PEL,
SP). Figures 1, 2.

Species haec V. polystachyae et V. kurtzianae proxima,
a quarum habitus arborescenti, ramificatione sympodialis
foliisque apici ramorum aggregatis distincta est.

Gynodioecious shrub 1-3 m high, profusely sym-
podially ramified, the branches bearing conspicu¬
ous leaf scars; canopy 2.5—3 m diam. Leaves op¬
posite, somewhat tufted at the apex of the branches;
blades pinnatisect with elliptic-obovate profile,
8.5-18 cm long, 4-7 cm wide, 15-24-lobed (dis¬
counting 4—5 strongly reduced proximal lobes at
the base of pseudopetioles), the abaxial face very
sparsely glandular-punctate, the glands less than
0.1 mm diam., lobes with the margin slightly rev¬
olute, unequal in size, the 4—5 proximal pairs at
the base of the pseudopetioles ( 1— )3— 10 mm long,
(0.3)1— 1.5 mm wide, the others 30-45 mm long and

Novon 9: 114-117. 1999.

4—5 mm wide, markedly asymmetrical, the terminal
30-40 mm long and 3-3.5 mm wide, symmetrical;
pseudopetioles 4—15 mm long, each pair of leaves
connected at base forming a continuous nodal
sheath. Inflorescences paniculiform with the final
units dichasial, in hermaphrodite plants 20-30 cm
long and 10—15 cm wide, in pistillate ones 15-20
cm long and 5-8 cm wide, due to the reduction of
secondary ramifications (paracladia sensu Larsen,
1989); proximal bracts similar to leaves, about 6
cm long and 3 cm wide; distal bracts simple and
lanceolate, 8-10 mm long and 1—1.5 mm wide;
bracteoles lanceolate, 1.5-1. 8 mm long and 0.5 mm
wide, sometimes with sparse tufts of hairs 0. 1 mm
long in the base. Flowers sessile, pentamerous, the
calyx with 4—5 loosely distinct to completely fused
segments, forming a hyaline ring 0.2— 0.4 mm high;
corolla white, campanulate, with straight tube. Her¬
maphrodite flowers with corolla about 2 mm long
and 2 mm wide, lobes 0.8 mm long and 0.6-0. 7
mm wide; stamens attached at about the middle of
the corolla, 2—4, generally 3, exserted, filaments
about 2 mm long, anthers 2-lobed, globose, the loc-
ules in a same anther slightly unequal, about 0.3
mm long; style included, about 1 mm long, stigmas
3, 0.2-0. 3 mm; ovary inferior, triquetrous in shape,
filled with parenchymatous tissue, unilocular, uni-
ovulate; ovule pendulous. Pistillate flowers with co¬
rolla 0.5— 0.8 mm long and 0.5 mm wide, the lobes
about 0.3 mm long and wide; stamens 3, included,
0.2-0.3 mm long, anthers globose with somewhat
unequal locules, about 0.2 mm long, not producing
pollen in the flowers examined; style 1.3-1. 5 mm
long, stigmas 3, exserted, 0. 1-0.2 mm long; ovary
the same as in hermaphrodite flowers. Fruits tri¬
quetrous, asymmetrical in transversal view, 2-2.3
mm long and 1.1-1. 3 mm wide, uniseminate,
sparsely pilose, eventually somewhat more densely
so at the edges. Seeds examined immature.

Arborescent species of Valeriana are quite com¬
mon in the Andean highlands of Colombia, Peru,
anil Venezuela (Killip, 1925, 1928; Cuatrecasas,
1941; Xena, 1992), the region where Valeriana at-

Volume 9, Number 1
1999

Sobral

Valeriana tajuvensis from Brazil

115

Figure 1. Valeriana tajuvensis. Branch of hermaphrodite plant, drawn from holotype (Jarenkow 2708).

tains its highest diversity in the Neotropics (Meyer,
1951). Nevertheless, woodiness was not known in
eastern South American species (Muller, 1885;
Borsini, 1962, 1963).

The leaves of V. tajuvensis resemble those of

southern Brazilian and Argentine V. polystachya
and V. kurtziana, hut it may be distinguished from
these two species by the following characters:

Habit herbaceous, sometimes rhizomatous, never

116

Novon

Figure 2. Flowers and fruits from Valeriana tajuvensis. — A. Hermaphrodite flower ( Jarenkow 2708). — B. Feminine
flower ( Jarenkow 2700). — C. Fruit in frontal view. —I). Fruit in apical view. — E. Immature fruit, transversal section,
schematic: I = locale, o = ovule, p = parenchymatous tissue (C— F, Sobral 7945b).

woody; ramification monopodial; leaves equally

distributed along the entire plant .

. Valeriana kurtziana and V. polystachya

Habit arborescent, woody, never rhizomatous; rami¬
fication sympodial; leaves concentrated at the apex
of the branches . Valeriana tajuvensis

Phenology. Flowering and simultaneously fruit¬
ing specimens were collected in July, August, and
December.

Distribution. The species has been collected in
Serra da Tajuva, in the state ol Rio Grande do Sul,
and Serra do Faxinal, state of Santa Catarina.

Ecology. Heliophilous and subxerophilous
shrub growing on the eastern slopes ol the southern
Brazilian massif named Serra Geral, at altitudes be¬
tween 500 and 700 m above sea level; at the col¬
lection site in Tajuva, it occurs either sporadically
at the edge of submontane forests or in great den¬
sities of hundreds of individuals in more openly
insolated and well-drained rocky fields; at Serra do

Faxinal only two individuals were observed on a
rocky roadside (Falkenberg, pers. comm.).

I'aratypes. BRAZIL. Rio Grande do Sul: Serra da
Tajuva, municipality of Morrinhos do Sul (29°2LS,
49°58'W), July 1995 (pistillate), Sobral 7924a (ICN), July
1995 (hermaphrodite). Sobral 7924b (ICN). Aug. 1995
(pist.). Sobral 7945a <£: Miro (ICN. RB), Aug. 1995
(herm.), Sobral 7945b Miro (ICN). Aug. 1995 (pist.).
Jarenkow 2702 & Sobral (ICN. MBM, PEL), Aug. 1995
(pist.) Jarenkow 2700 & Sobral (FLOR, ICY MBM. PEL,
RB, SP), Dec. 1995 (pist.). Sobral 8010 & Jarenkow
(1 LOB. MBM. MO). Santa Catarina: Serra do Faxinal,
municipality of Praia Grande, Mar. 1997 (herm.), Falken¬
berg 9797 (FLOR).

Acknowledgments. I am grateful to Angela Ri-
beiro and Paulo Hafner for introducing me to Serra
da Tajuva; to Cfntia Miro and Joao Andre Jarenkow
for their kind and valuable help with the collec¬
tions; and to Daniel B. Falkenberg and Donna M.
E. Ware for keen and useful observations on the
manuscript.

Volume 9, Number 1
1999

Sobral

Valeriana tajuvensis from Brazil

117

Literature Cited

Borsini, 0. K. 1962. Revisi6n de las valerianaceas de
Brasil. Lilloa 31: 149-170.

- . 1963. Valerianaceas del estado de Santa Catarina

(Brasil). Sellowia 15: 123—136.

Cuatreeasas, J. 1941. Notas a la flora de Colombia, IV.

Revista Acad. Colomb. Ci. Exact. 4: 337—348.

Killip, E. P. 1925. Twelve new species of Valeriana from
the Andes of South America. J. Washington Acad. Sei.
15: 450-456.

- . 1928. Seven new species of Valeriana from Co¬

lombia and Peru. .). Washington Acad. Sci. 18: 498—
501.

Larsen, B. B. 1989. A taxonomic review of Phyllaetis and
Valeriana sect. Bracteata (Valerianaeeae). Nordic J. Bot.
6: 427-446.

Meyer, F. 1951. Valeriana in North America and the West
Indies (Valerianaeeae). Ann. Missouri Bot. Card. 38:
377-485.

Muller. C. A. 1885. Valerianaeeae. In: k. Martins (editor).

Flora Brasiliensis 6(4): 339—350.

Xena de Enrech, N. 1992. Valerianaeeae. In: Flora de
Venezuela 5(1): 221-267.

A New Species of Psychotria Subg. Heteropsychotria (Rubiaceae)
from the Adolpho Ducke Forest Reserve, Brazil

Charlotte M. Taylor

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166, U.S.A.
Marina Thereza V do A. Campos

Universidade Federal do Acre, Parque Zoobotanico, Caixa Postal 1035, 69908-210,

Rio Branco, Acre, Brazil

Abstract. The new species Psychotria rhombi-
bractea C. M. Taylor & M. T. Campos, from wet
forest in the Adolpho Ducke Forest Reserve, Ma¬
naus, Amazonas, Brazil, is here described and il¬
lustrated.

Re SUMO. A nova especie Psychotria rhombibrac-
tea C. M. Taylor & M. T. Campos, com ocorrencia
na florestal tropical umida, na Reserva Florestal
Adolpho Ducke, Manaus, Brazil, 6 descrita e ilus-
trada.

During review of the Rubiaceae collected from
the Adolpho Ducke Forest Reserve, in Manaus,
Brazil, the following undescribed species was en¬
countered.

Psychotria rhombibraetea C. M. Taylor & M. T.
Campos, sp. nov. TYPE: Brazil. Amazonas: Re¬
serva Florestal Ducke, Manaus-Itacoatiara
road. Km 26, 02°53'S, 59°58'W, Floresta de
Plato, 28 setembro 1994 (fl), M. J. G. Hopkins,
E. da C. Pereira & C. F. da Silva 1483 (holo-
type, INPA; isotype, MO-5056257). Figure 1.

Haec species a Psychotria officinali (Aublet) Sandwith
inflorescentiae bracteis ellipticis oblanceolatis vel ple-
rumque rhombicis, corolla ex tubo 9-10 mm longo et lobis
5—8 mm longis constante atque fructus 4—6 X 6—8 mm
pyrenis rugosis distinguitur.

Shrubs or small trees to 5 m tall; vegetative or¬
gans glabrous. Leaves opposite; blades elliptic to
elliptic-oblong, 7—18 X 2—7 cm, at apex acuminate
to rather long-acuminate, at base acute to cuneate,
drying papyraceous; secondary veins 6—10 pairs,
often uniting into a looping submarginal vein at
least in distal part of blade, on both surfaces plane
or sometimes prominulous; petioles 5—15 cm long;
stipules persistent, united around the stem into a
truncate or concave sheath 0.5—1 mm long, with
lobes 2 on each side, deltoid, 0.5—1 mm long, cil-
iate at least when young. Inflorescences terminal.

glabrous to minutely puberulous, with peduncle 8—
20 mm long, flowering portion cymose, in outline
rounded-corymbiform, 1—1.5 X 2^1 cm (excluding
corollas), with secondary axes 1-3 pairs, paired but
often appearing subverticillate, generally dichasial,
with cymules subtended by bracts, these elliptic,
oblanceolate, or usually rhombic, 3-10 mm long,
obtuse to acute, white; flowers apparently distylous
(this state not confirmed with field breeding study),
sessile in cymules of 5—7, not individually bracte-
ate; hypanthium (ovary portion) cylindrical to ellip¬
soid, 0.5-0.8 mm long, glabrous to minutely pu¬
berulous; calyx limb 0.2-0. 5 mm long, minutely
puberulous, shortly dentate, entire or ciliolate; co¬
rolla salverform to narrowly infundibuliform, white,
externally and internally glabrous, tube 9-10 mm
long, ca. 1 mm diam. near middle, lobes narrowly
triangular to narrowly lanceolate, 5—8 mm long, 1—
1.5 mm wide, at apex shortly rounded and a little
thickened; stamens inserted above middle of corolla
tube, in short-styled form with anthers ca. 2 mm
long exserted on filaments ca. 4 mm long, in long-
styled form with anthers ca. 2 mm long, subsessile,
these anthers positioned near middle of corolla
tube; stigmas linear, in short-styled form ca. 2 mm
long and positioned near middle of tube, in long-
styled form ca. 1 mm long and shortly exserted;
disc annular, smooth, slightly exceeding calyx limb.
Infructescences with bracts becoming purple to
wine-red, magenta, red, or orange; fruits subglo-
bose, somewhat didymous, 4—6 X 6—8 mm, red or
purple-red; pyrenes 2, hemispheric to subglobose,
on abaxial (dorsal) surface faintly longitudinally
ridged and somewhat rugose throughout, on adaxial
(ventral) face planar with a slight central ridge.

This species is known only from the Manaus
area, where it has been collected primarily in the
Adolpho Ducke Forest Reserve. Here it has been
collected with flowers September-November, with

Novon 9: 118-119. 1999.

Volume 9, Number 1
1999

Taylor & Campos

Psychotria rhombibractea from Brazil

119

Figure 1. Psychotria rhombibractea C. M. Taylor & M. T. Campos, based on Hopkins et al. 1483. — A. Flowering
branch. — B. Corolla and style of long-styled flower, partially dissected. — C. Cymule with bracts. B, C to I -cm scale.

fruits February-April. June, and August. The com¬
mon name is reported to he “mata-calado.”

This new species is similar to Psychotria offici¬
nalis (Aublet) Sandwith, with which it has been
confused; P. officinalis can be distinguished by its
subulate to triangular bracts 4 — 7( — 13) mm long, vs.
elliptic, oblanceolate, or usually rhombic bracts 3—
10 mm long in P. rhombibractea ; corolla with the
tube 2.5—3 mm long and lobes ca. 2 mm long, vs.
the tube 9-10 mm long and lobes 5—8 mm long in
P. rhombibractea ; disc that is shorter than the calyx
limb, vs. longer than the calyx limb in P. rhombi¬
bractea ; and iruits 3 X 3—5 mm, vs. 4—6 X 6—8
mm in P. rhombibractea. The specific epithet of P.
rhombibractea refers to the most common and dis¬
tinctive form of the inflorescence bracts. The py¬
renes of P. rhombibractea are relatively large, and
their rugose surface is unusual.

Psychotria rhombibractea belongs to subgenus
Heteropsychotria Steyermark, section Pseudoce-
phaelis Steyermark (Steyermark, 1972); within this
section, it appears to belong to Series B, “Pseu-
docephaelis.”

Paratypes. BRA/11,. Amazonas: Reserva Florestal
Ducke, Manaus-ltacoatiara road. Km 26, 02°53'S,
59°58'W, 8 fev. 1996 (fr), Campos et al. 461 (INPA). 6
mar. 1996 (fr), Campos & Pereira 536 (INPA, MO). 6 nov.

1961, Rodrigues & Lima 3534 (INPA, NY), 9 jun. 1963
(fr). Rodrigues A Freitas 5322 (INPA, NY), 23 ago. 1994
(fr), Sothers & Silva 124 (INPA, MO); na beira da estrada
para a torre, perto da placa de 0.5 km. 29 mar. 1996 (fr).
Campos & Silva 586 (INPA, MO); along path, 25 out. 1977
(fr). Keel 189 (NY); prdximo a Torre, 18 mar. 1991 (fr).
Mota A- Santana 3 (INPA, MO); plantio do Pan Rosa na
direyao Acara, 14 jun. 1994 (fr). Ramos 2826 (INPA. MO);
a margem do caminho para o aeampamento da Silvicul-
tura, 5 jun. 1963 (fr), Rodrigues 5254 (INPA, NY); perto
da estayao Meterologica, abr. 1973 (fr), Rodriguez & Silva
9110 (INPA, MO); trilha a esquerda do Km 0.35 da es¬
trada A lojamento— Torres, 11 mar. 1994 (fr), Vicentini &
Silva 418 (INPA, MO); prdximo a torre de observayao,
trilha Torre-Acara, 12 nov. 1993 (fl), Vicentini & Assunydo
372 (INPA. MO).

Acknowledgments. We thank the Adolpho
Ducke Forest Reserve Flora Project and the cura¬
tors of INPA and NY for access to specimens; the
Overseas Development Agency for financial sup¬
port; Paulo Lopes Assunyao and Everaldo Pereira
for field assistance; R. E. Gereau for preparation of
the Latin diagnosis; and Brian Boom for his counsel
and help.

Literature Cited

Steyermark, J. A. 1972. Psychotria. In: B. M. Maguire &

Collaborators, Flora of the Guayana Highlands. Mem.

New York Bot. Card. 23: 406—717.

Carex tiogana (Cyperaceae), a New Sedge from the
Sierra Nevada, California

Dean W. Taylor

University Herbaria, University of California, Berkeley, California 94720, U.S.A.

Joy Mastrogiuseppe

Department of Anthropology, Washington State University, Pullman,
Washington 99164-4910, U.S.A.

ABSTRACT. Carex tiogana is described as a new
species in section Capillares from the Sierra Ne¬
vada, California, similar to C. capillaris hut with
culms usually shorter than or equal to the leaves,
serrulate leaf midribs, conspicuous riblike veins on
abaxial leaf surfaces, no enlarged vein at leaf mar¬
gin, long prickles on leaf margins and pistillate
scale midribs, perigynia fewer per spike and usu¬
ally shorter, with longer prickles toward beak, usu¬
ally shorter beaks, and shorter achenes.

A diminutive plant belonging to Carex sect. Ca¬
pillares Ascherson & K. Graebner was discovered in
the Tioga Pass region on the eastern border of Yo-
semite National Park, Mono County, California. No
member of section Capillares has previously been
reported from California (Mackenzie, 1922; Howell,
1959). The section includes two species known to
occur in North America: Carex capillaris L. and C.
williamsii Britton (Mackenzie, 1931-1935: 295-
297). Carex williamsii is known only from calcareous
substrates in the high Arctic, while C. capillaris is
a widely distributed cireumboreal plant often sub¬
divided into a number of infraspecific taxa or seg¬
regate species on the basis of growth habit, leaf
width, the sex of flowers in the terminal spike, spike
length, and perigynium size (Packer, 1983; Porsild
& Cody, 1980; Scoggan, 1978; Polunin, 1943; Ko¬
marov, 1935). These segregate taxa, here treated as
subspecies, are C. capillaris subsp. chlorostachys
(Steven) Love, Love & Raymond, C. capillaris subsp.
krausei (Boeckler) Bocher, and C. capillaris subsp.
porsildiana (Polunin) Bocher.

Carex tiogana D. Taylor & J. Mastrogiuseppe, sp.
nov. [sect. Capillares]. TYPE: U.S.A. California:
Mono County, headwaters of Parker Creek drain¬
age, Mono Basin, along trail between Parker
Pass and Koip Pass, Inyo National Forest
(37°49'28"N, 119°11'25"W), 3260 m, 17 July
1988, D. W. Taylor & K. A. Teare 9981 (holotype,
UC; isotypes, COLO, NY, US). Figure 1.

Pianta cespitosa; folia longiora eulrnis vel aequa, ar-
cuata vel faleata, crassa, nervata abaxialiter, margines fo-
liorum serrulatae, 8-12 aculei per mm, aculeus 0.1-0.14
mm longus, costa abaxialis serrulata. Squamae feminae
costae serrulatae, 7-15 aculei per mm. Periginia 2-5(-8)
per spicam, (1.2— )1.3— 1.8(— 1.9) mm longa, binervata, ner-
vi serrulati versus rostrum, aculei 6-15, aculeus 0.05-0.2
mm longus; rostrum (0.1-)0.3-0.5(-0.7) mm longum ail
achenium. Achenia 0.8-1 . 1 mm longa.

Perennial, densely caespitose in small clumps.
Culms 1.7— 6(— 7.3) cm, shorter than or about equal¬
ing tin1 leaves (occasionally slightly longer). Leaves
generally clustered on the lower one-quarter of the
culm, blades thick, stiff, bright green, arching to
falcate, channeled along the midrib, (1 .3—) 1 .7 — 6.5
cm X 0.9— 2.2(— 3) mm, with 3—5 prominent veins
abaxially but without thick vein at margin, margins
and midrib with 8-12 stifl prickles per mm, prick¬
les 0.1—0.14 mm long. Inflorescence bracts with
sheath 3-5 mm long, blade 4-26 mm long. Ter¬
minal spike staminate, 5—8 mm long; staminate
scales straw-colored with yellow serrulate midrib.
Pistillate spikes usually 2 or 3, 3-5(-7) mm long,
on slender ± nodding peduncles 8—11 mm long;
pistillate scales ovate, shorter and wider than peri¬
gynia, white-hyaline with green midrib, the apex
rounded or with a short-apiculate extension of the
midrib, midrib of at least some lower scales bearing
7-15 prickles per mm (0.025— )0. 5— 0.7 mm long,
scales deciduous as perigynia mature. Perigynia
(2— )5 — 8 per spike, obovate, (1 .2— )1 .3— 1 .8(— 1 .9) X
(0.4— )0. 7— 0.9 mm, trigonous, glossy, body green to
brown (chestnut), with usually two yellow-green
ribs bearing 6-15 prickles toward the beak, prick¬
les 0.05-0.2 mm long; beak (0.1-)0.3-0.5(-0.7)
mm from tip to achene. Aehene trigonous, 0.8-1. 1
X 0.4— 0.8 mm, filling perigynium body; stigmas 3.
The plants are more vigorous when garden-grown
(in Berkeley, California, and Moscow, Idaho) but
are still small with culms shorter than to somewhat
longer than the leaves and retain the other char¬
acteristic features (Table 1).

Novon 9: 120-123. 1999.

Volume 9, Number 1
1999

Taylor & Mastrogiuseppe
Carex tiogana from California

121

mm

D

Figure 1. Carex tiogana D. Taylor & J. Mastrogiuseppe. — A. Habit. — B. Surfaces of leaf blade showing adaxial
channeling, serrulate margins, abaxially riblike veins, and serrulate midrib. — C. Pistillate scale with perigynium
(abaxial view). — I). Adaxial view of perigynium.

Carex williamsii is distinguished from the Sier-
ran plants and from C. capillaris by having very
narrow folded or involute leaves with 12-15 mar¬
ginal prickles per millimeter and by perigynia usu¬
ally having veins between the two ribs (Table 1).
Characters distinguishing the subspecies of C. ca¬
pillaris are listed in Table 1. The Sierran plants can
be separated lrom C. capillaris without reference to
geographic origin on the basis of culm length rel¬
ative to leaves; leaf texture, venation, and prickles
on midrib and margins; prickles on pistillate scale
midribs; number of perigynia per spike; perigynium
length, beak length, prickles, and stipe; and achene
length (Table 1). Certain individual morphologic
features of the Sierran plants occur in some popu¬
lations of C. capillaris , but not in combination with
other features characteristic of the Sierran plants.

The Sierran plant is currently known only from
three small populations in Mono County, California,
all beside lakes fed by glacial meltwaters and un¬
derlain by calcareous substrates. The clones of C.
tiogana occupy small areas (no more than 400 m2),
making the species extremely vulnerable to human
disturbance. Carex tiogana is both moqihologically

distinct and highly disjunct from other members of
section Capillares. The closest stations are for C.
capillaris subsp. capillaris , in the Ruby Mountains
of east-central Nevada (Lewis, 1971) — 400 km dis¬
tant, and subspecies chlorostachys in the Wallowa
Mountains of northeastern Oregon (Mason, 1975) —
550 km distant.

Paratypes. U.S.A. California: Mono County, at the
type locality, 10 Aug. 1986. I). W. Taylor 8831 (UC, WS),
1 Aug. 1987, I). 11. Taylor 9198 (CAS, RSA, UC); marginal
meadow at outlet of Cooney Lake, Virginia Lakes basin,
Toiyabe National Forest (38°02,52”N, 1 19°17'01"W), 3102
m. 28 July 1988. I). W. Taylor & K. A. Teare 9994 (UC,
WS), 28 July 1988, Glenn L Clifton 18.329 (PUA); turf bank
at outlet of Upper Sardine Like, Bloodv Canyon, 3154 m,
24 July 1992, IJ. W. Taylor 13061 (UC).

la. Leaves 0.5— 0.8 mm wide, most tightly folded or
involute, margins with 12-15 prickles per mm

. Carex williamsii

lb. Leaves 0.9—3 mm wide, most flat or shallowly
channeled, margins with 0—12 prickles per mm.

2a. Culms usually longer than leaves, sometimes

equaling leaves; leaf midribs abaxially
smooth (occasionally sermlate at tip), prick¬
les on leaf margins 0— 0.08(— 0.1) mm long;
pistillate scale midribs usually without

Sierran Carex capillaris

plants subsp. capillaris subsp. chlorostachys subsp. krausel subsp. porsildiana Carex williamsii

Growth habit densely tufted densely tufted loosely tufted, rhizoma- densely tufted densely tufted loosely tufted, rhizoma-

tous tous

Leaf blades thick, ± falcate thick or thin. us. not thin, “grassy” thick, not falcate thick, occ. falcate thick, not falcate

122

Novon

1 3

. . i

> «

t _

_2

_2 i

_* £

o

E c

■i S

-r -r 4J

2 7

7

O -

o ^

O d

lA 5

S j 3

oi

-7

2 i I

<0 V

I

2 5

ci o

_ ^E

1 =

sd JJ «

a -x "5

C C _Q

'5b '5c d

a! cfl

E E E

A £
A S

cO A

iA _

JJl

A

u J »

X A
2 A

w A »

uA

£1 A

_ A

ro ^

-a £

i -c 5

sq .5? -S
T ts E
r- cd

q vi S

JS2 1

SC bC *

eo.~'o3''o3va3_!E-5-5 E

i Cl d 3 3 ji

■ — 1 j j u u H

T 7

pc —

o

r 7

^ ©

7 7

J,

7 7

cA

2 J

q i

3

si 3

ci si d

CN

uA

O rA
cA g oo ~
io <n q d- i

O O LC (N
I r — ■ I — !

O'

7

_ _ _ r-

os ©

7 4

E

E

4) ' — '

_£
a. wd

-c -c

E c
E d-

4) 03 £ £

S 7

a. r-

2 d o

&U

!S — ’

7 7 7

CM — up

3 7 2 7

S d -1 d

CO 00

7 7

CM sO

7

io q «?

7 5 1 2 1

id sO „

• d “

d t q *p>

X--L7 o

i/5 t A C

° 7 qj

A 2 ^ -A ^

12 -c 2

£ £

_E - c E
-* -* d E c
u a a -s

? *p _c
^ a * Ec -

I I I Jl

•c

— — S£ SB SB So Jo 5d 5d

•c d d 'C d

E .2?

& Sc li

C d -=

a. a. a. a. a.

a. a. a. < <

Volume 9, Number 1
1999

Taylor & Mastrogiuseppe
Carex tiogana from California

123

prickles (occasionally bearing prickles up to
0.05 mm long); achenes 1—1.9 mm long . .

. Carex capillaris

2b. Culms shorter than or about equaling leaves;
midribs of at least some leaves abaxially ser¬
rulate at midlength, prickles on leaf margins
0.1—0.14 mm long; pistillate scale midribs
with 7—15 prickles (0.025— )0. 5— 1 0.7 mm

long; achenes 0. 8-1.1 mm long .

. Carex tiogana

Acknowledgments. We thank Hack Chin Kim,

Sarah Moore, and the herbaria loaning specimens:
CAN, CAS/DS, MT, and RSA/POM.

Literature Cited

Howell, J. T. 1959. Carex. In: P. A. Munz & D. D. Keek.

A California Flora. Univ. California Press, Berkeley.

Komarov, V. L. [Editor |. 1935. Flora of the U.S.S.R. Vol.

III. Botanical Institute, Academy of Sciences, U.S.S.R.

[Translated from Russian, Israel Program for Scientific

Translations (1964). |

Lewis, M. E. 1971. Flora and Major Plant Communities
of the Ruby-East Humboldt Mountains, Nevada. U.S.
Forest Service, Ogden, Utah.

Mackenzie, K. K. 1922. A monograph of the California
species of the genus Carex. Erythea 8: 1-95.

- . 1931-1935. Carex. North American Flora 18: 9-

478.

Mason, G. 1975. Guide to the Plants of the Wallowa
Mountains of Northeastern Oregon. Museum of Natural
History, University of Oregon. Eugene.

Packer. J. G. 1983. Flora of Alberta. 2nd Ed. Univ. To¬
ronto Press, Toronto.

Polunin, N. 1943. Contributions to the flora and phyto¬
geography of southwestern Greenland. J. Linn. Soc.,
Bot. 52: 349-406.

Porsild, A. E. & W. J. Cody. 1980. Vascular Plants of
Continental Northwest Territories, Canada. National
Museum of Natural Sciences Publications in Botany,
Ottawa.

Scoggan, H. J. 1978. The Flora of Canada. Part 2, Pteri-
dophyta, Gymnospermae, Monocotyledoneae. Publ. Bot.
(Ottawa) 7(2).

Transfer of Four Peruvian Altamiranoa Species to Sedum

(Crassulaceae)

Joachim Thiede

Botanical Institute, University of Cologne, GyrhofstraBe 15, D-50931 Cologne, Germany.

jthiede@novell.biolan.uni-koeln.de

Henk V Hart

Department of Plant Ecology and Evolutionary Biology, Utrecht University, Padualaan 8, NL-
3584 CH Utrecht, The Netherlands. HtHart@bio.uu.nl

Absthact. Recently Altamiranoa Rose ( = VUIadia
sect. Altamiranoa (Rose) R. T. Clausen) was merged
with Sedum L. For four Peruvian Altamiranoa species
without valid names in Sedum we propose three new
combinations: Sedum decipiens (Baker) Thiede & 't
Hart, Sedum reniforme (H. Jacobsen) Thiede & 't
Hart, and Sedum weberbaueri (Diels) Thiede & 't Hart,
and one new name, Sedum plicatum Thiede & ’t Hart.
A lectotype is designated for Cotyledon decipiens Bak¬
er, the basionym of Sedum decipiens.

Rose (in Britton & Rose, 1903: 3) described the
genera Altamiranoa Rose and Villadia Rose to ac¬
commodate some Mexican Crassulaceae with sym¬
petalous flowers. Previously these species had mostly
been classified in Cotyledon L., the hold-all of sym¬
petalous Crassulaceae. Berger (1930) included both
genera in subfamily Echeverioideae A. Berger, but
noted that several Altamiranoa species are very sim¬
ilar to Sedum. Berger (1930) considered Altamiranoa
and Villadia to be closely related though they differ
in the structure of the inflorescences. Altamiranoa
has predominantly eymose inflorescences, whereas
those of Villadia are usually spicate, racemose, or,
most frequently, thyrsoid. Froderstrom (1936) more
or less accepted Rose’s concept, but transferred sev¬
eral Altamiranoa species to Sedum. Baehni and
Macbride (in Baehni, 1937), on the other hand, unit¬
ed Altamiranoa and Villadia, and Clausen (1940)
distinguished Altamiranoa as a section of Villadia.
Baehni s and Clausen’s classifications have been
widely accepted until recently, when Moran (1996)
merged Villadia seel. Altamiranoa (Rose) R. T. Clau¬
sen (= Altamiranoa) with Sedum. In particular the
occurrence of intermediate forms bridging the gap
between Altamiranoa and Sedum , and his doubts
about the monophyly of Villadia in the sense of
Baehni and Macbride and Clausen, prompted Mor¬
an's decision. Moran (1996) already made the nec¬
essary new combinations under Sedum for three

Mexican Villadia (sensu lato) species. Here we com¬
plete the transfer ol Altamiranoa to Sedum with three
new combinations and one new name for four Pe¬
ruvian species.

Sedum decipiens (Baker) Thiede & 't Hart, comb,
nov. Basionym: Cotyledon decipiens Baker, Re¬
fug. Bot. 3, t. 200. 1870. Echeveria decipiens
(Baker) E. Morren, Belgique Hort. 24: 159.
1874. Altamiranoa decipiens (Baker) Froder¬
strom, Acta Horti Gothob. 10, App.: 145.
1936. Villadia decipiens (Baker) H. Jacobsen,
Natl. Cact. Succ. .1. 13: 76. 1958. TYPE: Peru.
Sine loco, sine datum, Farris s.n. (holotype,
not extant); Refug. Bot. 3, tab. 200. 1870 (lec¬
totype, designated here).

The species is known only from the type collec¬
tion and is as yet not validated by additional col¬
lections (Brunner, 1993: 376). No type material is
extant at Kew (where J. G. Baker worked), at the
British Museum (Baker described many species
from specimens deposited at BM), or at Oxford (ac¬
cording to a note in the protologue Baker appar¬
ently obtained living specimens of Farris’s collec¬
tion via W. W. Saunders, whose herbarium is now
kept at OXF). Therefore, we designate the excellent
figure (tab. 200) accompanying the diagnosis as the
lectotype of Cotyledon decipiens Baker.

Sedum plicatuni Thiede & *t Hart, nom. nov. Re¬
placed name: Cotyledon stricta Diels, Bot. Jahrb.
Syst. 37: 410. 1906. Altamiranoa stricta (Diels)
A. Berger, in Engler & Prantl, Nat. Pflanzenfam.
ed. 2, 18a: 470. 1930. Villadia dielsii Baehni &
.1. F. Macbride, Candollea 7: 285. 1937 (nom.
nov. pro Cotyledon stricta Diels due to Villadia
stricta Rose, 1905). TYPE: Peru. Dep. Ancahs
(Ancachs): pr. Caraz in rupestribus eamporum,
2200-2500 m s.m., flor. m. Maj. 1903, Weber-
bauer 3(XX) (holotype, B).

Novon 9: 124—125. 1999.

Volume 9, Number 1
1999

Thiede & ’t Hart
Peruvian Altamiranoa

125

When Cotyledon stricta is transferred to Sedum
neither the name Sedum strictum nor Sedum dielsii
can be used, because of the earlier homonyms Se¬
dum strictum K. Koch (1847) and Sedum dielsii Ha-
met (1913). The new name refers to the costately
plicate petals mentioned by Diels in the original
description of Cotyledon stricta.

Sedum reniforme (H. Jacobsen) Thiede & 't
Hart, comb. nov. Basionym: Villadia renifor-
mis H. Jacobsen, Natl. Caet. Succ. J. 13: 76.
1958 (nom. nov. pro Cotyledon imbricata
Diels due to Villadia imbricata Rose, 1903).
Cotyledon imbricata Diels, Bot. Jahrb. Syst.
37: 411. 1906. Altamiranoa imbricata
(Diels) A. Berger, in Engler X Prantl, Nat.
Pflanzenfam. ed. 2, 18a: 470. 1930. Villadia
imbricata (Diels) Baehni & J. F. Maebride,
Candollea 7: 286. 1937. Nom. illeg. (Art.
53.1 Tokyo Code, non Villadia imbricata
Rose, Bull. New York Bot. Card. 3: 3. 1903).
TYPE: Peru. Dep. Cajamarca: pr. Hualgay-
oc, juxta praedim La Tahona in rupibus
2600 in s.m., flor. m. Maj. 1904, Weberbauer
4053 (holotype, B).

When Cotyledon imbricata is transferred to Se¬
dum the epithet of the second oldest name of this
taxon, the synonym Villadia reniformis H. Jacobsen,
must be used because of the earlier homonym Se¬
dum imbricatum (Edgeworth) Walpers (1848—
1849).

Sedum weberbaueri (Diels) Thiede & t Hart,
comb. nov. Basionym: Cotyledon weberbaueri
Diels, Bot. Jahrb. Syst. 37: 411. 1906. Alta¬
miranoa weberbaueri (Diels) A. Berger, in En¬
gler & Prantl, Nat. Pflanzenfam. ed. 2, 18a:

470. 1930. Villadia weberbaueri (Diels) Baehni
& J. F. Maebride, Candollea 7: 286. 1937.
TYPE: Peru. Dep. Amazonas: Prov. Chacha-
poyas ad altera orientalia vallis fluminis Mar-
anon supra Balsas in graminosis siccis 2300
m s.m., Hor. m.. Jim. 1904, Weberbauer 4282
(holotype, B).

Acknowledgments. The authors are indebted to
Urs Eggli (Zurich, Switzerland) and Reid Moran (San¬
ta Rosa, U.S.A.) lor their comments, to Susan Carter-
Holmes (Kew, U.K.), Roy Vickery (London, U.K.), and
Susan Mamer (Oxford, U.K.) for checking specimens
at K, BM, and OXF, and to Paul Hiepko and Manfred
Bassler for providing access to specimens at B.

Literature Cited

Baehni, Ch. 1937. Vdladia et Altamiranoa. Etude sur la
fusion de deux genres de Crassulacees. Candollea 7:
283-286.

Berger. A. 1930. Crassulaceae. Pp. 352-483 in A. Engler
& K. Prantl (editors). Die natiirlichen Pflanzenfamilien
ed. 2, 18a. W. Engelmann, Leipzig.

Britton. IN. I.. & .1. N. Rose. 1903. New and noteworthy
North American Crassulaceae. Bull. New York Bot.
Card. 3: 1—45.

Brunner. I). R. 1993. Crassulaceae. Pp. 375—377 in E.
Brako X J. L. Zarucchi (editors). Catalogue of the f low¬
ering Plants and Gymnosperms of Peru. Monogr. Syst.
Bot. Missouri Bot. Card. 45: 1—1286.

Clausen. R. T. 1940. Studies in the Crassulaceae: Vdla¬
dia. Altamiranoa anil Thompsonella. Bull. Torrey Bot.
Club 67: 195-198.

Eroderstrom, II. 1936. The genus Sedum. A systematic
essay. Part IV. Acta Horti Gothob. 10: 2-262.

Hamet, R. 1913. Sur deux Sedum nouveaux de l'herbier
Royal <le Firenze. Malpighia 29: 57—63.

Koch, K. 1847. Beitrage zur Flora des nordliehen Kiisten-
landes von Kleinasien. Linnaea 19: 1—67.

Moran, R. 1996. Altamiranoa into Sedum. Haseltonia 4: 46.
Walpers. G. G. 1848—1849. Crassulaceae. Pp. 322—326 in
Annales Botanices Systematical1. I. Er. Hofmeister, Lip-
siaea | = Leipzig],

Volume 9, Number 1, pp. 1-126 of NOVON was published on 16 March 1999.

Volume 9
Number 2
1999

NOVON

Heterocentron evansii (Melastomataceae): A New Species from
Pico Bonito National Park, Honduras

Frank Almeda

Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco,
California, 94118-4599, U.S.A. falmeda@cas.calacademy.org

Abstract. Heterocentron evansii, known only
from the higher slopes of Pieo Bonito in northern
Honduras, is described, illustrated, and compared
with the species of subgenus Schizocentron to which
it is here assigned. It is unique among congeners
in having a combination of penninerved leaves,
densely pubescent foliage and cauline internodes,
glandular hairs on the prolonged connective of the
larger anthers, and in occupying rocky ridgetops
above tree line. This new species is the only Het¬
erocentron restricted to Honduras and one of three
species of Melastomataceae thought to be endemic
to the country.

Heterocentron, with about 28 species, is one of two
genera in the Melastomataceae with a distribution
restricted to Mexico and Central America (Almeda,
1993). Since it has been the subject of a recent re¬
vision (Whiffin, 1972), Heterocentron has generally
been regarded as one of the better known genera of
neotropical Melastomataceae. Species of this family
frequently colonize windswept ridgetops throughout
the Neotropics, so it comes as no surprise that ex¬
ploration of these sites continues to yield new and
unusual taxa. Such is the case with the new species
described below, which stands apart from all de¬
scribed species of Heterocentron by virtue of its co¬
piously hirsute cauline indument and densely pu¬
bescent foliage. This, coupled with its apparent
restriction to an exposed, rocky ridgetop on a pre¬
viously unexplored mountaintop in the largest na¬
tional park in Honduras, makes it especially notable.
Like Henriettella hondurensis Wurdack and Miconia
celaquensis Almeda, the other two species of Melas¬

tomataceae believed to be endemic to Honduras (Al¬
meda, 1996), the affinities of Heterocentron evansii
are with congeners restricted to montane areas of
Guatemala and southern Mexico.

Heterocentron evansii Almeda, sp. nov. TYPE:
Honduras. Atlantida: narrow crest of ridge
leading up to Pico Bonito from the NE (from
near confluence, at ca. 500 m elevation, of Rfo
Bonito and a large quebrada flowing from the
SW), 15°38'N 86°52'W, 1900 m, 23 Apr. 1996
(fl & fr), R. J. Evans 2553 (holotype, CAS;
isotypes, EAP, MEXU, MO). Figure 1.

Frutex humifusus, ramuli quadrangulati demum rotun-
dato-quadrangulati sicut petioli foliorum venae primariae
subtus hypanthiaque densiuscule setosi pilis plerumque
0.5—1 mm longis laevibus pro parte glanduliferis. Lamina
1.4— 2.9 X 0.9— 1.9 cm elliptica vel ovato-elliptica, pen-
ninervis, supra dense strigosa pilis laevibus 0.5—1 mm
longis persistentibus, subtus dense strigosa pilis laevibus
0.5—1 mm longis. Flores 4-meri in axillis foliosis super-
ioribus solitarii, pedioellis 1.6— 2.6 cm longis. Stamina di-
morphica; filamenta 7 mm vel 5 mm longa. Stamina
maiora: thecae 4.5-5 X 0.5—0.75 mm; conneetivum 2 mm
prolongatum ad basim glandulis 0.25—0.5 mm stipitatis
ca. 2 ornatum. Stamina minora: thecae 3 X 0.5 mm; con-
nectivum ca. 0.1 mm prolongatum. Ovarium 4-loculare:
semina 0.5 mm longa.

Low, procumbent, perennial subshrub with trailing
branches to 1 m long and upright leafy axillary
branchlets typically less than 8-9 cm in length. The
quadrate to rounded-quadrate cauline internodes
densely hirsute with spreading, smooth, brown hairs
0.5-1 mm long. Leaves of a pair essentially equal in
size; petioles 3—12 mm long, densely hirsute like the

Novon 9; 127-130. 1999.

128

Novon

figure 1. Heterocentron evansii Almeda. — A. Habil with enlargement of pubescence details at a node. — B. Repre¬
sentative leal, abaxial surface. — (.. Hypanthium and bibracteolate pedicel. — I). Hypanthium (top view) showing calyx
lobes and ciliate scales on ovary summit. — E. Petal, abaxial surface. — F. Antepetalous stamen (left) and antesepalous
stamen (right). — -G. Seeds. Scale: for A. bar = 1.5 cm; for B. bar = 4.5 mm; for C, bar = 5 mm; for I). bar = 2.25
mm; for E, bar = 2.4 mm; for E. bar = 2.5 mm; for G, bar = 0.4 mm. (A-G from the holotype.)

Volume 9, Number 2
1999

Almeda

Heterocentron evansii from Honduras

129

cauline intemodes; mature blades 1.4— 2.9 X 0.9— 1.9
cm, elliptic to elliptic-ovate, penninerved, abruptly
acute at the apex, acute at the base, the margin entire
to subentire, filiate, adaxial surface densely covered
with ± appressed, smooth hairs mostly 0.5—1 mm
long, the abaxial surface also copiously but not as
densely covered with antrorsely spreading, smooth
hairs 0.5—1 mm long on and between the elevated
veins. Inflorescence consisting of solitary flowers,
sometimes crowded and appearing terminal or pseu¬
doterminal on short ascending lateral branchlets. Ped¬
icel 1. 6-2.6 cm long, copiously covered with spread¬
ing glandular hairs mostly 0.25-0.75 mm long;
bracteoles at the base of each pedicel sessile, 3-6 X
2—3.5 mm, ovate to cordate, acute at the apex, the
margin entire; adaxiallv glabrous or with smooth ap¬
pressed hairs ± restricted to the margins, abaxial sur¬
face sparsely to moderately strigose with smooth ±
appressed hairs 0.5 mm long. Hypanthium (at anthe-
sis) 5—6 X 4—6 mm, typically flushed with red like
the calyx lobes, campanulate, copiously pubescent
with spreading mostly glandular hairs 1-2.5 mm long.
Calyx lobes 4, narrowly triangular, 7-8 X 2 mm,
acute at the apex with a terminal, typically gland-
tipped hair, glabrous adaxially and sparingly beset
abaxially with a few appressed, simple (sometimes
gland-tipped) hairs 0.25-0.5 mm long, the margin
glandular-ciliate (in part). Petals 4, reportedly violet,
7-10 X 5—8 mm, glandular-ciliolate. Stamens 8,
markedly dimorphic, differing in size and form, anther
thecae glabrous, linear-oblong, yellow (fide label data
on the type). Large (antesepalous) stamens: filaments
glabrous, 7 mm long, thecae 4.5—5 mm long, 0.5—
0.75 mm wide, ± horizontal to somewhat ascending;
connective prolonged 2 mm below the thecae, modi¬
fied ventrally at the filament insertion into a bifid ap¬
pendage 1.5-1.75 mm long, the connective beset dor-
sally with two (rarely one) spreading glandular hairs
positioned 0.25 mm above the filament insertion.
Small (antepetalous) stamens: filaments glabrous, 5
mm long, thecae 3 mm long and 0.5 mm wide, erect,
connective prolonged below the thecae ca. 0.10 mm
and modified ventrally into a bifid appendage ca.
0.25—0.5 mm long. Ovary elliptic-, glabrous but
crowned with four ciliate scales surrounding the stylar
scar. Style 8-10 mm long, declined to one side of the
flower opposing the larger stamens. Fruiting hypan¬
thium campanulate, 7-8 mm long to the torus and 5—
6 mm wide with a somewhat pustulate or muriculate-
tuberculate surface formed by the enlarged hair bases.
Seeds cochleate and tuberculate, 0.5 mm long.

Distribution. Known only from the type locality
on a dry, windswept ridgetop with thin, rocky soil
and exposed outcrops above primary lower montane

moist forest of Pinus, Liquidambar, and Clusia in
Pico Bonito National Park in northern Honduras.
This largest of Honduran parks, with 168,000
acres, protects an elevational range from near sea
level to 8000 ft. (Wallace, 1992).

In the most recent taxonomic treatment of Het¬
erocentron, Whiffin (1972) recognized two subgen¬
era. Heterocentron subg. Heterocentron , which in¬
cludes the majority of described species, is
distinguished by its erect suffrutescent to subshrub-
by habit, penninerved leaves, few- to many-flow¬
ered, paniculate inflorescences, and white or pink
(rarely purple) petals. For those species character¬
ized by a procumbent suffrutescent to shrubby hab¬
it, trinerved (rarely penninerved) leaves, solitary
flowers, and purple petals, Whiffin proposed the as
yet unpublished subgenus “ Schizocentron In this
latter grouping he included H. elegans (Schlechten-
dal) Kuntze, H. hirtellum (Cogniaux.) L. 0. Wil¬
liams, //. purpureum S. Winkler, and H. suffruti-
cosum Brandegee, all of which have patchy or
restricted distributions in southern Mexico and/or
northern Central America.

By virtue of its habit, inflorescence type, and
petal color, H. evansii is here assigned with cer¬
tainty to subgenus Schizocentron. It differs from oth¬
er species in this subgenus by a diagnostic suite of
characters that includes copiously hirsute cauline
internodes, a dense foliar indumentum of smooth
hairs, consistent presence of two (rarely one) glan¬
dular hairs on the prolonged connective of the larg¬
er anthers, and persistent pustulate or muriculate-
tuberculate hair bases on fruiting hypanthia.
Among the species enumerated above, //. evansii
appears to be most similar to //. suffruticosum of
Chiapas, Mexico, and adjacent Guatemala. These
are the only two species in subgenus Schizocentron
that share the penninerved leaves of subgenus Het¬
erocentron. In other features, Heterocentron suffru¬
ticosum differs markedly from H. evansii in its erect
or suberect habit, and in having appressed pubes¬
cent or subglabrous hypanthia and cauline inter¬
nodes, basally roughened hairs on hypanthia and
intemodes, modally larger leaves (15—50 X 12—45
nun), and glabrous anther connectives.

The distinctly trinerved leaves of the other three
species of subgenus Schizocentron readily set them
apart. Only H. elegans, which also occurs in Hondu¬
ras, has the habit and mostly glandular, spreading
hypanthial hairs characteristic of //. evansii. In //. ele¬
gans, however, the trailing branches commonly root
at the nodes, and the mature, ovate to oblong-ovate
leaf blades are sparsely to moderately appressed-stri-
gose on both surfaces. One other species, //. hondu-
rense Gleason, is also recorded from Honduras. The

130

Novon

following key is provided to facilitate identification of
the three species now known from that country.

Key to the Species ok Hetkrocentron in Honduras

la. Flowers solitary; petals purple or deep magenta;
southern Mexico, Guatemala, and Honduras.

2a. Principal leaves trinerved with all primary

nerves arising from a common point at the
base of the blade; distal internodes sparsely
to copiously covered with smooth, appressed
hairs; southern Mexico (Hidalgo, Puebla, Ve¬
racruz, and Chiapas), Guatemala, and Hon¬
duras .... H. elegans (Schlechtendal) kuntze
2b. Principal leaves penninerved with the pri¬
mary nerves diverging from the median
nerve in subopposite to irregularly alternate
fashion at successive points above the base
of the blade; distal internodes densely cov¬
ered with smooth, spreading hairs; Honduras
. H. evansii Almeda

lb. Flowers borne in few- to many-flowered dichasia;

petals white or pink; El Salvador, Honduras, and
Nicaragua . H. hondurense Gleason

I take pleasure in naming this species for Rand¬
all J. Evans, collector of this and other noteworthy
species of flowering plants in the course of his re¬
cent fieldwork in tropical America.

Acknowledgments. I thank Jenny Speckels for
preparing the line drawings, Gerrit Davidse for
tracking down duplicate collections at MO and hav¬
ing them dispatched to me in expeditious fashion,
and Randall J. Evans for providing additional in¬
formation on the type locality.

Literature Cited

Almeda. F. 1993. Slanmnrkia, a new genus of Melasto-
mataceae from the volcanic highlands of western Gua¬
temala and adjacent Mexico. Brittonia 45: 187-203.

- . 1996. A new Miconia (Melastomataeeae) from

Celaque National Park. Honduras. Novon 6: 319—322.
Wallace, I). R. 1992. Unfathomed f orests. Pacific Discov¬
ery 45: 30-35.

Whiffin, T. 1972. A Systematic Study of the Genus Het-
erocentron (Melastomataeeae). Ph.D. Dissertation, Uni¬
versity of Texas. Austin.

Merger of the Monotypic Genus F arringtonia with Siphanthera

(Melastomataceae)

Frank Almeda and Orbelia R. Robinson

Department of Botany, California Academy of Sciences, Golden Gate Park, San Francisco,

California 94118-4599, U.S.A.

Abstract. Phylogenetic studies based on mor¬
phological data show that F arringtonia, a monotyp¬
ic genus originally placed in the Microlicieae but
recently assigned to the Melastomeae, forms part of
a small clade nested within the genus Siphanthera.
Because these genera form a monophyletic group,
Farringtonia and Siphanthera are merged and the
new combination, Siphanthera fasciculata, is pro¬
posed.

The genus Farringtonia was established by
Gleason (1952) for a single Venezuelan collection
made by Julian A. Steyermark at the southeastern
base of Cerro Duida in the state of Amazonas. In
the protologue, Gleason suggested a close relation¬
ship between Farringtonia and Siphanthera Pohl
based on such shared characters as 4-merous flow¬
ers and rostrate anthers. The distinctive features
that, in Gleason's view, set Farringtonia apart from
other inelastome genera and species known to him
at the time included its stoutly subulate anthers,
large anther appendages, malpighiaceous cauline
hairs, and fascicled Empetrum-Uke leaves.

In the course of preparing a monograph of Si¬
phanthera we have conducted a comprehensive
character analysis of Siphanthera, Farringtonia,
and numerous possible sister genera in the Microli¬
cieae and Melastomeae, the two tribes to which Si¬
phanthera has been assigned (Gleason, 1952; Ren¬
ner, 1993). Our phylogenetic studies of these
genera using vegetative and reproductive charac¬
ters provide consistent and unequivocal evidence
that Farringtonia and Siphanthera form a mono¬
phyletic group, because F. fasciculata Gleason and
S. cordifolia (Bentham) Gleason are sister taxa nest¬
ed in a small clade within Siphanthera. Among the
characters Gleason used to emphasize the generic
distinctiveness of Farringtonia, only the apomorph-
ic Ernpetrum-\ike leaves appear to be noteworthy
in circumscribing the species, but none of these or
any other characters have diagnostic value in the

continued recognition of a monotypic genus. The
malpighiaceous cauline hairs of F. fasciculata are
very similar to those found in another, as yet un¬
described species of Siphanthera from Amazonian
Brazil. Thus, continued recognition of Farrington¬
ia, a monotypic genus nested within a clade of the
larger, diverse Siphanthera is inconsistent with our
attempt to provide a phylogenetic classification.

The single species of Farringtonia is here trans¬
ferred to Siphanthera, a monophyletic South Amer¬
ican genus defined by a 4-merous flower, 2-locular
ovary, and lacrimiform to vaguely reniform, areolate
seeds. A description, distributional data, and hab¬
itat notes are included in our treatment of Si¬
phanthera for the Flora of the Venezuelan Guayana
and will be amplified in our forthcoming monograph
(Almeda & Robinson, in prep.).

Siphanthera fasciculata (Gleason) Almeda & O.
Robinson, comb. nov. Basionym: Farringtonia
fasciculata Gleason, Fieldiana, Bot. 28: 426.
1952. TYPE: Venezuela. Amazonas: between
Esmeralda Savanna and the southeastern base
of Cerro Duida, 22 Aug. 1944, Steyermark
57837 (holotype, F; isotypes, GH, NY, US).

Acknowledgments. We thank Paul E. Berry and
Kay Yatskievych for their assistance in arranging,
selecting, and dispatching a loan of Siphanthera
from VEN via W IS and to the curators and staffs
of the following herbaria for making material under
their care available for study: A, BM, BR, C, CAS,
DS, F, G, GH, GOET, HB, INPA, K, LAM, LE, M,
MG, MO, NY, P, R, RB, S, SP, U, UB, UEC, US, W.

Literature Cited

Gleason, ft. A. 1952. Melastomaceae. Pp. 424—438 in J.

A. Steyermark, Botanical exploration in Venezuela — II.

Fieldiana, Bot. 28: 243-447.

Renner, S. S. 1993. Phytogeny and classification of the

Melastomataceae and Memecylaceae. Nordic J. Bot. 13:

519-540.

Novon 9: 131. 1999.

Twisselmannia (Brassicaceae), a Remarkable New Genus from

California

Ihsan A. Al-Shehbaz

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

ABSTRACT. Twisselmannia californica, a new ge¬
nus and species from Kings County, California, is
described and illustrated. Its relationships to other
mustard genera are discussed.

To the foreign reader, the discovery of a new ge¬
nus from the United States would seem rather un¬
usual, especially lor a family like the Brassicaceae
(Cruciferae), which has recently been treated for
North America in the outstanding monograph by
Rollins (1993). Although most of the North Amer¬
ican flora north of Mexico is well known, numerous
discoveries continue to be made, and it is estimated
that several hundred new taxa remain to be discov¬
ered and named from the United States alone (Ert-
ter, in press; Milius, 1999). Hartman and Nelson
(1998) indicated that 1197 new taxa of vascular
plants, including five genera based entirely on new¬
ly described species, were discovered in the United
States and Canada between 1975 and 1994. Of
these, 91 taxa belong to the Brassicaceae, and 217
(ca. 18%) were discovered from California alone.
Recent outstanding discoveries in the Brassicaceae
from California include the new genus and species
Sibaropsis hammittii S. Boyd & T. S. Ross (Boyd &
Ross, 1997) and Arabis hirshbergiae S. Boyd (Boyd,
1998).

The remarkable new discovery, hereafter known
as Twisselmannia californica, is based on a single
specimen collected in 1994 and was nearly passed
over by its collector as the common weed Capsella
bursa-pastoris (L.) Medikus (Andrew Sanders, pers.
comm.). It was sent initially to Reed C. Rollins who
for reasons of deteriorating health never had the
time to study it. The specimen, along with many
others left in Dr. Rollins’s office after he died, were
loaned to me in November 1998. Upon a casual
look at the specimen, it was immediately evident
that it does not belong to any known genus in the
Brassicaceae. Critical further study, as well as a
thorough checking of Schulz (1936) and all the ma¬
jor floras of the world, supported the fact that the
plant is a native new species that belongs to an
undescribed monotypic genus. The saltbush scrub
habitat in which T. californica grows belongs to a

No VON 9: 132-135. 1999.

private property of ca. 1700-acre parcel land in
Kings County that is up for development into an
“instant city," an action that would have resulted
in the total destruction of the habitat (Andrew
Sanders, pers. comm.) and perhaps the extinction
of the species. It was planned that I and several
colleagues would look for T. californica in late Feb¬
ruary 1999 before I formally described it. However,
due to the drought this year in that part of Califor¬
nia and because of the immediate need to protect
the habitat in which it grows, it became necessary
to publish a formal account without delay.

The genus, anecdotally referred to as the Kings
Gold, is named in honor of Ernest C. Twisselmann
(1917-1972), rancher and tireless botanical ex¬
plorer of the arid lands of south-central California,
whose name is synonymous with floristic works on
the Inner South Coast Ranges. His initial attempt
to compile a simple list of the plants on the Twis¬
selmann family ranches culminated in “A flora of
the Temblor Range and the neighboring part of the
San Joaquin Valley,” all potential habitat for addi¬
tional populations of Twisselmannia. He subse¬
quently authored “A flora of Kern County, Califor¬
nia.” His floras (Twisselmann, 1956, 1967)
continue to inspire several of the most enthusiastic
floristicians in California.

Although the opportunity to honor Ernest Twis¬
selmann with his own genus could not be passed
by, credit must also be given to the discover, Ed
LaRue, whose keen-sighted recognition of this
anomalous mustard resulted in the fortuitous col¬
lection of the sole specimen currently known.

Twisselmannia Al-Shehbaz, gen. nov. TYPE: Twis¬
selmannia californica Al-Shehbaz.

Herba annua; folia caulina pinnatisecta; racemi ad ap-
ieem bracteati, valde elongati; sepala oblonga, nonsaccata;
petala lutea; stamina 6. tetradynama; fructus obdeltoidei,
valde compressi, puberuli, valvis in dimidio inferiore car-
inatis, laevibus et tenuiter coriaceis, in dimidio superiore
rotundatis, tuberculato-rugosis et crasse coriaceis vel sub-
lignosis; septum valde angustatum; semina 4-8. oblonga,
nonmucilaginosa; cotyledones incumbentes.

Herbs annual. Trichomes unicellular, eglandular,
simple and minutely forked. Stems erect, few

Volume 9, Number 2
1999

Al-Shehbaz

Twisselmannia from California

133

branched basally. Basal leaves not seen; cauline
leaves short petiolate, not aurieulate, pinnatisect.
Inflorescences many-flowered, somewhat dense ra¬
cemes, bracteate to their apices, elongated consid¬
erably in fruit; rachis straight; bracts similar to cau¬
line leaves, progressively smaller upward. Fruiting
pedicels rather slender, terete, ascending-divari¬
cate. Sepals oblong, deciduous, erect, glabrous,
base of inner pair not saccate, margins narrowly
membranous. Petals yellow, slightly longer than se¬
pals, oblanceolate-obovate, apex rounded; claw ob¬
scurely distinct from blade. Stamens 6, tetradyna-
mous; filaments filiform, dilated at base; anthers
ovate, obtuse at apex. Nectar glands confluent, nar¬
rowly subtending bases of all stamens. Ovules 4—8
per ovary. Fruit dehiscent silicles, obtriangular,
strongly flattened and angustiseptate, sessile;
valves veinless, puberulent with simple and forked
hairs, keeled, thin leathery and smooth on proximal
half, rounded, thick leathery or subwoody, and tu-
berculate-rugose distally, wingless; gynophore ab¬
sent; septum complete, very narrow; style slender,
persistent, glabrous; stigma capitate, entire. Seeds
2—4 per locule, uniseriate, wingless, oblong, plump;
seed coat minutely reticulate, not mucilaginous
when wetted; cotyledons incumbent.

Twisselmannia californiea Al-Shehbaz, sp. nov.
TYPE: U.S.A. California: Kings County, S of
Kettleman City, E side of Interstate Flwy. 5, 3
mi. N of the Kern County line, grazed valley
saltbush scrub with Atriplex confertifolia ,
Frankenia salina, Astragalus oxyphysus, Bro-
mus hordeaceus, Deschampsia danthonioides,
and Hordeum depressum, 30 Mar. 1994, Ed
LaRue s.n. (holotype, UCR). Figure 1.

Herba annua; caules 7—25 cm longi, pilosi, trichoma-
tibus ad usque 1.5 mm longis; folia caulina et bracteae
petiolatae, pilosae, pinnatisectae, lobis lateralibus oblon-
gis vel oblongi-linearibus, 2—8 X 0.5—2 mm, integris; se-
pala oblonga, 1.2— 1.5 X 0.7— 0.9 mm; petala lutea, oblan-
ceolato-obovata, 1.6— 2.1 X 0. 7-0.9 mm; pedicelli
fructiferi tenues, pilosi, inferiores ca. 28 mm longi, su-
periores ca. 3 mm longi; fructus 4—5 X 4—5 mm, puberuli,
trichomatibus 0.1—0.25 mm longis; replum anguste line-
are, 4—5 X 0.3— 0.4 mm; stylum 0.3— 0.9 mm longum; sem-
ina oblonga 1.2— 1.5 X 0.6— 0.9 mm.

Herbs annual. Stems 7—25 cm tall, slender,
erect, terete, pilose throughout with simple tri-
chomes to 1.5 mm long. Basal leaves not seen; cau¬
line leaves petiolate, pilose as on stem; petiole ca.
1 cm long, shorter upward; leaf blade 2.5— 4.5 cm
long, pinnatisect; lateral lobes 2^1 on each side,
oblong to oblong-linear, 2—8 X 0.5—2 mm, entire;
terminal lobe 1.5— 2 X larger than lateral ones.
Fruiting racemes bracteate throughout, elongated

considerably; bracts similar to cauline leaves, grad¬
ually reduced in size upward. Sepals oblong, 1.2—
1.5 X 0.7— 0.9 mm, glabrous, narrowlv membranous
at margin. Petals yellow, oblanceolate-obovate, 1.6—
2.1 X 0.7— 0.9 mm, cuneate into a short claw. Fil¬
aments white, median pairs 1.2— 1.4 mm long, lat¬
eral pair 0.9—1 mm long; anthers ovate, 0.3— 0.4 mm
long. Fruiting pedicels divaricate-ascending, slen¬
der, pilose, lowermost to 28 mm long, uppermost to
3 mm long. Fruit a silicle, 4—5 X 4-5 mm, strongly
angustiseptate, obdeltoid, veinless, puberulent with
simple and minutely 1 -forked trichomes 0.1-0.25
mm long; valves strongly keeled, smooth, and thin
leathery on lower half, rounded, tuberculate-rugose,
and thick leathery to subwoody on the outer part of
distal half, minutely ciliate near margin along line
of attachment to replum; replum narrowly linear, 4-
5 X 0.3-0.4 mm; septum to 0.1 mm wide; style
slender, 0.3— 0.9 mm long. Seeds 2^4 and in upper
half of each locule, oblong, brown, 1.2-1. 5 X 0.6-
0.9 mm.

Although the fruit shape of Twisselmannia some¬
what resembles that of Capsella Medikus, the two
genera are quite different in many characters. Cap¬
sella has stellate trichomes, ebracteate inflores¬
cences, aurieulate cauline leaves, smooth, distinct¬
ly veined, and thin fruit valves, and up to 40 seeds
per fruit. By contrast, Twisselmannia has simple
and minutely 1-forked trichomes, bracteate inflo¬
rescences, nonauriculate cauline leaves, tubercu¬
late-rugose, veinless, and thick fruit valves, and up
to 8 seeds per fruit.

Because of its pinnatisect leaves and thick, tu¬
berculate-rugose, angustiseptate fruit valves, Twis¬
selmannia bears some resemblance to Coronopus
Zinn. Coronopus differs in having ebracteate inflo¬
rescences, 2-seeded fruits, completely closed
valves, rudimentary septum, and exclusively simple
trichomes, whereas Twisselmannia has bracteate in¬
florescences, 4— 8-seeded fruits, open valves, dis¬
tinct (though narrow) septum, and simple and
forked trichomes.

Both Capsella and Coronopus are represented in
North America by introduced weeds, and these
genera are definitely unrelated to Twisselmannia.
None of the New World Brassicaceae genera ap¬
pears to be closely related to Twisselmannia, and
the remarkable fruit morphology (with valves thin
leathery, keeled, and smooth on the lower half and
thick leathery or subwoody, rounded, and tuber¬
culate-rugose on the upper) clearly sets it apart
from all genera of the family. Twisselmannia is per¬
haps related to Mancoa Weddell, a genus of about
ten species distributed in Texas, Mexico, and dis-

134

Novon

figure I. Twisselmannia californica Al-Shehbaz. — A. Plant . — B. Trichomes: long ones of leaves, stems, and pedicels;
short ones of fruit valves. — C. Sepal. — I). Petal. — E. Stamen. — F. Fruit. — G. Fruit after removal of valves and
seeds. Seale: A = 1 cm; B-G = 1 mm. Drawn by Al-Shehbaz from the holotype.

Volume 9, Number 2
1999

Al-Shehbaz

Twisselmannia from California

135

junctly into South America (Argentina, Bolivia,
Peru). However, the differences between these two
genera are substantial. Mnncoa has oblong fruits,
papery, smooth, rounded, readily dehiscent, and
widely opened valves, biseriate seeds (15— )40-100
per fruit, filiform and long funicles, and broad re¬
plum to 2.5 mm wide. Twisselmannia has obtrian-
gular fruits, leathery, distally tuberculate-rugose,
basally keeled, tardily dehiscent, and nearly closed
valves, uniseriate seeds 4—8 per fruit, stout and
short funicles, and narrow replum to 0.4 mm wide.

Other genera with angustiseptate fruits, which
are widely distributed in California, the adjacent
southwestern U.S., and Mexico, include Dithyrea
Harvey and Dimorphocarpa Rollins, but these are
readily distinguished from Twisselmannia by their
didymous fruits, 1 -seeded valves, ebracteate inflo¬
rescences, decurrent stigmas, dendritic trichomes,
obsolete septum, and undivided leaves (for com¬
parison, see Rollins, 1979).

Acknowledgments. I am grateful to Henk van
der Werff for correcting the Latin. I thank Peter H.
Raven, Andrew Sanders, Emily Wood, and Kathryn
Rollins for help and support. I am much indebted
to Barbara Ertter for suggesting the generic name
and for providing information on E. C. Twissel-
mann.

Literature Cited

Boyd, S. 1998. Arabis hirshbergiae (Brassicaceae), a nar¬
row endemic from the Cuyamaca Mountains, San Diego
County, California. Aliso 17: 203—205.

- & T. S. Boss. 1997. Sibaropsis (Brassicaceae), a

new monotypic genus from southern California. Madro¬
no 44: 29—47.

Ertter, B. In press. Floristic surprises in North America
north of Mexico. Ann. Missouri Bot. Card.

Hartman, R. L. & B. E. Nelson. 1998. Taxonomic Nov¬
elties from North America North of Mexico: A 20-year
Vascular Plant Diversity Baseline. Monogr. Syst. Bot.
Missouri Bot. Card. 67.

Milius, S. 1999. Unknown plants under our noses: How
much backyard botany remains to be discovered? Sci.
News 155(1): 8-10.

Rollins, R. C. 1979. Dithyrea and a related genus (Cru-
ciferae). Publ. Bussey Inst. Harvard Univ. 1979: 3-32.
- . 1993. The Cruciferae of Continental North Amer¬
ica. Stanford Univ. Press, Stanford.

Schulz, 0. E. 1936. Cruciferae. In: A. Engler & K. Prantl
(editors), Nat. Pflanzenfam. ed. 2. 17B: 227—658.
Twisselmann, E. C. 1956. A flora of the Temblor Range
and the neighboring part of the San Joaquin Valley.
Wasmann .1. Biol. 14: 161—300.

- . 1967. A flora of Kern County, California. Was¬
mann J. Biol. 25: 1—395. [Reprinted by the California
Native Plant Society, 1995.]

Note added in proof. Dean W'm. Taylor and Barbara
Ertter are thanked for making a special effort to re-collect
Twisselmannia californica. Their search for the plant in
the type locality and five other sites on 25 March 1999
resulted in the discovery of several small plants in an area
of perhaps two square meters in association with Atriplex
spinifera, not A. confertifolia, as stated in the collection
data of the holotype. Of these, only a few plants were
collected, and they represent the following paratypes:

CALIFORNIA. San Joaquin Valley, Kings County, ca.
13 miles S of Kettleman City, along pipeline access road
lying 0.3 mi E of Interstate 5, at a point along the freeway
3.1 road mi. N of Kern County line, 217 ft. alt.,
T24S R20E SE1/4 NE1/4 Section 22. 35°49'41"N,
1 19°48'16''W, West Camp 7 1/2' USGS quadrangle, sub-
alkaline, sandy clay in Atriplex spinifera scrub, flowers
yellow, 25 Mar. 1999, Dean Win. Taylor 17098 & Barbara
Ertter (MO, UC).

Adenocalymma ubatubensis Assis & Semir, a New Species of
Bignoniaceae from Ubatuba, Sao Paulo State, Brazil

Marco Antonio Assis

Depto. Botanica-I B/UNESP, C. P. 199, Rio Claro-13.506-900, SP, Brazil

Joao Semir

Depto. Botanica-IB/UNICAMP, C. P 6109, Campinas-13.083-970, SP, Brazil

ABSTRACT. Adenocalymma ubatubensis, a new
species from southeastern coastal Brazil, is de¬
scribed and illustrated. The species is distin¬
guished by its large pseudostipules and calyx with¬
out nectaries.

Adenocalymma is the second largest genus of the
tribe Bignonieae (Bignoniaceae), with ca. 50 spe¬
cies in tropical America. Most species have a re¬
stricted distribution, and in a recent study (Gentry,
1993) six new species were described from eastern
South America.

A study of recent collections of Adenocalymma
for the “Flora Fanerogamica do Estado de Sao Pau¬
lo" project revealed a distinct new species related
to A. hatschbachii A. H. Gentry and A. ackermanii
Bureau & K. Schumann. The new species and Ad¬
enocalymma hatschbachii occur in the Atlantic
rainforest of southeastern coastal Brazil, while A.
ackermanii is known oidy from the type collected
in Minas Gerais state.

Adenocalymma ubatubensis Assis & Semir, sp.
nov. TYPE: Brazil. Sao Paulo: Ubatuba, Picin-
guaba, 9 May 1990 (fl), R. Romero et al. 74
(holotype, HRCB: isotypes, MO, SP, UEC).
Figure 1.

Scandens, ramis junioris dense tomentosis, glabrescen-
tis. pseudostipulae magnae foliaceae. Folia bifoliolata, in-
terdum cirrho simpliei, foliola ovato-oblonga vel subellip-
tica. apice acuto vel acuminato, base rotundata,
subeordata, supra subvillosa tandem glabrata, subtus
densa subvillosa. pilosa, pi 1 is, simplicibus vel ramosis. ln-
florescentiae axillares racemosae, braeteis et bracteolis to¬
mentosis et eglandulosis. Flores calyce campanulatos ex-
tus tomentosos, eglandulosos, corolla campanulata, lutea,
basi glabra, supra dense tomentosa, stamina inclusa. Cap-
sula ignota.

Lianas, branchlets densely tomentose when
young, soon glabrate when older, without interpe-
tiolar nectaries (glandular fields), lenticellate; the
two basal leaflets of each leaf forming pseudosti¬
pules, 1.1— 3.5 X 0.4— 1.4 cm, persistent in inflo¬

rescence base, elliptic, obovate-elliptic to spatu-
late, acute to slightly rounded, mucronulate, base
attenuate, enlarged, both faces tomentose, with ve¬
nation subparallelodromous to mixed craspedod-
romous, 5-7 main nerves raised below, few incon¬
spicuous nectaries. Leaves 2-foliolate, the terminal
leaflet replaced by a simple tendril; petioles (1.2—
2.6 cm) and petiolules (0.6-4 cm) tomentose to
subvillose; blades 5.5—17.5 X 2-6.7 cm, ovate-ob¬
long to subelliptic, apex acute to shortly acuminate,
base rounded to subcordate, margin slightly revo¬
lute, intramarginal vein inconspicuous, charta-
ceous, tomentose, subglabrate, scabridulous and
subpilose above, muriculate on nerves, venation
gradually raised, densely tomentose to subvillose
below, simple or forked slightly uncinate trichomes,
few punctate, nerves raised. Inflorescence ca. 10
cm, axillary, racemiform to subcorymbiform; bracts
and bracteoles 7—10 X 2—1 mm, tomentose, without
nectaries. Flowers yellow, calyx 1.3— 1.5 X 0.7 cm,
margin 5-denticulate, nerves raised above, thick,
tomentose outside, without nectaries, papillose in¬
side; corolla 5—6 cm, infundibuliform, basal tube
(ca. 1.5 cm) glabrous, toward apex densely tomen¬
tose outside, lobes ca. 1.5 cm, subtomentose faces,
inside at the level of stamen insertion with glan¬
dular trichomes; stamens included, longest pair
1. 7-2.6 cm, shortest pair 0.5-1. 9 cm, anthers 4 mm
long, with thecae divergent; ovary oblong, subtetra-
gonal, glabrous or with scattered trichomes, simple
or forked, punctate, glandular; the ovules biseriate
in each locule, ca. 18 to series. Fruit not seen.

Distribution. Known only from the types col¬
lected in coastal plains (restinga) of northern Sao
Paulo state.

Adenocalymma ubatubensis has a calyx without
nectaries and long pseudostipules (1.1-3. 5 cm
long), densely tomentose and with conspicuous
nerves raised, more or less parallel, that differen¬
tiate this new species from all species circum¬
scribed in the genus. Two related species, Adeno-

Novon 9: 136-138. 1999.

Volume 9, Number 2
1999

Assis & Semir

Adenocalymma ubatubensis from Brazil

137

Figure I. Adenocalymma ubatubensis Assis & Semir. —A. Flowering branch. — B. Pseudostipule. — C. Bud showing
calyx and bracteoles without nectaries. — 1). Ovary and disc. — F. Ovary cross section. ( R . Romero el al. 74.)

calymma ackermanii and A. hatschbachii, also have mentose pseudostipules (Gentry, 1993). Moreover,
eglandular calyxes. Adenocalymma hatschbachii , A. hatschbachii has membranaceous and only pu-
which also occurs in the Atlantic rainforest, differs berulous, long-acuminate leaflets, with stamens
in having smaller (5 mm), subulate, and non-to- more or less exserted. The other species, Adeno-

138

Novon

calymma ackermanii, has a similar form of pseu¬
dostipules, hut these are more acute and puberu-
lous (Bureau & Schumann, 1896-1897). It also
differs in the smaller leaflets and in the fewer num¬
ber of ovules in the loeule (only 26).

Another species with conspicuous pseudostip¬
ules is Adenocalymma prancei A. H. Gentry. How¬
ever, the pseudostipules of that species are linear-
oblong and measure up to 1.3 cm long (Gentry,
1978). It also differs in having a glandular calyx
and in occurring in central and upper Amazonia.

Para type. BRAZIL. Sao Paulo: Ubatuba, Picingua-
ba, 6 May 1988 (fl), R. Costa et al. 14 (HRCB, UEC).

Acknowledgments. We thank Antonio Furlan,
Reinaldo Monteiro, and Angela B. Martins for com¬
ments on earlier drafts, and Lindolpho Capellari,
Jr., for preparing the line drawings. We also ac¬
knowledge financial assistance from grants from
Fundunesp and FAPESP.

Literature Cited

Bureau, E. & K. Schumann. 1896-1897. Bignoniaceae.
In: C. F. P. Martius. Flora Brasiliensis 8(2): 1 — 4-S2 .
Monachii.

Gentry. A. H. 1978. Studies in Bignoniaceae 31: New spe¬
cies and combinations from Amazonian Peru and Brazil.
Ann. Missouri Bot. Card. 65: 725-735.

- . 1993. Six new species of Adenocalymma (Bignoni¬
aceae) from eastern South America. Novon 3: 137-141.

Th ree New Species of Huberia (Melastomataceae) from Peru

Jose Fernando A. Baumgratz

Instituto de Pesquisas Jardim Botanico do Rio de Janeiro, Rua Pacheco Leao 915,
22460-030, Rio de Janeiro, RJ, Brazil

ABSTRACT. Three new Andean species of the ge¬
nus Huberia DC. are described and illustrated.
These species were found between 1500 and 3350
m in restricted areas of northwestern Peru. The
presence of staminodes in H. staminodia represents
a new record for this fforal structure for Huberia. A
key is also provided for the four species of Huberia
occurring in the Andean region: H. cogniauxii, H.
peruviana , //. staminodia , and //. weberbaueriana.

Huberia DC. is a neotropical genus of 16 species,
the majority of which occur in eastern Brazil. Only
one species, //. peruviana, was known to occur in
Andean Ecuador and Peru. During a recent taxo¬
nomic revision of this genus (Baumgratz, 1997),
species new to science were discovered, of which
three, //. cogniauxii, //. staminodia, and //. weber¬
baueriana, from the Peruvian Andes are described,
illustrated, and mapped (Fig. 1).

Huberia staininodia Baumgratz, sp. nov. TYPE:
Peru. Depto. Amazonas: Province of Bongara,
between rfo Utcubamba and Shipasbamba, 4
km from Campomiento Ingenio, altitude 1520
m, 2 Feb. 1964, P. C. Hutchinson & J. K.
Wright 3995 (holotype, NY; isotypes, F, K, US,
USM). f igure 2.

Frutices ca. 1 m alta, indumento glanduloso-furfuraceo
etiam gemmis vegetativis glanduloso-vilosis; ramis tetra-
gonis. Folia membranacea, anguste ovata vel elliptica, tri-
nervia, margine integra vel apice crenata; domatia petio-
laria tubulosa interdum leviter inflata. Flores saepe
4-meri, interdum 5-meri, bypanthio 8-anguloso, interdum
10-, calycis laeiniis aequi-inaequilongisve, anguste trian-
gularibus; petalis ellipticis vel angusto-obovatis, apice
acutis, staminibus fertilibus et staminodiis, loculis an-
therarum sinuatis. Huberia staminodia affinis II. peruvi-
anae, H. weberbauerianae et //. cogniauxii sed praesertim
staminodiis praesentibus differt.

Shrub ca. 1 m high; branches, leaves, inflores¬
cences, bracts, prophylls, pedicels, hypanthium,
and calyx glandular-furturaceous, the vegetative
shoots and their adjacent axillary regions also glan¬
dular-villous. Branches quadrangular, gray and sul-
cate, brownish when young. Petioles 0.5— 1.5 cm
long, the apex with one pair of tubular domatia,
sometimes slightly inflated. Leaf blades 2.5— 6.4 X

1.2— 2.4 cm, membranaceous, narrowly ovate to el¬
liptic, the base acute, the apex acute-acuminate,
the margin entire or basal 2/s entire and distal Vs
crenulate, 3-nerved with the lateral veins arising at
the base, the central secondary veins 10-13 pairs,
transverse to oblique-ascendant. Inflorescences ter¬
minal, sessile, cymoids (Troll, 1969), corymbose or
not, 2.5— 3.5 cm long, with 15—25 flowers, the
branches 2( — 4), as triads or sometimes umbelliform
cymes with 5 flowers, distal node a triad or um¬
belliform cyme with 9 or 13 flowers; bracts 2-21
X 0.5— 9.0 mm, foliaceous, petiolate to sessile,
membranaceous to thickened, ovate to elliptic or
linear, slightly concave, the apex acute to obtuse,
the margin entire or sometimes distal Vs crenulate;
prophylls 1-2, 1—3 X 0.2-0. 5 mm, thickened, ob¬
long to linear. Flowers 4(— 5)-merous; pedicel 6.0-
7.5 mm long, obtusely quadrangular; hypanthium

3. 1— 4.0 X 2.0— 2.5 mm, tubular, 8-angled, the apex
slightly constricted; calyx lobes equal or unequal
in size, (0. 5-0.8) 1.0-1. 5 X (0.6-)0.8— 1.1 mm, nar¬
rowly triangular, angled at abaxial surface, the base
laminar, rigidly membranaceous, the apex thick¬
ened, subterete, oblong, rounded; petals white, 10—
11 X 3. 0^1.3 mm, elliptic to obovate, the apex
acute; fertile stamens 4— 5(6— 7), unequal in size,
antisepalous and/or antipetalous, yellow, the fila¬
ments 5.8— 6.3 mm long, the anthers narrowly tri¬
angular, the base symmetric or asymmetrically
lobed, the thecae 4.5— 5.6 mm long, undulate, bi¬
locular to pseudo-bilocular to the apex, the tube
unilocular, 0.3— 0.4 mm long, the connective ap¬
pendage 1.5— 2.1 mm long, dorsal, filiform, straight
or zigzag-shaped, sometimes absent; staminodia (1—
2)3 — 4, antipetalous, subisomorphic, yellow, 7.5— 9.5
mm long, filiform, the apex subulate, sigmoid to
sinuous or sometimes zigzag-shaped; ovary %—%(—%)
free, 3. 8-4.0 X 2.0-2. 1 mm, 4-loeular, the ovules

1.1— 1.3 mm long; style 12.0—12.5 mm long, sub-
terete, the stigma punctiform. Fruit a ruptidium (a
capsular fruit type; Baumgratz, 1997), 18-19 X

5.2— 5.5 mm, the hypanthium and calyx persistent,
urceolate, the apex slightly 8-angled, the pedicel
9-10 mm long; seeds 300—350, 2. 2^4.0 X 0. 4-0.7
mm, winged, linear to narrowly elliptic or ovate;
embryo 0.4— 0.7 mm long.

Novon 9: 139-146. 1999.

140

Novon

staminodia (★), and H. weberbaueriana (I).

Distribution. Huberia staminodia is known only
from the type collection. It is endemic to north¬
western Peru, between the Utcubamba and Shipas-
bamba rivers, in the cloud forests of Cordillera
Central or Oriental, Prov. Bongard, at 1520 m. This
phytogeographic region is named Ceja de la Mon¬
tana, more precisely Ceja del Marandn, where the
presence of clouds is observed throughout the year
(Weberbauer, 1945). It is described by Young and
Le6n (1993) as a region of moist forest of the ori¬
ental slopes.

Huberia staminodia is unique in the genus be¬
cause of its staminodia. The tubular domatia in the
petioles and 5-merous flowers also distinguish this
species from the other Peruvian species. The inflo¬
rescences are very condensed and short, with 15,
19, 21, and 25 flowers. The most frequent androe-
cium pattern in the tetramerous flowers is five fer¬
tile stamens and three staminodia, but the propor¬
tion 4:4 and sometimes 6-7:1— 2 is also
encountered in the same specimen. In the pentam-
erous flowers the androeeium has seven fertile sta¬
mens and three staminodia. In both types of flowers
the staminodia are always antipetalous, alternating
with the fertile stamens. As observed in all species
of Huberia, the fruit is an interesting capsule type:
it was termed a ruptidium by Baumgratz (1997). It

is composed of the mature ovary and the urceolate
hypanthium and calyx. In the mature fruit, it is
necessary lor the hypanthium and calyx to open by
irregular dehiscence in order for the seeds to be
dispersed; the mature ovary has loculicidal and
septifragal dehiscence. Flowering and fruiting in
February.

Huberia weberbaueriana Baumgratz, sp. nov.
TYPE: Peru. Depto. Amazonas: Prov. Bongard,
below San Carlos, 1700 m, June 1915, A. We¬
berbauer 7145 (holotype, US; isotypes, F, GH).
Figure 3.

Plantae indumento glanduloso-furfuraceo, gemmis ve-
getativis, petiolis et supra junioribus foliorum indumento
glanduloso-vilosis omatis; ramis tetragonis. Folia rigido-
membranacea, anguste ovata vel elliptica, margine basi
integro apiceque serrulato, trinervia. Domatia laminaria
marsupiformia interdum junioribus foliis carenlia. Flores
4-meri. hypanthio 8-anguloso, calycis laciniis anguste
triangularibus, petalis oblongis vel ellipticis, stigma cap-
itatis. Huberia weberbaueriana affmis //. cogniauxii sed
praesertim marginalibus nervis secundariis in basi non
confluentibus cum lateralibus nervis principalibus, flori-
bus et fructibus longioribus, petalis apice acuto vel ob-
tuso-apiculato, antherarum thecis manifeste sinuatis dif¬
fer!.

Plants with branches, leaves, inflorescences,
bracts, prophylls, pedicels, hypanthium, and calyx
glandular-furfuraceous, the vegetative shoots and
their adjacent axillary regions, and the petioles and
adaxial surface of the foliar blade when young also
glandular-villous. Branches subquadrangular, stri¬
ate when young. Petioles 0.6-1. 1 cm long. Leaf
blades 2.6— 5.8 X 1.0-1. 9 cm, rigidly membrana¬
ceous, narrowly ovate to elliptic, the base acute or
sometimes obtuse, the apex acuminate, the basal
margin Vs entire and distal % conspicuously to
slightly serrulate, 3-nerved with the lateral veins
arising at the base, the marginal secondary veins
inconspicuous and confluent to lateral primary
veins at the base, the central secondary veins 12—
15 pairs, oblique-ascendant; abaxial surface with
marsupiform laminar domatia, sometimes absent in
young leaves, apiculate at border of the membrane.
Inflorescences terminal, sessile, corymbose cy-
moids (Troll, 1969), 2. 5-3.0 cm long, with 11-19
flowers, the branches 2—4, frequently as triads or
sometimes umbelliform cymes with 4 flowers, distal
node usually an umbelliform cyme with 5 flowers
or sometimes a triad; bracts membranaceous, foli-
aceous, 9—23 X 1.5—10.0 mm, petiolate, narrowly
elliptic, the apex acute, the margin entire or sparse¬
ly serrulate to the apex, or bracts thickened, 1.4-
5.0 X 0.2-0. 5 mm, sessile, oblong, slightly con¬
cave, the apex obtuse, the margin entire; prophylls

Volume 9, Number 2
1999

Baumgratz

New Huberia from Peru

141

Figure 2. Huberia staminodia Baumgratz. — a. Flowering branches. — b, c. Details of petiolar domatia. — d. Transverse
petiole sections at levels in c, illustrating the cavities of the domatia. — e. Flower. — f, g. Pedicel, hypanthium, and
the calyx with lour and five lobes, respectively. — h, i. Petals. — j, k. Fertile stamens. — 1. Transverse anther sections
at levels in j. illustrating loculi and septa. — m. Detail ol the connective appendage. — n. Staminodium. — o. Longi¬
tudinal ovary section showing the partial adnation to the hypanthium. — p, q. Fruits. — r. s. Seeds.

142

Novon

Figure 3. Huberia weberbaueriana Baumgratz. — a. Flowering branches. — b. Detail of the foliar indument. — e-e.
Details of the laminar domatia. — f. Transverse leaf base sections at levels in c. illustrating the cavities of the dornatia.
— g, h. Bracts: foliaceous and petiolate, and thickened and sessile, respectively. — i. Prophyll. — j. Flower. — k. Pedicel,
hypanthiurn, and calyx. — 1. Calyx lobe. — m, n. Petals. — o, p. Longer and shorter stamens, respectively. — q. Ovary.
— r. Style.

Volume 9, Number 2
1999

Baumgratz

New Huberia from Peru

143

1-2, 0.4— 0.6 X 0. 1-0.2 mm, linear. Flowers 4-mer-
ous; pedicel 9—11 mm long, quadrangular; hypan-
thium 5.5— 6.2 X 2.5— 3.0 mm, tubular, 8-angled,
the apex slightly constricted; calyx lobes 1.5— 1.9
X 0.5— 0.6 mm, narrowly triangular, angled at ab-
axial surface, the base laminar, rigidly membrana¬
ceous, the apex thickened, subterete, oblong,
rounded; petals white, 12—14 X 4.5— 5.8 mm, ob¬
long to elliptic, the apex acute to obtuse-apiculate;
stamens of two sizes, yellow, the antipetalous sta¬
mens longer than the antisepalous, the filaments of
the longer 6.0— 6.8 mm long, of the shorter 5.2— 6.1
mm long, the anthers narrowly triangular, the the¬
cae of the longer 6.5— 7.3 mm long, of the shorter
5.6-6.2 mm long, conspicuously undulate, bilocu¬
lar to unilocular to the apex, the tube unilocular,
0.3— 0.4 mm long, connective appendage 3.5— 4.4
mm and 4. 2^4. 4 mm long, dorsal, filiform, straight
or zigzag-shaped, the apex apiculate or sometimes
asymetrically uni- or bi-apiculate; ovary %— X, free,
4.0— 4.8 X 2.3-2. 7 mm, 4-locular, the ovules 0.6-
1.6 mm long; style 14—16 mm long, the stigma cap¬
itate. Mature fruit not seen; young fruit 17.0-17.5
X 2.8— 3.0 mm, ureeolate, 8-angled, the pedicel 9—
10 mm long, striate.

Distribution. Huberia weberbaueriana is known
only from the tvpe collection. It is endemic to
northwestern Peru, where it occurs near the city of
San Carlos at 1700 m and in the same phytogeo¬
graphic region as the other Peruvian species of
Huberia. Given the altitude and based on the study
of Young and Leon (1993), this new species may
occur on the rainforest slopes of Ceja de la Montana
or in cloud forests situated in moist forest at higher
elevations ol the oriental slopes.

Huberia weberbaueriana may be distinguished
from the other Andean species of Huberia by its
marsupiform laminar domatia, longer size of the
flowers, fruits and their pedicels, the stamens of two
sizes, the thecae conspicuously undulate, and the
stigma capitate. The long glandular hairs in the
vegetative shoots and their adjacent axillary regions
are early caducous and they may be obscured by a
viscous substance in dried specimens. Usually each
leaf has a pair of domatia, but some specimens have
only one and on young leaves the domatia may not
be observed. The inflorescences are condensed and
short, with 11, 15, 17, and 19 flowers. This new
species seems most closely related to H. cogniauxii,
but the latter has leaves with marginal secondary
veins confluent to the lateral primary veins above
the base, shorter flowers and flower parts, and the
thecae with its surface smooth to slightly undulate.
Flowering and with young fruits in June.

lluhcria cogniauxii Baumgratz, sp. nov. TYPE:
Peru. Depto. Amazonas: Prov. Bongara, Lagu¬
na Pomacocha, ridge NE of lake, 77°52'W,
5°50'S, elev. 3350-2450 m, 7 Feb. 1985, B.
Stein & C. Todzia 2116 (holotype, MO; iso-
types, CAS, BB). Figures 4, 5.

Arbores, indumento glanduloso-furfuraceo etiam gem-
mis vegetativis glanduloso-vilosis. ramis tetragonis vel
subtetragonis. Folia rigido-membranacea, anguste ellipti-
ca, margine basi integro apiceque serrulate), Irinervia.
Domatia laminaria marsupiformia. Flores 4-meri, hypan-
tliio leviter 8-anguloso, calyeis laeiniis anguste triangu-
laribus, petalis ellipticis. Huberia cogniauxii affinis H. we-
berbauerianae sed praesertim marginalibus nervis
seeundariis et lateralibus principalibus foliorum basi oon-
fluentibus, floribus et fructibus minoribus, petalis apiee
acuminatis, antherarum thecis planis vel leviter sinuatis
el stigma capitatis differt.

Small tree ca. 3 m high. Branches, leaves, inflo¬
rescences, bracts, prophylls, pedicels, hypanthium,
and calyx glandular-furfuraceous, the vegetative
shoots and their adjacent axillary regions also glan¬
dular-villous. Branches quadrangular to subquad-
rangular, striate when young. Petioles 0.6— 1.0 cm
long. Blades 3. 4-5.0 X 1. 1-1.8 cm, rigidly mem¬
branaceous, narrowly elliptic, the base acute, the
apex acuminate, the basal margin Vs entire and dis¬
tal % serrulate or sparsely serrulate, 3-nerved with
the lateral veins arising at the base, the marginal
secondary veins tenuous and confluent to lateral
primary veins above the base, the central secondary
veins 10—14 pairs, oblique-ascendant; abaxial sur¬
face with marsupiform laminar domatia, apiculate
at the border of the membrane. Inflorescences ter¬
minal, sessile, cymoids (Troll, 1969), corymbose or
not, 2. 5-3. 7 cm long, with 19—27 flowers, the
branches 4—6, frequently as triads or sometimes
sessile monads or triads or umbelliform cymes with
5 flowers, distal node a triad or umbelliform cyme
with 7 flowers; bracts membranaceous, foliaceous,
10-20 X 2.2— 4.8 mm, petiolate, narrowly elliptic,
the apex acuminate, the margin inconspicuous and
sparsely serrulate to the apex, or bracts thickened,
1.4— 1.5 X 0.3-0. 4 mm, sessile, oblong, slightly
concave, the apex rounded, the margin entire; pro-
phylls 1—2, 0.4— 0.8 X 0.1— 0.2 mm, oblong. Flowers
4-merous; pedicel 5.5— 6.0 mm long, terete; hypan¬
thium 3. 8-4. 5 X 1. 8-2.1 mm, tubular, slightly 8-
angled, the apex slightly constricted; calyx lobes
1.1— 1.3 X 0.3— 0.5 mm, narrowly triangular, angled
at abaxial surface, the base laminar, rigidly mem¬
branaceous, the apex subterete, oblong, thickened,
rounded; petals white, 9.2— 9.8 X 3.1— 3.5 mm, el¬
liptic, the apex acuminate; stamens of two sizes,
yellow, the antipetalous stamens longer than the an¬
tisepalous, the filaments of the longer 5. 7-6.0 mm

144

Novon

Figure 4. Huberia cogniauxii Baumgratz. — a. Flowering branches. — b. Adaxial surface of the leaf showing the
marginal secondary veins confluent to lateral primary veins above the base. — e. Abaxial surface of the leaf showing
the pair of domatia and the marginal secondary veins confluent to lateral primary veins above the base. — d. e. Bracts:
foliaceous and petiolate, and thickened and sessile, respectively. — f. Prophyll. — g. Flower bud. — h. Flower.

Volume 9, Number 2
1999

Baumgratz

New Huberia from Peru

145

Figure 5. Huberia cogniauxii Baumgratz. — a. Pedicel, hypanthium, and calyx. — b. Calyx lobe. — c. Petal. — d, e.
Longer and shorter stamens, respectively. — f, g. Anther: adaxial surface and transverse sections at levels in f, illus¬
trating loculi and septa, respectively. — h. Detail of the pore. — i. Ovary. — j. Longitudinal ovary section showing the
partial adnation to the hypanthium. — k. Fmit. — 1-n. Seeds.

3 mm

146

Novon

long, of tlio shorter 5. 0-5.2 mm long, the anthers
oblong-subulate, the thecae of the longer 5. 1-5.5
mm long, of the shorter 4.5-4. 7 mm long, smooth
to slightly undulate, pseudo-bilocular to unilocular
to the apex, the tube unilocular, 0.2— 0.3 mm long,
the connective appendage 1.3-1. 8 mm and 1.5— 2.4
mm long, dorsal, filiform, zigzag-shaped; ovary %— /-
free, 3.5— 3.8 X 1.7— 1.8 mm, 4-locular, the ovules
0.6-0.9 mm long; style 1 1-13 mm long, the stigma
capitate. Fruit a ruptidium (a capsular fruit type;
Baumgratz, 1997), 11-13 X 4. 5-5.0 mm, the hy-
panthium and calyx persistent, urceolate, incon¬
spicuously 4-angled, the pedicel 5.5— 6.5 mm long;
seeds 275-295, (1.4)2.4-4.5 X 0.3-0.6(0.7) mm,
winged, linear, rarely narrowly ovate, sometimes
sinuous; embryo (0.5)0. 6-0. 8 mm long.

Distribution. Huberia cogniauxii is known only
from the type collection. It is endemic to north¬
western Peru, occurring at 2450—3350 m in Bon-
garA Province. This region is characterized by cloud
forests with rocky places, according to the collec¬
tors B. Stein and C. Todzia. The new species is
situated in the moist forest of the oriental slopes of
Cordillera Central, as described in the study of
Young and Ledn (1993).

Huberia cogniauxii is distinguished by leaves
with marginal secondary veins confluent to the lat¬
eral primary veins above the base, marsupiform
laminar domatia, short length of the pedicels, pet¬
als, and lruits, short width of the calyx lobes, an-
droecium only with fertile stamens of two sizes, the¬
cae smooth to slightly undulate, stigma capitate,
and lesser number of seeds per fruit. Frequently,
as observed in the other species of this genus
(Baumgratz, 1997), all the indument of the dried
specimen is obscured by a viscous substance, in¬
clusively the glandular hairs of the vegetative
shoots and their adjacent axillary regions, that are
also early caducous. The inflorescences are very
condensed and short, with 19, 21, 23, and 27 flow¬
ers. Flowering and with fruits in February.

Key to the Andean Species ok Hvhkkia

la. Leaves with tubular petiolar domatia; flowers
generally 4-merous or sometimes 5-merous;
androecium with stamens and staminodia

. H. staminodia Baumgratz

lb. leaves with marsupiform, petiolar. or laminar
domatia, flat or like two inflated pockets on the
petioles; flowers always 4-merous; androecium
with all stamens fertile.

2a. Leaves with petiolar domatia forming two in¬
flated pockets at the apex; calyx lobes 2.1-
2.5 mm long; stigma punctiform; fruits 20-
26 X 5-7 mm, the pedicel 1 1—15 mm long

. H. peruviana Cogniaux

2b. Leaves with laminar domatia; calyx lobes

I. 1— 1.9 mm long; stigma capitate; fruits

II. 0—17.5 X 2.8—5 mm, the pedicel 2.3—

10.0 mm long.

3a. Leaf blades narrowly ovate to elliptic,
the marginal secondary veins confluent
to lateral primary veins at the base;
pedicel 9.0-11.0 mm long; hypanthium
5. 5-6. 2 mm long; calyx lobes 1.5— 1.9
mm long; petals 12.0-14.0 X 4.5— 5.8
mm; thecae 5.6— 7.3 mm long, conspic¬
uously undulate; style 14.0-16.0 mm

long . H. weberbaueriana Baumgratz

3b. Leaf blades narrowly elliptic, the mar¬
ginal secondary veins confluent to lat¬
eral primary veins above the base; ped¬
icel 5. 5-6.0 mm long; hypanthium 3.8—

4.5 mm long; calyx lobes 1. 1-1.3 mm
long; petals 9.2-9.8 X 3. 1-3.5 mm;
thecae 4.5— 5.5 mm long, smooth to
slightly undulate; style 11.0-13.0 mm
long . H. cogniauxii Baumgratz

Acknowledgments. This article is based on my
doctoral study at the University of Sao Paulo. My
special thanks to Jose Rubens Pirani for the ori¬
entation and suggestions, Jorge Fontella Pereira,
Maria do Carmo Mendes Marques, and Graziela
Maciel Barroso for revision of the Latin diagnoses,
Andre Maurfcio de Carvalho for review of the En¬
glish manuscript, Frank Almeda for a critical re¬
view and for sending an isotype of Huberia cog¬
niauxii as a gift to the RB Herbarium, Walter S.
Judd for helpful suggestions, the curators of the cit¬
ed herbaria for making available the collections,
the botanical illustrator Maria Helena Pinheiro for
the drawings, anil CNPq for the doctoral grant.

Literature Cited

Baumgratz, J. F. A. 1997. Revisao taxonomica do genero
Huberia DC. (Melastomataceae). Tese de Doutorado,
Universidade de Sao Paulo, Sao Paulo.

Troll, W. 1969. Die Infloreszenzen. Typologie und Stellung
in Aufbau des Vegetationskorpers, Vol. 2. Stuttgart,
Jena.

Weberbauer, A. 1945. El Mundo Vegetal de los Andes
Peruanos. Talleres Graficos de la Editorial Lumen,
Lima.

Young, K. R. & B. Le6n. 1993. Distribution and conser¬
vation of Peru’s montane forests: Interactions between
the biota and human society. Pp. 237—246 in L. S.
Hamilton. J. 0. Juvik & F. N. Scatena (editors). Tropical
Montane Cloud Forests, Proceedings of an International
Symposium at San Juan, Puerto Rico.

Stictocardia sivarajanii, a New Species of Convolvulaceae from

Kerala, India

5. D. Biju and P. Pushpangadan

Tropical Botanic Garden & Research Institute, Palode P.O., Thiruvananthapuram 695 562,

Kerala, India

Philip Mathew

Department of Botany, University of Calicut, Calicut University P.0. 673 635, Kerala, India

ABSTRACT. Stictocardia sivarajanii , a new species
of Convolvulaceae from Kerala, India, is described
and illustrated. A tabular comparison with its re¬
lated species Stictocardia tiliifolia is given.

Stictocardia is comprised of about 12 species, of
which some are pantropical and others appear to
be native to Africa (Verdcourt, 1963). Perhaps the
genus originated in Asia or Africa (Austin, 1980).
In India, the genus is represented bv two species,
Stictocardia tiliifolia (Desrousseaux) Hallier f. and
the new species described here. Plant exploration
trips conducted in 1983 and 1996 in Rajamala and
Kallar on the high ranges of Western Ghats in the
Idukki District of Kerala State, India, yielded a tax¬
on quite different from the species of Stictocardia
hitherto described (see Table 1). This is presented
here as a new species, and is named alter the late
V. V. Sivarajan in appreciation of his great contri¬
bution to Indian taxonomy; S. D. B. and P. M. also
gratefully acknowledge his valuable guidance in
their research.

Stictocardia sivarajanii Biju, Pushpangadan & P.
Mathew, sp. nov. TYPE: India. Kerala: Raja¬
mala shola edge, 28 Sep. 1996, Biju 40182
(holotype, K; isotypes, CALI, Mil, TBGT, US).
Figure 1.

Stictocardia sivarajanii sp. nov. affinis 5. tiliifoliae
(Desrousseaux) Hallier i. set! capsula depresse globosa
laevi, calycis lobis fructiferis capsulam baud involventi-
bus, pedicellis longioribus, loliis subtus purpureo-viola-
ceis facile distinguenda.

Perennial; stem woody at base, trailing or twin¬
ing, terete, herbaceous toward tip. pubescent; latex
colorless, gummy. Leaves simple, ovate to ovate-
deltoid, 4—20 X 3-17 cm, apically acuminate and
shortly mucronulate, basally cordate, sparsely pu¬
bescent above, young leal with dorsal side purplish
violet, pubescent or tomentose below, covered with
minute black glands; midrib and lateral veins
raised beneath, lateral veins 8—13 pairs; petiole up
to 10 cm long, pubescent, younger ones tomentose.
Flowers axillary, solitary or in 2— 3-flowered cymes;
peduncle up to 15 cm long, terete to angular, pu-

Table 1. Distinction between Stictocardia tiliifolia and S. sivarajanii.

Stictocardia tiliifolia Stictocardia sivarajanii

Leaf dorsal side purplish violet, apically acuminate and

Leaf dorsal side pale green, apically acute to short-acu¬
minate and not mucronulate

Peduncle short, up to 7 cm long

Sepals suborbicular to orbicular and scarious to shortly
ciliate along the margins

Capsule globose, 4-lobed, 2.4— .'4.2 cm long

Fruiting sepals engulfing the capsule, 2.5 X 4.2 cm.
smooth

Seed obscurely angled, obovoid, prominently hairy at hi-
lum, grayish brown

Usually found along sea shores

mucronulate

Peduncle long, up to 15 cm long

Sepals widely ovate to orbicular with nonciliate margins

Capsule depressed globose, entire. 1 — 1.8 cm long

Fruiting sepals not engulfing the capsule, 1.3— 1.6 X
1.4-2 cm, striate

Seed globose, sparsely pubescent at hilum. black

Usually found along stream banks in semi-evergreen for¬
est and Ghat roadsides

Novon 9: 147-149. 1999.

148

Novon

Figure I . Stictocardia sivarajanii Biju, Pushpangadan & P. Mathew. — A. Flowering and fruiting twig. — B. Bracts.
— C, D. Outer sepals. — E-G. Inner sepals. — H. Stamen. — I. Pistil. — J. Fruit. — K. Seed. Based on Biju 22013
(TBGT).

Volume 9, Number 2
1999

Biju et al.

Stictocardia sivarajanii

149

Descent. mostly longer than petiole; bracts 2, ovate-
elliptic, small, 5—8 X 2-3 mm, apically acute to
slightly acuminate, pubescent outside, glabrous
within; pedicels up to 2-4 cm long, pubescent, an¬
gled and enlarged in fruits; sepals 5, subequal, out¬
er 2 larger, 1-1.2 X 0.7— 0.9 cm, widely ovate to
orbicular, apically obtuse to emarginate, densely
black glandular on both sides, sparsely pubescent,
middle 2 medium-sized, 0.8-1 X 0.5-0. 6 cm,
broadly ovate-elliptic, apically emarginate and
shortly apiculate, glandular dotted, glabrous to pu-
berulent outside, innermost one small, 0.6— 0.8 X
0.3-0. 4 cm, suborbicular, apically emarginate and
shortly apiculate, glandular dotted, glabrous; corol¬
la reddish purple with a darker center, funnel-
shaped, tube up to 3.5 cm long, limb slightly 5-
lobed, 5—7 cm across, mid petaline bands glabrous
and with minute black glands; stamens inserted;
anthers up to 7 mm long, straight after dehiscence;
filaments attached up to 5 mm above the corolla
base, 2 long, up to 2.5 cm, 3 short, up to 2.1 cm
long, filiform, purplish red ciliolate at dilated base;
disc small, ± 1 mm long, annular, slightly lobed;
style subexserted, up to 3.5— 3.7 cm long, glabrous,
filiform, stigma biglobose, papillose; ovary conical,
1.2 X 2 mm, glabrous, 4-celled. Fndts capsular,
indehiscent, depressed globose, not 4-lobed, 1-1.8
X 0.8-1. 2 cm, glabrous, not engulfed bv the en¬
larged sepals, which are striate cartilaginous, 1.3—
1.6 X 1.4—2 cm; seeds 1-4, globose, 6—8 X 6—7
mm, covered with minute hairs, sparsely pubescent
at hilum, pale black; seed germination epigeal. hy-
pocotyl up to 6.5 cm long, glabrous, apically deeply
cordate, sinus up to 3.5 cm deep, basally truncate
to slightly cordate, petiole up to 1 cm long.

Phenology. Flowering Oct.— Dec. Flower open¬
ing 5 a.m.-6:30 a.m., closing after 5 pan. Fruiting.
Dec.— Feb.

Distribution. Stictocardia sivarajanii is known
from only two localities in Western Ghats, the Er-
avikulam National Park and Kallar near Munnar,
Kerala.

Ecology. The new species is found near stream
banks in the semi-evergreen forests and Ghat road¬
sides generally above 4000 ft.

Etymology. The species is named alter the late
V. V. Sivarajan.

Paratype. INDIA. Kerala: Mukki (list.. Kallar, ±
4000 ft.! 30 Oct. 1983. Hiju 220 Pi (TBGT).

Acknowledgments. We express our sincere
thanks to Bernard Verdcourt (K) lor his continuous
help during the study and for excellent comments
that improved the manuscript. Sincere thanks are
also due to Daniel F. Austin for his help in con¬
firming the novelty of the species, to the Director,
Kew, for a photocopy of the type, and to Suresh
Elamon for help during the collection of materials.
Financial support was received from Kerala State
Forests and Wildlife Department; we are grateful to
T. M. Manoharan, Chief Conservator of Forests
(Wildlife), for his cooperation and keen interest in
our studies.

Literature Cited

Austin, I). F. 1980. Convolvulaceae. Pp. 288—363 in M.
D. Dassanavake & F. R. Fosberg (editors), A Revised
Handbook of the Flora of Ceylon. Vol. I. Oxford & IBH.
New Delhi.

Verdcourt, B. 1963. Convolvulaceae. Pp. 1 — ltd in C. K.
Hubbard & E. Milne-Redhead (editors). Flora of Trop¬
ical East Africa. Balkema, Rotterdam.

Two New Distinctively Large-leaved Species of Tovomita
(Clusiaceae) from the Venezuelan and Peruvian Amazonian Region

Nidia L. Cuello A.

Herbario Universitario (PORT), Programa R.N.R., UNELLEZ-Guanare, Mesa de Cavaeas, Edo.

Portuguesa 3323, Venezuela

ABSTRACT. Two new large-leaved species of To¬
vomita are described: T. auriculata from Venezue¬
lan and Peruvian Amazon lowland forests, and T.
foldatsii from Venezuelan Amazon lowland forests.
The morphological relationships to related species
are discussed.

RESUMEN. Se describen dos especies nuevas de
Tovomita: T. auriculata de los bosques de tierras
bajas de la Amazonia de Peru y Venezuela y T.
foldatsii de los bosques de tierras bajas de la Ama¬
zonia venezolana. Se discuten sus relaeiones mor-
folbgieas con especies afines.

Tovomita Aublet (Clusiaceae), with about 45 spe¬
cies, is known exclusively from the New World
tropics. This genus, one of the most poorlv known
of the Clusiaceae, is represented in Venezuela by
at least 30 species. The two undescribed species
were detected during preparation of a treatment of
the Tovomita for the Flora of the Venezuelan Gua-
yana.

Tovomita auriculata Cuello, sp. nov. TYPE: Ven¬
ezuela. Amazonas: Dpto. Atures, bosque alto
de tierra firme en el sector “La Culebra” a 26
km al SE de Pto. Ayacucho, por la carretera
Pto. Ayacucho— El Gavilan, 5°31'N, 67°24'W.
80—150 m, ene. 1989 (pist. H, fr), N. Cuello &
Y. Fernandez 480 (holotype, PORT; isotype,
MO). Figure 1.

Haec species T. speciosae Ducke affinis sed foliis ses-
silibus, laminis attenuatis, auriculatis, obovatis, petiolis
minus quam 5 mm longis, inflorescentia 3-6 mm longa,
sepalis extends triangularibus, sepalis internis oblongis,
petalis 5. ovario 5-carpellato, atque fruotibus rotundatis
differt.

Dioecious trees, 3 to 10 m tall. Branches terete,
bark corky, ringed by leal scars. Leaves opposite,
decussate, short-pet iolate or subsessile, the petioles
less than 5 mm long; the blades coriaceous, largely
obovate, 9-16 X 30-45 cm, attenuate and aurie-
ulate at base, obtuse to rounded at apex, base 1-2
cm wide, lateral veins 12 to 20 pairs, well differ¬
entiated and spaced 1-2.2 cm apart, curving toward

the margin. Inflorescences cymose, terminal, len-
ticellate. Pistillate inflorescences 3-flowered, pe¬
duncle 1—1.2 cm long, pedicels of lateral flowers

3.4— 3.6 cm, pedicel of terminal flower 3.5-4 cm
long. Staminate inflorescences 13-39-flowered, 3-
6 cm long, pedicels of lateral flowers 5-10 mm
long; pedicel of terminal flower 13-16 mm long.
Staminate buds 5-9 X 4—6 mm, external sepals 2,
rugose and coriaceous, 6—7 X 8-9 mm, internal
sepals smooth and membranaceous. Petals 5, ob¬
long, 7—8 X 8—9 mm. Stamens numerous, ca. 60.
Pistillate buds oblong-elliptic, 10-15 X 6-8 mm;
pistillate flower 15-18 mm long at anthesis; exter¬
nal sepals 2, triangular, 13-15 mm long, 12-16 mm
wide at base; internal sepals 2, oblong, 14-16 X
8-10 mm. Petals ovate, 15-18 X 3. 5-4. 5 mm.
Staminodes 40—60, 8-10 mm long. Ovary 5-car-
pellate, 5 mm long, style 5 mm, stigmas capitate,

2.5- 3 mm long. Ovules 1 per carpel, 4-4.3 mm
long, axile. Fruits rounded, 3—3.5 cm diam., with
persistent stigmas, 5—9 mm long.

Distribution and ecology. Tovomita auriculata
is only known from the type and paratype collec¬
tions, and is apparently a common tree at the lower
and medium strata in wet lowland to foothill forests.
In Venezuela, it is known by the common name of
“palo de tinaja.”

Due to its large leaves, T. auriculata appears to
be most closely related to T. speciosa Ducke
(Ducke, 1925, 1935) from Brazilian Amazon low¬
land forests in the states of Amapa and Para. How¬
ever, T. auriculata has sessile to subsessile leaves,
the bases of the obovate blades are attenuate and
auriculate (petiole less than 5 mm long), the inflo¬
rescences are 3-6 cm long, with triangular and ob¬
long sepals, 5 ovate or oblong petals, ovary with 5
carpels, and rounded fruits, 5—9 mm long. In con¬
trast, T. speciosa has the bases of oblong-obovate or
elliptic-obovate blades cuneate-attenuate, the pet¬
ioles are 3-4 cm long, the inflorescences are 2—2.5
cm long, with wide orbicular sepals, 6 obovate-el-
liptic petals, ovary with 6 carpels, and pyriform
fruits, 2.5—3 cm long. In addition, T. auriculata is
also similar to T. excelsa Andrade-Lima & Mariz

Novon 9: 150-152. 1999.

Volume 9, Number 2
1999

Cuello A.

Tovomita (Clusiaceae)

151

Figure 2. Tovomita auriculala Cuello (stum. II. Sanoja <£'
Fernandez 3265, paratype, PORT).

Haec species T. fanshawei Maguire similis sed petioli
alis longioribus (6—9 mm vs. 4-5 mm), lamina foliari long-
iore (30—45 X 20—25 cm vs. 15—22 X 6—9 cm) atque
stigmate in fructu lineari (vs. subpeltato) differt.

Dioecious tree, 4—5 m tall. Branches quadran¬
gular, glabrous, with smooth bark. Leaves decus¬
sate, the petioles thick, semi-terete, striate, 1.5—2
cm long, 4—5 mm wide; axillary petiolar flaps con¬
spicuously raised, 6—9 mm diam., the blades cori¬
aceous, broadly oblong-elliptical, 3CM15 X 20—25
cm, shortly attenuate at base, obtuse to rounded at
apex, lateral nerves prominent, 17—21 pairs, 2—3.2
cm apart. Inflorescences and flowers unknown.
Fruit a fleshy stipitate capsule, ellipsoidal to fusi¬
form, 8—10 X 2.5—3 cm; with a style 1.2—2 cm long
and 5 linear stigmas; stipe 1—2 cm long; staminodes
usually persistent in fruit. Seeds reniform, 3.5^4 X
0.9-1. 2 cm, 1 per loeule, apparently with a fleshy
aril-like covering.

Distribution and ecology. Tovomita foldatsii is
found in lowland wet forests of the eastern Ama¬
zonas in Venezuela and is known only from the type
collection.

Tovomita foldatsii is a distinctive species be¬
longing to the large-leaved complex ( T. macrophylla
(Poeppig) Walpers, T. fanshawei Maguire, T. morii

Figure 1. Tovomita auriculata Cuello (pist. (I and fr, A.
Cuello & ). Fernandez 430. holotype. PORT).

(Mariz, 1974) from wet forests of the state of Ala-
goas, Brazil. However, T. auriculata has coriaceous
leaves, petioles less than 5 mm long, and oblong
petals, 8—9 mm wide in staminate flowers. Tovomita
excel-sa differs in its membranaceous leaves, 2.5-3-
cm-long petioles, and lanceolate petals, 3—5 mm
wide in staminate flowers.

Faratypes. PERU. Loreto: Prov. Maynas, Ouebrada
Santa Cruz (tributario del Rio Nanay). 03°50'S. 73°35,\X.
130 in. Mar. 1982 (tr). R. Ydsquez <£■ J. Ruiz 2945 (MO).
VENKZUELA. Amazonas: Dpto. A lures. Laguna El Si 1-
lon y Cano Mariguaca, 78 km NK Puerto Ayacueho,
05°49'N. 66°50'W, 400 m. Oct. 1989 (buds). Proveeto In-
ventario de los Recursos Naturales de la Guayana Vene-
zolana-CVG- l ECM IN, Fernandez 6109 (MO. PORT); Ser-
rania Batata, 2 km al INK de Salto Colorado. Cano
Colorado sobre corazas laterfticas, 55 km al 5E de Pto.
Ayacueho, 5°33'N, 67°08'W. 550 m. Oct. 1989 (slam. fl).
Sanoja 3240 (POR I’ unicate), 3265 | MO. PORT (Fig. 2)|,
3276 (PORT unicate); Rio Cataniapo. 56 km al SE de Pto.
Ayacueho, 5°38,N, 67°12'W. 120 m. ago. 1981 (slam. fl).
Gudnchez 1425 (TFAV unicate).

Tovomita foldatsii Cuello, sp. nov. TYPE: Vene¬
zuela. Amazonas; Dpto. Atures, bosque alto de
tierra firme en el sector “Campamento Remo,”
alto Rio Sipapo, 4°34'N, 67°18'W, 80-150 m,
junio 1989 (fr), E. Foldats & J. Veldzco 9577
(holotype, PORT; isotype. MO). Figure 3.

152

Novon

Figure 3. Tovomita foldalsii Cuello (fr, E. Foldats & ./•
Veldzco V577, hololype, PORT).

Maguire, T. speciosa Ducke, T. froessi Maguire, and
T. auriculata Cuello) from Panama, the Amazon Ba¬
sin, and Guiana Shield. Among those, T. foldatsii
is most closely related to T. fanshawei Maguire (Ma¬
guire, 1948) from Guyana, Colombia, Peru, Brazil,
and Bolivia (Cuello, 1998), both having raised ax¬
illary petiolar Haps. However, in T. foldatsii , the
blades (30-45 X 20—25 cm) and axillary petiolar
flaps (6-9 mm) are much larger than in T. fanshaw¬
ei (15-22 X 6—9 cm and 4—5 mm, respectively),
and the stigmas in the fruits ol T. foldatsii are linear
and not subpeltate as are those in T. fanshawei. In
addition, T. foldatsii is similar to T. morii Maguire

(Maguire, 1978), from Panama, having quadrangu¬
lar branches in common. However, T. foldatsii is a
tree, with petioles 1.5—2 cm long, a broadly oblong-
elliptical lamina, 30—45 X 20-25 cm, and stipitate
fruit, 8—10 X 2.5—3 cm. In contrast, T. morii is an
epiphytic shrub, with petioles less than 0.5 cm
long, broadly oblanceolate lamina, 18-25 X 10-14
cm, and non-stipitate fruit, 4—5 X ca. 2 cm.

This species is named in honor of Ernesto Fol¬
dats due to his significant contributions to our
knowledge of the Venezuelan flora, and for collect¬
ing the type specimen.

Acknowledgments. I thank the Missouri Botan¬
ical Garden and J. J. Pipoly III (BRIT) for making
their facilities available for research. I thank G.
Aymard (PORT) for his guidance and encourage¬
ment during the preparation of the manuscript. I
am grateful to P. Berry, B. Hammel, C. Vargas, and
K. Yatskievych (MO) for their valuable comments
and observations on the manuscript. I thank Jim
Solomon (MO) for his help with the photographs,
and C. Vargas (MO) and Roy Gereau (MO) for help
with the Latin diagnoses. I also thank the anony¬
mous reviewers for their assistance with the man¬
uscript.

Literature Cited

Cuello, N. 1998. Tovomita (Clusiaceae). Pp. 313—322 in
P. E. Berry, B. k. Moist & k. Yatskievych (volume ed¬
itors), Flora of Venezuelan Guayana, Vol. 4 (Caesalpi-
niaceae to Ericaceae). Missouri Botanical Garden Press,
St. Louis.

Ducke, A. 1925. Plantes nouvelles ou peu connues de la
region amazonienne (III partie). Arq. Jard. Bot. Rio de
Janeiro 4: 134—140.

- . 1935. Plantes nouvelles ou peu connues de la

region amazonienne (VIII partie). Arch. Inst. Biol. Ve-
get. Rio de Janeiro 2: 27—64.

Maguire. B. 1948. Plant explorations in Guiana in 1944.
chiefly to the Tafelberg and Kaieteur Plateau IV. Bull.
Torrey Bot. Club. 75: 374—438.

- . 1978. Notes on the Clusiaceae, chiefly of Panama

III. Phytologia 39: 65—77.

Mariz, G. 1974. Two new species of Tovomita Aublet (Gut-
tiferae) from northeastern Brazil. Bull. Torrey Bot. Club
101: 367-371.

The Identity of Grimmia involucrata Cardot (Grimmiaceae, Musei)

Claudio Delgadillo M.

Departamento de Botanica, Instituto de Biologfa, UNAM, Apartado Postal 70-233,

Mexico, D.F. Mexico

ABSTRACT. Grimmia involucrata is a Mexican en¬
demic that has been confused with G. affinis, G.
americana, and G. tergestina. It is recognized here
as a valid, morphologically distinct species char¬
acterized by clavate stems, a gonioautoicous con¬
dition, differentiated perichaetial leaves, immersed
sporophytes, and the smooth basal thin! of the outer
plate of the peristome teeth. It is illustrated herein.

Grimmia involucrata was described by Cardot
(1909) on the basis of a specimen collected in the
state of Hidalgo, Mexico. Reference to this species
was made by Theriot (1931), who listed specimens
collected by Amable in Mexico City; Sayre (1952)
and Whitehouse and McAllister (1954) treated it as
conspecific with G. americana Bartram, a species
described from Texas (Bartram, 1929). Crum and
Anderson (1981: 424) disagreed and concluded
that G. involucrata was not “even a member of the
subgenus Gastero grimmia Schimp.” In recent
years, Crum (1994) treated it as a synonym of G.
affinis Hornschuch; he also studied a mixed spec¬
imen from the type locality of G. involucrata that
he referred to G. tergestina Tommasini, an Old
World taxon.

Except for occasional reference in floristic lists
or indices, information on G. involucrata is scarcely
available. Descriptions other than the original have
been lacking, as have illustrations. In addition, the
type specimen is a mixed collection, which has
contributed to confusion regarding the taxonomic
identity and nomenclature of G. involucrata and
other Mexican Grimmias. This paper attempts to
clarify its status.

Grimmia involucrata Cardot, Rev. Bryol. 36:
105. 1909. TYPE: Mexico. Hidalgo: Cuyama-
loya, on rocks, 17 Sep. 1908, Pringle 10598
p.p. (holotype, PC not seen; isotypes, FH, H-
BR). Figure 1.

Stem clavate, dark green or black, up to 9 mm
high, with a central strand; in cross section with an
epidermal layer of 1-2 layers of smaller thick-
walled cells. Leaves concave in proximal and mid¬
leaf sections, carinate at extreme apex, imbricate

when dry, erect-spreading when moist, oblong-
ovate, 0.9— 1.3 mm in length; apex obtuse; distal
leaf margins bistratose, plane, entire. Distal leaf
cells quadrate to short rectangular, in 2(— 3) layers
of thick-walled, slightly sinuose cells, 5—12 pun
diam.; marginal leaf cells at base quadrate to short
rectangular or transversely elongated, with thicker
cross walls, 7—17 pan diam.; other basal cells uni-
stratose, quadrate to rectangular, thick-walled,
slightly sinuose upward; basal juxtacostal cells not
forming a distinct area, mostly quadrate and thin-
walled. Costa excurrent, in section with one dorsal
substereid layer and 2—6 guide cells around or ven¬
tral to an internal hydroid group. Hair-point 0.6-
1.3 mm long, denticulate, slightly decurrent at
base. Gonioautoicous. Perigonial bracts convolute,
0.6 mm long. Perichaetial leaves strongly differ¬
entiated, the outer with a hyaline marginal area, the
inner mostly thin-walled and hyaline; oblong-ovate
to convolute, 1.4— 1.9 mm long; hair-point 0.9— 1.6
mm long. Seta straight, 0.5 mm long, not twisted.
Capsule symmetrical, ellipsoid, immersed, 1.2-1 .4
mm long; peristome teeth somewhat cribrose, del¬
toid, 287-300 X 62—75 pun, straw-colored to or¬
ange, outer plate nearly smooth in the lower third,
trabeculate, distally papillose; inner plate finely pa¬
pillose throughout; annulus revoluble; exothecial
cells prismatic, mostly longer than wide, thin-
walled, 25—60 pun diam.; stomata several, basal;
operculum conic, short-rostrate, 0.7 mm long; ca-
lyptra mitrate, smooth, 0.9—1 mm long. Spores 10-
15 pun diam., spore wall perforated.

Distribution. Endemic to Mexico (Distrito Fed¬
eral, Hidalgo, Mexico, Zacatecas).

Habitat. On rocks or soil-covered rocks, in Pi-
nus forest, dwarf oak forest, or desert scrub, 2180—
2590 m elevation.

The costa in Grimmia involucrata is reported
here as excurrent because there is morphological
continuity of the inner costal cells with those of the
hair-point. The latter consists, in addition, of hya¬
line laminal cells in this and other species of the
genus, but its ontogeny is still unknown.

As noted by Cardot (1909), G. involucrata is
similar to G. tergestina. Both taxa have clavate

Novon 9: 153-155. 1999.

154

Novon

Figure 1. Grimmia involucrata Cardot. — A. Vegetative leaves. — B. Leaf basal cells. — C. Leaf cross section in distal
half. — I). Perichaetial leaves. — E. Portion of peristome showing smooth outer plate and papillose distal end of broken
teeth. Drawings from the type specimen, except A {Cardenas 3018) and K ( Cardenas 3912). Scale = 1 mm (A. I)), =
0.1 mm (B. C, E).

stems due to large differentiated perichaetial
leaves that enclose an immersed sporophyte; their
vegetative leaves are bistratose in the distal half
bearing a decurrent denticulate hyaline hair; the
basal leaf cells are quadrate to short rectangular
and, in section, the costa has several guide cells
frequently surrounding a single hydroid group. Al¬
though similar in general morphology, the game-
tophore in G. involucrata is consistently gonicau-
toicous and the peristome teeth are papillose
except at the outer base and have conspicuous tra¬
beculae. In G. tergestina the gametophore is di-
oicous and its peristome teeth virtually lack basal
trabeculae and are papillose throughout. The tax¬
onomic significance of these differences remains
problematic; the peristome structure in G. terges¬
tina is comparable to that of G. involucrata so that
their only reliable morphological difference is the
sexual condition.

According to Churchill and Linares (1995),
Grimmia affinis is not present in the Neotropics and
may be a synonym of G. longirostris Hooker. In any
event, neither G. affinis nor G. longirostris is closely
related to G. involucrata. Their leaf cross sections
show an elliptical to reniform costa with a ventral
u-shaped sinus distally; at mid-leaf, there is a row
of 2-4 ventral guide cells overlying a central hy¬
droid group and a dorsal epidermal layer of sub-
stereid cells; the capsule is exserted in G. affinis.
Thus, the proposed synonymy between this and G.
involucrata (Crum, 1994) does not have a taxonom¬

ic basis, for they are clearly distinct. The confusion
apparently derives from plants of a second Grimmia
in the type specimen of G. involucrata. Cardot
(1909) stated that this contained G. ovata F. Weber
& D. Mohr, a taxon currently considered a synonym
of G. affinis. A similar mixture was observed in
Pringle 10599 (MEXU), also from the type locality.

At present, Grimmia involucrata is considered a
valid species endemic to Mexico. Its presumed dis¬
tribution in Texas is based on collections of Grim¬
mia americana , a species that is distinguished by
its ventricose sporophytes.

Specimens examined. MEXICO. Distrito Federal:
Zacatenco, Amable 1352 (MEXU); Tlalpan, Amable I44H
(MEXU). Hidalgo: Cuyamaloya, Sierra de Pachuca, Prin¬
gle 1 0599 p.p. (MEXU); alrededores de Ciudad Sahagun,
Garde nos 3103 (MEXU). Mexico : Sierra de Alcaparrosa,

4 km NW de Tepotzotlan, I9°44'N, 99°15'W, Cardenas
3912 (MEXU). Zacatecas: 1 km S de Troncoso, Cardenas
3092 (MEXU); Gerro Gordo, ,4 km SE de La Colonia Hi¬
dalgo. 22°30'N. 102°13'W, Cardenas 736 (MEXU); 2 km

5 de San Juan de los Homillos, cerca de Presa Homos,
23°()1 'N. 1 03° 1 1 ' W, Cardenas 752 (MEXU).

Acknowledgment. The loan of type specimens
by the curators of FH, H-BR, and NY is gratefully
acknowledged.

Literature Cited

Bartram. E. B. 1929. Mosses from western Texas collected
by Mr. C. B. Orcutt. Bryologist 32: 7-1 1.

Cardot, J. 1909. Diagnoses pr^liminaires de mousses mex-
icaines. Rev. Bryol. 36: 105—1 15.

Volume 9, Number 2
1999

Delgadillo M.
Grimmia involucrata

155

Churchill, S. P. & E. L l.inares C. 1995. Prodromus
Bryologiae Novo-Granatensis. Introdueeion a la flora de
musgos de Colombia. Parte 2. Instituto de Ciencias Na-
turales, Universidad Nacional de Colombia, Bogota.

Crum, H. A. 1994. Grimmia. Pp. 386-408 in A. J. Sharp
et ah. The Moss Flora of Mexico. Mem. New \ork Bot.
Card. 69(1).

- & L. E. Anderson. 1981. Mosses ol Eastern North

America, Vol. 1. Columbia Univ. Press, New York.

Sayre, G. 1952. Key to the species of Grimmia in North
America. Bryologist 55: 251—259.

Theriot, 1. 1931. Mexican Mosses Collected by Brother
Arsene Brouard — III. Smithsonian Miscellaneous Col¬
lections 85: 1-44.

Vi hitehouse, E. & E. McAllister. 1954. The mosses ol Tex¬
as. Bryologist 57: 63—146.

Submitted 26 Feb. 1998.

Caxamarca, a New Monotypic Genus of Senecioneae (Asteraceae)

from Northern Peru

Michael 0. Dillon

Department of Botany, The Field Museum, Chicago, Illinois 60605-2496, U.S.A.

Abundio Sagdstegui Alva

Museo de Historia Natural, Universidad Antenor Orrego, Casilla 1001, Trujillo, Peru

Abstract. A new monotypic genus of Astera¬
ceae— Senecioneae, Caxamarca, is described from
northern Peru. The single herbaceous perennial
species, Caxamarca sanchezii M. 0. Dillon & Sa¬
gdstegui, is endemic to the seasonally dry valleys
ol southwestern Cajamarca. The potential relation¬
ships of the genus are discussed.

RESUMEN. Un gdnero monotfpico nuevo de Aster¬
aceae— Senecioneae, Caxamarca se describe para el
norte de Peru. La unica especie, la herbacea per-
enne, Caxamarca sanchezii M. 0. Dillon et Sagas-
tegui, es endemica de los valles estacionahnente
secos del suroeste del Departamento de Cajamarca.
Se discuten las relaciones potenciales del genero.

During continuing detailed studies on the flora
of northern Peru, the following monotypic genus of
Senecioneae (Asteraceae) was encountered in the
seasonally dry river valleys in southwestern Caja¬
marca. This new genus is proposed after an ex¬
haustive examination of all available genera within
the tribe and detailed studies on many neotropical
members of the large and complex genus Senecio
L. While generic concepts within the Senecioneae
have been reviewed (Barkley, 1985a; Barkley et al.,
1996; Bremer, 1994; Nordenstam, 1978), future
changes are to be expected as taxonomies are tested
with data from other disciplines, such as molecular
systematics.

Caxamarca (Asteraceae tribe Senecioneae) M. 0.
Dillon & Sagastegui, gen. nov. TYPE: Caxa¬
marca sanchezii M. O. Dillon & Sagastegui.

Herba rhizomatosa usque ad 1.5 m alta, radices carnosi
foetidi; caules erecti simplexalata medullosus. Folia ba-
sale 40-50 cm longa, profunde pinnatisecta. Folia caulina
alterna, herbacea, sessilia, basi decurrentis; lamina plana
herbacea 20—35 cm longa, 7—1 1 cm lata, margine distinc-
te denticulata. Capitulescentiae corymboso-cymosae 3-20
capitulis (vulgo 3—7). Capitula heterogama radiata ca. 10
mm alta, ca 10 mm lata; phyllaria subbiseriatis 17-28
(vulgo 22-28). Flores radiati pistillati 10-22: ligulae lu-
teae, 10-25 mm longae, 4-8 mm latae. Flores disci her-

maphroditi 80—160 (vulgo 80-130); corollae tubulosae lu-
teae, 12-15 mm longae; limbus profunde quinquelobus.
Antherae basis sagittata; collum filamenti basi dilatatum.
Styli rami apice conico pilis penicillum centrale. Achaen-
ia decemcostata pilosus; pappi setae biseriate 6-7 mm
longus barbellatae persistentes.

Caxamarca sanchezii M. 0. Dillon & Sagdstegui,
sp. nov. TYPE: Peru. Dept. Cajamarca: Prov.
Contumaza, Alrededor de San Benito, 1200 m,
12 May 1994, A. Sagdstegui A., S. Leiva & C.
Sagdstegui 15185 (holotype, HAO; isotypes,
CPUN, F, MO, US). Figures 1, 2.

Perennial herbs to 1.5 m tall; roots fasciculate,
tuberous, malodorous, individual tubers 8-12 cm
long, 2—3 cm diam., white; stems fistulose, 1-2.5
cm diam. at the base, unbranched, glabrous,
winged by the decurrent leaf bases, the wings 4—7
mm wide, glabrous, reticulate-nerved, denticulate,
ciliolate. Leaves basal and cauline; basal leaves ro-
sulate, the blades profoundly dissected, odd pin¬
nate, the rachis 40—50 cm long, 8-10 pairs of leaf¬
lets, the leaflets ovate to lanceolate, 5—10 cm long,
2-6 cm wide, the margins dentate-serrate; cauline
leaves alternate, sessile, decurrent, the blades ob¬
long-lanceolate, 20-35 cm long, 7-11 cm wide,
membranous, apieally acuminate, the margins ir¬
regularly dentate-serrate, reticulate-nerved, gla¬
brous, the nerves sparsely pilose. Capitulescences
cymose-corymbose. Capitula 3— 7(— 20), radiate;
pedicels 1—8 cm long, striate, swollen apieally,
densely pilose; calyculus (10— )16-22-braeteolate,
the bracts narrowly linear, 1-2 cm long, 0.5-1 mm
wide, acuminate apieally, 1-nerved, greenish pur¬
ple; involucres hemispherical, ca. 10 mm tall, ca.
10 mm wide; phyllaries (17-)22-28, weakly bi¬
seriate, equal, lanceolate, 8-10 mm long, 2-3 mm
wide, fused at the base, glabrous or pilose, acu¬
minate; ray florets 10—22, pistillate, the corolla yel¬
low, the tube 0. 5-0.6 mm long, glabrous, the ligule
oblong-linear, 10-25 mm long, 4—8 mm wide, gla¬
brous, 10— 12-nerved, 2-3-dentate apieally; disc

Novon 9: 156-161. 1999.

Volume 9, Number 2
1999

Dillon & Sagastegui
Caxamarca from Peru

157

Figure 1. Caxamarca sanchezii VI. (). Dillon & Sagastegui. —A. Flowering branch. — B. Portion of basal leaf. — C.
Fleshy roots. — D. Capitulurn. — E. Involucre. — F. Ray floret. — G. Disc floret. — H. Corolla lobes. — 1. Stamen. — J.
Style branches of disc floret. — k. Achene with pappus. — L. Enlargement of achene cross section. Illustration based
on Sagastegui , Leiva & Sagastegui 15185 (HAO).

158

Novon

Figure 2. Caxamarca sanchezii. — A. Basal leaves in greenhouse grown plant (scale in inches). — B. Flowering stem
photographed in the field. — C. Capitulescence.

florets 80-130(-160), hermaphroditic, the corollas
yellow, tubular, 12-15 mm long, the throat slightly
expanded, deeply 5-lobed, the lobes oblong, 2-2.5
mm long, 0.8-1 mm wide, apically obtuse, reflexed;

anther bases sagittate, the filaments basal collar
swollen or balusterform; style branches with par¬
allel stigmatic lines, the apices conical with a ter¬
minal crown of penicillate trichomes. Achenes cy-

Volume 9, Number 2
1999

Dillon & Sagastegui
Caxamarca from Peru

159

lindrical, 2.5-4 mm long, 0.5— 0.6 mm wide,
10-costate, densely pilose with biseriate trichomes;
pappus biseriate, persistent, the bristles subequal,
barbellate, 6—7 mm long.

Distribution and ecology. Caxamarca sanchezii
grows in clay soils in semi-arid open habitats with
scattered deciduous small trees and evergreen
shrubs. It is confined to the adjacent cuencas or
drainage basins of the Rio Chicama and Rio Je-
quetepeque (1200—2200 m) in southwestern De¬
partment of Cajamarca. These areas are a part ol
the western or Pacific slope with the rivers draining
to the Pacific Ocean. The prevailing weather pat¬
tern, and hence available moisture, is associated
with coastal weather systems and unpredictable but
recurrent El Nino conditions (Dillon & Rundel,
1990). Flowering is between March and May in nor¬
mal years.

In favorable years, the associated floristic com¬
munity is dominated by several annual Poaceae
genera. Other species recorded from the collecting
localities include Acacia angustissima (Miller)
Kuntze (Fabaceae), Caesalpinia paipai Ruiz & Pa-
von, C. spinosa (Molina) Kuntze (Fabaceae), Erio-
theca discolor (HBK) Robyns (Bombacaceae), Cor-
dia macrocephala (Desvaux) Kunth, C. munda I. M.
Johnston (Boraginaceae), Solarium sp. (Solanaeeae),
Verbesina saubinetioides S. F. Blake (Asteraceae),
Zinnia peruviana (L.) L. (Asteraceae). The region
also contains an association of several geophytes:
Ismene amancaes (Ruiz & Pavon) Herbert (Amar-
yllidaceae), Leptochiton helianthus (Ravenna) Ger-
eau & Meerow (Amaryllidaceae), Stenomesson sp.
(Amaryllidaceae), and Trichlora peruviana Baker
(Liliaceae).

Phenology. The life history of Caxamarca in
the seasonally dry valleys of southwestern Caja¬
marca is closely connected with the yearly cycles
of rain and drought. In the region where it is en¬
demic, the rainy season typically begins in late De¬
cember and rainfall increases steadily until March.
By April the rains begin to decrease and by the
end of May the area is completely dry once again.
Rainfall between June and November is infrequent
and of short duration. Soon after the rainy season
begins, a series of long-petioled, basal leaves
emerge sequentially, one every 4—5 days. Basal
leaves are unknown from field-collected material
since they are extremely ephemeral, but were ob¬
served in cultivated plants (Fig. 2A). In nature, the
basal leaves are completely absent in flowering in¬
dividuals; they develop and die hack to the ground
prior to the development of the flowering stems. The
flowering shoot develops rapidly, and vegetative

growth, flowering, and fruiting are completed within
three months. After flowering, the stems die back
and leave little trace of their existence until the
next year. Some aspects of its auteeology are similar
to geophytes which begin their development with
the coming of the rainy season and flower after the
vegetative growth period.

Etymology. The generic name is derived from
the original Spanish rendering of Cajamarca, the
native Quechua name for this area in northern
Peru. It is a pleasure to dedicate this new species
to its first collector, Isidoro Sanchez Vega, the
founder and current director of the herbarium at
Universidad Naeional de Cajamarca (CPUN).

Paratypes. PERU. Dept. Cajamarca: Prov. Contu-
maza, Travesfa de Andaloy (San Benito— Yeton), Sagdste-
gui & Leiva 15491 (F. HAO), Sagastegui & h-iva 15536
(F, HAO); S. 0. de El Rupe. Km 13 de la carretera Chi-
lete-Contumaza. Sanchez 2243 (CPUN); Chilete-Rupe,
Sanchez 4219 (CPUN, HAO); Cushion (Chilete— Contu-
maza), Sagastegui et al. 15628 (F, HAO), Sagastegui et
al. 15857 (F, HAO); Tambo Ua Eima (Cascas— Contumaza),
Sagastegui et al. 15548 (F, HAO); Prov. San Pablo, San
Bernardino— Sangal, Sagastegui et al. 15597 (F, HAO).

Discussion. Our studies in the Andean Sene-
cioneae (Dillon & Sagastegui, 1988, 1996; Vision
& Dillon, 1996; Sagastegui & Dillon, 1985) have
revealed rich generic and species diversity. With
this addition, Peru contains no fewer than 15 de¬
scribed genera and over 320 species within the
tribe Senecioneae (Dillon & Hensold, 1993). This
new genus belongs to the subtribe Senecioninae
Dumortier, diagnosed by a series ol microcharac¬
ters, including balusterform anther collars and style
branches with separate stigmatic lines (Bremer,
1994).

The combination of characters displayed by Cax¬
amarca does not allow its placement within any
described genus known to us within the Seneci¬
oneae: the distinctive herbaceous perennial habit,
with enlarged fleshy or tuberous roots, precocious
basal leaves (Fig. 2A), scapose fistulose stems with
decurrent cauline leaves (Fig. 2B). lax and few¬
headed cymes, and large capitula with conspicuous
reflexed, bright yellow ray corollas (Fig. 2C). The
style branches of the disc florets have parallel stig¬
matic lines and convex to triangular apices with a
terminal crown of penicillate trichomes. This type
of style apex is unlike those found in the large and
diverse genus Senecio L.

Principal basal leaves and fibrous roots similar
to those present in Caxamarca are known from oth¬
er genera within the Senecioneae, for example, Pip-
penalia, a Mexican radiate genus once considered
intermediate between Cacalia and Senecio (Barkley,

160

Novon

1985b; Barkley et al., 1996). Some species of Se-
necio develop prominent basal leaves, for example,
the Brazilian Senecio pulcher Hooker & Arnott. The
morphology of the basal leaves in Caxamarca is
still an open question since basal leaves have not
been collected in the field. While not common,
fleshy roots with greatly thickened tubers do occur
in some Mexican Rolandas (T. Barkley, pers.
comm.) and in other tribes such as the Liabeae
(e.g., Munnozia sagasteguii H. Robinson) occurring
in this region.

In the key to genera provided by Vision and Dil¬
lon (1996), Caxamarca would key to Dorobaea Cas¬
sini. The large radiate hemispheric capitula with
well-developed calyculus and numerous phyllaries
are reminiscent of those found in Dorobaea, a small
Andean genus distributed in northern Peru, Ecua¬
dor, and southern Colombia (Nordenstam, 1978;
Nordenstam & Pruski, 1995). Dorobaea was estab¬
lished as distinct from Senecio on character differ¬
ences in the habit, capitulescences, and floral mor¬
phology. The distinctness of the genus has been
confirmed by cpDNA variation studies (Kadereit &
Jeffrey, 1996). Dorobaea has deeply lobed or pin¬
nate, basally rosulate leaves present at the time of
flowering. It has a leafless, scapose capitulescence
bearing a solitary, radiate capitulum. Dorobaea has
slightly thickened fibrous roots, not unlike those in
Caxamarca, but in Dorobaea the roots are typically
no more than 7-10 cm long and ca. 1 cm in di¬
ameter. Within its range, Dorobaea is distributed in
mesic habitats at elevations above 2000 m, and
while Dorobaea and Caxamarca are both distrib¬
uted in Cajamarca, they are ecologically and geo¬
graphically isolated. The chromosome number for
Dorobaea has been reported as n — ca. 50 (Dillon
& Turner, 1982). A chromosome count for Caxa¬
marca should be useful in efforts to determine the
relationships of this genus (Robinson et al., 1997).

The following key is provided for Andean genera
with herbaceous or suffrutescent habits, radiate ca¬
pitula with a prominent calyculus, and conical to
triangular style branch apices.

Key to Caxamarca and Potentially Related Andean
Genera

la. Scandent, herbaceous perennials or suffrutescent
vines; leaves strictly cauline; petioles not long-
decurrent on stems; capitula radiate (rarely dis¬
coid).

2a. Reclining, herbaceous perennials (typically
hydrophilic); capitula radiate, the corollas
yellow or orange; anther collar filament cy¬
lindrical, the theca bases rounded ....

. Garcibarrigoa Cuatrecasas

2b. Robust, suffrutescent vines; capitula radiate
or discoid, the corollas orange or occasion¬

ally rose to purple; anther collar filament

swollen or balusterform .

. Pseudogynoxys (Greenman) Cabrera

lb. Erect, herbaceous perennials; leaves basal, or
basal and cauline with long-decurrent petioles;
capitula radiate.

3a. Leaves strictly basal, no cauline leaves evi¬
dent; capitulescence a long-scapose, solitary

capitulum . Dorobaea Cassini

3b. Leaves basal and cauline; capitulescence

cymose-corymbose .

. Caxamarca M. O. Dillon & Sag&stegui

Acknowledgments. We are grateful to the Univ-
ersidad Privada Anterior Orrego for supporting field
studies. Theodore Barkley is acknowledged for
valuable review comments, which greatly improved
the manuscript. We thank Fred Barrie and Nancy
Hensold for comments on an early version of the
manuscript. The illustration was prepared by Se-
gundo Leiva Gonzales, Professor of Botany at
UPAO. Lastly, the authors wish to acknowledge and
thank Theodore Barkley, Harold Robinson, John
Pruski, Bertil Nordenstam, and especially, the late
Jos6 Cuatrecasas, for sharing their ideas concerning
generic concepts in the Senecioneae.

Literature Cited

Barkley, T. M. 1985a. Generic boundaries in the Seneci¬
oneae. Taxon 34: 17-21.

- . 1985b. Infrageneric groups in Senecio, s.l., and

Cacalia, s.l. (Asteraceae: Senecioneae) in Mexico and
Central America. Brittonia 37: 211-218.

- , B. L. Clark & A. M. Lunston. 1996. The segre¬
gate genera of Senecio sensu lato and Cacalia sensu lato
(Asteraceae: Senecioneae) in Mexico and Central Amer¬
ica. Pp. 613—620 in D. J. N. Hind & H. .). Beentje
(editors), Compositae: Systematics. Proceedings of the
International Compositae Conference, Kew, 1994, vol.

1 . Royal Botanic Gardens, Kew.

Bremer, K. 1994. Asteraceae. Cladistics and Classifica¬
tion. Timber Press, Portland, Oregon.

Dillon. M. (). & N. Hensold. 1993. Family Asteraceae. In:
L. Brako & J. L. Zarucchi, Catalogue of the Flowering
plants and Gymnosperms of Peru. Monogr. Syst. Bot.
Missouri Bot. Gard. 45: 103-189.

- & P. W. Rundel. 1990. The botanical response of

the Atacama and Peruvian Desert Flora to the 1982-
83 El Nino event. Pp. 487-504 in P. W. Glynn (editor).
Global Ecological Consequences of the i 982-83 El
Nino-Southern Oscillation. Elsevier Science Publishers,
New York.

- & A. Sagastegui A. 1988. Additions to South

American Senecioneae (Asteraceae). Brittonia 40: 221-
228.

- & - . 1996 [1997], Revision of the dioecious

genus Chersodoma Phil. (Senecioneae, Asteraceae), in¬
cluding a new species and status change. Brittonia 48:
582-604.

- & B. L. Turner. 1982. Chromosome numbers of

Peruvian Compositae. Rhodora 84: 131-137.

Kadereit, J. W. & C. Jeffrey. 1996. A preliminary analysis
of cpDNA variation in the tribe Senecioneae (Compos¬
itae). Pp. 349-360 in D. J. N. Hind & H. J. Beentje

Volume 9, Number 2
1999

Dillon & Sagastegui
Caxamarca from Peru

161

(editors), Conipositae: Systematics. Proceedings of the
International Compositae Conference, Kew, 1994, vol.
1. Royal Botanic Gardens, Kew.

Nordenstam, B. 1978. Taxonomic studies in the tribe Se-
necioneae (Compositae). Opera Bot. 44: 1—83.

- & J. F. Pruski. 1995. Additions to Dorobaea and

Talamancalia (Compositae— Senecioneae). Comp.
Newsl. 27: 31-42.

Robinson. H., G. I). Carr, R. M. King & A. M. Powell.
1997. Chromosome numbers in Compositae, XVII: Se¬
necioneae III. Ann. Missouri Bot. Card. 84: 893—906.
Sagastegui-Alva, A. & M. 0. Dillon. 1985. Four new spe¬
cies of Asteraceae from Peru. Brittonia 37: 6—13.
Vision, T. J. & M. 0. Dillon. 1996. Sinopsis de Senecio
L. (Senecioneae, Asteraceae) para el Peru. Amaldoa 4:
23^16.

New Taxa of Ericaceae from China

Fang Rhui-cheng

Department of Phytotaxonomy and Phytogeography, Kunming Institute of Botany,
Chinese Aeademy of Seiences, Kunming, Yunnan 650204, People’s Republic of China

ABSTRACT. Nine new species (Cassiope membran-
ifolia, Gaultheria heteromera, G. jingdongensis, G.
longibracteolata, G. pseudonotabilis, G. purpurea, G.
trigonoclada, Vaccinium rubeseens, V. tenuiflorum ),
tour new varieties ( Enkianthus deflexus var. gla-
brescens, Gaultheria dumicola var. hirticaulis, G.
griffithiana var. insignis, G. wardu var. elongata ),
a new name (G. straminea), and four new combi¬
nations ( Enkianthus serrulatus var. sichuanensis,
Gaultheria leucocarpa var. yunnanensis, G. leuco-
carpa var. psilocarpa, G. brevistipes) are proposed.

As a result of a recent visit to several herbaria
in the United States, as well as the examination of
additional material that has accumulated in Chi¬
nese herbaria since the publication of Fang’s (1991)
treatment of the Ericaceae for Flora Reipublicae
Popularis Sinicae , the tollowing novelties are pro¬
posed to make the names available for the forth¬
coming account of the Ericaceae for the Flora of
China.

Cassiope meiubranifolia R. C. Fang, sp. nov.
TYPE: China. W Yunnan: Lushui Xian, 3600
m, 25 Aug. 1964, S. G. Wu 8423 (holotype,
KUN). Figure 1.

Species nova C. myosuroidae W. W. Smith affinis. a quo
foliis plerumque suborbicularibus raro ellipticis, dorsis
saepe glabris, acuminibus membranaceis laminis aequal-
ibus vel longioribus, 1-2 mm longis differt.

Procumbent slender shrubs. Stems elongated, ca.
26 cm long. Leaves assurgent, densely imbricate
on stem, appressed in 4 ranks; leaf blade subor-
bicular, rarely elliptic, scale-like, 1.2-1. 8 X 1-1.5
mm, coriaceous, adaxially concave, glabrous, abax-
ially not furrowed, convex, glabrous or sometimes
pilose at base, base arched, margin distinctly mem¬
branous, membranous rim upward forming an elon¬
gated-triangular hyaline acumen as long as or lon¬
ger than blade. Flowers solitary, axillary. Pedicels
1—2.5 cm long, densely crisped-tomentose, base
with 3 lacerate bracts; bracteole absent. Calyx pur¬
ple, glabrous; lobes subtree, oblong, ca. 2.5 mm
long, margins membranous, fimbriate. Corolla
white, campanulate, 5—7 mm long, glabrous out-

Novon 9: 162-178. 1999.

side, 5-lobed; lobes ovate, ca. 2 mm long. Stamens
10, ca. 1.5 mm long; filaments linear, flattened, pi¬
lose or glabrous; anthers ovate, with 2 recurved
awns ca. 1 mm long. Ovary glabrous. Capsule not
seen. Flowering August.

Distribution and ecology. China, western Yun¬
nan Province. Alpine meadows among moss on
rocks; ca. 3600 m.

Cassiope membranifolia appears to be closely re¬
lated to C. myosuroides W. W. Smith, which has a
similar membranous leaf blade margin. It differs in
that the leaf margin is membranous with the distal
portion of the rim forming a hyaline elongated-tri¬
angular acumen, and in the leaf blades being sub-
orbicular, rarely elliptic, adaxially concave, rough¬
ened, and abaxially subglabrous.

Paratypes. CHINA. Yunnan: without locality, G. For¬
rest 79H2 (A); Lushui Xian, Bijiang Exped. 1772 (KUN).

Enkianthus deflexus (Griffith) C. K. Schneider
var. glabrescens R. C. Fang, var. nov. TYPE:
China. Gansu: Wen Xian, 1640 m, 21 July
1991, Y. F. Wang & X. L. Chen 911007 (ho¬
lotype, MO).

A var. deflexo foliis utrinque glabrescentibus vel abax-
ialiter ad costas sparse hispidulis; petiolis glabris; rhachi-
dibus et pedicellis pubescentibus et pilis glanduliferis ob-
tectis differt.

Enkianthus deflexus var. glabrescens differs from
variety deflexus in having leal blades glabrescent
on both surfaces or abaxially sparsely hispidulous
along the midrib, glabrous petioles, and glandular-
pubescent rachises and pedicels. Except for this
new variety, all of the other Chinese taxa of En¬
kianthus are found in provinces south of the
Changjing (Yangtze River).

Enkianthus serrulatus (Wilson) C. K. Schneider
var. sichuanensis (T. Z. Hsu) R. C. Fang, stat.
et comb. nov. Basionym: Enkianthus sichu¬
anensis T. Z. Hsu, Acta Bot. Yunnan. 4: 358.
1982. TY PE: China. Sichuan: Fengjie Xian,
M. Y. Fang 24515 (holotype, HIB not seen;
isotype, KUN).

Volume 9, Number 2
1999

Fang

New Ericaceae from China

163

Figure 1. Cassiupe membranifolia R. C. Fang. — A. Habit. — B. Flower. — C. Opened corolla. — I). Stamens and
pistil. — E. Stamen. — F. Adaxial leaf surface. — G. Abaxial leaf surface. Scale bars: A = 1 cm, B— G = 1 mm. Drawn
from holotype (S. G. Wn 8423).

164

Novon

Enkianthus sichuanensis was based on a single
collection, which is indistinguishable from material
of E. serrulatus, except for the hairy ovary. Enkian¬
thus sichuanensis is reduced to a variety of E. ser¬
rulatus since the type collection of the former was
collected from within the range of the latter.

Gaultheria brevistipes (C. Y. Wu & T. Z. Hsu) R.
C. Fang, comb. nov. Basionym: Leucothoe brev-
istipes C. Y. Wu & T. Z. Hsu, Act. Phytotax.
Sin. 23: 315. 1985. Leucothoe griffithiana C.
B. Clarke var. sessilifolia C. Y. Wu & T. Z.
Hsu, FI. Xizang. 3: 684. 1986. TYPE: China.
Xizang: Medog, Qinghai-Xizang Complex Ex-
ped. 74-3951 (holotype, KUN).

Both Leucothoe brevistipes and L. griffithiana var.
sessilifolia are based on the same collection num¬
ber. Hsu (1991) incorrectly listed the species as L.
sessilifolia C. Y. Wu & T. Z. Hsu [sic] with the exact
bibliographical citation ot L. brevistipes, and cited
L. griffithiana var. sessilifolia as a synonym.

Additional specimens examined. CHINA. Xizang: Me¬
dog, 8. S. Li & S. Z. Cheng 1289. 2496. 3358 . 3741,
4757, 5082 (KUN), H. Sun, Z. K. Zhou & H. Y. Yu 516,
2751, 3655, 3713, 4827 (KUN).

Gaultheria dumicola W. W. Smith var. hirticaulis

R. C. Fang, var. nov. TYPE: China. SE Yun¬
nan: Malipo Xian, 2000 m, 6 Dec. 1962, Q.
A. Wu 62—134 (holotype, KUN).

A var. dumicola recedit caulibus densissime brunneo-
hirsutis, foliis subtus hispidis.

Juvenile and adult branches densely brown-hir¬
sute. Leal blade ovate, 3-8 X 2^1 cm, base rotund
to broadly cuneate, adaxially glabrous, abaxially
short brown-hispid.

This new variety is readily distinguished from
the other varieties of Gaultheria dumicola by hav¬
ing densely brown-hirsute juvenile and adult
branches. All of the remaining varieties of G. dum¬
icola have glabrous leaves, except lor variety aspera
Airy-Shaw, which has abaxially hispid leal blades.

Gaultheria griffithiana Wight var. insignis R. C.

Fang, var. nov. TYPE: China. SE Xizang: Zayu
Xian, 2800 m, 18 July 1973, Qinghai-Xizang
Complex Exped. 73-781 (holotype, KUN; iso¬
type, PE).

A var. griffithiana recedit rhachidibus robustis, lanu-
ginosis; pedicellis robustis, tomentosis; bracteis rotundo-
ovatis, 5-7 mm longis, extus infeme et intus omnino den¬
sissime sericeis; calycibus intus densissime sericeis.

Plants robust, 1—1.2 m tall. Stems narrowly

winged. Leal blade oblong or narrowly ovate, 13—
14 X 4.5—6 cm, base subcordate, apex short cau¬
date. Racemes 4—5 cm long; rachis robust, lanu-
ginose; pedicels robust, 6-9 mm long, tomentose;
bracts lulvous, rotund-ovate, 5-7 mm long, densely
sericeous adaxially and abaxially below.

Gaultheria griffithiana var. insignis is most easily
distinguished from variety griffithiana by having
lanuginose rachises, tomentose pedicels, densely
sericeous bracts and calyx interiors, and subcordate
bases ol the leaf blades. By contrast, variety grif¬
fithiana has pubescent rachises and pedicels,
subglabrous bracts and calyx interiors, and obtuse-
rounded bases of the leal blades.

Gaultheria heteromera R. C. Fang, sp. nov.
TYPE: China. Xizang: Medog Xian, 3900 m,
1 Aug. 1974, Qinghai-Xizang Complex Exped.
74-3876 (holotype, KUN; isotypes, KUN, PE).
Figure 2.

Species nova G. prostratae W. W. Smith affinis, sed ca¬
lycibus tetrasepalis, corollis globoso-urceolatis, ramulis
puberulis et hirtellis. A G. pyrolaefolia J. I). Hooker ex
C. B. Clarke foliis parvioribus, ovariis hispidulis, calyci¬
bus tetrasepalis differt.

Small shrubs, ea. 10 cm tall. Stems procumbent,
terete, slender, branchlets densely puberulent
mixed with long brown hairs. Leaves scattered
along stem; petioles ca. 1 mm long, glabrous or
sparsely hirtellous; leaf blades elliptic, 0.8— 1.7(—
2.5) X 0.6-1 cm, chartaceous, base broadly cune¬
ate, margin dentate, teeth apiculate, apex obtuse to
acute, mucronulate, both surfaces drying brown,
glabrous or sparsely hirtellous abaxially; veins
slender, secondary ones in 3 pairs, ultimate veinlets
conspicuous, all immersed adaxially, raised abaxi¬
ally. Inflorescences shortly racemose, terminal or
axillary on branchlet apices, 1-4-flowered; rachis
3-6 mm long, puberulent; bracts broadly ovate,
1.5-2 mm long, glabrous. Pedicels 4-6 mm long,
puberulent; bracteoles 1 or 2 pairs, at middle or
upper part of pedicel, somewhat remote from calyx,
ovate, 1-1.5 mm long, glabrous. Calyx glabrous;
lobes 4, triangular-ovate, 2.5—3 mm long. Corolla
white, globose-urceolate, ca. 4 mm long, glabrous
abaxially; lobes 4, very short, reflexed. Stamens 8
or 9, ca. 2 mm long; filaments dilated basally, gla¬
brous, papillose; theca 2-aristate. Ovary hispidu-
lous. Capsule unknown.

Distribution and ecology. China, southeastern
Xizang. Alpine cushion thicket meadows; ca. 3900 m.

Gaultheria heteromera is closely related to G.
prostrata W. W. Smith from which it differs by hav¬
ing a 4-lobed calyx, a globose-urceolate corolla.

Volume 9, Number 2
1999

Fang

New Ericaceae from China

165

Figure 2. Gaultheria heteromera R. C. Fang. — A. Habit. — B. Flower with two pairs of bracteoles. — C. Abaxial leaf
surface. — D. Opened calyx showing pistil. — E. Stamen ventral view. — F. Stamen dorsal view. Scale bars: A = 1 cm,
B, D-F = 1 mm, C = 5 mm. Drawn from holotype ( Qinghai-Xizang Complex Exped. 74-3876).

166

Novon

and puberulent and hirtellous branchlets. Gaulthe¬
ria prostrata has a 5-lobed calyx, a broadly cam-
panulate corolla, and pubescent branches that be¬
come glabrescent and cinereous with age.
Gaultheria heteromera differs from the closely re¬
lated G. pyrolaefolia in having smaller leaves (0.8-
1.7(— 2.5) X 0.6-1 cm vs. 1.3-5 X 0.8-2.5 cm), a
4- instead of 5-lobed calyx, and hairy instead of
glabrous ovary.

Paratypes. CHINA. Xizung: Medog, Qinghai-Xizang
Complex Exped. 74-3896 (KUN, PE).

Gaultheria jingdongensis R. C. Fang, sp. nov.
TYPE: China. WC Yunnan: Jingdong Xian,
2700 m, 1 May 1959, S. G. Wu 4774 (holotype,
KUN; isotype, KUN). Figure 3.

Proxima G. trichophyllae Royle, sed foliis majoribus,
subtus dense hispidulis, margine setis caducis, floribus
persaepe majoribus, lobis calycis 3-5 mm longis, dense
eiliolatis, corollis 6—9 mm longis, profunde divisis, lobis
3—6 mm longis differ!.

Small shrubs, ea. 15 cm tall. Stems slender, te¬
rete, slightly prostrate, densely cinnamon- or
brown-hirsute. Petioles 1—2 mm long, hirsute; leal
blades obovate-oblong, 1-2 X 0.5-1 cm, charta-
ceous to coriaceous, base euneate-attenuate, mar¬
gin denticulate, teeth setose or setae caducous,
apex obtuse or rotund, mucronate, adaxially drying
brown-green, glabrous, abaxially cinnamon to
brown, densely hispidulous; midrib slightly raised
below, secondary veins and ultimate veinlets ob¬
scure on both surfaces. Flower solitary, axillary;
pedicels 2-3 mm long, hirtellous; bracts lacking;
bracteoles 2, broadly ovate, ca. 3 mm long, coria¬
ceous, cucullate, close to calyx, glabrous. Calyx
lobes unequal, triangular-ovate, 3—5 mm long,
densely white ciliolate. Corolla white, 6-9 mm long,
deeply 5-lobed; lobes oblong-ovate, 3—6 mm long,
erect or apically recurved. Stamens ca. 3 mm long;
filaments ca. 1 .5 mm long, rhomboid, glabrous, pa¬
pillose; thecae 1-aristate. Ovary glabrous. Fruit un¬
known.

Distribution and ecology. China, west-central
Yunnan. Shrub forests or roadsides on slopes; ca.
2700 m.

Gaultheria jingdongensis differs from the related
G. trichophylla in having larger (10-20 X 5-10 mm
vs. 5—13 X 2—5 mm) leaves, abaxially densely his¬
pidulous instead of glabrous leaf blades, and ca¬
ducous setae instead of persistently long-ciliate leaf
margins. The flowers are also larger. The densely
eiliate calyx lobes are 3—5 mm long instead of gla¬
brous and 2—2.8 mm long. Gaultheria jingdongensis

also has a larger corolla (6-9 mm vs. ca. 6 mm
long) with lobes 3-6 instead of 2-3 mm long.

Paratypes. CHINA. Yunnan: Jingdong, M. K. Li 3545
(KUN), B. Y. Qiu 52915 (KUN), {). A. Wu 9234 (KUN),
Q. A. Wu 9399 (KUN).

Gaultheria leucoearpa Blume var. psilocarpa

(H. F. Copeland) R. C. Fang, stat. et comb,
nov. Basionym: Gaultheria psilocarpa H. F.
Copeland, Philipp. J. Sci. 47: 62. 1932. TYPE:
Philippines. Mindanao, Bukidnon Province,
Bur. Sci. 38903, Ramos & Edano (holotype,
M? (not seen) or PNH? (not seen); isotype. A).

Gaultheria leucoearpa Blume var. yunnanensis

(Franchet) T. Z. Hsu & R. C. Fang, stat. et
comb. nov. Basionym: Vaccinium yunnanen.se
Franchet, J. Bot. (Morot) 9: 368. 1895. TYPE:
China. Yunnan: Tchen-fong-chan (Cheng-feng-
shan). J. M. Delavay 3069 (holotype, P).

kKY TO THK VARIKTIKS OK GAULTHERIA LEUCOCARI'A

la. Ovary and filaments glabrous; stems sparsely pu¬

berulent and glandular-pilose or glabrous; China
(Taiwan), Philippines .

. var. psilocarpa (H. F. Copeland) R. C. Fang

lb. Ovary sericeous, filaments variously hairy; stems
glabrous or variously hairy.

2a. Mature fruit white; horns of anther tubes 0.3

mm long; Malay Peninsula, Java .

. var. leucoearpa

2b. Mature fruit black, blue, or black-purple;
horns of anther tubes 0.5 mm long.

3a. Stems and leaves glabrous; China (Fu¬
jian, Guangdong, Guangxi, Guizhou,
Hubei, Hunan, Jiangxi, Sichuan, Tai¬
wan, Yunnan), Indo-Chinese Peninsula
. . . var. yunnanensis (Franchet) T. Z. Hsu
& R. C. Fang

3b. Stems and flowering branches hairy.

4a. Stems, petioles, rachis or pedicels
glandular-hirsute; leaf blade abaxi¬
ally short-hispid, sometimes adaxi¬
ally glandular-hirsute, margin seta-
eeous-ciliate; China (Guangxi, C

and SE Yunnan) .

. var. crenulata (Kurz) T. Z. Hsu

4b. Stems with glandular bristles, ra¬
chis and pedicels glabrous, leaf
blade margin spinulate-serrulate or
cal loso-serru late, glabrous; Philip¬
pines .

. . . var. cumingiana (Vidal) T. Z. Hsu

Gaultheria longibracteolata R. C. Fang, sp. nov.
TYPE: China. S Yunnan: Luchun Xian, 1700
m, 17 Sep. 1973, D. D. Tao 68 (holotype,
KUN; isotype, KUN). Figure 4.

Species valde affinis G. discolori Nuttall ex J. I). II ook-

Volume 9, Number 2
1999

Fang

New Ericaceae from China

167

Figure 3. Gaultheria jingdongensis R. C. Fang. — A. Habit. — B. Abaxial leaf surface. — C. Flower with bracteoles.
— I). Opened corolla showing stamens. — E. Pistil, calyx, anil bracteoles. — F. Stamen dorsal view. — G. Stamen ventral
view. — -H. Portion of stem. Scale bars: A = 1 cm, B = 5 mm, C— G = 1 mm, H = 4 mm. Drawn from holotype (S.
G. Wu 4774).

168

Novon

Figure 4. (kiultheria longibracteolata R. C. Fang. — A. Habit. — B. Abaxial leaf surface. — C. Portion of stem. — D.
flower with bracteoles. — E. Capsule with persistent calyx and bracteoles. — F. Bract. — G. Stamen ventral view. Scale
bars: A, B = 1 cm. C = 6 mm, D— G = 1 mm. Drawn from holotype (D. D. Tao 68).

Volume 9, Number 2
1999

Fang

New Ericaceae from China

169

er. sed ramulis dense hispidis et puberulis, foliis subtus
appresse-setulosis, stylis pilosis, capsulis calycibus car-
mesinis vel purpureis tectis, aristis antherae 0.8 mm lon-
gis, non bicuspidatis differl. et a G. straminea R. C. Fang
differt foliis ellipticis vel oblongo-ellipticis, bracteolatis
calycem subaequantibus, stylis pilosis, capsulis calycibus
carmesinis vel purpureis tectis.

Shrubs ca. 50 cm tall. Stems terete, slender an¬
gular, densely brown hispid and puberulent. Peti¬
oles red, 2-4 mm long, slightly hispid; leaf blades
elliptic or oblong-elliptic, 2.5—5 X 1—2 cm, thin
coriaceous, base broadly cuneate to obtuse, margin
dentate, recurved, apex acute to obutse, mucronu-
late, adaxially drying gray-green, glabrous, abaxi-
ally silvery-white, drying gray-white or pale brown,
sparsely and distinctly appressed-setulose; second¬
ary veins 2 or 3 pairs, ultimate veinlets obscure.
Inflorescences from axils of upper leaves and ter¬
minal, ca. 2.5 cm long; raehis densely pubescent;
bracts orbicular-ovate, 4—5 mm long, margin fim-
brillate, abaxially subglabrous, adaxially pubes¬
cent. Pedicels 3—4 mm long, densely pubescent;
bracteoles 2, ovate, similar to bract, ca. 5 mm long,
close to calyx. Calyx pink, 4—5 mm long, glabrous
abaxially, pubescent adaxially; lobes oblong-ovate,
2.5—3 mm long, ciliolate, apex apiculate. Corolla
white, urceolate, ca. 6 mm long, glabrous outside,
upper part ol inside pilose; lobes ovate-triangular,
ca. 0.8 mm long. Stamens 10, ca. 3 mm long; fil¬
aments flattened, adaxially glabrous, abaxially pi¬
lose; thecae with 2 long aristae ca. 0.8 mm long.
Ovary tomentose; style pilose. Fruit globose, ca. 5
mm diam.; capsule enclosed by crimson or purple-
green mucronate calyx. Flowering June— September,
fruiting August— October.

Distribution and ecology. China (southern Yun¬
nan Province) to Thailand. Evergreen forest mar¬
gins, on open slopes or among thickets; 1200-
2000(-2700) m.

Gaultheria longibracteolata is most closely relat¬
ed to the Bhutanese G. discolor Nuttall ex J. D.
Hooker and Tibetan G. straminea (C. Y. Wu & T.
Z. Hsu) R. C. Fang, which it resembles in the out¬
line of the leaf blade, few lateral veins, and grayish
white or silvery-white abaxial leaf surface. It differs
from G. discolor by having densely hispid and pu¬
berulent branchlets, abaxially sparsely appressed-
setulose leaf blades, pilose styles, and anthers with
aristae ca. 0.8 mm long. It differs from G. straminea
by having elliptic or oblong-elliptic instead of ob¬
long-lanceolate leaf blades, bracteoles almost as
long as instead of shorter than calyx, pilose instead
of glabrous styles, and capsules enclosed by a crim¬
son or purplish instead of white calyx.

Paratypes. CHINA. Yunnan: Yuanyang Xian, Feng

Chen Lin, A. Henry 9460B (A); without locality, A. Henry
9460C (A); Kuan Yin Shan, Red River from Manmei,
6(MM) ft., A. Henry 9761 (A); Wenshan Xian, //. T. Tsai
51543 (A, KUN); Mienning (Lincang Xian), T. T. Yii
17726 (KUN); Jingdong Xian, M. K. Li 1792 (KUN); Lin¬
cang Xian, J. S. Xin 601 (KUN); Xingping Xian, 5. G. Wu
474 (KUN); Yuanjiang Xian, S. G. Wu 761 (KUN); Wen¬
shan Xian, S. G. Wu 61-3832 (KUN); Mengla Xian, Z. H.
Tsi 91-214 (A); Jingdon Xian, H. Peng & H. Hai 649
(KUN); Luchun Xian, S. G. Wu, Y. M. Shui, J. Murata et
al. 47, 861, 2577 (KUN, MAK). THAILAND. Chiang
Mai: Doi. Intanon National Park, J. F. Maxwell 89-1613
(A).

Gaultheria pseudonotabilis H. Li ex R. C. Fang,
sp. nov. TYPE: China. NW Yunnan: Gongshan
Xian, Dulongjian region, 1350 m, 9 Mar. 1991,
Dulongjian Bot. Exped. 4446 (holotype, KUN;
isotype, KUN). Figure 5.

Species valde affinis G. notabili Anthony, sed foliis ma-
joribus, 8—15 X 3-7 cm, basi cordato-rotundatis, floribus
rubris, filamentis glabris differt.

Shrubs l-2(-3) m tall. Stems terete, densely and
patently ferrugineous-setose, setae 2-5 mm long.
Petioles 3—7 mm long, densely and patently ferru¬
gineous-setose; leaf blades elliptic-ovate, ovate, or
oblong-lanceolate, (5— )8— 15 X 3—7 cm, charta-
ceous-coriaceous, base shallowly cordate to round¬
ed, margin sparsely dentate, slightly revolute,
densely setiform-ciliate, apex acuminate, caudate,
or acute, adaxially bullate, glabrous, abaxially
sparsely long setose along midrib and secondary
veins, otherwise glabrous; veins adpressed adaxi¬
ally, distinctly raised abaxially, secondary veins in
2 pairs arising from leaf base. Inflorescences short¬
ly corymbose-racemose, axillary; rachises 2—8 mm
long, glabrous; bracts leathery, rhomboid-triangular,
2-4 mm long, glabrous, ciliolate. Pedicels 0.8— 1.2
cm long, glabrous; bracteoles 2, basal, similar to
bracts but smaller. Calyx cupular, 4—5 mm long;
lobes triangular, 2.5—3 mm long, glabrous. Corolla
red, broadly campanulate, 7—9 mm long, glabrous
outside; lobes triangular-ovate, 3—5 mm broad. Sta¬
mens 10, 3-4 mm long; filaments 1.5—2 mm long,
glabrous, papillose, base dilated; thecae 2—2.5 mm
long, shortly 2-aristate. Ovary glabrous. Fruit de¬
pressed globose, ca. 5 mm diam.; capsules enclosed
by dark purple calyx. Flowering March, fruiting
May.

Distribution and ecology. China, northwestern
Yunnan Province. Under evergreen broad-leaved
forest, on rocks or among thickets; 1300-2000 m.

Gaultheria pseudonotabilis is allied to and more
robust than G. notabilis Anthony. Most of the es¬
sential characters and geography of the two are ba¬
sically the same. Gaultheria notabilis is a subshrub

170

Novon

Figure 5. Gaultheria pseudonotabilis H. Li ex R. C. Fang. — A. Habit. — B. Portion of stem. — C. Portion of abaxial
leaf surface. — D. Flower. — E. Stamen ventral view. — F. Stamen dorsal view. Scale bars: A, C = 1 cm. B = 4 mm,
D— F = 1 mm. Drawn from holotype ( Dulongjian Bot. Exped. 4446).

Volume 9, Number 2
1999

Fang

New Ericaceae from China

171

30-40 cm tall with leaves 2-3.5 X 1.5—2 cm,
bracts 1-2 mm long, pedicels ca. 6 mm long, calyx
lobes ca. 3 mm long, and corolla lobes 2-3 mm
long. Gaultheria pseudonotabilis can be separated
from the related G. codonantha Airy-Shaw by its
adaxially glabrous instead of densely short setulose
leaf blades, abaxially sparsely long setose instead
ol very densely ferrugineous-setulose especially
along the midrib and secondary veins, glabrous in¬
stead ol densely tomentose bracts and inflorescence
rachises, short (2-8 mm vs. 10^40 mm) rachises,
and red instead of white flowers.

The name Gaultheria pseudonotabilis first ap¬
peared as a nomen nudum in Li (1993), and is
herein validated.

Paratypes. CHINA. Yunnan: Gongshan Xian. A. M.
Feng 24731 (KI N), Q. Lin & X. F. Deng 79-1170 (KUN),
Qinghai-Xizang Complex Exped. 8144 (KUN, PE), Du-
longjian Hot. Exped. 915, 8149, 8406, 4152, 4115, 4208.
4221. 4624. 4669. 5395 , 6738, 7013 (KUN).

Gaultheria purpurea R. C. Fang, sp. nov. TYPE:
China. Xizang: Medog Xian, 2400 m, 2 Dec.
1982, B. S. Li & S. Z. Cheng 1994 (holotype,
KUN; isotype, PE). Figure 6.

Haec species differt a congeneris foliis in sicco subtus
purpuratis, secus costas, nervos laterales et reticulationes
dense hirsutis, racemis summis caulis, pedunculis 3.5—
4.5 cm longis.

Shrublets, 5—10 cm tall. Stems procumbent, te¬
rete, slender, branchlets slightly puberulent, dense¬
ly brown hirsute. Leaves scattered; petioles 1-2
mm long, hirsute; leaf blades elliptic, 1.5—3 X 1—
2 cm, coriaceous, base broadly cuneate to rotund,
margin denticulate, teeth setose, apex obtuse to
acute, mucronulate, adaxially drying gray-green,
glabrous, abaxially purple, hirsute along veins; sec¬
ondary veins 3-5 on each side of midrib, midrib,
secondary veins, and ultimate veinlets immersed
adaxially and raised abaxially. Inflorescences ra¬
cemose, on branchlet apices, terminal or axillary,
5—7 cm long in fruit; peduncle 3.5— 4.5 cm long,
glabrous; rachises glabrous; bracts ovate, coria¬
ceous, 2—3 mm long, glabrous. Bracteoles 2, at
middle or upper part of pedicel, ovate-triangular,
ca. 2 mm long, ciliolate. Calyx glabrous; lobes tri¬
angular-ovate, 1.5—2 mm long, ciliolate. Corolla
and stamens not seen. Ovary hispidulous. Fruit glo¬
bose, 6—7 mm diam.; capsule enclosed by dark pur¬
ple fleshy calyx; fruit stalk 4-9 mm long, glabrous.
Fruiting November-December.

Distribution and ecology. China (southeastern
Xizang). On rocky slopes near the summit of moun¬
tain; 2000-3400 m.

Gaultheria purpurea is readily distinguished from
the other Chinese species of the genus by having
purple, hirsute leaves, terminal or axillary inflores¬
cences on branchlet apices, and peduncles 3. 5—4. 5
cm long.

Paratypes. CHINA. Xizang: Medog, H. S. Li & S. Z.
Cheng 1862 (KUN. PE).

Gaultheria straminea R. C. Fang, nom. nov. Re¬
placed name; Gaultheria wardii Marquand &
Airy-Shaw var. serrulata C. Y. Wu & T. Z. Hsu,
FI. Xizang. 3: 699. 1986; not G. serrulata Her¬
zog, Meded. Herb. Leiden 27: 19. 1915; not
G. serrulata Danguy & Chernt., Bull. Mus.
Hist. Nat. Paris 28: 435. 1922. TYPE: China.
Xizang: Medog Xian, 1700 in, Qinghai-Xizang
Complex Exped. 73-890 (holotype, KUN). Fig¬
ure 7.

Shrubs, 1—1.5 m tall. Stems brown, terete;
branchlets angular, rather slender, densely brown-
hispid and puberulent. Leaves scattered; petiole 4—
5 mm long, hispid, puberulent or glabrous; leaf
blades oblong-lanceolate or elliptic, 4—7 X 1-2.5
cm, coriaceous, base cuneate or broadly cuneate,
margin dentate, apex acuminate or acute, mucron¬
ulate, adaxially drying gray-green to dark green,
glabrous, abaxially stramineous, appressed setu¬
lose; secondary veins 3 or 4 pairs, ultimate veinlets
slightly distinct. Inflorescences racemose, axillary
and terminal, 1.5— 3.5 cm long; rachises densely
pubescent; bracts ovate, 3^4 mm long, both surfac¬
es puberulent, ciliolate. Pedicels 1-2 mm long,
densely pubescent; bracteoles 2, 2-3 mm long,
close but not touching calyx, abaxially puberulent,
margins membranous, ciliolate, adaxially glabrous.
Calyx ca. 3 mm long, both surfaces puberulent;
lobes subulate-triangular, ca. 2 mm long, ciliolate.
Corolla white, urceolate, 3—4 mm long, glabrous
outside, pilose inside; lobes very short, ovate-tri¬
angular. Stamens 10, ca. 1.5 mm long; filaments
flattened, spindle-shaped, papillose; thecae short 2-
aristate. Ovary tomentose; style glabrous. F ruit glo¬
bose, 4—5 mm diam.; capsules enclosed by white,
rugose calyx. Flowering August— October, fruiting
October-March.

Distribution and ecology. China, Xizang. Flood
plains, among Salix scrub forests, evergreen forest
margins or on slopes; 600-2100 m.

Specimens examined. CHINA. Xizang: Medog Xian,
Qinghai-Xizang Complex Exped. 1682 (KUN, PE), Qing¬
hai-Xizang Complex Exped. 74-1712 (KUN, PE), Qinghai-
Xizang Complex Exped. 74-4535 (KUN, PE), H. S. Li X
S. Z. Cheng 973, 1471, 1788, 1907. 2261. 2991, 3125,
3527 (KUN, PE), H. Sun, Z. A. Zhou X 11. Y. Yu 221,

172

Novon

Figure 6. Gaultheria purpurea K. C. Fang. — A. Habit. — B. Portion of stem. — C. Leaf abaxial surface. — D. Young
capsule with persistent calyx and bracteoles Scale bars: A, C = 1 cm. B = 4 mm. D = 1 mm. Drawn from holotype
(B. S. Li & S. Z. Cheng 1994).

Volume 9, Number 2
1999

Fang

New Ericaceae from China

173

Figure 7. Gaultheria straminea R. C. Fang. — A. Habit. — B. Abaxial leaf surface. — C. Portion of stem. — D. Flower
ith bract and bracteoles. — E. Opened calyx. — F. Bract. — G. Stamen ventral view. Scale bars: A, B = 1 cm, C =
mm. D-G = I mm. Drawn from Li & Cheng 973 (KUN).

174

Novon

230, 1832, 2335, 2553, 2787, 4203, 4242, 4872, 6070
(KUN).

Gaultheria straminea is closely related to the
Bhutanese G. discolor from which it differs in hav¬
ing densely brown-hispid and puberulent branch-
lets, abaxially appressed setulose leal blades, 3 or
4 secondary veins on each side of the midrib,
densely pubescent raehises and pedicels, and pa¬
pillose filaments. It can he distinguished from the
related G. longibracteolata by having the bracteoles
at some distance from instead of adjacent to the
calyx, subulate-triangular instead of oblong-ovate
calyx lobes, glabrous instead of pilose styles, and
capsules enclosed by a white and rugose instead of
crimson, purplish green, or purple-black glaucous
calyx.

Gaultheria trigonoelada R. C. Fang, sp. nov.
TYPE: China. Xizang: Medog Xian, 2300 m,
26 Oct. 1992, //. Sun, Z. K. Zhou & H. Y. Yu
550 (holotype, KUN; isotype, KUN). Figure 8.

Species valde iusignis ramulis trigonis usque ad alatos,
glabris, foliis rhombico-ellipticis vel ellipticis, nervis la-
teralibus 2-jugis, inflorescentiis foliis multo brevibus, flo-
ribus densis, bracteis et braoteolatis carinatis, 1 costis om¬
nibus.

Shrubs, 1-2 m tall. Stems brownish, slightly 3-
angled, branchlets distinctly 3-angled or with nar¬
row wings, glabrous. Petioles 5-10 mm long, gla¬
brous; leal blades elliptic or rhombic-elliptic, 4—
8.5 X 2.5^4 cm, thick coriaceous, base cuneate,
margin dentate, apex acute to acuminate, adaxially
glabrous, abaxially sparsely shortly brown glandu¬
lar-setose; secondary veins 2 on each side of mid¬
rib, ultimate veinlets subparallel, impressed adax¬
ially, raised abaxially. Inflorescences racemose,
axillary, 2—3 cm long; rachis angular, densely pu¬
bescent; bracts ovate-lanceolate, carinate, 5—6 mm
long, coriaceous, distinctly 1-costate, both surfaces
glabrous or puberulent abaxially. Pedicels 2-4 mm
long, pubescent; bracteoles 2, similar to bracts,
ovate, ca. 3 mm long, close to calyx. Calyx 2.5-3
mm long, divided to near base; lobes triangular-
lanceolate, glabrous. Corolla while, urceolate, ca. 5
mm long, glabrous outside, pilose inside; lobes tri¬
angular-ovate, ca. 0.5 mm long. Stamens 10, ca. 2
mm long; filaments flattened, dilated below middle,
papillose; thecae with 2 slender aristae. Ovary to-
mentose; styles as long as stamens. Capsules glo¬
bose, 4—6 mm diam., enclosed by a purple calyx.
Flowering October, fruiting June— August.

Distribution and ecology. China, southeastern
Xizang. Forests, river terraces among thickets;
2000-2300 m.

Gaultheria trigonoclada is readily distinguished
from the other Chinese species of the genus by its
trigonous or winged, glabrous branchlets, rhombic-
elliptic or elliptic leaves with 2 secondary veins on
each side of the midrib, short atul densely flowered
inflorescences, and carinate, 1 -costate bracts and
bracteoles.

Paratypes. CHINA. Xizang: Medog. 2300 m, 3 Aug.
1974. Qinghai- Xizang Complex Exped. 74-3055 (KUN.
PE), 8. S. Li & S. Z. Cheng 5081 (KUN. PE).

Gaultheria wardii Marquand & Airy-Shaw var.
elongata R. C. Fang, var. nov. TYPE: China.
Yunnan: Gongshan Xian, 2000 m, 28 July
1982, Qinghai-Xizang Complex Exped. 8858
(holotype, KUN; isotypes, KUN, PE).

A var. ivardii rhachidi elongata, floribus vel fructibus
distanlibus in rhaehidi, rhachidi visibili, ealycibus glabris
extra, bracteis et bracteolis subglabris differt.

Gaultheria wardii var. elongata differs from va¬
riety wardii by its elongated raehises, flowers or
fruits distant and the raehises distinctly visible, and
calyx, bracts, anti bracteoles subglabrous abaxially.
By contrast, variety wardii has shortened raehises,
crowded flowers and fruits and the raehises invis¬
ible, and calyx, bracts, and bracteoles densely se¬
riceous or velutinous abaxially.

Paratypes. CHINA. Yunnan: Gongshan Xian, T. T.
Yii 21060 (KUN), Qinghai-Xizang Complex Exped. 8925
(KUN, PE); Dulongjian Valley, Dulongjian Hot. Exped.
826, 1087, 1368, 1441, 1608, 1941, 3210, 4255, 4788,
4955, 5082, 5160, 6981 (KUN).

Vaceiniuin rubescens R. C. Fang, sp. nov. TYPE:
China. W Yunnan: Ximeng Xian, 2000 m, 8
Mar. 1958, E C. Du D580156 (holotype, KUN;
isotype, KUN). Figure 9.

Species nova V. ardisioidi J. I). Hooker & C. B. Clarke
valde affinis, sed foliis oblongo-lanceolatis vel oblongo-
ovatis, pedicellis rubris et clavatis, corollis rubris, tubu-
losis, filamentis pubescentibus, antheris calcaribus valde
brevibus differt.

Evergreen shrubs, almost glabrous, terrestrial or
epiphytic, 1-2 m tall. Branchlets terete. Leaves 3-
8 in pseudoverticils, sessile; leaf blades oblong-lan¬
ceolate or oblong-ovate, 7.5—12 X 2-4 cm, char-
taceous, base narrowly obtuse-rotund, subcordate,
margin entire, apex acuminate; secondary veins
slender, more than 15 on each side of midrib, mid¬
rib raised on both surfaces. Inflorescences race¬
mose, 2- or 3-umbelled in whorl of leaves, 7—12
cm long, flowers numerous, secund; raehises red,
angulate-winged; bracts and bracteoles wanting.
Pedicels red, davate, ca. 1.5 cm long, thickened

Volume 9, Number 2
1999

Fang

New Ericaceae from China

175

Figure 8. Gaultheria trigonoclada R. C. Fang. — A. Habit. — B. Portion of stem. — C. Inflorescence in flower showing
bracts. — D. Flower with bract and bracteoles. — K. Stamen ventral view. — F. Stamen dorsal view. — G. Capsule
enclosed by calyx. Scale bars: A— C = 1 cm. D— G = 1 mm. Drawn from holotype ( H . Sun, Z. K. Zhou & H. Y. Yu
550).

176

Novon

Figure 9. Vaccinium rubescens R. C. Fang. — A. Habit. — B. Flower. — C. Opened corolla showing stamens. — D.
Berry. — E. Stamen ventral view. — F. Stamen dorsal view. Scale bars: A = 1 cm. B, C, E = 1 mm. D = 2.5 mm. A-
C, E. E drawn from holotype (E C. Du D580156), I) drawn from Li 4988 (KUN).

Volume 9, Number 2
1999

Fang

New Ericaceae from China

177

Figure 10. Vaccinium tenuifiorum R. C. Fang. — A. Habit. — IF Flower bud. — C. Opened corolla showing stamens.
— I). Stamen ventral view. — E. Stamen dorsal view. Scale bars: A = I cm. B— E = 1 mm. Drawn from holotype (//.
Sun, /. k. /hou & H. Vi Yu 2bl7).

178

Novon

upward, distinctly articulated with calyx. Calyx
tube shortly terete, ea. 4 mm long, upper parts tree,
not adnate to ovary; teeth triangular, ca. 1 mm long.
Corolla orange to pink, tubular, 7-8 mm long; teeth
less than 1 mm long. Stamens 10, 5—7 mm long;
filaments 2-4 mm long, pubescent especially at
apex; anthers spinulose, with 2 short spurs; tubules
as long as thecae. Berry globose, scarlet, ca. 5 mm
diam.; fruit stalk scarlet, ca. 2 cm long. Flowering
March, fruiting May.

Distribution and ecology. China, western Yun¬
nan. Dense forests, roadsides, moist places; 2000—
2200 tn.

Vaccinium rubescens is most closely related to V.
ardisioides J. D. Hooker ex C. B. Clarke from which
it can be distinguished by the oblong-lanceolate or
oblong-ovate instead of the ovate or elliptic leaves,
red and elavate instead of purple and slender ped¬
icels, orange or pink and tubular instead of white-
green and urceolate-tubular corollas, pubescent in¬
stead of subglabrous filaments, and shortly
2-spurred instead of spurless anthers.

Paratypes. CHINA. SVi Yunnan: I.uxi Xian. //. T.
Tsai 56424 (KUN); Gengma Xian. P. Y. Mao 5574 (KUN).
Y. H. Li 4983 (kUN).

Vaccinium tenuiflorum R. C. Fang, sp. nov.
TYPE: China. SE Xizang: Medog Xian, 1800
m, 26 Dec. 1992, II. Sun , Z. K. Zhou & //. Y.
Yu 2617 (holotype, KUN; isotype, KUN). Fig¬
ure 10.

Species nova V. vacciniacei (Roxburgh) Sleumer valde
affinis, sed foliis coriaceis non chartaceis, 4—1 1 X 0.9-2
cm, pedicellis usque ad tubos calycum dense pubescen-
tibus, corollis albis, anguste-tubulosis differt.

Evergreen shrubs or climbers. Branches sca¬
brous, juvenile ones pubescent and shortly glan¬
dular-setose, with scattered laneeolate leaf bud
scales. Leaves 9 or 10 in pseudoverticils; petioles
1-2 mm long, juvenile pubescent; leaf blades lan¬
ceolate to oblong-lanceolate, 4-11 X 0.9-2 cm, co¬
riaceous, base narrowly cuneate or rotund, margin
sparsely serrate, recurved, apex acuminate, adaxi-
ally drying dull green, abaxially olive green, both
surfaces glabrous, juvenile puberulent along mid¬
rib, secondary veins 12 or 13 on each side of mid¬
rib. Inflorescences from top of branches, axillary,
racemose, 5—6 cm long; rachises slender, densely
pubescent, with numerous lanceolate flower bud

scales at base; bracts lanceolate, 2-3 mm long, pu¬
berulent, ciliolate. Pedicels slender, 5-6 mm long,
densely pubescent, slightly thickened upward, ar¬
ticulated with calyx; bracteoles 2, linear, ca. 1 mm
long. Calyx 2-2.8 cm long, tube densely pubescent;
teeth 1-1.8 mm long, triangular, pubescent or gla¬
brous, 1-veined. Corolla white, narrowly tubular, 4 —
5 mm long, both surfaces glabrous except 5
pubescent ridges outside; lobes triangular, ca. 0.6
mm long. Stamens 10, shorter than corolla; fila¬
ments 0.8—1 mm long, flattened, widened at base,
glabrous; anthers spinulose, without spurs, tubules
longer than thecae. Berry subglobose, ca. 3 mm
diam. Flowering December, fruiting April.

Distribution and ecology. China, southeastern
Xizang. In forest, ca. 1800 m.

Vaccinium tenuiflorum is closely related to V.
vacciniaceum (Roxburgh) Sleumer, from which it
can be distinguished by having coriaceous leaves,
densely pubescent pedicels and calyx tubes, and
white and narrowly tubular corollas. Vaccinium vac¬
ciniaceum has chartaceous leaves, glabrous pedi¬
cels and calyx tube, and greenish yellow urceolate
corollas.

Paratype. CHINA. Xizang: Medog Xian. //. Sun, X.
K. Zhou & //. Y. Yu 5142 ( KUN).

Acknowledgments. I am grateful to the editorial
committee of the Flora of China for supporting my
visit to the United States. I am thankful to the staff
of the Missouri Botanical Garden and Harvard Uni¬
versity Herbaria for their help during my visits, es¬
pecially to D. E. Boufford, Hu Shiu-ying, and E.
W. Wood. I am grateful to Ihsan Al-Shehbaz for his
help with the manuscript. I thank the keeper of the
Royal Botanic Gardens, Kew, for the loan of spec¬
imens. Thanks are also extended to my colleague
Hsu Ting-zhi for his leading work on most of the
genera dealt with in this paper. My special grati¬
tude is expressed to P. F. Stevens of Harvard Uni¬
versity for providing much needed literature and for
sharing his views on various groups. Wu Xi-lin is
thanked for his outstanding illustrations.

I.iteruture Cited

tang, R. C. (Editor). 1991. Ericaceae. FI. Reipubl. Po-

pularis Sin. 57(3): 1-164.

Hsu, T. Z. 1991. Leucothoe. In: R. C. Fang (editor), FI.

Reipubl. Popularis Sin. 57(3): 19-22.

Li, II. 1993. Flora of Dulongjiang Region. Yunnan Science

and Technology Press, Kunming.

Asclepias jorgeana (Asclepiadaceae), a New Milkweed from
Montane Western Mexico

Mark Fishbein

Herbarium and Department of Ecology and Evolutionary Biology, University of Arizona,
Tucson, Arizona 85721, U.S.A. Current Address: Department of Botany,
Washington State University, Pullman, Washington 99164-4238, U.S.A.

Steven P. Lynch

Department of Biological Sciences, Louisiana State University, One University Place,
Shreveport, Louisiana 71115, U.S.A.

Abstract. A new milkweed, Asclepias jorgeana ,
from pine-oak forests of Chihuahua, Sonora, Du¬
rango, and Guanajuato, Mexico, is described and
illustrated. It appears to he related to A. pringlei
(Greenman) Woodson, which occurs mostly to the
south and east of A. jorgeana ; however, they occur
sympatrically in Durango. The new species is dis¬
tinguished from A. pringlei by attenuate corona
hoods, erect, fusiform follicles, and comose seeds.
Although the new species has been collected sev¬
eral times since 1946, specimens have remained
unidentified or have been determined to be hybrids
or an undescribed species.

In preparing a treatment of Asclepiadaceae for
the revision of Howard Scott Gentry's Rio Mayo
Plants (Martin et ah, 1998), the first author exam¬
ined several collections of Asclepias that could not
be assigned to any known species. Contemporane¬
ously, the second author also determined that sev¬
eral herbarium specimens under study from Duran¬
go represented an undescribed species of Asclepias.
Independent study of this plant in the field by both
authors and subsequent discussions confirmed the
identity and distinctiveness of these populations,
which we describe here as a new species.

Asclepias jorgeana Fishbein & S. P. Lynch, sp.
nov. TYPE: Mexico. Chihuahua: Caseada de
Basaseachic, 28°10'20"N, 108°13'W, 1950 m,
approx. 100 m upstream from the top of the
falls, above the trail, on steep, shaded, moist,
SE-facing, roekv slope, 5 July 1994, M. Fish¬
bein 1789, S. McMahon, G. Ferguson , R. Fel-
ger, M. Wilson (holotype, ARIZ; isotype,
MEXU). Figure 1A-H.

Ilerha e radice fibrosa. Cawlis simplex credos. Folia
opposita petiolata parce puberula: lamina elliptico-ovali

vel ovata. maxima 45—123 mm longa, 15—61 mm lata;
petiolo 5—16 mm longo. Umbellae 1-5 extra-axillares vel
ut videtur terminales, 7— 26-florales: pedunculo 10—40 mm
longo; pedicello 12.5—22.5 mm longo. Calyx viridis; lobis
reflexis, 1.75 mm longis. Corolla viridis saepe obscure ro-
seolo-brunnea sulfusa; lobis reflexis, 4.75—6.5 mm longis.
Corona e 5 cucullis pallido-lavandulis vel albis, stipitatis
fere erectis, 3.75—4.75 mm longis composita. Corniculi ad
medium cucullis adnati, arcuati exserti. Columna 0.8—
1.25 mm longa. Colliculus erectus in pedicello reflexo,
fusiformis. Semina ovalia comosa.

Perennial herb 20—50 cm tall from fibrous root-
stock. Stem single from base, unbranched, puber-
ulent in lines or hairs scattered around the stem.
Leaves opposite, decussate, simple; lowest pair
17.5—62.5 mm long and 10—27.5 mm wide, the
largest 45—123 mm long and 15—61 mm wide, be¬
coming reduced and subsessile in the inflores¬
cence; blades elliptic-oval to ovate, the apex acute,
the margin entire, ciliate, decurrent on the petiole
(especially pronounced in southern populations;
Fig. 1A), the base acute to rounded, appearing sub-
cordate in southern populations (Fig. 1A), puber-
ulent on veins below, sparsely pubescent through¬
out above, the venation reticulodromous, with 3—10
tan to dark brown, glandular, multicellular tri-
chomes (“colleters”) borne at the base of the mid¬
vein above, membranous, bright green above, paler
below; petioles 5—16 mm; stipules absent, several
glands (colleters) similar to those on the blade in
their position. Inflorescence of 1-5 umbels adjacent
to upper leaf axils and also apparently terminal, on
peduncles 10—40 mm long; umbels of 7—26 flowers
on puberulent pedicels 12.5-22.5 mm long. Flow¬
ers hermaphroditic, 10—12.5 mm long from tip of
reflexed corolla lobe to corona tip; calyx synsepal-
ous, green, the tube very short, the 5 reflexed lobes
1.75 mm long; corolla sympetalous, green often suf¬
fused with dull pinkish brown, the tube short, the
5 reflexed lobes 4.75—6.5 mm long; stamens 5, ad-

Novon 9: 179-184. 1999.

180

Novon

cm

cm

Hgure 1. Aselepias jorgeana Fishbein & S. P. Lynch. — A. Habit, showing broad leaves characteristic of Durango
populations. — B. Roots. — (.. Flower. — D. Corona segment (hood with protruding horn), lateral view. — E. Pollinarium.
— F. Narrow leaf characteristic of northern populations. — G. Follicle. — H. Seed. Based on Lynch 14175 (I.SUS),
Fishbein I7HV (AH1Z, MEXU), and photos.

Volume 9, Number 2
1999

Fishbein & Lynch

Asclepias jorgeana from Mexico

181

nate to the common style apex of the two carpels
to form a gynostegium, the column composed of
connate stamen filaments 0.8-1.25 mm long, cylin-
dric or slightly obconic; corona of 5 distinct seg¬
ments (hoods) arising from the top of the column at
the attachment of the anthers, the hoods nearly
erect to spreading, stipitate, cup-shaped, dorsally
flattened, with entire margin, truncate basally, ob¬
tuse and rounded apieally, 3.75—4.75 mm long, sur¬
passing the anther-style head, pale lavender to
white, bearing a narrowly falcate appendage (horn)
within, the horns adnate to hoods from the base to
midway up the segment and abruptly curving over
the anther-style head, sometimes geniculate near
the point of egress from the hood; anther-style head
narrowly cylindric, 2.25—2.5 mm long, the anthers
bithecous with one fertile cell per theca, bearing
corneous marginal appendages (wings) widest at
base and scarious, ovate-lanceolate, apical append¬
ages, 0.5 mm long, pollen grains in each anther cell
embedded in a waxy matrix forming a pollinium;
pollinia lacrimiform, 1.25 mm long, those of adja¬
cent anthers united by a translator with arms 0.3
mm long; gynoecium of 2 carpels, the ovaries free,
the styles connate their entire length, abruptly ex¬
panded apieally and fused to anthers, the stigmas
5, borne below the style apex. Fruits follicular, 8
cm long, fusiform, apieally acuminate, borne erect
on deflexed pedicels. Seeds narrowly ovate, com¬
pressed, narrowly winged, 6.2— 6.5 mm long, bear¬
ing an apical tuft of bright white hairs 2. 5-4. 5 cm
long.

Etymology. This species is named in honor of
George Ferguson, a broadly knowledgeable natu¬
ralist who has introduced the first author to many
fine places in the Rio Mayo region. He has col¬
lected several specimens of the new species, which,
like George, is at home in the wild, rugged canyons
of the pine-oak forest.

Distribution and habitat. Asclepias jorgeana is
known from west-central Chihuahua and adjacent
Sonora, western Durango (prominently along the
Durango-Mazatlan highway), and as a disjunct in
eastern Guanajuato (Fig. 2). In Chihuahua and So¬
nora, populations of the new species have been
found from ca. 1825—2250 m in pine-oak forest and
riparian forest dominated by Pinus spp., Quercus
spp., Cupressus, Abies , and Alnus. At these sites, A.
jorgeana occurs mainly in the shade of canopy trees
in steep ravines and in canyon bottoms along pe¬
rennial streams. At Pinos Altos, Chihuahua, the
species was found on a disturbed hill with little
topsoil in sparse vegetation dominated by Junipe-
rus. This site is considerably more xeric than the

other known localities and is the site of the only
previously collected population in the vicinity of
Cascada de Basaseachic that could not be relocated
in 1994.

In Durango, populations of A. jorgeana have
been found from ca. 1700 to 2525 m in a variety
of vegetation types. At the lowest elevations, the
new species has been collected at the lower limit
of pine-oak forest at the ecotone with tropical de¬
ciduous forest or thorn scrub. Along streams at
these elevations, it has been found in pine-oak for¬
est with Abies. At middle elevations, it has been
found in pine-oak forest on canyon slopes and in
cloud forest with Magnolia. At the highest eleva¬
tions, A. jorgeana has been collected in meadows
within pine-oak forest and on steep slopes in mixed
conifer forest with Pinus , Abies, Cupressus, and
Pseudotsuga.

In Guanajuato, A. jorgeana has been collected
on a hillside at ca. 1835 m. The vegetation at this
locality was not reported by the collectors.

Ecology and phenology. Asclepias jorgeana is
unusual among North American milkweeds by in¬
habiting dense forest understories rather than more
typical open woodland and grassland sites. Other
species of Asclepias that are found in similar hab¬
itats include A. quadrifolia Jacquin and A. exaltata
L. in the eastern United States, A. similis Hemsley
in eastern and southern Mexico, A. hypoleuca (A.
Gray) Woodson (sympatric with A. jorgeana in Chi¬
huahua and Sonora), and A. ovata M. Martens &
Galeotti. Presumably these species are tolerant of
the shade and acidic soils characteristic of these
habitats. Unlike A. quadrifolia, A. exaltata, A. sim¬
ilis, and A. hypoleuca, which possess large storage
roots that may be advantageous to plants of the for-

182

Novon

est understory (Chaplin & Walker, 1982; M. Fish-
bein, pers. obs.), A. jorgeana has fibrous roots, like
A. ovata (M. Fishbein, pers. obs.). Despite the laek
of large roots, individual plants of A. jorgeana have
persisted for several years in cultivation (M. Fish¬
bein, unpublished data).

Based on label data, flowering dates for the new
species range from late June through August.
Southern populations in Durango appear to have a
more extended flowering season lasting through Au¬
gust, whereas the latest flowering specimens from
Chihuahua were collected on the first of August.
Fruiting specimens in Chihuahua and Durango
have been collected in mid August.

Several kinds of insects have been observed vis¬
iting the flowers of A. jorgeana in Chihuahua. Along
the Arroyo Durazno, small butterflies of the families
Hesperiidae and Lycaenidae were observed. The
flowers emitted a strong violet-like scent when
these observations were made on a warm sunny
morning; during all other field observations, how¬
ever, no scent could be detected. At several sites
in Chihuahua, tiny Diptera or Hymenoptera were
observed visiting flowers in great numbers. Among
observed visitors, hesperiid and lycaenid Lepidop-
tera were probably marginally large enough to be
effective pollinators of this species, although no
pollinia were observed on these visitors (cf. effec¬
tive pollinators of A. tuberosa L., which has flowers
comparable in size and shape to those of A. jor¬
geana [Fishbein & Venable, 1996]). We have not
observed flower visitors that we deem to be regular,
effective pollinators of this species. The small flies
were certainly too small to be pollinators, being ap¬
proximately equal in size to a pollinium.

In Durango, woolly caterpillars (Lepidoptera:
Arctiidae) and Queen butterfly caterpillars (Lepi¬
doptera: Danaiidae) were observed feeding exten¬
sively on leaves and young fruit. No herbivory was
observed at any of the sites in Chihuahua.

Phylogenetic relationships. Asclepias jorgeana
has stipitate, cup-shaped corona segments that are
similar to those of species included by Woodson in
subgenus Asclepias (Woodson, 1941b, 1954). With¬
in this subgenus, A. jorgeana is most similar to spe¬
cies assigned by Woodson (1954) to the heteroge¬
neous series Syriacae, Purpurascentes, and
Tuberosae. Among these, A. jorgeana is most similar
to A. pringlei (Greenman) Woodson of series Syr¬
iacae. The two species are nearly identical in veg¬
etative characteristics, and we have been able to
determine only with great difficulty the identity of
non-reproductive specimens from areas where the
two species co-occur. However, the species are eas¬
ily distinguished in flower or fruit. Flowers of A.

pringlei have smaller corona segments (hoods) with
less attenuate and more rounded apices. The ad-
axial hood appendages (horns) of A. pringlei are
less arching than those of A. jorgeana or may be
absent altogether. Arching corona horns are present
in all specimens of A. jorgeana that we have ex¬
amined. Fruits of A. pringlei are distinctive: they
are pendulous, ovoid, and have a thick spongy layer
in the wall of the follicle. The seeds are also un¬
usual in lacking the apical triehomes (coma) com¬
mon to most species in the family (cf. A. perennis
Walter of the southeastern U.S.). Asclepias jorgeana
produces fruits that are more typical of those of
other North American species of Asclepias : erect,
relatively slender with a long, tapering apex, and
thin-walled. The seeds possess a typical coma.

Other species with which A. jorgeana shares sim¬
ilar characteristics include A. variegata L. (eastern
U.S.) and A. hallii A. Gray (southwestern U.S.) of
series Purpurascentes and A. tuberosa (widespread
in the U.S.) of series Tuberosae. Asclepias variegata
is similar vegetatively, but differs in having com¬
pact hemispherical umbels and small hoods with
rounded apices reminiscent of those of A. pringlei.
Asclepias hallii is quite similar in floral morphology
and coloration, but differs from A. jorgeana in bear¬
ing alternately arranged, narrowly lanceolate
leaves. Asclepias tuberosa is also similar in floral
morphology, but differs in having brightly colored
yellow, orange, or red flowers, branched inflores¬
cences, alternate phyllotaxy, lack of white latex,
and densely hirsute vestiture.

Preliminary phylogenetic study of Asclepias
based on a broad sampling of morphological char¬
acters suggests that A. jorgeana is a member of a
large clade that includes the majority of species
comprising subgenus Asclepias (Fishbein, 1996a, b,
unpublished). Relationships within this large clade
are poorly resolved in the analyses of Fishbein
(1996a, b, unpublished), with members of series
Syriacae , Purpurascentes, and Tuberosae intermin¬
gled. In the set of most parsimonious trees, A. jor¬
geana is usually found to be a member of a sub-
clade containing A. variegata, A. hallii, and A.
tuberosa. Placement of A. jorgeana in this subclade
is supported by possession of an attenuate hood
apex and appressed anther margins that tightly re¬
strict access to the underlying stigmatic chamber.
In some analyses, A. pringlei is found to be a mem¬
ber of a closely related subclade, and placement of
A. jorgeana in the same subclade as A. pringlei is
only slightly less parsimonious than placement in
separate subclades (Fishbein, unpublished). The
highly apomorphie fruit characters that readily dis¬
tinguish A. pringlei from A. jorgeana do not link A.

Volume 9, Number 2
1999

Fishbein & Lynch

Asclepias jorgeana from Mexico

183

pringlei with any other species. These characters
clearly have arisen independently in other lineages,
such as pendulous fruits in the desert shrub .4. sub-
ulata Decaisne, and seeds lacking a coma in .4.
perennis. We consider it likely that more refined
analyses of A. jorgeana and related species will find
a closer relationship between A. jorgeana and A.
pringlei.

Specimens of Asclepias jorgeana have been iden¬
tified previously as A. ovata, which occurs sympat-
rieally, hut often at lower elevations and in more
open habitats (e.g., Spellenberg 9603, NMC; Gentry
& Gilly 10615, MICH; LeDoux 1947, ENCB; Bravo
1076, IEB). The two species share an erect, un¬
branched habit with sparse pubescence, ovate
membranous leaves, fruits borne erect on deflexed
pedicels, and fibrous roots. Compared to A. jor¬
geana, the umbels of A. ovata are nearly spherical,
rather than flat-topped to nearly hemispherical, and
the corona segments are bright white and apieally
truncate, rather than pale lavender and apieally at¬
tenuate. Specimens of A. jorgeana have also been
identified as A. contrayerba Sesse & Mogino (= A.
jaliscana B. L. Robinson; Maysilles 7434, MO), A.
glaucescens HBK (Warnoch 2156, TEX 2 sheets),
and A. lanuginosa HBK. non Nuttall (= A. otarioi-
des E. Fournier; Waterfall 13654, MO). Asclepias
jorgeana has not been confused with A. pringlei by
collectors, despite vegetative similarities, probably
because they are easily distinguished in flower and
fruit. Some specimens of .4. jorgeana have been de¬
termined to represent an undescribed species (e.g.,
Pennington 93, TEX; Tenorio 1150, ENCB 2
sheets, MEXL ).

Specimens attributable to A. jorgeana also have
been considered to represent hybrids. A specimen
of A. jorgeana examined by Woodson ( Maysilles
7434, MICH) was annotated initially as a putative
hybrid between A. otarioides E. Fournier (as A. lan¬
uginosa HBK) and A. pringlei (as A. humilis (sic),
i.e., Acerates humilis Bentham; see Woodson, 1954:
123). Woodson later determined specimens of A.
jorgeana to be hybrids between A. otarioides and A.
ovata (e.g., Maysilles 7852, MICH; Maysilles
7860A, MICH; Maysilles 8477, MICH). The simi¬
larities of A. jorgeana to A. pringlei and A. ovata
already have been discussed. However, we find no
compelling reason to invoke the parentage of the
rather dissimilar A. otarioides in the origin of A.
jorgeana. No evidence of intermediacy exists in lo¬
cations where A. jorgeana is sympatric with A. otar¬
ioides and A. pringlei in Durango or with A. ovata
throughout its range. We cannot disprove the hybrid
origin of A. jorgeana with existing data. However,
the species is (dearly composed of self-reproducing

populations with no evident gene flow from the pu¬
tative parents.

Of other specimens that Woodson (1954: 123)
suggested to be part of hybrid swarms between A.
otarioides and either A. ovata or A. pringlei, some
have been recognized as representing a distinct
species, namely, A. notha W. D. Stevens (Stevens,
1983, 1985). Others should perhaps be recognized
as an additional species endemic to Michoacan,
viz., A. scheryi Woodson (Woodson, 1941a). Wood-
son (1954) subsequently included these specimens
in his concept of A. lanuginosa HBK (= A. otar¬
ioides). Most of the remaining specimens consid¬
ered by Woodson (1954) to belong to this complex
of putative hybrids were collected in Puebla and
Oaxaca and appear to represent yet another species
(S. Lynch, unpublished data). Although it may be
surprising that a new species of Asclepias should be
described from North America, it is evident that
several more will be named from relatively poorly
collected areas in the rugged Mexican sierras.

Paratypes. MEXICO. Chihuahua: Cascada de Basa-
sedchic, 28°10'20"N, 108°13'W, 2000 m. top of falls, 23
July 1986, P. Jenkins s.n. (ARIZ); 28°11'W. 108°12'30"N,
in sand along the Rio Basaseachic between the camp¬
ground and the top of the falls, I Aug. 1988. 8. Spellen¬
berg 9603, 8. Corral, J. Brunt, L. Huenneke (NMC); Pinos
Altos, on trail from Santo Nino shaft to lower tunnel,
28°15'N, 108°17'W. 2075 m, steep moist, W-facing an¬
desitic cliffs, 21 July 1946. 1C P. Hewitt, 144 (TEX);
28°16'20"N, 108° 18' 10"W, 2250 m, 26 July 1986. G. Fer¬
guson s.n. (ARIZ); 1.5 mi. NW of Pinos Altos along road
to La Bateria. 28°16'40"N, 1 08° 1 8'40"W . 2100 m. tall
pine-oak forest, 27 June 1986, G. Ferguson s.n. (ARIZ);
2.0 mi. NW of Pinos Altos along road to La Baterfa,
28°16'40"N. 108°18'40"W, 2025 m, steep, rocky. N-facing
ravine in pine-oak forest, 6 July 1994, M. Fishbein 1804,

S. McMahon, G. Ferguson, 8. Felger, M. Wilson (ARIZ),
19 Aug. 1994, M. Wilson s.n. (ARIZ); Arroyo Durazno at
I ,a Bateria road crossing, 28CT7'30”N. 108°20'W, 1730 m,
fir-alder riparian forest, 28 June 1986, G. Ferguson s.n.
(ARIZ, MO); near river in shade of oaks and pines, 27
July 1988. P. Jenkins s.n. (ARIZ); Arroyo Durazno, approx.
2 km upstream from La Bateria road crossing, 28°17'N,
108°20'W, 1730 m, riparian forest, 7 July 1994, M. Fish¬
bein 1826, S. McMahon, G. Ferguson, 8. Felger, M. Wilson
(ARIZ, MO); Lagotera, 1525-1825 m. 21 July 1965. C.
1C Pennington, 93 (TEX). Durango: Coyotes Hacienda,
63 road mi. W-SW of Cd. Durango, 2400-2500 m, open
meadows in pine forests, 26 July 1950, J. H. Maysilles,
7434 (MICH, MO), 16 July 1955, J. H. Maysilles,' 7852,
7860A (MICH), 10 Aug. 1955, ./. II. Maysilles, 8477
(MICH); just SW of Coyotes Hacienda, slight NE-facing
slope, edges of pine woods, 25 Aug. 1980, M. J. Warnock
2156, S. McCormick. J. Norris (TEX 2 sheets); 10-12 mi.
W of La Ciudad (38—40 mi. W of El Salto); Mazatldn—
Durango Rd, 2525 m, steep, moist canyon slope in pine-
oak forest, 15 June 1951, //. S. Gentry & C. L Gilly,
10615 (ARIZ, MICH); 5.5 mi. E of El Salto, about 55 mi.
SW of Durango, open pine-oak woods, 12 Aug. 1957. U.

T. Waterfall 13654, C. S. Wallis (MO); 6.0 mi. NE of El

184

Novon

Salto, along llwy. 40, 2300 m, pine-oak forest, 10 Aug.
1994, S. P. Lynch 141 75 (LSUS); lull overlooking town of
Llano Grande, 32 mi. W of Durango, on sandy loam, 22
July 1975, D. LeDoux & Dunn 1947 (ENCB, LL, MO.
NY); Mpio. El Salto, “El Capulin,” al 4 km noroeste de
El Pueblo de 1010, bosque pino-encino, suelo negro for-
estal, 28 jun. 1982, P. Tenorio L. 716, C. Romero de T.
(MEXU); 2 km al oeste de Revolcaderos, Carr. Mazatlan-
Durango, 2100 m, canadas escarpadas, bosque mesofilo
de montana, 26 jul. 1989, A. Vazquez 4905, R. Alvarez
(IEB, I.SUS, Vt IS); Mpio. Tepehuanes, Cerro Boludo, Si¬
erra de la Candela. 20 km al este de Tepehuanes, 1700
m, 21 jul. 1982, R. Hernandez M. Hill, et al. (MEXU,
TEX); 18 km al noroeste de Tepehuanes, bosque pino-
encino, suelo negro forestal, 21 jul. 1982, P. Tenorio L
1150, C. Romero de T. (ENCB 2 sheets, MEXU, MO); 9
km al este de Tabahueto, camino a Tepehuanes, ecotonfa
de selva baja caducifolia y bosque de pino-encino, alter-
ado, suelo pedregoso, 31 ago. 1983, R. Torres C. 3586, P.
Tenorio L, E. Torreeillas (MEXU, MO 3 sheets); Mpio.
Iayoltita, El Pino, 20 km de el entronque a Sapioris con
la Brecha Coyotes-San Miguel de Cruces, 24°31'N,
105°49'W, 181)0 m, ecotonfa bosque de pino-matorral es-
pinoso, 8 jul. 1984, P. Tenorio /.. 6 319, C. Romero de T..
T. P. Ramamoorthy (MEXU, MO); Mpio. Canelas, predio
particular Ijis Cebollitas, 1 ago. 1990, 0. llravo B. 1076
(IEB). Guanajuato: Mpio. Atarjea, Cueillos, 23(H) m, lad-
era de cerro, vegetacitfn boscosa, 14 jul. 1990, E. Ventura
A- E. Ldpez 8285 (IEB). Sonora: Mpio. Y^cora, El Divi-
sadero (El Bordo), 1 km SE of El Llano on road to Ber¬
mudez. Mesa del Campanero, 28°19'41"N. 109°02'38"W,
21(H) m, pine-oak forest. 14 July 1997, T. R. Van Devender
97-647, A. L. Reina, R. L. Belsey, I). Larson, M. J. Mar¬
tinez C„ P. Larson, A. Burquez A/.. M. Quintana (ARIZ).

Acknowledgments. We appreciate the helpful
comments of Eric Sundell, Ashley Nicholas, Rich
Spellenberg, and Shelley McMahon on this and ear¬
lier versions of the paper. H. Glen commented on
the Latin description. We are very grateful to Car¬
olyn Crawford for her careful and accurate work in
the preparation of Figure 1. We thank the curators

of the following herbaria for the loan of specimens
critical to this work: ENCB, F, GH, IEB, MEXU,
MICH, MO, NMC, NY, TEX-LL, US, WIS. M. Fish-
bein thanks Lucinda McDade and the staff of the
University of Arizona Herbarium for the handling
of loan specimens. This paper is dedicated to the
memory of Ed Bryant.

Literature Cited

Chaplin, S. J. & J. L. Walker. 1982. Energetic constraints
and adaptive significance of the floral display of a forest
milkweed. Ecology 63: 1857—1870.
fishbein. M. 1996a. Morphological analysis of infragener¬
ic relationships of Asclepias L. (Asclepiadaceae). Amer.
J. Rot. 83 (suppl.): 156 [abstract].

- . 1996b. Phylogenetic Relationships of North

American Asclepias L. and the Role of Pollinators in the
Evolution of the Milkweed Inflorescence. Ph.D. Disser¬
tation, University of Arizona, Tucson, Arizona.

- & I). L. Venable. 1996. Diversity and temporal

change in the effective pollinators of Asclepias tuberosa.
Ecology 77: 1061—1073.

Martin, P. S., I). Yetman, M. Fishbein, P. Jenkins, T. R.
Van Devender & R. k. Wilson (Editors). 1998. Gentry’s
Rio Mayo Plants. The Tropical Deciduous Forest and
Environs of Northwest Mexico. University of Arizona
Press, Tucson, Arizona.

Stevens, W. I). 1983. New species and names in Apoey-
naceae, Asclepiadoideae. Phytologia 53: 401-^105.

- . 1985. Asclepiadaceae. Pp. 228-241 in J. Rze-

dowski & G. C. de Rzedowski (editors). Flora Fanero-
gamica del Valle del Mexico. Escuela Nacional de Cien-
cias Biologicas, Instituto de Ecologfa, Mexico. D.F.
Woodson, R. E., Jr. 1941a. Miscellaneous new Asclepia¬
daceae and Apocynaceae from tropical America. Ann.
Missouri Rot. Card. 28: 271-286.

- . 1941b. The North American Asclepiadaceae. I.

Perspective of the genera. Ann. Missouri Bot. Gard. 28:
193-244.

- . 1954. The North American species of Asclepias

L. Ann. Missouri Bot. Gard. 41: 1-211.

Validation of the Name Cephalotaxus latifolia (Cephalotaxaceae), a

Species from Southeast China

Fu Li-kuo

Herbarium, Institute of Botany, Chinese Academy of Sciences, 20 Nanxincun, Xiangshan,
Beijing 100093, People’s Republic of China

Robert R. Mill

Royal Botanic Garden Edinburgh, 20a Inverleith Row, Edinburgh EH3 SLR,

Scotland, United Kingdom

Nicholas J. Turland

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A.

ABSTRACT. The name Cephalotaxus latifolia (W.
C. Cheng & L. K. Fu) L. K. Fu is an invalid com¬
bination because it was based on C. sinensis (Reh-
der & E. H. Wilson) H. L. Li var. latifolia W. C.
Cheng & L. K. Fu, which is itself invalid because
two specimens (male and female) were indicated as
types in the protologue. The name is here validated
at the rank of species, with the female specimen as
the holotype.

Cephalotaxus latifolia W. C. Cheng & L. K. Fu
ex L. K. Fu & R. R. Mill. sp. nov. TAPE:
China. Sichuan: Nanchuan County, Jinfe
Mountain, 1760 m (female), X. (A Chen & K.
Y. Liang 2463 (holotype, PE).

A Cephalataxo sinensi (Rehder & E. H. Wilson) H. E.
I j ramulis robustis et foliis latioribus, apice abrupte mu-
cronatis, margine recurvatis recedit. A C. harringtonii
(Knight ex K. Koch) J. Forbes foliis saepissime breviori-
bus (16—30 mm longis. non 25—50 mm) et ratione longi-
tudinis latitudini minore (5—7 : I. non 10—15 : I) basi cu-
neata (non truncata vel obtusa) et a speciininibus typicis
illius speciei apice folii abrupte mucronato tmicronc 0.2—
0.4 mm longo (non in cuspidem 0.5—1 mm gradatim an-
gustato), zonis stomatiferis etiam paulo angustioribus (0.8—
1 mm latis, non 1.1— 1.3 mm) recedit.

Shrubs or small trees to 5 m; bark grayish brown:
branchlets stout. Branchlets oblong itt outline,
plane, 5.5—9 X 4—5 cm. Leaves borne at (60°-)65°-
80°(— 85°) to branehlet axis, sessile but decurrent
for whole length of internode; blade dark olive
green abaxially, linear, ± straight, 1.6-3 cm X 2.8—
4(— 6) mm, thick and leathery, midvein raised abax¬
ially and bordered on each side by a paler, longi¬
tudinal channel, strongly raised adaxially, 0.25—0.5
mm wide, stomatal bands white, 0.8—1 mm wide,
of 11—15 rows of stomata, ca. 2— 3X as wide as

midvein, marginal bands green, ca. 0.3 mm wide,
base cuneate, slightly asymmetric, margin flat
(slightly revolute when dry), apex abruptly mucro-
nate (mucro 0.2— 0.4 mm). Pollen cone capitula ax¬
illary on adaxial side of branehlet axis; peduncle
1.5— 2.5 mm, scaly; bracts ca. 9, ovate, in 4 rows,
apical bracts ca. 0.7 X 0.3 mm, apex mucronate.
Seed cones borne 2-6 together; peduncle 2-4 mm,
scaly. Seed obovoid, 1.8-2 cm, apex with small mu¬
cro at center. Pollination May.

This species is endemic to China, occurring in
thickets in mountainous areas from 900 to 2400 m
in Fujian, Guangdong, Guangxi, Guizhou, Hubei,
Jiangxi, and Sichuan provinces.

Cephalotaxus latifolia was originally described
as C. sinensis var. latifolia by W. C. Cheng and L.
K. Fu (in Cheng et al., 1975: 86). However, that
name is invalid under the Tokyo Code (Arts. 8.1,
37.1) because two types (a male collection, J. H.
Xiong & Z. L. Zhou 90227, and a female collection,
X (A Chen & K. Y. Liang 2463) were designated.
The name was later raised to the rank of species
by L. K. Fu (Fu, 1984: 280), but no validation was
effected at that time either. The female collection
is here designated as the type specimen.

Cephalotaxus latifolia is more similar to C. har¬
ringtonii (Knight ex J. Forbes) K. Koch ["harring-
tonia" | (= C. drupacea Siebold & Zuccarini), from
Japan and Korea, than to C. sinensis. Cephalotaxus
latifolia has leaves shorter (16—30 mm), with a
smaller length : width ratio (5-7 : 1), stomatal
bands slightly narrower (0.8-1 mm wide), and apex
abruptly mucronate, with mucro 0.2— 0.4 mm. Ce¬
phalotaxus harringtonii has leaves longer (25—50
mm), with a larger length : width ratio (10-15 : 1),
stomatal bands slightly wider (1.1-1. 3 mm), and

Nov ON 9: 185-186. 1999.

186

Novon

apex gradually narrowed into a relatively long cusp
0.5—1 mm (as in C. sinensis). Cephalotaxus har-
ringtonii has often been recorded from China (e.g.,
by Hu, 1964: 25—26), hut apparently always as a
misidentification of either C. latifolia or C. sinensis.

Cephalotaxus sinensis has relatively longer leaves
(7-10X as long as wide vs. 5-7X as long as wide
in C. latifolia ), which are gradually acuminate into
a cusp that is 0.5-1 mm long. In C. latifolia the
leaf apex ends abruptly in a short mucro only 0.2-
0.4 mm long. In C. latifolia the pollen cone capit-
ula are borne on scaly (hracteate) peduncles; in C.
sinensis these are naked.

Acknowledgments. We thank Dai Lunkai (PE)
and Guanghua Zhu (MO) lor their assistance, and
an anonymous reviewer for helpful comments. The
Royal Botanic Garden Edinburgh is supported by
the Scottish Office of Agriculture, Environment and
Fisheries Department.

Literature Cited

Cheng, W. C., L. K. Fu & C. Y. Cheng. 1975. Gymnosper-
mae sinicae. Acta Phytotax. Sin. 13(4): 56-89.
lu. L. K. 1984. A study on the genus Cephalotaxus Sieb.

et Zucc. Acta Phytotax. Sin. 22: 277-288.

Hu, S. Y. 1964. Notes on the Flora of China IV. Taiwania
10: 13-62.

Validation of an Invalidly Described Species of Gnetum
(Gnetaceae) from Hainan Island, China

Fu Li-kuo and Yu Yong-fu

Institute of Botany, Chinese Academy of Sciences, 20 Nanxincum, Xiangshan, Beijing 100093,

People’s Republic of China

Michael G. Gilbert

Flora of China Project, Missouri Botanical Garden, c/o Department of Botany, Natural History
Museum, Cromwell Road, London SW7 5BD, United Kingdom

ABSTRACT. Gnetum hainanense, previously inval¬
idly described, is validated.

A nomenclatural problem was noticed during
work on the Gnetaceae for volume 4 of the Flora
of China. The original description of Gnetum hain¬
anense C. Y. Cheng (Cheng et al., 1975) is invalid
because two types (a male collection, McClure
20084, and a fertile collection, Tsang & Fung 554 )
were designated. It should be noted that the diag¬
nosis given in the original publication is of ques¬
tionable validity, as it gave no direct information on
the new species but only indicated how another,
known, species, Gnetum parvifolium (Warburg)
Chun, differed from it. The species is well defined
and moderately common and needs to be validated.
The fertile collection is here selected as ihe type
collection.

Gnetum cleistostachyum C. Y. Cheng was de¬
scribed in the same paper as G. hainanense, again
with male and female types, and thus this name is
also invalid. It was based on two collections at early
anthesis. We have seen very little material at a
comparable stage of development. What we have
seen suggests that the urceolate involucral collars
used to characterize G. cleistostachyum might be a
developmental stage that occurs more widely in the
genus, and thus are of questionable taxonomic val¬
ue. When this is coupled with the absence of seeds,
which provide the most reliable diagnostic char¬
acters for the group, we prefer not to take up this
name. Similarly, we do not intend to recognize any
infraspecific taxa within Gnetum pendulum . Thus,
we are not validating Gnetum pendulum f. inter¬
medium, also described with male and female
types. The question of the validity of an essentially
negative diagnosis also applies to Gnetum pendu¬
lum: again, the diagnosis gave no information on
the new species, stating only how G. oblongum
Markgraf differed from it. However, because all oth¬

er parts of the protologue are valid and a reasonable
illustration is provided, we regard that name as val¬
id.

Gnetum hainanense C. Y. Cheng ex Fu Li-kuo,
Yu Yong-fu & M. G. Gilbert, sp. nov. TYPE:
China. Hainan Island: Hung Mo Shan, Lai
[Loi] area, 22 July 1929 (fr), Tsang & Fung
554 (holotype, PE 100207; isotypes A, K, PE
1188347). Figures 1, 2.

Species Gneto lofuense C. Y. Cheng affinis, sed foliis
nervatis lateralibus utroque latere 6— 8 (non 9-1 I), spici-
bus maribus longioribus, involucris 12—20 (non 9—11),
seminibus in sicco laevibus vel obscure retieulatis (non
grosse retieulatis), minoribus ((1.5— )1 .9—2.1 (-2. 5) X 1.1 —
1.4(— 1.6) cm, non 2.4— 2.8 X 1.5— 1.6 cm); a Gneto par-
vifolio (Warburg) Chun foliis maioribus (10— 15(— 30) X 3—
7.5 cm, non 2.5— 10(— 13) X 1.5—5 cm), spicibus maribus
brevioribus, involucris 12—20 (non 5— 10(— 12)), pi I is in¬
train volucrali bus dense conspicuis (non paucis brevissim-
is), seminibus in sicco late ellipsoideis non striatis (non
fusiformibus longitudine striatis) manifeste differt.

Vines, slender and fragile, glabrous except with¬
in the fertile spikes. Petiole 8-12 mm; leaf blade
oblong-elliptic or oblong-ovate, 10— 15(— 30) X 3—
7.5 cm, leathery, glossy; lateral veins 6—8 each
side. Male inflorescences simple or cymosely once-
branched. Male spike 15—30 X 3^4 mm; involucres
12-20; each subtending (40— )60— 80(-90) male
“flowers” in several ranks plus an uppermost whorl
of 15— 20(— 30) sterile female “flowers”; basal hairs
between the “flowers” very numerous and forming
a dense brown cushion. Female inflorescence ax¬
illary on old branches, cymosely once-branched,
10—15 cm long when fertile. Female spike with 10—
20 involucres, rachis thick; each node with 8 or 9
female “flowers.” Seed cylindric to broadly ellip¬
soid, red, outer integument smooth to ± retieulately
wrinkled when dry, (1.5— ) 1 .9 — 2.1(— 2.5) X 1.1—
1.4(— 1.6) cm, sessile or nearly so; apex acute or

Novon 9: 187-189. 1999.

188

Novon

Figure 1. Isotype (Tsang & Fung 554) of Gnetum hainanense C. Y. Cheng ex Fu l.i-kuo, Yu Yong-fu & M. G. Gilbert
(erroneously written as Yu et al. in photograph).

Volume 9, Number 2
1999

Fu et al.

Validation of Gnetum hainanense

189

IIIIIIIIIIIIIIIIIUIIIIIIIII

Figure 2. Detached seed of Gnetum hainanense, taken
from the packet attached to the isotype sheet (scale in
millimeters).

with a small pointed umbo in center. Flowering
February— July, producing seeds July-December.

Gnetum hainanense is probably most closely re¬
lated to G. lofuensis, which has rather similar retie-
ulately wrinkled seeds; the seeds differ in G. lo-
Juensis by being distinctly larger and much more
prominently and coarsely wrinkled, perhaps indi¬
cating a thicker fleshy layer. Gnetum parvifolium,
to which G. hainanense was originally likened, is
clearly differentiated by the much smaller seeds
that are finely longitudinally striate when dried and

the smaller male spikes with only 8-12 involucral
whorls and with few, short hairs between the male
“flowers.”

Paratypes. CHINA. Guangxi: Far Pan, W of Hoo-
chih, 22 July 1928. R. C. Clung 6503 (A); Seh-feng. I)ar
Shan, S of Nanning, 21 Oct. 1928, R. C. Citing R093 (A.
NAN); Shang-sze Distr. (near Guangdong border), Na Wai
Village. Shap Man Taai Shan, June 1934, IF T. Tsang
23948 (A); Shang-sze Distr., Shap Man Taai Shan, 31 Aug.
1934. IF. T. Tsang 24161 (A, MO). Hainan: Dung ka to
Wen Fa Shi. 520 m, 1932. Ar. A. Chun & C. 1.. Tso 43785
(A. B. k); Pak Shik Ling, ku Tung village (Ching Mai
Distr.), 22 July 1933, Fung H. 20175 (A. BM. E, k. MO.
PE); Ch’ang-kiang Distr., ka Chik Shan. 28 Mar. 1933.
Yaichow, 700 m. 1933, F. C. How 70550 (MO. PE); no
locality, 29 Apr. 1933. F. C. How 70614 (A. PE), S. A.
Lau 1410 (A. BM); Ue Lung Shan (Ch'ang-kiang Distr.).
25 Jan. 1934, S. A. Lau 3214 (A); Kan-en Distr., Chim
Fung Mt.. near Fong Ngau Po Village. Jan. 1935, S. A.
Imu 5256 (A. E); Fat Lo Shi, Yai-hsien. Mar. 1935, S. A.
Lau 5775 (A); Yeung Lam shan, Yai-hsien, Mar. 1935, S.
A. Lau 6282 (A. E); Ling-shui (Ling-tui) Distr.. Chim
Shan. Fan Maan Ts'uen, May 1932, S. A. Imu 26455 (A);
Bak-sa, 29 Apr. 1936. S. A. Imu 26549 (A): Loktung, 24
May 1936, S. A. Giu 26888 (A. kUN); Pak Shik Ling, ku
Tung village (Ching Mai Distr.), 22 July 1933. C. I. 875
Isi (B. k, PE); no locality, 23 Feb. 1934, II. Y. Liang
65151 (E, PE). McClure 20084 (GH. k. MO. PE); Paak
Po Shan, Taam-chau Distr., 1 Sep. 1927, W. T. Tsang 727
(A, k. PE).

Acknowledgments. We thank the curators and
directors of the following herbaria for access to
their material, directly or via loans: A. B, BM. E.
K, KUN, MO, NAN, PE.

Literature Cited

Cheng W. C., Fu L. k. & Cheng C. Y. 1975. Gymnosper-
mae Sinicae. Acta Phytotax. Sinica 13(4): 56—90, figs.
1-66.

Notes on the Piperaceae of China

Michael G. Gilbert

Flora of China Project, Missouri Botanical Garden, c/o Department of Botany, Natural History
Museum, Cromwell Road, London SW7 5BD, United Kingdom

A in Nian-he

Taxonomy Department, South China Institute of Botany, Chinese Academy of Science, Wushan,
Guangzhou, Guangdong SI 0224, People’s Republic of China

ABSTRACT. Four new species of Piper are de¬
scribed from China: P. dolichostachyum , P tseng-
ianum , P. wangii, and P. yui ; one new combination
is made: Piper boehmeriifolium var. glabricaule ; a
new name. Piper cathayanum, is proposed for a lat¬
er homonym; the identities of Piper arboricola and
P. philippinum are discussed; and new synonyms
and lectotypes are proposed for various other spe¬
cies of Peperomia and Piper.

The following notes are a precursor to the ac¬
count of the Piperaceae to be published in volume
4 of the Flora of China. The Asian members of this
family have been largely ignored by European tax¬
onomists, and our studies have shown the need for
a significant number of changes from the first flo-
ristic account of the family within China (Tseng in
Chen et al., 1982). This account was produced
without access to most of the types, or even pub¬
lished assessments of their identities, or to very
much material from neighboring countries.

Peperomia blunda (Jacquin) Kunth, in HBK,
Nov. Gen. Sp. 1: 67. 1816. Piper blandum Jac¬
quin, Collectanea 3: 211. 1789. TYPE: Jac¬
quin, Ic. PI. Rar. 2, t. 218.

Peperomia arabica Deeaisne ex Miquel, Syst. piperac. 1:
121. 1843. TYPE: Yemen. P. A. Botta s.n. (lectotype,
here designated, P). Rejected syntype: South Africa.
Cape of Good Hope to Port Natal, Drtge s.n. (G).
Peperomia dindygulensis Miquel, Syst. piperac. 1: 122.
1843. TYPE: India. Habitat in rupibus Prov. Din-
dygul. Wig hr in /V. Wallich 666 3B (lectotype, here
designated, K-WALL; isolectotype, P).

Peperomia esquirolii 11. L^veillG Repert. Spec. Nov. Regni
Veg. 10: 149. 1912. Syn. nov. TYPE: China. Guizhou
(“Kouy Tch£ou”): 18 Mar. 1910, Esquirol 710 (ho-
lotype, E; isotype, K).

Peperomia formosana C. DC., Annuaire Conserv. Jard. Rot.
Genfcve 21: 223. 1920. Syn. nov. TYPE: Taiwan: kel-
nag Samtianneapass, Warburg 9338 (holotype, B).

Novon 9: 190-198. 1999.

Peperomia japonica Makino, Bot. Mag. (Tokyo) 15: 145.
1901. Syn. nov. SYNTYPES: Japan. Loochoo, 1876,
Coll. Imp. Mus. s.n. (in hb. Makino) & H. Kuroiwa
s.n. (in hb. Makino); Okinawa, Nakizin, Mar. 1887,
S. Toshiro s.n.; Prov. Tosa in Isl. Shikoku: Isl. Hesh-
ima, K. Naganuma s.n. & 17 June 1887, T. Makino
s.n. & Y. Yoshinaga s.n. (in hb. Makino); Prov. Mu-
sashi: Tokyo, Bot. Gard. Koishikawa, cult, from
Amami-Oshima, Nov. 1901, Uchiyarna in Makino
s.n. (all Tl not seen).

Peperomia laticaulis C. DC., Annuaire Conserv. Jard. Bot.
Geneve 21: 223. 1920. Syn. nov. TYPE: Taiwan: in
rupibus montium Kushaka, Faurie 481 (lectotype,
here designated, G-DC; isolectotypes, B. BM, P). Re¬
jected syntypes: Taiwan: Urai, Faurie 626 (B, BM, P).

A decision has been made to use a rather wide
species concept for this pantropical Peperomia
blanda complex. The variation in stature, leaf
shape, coloring, and details of inflorescence mor¬
phology are considerable, but we believe that there
are not sufficient discontinuities to justify the rec¬
ognition of distinct taxa, at least in mainland Asia.
Material from Japan is very uniform but not distin¬
guishable from some of the collections from the
mainland, and there is not a good case for keeping
Peperomia japonica as a distinct species. The po¬
sition within Taiwan is more perplexing. There are
two distinct forms on the island: a very robust plant
without any red pigmentation corresponding to Pep¬
eromia japonica and including P. laticaulis , and a
much more delicate plant with red stems and un¬
dersides of the leaves and a distinctive epidermis,
recently described as P. sui (Lin & Lu, 1995). How¬
ever, the dividing line between these and the main¬
land plants is not clear, and for now we prefer to
include everything within the one taxon. This does
not do justice to the Taiwan plants, which taken in
isolation would have to be treated as two distinct
species. None of the syntypes of Peperomia japon¬
ica have been seen, but material from Japan proper
is very uniform and all belongs to the one taxon of
Peperomia. There can be no reasonable doubt as to
the identity of P. japonica.

Volume 9, Number 2
1999

Gilbert & Xia
Piperaceae of China

191

The opportunity has been taken to designate lec-
totypes for those taxa for which we have seen syn-
type material. In the case of Peperomia arabica, the
specimen from Arabia (Yemen) is clearly in accord
with the protologue and, in view of the specific ep¬
ithet, it seems logical to select this as lectotype in
preference to Miquel's other syntype from South Af¬
rica. The original material of P. dindygulensis in
the Wallieh herbarium is quite variable. Again, we
have chosen a collection clearly associated with the
locality from which the taxon was named and one
representing one of the more clearly defined forms
within the species (drying a rather pale green and
with many terminal and subterminal inflorescenc¬
es). There are two cited syntypes of P. laticaulis,
Faurie 481 and Faurie 626. These syntypes should
both be in Geneva, but Faurie 626 could not be
located there and thus we feel obliged to select
Faurie 481 as the lectotype. Unfortunately, dupli¬
cates seen in B, BM, and P show Faurie 626 to
have better developed inflorescences.

Peperomia heyneana Miquel, Syst. piperac. 1:
123. 1843. TYPE: Nepal. Kandrang Garhi,
near Kathmandu [“Chandaghiry”], Feb. 1821,
N. Wallieh 6663 C (holotype, K-WALL).

Peperomia duclouxii C. DC., Notul. Syst. (Paris) 3: 41.
1914. Syn. nov. TYPE: China. Yunnan: Hay-y, pres
Lou-Lo, Ducloux 4760 (holotype, P).

The characters used to distinguish Peperomia
duclouxii from P. heyneana are quantitative, and
similar plants have been seen from throughout the
range of that species. Moreover, there does not
seem to be any discontinuity between the two taxa,
and we feel confident that P. duclouxii was based
on stunted material of P. heyneana.

Piper boehmeriifolium (Miquel) C. DC., in A.
DC., Prodr. 16: 348. 1868. Chavica boehmer-
iifolia Miquel, Syst. piperac. 1: 265. 1843.
TYPE: Bangladesh. Sylhet [“Sillet” or “Sil-
het”], N. Wallieh 6654A (holotype, K; isotype,
K-WALL).

Piper boehmeriifolium var. boehmeriifolium

Piper boehmeriifolium var. tonkinen.se C. DC., in Lecomte,
El. Indo-Chine 5: 81. 1910. Syn. nov. TYPE: Viet¬
nam. Vallee de Lankok (Mont-Bavi), dans les hois.
9 Mar. 1888. Halansa 3628 (holotype, P).

Piper spirei C. DC., in Lecomte, FI. Indo-Chine 5: 87.
1910. Syn. nov. TYPE: Laos. Phon thane. Spire 258
(holotype, P; isotype, P).

Piper spirei var. pilosius C. DC., in Eecomte, El. Indo-Chine
5: 88. 1910. Syn. nov. TYPE: Cambodia. Erequens in
montibus Krewanh (“krewwaoh" on holotype), June
1870. Pierre 4817 (holotype, P; isotvpes, P).

Piper terminaliflorum Y. C. Tseng, Acta Phytotax. Sin. 17:
30. 1979. Syn. nov. TYPE: China. Yunnan: Feug-
qing, 2200 m. 24 June 1938, T. T. Yii 16454 (holo¬
type. PE; isotypes, A, E).

Piper boehmeriifolium var. glabricaule (C. DC.)
M. G. Gilbert & N. H. Xia, comb, et stat. nov.
Basionym: Piper glabricaule C. DC., Notizhl.
Konigl. Bot. Gart. Berlin-Dahlem 6: 477.
1917. TYPE: China. Yunnan: Mengzi Xian,
Mengzi [“Mbngtse”] In silvis montium meri-
dionali-orientialum, A. Henry 9482A (holotype,
B; isotypes. A, K).

Piper boehmeriifolium is the most commonly col¬
lected erect species of Piper in China and Indo¬
china. There is considerable variation in leal width;
associated with this is variation in the number of
leal veins, peduncle length, and bract diameter, the
principal characters used to distinguish P. boeh¬
meriifolium var. tonkinen.se from P. glabricaule. We
are unable to define any discontinuity between P.
boehmeriifolium s. str. and variety tonkinense that
should be included within the species. Male plants
appear to have effectively determinate growth, with
the uppermost leaves very poorly developed so as
to leave the inflorescences in a terminal position.
Y. C. Tseng treated such plants as a distinct spe¬
cies, P. terminaliflorum. However, there are a few
male plants from throughout the distribution of P.
boehmeriifolium that resemble the type ol P. ter¬
minaliflorum, and we believe that this species is
also better included within P. boehmeriifolium s. str.
The small fruits of P. glabricaule are more distinc¬
tive, but again the variation is almost continuous.
We therefore prefer to treat P. glabricaule as a va¬
riety of P boehmeriifolium rather than as a lull spe¬
cies.

The type collections of Piper spirei and P. spirei
var. pilosius are not distinguishable from P. boeh¬
meriifolium. However, the Chinese collection that
was the basis of the record of Piper spirei in the
Flora Reipublicae Popularis Sinicae account (Tseng
in Chen et al., 1982) has not been matched with
any other species and is described here as the new
species P. dolichostachyum.

Piper eat hay a in i n i M. G. Gilbert & N. H. Xia,
nom. nov. Basionym: Chavica sinensis Cham¬
pion ex Bentham, Hookers J. Bot. Kew Garil.
Misc. 6: 116. 1854. Piper sinense (Champion
ex Bentham) C. DC., in A. DC., Prodr. 16: 361.
1868, nom. illegit. Blocking name: Piper chi-
nense Miquel, London J. Bot. 4: 439. 1845.
TYPE: Hong Kong: Champion 491 (holotype,
K, ex herb. Bentham; isotype, K, ex herb.
Hooker).

192

Novon

It is with distinct regret that we designate a new
name lor Piper sinense. This is a much better known
species than that with the blocking name, P. chi-
nense, which is certainly extremely rare or even ex¬
tinct, known to us only from the type. Unfortunately
the ICBN, Article 53.3, Example 8 (Greuter et al.,
1994), is quite clear that “chinense” and “sinense”
must be treated as homonyms; thus the later hom¬
onym must be given a new name.

K has two sheets of Champion 491, but the one
from Bentham's personal herbarium must be pre¬
sumed to be the holotype, while the slightly better
sheet from Hooker’s personal herbarium is an iso-
type.

Piper (iolichostacliyum M. G. Gilbert & N. H.
Xia, sp. nov. TYPE: China. Yunnan: Jingping,
16 Apr. 1956, Sino-Soviet Botanical Exped. 67
(holotype, SCBI; isotype, KUN).

A Piper boehmeriifolio, nervis paginis foliis abaxillari-
bus dense rufipilosis, non glabris vel sparse pallidipub-
erulis, infructescentibus femineis longissimis (27-30 cm,
non 6-12 cm), inanifeste differt.

Erect, shrubby herb; most parts with reddish
brown hairs. Stems 3-4 mm thick, furrowed when
dry, glabrescent. Petiole 5—13 mm, densely pubes¬
cent, prophyll to 3 cm, glabrous; leaf blade elliptic-
lanceolate to obovate, strongly asymmetrical, to 14—
25 X 6—11 cm, thinly papery, base strongly
obliquely cordate, basal lobes overlapping, bilateral
difference to 3 mm, apex long acuminate, veins 8-
10, 3 on the narrow side, up to 7 on broad side,
apical pair arising 2-6 cm above base, alternate,
nearly reaching leaf apex, next pair often also
above base, reticulate veins lax, transversely ob¬
long, slightly raised abaxially, without evident
glands, abaxially densely brown-pubescent, almost
tomentose on veins, adaxially sparsely minutely
scabrid. Plants dioecious. Spikes leal-opposed.
Male spike not seen. Female spikes 27-30 X 0.6-
0.7 cm in fruit, peduncle 4r-4. 5 cm, glabrous; ra-
chis pubescent; bracts orbicular, peltate, margin
pale when dried, 1.5— 1.7 mm diam. Ovary ± cy¬
lindrical; stigmas 3 or 4, reflexed, very short and
inconspicuous. Drupes densely packed, prismatic-
cylindrical, ca. 2 X 1.5 mm.

This material was first identified as Piper spirei
C. DC. (Tseng in Chen et al., 1982), but examina¬
tion of the holotype of that name has shown it to
be inseparable from P boehmeriifolium and has re¬
vealed that this Chinese specimen was without a
name. One of the more distinctive features is the
extremely long infructescence, and it has been
named accordingly (Greek: dolichos = long, stachys

= spike, as in the inflorescence of wheat). The oth¬
er distinctive feature is the indumentum, which is
quite dense and reddish brown, in contrast to P.
boehmeriifolium, which is glabrous or only sparsely
and inconspicuously puberulent.

Piper hongkongense C. DC., in A. DC., Prodr.
16: 347. 1868. TYPE: Hong Kong: Seemann
(holotype, G-DC not found).

Chavica puberula Bentham, FI. Hongk.: 335. 1861. Syn.
nov. Piper puberulum (Bentham) Maximowicz, Bull.
Acad. Imp. Sci. Saint- Peters bourg, ser. 3, 31: 94.
1887, nom. illegit. Blocking name: P puberulum
(Bentham) Seemann, FI. Vit. 268. t. 75. 1868, based
on Macropiper puberulum Bentham, London J. Bot.
2: 235. 1843. TY PK: Hong Kong: Hance 10159 (ho¬
lotype, BM).

The hairy species of Piper in southeastern China
have been subjected to more than their fair share
of nomenclatural problems. One of the more widely
used names is P. puberulum (Bentham) Maximow¬
icz, which is based on material collected on Hong
Kong Island. Unfortunately, this is a later homonym
of P. puberulum (Bentham) Seemann, which is
based on a very different species from Fiji. Thus a
name regarded as a synonym in all works subse¬
quent to publication must be resurrected. A prob¬
able factor in the confusion between the two hom¬
onyms must have been the fact that both were
based on taxa described by Bentham, the Fijian
species in Macropiper and the Chinese species in
Chavica. The next available name is P. hongkon¬
gense, which lias been consistently ignored (Bret-
sehneider, 1898) or treated as a synonym of P. pub¬
erulum (Bentham) Maximowicz (Hemsley, 1891)
ever since it was published. There is a problem
with the typifieation of P. hongkongense in that the
protologue cited only one element, “Hongkong
(Seemann! in h. DC.),” which was not found in the
De Candolle herbarium in Geneva. Hemsley (1891)
cited a collection by Seemann from Hong Kong un¬
der P. puberulum (Bentham) Maximowicz apparent¬
ly in Kew, but no material attributed to Seemann
was found there either. If another search of the Ge¬
neva herbarium fails to locate this material it is
suggested that Hance 10159, the holotype of Chav¬
ica puberula, be designated the neotype of P hong¬
kongense.

The protologue of Chavica puberula gave no de¬
tails of the type beyond “Hongkong, Hance” but
there is only one collection in K and BM, Hance
10159 in BM, that has been annotated by Bentham
and it is assumed that this is the holotype.

The species varies from very densely hairy to
quite thinly hairy. Material with a thinner indu-

Volume 9, Number 2
1999

Gilbert & Xia
Piperaceae of China

193

mentum has often been regarded as conspecific
with material from Taiwan and placed within “ Piper
arboricola .” The Taiwan material is here regarded
as belonging to an endemic taxon, P. sintenense
(q.v.), while the mainland material seems to be bet¬
ter regarded as belonging to a single variable spe¬
cies.

Piper kadsura (Choisy) Ohwi, Acta Phytotax.
Geobot. 3: 81. 1934. Ipomoea kadsura Choisy,
Mem. Soc. Phys. Geneve 6: 475. 1833. TYPE:
Japan. “Iaponicae: Karami Kadfura, it. Saifin."
(holotype, UPS-THUNB).

Piper arboricola C. DC., Annuaire Conserv. Jard. Bot. Ge¬
neve 21: 221. 1920. Syn. nov. TYPE: Taiwan: in
arboribus Ke-Lung, 13 May 1903. Faurie 480 (lec-
totype, here designated, G; isolectotypes. BM, P(2
sheets)).

Until recently the name Piper arboricola had
been applied almost entirely to material of what is
here named as P sintenense. Examination ol dupli¬
cates of the syntypes of P. arboricola in BM and P
shows that this represents a rather broad-leaved
form of Piper kadsura and that P. arboricola must
be treated as a synonym ol that species. Japanese
authors have spelled the epithet as “ kadzura ,” pre¬
sumably in accordance with current transliteration
of the vernacular name upon which the epithet is
based. This is not in accordance with the spelling
used in the protologue and associated literature,
which should be retained. Faurie 480 is selected
as the lectotype of P. arboricola in preference to
Faurie 479, because although the two collections
are very similar in quality there seem to be more
sheets of Faurie 480 available. Records ol this spe¬
cies from Fujian and Zhejiang are based on mate¬
rial that seems better regarded as a form of Piper
wallichii.

Piper kwashoense Hayata, J. Coll. Sci. Imp. Univ.
Tokyo 30: 235. 1911. TYPE: Taiwan: Kwash-
oto, 1907, Kawakami & Kobayashi 475 (ho¬
lotype, “ Tokyo Univ.").

This taxon has been named as Piper philippinum
Miquel, following the account of Philippine Piper¬
aceae by Quisumbing (1930: 110). However Mi-
quel’s (1843) protologue clearly excludes the taxon
described by Quisumbing, as it refers to a plant
with 5-veined leaves, lax female inflorescences 10-
12 cm long, and ovoid fruits 4—5 mm long, appar¬
ently free from the rachis. Quisumbing used the
name for a plant with 7-veined leaves, female
spikes 3—9 cm long, and relatively dense and
subglobose fruit, 2—3 mm in diameter, partly con¬

nate to the rachis. The mistake seems to have come
from Miquel, who based the protologue proper on
a female plant, “ Cuming in herb, de Lessert 1642,”
and then added a note that another collection,
Cuming 912, might be the male of the same species
though it did have a number of differences. This
latter collection is much more widely distributed,
and P. philippinum has been interpreted as if Cum¬
ing 912 were the type and the description in the
protologue ignored.

Piper macropodum C. DC., Bull. Herb. Boissier,
ser. 2, 4: 1026. 1904. TYPE: China. Yunnan:
Simao Xian, Simao [“Szemao”], 1370 m, A.
Henry 12210D (holotype, Z not seen; isotype,
K).

Piper szemaoense C. DC., Notizbl. Konigl. Bot. Gart. Ber-
lin-Dahlem 6: 481. 1917. Syn. nov. TYPE: China.
Yunnan: Simao Xian, Simao [“Szemao”], 1370 m, A.
Henry 12210B (lectotype, here designated, B; isolec¬
totypes, A, E, K, MO).

Piper szemaoense was separated from the com¬
pletely glabrous P. macropodum by the presence of
an indumentum. The two taxa are sympatric and,
from the numbering used, were regarded by the col¬
lector as just one species. All other variation is
continuous, and we have no hesitation in maintain¬
ing only one taxon. Two collections were listed in
the protologue of P. szemaoense, Henry 12210A and
12210B (as Henry 1210A and 1210B). The latter
collection is selected as lectotype because it is fe¬
male, and thus taxonomically more informative, and
also because it seems to be slightly more widely
distributed.

Piper pedicellatum C. DC.. J. Bot. 4: 164. 1866.
TYPE: Bangladesh [“Bengalia orient.”]. Grif¬
fith 4404 (lectotype, here designated, K not
seen; isolectotype, P).

Piper curtipedunculum C. DC., Notizbl. Konigl. Bot. Gart.
Berlin-Dahlem 6: 481. 1917. Syn. nov. TYPE: Chi¬
na. Yunnan: Mengzi Xian, Mengzi [“Mongtse”], A.
Henry 10438 (holotype. B; isotypes. A, E. K).

The protologue of Piper pedicellatum lists three
elements. Hooker & Thomson s.n. from Sikkim, in
Geneva (“Herb. Cand.”), and two collections by
Griffith, 4404 and 4418, from “Bengalia orient.,”
now Bangladesh, in Kew. The Geneva sheet of the
Hooker and Thomson collection was not located
during a visit to Geneva and, because most exsic-
catae of these collectors were distributed with very
incomplete label data, there is often doubt as to
which sheets belong to the same collection. Thus,
it does not seem advisable to select this collection

194

Novon

as the type. Griffith 4418 is ol female material, and
De Candolle indicated some doubts (probably un¬
justified) about the identity of this collection in a
later account of this species (C. De Candolle, 1869:
350) so this syntype also seems best rejeeted. Grif¬
fith 4404 is of male material, and as the epithet
alludes to the distinctively long-stalked male
bracts, it seems best to select this as the lectotype.

The minor quantitative differences in leaf shape,
peduncle length, and bract size used to separate
Piper curtipedunculum from P. pedieellatum have
proved too ill-defined to justify the maintenance of
two taxa. especially when material from the full
range is considered.

Piper puhicatuluin C. DC., in Lecomte, FI. Indo-
Chine 5: 74. 1910. TYPE: Vietnam. “In ne-
more Van XA, 4 May 1886, Bon 3139bis ” (ho-
lotype, P).

There is some doubt as to the identity of the type
of this species. The above specimen is the only
sheet so named found in the Paris herbarium. It
was determined by C. De Candolle. The protologue
gives the type as Bon s.n. from “Tonkin, environs
de Ninh Binh,” but no such material was seen and
we have assumed that Bon 3139bis must be the
collection upon which the species was based. The
material has 5-veined leaves with more or less sym¬
metrical, cuneate bases, glabrous bracts, and very
densely packed, globose fruit ca. 1 mm in diameter.
The protologue describes the leaves as 5-veined
with slightly oblique bases and glabrous bracts. In
distinct contrast, Chinese material so named has 7—
9-veined leaves with distinctly oblique bases,
rounded to almost cordate on one side, and hairy
bracts. It seems certain that the Chinese material
is not correctly named. Because it has not been
possible to match it with any other species, it is
described below as a new species. Piper wangii.

Piper rhytidocarpum J. D. Hooker, FI. Brit. In¬
dia 5: 92. 1886. TYPE: Bangladesh. Chitta¬
gong, 22 Sep. 1850, Hooker & Thomson s.n.
(lectotype, here designated, K; isolectotype,

K).

Piper madid urn Y. C. Tseng, Acta Phytotax. Sin. 24(5):
382, fig. 1. 1986. Syn. nov. TYPE: China. Xizang:
Medog, Beiheng. 850 m, 9 Apr. 1983, It. S. Li & .S'.
/. Cheng 3990 (holotype. PE not seen; isotypes, (2.
not annotated) PE).

Rejected syntypes. BANGLADESH (“East Bengal”).
Griffith 4423 (k); Chittagong, 17 July 1850. Hooker &
Thomson s.n. ( K). F. de Silva & It. Gomez in Wallich
6 6588 (k).

The collection designated here as lectotype is
annotated by Hooker and is the only collection cit¬
ed in the protologue with apparently mature fruit.
The material of Piper madidum from southeastern
Xizang is a good match in all features, including
the peculiar sessile scales on the undersides of the
leaves, though the fruits are too immature to show
the granulose surface of the fruits upon which the
specific epithet was based. No material annotated
by the author was seen, but two unnamed sheets
matching in all details and fitting the description
have been seen.

Piper sarmentosum Roxburgh, in Hunter, Asiat.
Res. 11: 565. 1810. Chavica sarmentosa (Rox¬
burgh) Miquel, Syst. piperac. 1: 242. 1843.
TYPE: “cultivated in Calcutta," Roxburgh tab.
1267 (lectotype, here designated, K).

Piper albispicum C. DC., in Lecomte, FI. Indo-Chine 5:
85. 1910. Syn. nov. TYPE: Vietnam. Tonkin merid¬
ional: In sepib. Ninh Bhinh.. 4 Nov. DUO. Hon 982
(holotype, P).

Piper brevicaule C. DC., Annuaire Conserv. Jard. Bot. Ge¬
neve 2: 272. 1898. Syn. nov. TYPE: Vietnam. Ton¬
kin: village de Tchontiao, a la base du Mont-Bavi,
July 1886. Balansa 3631 (lectotype, here designat¬
ed, P).

Piper gymnostachyum C. DC., in Lecomte, FI. Indo-Chine
5: 72. 1910. Syn. nov. TYPE: Cambodia. Phuoc-
ihan, Thorel “T” (lectotype. here designated, P).
Piper lolot (7 DC., Annuaire Conserv. Jard. Bot. Geneve
2: 272. 1898. Syn. nov. TYPE: Vietnam. Tonkin, En¬
virons de Quang-yen. Sep. 18155. Balansa 539 (ho¬
lotype. P).

Piper pierrei C. DC., in Lecomte. FI. Indo-Chine 5: 78.
1910. Syn. nov. TV PE: Vietnam. In svlvis ad Bao-
chiang in austro Cochinchine, Sep. 1869, Pierre
4814 (holotype, P; isotype, P).

Piper saigonense C. DC., in Lecomte, FI. Indo-Chine 5:
79. 1910. Syn. nov. TYPE: Vietnam. Ad urbem Sai¬
gon in austro Cochinchine, Aug. 1872. Pierre 1142
(holotype, P: isotypes, MO. P).

Piper sarmentosum is a distinctive species, easily
recognized by the creeping, more or less terrestrial
habit, virtually unmatched among Asiatic Piper,
combined with the very distinctive leaves, which
are usually palmately veined or almost so with a
very minute “powdery puberulent” indumentum,
and infructescences white at anthesis and with
fruits fused to the rachis when mature.

The earliest literature reference to this taxon ap¬
pears to be by Hunter (1807) who used the provi¬
sional name “Piper latifolium” for a sterile plant.
In subsequent correspondence with Roxburgh, it
was established that this was the same as a species
in cultivation in Calcutta and provisionally named
by Roxburgh. Hunter published the name and de¬
scription that Roxburgh had supplied in an appen-

Volume 9, Number 2
1999

Gilbert & Xia
Piperaceae of China

195

dix to volume 11 of “Asiatick Researches, but this
has been largely overlooked; the name is more often
given as validated in 1820 in Roxburgh’s Flora
Indica. No Roxburgh herbarium material has been
located, but the species was illustrated for Rox¬
burgh (painting 1267) and this makes an adequate
lectotype. The leaves are often used as a condiment
and also medicinally, and it is likely that it has
been spread by humans: many of the records are
from around habitation. It was therefore surprising
to find that so many collections from Vietnam and
Cambodia have been described as distinct species.

It seems that C. De Candolle failed to recognize the
extremely distinctive habit anti based his descrip¬
tions on minor variations in leaf size and shape,
due largely to variation between the lower leaves
and the uppermost leaves associated with the inflo¬
rescences, and also variation in size and color of
the inflorescence with age.

Two collections annotated as Piper brevicaule by
C. De Candolle were found in the Paris herbarium:
Balansa 3631 from “Tonkin: village de Tchontiao,
a la base du Mont-Bavi” and Godefroy in Harmand
s.n. from Saigon. The protologue cited two syntypes,
Balansa 3631 with details almost exactly as above,
and “Cambodje ( Harmand in h. Mus. Par. Spec,
fructiferum).” In view of the apparent discrepancy
between protologue and herbarium specimen, it
seems best to designate the Balansa collection as
lectotype, especially as De Candolle himself wrote
“sp. nov.” on the sheet. Piper gymnostachyum was
based on two collections by Thorel: “T,” a male
collection from Phuoc-than, Cambodia, and V, a
fruiting collection from Nha-met, Vietnam. The lat¬
ter collection, particularly the infructescences, has
been infested by fungi, making it difficult to see
the indumentum, particularly the “naked spikes
alluded to by the epithet. It thus seems preferable
to select the better preserved male collection as the
lectotype.

Piper sintenense Hatusiina, Acta Phytotax. Geo-
bot. 4: 210. 1935. TYPE: Taiwan. Prov. Tai-
hoku, in silvis districtus Kanko, Nov. 1932, S.
Hatusima s.n.

Piper hispidum Hayata, J. Coll. Sci. I ni|>. Univ. Tokyo .10:
234. 1011. Not P. hispidum Kunth, in HBk. Nov.
Gen. Sp. 1: 50. 1816. TYPE: Taiwan: Koshun, Gar-
anbi, 1896, Y. Toshiro s.n. (holotype, TAI: [Tokyo,
photo] TAI).

Most material of this taxon has been named as
Piper arboricola (see Tseng in Chen et al., 1982).
Examination of type material of P. arboricola has
shown clearly that it belonged to P. kadsura and
that the name was incorrectly applied to this spe¬

cies. This taxon is most easily recognized by the
membranous leaves, usually drying dark green, and
by the coarse curved hairs on the stems. The only
available name is Piper sintenense.

Piper thomsonii (C. DC.) J. D. Hooker, FI. Brit.
India 5: 87. 1886. Chavica thomsonii C. DC.,
Prodr. 16: 389. 1868. TYPE: India. Khasia
Mountains, 20 June 1850, Hooker A* Thomson
18 (holotype, B; isotype, K).

Piper bavinum C. DC., Annuaire Conserv. Jard. Bot. Ge¬
neve 2: 270. 1898. Syn. nov. TYPE: Vietnam. Mont
Bavi. vallee de Lankok, entre Moc-ha et Lang-nuong,
Balansa 3630 (lectotype. here designated. P).

Piper punetulivenuin C. DC., in Lecomte, FI. Indo-Chine
5: 77. 1910. Syn. nov. TYPE: Taos. Euang Prabang,
Thorel s.n. (lectotype, here designated, P).

Piper punctulivemim v ar. parvifolium C. DC., in Lecomte.
FI. Indo-Chine 5: 77. 1910. Syn. nov. TYPE: Laos.
Paklai. Thorel s.n. (holotype, P).

This species is very common in southwestern
China and can be recognized instantly by the char¬
acteristic dark red glands of the leaves; the long-
acuminate leaves with pale undersides and shorl
infructescences are also distinctive. Piper bavinum
has been distinguished primarily by having shorter
inflorescences and lacking the very finely powdery
pubescent indumentum of P. thomsonii s. str., but
it is now clear that there is a continuity in variation
such that it is not possible to justify treating these
two taxa as distinct. Piper punctulivenum and its
variety are clearly this species with even more
prominent leaf glands than in the rather poorly pre¬
served types of P. bavinum.

Piper thomsonii has sometimes been included
within Piper sylvaticum Roxburgh, most notably by
Long (1984), but there are discrepancies between
the taxon as accepted here and the protologue of
Roxburgh's species. Most notably, the protologue
mentions the leaves as being “broad-cordate, ob¬
tuse, lobes of the base large, equal, circular," which
coupled with the mention of stems creeping on the
ground is much more suggestive of another one of
Roxburgh’s species. Piper sarmentosum. The type
illustration includes an enlargement of a leal tip
that is clearly bluntly acuminate; it must be as¬
sumed that the “obtuse” in the protologue does not
refer to the leaf apex as a whole, which is shown
as being very distinctly acuminate in overall form.

The protologue of Piper bavinum cites two col¬
lections, Balansa 3627 and Balansa 3630. Neither
collection is of very high quality, both being so
poorly dried that the usually distinctive dark red
glands can only be seen with careful examination.
Choice of a lectotype is somewhat arbitrary, but it

196

Novon

could he argued that the specimen citation by C.
De Candolle (1910), “Tonkin: mont Bavi, valine de
Lankok, entre Moc-ha et Lang-nuong ( Balansa ),”
clearly refers to Balansa 3630 from exactly that
locality. This could he regarded as a move toward
lectotypification by the original author and, as the
quality of the material is uniform, it is best to con¬
firm this selection as the lectotype. The material of
P. punctulivenum is of much better quality with
nothing to pick between the two syntypes of P.
punctulivenum s. str., both unnumbered Thorel col¬
lections, one from Luang Prabang and the other
from the same area as the holotype of variety parvi-
folium (Paklai). The collection from Luang Prabang
is selected as lectotype on the grounds that it is
more likely that there could be significant differ¬
ences between the material from different localities
than between the two collections from the same
area (Paklai).

Presumed paratype of P. bavinum. VIETNAM. Forets
du Mont-Bavi, au-dessus de la pagode de Dfein-Touan, 18
Mai 1888. Balansa 3627 (P).

Piper tsengianum M. G. Gilbert & N. H. Xia, sp.
nov. TYPE: China. Yunnan: flanks of Shweli
valley, ca. 24°20'N, 98°33'E, 2100-2400 m.
May 1925, Forrest 26411 (holotype. A; iso¬
type, E).

A Piper flavifloro, paginis foliis abaxillaribus squamis
griseo-albis obtectis, non pall id i vi ridis nudis, inflorescen-
tibus maribus longissimis griseo-viridis, non flavis, man-
ifeste differt.

Shrubs, climbing, 1.2-2 m, apparently glabrous.
Stems pale brown when dry, 2-2.5 mm thick, te¬
rete, striated. Prophylls ca. 15 mm; petioles 8—13
mm; leaf blade lanceolate, 8-10.5 X 3.2-4. 5 cm,
papery, base ± rounded to subcuneate, symmetri¬
cal, apex long acuminate; veins 5, apical pair aris¬
ing 0.5— 1.5 cm above base, alternate, other pair
basal, reticulations raised adaxially, inconspicuous
abaxially, without evident glands, abaxially with
dense layer of ?sessile scales. Plants dioecious.
Spikes leaf-opposed. Male spikes 7-20 cm, 2-2.5
mm thick; peduncles to 1.8 cm; rachis densely
hairy, bracts orbicular, 1.3-1. 5 mm wide, peltate,
obscurely gland-dotted, subsessile. Stamens 3 or 4,
filaments longer than anthers, anthers ovoid, api-
cally confluent. Female spike (only one seen) short¬
er and thicker than male, ca. 2.5 cm long, 3.3 mm
thick; peduncle ca. 2 cm; rachis hairy, bracts sim¬
ilar to those of male, 1—1.2 mm wide. Ovary globose
to slightly oblate, slightly wider than bracts, free;
stigmas 4, linear-lanceolate, tightly adpressed to
ovary. Fruit not seen. Flowering May-June.

We are hesitant to describe collections lacking
fruits as a new species, but we believe that these
can only be related to Piper flaviflorum. The dif¬
ferences from that species are so clearly marked
that the two species must be treated as distinct. The
exact nature of the covering of the abaxial leaf sur¬
faces is difficult to make out but seems best re¬
garded as a layer of sessile scales.

Paratypes. CHINA. Yunnan: Shweli-Salwin divide,
ca. 25°45'N, 98°40'E, 2100-2450 m, June 1924, Forrest
24388 (E, K).

Piper wallichii (Miquel) Handel-Mazzetti, Symb.
Sin. 7: 155. 1929. Chavica wallichii Miquel,
Syst. piperac. 2: 254. 1843. Piper aurantiacum
Wallich ex C. DC., in A. DC., Prodr. 16: 357.
1868, nom. illegit. (cited Chavica wallichii as
a synonym). TYPE: Nepal. “In Nepalia,”
1821, N. Wallich 6658 A (lectotype, here des¬
ignated, K; isolectotypes. A, BM, P).

Piper aurantiacum var. hupeense C. DC., Notizbl. Konigl.
Bot. Cart. Berlin-Dahlem 6: 478. 1917. Piper wal¬
lichii var. hupeense (C. DC.) Handel-Mazzetti, Symb.
Sin. r. 155. 1929. TYPE: China. Hubei: A. Henry
3893 (lectotype, here designated. B; isolectotypes,
GH, K).

Piper emeiense Y. C. Tseng. Acta Phytotax. Sin. 24: 385,
fig. 3. 1986. Syn. nov. TYPE: China. Sichuan: Mt.
Emei, C. H. Hsiung 32838 (holotype, I BSC).

Piper ichangense C. DC., Notizbl. Konigl. Bot. Cart. Ber¬
lin-Dahlem 6: 480. 1917. TYPE: China. Hubei: Hu¬
peh occidentalis: prope Ichang, Oct., Wilson 499
"489" (holotype. B; isotypes, A, K, P).

Piper martinii C. DC., Notul. Syst. (Paris) 3: 41. 1914.
Syn. nov. TY PE: China. Guizhou (“Kouy Tch£ou”):
environs de San-pin. rocailles au-dessus de la gran¬
de grotte, L. Martin A- E. Bodinier 2298 (lectotype,
here designated, P; isolectotype, E).

This taxon was first named as “ Piper aurantia¬
cum'' by Wallich in his Catalog, under his number
6658 and including two collections, one from Nepal
collected by himself and the other from Sillet, now
in Bangladesh, collected by F. de Silva and W.
Gomez, but no description was given and that name
is invalid. The first valid name was that of Miquel,
who cited the original Wallich collections but did
not take up Wallich’s epithet. The collection from
Sillet was subsequently included by Hooker as a
syntype of P. rhytidocarpum, and thus Wallich’s own
collection from Nepal is the only sensible choice
as lectotype. The protologue of Piper aurantiacum
var. hupeense cited two collections: Wilson 499 in
CAL and A. Henry 3893 in B. Both are well rep¬
resented by duplicates in a number of major her¬
baria, but because it has not been possible to see
the cited collection of Wilson 499, and another
sheet of this collection was made the type of Piper

Volume 9, Number 2
1999

Gilbert & Xia
Piperaceae of China

197

ichangense, it seems advisable to select the Henry
collection as lectotypc. The type of P. ichangense
was given as Wilson “489.” There are rather com¬
plete sets of Wilson’s collections in A and K, so it
was puzzling when it was not possible to locate any
Piper with this number. The problem was solved by
examination of the holotype sheet in B, which has
two labels: an original with “499” and a second
label added in B with “489” clearly a copying er¬
ror. There are two collections of P. martinii cited
in the protologue, both in Paris and annotated by
De Candolle: Cavalerie 2387 and Martin & Bodi-
nier 2298. Other things being equal, we have se¬
lected the latter as lectotype in view of the epithet
used for the taxon, and because there is a duplicate
specimen in Edinburgh; Cavalerie 2387 seems to
be a unicate.

Tseng (in Chen et al., 1982) distinguished Piper
wallichii and P. martinii as follows:

la. Male spikes ea. 2X as long as leaf blades, pe¬

duncles 2.5— 3 X as long as petioles: petioles of
female bracts barely elongated in fruit, sparsely
pubescent . P. martinii

lb. Male spikes nearly as long as leaf blades, pe¬

duncles nearly as long as petioles; petioles of
female bracts elongated to 2 mm in fruit, densely
long white pubescent . P. wallichii

The differences in inflorescence dimensions are
not well defined and do not justify recognition of
distinct taxa. The bract characters sound extremely
distinctive, but we have not seen any material that
matches the description given for Piper wallichii.
This includes the types of P. wallichii and P. mar¬
tinii, which key out together in the above key and
are certainly nonspecific.

Piper emeiense was distinguished from P. walli¬
chii by the small leaves, which have palmate lateral
veins and a very thin indumentum, and by the short
peduncles. Other collections seen from near the
type locality. Mount Emei, show a transition be¬
tween juvenile climbing stems with foliage very like
that seen in the type collection of P. emeiense and
flowering stems with adult foliage typical of P. wal¬
lichii. We believe that there can now be little doubt
that C. H. Hsiung 32838 is material of P. wallichii
that has flowered in an unusually juvenile stage,
directly from the climbing stems without the pro¬
duction of normal flowering stems.

Piper wangii M. G. Gilbert & N. H. Xia, sp. nov.
TYPE: China. Yunnan: Jing Hong, Nan-hsien-
ho, 800 m, Oct. 1936, Wang C. W. 79466 (ho¬
lotype, A; isotypes, KUN, PE).

A Piper pubicatulo, foliis (5— )7(— 9)-nervis (non 5-ner-
vis) basibus manifeste obliquis (non aequilateris), stipiti-

bus bracteorum pilosis (non glabris), fruetibus maioribus
(plus quam 1.5 mm, non ca. 1 mm) differt: a P. boehmer-
iifolio et P. pedicellato , habitu scandentibus, foliis latior-
ibus (7—12 cm, non (2.5— )4— 9.5 cm), inflorescentibus fem-
ineis brevioribus (ca. 3 cm, non 6—14 cm) differt.

Climbers, glabrous except for rachis. Stems fine¬
ly striated, tuberculate. Prophylls ca. as long as
petioles; petioles 1—1.5 cm; leaf blade broad-ellip¬
tic to ovate, (12. 5 — ) 1 5 — 2 1 X 7-12 cm, papery, base
obliquely rounded-cuneate, apex acute to acumi¬
nate; veins (5 — )7( — 9), apical pair arising (2— )3— 6.5
cm above base, almost reaching leaf apex, l(or 2)
pair(s) basal, reticulate veins conspicuous, densely
finely glandular. Plants dioecious. Spikes leaf-op-
posed. Male spike not seen. Female spikes ca. 3
cm in young fruit; peduncle 1-1.4 cm; bracts pel¬
tate, orbicular, stalk pilose, 0.7-1 mm diam., mar¬
gin not entire, glabrous abaxially. Ovary distinct;
stigmas 3 or 4, short, ovate-lanceolate. Unripe
drupe subglobose, ca. 1.5 mm diam. Flowering
May-June, October.

Material of this species was originally identified
as Piper puhicatulum C. DC. Examination ol the
presumed holotype of P. puhicatulum, Bon 3139
bis, shows this to be a distinct species, easily dis¬
tinguished by the 5-veined leaves with symmetri¬
cal, cuneate bases, glabrous bracts, and densely
packed, globose, lruits ca. 1 mm in diameter. The
Chinese material is related to species such as P.
boehmeriifolium and P. pedicellatum, which have
similarly shaped and textured leaves, but it is eas¬
ily recognized by the combination of the climbing
habit, broader leaves, and short inflorescences. One
sheet ( Wang 76S84) states that the leaves are edi¬
ble.

Paratypes. CHINA. Yunnan: Lan Chang Xian, 15(H)
m, May 1936. Rung C. W. 76584 (A. II1SC).

Piper yui M. G. Gilbert & N. H. Xia, sp. nov.
TYPE: China. Yunnan: Kiukiang Valley, Chin-
gontum, 1300 in, 25 Sep. 1938, T. T. Yii 20434
(holotype. A; isotypes, E, PE).

A Piper nudibaccato, apicibus foliis rotundatis vel su-
bacutis, non acutis vel acuminatis, et bracteis obovatis ca.
1.5 mm longis, non orbicularis ca. 1 mm diam., differt.

Climbers, woody. Stems dark brown when dry,
striated, minutely hispidulous when young. Pro¬
phylls ca. as long as petioles; petiole 0.8— 1.4 cm,
hispidulous; leaf blade elliptic, to 9 X 4.5 cm, pa¬
pery, drying dark green, base obliquely cordate to
± rounded, one side to 3 mm higher, almost sym¬
metrical in uppermost leaves, apex rounded to
broadly acute; veins 5(-7), apical pair arising 0.3—
0.8(-1.2) cm above base, others basal, reticulations

198

Novon

slender, raised both sides in dry material, finely
glandular, very finely and sparsely puberulent on
veins, otherwise glabrous. Plants dioecious. Spikes
leaf-opposed. Male spike not seen. Female spikes
to 15 cm in fruit; peduncle to 5.3 cm, rachis and
undersides of bracts densely brownish hairy; bracts
obovate, ca. 1.5 X 1 mm. Ovary ovoid, partly con¬
nate to rachis; stigmas 3 or 4, ellipsoidal, reflexed.
Drupe (immature) ellipsoid, ca. 6X3 mm, slightly
rugulose, partly connate to rachis. Young fruit Sep-
tember-October.

The strictly elliptical leaves with rounded apices
of this species are immediately distinctive and are
not matched by any species from China and Indo¬
china. Piper hamiltonii C. DC. is the only species
with leaves similar in shape, but this poorly known
species from northern India has completely differ¬
ent inflorescences and the similarity is probably
due to convergence.

Acknowledgments. The authors thank the staff
of A, E, K, KUN, P, PE, and TAI for facilitating
access to their collections and for the loans of se¬
lected specimens. They also thank two anonymous
reviewers and the editors of Novon for numerous
minor corrections and suggestions.

Literature Cited

Hretschneider, K. 1898. History of European Botanical

Discoveries in China. Imperial Russian Academy of

Sciences, St. Petersburg.

Candolle, C. De. 1869. Piperaceae. In: A. I)e Candolle,
Prodromus Systematis Naturalis Regni Vegetabilis
16(1): 235—471. Masson, Paris.

- . 1910. Piperaceae. In: H. Lecomte (editor), Flore

general de I’lndo-Chine 5: 62-92. Masson et Cie, Paris.
Chen Pei-shan. Tseng Yung-chien & Zhu Pei-zhi. 1982.
Piperaceae. In: Tseng Yung-chien (editor), Flora Rei-
publieae Popu laris Sinicae 20(1): 11-77. Academia
Sinica, Beijing.

Creuter, W., F. R. Barrie, H. M. Burdet, W. G. Chaloner,
V. Demoulin, D. L. Hawksworth, P. M. Jprgensen, D. H.
Nicolson. P. C. Silva. P. Trchane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Hemsley, W. R. 1891. Piperaceae. In: F. R. Forbes & W.
R. Hemsley. 1886— 1904. An enumeration of all the
plants known from China proper, Formosa, Hainan, Co¬
rea, the Luchu Archipelago, and the Island of Hong¬
kong, together with their distribution and synonymy. J.
Linn. Soe. Rot. 26: 363-367.

- . 1905. List of the ( ienera and Species discovered

in China since the publication of the various parts of
the “Enumeration,” from 1886 to March 1904. alpha¬
betically arranged. J. Linn. Soc. Bot. 36: 451-530.
Hunter, W. 1807. Remarks on the species of Pepper which
are found on Prince of Wales Island. Asiat. Res. 11:
383-392.

Lin Tzer-tong & Lu Sheng-you. 1995. Peperomia sui and
Piper taiwanense, new species of Piperaceae from Tai¬
wan. Taiwania 40: 353—358.

Long, I). G. 1984. Piperaceae. In: A. J. C. Grierson & I).
G. Long, Flora of Bhutan 1(2): 342-351. Royal Botanic
Garden Edinburgh.

Miquel, K A. G. 1843—1844. Systema Piperaeearum. Kra¬
mers, Rotterdam.

Ouisumbing, F. 1930. Philippine Piperaceae. Philipp. J.
Sci. 43: 1-246.

Roxburgh, W. 1820. In: W. Carey (editor). Flora Indica:
Or descriptions of Indian Plants, vol. 1. Serampore.

Licaria caribaea (Lauraceae): A New Species from the Caribbean

Lowlands of Costa Rica

Jorge Gdmez-Laurito

Eseuela de Biologfa, Universidad de Costa Rica, Ciudad Universitaria 2060,

San Jose, Costa Rica

Alfredo Cascante

Herbario Nacional de Costa Rica, Museo Nacional, Ap. 749-1000 San Jose, Costa Rica

ABSTRACT. A new Licaria (Lauraceae) from the
Caribbean lowlands of Costa Rica is described and
illustrated here.

Resum EN. Una Licaria (Lauraceae) de las tierras
hajas del Caribe costarrieense se describe y se ilu-
stra en este artfculo.

Licaria Aublet (Lauraceae) is a Neotropical ge¬
nus of trees represented in Costa Rica by at least
eight species, some of them very poorly understood
(Burger & van der Werff, 1990). It is well charac¬
terized hy its androecium of three stamens with two
thecae each and the deeply cupulate fruiting re¬
ceptacle with two usually distinct margins on its
rim. An additional species from the Caribbean low¬
lands is described below.

Licaria caribaea Gomez-Laurito & Cascante, sp.
nov. TYPE: Costa Rica. Limon: Lfmite norte
del Parque Nacional Tortuguero, Cano Suerte,
9°36'42"N, 83°45'00"W. 0-10 m, 2 set. 1995
(fl, fr), A. Cascante & A. Ruiz 70S (holotype,
CR; isotypes, F, MO, USJ). Figure 1.

Species inflorescentia paniculata usque 32 cm longa,
multiflora; stamina 2-locellata. locellis extrorsis, apicali-
bus; cupula late conica. usque 42 mm longa, longitudi-
naliter cristata a eongeneribus tliversa.

Trees 12-15 m tall, 20—30 cm DBH; bark len-
ticellate; inner bark yellowish, hard; leafy branch-
lets minutely puberulous, grayish, solid, terete.
Leaves alternate; petioles 10—20 mm long, sulcate
above, glabrous; leaf blades 18—24 cm long, 8—11
cm broad, ovate-elliptic, tapering to an acuminate
apex, obtuse at the base, drying chartaceous and
brownish green, glabrous on both surfaces, with 8—
10 major secondary veins on each side, margin en¬
tire. Inflorescences paniculate, axillary and extra-
axillary to distal leaves, 13 — 18( — 32) cm long, 14-
lb cm wide, many-flowered, minutely puberulous,
peduncles 5—12 cm long, 1.5-2 mm thick with al¬

ternating lateral branches, tertiary branches 6—20
mm long, crowded by 6—9 flowers on pedicels 2-4
mm long. Flowers yellow, drying dark, 2-2.2 mm
long, 6 tepals, externally very short puberulous, in¬
ternally villous with whitish hairs; 3 fertile stamens,
1 mm long, strongly connivent forming a thick col¬
umn, filaments not differentiated, anthers ovoid
with 2 apical thecae and the small (0.25 mm)
valves with extrorse dehiscence near the apex of
the stamen; staminodes and glands minute, 0.3— 0.5
mm long, tightly pressed together and appearing
fused; ovary narrowly ovate, glabrous, pistil 1 mm
long, glabrous, style slender, simple, slightly ex-
serted between the fertile stamens. Fruits borne on
a broadly conical receptacle 3 1 —12 mm long, 15—
30 mm broad at the top, ca. 18 mm deep, longi¬
tudinally ridged, with two distinct margins on its
rim, lenticellate, green turning red; fruiting pedi¬
cels 5—20 mm long, 3^1 mm diam.; berry narrowly
ovoid to ellipsoid, 20—25 mm long, 10 mm diam.,
red to purple at maturity.

Licaria caribaea is currently known only from the
Caribbean coastal plain of Costa Rica (hence its
name), from 0 to 10 m elevation. It is found in the
very wet forest commonly called “yolillales,” “Trop¬
ical, tropical, very wet with no dry months” biotic
unit according to Herrera and Gomez (1993). This
new species is associated with Raphia taedigera
Martius (Arecaceae), Carapa guianensis Aublet
(Meliaceae), and the legumes Pentaclethra macro-
loba (Willdenow) Kuntze, Prioria copaifera Grise-
bach, and Crudia acuminata Bentham, among oth¬
ers. According to Kurz (1982), this new taxon
belongs to his subgenus Guianensis (which he
should have named subg. Licaria , because the type
of the genus is included in the subgenus, according
to van der Werff, in litt., 1996) and keys to L. ex-
celsa Kostermans. Kurz’s description of L. excelsa
is rather broad; however, the two species can be
readily separated hy the following key:

Novon 9: 199-201. 1999.

200

Novon

1 igure 1. Licana canbaea Gdmez-Laurito & Cascante (A. Cascante & A. Ruiz 705). — a. Flowering branch. — b.
Flower. — c. Fmit.

la. Trees of medium- to high-altitude habitat, (600-)

1 I (X)— 2300 m; leaves drying dark in color; inflo¬
rescences 5-20 cm long; fruits borne in a cu-
pulate or hemispherical receptacle 13—20 mm
long, 18-12 mm broad, not ridged . L. excelsa

lb. Trees of low-altitude habitat, 0-10 m; leaves dry¬
ing brownish green in color; inflorescences 13—

32 cm long, many-flowered; fruits borne in a

broadly conical receptacle 31-42 mm long, 15—

30 mm broad, longitudinally ridged caribaea

Paratypes. COSTA RICA. Limdn: same locality as
type, 3 Mar. 1996 (ster.). A. Cascante A: P. Opay 987 (CR,
USJ), 3 Mar. I *196 (ster.), A. Cascante & P. Opay 986 (CR,
USJ), 3 Mar. 1996 (fr), A. Cascante A: P. Opay 985 (CR,
USJ).

Volume 9, Number 2
1999

Gomez-Laurito & Cascante
Licaria caribaea from Costa Rica

201

Acknowledgments. We are grateful to Henk
van der Werff (MO) and to William C. Burger (F)
for critically reading the manuscript, and to Car¬
men Marin for the illustration. We also thank the
Canadian Organization for Tropical Education and
Rainforest Conservation (COTERC) for its support
during field trips to Cano Palma Biological Sta¬
tion.

Literature Cited

Burger, W. C. & H. van der Werff. 1990. Flora costari-
censis. Family #80. Lauraceae. Fieldiana Bot. New Ser.
23: 1-129.

Herrera, W. & L. 1). Gomez. 1993. Mapa de unidades
bidticas de Costa Rica. Escala 1:685.000. Instituto
Geografico de Costa Rica.

Kurz, H. 1982. Fortpflanzungsbiologie einiger Gattungen
neotropischer Lauracean und Revision der Gattung Li¬
caria (Lauraceae). Ph.D. Thesis, Universitat Hamburg.

New Combinations and a New Species in the Brazilian
Genus Chaetostoma DC. (Mierolicieae: Melastomataceae)

Cristiana Koschnitzke and Angela Borges Martins
Departamento de Botanica, Universidade Estadual de Campinas, CP 6109, 13083 970,

Campinas, Sao Paulo, Brazil

Abstract. Four new combinations in Chaetosto¬
ma (C. albiflorum, C. cupressinum, C. selagineum,
C. stenocladon,) are proposed, and a new species,
Chaetostoma flavum , endemic to the Chapada dos
Veadeiros, Goids, Brazil, is described. Tbe new
species can be distinguished by its 4— 5-locular
ovary and antisepalous stamens with prolonged ap¬
pendages.

Chaetostoma DC. is a genus endemic to Brazil
comprising 1 1 species that occur in campo rupes-
tre, campo de altitude, and campo cerrado vegeta¬
tion in the states of Goids, Distrito Federal, Minas
Gerais, with a less considerable representation in
Bahia, Rio de Janeiro, Sao Paulo, and Parana. This
genus can be distinguished within the tribe Mi-
crolieieae, especially with regard to Microlicia D.
Don, to which it has the closest affinity, by the com¬
bination of the following diagnostic features: cari¬
nate, imbricate, and pungent leaves without trans¬
lucent dots; the mature capsule longer than the
hypanthium; and a crown of triehomes around the
outward apex of the hypanthium. The distinction
between Chaetostoma and Microlicia has been
made by Cogniaux (1883, 1891), in the last revision
of these genera, on the basis of just one character:
the morphology of the stamens, considered equal in
both cycles in Chaetostoma and very unequal in
Microlicia. This character has proven to be unre¬
liable, and the most striking difference between
these genera is in the possession of the crown of
triehomes on the hypanthium in all species of
Chaetostoma , which is absent in all species of Mi¬
crolicia. In a recent revision of Chaetostoma
(Koschnitzke, 1997), some taxonomic novelties
were recognized, and it was found necessary to pro¬
pose some new combinations and a new species for
the genus.

Chaetostoma alhifloruin (Naudin) Koschnitzke &
A. B. Martins, stat. nov. Basionym: Chaetosto¬
ma pungens DC. var. albiflorum Naudin, Ann.
Sci. Nat. Bot. Ser. 3, 3: 191. 1845. TYPE: Bra¬
zil. Minas Gerais: Serra de Carrancas, 1822,
A. St-/lilaire s.n. (holotype, P not seen).

Novon 9: 202-204. 1999.

Chaetostoma pungens DC. var. pallidiflorum Cogniaux, in
Martins, FI. Bras. 14: 32. 1883. Syn. nov. TYPE:
Brazil. Minas Gerais: Serra de Itatiaia, Mar. 1839.
Riedel s.n. [Martii Herbar. Florae Brasil no. 922|
(lectotype, selected here. BB: isolectotypes, BM. G,
K. W).

Chaetostoma cupressinum (D. Don) Koschnitzke
& A. B. Martins, comb. nov. Basionym: Mi¬
crolicia cupressina D. Don, Mem. Wern. Soc.
4: 302. 1823. TYPE: Brazil. Minas Gerais: 6
May 1819, Sellow 1329 (holotype, BM; iso¬
type, BR).

Chaetostoma luteum Cogniaux. in Martius, FI. Bras. 14:
589-590. 1888. Syn. nov. TYPE: Brazil. Minas Ger¬
ais: Sao Joao del Key. Serra do Lenheiro, 24 June
1887, Glaziou 16769 (holotype, B destroyed; lecto¬
type, selected here. BR; isolectotvpes, C. F. G, K.

MO. P. KB).

Chaetostoma Intern Cogniaux var. quadrifarium Cogniaux,
in A. DeCandolle & C. DeCandolle, Monogr. Phan.
7: 29. 1891. Syn. nov. TYPE: Brazil. Minas Gerais:
Sao Joao del Key, Serra do Lenheiro, 24 Jan. 1889,
Glaziou 17503 (holotype, BR; isotvpes. BR. C. F. G,
k. P. R).

Chaetostoma trauninense Cogniaux, in A. DeCandolle &
C. DeCandolle. Monogr. Phan. 7: 30, 1891. Syn. nov.
TV PE: Brazil. Minas Gerais: Sao Joao del Key. Serra
do Lenheiro, 24 June 1887. Glaziou 16769a (lecto¬
type, selected here, BR; isolectotypes, B. C, F. k,
KB).

Chaetostoma joannae Cogniaux, in A. DeCandolle & C.
DeCandolle, Monogr. Phan. 7: 28. 1891. Syn. nov.

I V PE: Brazil. Minas Gerais: Serra do Lenheiro,
1889. Glaziou 17504 (lectotype, selected here, BR;
isolectotypes, B. BR, C, F, G, k, P).

Chaetostoma selagineum (Naudin) Koschnitzke
& A. B. Martins, comb. nov. Basionym: Mi¬
crolicia selaginea Naudin, Ann. Sci. Nat. Bot.
Ser. 3, 3: 188. 1845. TYPE: Brazil. Goitls:
Chapadao, A. St-Hilaire 550 (holotype, P not
seen; isotype, F).

Chaetostoma stenocladon (Naudin) Koschnitzke
& A. B. Martins, comb. nov. Basionym: Mi¬
crolicia stenocladon Naudin. Ann. Sci. Nat.
Bot. Ser. 3, 3: 188. 1845. TYPE: Brazil. Goids:
Serra dos Pirineus, 17 June 1819,4. St-Hilaire
s.n. (holotype, P not seen; isotype, F).

Volume 9, Number 2
1999

Koschnitzke & Martins
Chaetostoma from Brazil

203

Figure 1. Chaetostoma flavum Koschnitzke & Martins. — A. Flowering branch. — B. Branch apex. — C. Leaf (adaxial
surface). — D. Flower. — E. Calyx lobes, apex of the hypathium outward with trichomes. — F. Petal. — G. Stamens,
antipetalous (left) and antisepalous (right). — H. Ovary and style. Drawn from holotype.

204

Novon

Microlicia amaroi Brade, Arq. Jard. Bot. Bio J. 16: 29.
1959. Syn. nov. TYPE: Brazil. Goi&s: Serra dos Pi-
rineus, 17 Feb. 1956, Macedo 4360 (holotype. Her¬
barium A. C. Brade & A. Macedo not seen; isotypes,
K, S).

Chaetostoma flavum Koschnitzke & A. B. Mar¬
tins, sp. nov. TYPE: Brazil. Goids: Alto Par-
afso, Chapada dos Veadeiros, 18 Mar. 1976,
Semir 712 (holotype, UEC; isotypes, UEC,
MO). Figure 1.

Haee species inter affines petalis llavis, hypantho sub-
tiliter striato, triohomatibus brevibus incrassatis superne
circumcirca ornato, connectivo antherarum maiorum longe
producto et praecipue ovario 4— 5-loeulari distinguitur.

Erect and caespitose subshrub 30-100 cm tall,
branching dichotomous or trichotomous, branches
fastigiate, glabrous, and articulate. Leaves cadu¬
cous with age, imbricate, adpressed, erect, sessile
and semiamplexicaul, blade 3-7 X 0.8-1 mm,
keeled, coriaceous, triangular-lanceolate, acute at
the apex, pungent, margins entire or serrulate-cil-
iate, abaxial surface glabrous, adaxial surface with
short trichomes, 5-nerved. Flowers 5-merous, soli¬
tary, terminal, without bracts, sessile; hypanthium
2. 5-4.5 mm long, oblong-campanulate, ribbed, gla¬
brous or rarely pubescent, apex outward with short
and thick adpressed trichomes; calyx lobes 3-4 mm
long, 1-2 mm wide at the base, persistent, erect,
subcarinate, triangular-lanceolate, acute at the
apex, pungent, margins inconspicuously serrulate-
ciliate, abaxial surface glabrous, adaxial surface
with short trichomes, 5-nerved. Petals 8-11 X 4—
7 mm, irregularly and broadly obovate, acuminate
at the apex, yellow; antisepalous stamens five, 9-
13 mm long, filaments 4—5 mm long, anthers te-
trasporangiate, 2-3 mm long, yellow, oblong with a
small and single pore, dorsally inclined in both se¬
ries, connective 3-5 mm prolonged below the the¬
cae, appendage 1-2 mm, bilobed; antipetalous sta¬
mens five, 5—7 mm long, filaments 4 mm long,
anther 1.5— 2.5 mm long, similar to the antisepal¬
ous, connective 1 mm prolonged below the thecae,
appendage small, tubereulate; ovary 4-5-locular,

2-4 mm long, oblong, glabrous, style 5.5-7 mm
long, stigma punctiform. Fruit capsular, 4—5 mm
long, completely covered by the hypanthium and
loculicidally dehiscent in the upper part; seeds nu¬
merous, 0.6-0.65 mm long, slightly curved, testa
reticulate-foveolate, anticlinal walls undulated.

Chaetostoma selagineum is closely related to C.
flavum because of its acute and pungent leaves,
slightly ribbed hypanthium, yellow petals, and con¬
nective of the antisepalous stamens with prolonged
appendages. Chaetostoma flavum has a sessile hy¬
panthium, 4— 5-locular ovary, and a seed testa with
undulated anticlinal walls, while Chaetostoma se¬
lagineum has a pedunculate hypanthium, 3-locular
ovary, and seeds with straight anticlinal walls.

Chaetostoma flavum is endemic to the Chapada
dos Veadeiros, Goids, Brazil, in areas with stony
and sandy soils. It has been collected flowering
from January until August and with fruits in Jan¬
uary and June.

Paratypes. BRAZIL. Goias: Alto Parafso: rio Pigarao,
Glaziou 21212 (BB, C, G, P); Chapada dos Veadeiros,
Anderson 6707 (F, MO, BB, US), Dawson 14692 (US),
Gates & Estabrook 13 (SP), Irwin et al. 24676 (F. MO,
US), Oliveira & Anderson 471 (CAS, MBM); Serra dos
Veadeiros, Glaziou 1895 (P). Niquelandia: Estrada da
mina de nfquel, Fonseca et al. 167 (UEC).

Acknowledgments. The authors thank Volker
Bittrich for critically reviewing the manuscript. The
first author was supported by a grant received from
Fundagao da Coordenagao de Aperfeigoamento de
Pessoal de Nfvel Superior (CAPES).

Literature Cited

Cogniaux, A. 1883. Melastomaceae. In: C. F. P. von Mar-
tius & A. G. Eichler (editors), Flora Brasiliensis 14(3):
1-510.

- . 1891. Melastomaceae. In: A. DeCandolle & C.

DeCandolle (editors), Monographiae Phanerogamarum
7: 1-1256.

Koschnitzke, C. 1997. Revisao taxonomica do genero
Chaetostoma DC. (Microlicieae: Melastomataceae).
Ph.D. Thesis, Universidade Estadual de Campinas, Sao
Paulo.

A Revised Taxonomy for Two North American Rhynchospora
(Cyperaceae) and for Two North American Xyris (Xyridaceae)

Robert Krai

Herbarium (VDB), % Botanical Research Institute of Texas, Inc.,
Fort Worth, Texas 76102-4060, U.S.A.

Abstract. Changes in taxonomic rank are pro¬
posed for two North American Rhynchospora (Cy¬
peraceae) and for two Xyris (Xyridaceae): Rhyn¬
chospora recognita (Gale) Krai, Rhynchospora
harveyi W. Boott var. culixa (Gale) Krai, Xyris lax-
ifolia Martius var. iridifolia (Chapman) Krai, and
Xyris stricta Chapman var. obscura Krai. The mor¬
phologic basis for such changes is given.

In the process of preparing treatments of Rhyn¬
chospora (Cyperaceae) and Xyridaceae for the Flo¬
ra of North America , and for the sake of a consistent
taxonomy in both, the following efforts are made.

1. Rhynchospora recognita (Gale) Krai, stat.
nov. Rhynchospora globularis (Chapman) Small
var. recognita Gale, Rhodora 46: 245. 1944.
TYPE: U.S.A. Virginia: Sussex Co., dry pine-
lands, 4 mi. NW of Waverley, 26 July 1936,
Fernald & Long 6070 (holotype, GH; isotype,
PH).

Whatever its name or rank, Rhynchospora recog¬
nita , as proposed herein, had, from the time of
Muhlenberg and Elliott to the Gale and Fernald
revisions, been thought to be R. cymosa. As care¬
fully pointed out by Gale (1944), the specific epi¬
thet “cymosa,” based on Schoenus cymosus Willde-
now (actually a specimen of Scirpus polyphyllus
Vahl), had been consistently misapplied by authors
since Elliott (1821), who used it for the already
published R. glomerata (L.) Vahl. Thus, it was
Gale’s decision to solve the problem by utilizing
Small’s (1933) R. globularis (Chapman) Small, itself
based on Chapman’s description of R. cymosa var.
globularis Chapman (1860: 525). This, in Gale’s
treatment and in others since, is the shorter, more
slender plant that provides a working concept for
typical R. globularis. Gale also then assigned a new
name, “recognita,” to the larger, coarser plant pre¬
viously (and erroneously) named R. cymosa, making
it a variety of R. globularis along with the south¬
eastern U.S. and West Indian plant formerly called
R. pinetorum Britton & Small. Ktikenthal (1950) in
his revision of R. sect. Globulares adopted the es¬

sentials of Gale’s treatment of Rhynchospora glob¬
ularis varieties, simply adding two more to the list,
with R. obliterate Gale and R. saxicola Small thus
becoming R. globularis var. obliterata (Gale) Kii-
kenthal and R. globularis var. saxicola (Small) Kti-
kenthal.

After many years of field and herbarium expe¬
rience with this complex of Rhynchospora, and after
recent examination of many large loans, I have con¬
tinued to note that Gale’s variety recognita is ac¬
tually a species distinct from R. globularis and its
varieties. Authors such as Muhlenberg, Torrey,
Gray, Chapman, and Small were well aware of it as
a species even if they did not realize what the cor¬
rect name for it was. Also, thanks to Gale and Kii-
kenthal, we have a better idea of what the real R.
globularis is.

I have: (1) in Table 1 prepared a list of several
character states by which Rhynchospora recognita
is compared with R. globularis and its varieties; (2)
illustrated an example of R. recognita (Fig. 1) and
R. globularis together with fruits of its varieties
(Fig. 2); and (3) shown the eastern North American
mapped distributions for the species and varieties
in Figure 3A-D (R. recognita, R. globularis var.
globularis, R. globularis var. pinetorum, and R.
globularis var. saxicola, respectively).

From Table 1, it can be seen that Rhynchospora
recognita is the more robust, taller, stilfer plant,
with bracts and bracelets more exserted, with rust-
red spikelets mostly longer, and with longer fruits
than is true for R. globularis and its varieties. From
the maps in Figure 3 it can be seen that R. recog¬
nita has the widest range, and that the R. globularis
varieties are largely nested within it. The maps in
Figure 3 do not show total distributions: R. recog¬
nita and R. globularis var. globularis have popu¬
lations in northern California, and the former is
well represented in the West Indies, Mexico, and
Mesoamerica; R. globularis var. pinetorum extends
south from Florida into the Antilles. However, the
area of greatest sympatry is within the Coastal Plain
of the eastern U.S., particularly in moist sandy
peats or sandy clays of flatwoods, savanna, ditches.

Novon 9: 205-219. 1999.

206

Novon

'fable 1 . Comparison of 1

l character states in Rhynchospora recognita and R. globularis.

Rh ynchospora recognita

Rhynchospora globularis and vars.

Habit

perennial, culms stiff, erect. 6-10 dm

perennial, culms slender, lax to stiff,
erect to excurved or spreading, (1—)
2-6(-8) dm

I, eaves

blades 2—5 mm wide

blades 1-3 mm wide

Inflorescence compounds,

compounds of clusters (fascicles) (1—)

compounds of clusters (fascicles) (1-)

spikelet clusters

2^4(-5), the terminal and the lateral
exceeded by subtending bract; spike¬
let clusters (fascicles) dense, exceed¬
ed by involucral bractlets giving
them a “bristly” look

2-3(-4), the terminal compound usu¬
ally exceeding the subtending bract;
spikelet clusters of fewer spikelets,
arranged in more diffuse compounds,
exceeding their subtending bractlets.
thus lacking a "bristly” look

Spi kelets

ovoid to narrowly ovoid, 5-4 nun. acute
to acuminate, red-brown

ovoid to subglobose, 2.5 — 3( — >1-) mm.
broadly acute to obtuse, brown to
deep brown

Fertile scales

ovate, 2.5—3 mm, convex, acute to ob¬
tuse or emarginate, midcosta often
excurrent as cusp or arista

broadly ovate to orbicular, 1.7— 2.3 mm,
deeply convex, obtuse to rounded or
emarginate, midcosta mostly includ¬
ed, sometimes excurrent as apiculus

Stamens, anthers

2—3. 2 mm

1-3, 1.5(— 2) mm

Fruit

body obovoid-lenticular, tumid

body broadly obovoid-lenticular to sub-
orbicular-lenticular, tumid

Fruit length (including tuber¬
cle)

2—2.3 mm

1 .5—2 mm

Tubercle

compressed-conic, (0.4— )0.5— 0.6(-0.8)
mm

depressed-conic to almost patelliform,
never exceeding 0.5 mm

Sculpture of fruit body

transverse rows ol vertical, narrowly
rectangular cancellae, contiguous
endwalls producing sharp, wavy
transverse rugosity

transverse rows of vertical, narrowly
rectangular cancellae, contiguous
endwalls producing sharp, wavy,
transverse rugosity, or merely isodi-
ametrically reticulate-cancellate

Perianth

bristles usually 6, extending to fruit
midbody

bristles usually 6, various lengths

and shorelines. All of these particular species and
varieties are aggressively weedy and are quick to
invade timbered lowlands, areas that have under¬
gone mechanical or fire disturbance, low areas in
abandoned fields, and sandy shores of artificial im¬
poundments. It is not rare to see a majority of these
taxa mingling in the same site; in fact, I have col¬
lected all of them together on and around outcrop¬
pings of the Altamaha Grit in southern Georgia
pinelands. In such cases there are no evidences of
intergradation, and R. recognita, taller and more
robust, puts its bristly-looking rusty spikelet clus¬
ters at a level well above the lower, more lax and
spreading culms of darker-spikeleted R. globularis.

Since the type chosen by Gale is represented by
only two specimens, I have distributed two repre¬
sentative sets, namely: U.S.A. Alabama: Cullman
Co., R. Kml 84838 (APSC, BM, BRCH, CLEMS,
CM, CTB, F, FLAS, FSU, GA, GH, K, KANU,

MICH, MEM, MO, MSC, NCU, NY, OS, TENN,
TEX, UNA, USCH, VDB, VSC, WAT, WILLI, WIS);
Shelby Co., R. Krai 80984 (BM, BRCH, CH, CM,
CTB, FSU, GH, K, KANU, MICH, MO, MU, M,
NCU, PH, SMU, VDB, VPI, VSC, WILLI, WIS).

2. Rhynchospora harveyi W. Boott var. eulixa
(Gale) Krai, comb. nov. Basionym: Rhyncho¬
spora eulixa Gale, Rhodora 46: 235. 1944.
Rhynchospora grayii Kunth var. eulixa Kiiken-
thal, Bot. Jahrb. Syst. 75(1): 120. 1950. TYPE:
U.S.A. Georgia: Tift Co., Irby, 28 Aug. 1890,
S. M. Tracy 1498 (holotype, US).

Gale described this plant as a species most
closely resembling Rhynchospora harveyi, differing
from it mostly in its more slender, shorter-leaved
culms, its smaller, more compact spikelet clusters,
and its fruits tending to be obovoid rather than el-

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

207

Figure I. Rhynchospora recognita (Krai 84838). — a. Habil sketch. — b. Leaf at sheath-blade junction, abaxial (left),
adaxial (right). — c. Leaf midblade, adaxial (left), abaxial (right). — d. Leaf apex. — e. Inflorescence, apical compound.
— f. Spikelet. — g. I,owest sterile scale (upper right); upper sterile scale (lower right); fertile scale (upper left). — h.
Fruit, achene showing perianth, tubercle, hypogynous stamen (right); fruit outline (left).

208

Novon

Figure 2. Rhynchospora globularis var. globularis ( Krai 84527). — a. Habit sketch; scale bar = 1 dm. — b. Leaf
blade, sheath junction. — c. Leaf' midblade, abaxial (left), adaxial (right). — d. Leaf apex. — e. Inflorescence, apical
compound. — f. Spikelet. — g. Fertile scale, adaxial view (right), cross section with enclosed fruit, sketch (left). — h.
Lowest sterile scale. — i. Distal part of filament, with anther. — j. Fruit. — k. Fruit of R. globularis v ar. pinetorum (Krai
7031). — 1. Fruit of R. globularis var. saxicola (Krai 80854).

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

209

Figure 3. Distribution maps. Dots indicate taxa occurrence by county in eastern North America. — A. Rhynchospora
recognita (Gale) Krai. — B. Rhynchospora globularis (Chapman) Small var. globularis. — C. Rhynchospora globularis
var. pinetorum (Britton & Small) Gale. — I). Rhynchospora globularis var. saxicola (Small) Kiikenthal. — E. Rhynchospora
grayii Kunth. — F. Rhynchospora harveyi W. Boott var. harveyi. — G. Rhynchospora harveyi var. culixa (Gale) Krai.

210

Novon

Table 2. Conq

>arison ol 10 character stales in

three Rhynchospora taxa

Rhynchospora harveyi

Rh ynchospora grayii

var. culixa

Rh ynchospora h arveyi

Plant height

5—13 dm

5-7 cm

5—1 1 dm

Principal leaves.

2-4 dm X (1 — ) 1 .5 — 3 mm

1 .5—3 dm X 1 .5—2 mm

1.5—3 dm X 1.8—3 mm

length X width

Spikelet clusters

1— f. turbinate to hemispheric.

2—4, turbinate to hemispheric,

2-^1, turbinate to hemispheric,

uppermost exceeding or ex-

uppermost sparse, exceeding

uppermost dense, exceeded

ceeded by subtending seta-

subtending setaceous bract

by 1 or more subtending

ceous bract

bracts

Spikelet shape.

ovoid to lanceolate-ovoid, 5—

narrowly ovoid, 5—6 mm acute.

ovoid to broadly ovoid, 3-3.5

dimensions.

6(— 7.5) mm narrowly acute.

pale brown

(-4) mm, broadly acute, pale

color

mostly dull pale brown

brown to red-brown

Fertile scales

broadly elliptic to ovate or or-

broadly ovate to suborbieulate.

broadly ovate to suborbieulate.

biculate, 4-5 mm

2.5— 3.5 mm

2.5— 3.5 mm

Stamen number.

3( — 6), ca. 3 mm

3, 2—2.2 mm

3. 1.5—2 mm

anther length
Fruit body

broadly obovoid to subglobose

obovoid, ca. 1.5— 1.6 mm len-

obovoid to broadly ellipsoid.

or broadly ellipsoid. 2—2.5

ticular distally

1.5— 1.7 mm, nearly round in

mm, mostly tumidly lenticu¬
lar to nearly round in cross

cross section

section

Fruit surface

dark brown, with transverse

brown, with transverse wavy

dark brown to brown, with un-

wavy lines of tiny, raised or

lines of small vertical, short-

dulant-transverse lines of

pitlike cancellae

rectangular pits alternating

small, pitlike or raised can-

with low. smooth wavy ridges

cellae, alternating with nar¬
row low ridges, or very finely
cancel late

Tubercle

buttressed, depressed-conic.

buttressed, conic. 0.3 mm high

buttressed, conic, 0.3-0.5(— 0.7)

0.3 — ().5( — ().7) mm high

mm high

Perianth bristles

mostly 6, reaching at least to

6, none reaching past fruit

(l-)3-6, 3 usually rudimentary.

tubercle base

midbody

none reaching past fruit
midbody

lipsoid, less swollen distally, and with a broader,
paler transverse rugosity.

Six years later, Kiikenthal took a much more
conservative approach, reducing the four taxa treat¬
ed by Gale in her series Harveyae to two species,
Rhynchospora megalocarpa A. Gray and R. grayii
Kunth, placing both R. harveyi and R. culixa under
the latter as varieties.

In my opinion, this shared not just hy Gale hut
by most of those who have field experience with
these plants, Rhynchospora harveyi is a well-rec¬
ognized, wide-ranging, and distinctive species. It is
frequent to abundant over many physiographic
provinces in North America, from the Great Plains
eastward through much of the Gulf and Atlantic
Coastal Plain, inland through some Appalachian
provinces into the Central Plains (see Fig. 3F).
I herefore, it is well represented in systematics col¬
lections and exhibits considerable variation. One of
these variants, consisting of two collections, one

from Georgia the other from northern Florida, was
noted by Gale as specifically distinct.

Some realignments appear to be in order. The
relationships of taxa within Gale’s series Harveyae
appear, on a basis of comparisons of dried speci¬
mens and of living plants, to be much as she had
them, and Rhynchospora harveyi should be extri¬
cated from R. grayii. The only change from Gale’s
work would be simply to show a closer taxonomic
relationship of R. culixa to R. harveyi than she pro¬
posed, namely that its relationship be varietal.

In Table 2, Rhynchospora grayii, R. harveyi var.
culixa, and R. harveyi var. harveyi are compared
morphometrically, and all three are illustrated in
Figures 4, 5, and 6, respectively. Data in Table 2
suggest that (1) Rhynchospora harveyi stands dis¬
tinctly apart from R. grayii as a species and (2) R.
culixa overlaps too much in character with R. harv¬
eyi to be considered as more than a variety of it.

Gale based Rhynchospora culixa upon only two

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

211

d

4mm

ldm

C \ \

IF

/

a '

if

/

JS mm

F igure 4. Rhynchospora grayii Kunth ( Krai 39617, 84512). — a. Habit sketch. — b. Leaf sheatli-blaile junction. — c.
Leaf midblade, abaxial (above), adaxial (below). — d. I,eaf apex. — e. Upper portion of inflorescence. — f. Spikelet. — g.
Sterile scale. — h. Fertile scale. — i. Fruit. — j. Distal part of filament with attached anther.

212

Novon

Figure 5. Rhynchospora harveyi var. culixa (Krai 80621, 80654). — a. Ilal>it sketch. — 1). I.eaf sheath-blade junction,
abaxial view. — c. Leaf sheath-blade junction, adaxial view. — d. Leaf midhlade, adaxial (left), abaxial (right). — -e. Leaf
apex. — f. Inflorescence branch, showing bract base and prophyll. — g. Apical inflorescence group. — h. Spikelet (right),
lowest sterile bract (left). — i. Upper sterile bract (left), fertile bract (right). — j. Fruit, face view (right), fruit outline
(left).

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

213

Figure 6. Rhynchospora harveyi var. harveyi (Krai 35104, 53164). — a. Habit sketch. — b. Leaf sheath-blade junction,
adaxial view. — c. Sectors of leaf midblade, adaxial (below), abaxial (above). — d. Ueaf apex, adaxial sector (below),
abaxial sector with tip (above). — e. Upper part of inflorescence, apical compound. — f. Spikelet. — g. Fertile scale. — h.
Fruit. — i. Distal part of filament, with attached anther.

214

Novon

Table 3. Comparison

i of 1 1 character states of three

Xyris taxa.

Xyris ambigua

Xyris stricta var. obscura

Xyris stricta var. stricta

Plant height

3-10 dm

2-9(-10) dm

5-9(-10.4) dm

Leaf orientation

in wide fans

in narrow fans

in narrow fans

Leaf length; blade

(5 — ) 1 () — 40(— 50) cm; (2-)3-7

( 1 5-)20-40(-44) cm; (2-)2.5-

20-60 cm; 3-8 mm

width

(-10) mm

3(— 5) mm

Leaf sheath color; leaf

stramineous to pale brown;

purplish to reddish; deep

purplish to reddish; deep

blade color

olive or yellow green

green

green

Leaf blade margins;

scabro-ciliolate; tuberculate-

strongly papillate; densely

smooth to low papillate; low

distal scape costae

scabridulous

papillate

papillate to smooth

Spike symmetry; spike

commonly ovoid to ellipsoid,

commonly ovoid to ellipsoid-

laneeolate-cylindric to cylin-

length, apex

rarely short cylindric; 1 —

cylindric, rarely cylindric;

dric; (1 .5— )2— 3(-3.5) cm.

2(— 3) cm, acute

l-2(— 2.7) cm, blunt

blunt

Fertile bracts; outline.

broadly obovate, rarely subor-

suborbicular, rarely broadly

suborbicular, rarely broadly

length

bicular; 5—8 mm

obovate; 5— 6.5(-7) mm

obovate; (5.5-)6-7(-7.5)

mm

Fertile bracts, color

matrix lustrous, base deep

matrix lustrous, base casta-

matrix lustrous, base casta-

brown shading distally to

neous, shading distally to

neous, shading distally to

pale brown or stramineous

brown

brown

lateral sepals

lanceolate-curvate, 4—7 mm.

lanceolate-curvate, (4.5-)5—

lanceolate-curv ate, 5—6.5

acuminate, keel equaling

6(— 7) mm, acute, keel wid-

mm, keel wider than wing

wing

er than wing

Petal blades

obovate, 8—10 mm

obtriangular, ca. 5 mm

obtriangular, ca. 5 mm

Seeds

ovoid to broadly ellipsoid.

narrowly ellipsoid to cylin-

narrowly ellipsoid to cylin-

0.5(-0.6) mm, translucent

dric, 0.6-0. 7(-0.8) mm far-

dric, (0.55-)0.6-0.8 mm.

inose

farinose

specimens. I have now discovered seven Georgia
populations, confirming Gale’s species. These, all
represented by mounted duplicates in VDB at
BRIT, are: Berrien Co. (Krai 24255), Colquitt Co.
(Krai 80599, 80621), Cook Co. (Krai 80652), Lee
Co. (Krai 65385), Worth Co. (Krai 80769, 80780);
see Figure 3G. Ranges for R. grayii and R. harveyi
var. harveyi are given in Figure 3E and F.

Rhynchospora grayii differs also in its ecology
from R. harveyi, being an inhabitant of droughty
sandhills, mostly in the longleaf pine-deciduous
scrub oak systems. Rhynchospora harveyi var. harv¬
eyi frequents a wide variety of ecological systems
so long as they are at least seasonally moist, from
basic prairie soils to upland oak-pine sands and
clays, pine savanna or flatwoods, occasionally also
in sandy ecotones between sandhills and boggy
slopes and bottoms. These last systems are the ones
in which I have encountered R. harveyi var. culixa.

3. Xyris laxifoiia C. Martius var. iridifolia

(Chapman) Krai, stat. nov. Xyris iridifolia
Chapman, FI. S. U.S. 501. 1860. TYPE:
U.S.A. Florida: Franklin Co., “Apalachicola,”
A. W. Chapman s.n. (holotype, NY).

Those who have seen living or preserved speci¬

mens of the neotropical Xyris laxifoiia C. Martius
[1841: 58; TYPE: Brazil. Mart. Herb. no. 540 (ho¬
lotype, M)] and the mostly North American X. iri¬
difolia Chapman are struck by how similar the two
are. In fact, Malme, long the authority on Xyrida-
ceae, made such observations years ago (Malme,
1913, 1937). Both are robust plants with leaf
blades at least 1 cm wide, with leaf sheaths strongly
tinged with red or purple; both have broad, some¬
what compressed or even acipital, smooth scapes
and large (1—4 cm) multiflowered brownish spikes,
with contrasting paler greenish dorsal areas. The
lateral sepals of both taxa are free, of comparable
color and length, oblanceolate, included, mostly el-
liptic-oblanceolate, with thin, narrow, lacerate-cil-
iate or lacerate-fimbriate keels. Flowers of both taxa
are small and open early in the morning. Seeds of
both taxa are opaque, mostly farinous. Thus, when
the two are compared, one sees that overlapping
limits of variation are significantly many. The fea¬
tures that distinguish the two, for example, spike
outline, seed outline and degree of opacity, relative
widths of sepal keels, and location of fimbriae and/
or ciliae along the keel, are such as to make vari¬
eties in a genus such as Xyris. What has been
called X. iridifolia ranges through the Coastal Plain

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

215

Figure 7. Xyris ambigua ( Krai 19318). — a. Habit. — b. Leaf apex. — c. Leaf blade-sheath junction (below), enlarged
leaf blade margin (above). — d. Base of leaf sheath. — e. Spike. — f. Fertile bract. — g. Lateral sepal. — h. Petal blade,
stamen. — i. Staminodium (right); enlarged view of beard hair apex (left). — j. Stylar apex. — k. Capsule. — 1. Seed.

216

Novon

o-6 mm

Figure 8. Xyris stricta var. obscurn ( Krai 87830). —a. Habit sketch. — b. Leaf apex. — c. Leaf blade-sheath junction.
— H. Leaf sheath base. — e. Spike. — 1. Scape apex, basal spike bracts (above); sector of scape at midscape (below),
g. Fertile bract. — h. Lateral sepal. — i. Petal, stamen. — j. Staminodium. — k. Capsule, adaxial view. — I. Seed.

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

217

Figure 9. Xyris stricta var. stricta (Krai 85336). — a. Habit sketch. — b. Leaf apex. — c. Leaf blade-sheath junction
(above); small sector of leaf blade margin (below). — d. Leaf sheath base. — e. Spike. — f. Sector of midscape. — g.
Fertile bract. — h. Lateral sepal. — i. Petal, stamen. — j. Staminodium (left); enlarged beard hair apex (right). — k.
Capsule (left); stylar apex (right). — 1. Seed.

218

Novon

of the southeastern U.S. from North Carolina to
northern Florida, west into eastern Texas, disjunctly
in contiguous physiography inland to Oklahoma,
Arkansas, Tennessee, and locally in the Piedmont
of Georgia and the Carolinas. It reaches the Neo¬
tropics in the states of Tobasco and Veracruz, Mex¬
ico, and has some recorded stations in Costa Rica
and Honduras. Plants identified as X. laxifolia
overlap X. iridifolia types only in Costa Rica, be¬
coming increasingly common southward in Panama,
thence southward through most of lowland South
America. The two have a similar habitat, namely
acidic wet sands and peats of sour swamps and
bogs, seeps, shallows, and banks of ponds and dis¬
turbed sandy wetlands.

There are a few constant differences, insufficient
to distinguish two species but adequate in light of
distributional information to propose them as geo¬
graphic co-varieties.

4. Xyris strieta Chapman var. obseura Krai, var.
nov. TYPE: U.S. A. Alabama: Washington Co.,
12.2 mi. NNW of Citronelle, E side of US 45,
fine sandy peat of pine-Willow oak-Nyssa bi¬
flora Hat adjacent to Longleaf pine-evergreen
send) Hats, 6 Sep. 1998, R. Krai 87830 (ho-
lotype, VDB; isotypes, BAYLU, BM, BRCH,
CLEMS, CM, CTB, DUKE, FLAS, FSU, GA,
GH, JSU, KANU, LSU, MICH, MO, NLU, NY,
OS, OSC, OSH, PH. TENN, TEX, US, USCH,
VSC, WILLI, WIS). Figure 8.

Xyris louisianica Bridges & Or/ell. Phytologia 64: 56.
1687. TYPE: U.S. A. Louisiana: Calcasieu Parish,
swales and depressions of cutover wetland pine sa¬
vannah on N side of paved road. ca. 1.3 mi. NW of
Ldgerly, elev. 23 it.. 24 Sep. 1987, Orzell & Bridges
5800 (holotype, TEX not seen; isotypes, FSU, GH.
LSU, MISSA, MO, NCU, NLU. NY, SMU. TEX.
TAMU, VDB).

Differt a Xyris strieta var. strieta statura minore, laminis
foliorum brevioribus, angustioribus, margine scabridiori-
Inis. costis scaporum plus scabridis, sepalis lateralibus et
seminis brevioribus.

Slender but stiff, solitary to caespitose perennial
30-80(-90) cm high, with base slightly dilated, en¬
folded in chaffy or fibrous brownish bases of older
leaves. Principal leaves ascending in narrow fans,

( 1 5 — )20— 35( — 40) cm, longer than the scape
sheaths; sheaths entire, V4-V3 as long as blades,
gradually narrowing from broad, brown to red-
brown bases to blade, eligulate; blades narrowly
gladiate-linear or linear, 2— 5(— 7) mm wide, narrow¬
ing above middle to an incurved-acute, slightly cal-
lused tip, the margins a narrow, pale, and scabri-
dulous band, surfaces faintly multinerved, dull

green shading to reddish brown or purplish. Scape
sheaths loosely tubular, with narrow, sharp costae
abaxially, adaxially scarious, pale, the open oblique
orihee narrowed to a short, flattened, blunt cusplike
blade. Scapes slightly compressed distally, 1.5-2(-

2.4) mm wide, often ancipital with two flattened
scabridulous costae making edges, and often with
additional low, papillate costae, medially and prox-
imally subterete or angulate, 1-1.5 mm thick, gen¬
erally multicostate, the costae usually papillate.
Spikes mostly ovoid to ellipsoid to cylindric, l-2(—

2.5) cm long, mostly blunt of many, tightly spirally
imbricate bracts, ihe sterile ones few, keeled, ca.
2^4 mm, grading larger into broadly obovate to sub-
orbicular or reniform, convex fertile ones (4— )4.5-
6(— 7) mm, with entire margins lacerate in age, sur¬
faces castaneous proximally, tan distally, and with
subapical ovate to triangular or reniform green dor¬
sal areas. Lateral sepals subequilateral, oblaneeo-
late, (4 — )4.5 — 6 mm, acute. Petal blades obtriangu-
lar, 3—4 mm, the very shallowly rounded apex
erose. Stamens with anthers broadly oblong, ca. 1
mm, deeply emarginate and sagittate on stubby fil¬
aments ca. 0.5 mm; staminodia birachiate, densely
bearded. Capsule gray, broadly obovoid, ca. 4 mm,
planoconvex; placentation marginal; seeds narrowly
ovoid to ellipsoid-cylindric, 0.5(0.6-0.7)-0.8 mm,
opaque or yellow-farinous.

Moist fine sandy peats or organic sandy clay, At¬
lantic and Gulf Coastal Plains, from south Georgia
and panhandle Florida west across Alabama to
eastern Texas. Flowering from late June to October
or until frost.

During Yym-related fieldwork in the late 1950s
and early 1960s. I encountered what then appeared
to be two forms of X. ambigua Beyrich ex Kunth
growing in mixed populations in the flatwoods and
savannas of the Gulf Coastal Plain in Mississippi,
Louisiana, and eastern Texas. One, a taller plant
with more spreading, linear-gladiate leaves of a yel¬
low-green color and with tan or brown bases, had
broadly obovate petal blades unfolding in early
morning that were nearly 1 cm long. The other form
was a shorter plant, with narrower, more ascending
leaves red or purplish-tinged at base, and its
blooms, opening later in the morning, were broadly
obtriangular and only 4-5 mm long. Yet, in my
treatment (Krai, 1966: 230) I still treated the two
as forms of X. ambigua.

During Heldwork in Alabama in the 1970s, I
again saw the same variation, noting that, in red¬
dish pigmentation, in corolla size and shape, and
particularly in the farinose and narrower seed, af¬
finities of this smaller plant with smaller flowers

Volume 9, Number 2
1999

Krai

Rhynchospora and Xyris

219

actually were more with Xyris stricta , a plant often
in the same general areas hut much more robust
and in much wetter habitats. In the 1980s I began
using the herbarium name “X. obscura” with col¬
leagues and assigned this nomen nudum to material
sent to me for identification. Concomitantly, two
other biologists published this plant as X. louisian-
ica Bridges & Orzell (1987).

Xyris ambigua and the two proposed varieties of
X. stricta are compared in Table 3; the three taxa
are illustrated in Figures 7, 8, and 9. By its mor¬
phology, X. ambigua is clearly distinguish ed. There
is considerable character overlap between the two
X. stricta varieties: these differences are consis¬
tently quantitative but at the varietal level.

It is perhaps unusual for the two varieties in .Yy-
ris stricta to have such different habitats. Xyris
stricta var. obscura always appears upslope or in
drier habitat than X. stricta var. stricta, and thus
invariably shares its habitat with X. ambigua.

Both varieties nest well within the geographic
range of Xyris ambigua, which is found on moist
sandy or peaty sites in flatwoods, seeps, bogs, and
moist open areas of Coastal Plain from Virginia to
south Florida, west into eastern Texas, inland to
Oklahoma, Arkansas, middle Tennessee, and the
Piedmont eastward. It is the only one of the three
to be found outside the United States, occurring in
the Antilles, Mexico, and Central America. Xyris
stricta var. stricta is found in acid swamps, deep
wet bogs, edges of Taxodium ascendens Brongniart—
llex myrtifolia Walter— Nyssa biflora Walter domes,
wet savannas, and flatwoods potholes, mostly on
peat-muck substrates, from Louisiana east to north¬
ern Florida, north in the South Carolina Coastal

Plain. Intermediate examples between X. stricta
var. stricta and variety obscura are to be found in
ecotonal and disturbed sites in southern Georgia,
at or near the eastern boundary of the latter.

Acknowledgments. The writer is indebted many
times over by the kindness of botanists and curators
who in recent years have provided loans of critical
material and/or allowed visits in the following in¬
stitutions: ARN, BRCH, CTB, FSU, GA, GH,
KANU, MICH, MO, MSC, NCU, NLU. OS, SMU-
BRIT, TENN, TEX, US, USCH, VPI, VSC, WILLI,
and WIS. These aids are gratefully acknowledged.

I .iterature Cited

Bridges, E. I.. X S. L. Orzell. 1987. A new species of
Xyris (sect. Xyris) from the Gulf Coastal Plain. Phyto-
logia 64: 56-61.

Chapman, A. W. I860. Flora of the Southern United
States. Cambridge, Massachusetts.

Elliott. S. 1821. A Sketch of the Botany of South Carolina
and Georgia, vol. 1. J. R. Schenek, Charleston, South
Carolina.

Gale, S. 1944. Rhynchospora, section Eurhynchospora, in
Canada, the United States and the West Indies. Rhodora
46: 227-249.

Krai, R. 1966. Xyris (Xyridaceae) of the Continental Unit¬
ed States and Canada. Sida 1 : 1 77—260.

Kiikenthal, G. 1950. Vorarbeiten zu einer Monographic
der Rhynchosporideae. Bot. Jahrb. Syst. 75(1): 119—
121; 75(2): 156-161.

Malme, (7 0. K. 1913. Die Amerikanisehen Spezies der
Gattung Xyris I.. Untergattung Euxyris (Endlicher). Ark.
Bot. 13(8): 1-32.

- . 1937. Xyridaceae in N. American Flora 19(1):

3-15.

Martins. C. F. P. von. 1841. Xyridaceae. Herb. FI. Bras.
34(2): 56, 58.

Small, J. K. 1933. Manual of the Southeastern Flora. New
York .

A New Species of Paederia (Rubiaceae) from China

Li Heng

Kunming Institute of Botany, The Academy of Sciences of China, Heilongtan, Kunming,
650204, Yunnan, People’s Republic of China

ABSTRACT. The new species Paederia spectatissi-
ma H. Li from southeastern Yunnan, China, one of
tew evergreen species in the genus, is here validly
published.

During study ot the Rubiaceae tor the Flora of
Yunnan Province, China, the following new species
was discovered. An English description, distribu¬
tion map, list of specimens, and discussion of this
species were provided hy Puff (1991: 285); the el¬
ements necessary for its valid publication and ad¬
ditional notes on its distribution and recognition are
provided below.

Paederia spectatissima H. Li, sp. nov. TYPE:
China. Yunnan: Xichou Xian, Ting-Mann, alt.
1200-1300 m, in mixed forest, 11 Oct. 1947,
K. M. Feng 12341 (holotype, KUN; isotypes,
KUN).

Haec species a Paederia verticllata Blume foliis oppos-
itis, inflorescentiae axibus dense albo-scabridis atque
fmctibus diametro 10-11 mm distinguitur.

Distribution and habitat. Endemic to south¬
eastern Yunnan (Pingien, Xichou, and Malipo), in
secondary valley torests and shrubby grasslands at
1200-1300 m.

Paederia spectatissima is similar to P. verticillata
Blume; P. spectatissima can be distinguished by its
opposite leaves, axillary inflorescences, and larger
fruits 10—11 mm in diameter.

Paratypes. CHINA. Yunnan: Pingpien Xian, Ma-Wei
village. 1250 m, 17 Oct. 1954, Feng 5070 (KIJN); Malipo
Xian, Moun-Cao-ping, 1250 m, in shrubs, grasslands,
Wenshan Group w660-0245 (KUN 2 sheets).

Acknowledgments. 1 am grateful to C. M. Taylor
and Guanghua Zhu for their help with preparation
of the manuscript.

Literature Cited

Puff, C. 1991. Revision of the genus Paederia L. (Rubi¬
aceae— Paederieae) in Asia. In: C. Puff (editor). The ge¬
nus Paederia L. (Rubiaceae— Paederieae): A multidis¬
ciplinary study. Opera Rot. Belg. 3: 207-289.

Novon 9: 220. 1999.

A New Species of Peritassa Miers (Hippocrateaceae) from
Southern Brazil, and Notes on Two Confused Species

Julio Antonio Lombardi and LCvia Godinho Temponi

Departamento tie Botanica, Instituto de Ciencias Biologicas, Universidade Federal de Minas
Gerais, Av. Antonio Carlos 6627, 31270—110, Belo Horizonte, Minas Gerais, Brazil

ABSTRACT. Peritassa hatschbachii, a new species
of Hippocrateaceae from southern Brazil, is de¬
scribed and illustrated. In addition, comments on
two other contused species, P. calypsoides and P.
mexiae, and a key tor identification of Peritassa
species from southern Brazil are provided. A lec-
totype is proposed for P. calypsoides, and complete
synonymy is presented.

In the initial phase of a taxonomic study of the
neotropical species of Hippocrateaceae, we no¬
ticed a confused interpretation of three southern
Brazilian taxa of Peritassa: Peritassa calypsoides
(Cambessedes) A. C. Smith, P. mexiae A. C. Smith,
and P hatschbachii sp. nov. These three species
have a long history of mistaken interpretation, dat¬
ing back to the original description of Salacia ca¬
lypsoides by Cambessedes (1829), the basionym of
the first of these species. The Museum National
d’Histoire Naturelle in Paris (P) houses two sheets
bearing the name Salacia calypsoides, the first
consisting of two branches, the right branch with
a fixed label numbered 129 and the left branch
without a label, herbarium label reading “Salacia
calypsoides Cambess., Bresil — Province de Rio de
Janeiro. Voyage d'Auguste de Saint-Hilaire, de
1816 a 1821. Catal B- n° 129“ (F Neg. 35941).
The second sheet has one branch and also a fixed
label numbered 129, herbarium label reading “4.
Salacia calypsoides” and another reading “Iso-
type.” The sheet with two branches is a mixed
collection, the left branch corresponding to S. ca¬
lypsoides, and the right to Peritassa mexiae A. C.
Smith; there is also a duplicate fragment in the
Field Herbarium (F 935829) with two leaves, one
from each branch on the original sheet. The sec¬
ond sheet at Paris, also corresponding to S. calyp¬
soides and described by Cambessedes (1829: 107)
under the number 4, is proposed here as the lec-
totype of Salacia calypsoides.

Martius (1837) proposed the name Calypso syl-
vestris, based on a specimen that he had collected
in Rio de Janeiro, that is clearly nonspecific with

Salacia calypsoides, though the leaves are larger
and more elliptic (F Neg. 23267) than the lecto-
type.

Scdacia calypsoides is the basionym of Raddia
calypsoides (Cambessedes) Miers, a name estab¬
lished by Miers (1872) without study of the original
material in Paris. Miers also described Clercia clu-
siifolia Miers, conspecific with S. calypsoides, based
on a plant collected by Gomez in Rio de Janeiro
and currently deposited at K, as well as Clercia
retusa Miers, also conspecific with S. calypsoides,
based on a specimen collected by Sellow (without
locality) and also currently at K.

Miers transferred Calypso sylvestris to his genus
Sarcocampsa, as Sarcocampsa silvestris (Martius)
Miers, but the single specimen he examined, col¬
lected in Rio de Janeiro by his son, is conspecific
with Peritassa mexiae.

The last author who treated these taxa was Smith
(1940), who established the name Peritassa calyp¬
soides (Cambessedes) A. C. Smith without having
seen either the Cambessedes or Martius types aside
from the photographs from the Field Negative Col¬
lection. Smith’s morphological description of this
species is based on the specimens collected by Sel¬
low, and on other specimens collected in the states
of Sao Paulo, Parana, and Rio Grande do Sul. How¬
ever, the taxon thus described as Peritassa calyp¬
soides by Smith is another species: Peritassa
hatschbachii, here described.

Smith transferred Clercia retusa Miers to the ge¬
nus Peritassa, as Peritassa retusa (Miers) A. C.
Smith, here considered to be conspecific with P.
calypsoides. He also described Peritassa mexiae A.
C. Smith, which corresponds to the circumscription
of Sarcocampsa silvestris by Miers (1872).

Peritassa calypsoides, P. hatschbachii, and P.
mexiae are easily distinguished by leaf shape and
consistency in the dry state, and the size of the
inflorescences, even without reference to the char¬
acteristics of the flowers (Figs. 1, 2) and the geo¬
graphic distributions.

Novon 9: 221-226. 1999.

222

Novon

Figure 1. Peritassa hatschbachii Lombardi. — A. Branch aspect. — B. lateral view of flower and buds. — C. Disc,
androecium, and gynoecium, petals removed. — I). Disc and gynoecium, petals and stamens removed. — E. Abaxial
and adaxial view of young stamen before anthesis. — F. Abaxial and adaxial view of stamen at anthesis. Scales: A =
1 cm, B-F = 1 mm. Drawings by J. A. Lombardi, based on Ribas 983.

Peritassa hatschbachii Lombardi, sp. nov. TYPE:
Brazil. Paran4: Sao Jos6 dos Pinhais, Rio Ar-
raial, 26 Jan. 1983, Hatschbach 46068 (holo-
type, MBM; isotypes, BM, CTES, MO, S, U not
seen). Figure 1.

Haec species a Peritassa calypsoide inflorescentiis lon-
gioribus, staminibus non reflexis, dehiscentia quasi ver-
ticali, filamentis longioribus, a P. laevigata antherarum
conneetivis non insigniter productis, et a P. mexiae an¬
therarum dehiscentia e positione differt.

Liana, glabrous, drying dark brown, branches te¬
rete, lenticellate, fissured, smooth, with seasonal
growth marked by scale-like cataphylls. Leaves op¬
posite, stipules 0.4— 1.0 mm long, caducous, trian¬
gular, apiculate, entire or minutely erose; petiole
0.25-1.0 cm, subcylindric or canaliculate, fissured
when dry; blade 2.9-13.6 X 1.1-7. 5 cm, elliptic,
chartaceous. the base cuneate, the apex acute and
sometimes slightly emarginate, the margin entire or
obscurely denticulate and thickened, often cream-

Volume 9, Number 2
1999

Lombardi & Temponi
Peritassa from Southern Brazil

223

colored when dry, venation brochidodromous with
veins prominulous on both sides. Inflorescence
thyrsoid-paniculate, 0.9-6. 3 (-7. 8) X (0.45— )1. 3-
5.2(— 7.2) cm, with 10—180 flowers, axillary, often
at leafless nodes; bracts 0.6— 1.0 mm long, trian¬
gular, the margin entire, scarious; peduncle 0.2-
3.4(— 4.1) cm long, minutely lenticellate-punctulate,
subeylindric to flattened; branches alternate or sub¬
opposite; pedicels 0.4-0. 9 mm long, cylindric, ar¬
ticulate at base; bracteoles 2, at base of pedicel,
triangular, 0.3— 0.5 mm long. Flowers yellowish,
(3.1— )3. 5-^4. 4 mm wide at anthesis; sepals 5, 0.6—
1.5 X 0.8— 1.5 mm, elliptic, unequal, ± carnose,
the margin entire, drying revolute; petals 5, 1.9—
3.1 X 1.3— 2.1 mm, obovate, ± carnose, ± ascend¬
ing at anthesis, the margin minutely erose, scarious,
plicate; disc short-tubular, slightly projecting, ca.
0.4 mm high, carnose, free from reproductive parts,
the margin sinuate; stamens 3, ca. 1 mm long, in
open flower appressed to gynoecium. the filaments
flattened, slightly enlarged at base, the connective
adaxiallv thickened, not surpassing the thecae or
slightly so, drying brown, the anthers oblong, ca.
0.3 X 0.3 mm, slightly emarginate at base, basifix-
ed, composed of two extrorse locules, dehiscing
longitudinally by oblique slits slightly divergent at
base; ovary slightly 3-lobed, 3-locular, with 2 sub-
apical ovules at each loeule, style ca. 0.25 mm
long, central, cylindric; stigmas obscure. Mature
drupe ca. 3.8 X 3.6 cm, subspherical, tuberculate,
drying glaucescent or brown; seeds 2.3-3. 4 X 1.1
X 0.9-1 cm, subreniform, triangular in section, tes¬
ta smooth, firmly embedded in juicy endoearp, dry¬
ing brown.

Vernacular. Bacupari-de-cipo, Fruta-de-maca-
co.

Peritassa hatschbachii occurs at altitudes of 20—
900 m in rainforests of the southern Brazilian At¬
lantic coast, to as far north as southern Rio de Ja¬
neiro in mountain forests in the Serra dos Organs.
This species can easily be distinguished from Per¬
itassa calypsoides by the larger flowers, the stamens
not reflexed with longer filaments, the position of
the dehiscence slits, the larger inflorescences, and
by the consistently elliptic leaves with an acute
apex. It can be distinguished from Peritassa laevi¬
gata, which is similar in leaf form and inflores¬
cence size, by the connective not conspicuously ex¬
ceeding the anther thecae, by the slightlv oblique
dehiscence of the anthers, and by the smaller ma¬
ture fruits.

Paratypes. BRAZIL. Parana: Ypiranga, 22 Mar.
1909, Dusen 8311 (S); Fortaleza. 20 Feb. 1910. Dusen
9561 (F, (41. L, MICH. S): Jaguariahyva, 3 Nov. 1910.

Dusen 10305 ((41, S); Volta Grande, 7 Aug. 191 1, Dusen
12023 (S); Volta Grande, 17 Aug. 1911, Dusen 12056
((41. S): Ypiranga, 28 Aug. 1911, Dusen 12094 (S); Car¬
valho. 13 Sep. 1911, Dusen 13000 (S): without locality,
1903, Dusen s.n. (S); Morretes, 3 Aug. 1975, Dziewa 15
(MBM); Guaratuba. Serra de Arayatuba. 31 Jan. 1960,
Hatschbach 6712 (HB. I„ MBM); Morretes, Grata Funda.
estrada da Graciosa. 12 Jan. 1964. Hatschbach 10836
(BHCB, HB, MBM); Quatra Barras, Ceme, 23 Feb. 1964.
Hatschbach 10990 (MBM); Rio Branco do Sul. Serra do
Cade, 10 Jan. 1978, Hatschbach 40694 (UB); Morretes,
Marurnbi. 4 Jan. 1979, Hatschbach 41837 (F, MBM); Mor¬
retes, Rio Bromado, 12 Feb. 1985, Hatschbach 48899 &
Zelma (MBM. S); Guaraqueyaba, caminho do Paruquara,
23 Jan. 1991. Hatschbach 54897 et al. (CTES, HUFFS,
MBM, W); Morretes, Mandufra, 1 Feb. 1996, J. M. Silva
1633 & Ribas (BHCB, MBM). Rio de Janeiro: Petrop-
olis, Meio da Serra. Serra da Estrela, 9 Mar. 1978. Mar-
tinelh 4034 (RB). Rio Grande do Sul: prope Porto Ale¬
gre. Canoas, 10 Nov. 1892. Malme 290b (S). Santa
Catarina: Elorianopolis, Morro do Ribeirao, 14 Feb.
1967. R. M. Klein 7159 (MBM); Elorianopolis, Bom Re-
tiro. Paulo Lopes, 19 Feb. 1973, R. M. Klein & Rresolin
10846 (P); Eloriandpolis, Morro Costa da Lagoa, 22 June
1967. R. .17. Klein & Souza Sobrinho 7482 (MBM); Itajaf,
Morro da Fazenda, 3 June 1954. Reitz & R. M. Klein 1863
(B, HB, L); Campo Alegre, Morro do Iquererim, 9 Jan.
1958, Reitz & R. M. Klein 6021 (S); Presidente Nereu,
Sabia. 29 Jan. 1958. Reitz & R. 17. Klein 6364 (B. HB.
L); Campo Alegre, Morro do Iquererim, 4 Fell. 1958, Reitz
& R. 17. Klein 6366 (B. (41. L. UC); Rio do Sul. Alto
Matador, 16 Apr. 1958. Reitz & R. 17. Klein 8724 (B. F,
G, L. UC); Campo Alegre, margens do Rio Negro. 17 Jan.
1996, Ribas 983 et al. (BHCB, MBM). S5o Paulo: Sao
Paulo, Parque do Estado, planta viva niimero 9, 20 Jan.
1932, Hoehne s.n. (F 895908, HB 9025); Santos, in ripa
annuis Butuvira, I Dec. 1874. Mosen 3039a (C. S-2), 25
Jan. 1875. Mosen 3039b (S-3); Santos. 1875. Mosen 3039c
(S).

Peritassa calypsoides (Cambessedes) A. C.
Smith, Brittonia 3: 522. 1940. Salacia calyp¬
soides Cambessedes, in A. Saint-Hilaire, A.
Jussieu & Cambessedes, FI. Bras. Merid. 2:
107. 1829. Tontelea calypsoides (Cambessedes)
Endlicher ex Walpers, Repert. Bot. Syst. 1:
400. 1842. Raddia calypsoides (Cambessedes)
Miers, Trans. Linn. Soc. London 28: 396.
1872. TYPE: Brazil. Rio de Janeiro: without
locality, 1816—1821, Saint-Hilaire 129 (lecto-
type, designated here, P [4. Salacia calypso¬
ides]; isolectotype, F 935816).

[Other original material, P [Salacia calypsoides
Cambessedes] not seen (left branch) = F Neg.
35941, F 935829 (elliptic leaf), both specimens
mixed in the sheets with Peritassa mexiae ]

Calypso sylvestris Martius. Flora 20: 96. 1837. Salacia syl-
vestris (Martius) Steudel, Nomencl. Bot., ed. 2, 2:
492. 1841. Sarcocampsa silvestris (Martius) Miers.
Frans. Linn. Soc. London 28: 409. 1872. TYPE: Bra¬
zil. Rio de Janeiro: “in sylvis ad Lagoa dos Indos,”

224

Novon

Martins Herbar. Florae Brasil. 97 (holotype, G = F
Neg. 23267; isotype, W).

Clercia clusiifolia Miers [clusiaefolia |, Trans. Linn. Soc.
London 28: 381. 1872. TYPE: Brazil. Kin de Janei¬
ro: without locality, Gomez 36 (holotype, K not seen;
isotype, BM).

Clercia retusa Miers, I’rans. Linn. Soc. London 28: 380.
1872. Syn. nov. Peritassa retusa (Miers) A. C. Smith,
Brittonia 3: 524. 1940. T\ PE: Brazil. Without lo¬
cality, Sellow s.n. (holotype, K not seen; isotype,
BM). Figure 2 A— I).

Liana, glabrous, drying rusty brown; branches te¬
rete to subquadrangular, lenticellate, striate,
smooth, with short prehensile lateral branches.
Leaves opposite; stipules 0.4— 0.6 mm long, trian¬
gular, entire, caducous; petioles 0.2-0.6 cm, can¬
aliculate; blade 1.9— 7.4(-12.4) X 0.5— 5.4 cm, ob-
ovate or elliptic, chartaceous, the base attenuate or
cuneate, the apex acute or rounded, the margin en¬
tire or obscurely crenulate and thickened; venation
brochidodromous with veins immersed on both
sides or prominulous on abaxial side. Inflorescence
thyrsoid-paniculate, 0.6-2.9 X 0.6-3. 1 cm, with
15-35 flowers, axillary; bracts 0.4— 1.0 mm long,
triangular, the margin entire or minutely erose or
denticulate, scarious, carnose; peduncles 0—15 mm
long, covered by imbricate bracts at least at base,
the peduncle and branches minutely lenticellate-
punctulate, subcylindric; branches opposite or al¬
ternate; pedicels ca. 0.7 mm long, cylindric; brac-
teoles 2, at base of pedicel, triangular, ca. 0.5 mm.
Flowers yellowish, 2. 4-2. 6 mm wide at anthesis;
sepals 5, 0.8-1. 2 X 0.8— 1.5 mm, triangular or ob¬
long, carnose, the margin entire, scarious; petals 5,
1.4-1. 6 X 1.0-1. 3 mm, obovate, ± carnose, de-
flexed at anthesis, the margin entire; disc short-
tubular, 0.3-0. 4 mm high, carnose, free from re¬
productive parts, the margin sinuate or entire;
stamens 3, ca. 0.6 mm long, in open flower exsert-
ed, the filaments flattened, the connective triangu¬
lar, adaxially slightly thickened at base, not sur¬
passing the thecae, drying brown, the anthers
oblong, ca. 0.2 X 0.3 mm, dorsifixed, reflexed, de¬
hiscence extrorse by oblique slits strongly divergent
at base in immature stamen, almost vertical in ma¬
ture stamen; ovary slightly 3-lobed, 3-locular, with
2 subapical ovules in each locule, style ca. 0.15
mm long, central, cylindric; stigma obscure. Mature
drupe ca. 5.3 X 3.3 X 3.3 cm, ellipsoid, drying
dark brown, minutely pitted; seeds ca. 2.1 X 1 X
0.7 cm, ellipsoid.

Peritassa calypsoides occurs in seashore vegeta¬
tion for a distance of ca. 140 km between the cities
of Rio de Janeiro and Cabo Frio, and in the Atlantic
rainforests of southern Bahia. Only the great leaf
size of Mart ius 97 and the poor condition of the

inflorescences made possible the incorrect inter¬
pretation of the sheet’s photograph by Smith (1940),
as this species and P. hatschbachii are very distinct
and easily distinguished by the size of the flowers
and inflorescences, as well as by flower and stamen
characteristics. The two specimens examined from
southern Bahia have leaves matching those of Mar-
tius’s specimen in size and also the leaves of P.
hatschbachii.

Additional specimens seen. BRAZIL. Bahia: U ruyuca,
nova estrada que liga Uruyuca it Serra Grande, a 28-30
km de Uniyuca, 16 July 1978. Mori et at. 10253 (RB);
Una, eslrada Una/Sao Jos£, ea. 17 km ao W do municfpio
sede. 20 July 1994, Jardim el al. 498 (BHCB, CEPEC).
Rio de J aneiro: perto de Mariea, Lagoa do Padre, 8 Sep.
1975, Armijo 764 & Peixoto 570 (BHCB, RB); Maricd,
Lagoa do Padre, 8 Sep. 1950. Bautista 150 (RB-2): Mar¬
iea. Barra de Mariea, 5 Mar. 1985, V. /,. G. Klein 261 et
al. (BHCB. RB); Lagoa de Piratininga, 8 Nov. 1922. Kuhl-
mann s.n. (BHCB 34486, RB 20868); Oct. 1832. Riedel
1084 (P. W); without date. Riedel s. n. (W.); Barra de Mar¬
iea. 25 Sep. 1978. Rizzini A Mattos 1036 (CTES); without
loeality, 1832, Lhotsky s.n. (G); Cabo Frio. 13 Oct. 1968.
Sucre 3950 (BHCB, RB).

Peritassa mexiae A. C. Smith, Brittonia 3: 517.
1940. TYPE: Brazil. Minas Gerais: Viyosa, off
road to Sao Miguel, 720 m, 22 Oct. 1930, Mex-
ia 5206 (holotype, US not seen; isotypes, BM,
F = F Neg. 40852, GH). Figure 2E-H.

Liana, glabrous, drying cinereous; branches sub-
terete, lenticellate, striate, smooth, ± glaucous.
Leaves opposite or subopposite; stipules ca. 0.7 mm
long, caducous, triangular; petioles 0.4— 1.1 cm,
canaliculate; blade 4.2-19.5 X 1.9— 8.8 cm, elliptic
or (rarely) almost orbicular, coriaceous, the base
cuneate or rounded, the apex acute, the margin en¬
tire or obscurely crenulate and thickened, drying
minutely pitted on both sides, opaque, venation
brochidodromous with veins immersed to plane
adaxially, secondary veins prominulous abaxially.
Inflorescence thyrsoid-paniculate, (0.8 — )2.3 — 5.0 X
(0.8— )2. 8-7.1 cm, with 30—200 flowers, axillary, of¬
ten at leafless nodes; bracts 0.5— 1.0 mm long, tri¬
angular, denticulate at base, erose, sometimes dis¬
tant from the ramifications; peduncles 0-6 mm
long, striate, cylindric; branches alternate or subop¬
posite; pedicels 1.9— 2.5 mm long, cylindric, brac-
teoles 2, at base of pedicel, triangular, sometimes
absent, ca. 0.4 mm long. Flowers yellowish, fra¬
grant, 3. 5-4.0 mm wide at anthesis; sepals 5, 0.6—
1.4 X 0. 7-1.1 mm, elliptic, ± carnose, minutely
papillose, cucullate, the margin fimbriolate; petals
5, 1. 7-2.1 X 1.0-1. 2 mm, elliptic, ± carnose, ±
deflexed at anthesis, the margin papillose, plicate;
disc annular, short tubular, ca. 0.4 mm high, car-

Volume 9, Number 2
1999

Lombardi & Temponi
Peritassa from Southern Brazil

225

Figure 2. A— D. Peritassa calypsoides (Cambessfedes) A. C. Smith. — A. Lateral view of flower. — B. Disc, androecium,
and gynoecium, petals removed. — C. Abaxial and adaxial view of young stamen before anthesis. — D. Abaxial and
adaxial view of stamen at anthesis. E— H. Peritassa mexiae A. C. Smith. — E. Lateral view of flower. — E. Disc, an¬
droecium, and gynoecium, petals removed. — G. Abaxial and adaxial view of young stamen before anthesis. — H.
Abaxial and adaxial view of stamen at anthesis. Scales: A-H = f mm. Drawings by J. A. Lombardi: A-D from Bautista
150 ; E— H from Lombardi 1305.

nose, free from reproductive parts, the margin sin¬
uate or entire, projecting, yellow in dry or fresh
state; stamens 3, ca. 0.9 mm long, in open flower
exserted and reflexed, the filaments flattened, the

connective adaxially thickened, not surpassing the
thecae, the anthers oblong, ca. 0.15 X 0.4 mm,
slightly emarginate in base and apex, basifixed, de¬
hiscence extrorse by very strongly almost horizontal

226

Novon

divergent apical slits; ovary slightly 3-lobed, 3-loe-
ular, with 1 subapical ovule in each locule, style
ca. 0.3 mm long, central, cylindric; stigmas ob¬
scure, yellow in dry and fresh state. Mature drupes
not seen, immature 0.9-1. 3 X 0.6-1. 1 X 0. 5-1.0
cm, subspherical, the epicarp apparently thin, dry¬
ing blackish to glaucescent; seeds fusiform, 1-3.

This species is restricted to rainforests of the At¬
lantic coast of Brazil and inland to the State Park
of Rio Doce in Minas Gerais. It is easily distin¬
guished from P. calypsoides and P. hatschbachii by
the leaves drying cinereous, the long-pedicelled
flowers, the spreading petals, and by the noteworthy
almost horizontal dehiscence of the anthers.

Additional specimens seen. BRAZIL. Espfrito Santo:
Linhares, canal da Lagoa Juparana, 13 Aug. 1965, Belem
1584 (UB). Minus Gerais: Caratinga, Fslayao Biologica
de Caratinga, 20 June 1988, Costa et al. 524 (BHCB), 30
Sep. 1995, Gomes 285 (BHCB); Marlieria. Parque Estad-
ual do Rio Doce, 16 July 1993, Lombardi 351 (BHCB,
SPF); Marlieria, Parque Estadual do Rio Doce, margens
da I.agoa da Barra, 17 July 1996, Lombardi 1305 (BHCB,
ESA, MRM, SPF). Rio de Janeiro: Rio de Janeiro, Cor-
covado, 4 Aug. 1959, Duarte 4950 (HB, RR [as 10 Oct.
I942|); without locality, Saint-Hilaire 129a (P not seen =
F Neg. 35941 [right branch], F 935829 [orbicular leaf]);
Ignassu, “Sarcocampsa silvestris,'” Miers s.n. (BM).

Key to thk Species ok Peritassa in Southern Bkazii.

la. Shrubs, in cerrado and rocky grasslands .

. P. campestris (Catnbessfedes) A. C. Smith

lb. Lianas, shrubs, or trees, in forests and seashore
vegetation.

2a. Connective apiculate and exceeding thecae.

3a. Petals 0.8— 1.3 mm long, pedicel 1.0—

1.5 mm long, leaf margin conspicuously
erenulate, inflorescence (0.7— )1. 1—2. 4
cm long, with ca. 25—50 flowers | Minas

Gerais, Rio de Janeiro, Sao Paulo] . . .

. P. flaviflora A. C. Smith

3b. Petals 2.3— 3.3 mm long, pedicel (1.3—)

2. 1-3.7 mm long, leaf margin entire or
erenulate, inflorescence (0.7— )2. 0—5. 3
(-9.5) cm long, with ca. 30— 200 flowers
[Espfrito Santo, Rio de Janeiro] . . .

. P. laevigata (Hoffmannsegg ex Link)

A. C. Smith

2b. Connective not or only slightly exceeding
thecae.

4a. Filaments conspicuously exceeding bor¬
der of disc.

5a. Anther locules dehiscing by slight¬
ly divergent slits; ovary locules
consistently with two ovules . .

. P. hatschbachii

5h. Anther locules dehiscing by very
strongly divergent slits, almost api¬
cal; ovary locules consistently with

one ovule . P. mexiae

4b. Filaments not exceeding the border of

disc . P calypsoides

Acknowledgments. The authors thank Alain
Chautems for photocopies of literature, and Douglas
A. Yanega for proofreading the English; this con¬
tribution is part of a research project financed by
FAPEMIG (grant number CBS-2080/96), PRPq-
UFMG (grant number 23072.928822/97-95), and
CNPq (research fellowship grant 350221/97-9).

Literature Cited

Cambessedes, J. 1829. XXVII. Hippocrateaceae. In A. de
Saint-Hilaire, A. H. L. de Jussieu & J. Cambessedes,
Flora Brasiliae meridionalis 2: 102—1 12. A. Belin Bib-
liopolam, Paris.

Martins, C. F. P. 1837. Herbarium Florae brasiliensis. Flo¬
ra 20(2): 1-128.

Miers, J. 1872. On the Hippocrateaceae of South America.

Trans. Linn. Soc. London 28; 3 1 9—432.

Smith, A. C. 1940. The American species of Hippocra¬
teaceae. Brittonia 3: 341-555.

Una Nueva Especie de Acaena (Rosaceae) en Sudamerica Austral

Alicia Marticorena

Departamento de Botanica, Universidad de Concepcion, Casilla 160-C, Concepcion, Chile.

amartic@udec.el

Resu MEN. Se describe una nueva especie de Acae¬
na, A. patagonica, de la zona patagonica de Chile y
Argentina. Se publica ademas una nueva seccion den-
tro del genera la que incluye a esta especie.

Abstract. A new species of Acaena, A. patagon¬
ica, from Patagonian Chile and Argentina, is de¬
scribed. A new section of the genus is erected to
include this species.

El genero Acaena Mutis ex L. posee alrededor
de 100 especies (Mahberley, 1987) con una mayor
distribucion en el hemisferio sur, presentando solo
pocas especies en el hemisferio norte, por ejemplo
Mexico, California, Hawaii (Hutchinson, 1967). En
Chile el genero posee 20 especies (A. Marticorena,
in prep.) y en Argentina unas 21 especies (Gron-
dona, 1964, 1984), la mayorfa de la zona austral.

El genero se divide en secciones basado en el
tipo de inflorescencia y caracterfsticas de la cupela.
La cupela corresponde al conjunto del talamo
acrescente y el aquenio protegido por el talamo. Sin
embargo, la nueva especie no presenta ninguna de
las combinaciones de caracteres de las secciones
existentes que permita incluirla en ellas, luego se
propone una nueva seccion.

Acaena secc. Patagonicae A. Marticorena, sect,
nov. TIPO: Acaena patagonica A. Marticorena.

Inflorescentia terminalis spiciformis; fruetus cupula
ovata. glabra vel pubescente, 3— 4-costata, apice costa acu-
leo solitario disposita. sine glochidibus; folia sine stipulis.

Inflorescencia espiciforme corta, alargada en estado
fructfero, con glomerulos dispersos a lo largo del pe-
dunculo. Cupela ovoide, glabra a pilosa, con 3—4 es-
pinas apicales, sin gloquidios. Hojas sin estfpulas.

Acaena patagonica A. Marticorena, sp. nov.
TIPO: Chile. Ultima Esperanza: Parque Na-
cional Torres del Paine, Cerro Diente, 900 m,
15.XII.1985, Arroyo & Squeo 850829 (holoti-
po, CONC). Figura 1.

Planta perennis, caulibus brevissimis. Folia ambitu li-
nearia. 1.3— 2.5 cm longa; foliolis 5— 8-jugis, 3.0— 6.5 mm
longis, obovatis vel obovato-orbicularibus, base cuneifor-
mibus, 3— 5-partitis, laciniis linearibus, margine revolutis.

superne glabrescentibus, inferne sericeis; vagina margine
trichomatibus unicellularibus longis vestita; sine stipulis.
Inflorescentia terminalis spiciformis pauciflora diametro
ca. 6.5 mm: ramo florifero terminali in pedunculum usque
ad 16 cm longum desinente; bracteis basalibus linearibus
vel triangularibus, 1.0— 1.5 mm longis, margine partitis.
Flos sepalis 4—5, lanceolatis, ca. 1.5 mm longis, supra
glabris, infra sericeis, apice trichomatibus floccosis orna-
tis; staminibus 4, antheris orbicularibus, ca. 0.3 mm lon¬
gis; stigmate rotundato, laciniato, ca. 0.5 mm diametro.
Fruetus cupula ovata vel rotundata, 1.5— 2.5 mm longa,
pubescente vel glabra, 3— 1-costata, quaque costa aculeo
solitario 0.5— 2.0 mm longo, trichomatibus retrorsis tecto
apice ornato, raro aculeo brevi rudimentario sub aculeo
apicali posito; achenio solitario.

Planta de hasta 20 cm de alto, con entrenudos
cortos. Hojas en contomo lineares, lamina 1.3— 2.5
cm de largo; folfolos 5—8 pares, 3.0— 6.5 mm de lar¬
go, obovados a obovado-orbiculares, 3— 5-partidos,
los segmentos lineares a linear-lanceolados, el mar-
gen revoluto, el haz glabrescente, con tricomas blan-
cos esparcidos, el enves serfeeo, la base cuneada;
vainas foliares con el margen con pelos largos,
apendices estipulares ausentes. Rama florifera ter¬
minal; pedunculos hasta 16 cm de largo, pubescen-
tes a serfeeos; inflorescencia espiciforme corta, ca.
6.5 mm de diametro, con pequenos glomerulos de
floras y/o frutos a lo largo del pedunculo; bracteas
basales lineares a triangulares, 1.0— 1.5 mm de largo,
el margen partido. Flor chasmogama con 4—5 sepa-
los lanceolados, ca. 1.5 mm de largo, internamente
glabros, extemamente serfeeos, tricomas en el apice;
estambres 4, las anteras orbiculares, ca. 0.3 mm de
largo; estigma orbicular, laciniado, ca. 0.5 mm de
diametro. Cupela ovada a casi redonda, 1.5-2. 5 mm
de diametro, pubescente a glabra, con 3—4 costillas,
cada costilla con una espina de 0.5— 2.0 mm de lar¬
go, con tricomas retrorsos desde el apice, dispuestas
en el tercio superior de la cupela, a veces con otra
espina rudimentaria bajo la principal; aquenio 1.

Distribucion y habitat. Esta especie habita en la
zona de las Torres del Paine (Ultima Esperanza) y
en el Estrecho de Magallanes, Chile, y al sur de la
provincia de Santa Cruz y Tierra del Fuego, Argen¬
tina. Crece entre rocas, en sitios secos y tambien en
vegas de altura, desde el nivel del mar hasta los
1030 m.

Novon 9: 227-229. 1999.

228

Novon

Figure 1. Acaena palagonica A. Marticorena. — A. Planta. — B. Hoja v detalle de los folfolos. — C. Cupela. (A. B
basado en Arroyo & Squeo 870189', C basado en Elvebakk 532.)

Esta especie es afin a A. pinnatifida Ruiz & Pa-
von, A. platyacantha Speg. y A. poeppigiana Gay.
De A. pinnatifida se diferencia por ser mbs pequena
y frbgil, con folfolos menos divididos y las cupelas
con 3-4 espinas apicales; de A. platyacantha por
los folfolos mbs linos, menos pilosos, cupela sin
costillas notorias, espinas no aladas y superficie de
la cupela lisa, la que generalmente es rugosa en
aquella especie; y de A. poeppigiana por presentar
folfolos lineares y cupela sin espinas pequenas en-
tre las costillas o bajo la espina apical. Otro car-
dcter que reafirma el reconoeimiento de esta es¬
pecie y de la seccibn, es el patron que presenta la

epidermis de la cupela. Originalmente el genero se
divide en secciones, las que estan delimitadas por
el tipo de inflorescencia y caracterfsticas de la cu¬
pela (Bitter, 1911). Segun el tipo de inflorescencia
de Acaena patagonica, esta deberfa incluirse en la
seccibn Acaena. Sin embargo, las caracterfsticas de
la cupela son diferentes al resto de las de la seccibn
y se asemejan a las de la seccibn Acrobyssinoideae,
por el numero de espinas y su posicibn. Las es-
pecies afines antes mencionadas pertenecen a la
seccibn Acaena , basada en caracteres macroscbpi-
cos, las que ademas comparten caracteres micro-
scbpicos que tienen relacibn con el tipo de cblulas

Volume 9, Number 2
1999

Marticorena
Acaena patagonica

229

Figura 2. Celulas epidermicas de la cupela. — A. Tipo celular de la seccidn Acaena. — B. Celulas de A. patagonica.
Escala: 100 pun.

de la epidermis de la cupela, los que discriminan
claramente y agrupan a las demas especies de la
seccion que crecen en Chile (A. Marticorena, in
prep.). El tipo celular de las especies que perte-
necen a la seccion Acaena corresponde a celulas
mas o menos isodiametricas, con paredes anticlin-
ales rectas y relieve celular levantado (Fig. 2A), en
tanto que las celulas de A. patagonica no presentan
limites definidos, las paredes anticlinales son ir-
regulares y sinuosas y el relieve celular es acana-
lado (Fig. 2B). Los caracteres de la cupela son los
que entregan mayor informacion y por lo tanto son
de gran utilidad para la determinacion de las es¬
pecies y delimitacion de las secciones.

Clave para A. patagonica y Espkciks Relacionadas

la. Plantas mayores de 50 cm de alto; folfolos 5-7-
partidos; cupela con espinas de tamano similar

. A. pinnatifida

lb. Plantas menores de 30 cm de alto; folfolos 2-5-
partidos; cupelas con espinas mayores en el ter-
cio superior.

2a. Segmentos de los folfolos ovados; cupela con
espinas pequenas entre las costillas o pe¬
quenas es[)inas triangulares bajo la espina

principal . A. poeppigiana

2b. Segmentos de los folfolos lineares; cupela
sin espinas pequenas entre las costillas.

3a. Folfolos pubescentes a serfceos; super-
ficie rugosa entre las costillas de la cu¬
pela; espinas triangulares, aplastadas

lateralmente . A. platyacantha

3b. Folfolos glabros a glabrescentes; super-
ficie lisa entre las costillas de la cupela;
espinas aciculares . A. patagonica

Paratipos. ARGENTINA. Santa Cruz: Giier Aike,
Estancia Las Viscachas, cono Pan de Aziicar, 50°45'S,

71°58'0, 1030 m, 28 ene. 1977. T.B.P.A. 2563 (HIP);
Estancia La Verdadera Argentina, S Cerro Leon, 50°53'S,
72°13'0, 16 ene. 1977, T.B.P.A. 2052 (HIP); Estancia
Cabo Buen Tiempo, 51°35'S, 69°07'0, 4 die. 1975,
T.B.P.A. 50 (HIP). Tierra del Kuego: Tennessse oil
camp, 11 die. 1965, Goodall 207 (LP). CHILE. Prov.
Ultima Esperanza: Estancia Dos de Enero, Valle de las
Chinas, 8 die. 1978, Pisano & Cardenas 4853 (HIP); Si¬
erra Baguales, cerro Santa Lucfa, 50°44'S, 72°20'0, 800
m, 2 feb. 1987, Arroyo & Squeo 870189 (CONC); Cerro
Donoso, Rfo de las Chinas, 50°44'S, 72°31'0, 70 m, 9
feb. 1987, Arroyo, Veloso & Penaloza 870265 (CONC);
Cerro Corona, La Victorina, Lago Paine, 50°48'S,
72°47'0, 500 m, 21 feb. 1982, Pisano 5631 (HIP): 4-5
km SE Estancia Guido, 50°55'S, 72°20'(), 150 m, 9 die.
1995, Elvebakk 532 (CONC). Prov. Magallanes: Punta
Delgada. Buque Quemado, 52°28'S, 69°32'0, 5 m, 3 die.
1978, Dollenz 495, 497 (HIP).

Agradecimientos. Deseo agradecer a Nelson
Moya por las delicadas ilustraciones, a los herba-
rios del Museo de La Plata (LP) y del Instituto de
La Patagonia (HIP) por el prestamo de importante
material para este estudio, al Proyecto Flora de
Chile a traves del cual se ha realizado este trahajo
y a Roy Gereau (MO) por la correcion de la diag¬
nosis en latfn y la lectura crftica del manuscrito.

Literatura Citada

Bitter, G. 1911. Die Gattung Acaena. Vorstudien zu einer
Monographic. Biblioth. Bot. 17(4): i— ii. 1—336,37 lam.
Grondona. E. 1964. Las especies argentinas del genero
Acaena (Rosaceae). Darwiniana 13: 209—342.

- . 1984. Rosaceae. En M. N. Correa. Flora Pata¬
gonica, parte 4— b: 48—88. Instituto Nacional de Tec-
nologfa Agropecuaria, Buenos Aires.

Hutchinson, J. 1964—1967. Hie Genera of Flowering Plants
(Angiospermae), Vols. 1, 2. Clarendon Press, Oxford.
Mabberley, I). J. 1987. The Plant-Book. A Portable Dic¬
tionary of the Higher Plants. Cambridge Univ. Press,
Cambridge.

New Boraginaceae from Tropical America 1: New Species of
Bourreria and Tournefortia from Costa Rica and a Note on the
Publication of Cordia collococca

James S. Miller

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. Bourreria rinconensis and Tournefortia
isabellina are described as new from Costa Rica.
Keys are provided to the Costa Rican species of
Bourreria and the Central American species of
Tournefortia sect. Cyphocema.

In preparing a treatment of Boraginaceae for .4
Manual to the Plants of Costa Rica , two collections
representing previously undescrihed taxa were dis¬
covered.

Bourreria rinconensis J. S. Miller, sp. nov.
TYPE: Costa Rica. Puntarenas: Canton de
Osa, Aguahuena, margen izquierda de Que-
brada El Campo, Rinctin, 8°42'45"N,
83°31'35"W, elev. 200 m, 22 Sep. 1990, Ger¬
ardo Herrera 4350 (holotype, MO 04660635).
Figure 1.

Arbor usque ad 8 m alta. Folia persistentia; lamina gla¬
bra, elliptica, 7.7-14 cm longa, 2.7-6 cm lata, apiee ac¬
uminata. basi acuta ad attcnuata; petiolo 1. 3-2.3 cm lon-
go. glabro. Inflorescentia terminalis, cymosa; pedunculo
12—18 mm longo, glabro. Flos gemmis 8—9 mm longis,
apiculatis; calyce 6.5-8 mm longo, extus glabro, intus stri-
goso; corolla alba, hypocrateriformi, ca. 16 mm longa, 5-
loba. lobulis late ovatis, 3—3.5 mm longis, ca. 3 mm latis;
staminibus 5, filamentis ca. 14 mm longis, ad insertionem
puberulis. Fructus non visi.

Tree 8 m tall, the hark corky, the twigs glabrous.
Leaves persistent; leal blades elliptic, 7.7—14 cm
long, 2.7-6 cm wide, the apex acuminate, the base
acute to attenuate, the margin entire, the adaxial
surface glabrous, lustrous, the abaxial surface gla¬
brous, the venation brochidodromous, the midrib
prominent, strongly impressed above, raised below,
the secondary veins arching, 6-8; petioles 1.3-2. 3
cm long, broadly canaliculate on the adaxial sur¬
face, glabrous. Inflorescences terminal, cymose, the
peduncles 12-18 mm long, glabrous. Flowers bi¬
sexual, the buds ellipsoid, 8-9 mm long, apiculate;
calyx leathery, tubular, 6.5-8 mm long, ca. 3 mm
wide at the mouth, glabrous, 5-lobed, the lobes
sometimes not separating completely and the calyx
appearing 2-4-lobed, triangular, 3-5 mm long, 1.5-
2 mm wide at the base, the interior surface of the

lobes densely strigose. Corolla white, salverform,
brown tomentulose, the tube ca. 16 mm long, 2.5—
3 mm wide at the mouth, the 5 lobes spreading,
widely ovate, 3—3.5 mm long, ca. 3 mm wide, the
apex obtuse to rounded; stamens 5, the filaments
ca. 14 mm long, the upper 8 mm free, puberulent
at and just beneath the point of insertion; ovary
ovoid, ca. 1.5 mm tall, 1 mm broad, glabrous, the
style ca. 9 mm long, the 2 stigmas capitate. Fruit
unknown.

Bourreria rinconensis is known only from the type
collected in lowland wet forest on the Osa Penin¬
sula.

Bourreria is a complex genus with questionable
generic boundaries (Thulin, 1987; Miller, 1989),
and there is no comprehensive treatment of the
species available. The approximately 50 species
have been considered by most authors to be re¬
stricted to the New World, occurring in the Carib¬
bean and adjacent south Florida, Mexico and Cen¬
tral America, and northern South America. Thulin
(1987) extended both the morphological and geo¬
graphic boundaries of the genus by including five
species of Ehretia L. from eastern Africa. The
Central American species seem to form two nat¬
ural groups, one with large corollas and leathery
tubular calyces (e.g., B. grandicalyx J. S. Miller
& Sirot and related species) and the other with
smaller corollas and small, membranaceous, cam-
panulate calyces (e.g., B. andrieuxii (A. DC.)
Hemsley, B. mollis Standley, and B. oxyphylla
Standley). Bourreria rinconensis is a member of
the group of species with large corollas and leath¬
ery tubular calyces that was treated recently by
Miller and Sirot (1997). Within this group, B. rin¬
conensis shares a narrow salverform corolla wi th a
narrow, nearly parallel-sided tube with B. quirosii
Standley and B. cumanensis (Loefling) 0. E.
Schulz. Bourreria cumanensis occurs along the
north coast of South America and differs in having
shorter, broader leaves and a much smaller corol¬
la. Bourreria quirosii occurs in dry forests in
Guanacaste, Costa Rica, and Nicaragua and dif-

Novon 9: 230-235. 1999.

Volume 9, Number 2
1999

Miller

Tropical American Boraginaceae

231

Figure 1. Bourreria rinconensis J. S. Miller. — A. Flowering branch. — B. Inflorescence with buds and open flower.
— C. Open calyx showing the strigose interior margin of the lobes. — 1). Open corolla showing staminal attachment.
All from Herrera 4350 (MO).

232

No von

fers in having pubescent, shorter, broader leaves.

Bourreria cumanensis and B. quirosii also are
characterized by a fruit that breaks into 4 pyrenes,
each attached to the gynobase with a slender fiber,
a group that has been segregated as the genus Cre-
matomia Miers (Miers, 1869). The fruit of B. rin-
conensis is unknown, but it would help indicate

Kky to the Costa Rican Species ok Bovrkf.ru

la. Calyx rotate to campanulate, membranaceous.

2a. Corolla less than 1.5 cm long; leaves scabrous above . Bourreria litoralis Standley

2b. Corolla greater than 2 cm long; leaves glabrous above . B. huanita (Llave & Lexarza) Hemsley

lb. Calyx tubular, leathery.

3a. Petioles greater than 10 mm long; leaf blade greater than 6 cm long, glabrous above; plants usually trees.

4a. Corolla less than 2 cm long; calyx less than 8 mm long . B. rinconensis J. S. Miller

4b. Corolla greater than 2 cm long; calyx greater than 13 mm long.

5a. Flowering calyx greater than 2.5 cm long; edges of the lobes pubescent; fruits completely enclosed

by the accrescent calyx . B. grandicalyx J. S. Miller & Si rot

5b. flowering calyx less than 1.8 cm long; edges of the lobes glabrous; fruits not enclosed by the

calyx . B. coslaricensis (Standley) A. Gentry

3b. Petioles less than 10 mm long; leaf blade less than 7.5 cm long, scabrous above; plants shrubs .

. B. quirosii Standley

whether this new species is more closely allied to
B. cumanensis and B. quirosii, species from dry
lorests with similar tubular corollas, or to B. cos-
taricensis (Standley) A. Gentry and B. grandicalyx
J. S. Miller & Sirot, which have large indehiscent
lruits. The following key will help distinguish the
species that are known from Costa Rica.

Tournofortitt isabellina J. S. Miller, sp. nov.
TYPE: Costa Rica. Puntarenas: Parque Inter-
nacional La Amistad San Vito coto Brus, Finca
Cafrosa, 8°54'15"N, 82°46'50”W, elev. 500 m,
4 July 1990, Roberto Delgado 47 (holotype,
MO 05022359). Figure 2.

Liana, ramis velutinis isabellinis. Folia persistentia;
lamina ovata. 8-12 cm longa, 4.9— 7.2 cm lata, apice ac¬
uminata, basi rotundata, supra sparse serieea, infra velu-
tina isabellina; petiolo 7—16 mm longo, supra canalicu-
lato, velutino isabellino. Inflorescentia terminalis, cymosa,
ca. 1 1 cm longa, 15 cm lata, ramulis velutinis isabellinis;
pedicellis in fructu 1-3 mtn longis. Flores non visi. Fruc-
tus 4-lobus, 2.5—3 mm longus, 3-4 mm latus, glaber.

Liana, the stems evenly dull golden brown ve-
lutinous, the hairs 1-1.5 mm long. Leaves alter¬
nate, persistent; leaf blade ovate, 8-12 cm long,
4.9— 7.2 cm wide, the apex acuminate, the base
rounded, the margin entire, the adaxial surface
sparsely sericeous, more densely so along the mid¬
rib, the hairs appressed, dull golden brown, ca. 1
mm long, the abaxial surface evenly velutinous,
more densely so on the midrib and secondary veins,
the hairs erect, dull golden brown, 1-1.5 mm long,
the venation brochidodromous, the midrib promi¬
nent, impressed above, rounded and raised below,
the secondary veins arching, 6—8; petioles 7-16
mm long, dull golden brown velutinous. Flowers
unknown. Infructescence terminal, a much-
branched cyme, ca. 1 1 cm long, 15 cm broad, the
branches dull golden brown velutinous. Fruits yel¬
low with green markings, strongly 4-lobed, 2.5—3
mm tall, 3—4 mm broad, glabrous, borne 4—9 mm

apart, on pedicels 1—3 mm long, the pedicels and
persistent calyx with sparse, spreading, dull golden
brown hairs, the 4 nutlets black, ovoid, ca. 1.5 mm
long, colliculate.

Tournefortia isabellina is a distinctive species,
and the epithet refers to the distinctive color of the
hairs that cover the plant (Stearn, 1973). It is a
member of section Cyphocema 1. M. Johnston, an
entirely Neotropical section that is characterized by
a strictly vining habit anti unusual fruits with four
distinct nutlets that are surrounded and held to¬
gether by a membranaceous epicarp (Johnston,
1930). The individual nutlets are angular on their
ventral surface with the curved embryo bent around
an intrusion of the ventral wall. The species of sec¬
tion Cyphocema typically have elongate corolla
lobes and anthers that are connate apically (Miller,
1988). Johnston (1930) pointed out that while the
section Cyphocema is well marked within Tourne¬
fortia, the species are notoriously difficult to sep¬
arate. However, the majority of confusion lies with
the South American species, and only two species
have been recognized in Central America by recent
authors (Nowicke, 1969; Gibson, 1970; Miller,
1988). Johnston’s (1935) reference to Tournefortia
paniculata Chamisso in Costa Rica probably refers
to plants better placed in T. maculata Jacquin. The
present species is quite distinctive in its terminal,
dense, cymose inflorescence and distinctive indu-
ment. It differs from the other Central American
species of the section in its much larger leaves and
golden brown indument. The Central American

3cm

Volume 9, Number 2
1999

Miller

Tropical American Boraginaceae

233

Figure 2. Tournefortia isabellina J. S. Miller. — A. Branch with cymose infructescence. -
Infructescence branch. — D. Petiole and leaf insertion on branch. All from Delgado 47 (MO).

B. Mature fruit. — C.

234

Novon

species of Tournefortia sect. Cyphocema can be
separated as follows.

Kky to thk Central American Species of Tournefor¬
tia sect. Cyphocema

la. Plants evenly covered with a golden brown pu¬
bescence 1 mm or more long; leaves greater than

4.9 cm wide . T. isabellina J. S. Miller

lb. Plants glabrous to tomentose, the hairs white to
yellowish, never golden brown; leaves less than
5 cm wide.

2a. Corolla tube 3.3—5 mm long; leaves glabrous

to evenly short-strigillose below .

. T. maculata Jacquin

2b. Corolla tube 2—2.3 mm long; leaves densely
white puberulent to tomentose below ....

. T. volubilis L.

The Original Publication of Cordia
collococca L.

Cordia collococca L. is widespread in tropical
America, and the name has been considered to
have been published in the second edition of Spe¬
cies Plantarum (Linnaeus, 1762) by most authors
treating Central American species (e.g., Johnston,
1940; Gibson, 1970; Nash & Moreno, 1971; Miller,
1988). Johnston (1949) discovered that the name
originated in a thesis of Carolus Gust. Sandmark,
but attributed it to a later reprint (Linnaeus, 1760)
rather than to its earliest publication. In 1759, Lin¬
naeus published two treatments of Jamaican spe¬
cies based on specimens given to him by Patrick
Browne and material collected by Hans Sloane cur¬
rently deposited at BM. The first of these works,
published 28 November 1759, Plantarum Jamai-
censis Pugillus (Linnaeus, 1759a) mentioned only
Cordia bourreria I ,. However, later in the same year,
on 22 December, he published Flora Jamaicensis
(Linnaeus, 1959b), the thesis of Carolus Gust.
Sandmark, which was a list of binomial names for
the plants collected by Browne and Sloane. Here
Linnaeus published the name Cordia callococca by
listing it with reference to page 166 of Browne’s
Civil and Natural History of Jamaica (Browne,
1756). The page reference provided by Linnaeus is
incorrect, but on page 167, Browne did provide a
description of COLLOCOCCUS I. Folii rugosis ven-
osis oblong-ovatis, floribus laxae racemosis, the
clammy cherry or turkey-berry tree. As Browne’s
COLLOCOCCUS 2. Platyphyllus major, racemis
umbellatis on page 168 clearly seems to be the ba¬
sis lor Cordia macrophylla L., the species on page
167 is the only logical source of the name Cordia
callococca. The name appears with the same or¬
thography in a reprint of the Flora Jamaicensis ar¬
ticle in volume 5 of Amoenitates Academicae (Lin¬
naeus, 1760), a series of reprints of the Linnaean

dissertations, but is finally corrected to the original
spelling of Browne as Cordia collococca in edition
2 of Species Plantarum (Linnaeus, 1762).

While the publication of Cordia collococca is
confusing, the 1759 Flora Jamaicensis does appear
to be valid publication of the name conforming to
relevant articles of the International Code of Bo¬
tanical Nomenclature (Greuter et al., 1994). Al¬
though Linnaeus merely listed the name, he did
provide a direct reference to a previously published
description (article 32.4). Although the reference to
Browne’s earlier description is to an incorrect page
number, article 33.3 states that “errors of biblio¬
graphic citation. . .do not invalidate publication of
a new combination or avowed substitute.” Also the
change in orthography in the second edition of .Spe¬
cies Plantarum in 1762 to conform with Brown's
spelling of Collococcus seems to be an acceptable
correction of an orthographic error as allowed by
article 60.1. While some authors have questioned
authorship of names in Linnaean theses, Rickett
(1955) and Steam (1957) have argued that names
published in these works should be attributed to
Linnaeus and not his students. Krok (1925) did not
include Flora Jamaicensis in his list of Linnaean
theses written by the students themselves, so it also
seems that the name should be attributed to Lin¬
naeus and not Sandmark. Thus the correct citation
of original publication is:

Cordia oolloeoeca L., FI. Jamaic. 14. [22 Dec.]
1757. TYPE: Jamaica. P. Browne s.n. (lecto-
type, designated by Miller (1988), LINN, Sav¬
age Catalog number 253.8).

Acknowledgments. I thank John Myers for the
illustrations, Roy Gereau for help with the Latin
descriptions, Amy Pool for drawing my attention to
problems with the name Cordia collococca, Mary
Stiffler for bibliographic assistance, and Lennart
Andersson for helpful review comments.

Literature Cited

Browne, P. 1756. The Civil and Natural History of Jamai¬
ca. London.

Gibson, K. N. 1970. Boraginaceae. In: Flora of Guatemala.

Fieldiana, Bot. 24(9): 111-167.

Greuter, W., F. B. Barrie. H. M Burdet, W. (7 Chaloner,
V. Demoulin, I). L. Hawksworth, P. M. j0rgensen, D. H.
Nicholson, P. (7 Silva, P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature. Regnum
Veg. 131.

Johnston, 1. M. 1930. Studies in the Boraginaceae 8. Ob¬
servations on the species of Cordia and Tournefortia
known from Brazil, Paraguay, Uruguay, and Argentina.
Contr. Gray Herb. 92: 3—89.

- . 1935. Studies in the Boraginaceae 10. The Bor-

Volume 9, Number 2
1999

Miller

Tropical American Boraginaceae

235

aginaceae of northeastern South America. J. Arnold Ar¬
bor. 16: I— 64.

- . 1940. Studies in the Boraginaceae 15. Notes on

some Mexican and Central American species of Cordia.
J. Arnold Arbor. 21: 336-355.

- . 1949. Studies in the Boraginaceae 18. Boragi¬
naceae of the southern West Indies. J. Arnold Arbor.
20: 111-138.

Krok. Th. 0. B. N. 1925. Bibliotheca Botanica Suecana.
Stockholm.

Linnaeus, C. 1759a. Plantarum Jamaicensium pugillus.
Uppsala.

- . 1759b. Flora Jamaicensis. Uppsala.

- . 1760. Flora Jamaicensis. Amoen. Acad. 5: 371 —

388.

- . 1762. Species Plantarum. Edition 2. Holmiae,

Stockholm.

Miers, J. 1869. On the Ehretiaceae. Crematomia. Ann.
Mag. Nat. Hist. 3: 300—313.

Miller, J. S. 1988. A revised treatment of Boraginaceae
for Panama. Ann. Missouri Bot. Card. 75: 456—521.

- . 1989. A revision of the New World species of

Ehretia (Boraginaceae). Ann. Missouri Bot. Card. 76:
1050-1076.

- & B. Sirot. 1997. A new species of Bourreria (Bor¬
aginaceae) from Costa Rica. Novon 7: 395—397.

Nash. I). E. & N. P. Moreno. 1971. Boraginaceae. In: Flora
de Veracruz, Fasciculo 18: 1—149.

Nowicke, J. W. 16. Boraginaceae. In: Flora ol Panama.

Ann. Missouri Bot. Card. 56: 33-69.

Rickett. H. W. 1955. Notes on the Linnaean dissertations.
Lloyd ia 18: 49-60.

Stearn, W. T. 1957. Introduction to Species Plantarum.
The Ray Society, London.

- . 1973. Botanical Latin. 2nd Ed. David & Charles,

Newton Abbot, U.k.

Thulin, M. 1987. Bourreria (Boraginaceae) in tropical Af¬
rica. Nordic J. Bot. 7: 413-417.

New Species of Stemmadenia and Tabernaemontana (Apocynaceae)
from Costa Rica, Panama, and Colombia

J. Francisco Morales

Instituto Nacional de Biodiversidad (INBio), Apto. 22—3100, Santo Domingo de Heredia,

Costa Rica

ABSTRACT. Stemmadenia abbreviata and S. si-
mulans are described from Costa Rica and Panama,
and Tabernaemontana leeuwenbergiana is de¬
scribed from Colombia.

Stemmadenia Bentham is a small genus of 10
species, confined primarily to Central America, al¬
though a few species occur in northern South
America (Leeuwenberg, 1994). The genus is closely
allied to Tabernaemontana L., but it is readily dis¬
tinguished by its corolla tube with five staminal
ridges below the anthers within. The presence of
one corona, composed of five epistaminal ridges,
within the corolla tube in some species of Stem¬
madenia and Tabernaemontana is an unusual char¬
acter within the tribe Tabernaemontaneae. In Stem¬
madenia the corona is found in most species (e.g.,

S. alfari (Donnell Smith) Woodson, S. donnell-smi-
thii (Rose) Woodson, S. litoralis (Kunth) L. Allorge),
while in Tabernaemontana it is rarely present (e.g.,

T. columbiensis (L. Allorge) Leeuwenberg).

Study of material included among specimens re¬
ceived as gifts for identification revealed an unde¬
scribed species of Tabernaemontana from Colom¬
bia. In addition, while writing an account of the
genus Stemmadenia for the Manual to the Plants
of Costa Rica, two new species were discovered;
they are described below.

Stemmadenia abbreviata J. F. Morales, sp. nov.
TYPE: Costa Rica. Heredia: in forest along
Rfo Peje, about 1 km S of La Selva (new prop¬
erty) S boundary, 13 Apr. 1982 (fl), Hammel
11680 (holotype, INB; isotype, DUKE).

Arbuscula vel frutex 2 m alta, rain is furcatis, teretibus.
folia 5—14.5 X 1.5—6 cm, elliptica, glabra, membranaoea,
apice acuminata. Calycis segmenta 11-18 X 6-8 mm,
tubo corollae 21-23 mm. salverformi, lobis 6-7 X 4 mm,
antheris 3-4 mm. Fructus ignotus.

Shrub or small tree 2 m high; branchlets terete
to subterete, forked, glabrous. Leaves petiolate; pet¬
iole glabrous, 3-9 mm long; blade <5—14.5 X 1.5—
6 cm, membranaceous, elliptic, acuminate at the

Novon 9: 236-239. 1999.

apex, cuneate at the base, entire, glabrous on both
sides, with 8-10 pairs of upeurved secondary veins.
Inflorescence pedunculate, 3-7-Howered, somewhat
agglomerate, glabrous; peduncle 3-5 mm long, with
several bracts; pedicels 5—9 mm; bracts scale-like,
2-3 mm; sepals white outside, green inside, folia-
ceous, erect, ovate to narrowly ovate, 1 1—18 X 6—
8 mm, acute, subequal, glabrous, bearing 6—7 col-
leters inside, entire; corolla salverform, the tube
white, lobes pale yellow, with 5 narrow staminal
ridges 0.5 mm wide, corona composed of five ep¬
istaminal ridges ca. 0.5 mm wide above the inser¬
tion of the stamens; tube 21-23 mm long, almost
cylindrical, ca. 3 mm wide above the base, twisted
around the anthers; lobes obliquely obovate, 5—7 X
4 mm, obtuse, somewhat spreading to reflexed; sta¬
mens with apex 8-9 mm below mouth of corolla
tube, inserted 7-8 mm from the base; anthers nar¬
rowly oblong, 3^4 X 1 mm, apex acuminate, sag¬
ittate at the base; ovary ovoid, 2-2.5 mm, glabrous,
with a disk-like, ring-shaped thickening less than
1 mm high, adnate to base; style 5-6 mm; pistil
head 1 mm. Fruits unknown.

Known only from the type collection, at 100-300
m. It flowers from March to May.

In Costa Rica, the type collection has been re¬
ferred to Stemmadenia robinsonii Woodson, but S.
abbreviata is easily recognized by its narrow sal¬
verform corolla, cylindrical and straight tube, with
the lobes 5-7 mm long. The specific epithet refers
to the short and inconspicuous corolla lobes.

Stemmadenia simulans J. F. Morales & Q.
Jimenez, sp. nov. TYPE: Costa Rica. San Jos6:
Cant6n de Puriscal, Zona Protectora La Can-
greja, Faja Costena del Valle de Parrita, Mas-
tatal de Puriscal, Rfo Negro, 290-320 m, 3
Mar. 1994 (fl), J. F. Morales, D. A. Smith & G.
Smith 2414 (holotype, INB; isotypes, B, BR,
C, CAS, CR, F, G, GH, INB, K, MEXU, MICH,
MO, NY, P, UPS, US, USF, W, WAG, WIS, Z).
Figure 1.

Volume 9, Number 2
1999

Morales

Stemmadenia and Tabernaemontana

237

Figure 1. A— E. Stemmadenia simulans J. F. Morales & Q. Jimenez {Morales et al. 2414, INB). — -A. Habit. — B.
Calyx and corolla. — C. Opened corolla tube. — I). Pistil with sepals. — E. Fruit.

238

Novon

Arbuscula vel frutex 1—1.5 m alta, ramis furcatis, ter-
etibus. Folia 8—37 X (3.2-)5— 14 cm, elliptica vel obovata,
glabra, rnembranacea, apice acuminata. Calyeis segmenta
4—6 X 4—4.5 mm. tubo corollae 26—28 mm. salverformi
vel anguste infundibuliformi, lobis 9—11 X 5—6 mm, an-
theris 5-4) mm. Fructus 22-32 X 9—12 mm.

Shrub or small tree 1-1.5 m high; branchlets te¬
rete to subterete, forked, glabrous. Leaves petiolate;
petiole glabrous, 5-22 mm long; blade 8-37 X
(3.2-)5— 14 cm, membranaceous, elliptic to obovate;
acuminate at the apex, cuneate at the base; entire,
glabrous on both sides, with 11-12 pairs of upcur-
ved secondary veins. Inflorescence inconspicuously
pedunculate, 17-26-flowered, agglomerate; pedun¬
cle 1— 2(— 4) mm long, with several bracts; pedicels

4— 10 mm; bracts scale-like, 1-2 mm, persistent;
sepals green-yellowish, erect, unequal, ovate, 4—6
X 4-4.5 mm, obtuse, glabrous, bearing 5—7 colle-
ters inside, entire; corolla yellow, salverform to very
narrowly infundibuliform, with 5 narrow staminal
ridges ca. 1 mm wide, corona composed of five ep-
istaminal ridges above the insertion of the stamens;
tube 26-28 mm long, almost cylindrical, ca. 3 mm
wide above the base, twisted around the anthers;
lobes obliquely obovate, 9-1 1 X 5—6 mm, obtuse,
spreading to somewhat reflexed; stamens with apex
13-15 mm below mouth of corolla tube, inserted at
9-11 mm from the base; anthers narrowly oblong,

5— 6 X 1 mm, apex acuminate, sagittate at the base;
ovary ovoid, ca. 2 mm, glabrous, with a disk-like,
ring-shaped thickening 1 mm high, adnate to the
base; style 6-7.5 mm; pistil head 1 mm. Fruit of 2
separate mericarps; mericarps green-yellow,
obliquely ellipsoid or pod-like, 22—32 X 9—12 mm,
recurved, apiculate, sometimes caudate, with lat¬
eral ridges, smooth; aril orange. Seeds 7—8 X 3
mm, with longitudinal grooves.

This species is known from wet forests in Costa
Rica and northwestern Panama, from 290 to 1000(—
1400) m. It flowers from March to May, and fruits
from July to September. Somewhat related to Stem-
madenia pauli Leeuwenberg, it is easily recognized
by its agglomerate 17-26-flowered inflorescences,
conspicuous and persistent bracts, and persistent
and shorter sepals.

The specific name recalls that Stemmadenia si-
mulans looks like a species of Tabernaemontana.
In fact, two collections of this species ( Busey 547
and Gdmez 19612) with immature flowers were pre¬
viously identified as Tabernaemontana columbien¬
sis. However, the deep yellow corolla and the five
narrow staminal ridges below the anthers within the
corolla tube, confirm that this species should be
included within Stemmadenia.

Stemmadenia simulans is morphologically very

similar to Tabernaemontana columbiensis ; the latter
is easily separated, however, by its floral characters
(larger sepals and inflorescences, persistent and
conspicuous floral bracts), and smaller fruits (22-
32 mm vs. 35-55 mm in T. columbiensis) and seeds
(7-8 mm vs. 9-12 mm).

Paratypes. COSTA KICA. Punt arenas: hi la de Cal,
between Las Cruces and Nelly, Gdmez 19612 (MO). San
Jose: Zona Protectora La Cangreja, Santa Rosa de Pur-
iscal. Morales 278 (INB); Mastatal de Puriscal. Rfo Negro.
Morales .3128 (INB. MO); Zona Protectora I a Cangreja.
Mastatal de Puriscal, Ramirez et al. 44 (CR, INB. MO);
Mastatal de Puriscal, Zamora et al. 1271 (INB. MO).
PANAMA. Chiriquf: Burica Peninsula, San Bartolo I. fin¬
ite, 20 km W of Puerto Armuelles. Busey 547 (MO).

Tabernaemontana leeuwenbergiana J. F. Mo¬
rales, sp. nov. TYPE: Colombia. Narino: trail
from La Planada to Pielapi, wet lower montane
cloud forest, 1600-1800 m, 22 July 1988 (fl).
Gentry et al. 63607 (holotype, INB; isotype,
MO).

Arbuscula vel frutex, 3—4 m alta, ramis furcatis, tere-
tibus. Folia 41 — 4-5 X 26-29 cm, elliptica, glabra, mem-
branacea, apice acuminata. Calyeis segmenta 3—4 X 2-
3.5 mm, tubo corollae 21-24 mm, salverformi, lobis 8-12
X 5-6 mm, antheris 5—6 mm. Fructus ignotus.

Shrub or small tree, 3^1 m high; branchlets te¬
rete to subterete, glabrous, forked. Leaves petiolate;
petiole glabrous, 26—30 mm long; blade membra¬
naceous, 41-45 X 26—29 cm, broadly elliptic;
acuminate at the apex, cuneate at the base; entire,
glabrous on both sides, with 19-21 pairs of upeur-
ved secondary veins; tertiary venation reticulate.
Inflorescence pedunculate, 18-23-flowered, cor¬
ymbose, lax; peduncle 24-26 mm long, with several
bracts; pedicels ( 1 0 — ) 1 5 — 20 mm; bracts few, scale¬
like, 2—3 mm; sepals erect, ovate, 3—4 X 2-3.5
mm, obtuse to rounded, glabrous, bearing many
colleters inside, entire; corolla salverform, white,
glabrous; tube 21-24 mm long, almost cylindrical,
ca. 3 mm wide above the base, somewhat twisted
around the anthers; lobes obliquely obovate, 8—12
X 5—6 mm, obtuse, somewhat spreading; stamens
with apex 5—6 mm below mouth of corolla tube,
inserted at 12-13 mm from the base; anthers nar¬
rowly triangular, 5-6 X 1 mm, apex acuminate,
sagittate at the base; ovary ovoid, 2—2.5 mm, gla¬
brous, with a disk-like, ring-shaped thickening 1
mm high, adnate to base; style 5-6 mm; pistil head
1 mm. Fruits unknown.

This species is known only from the type locality,
at elevations of 1600—1800 m. It resembles Taber¬
naemontana maxima Markgraf, but differs by its
conspicuously petiolate leaves (26—30 mm vs. 1-4

Volume 9, Number 2
1999

Morales

Stemmadenia and Tabernaemontana

239

mm), shorter sepals (3-4 mm vs. 6—12 mm), and
corolla lobes (8—12 mm vs. 4—5 mm). From T.
markgrafiana J. F. Macbride it differs by its broadly
elliptic leaves (vs. elliptic to narrowly elliptic in T.
markgrafiana ), lax inflorescence (vs. dense and ag¬
glomerate), and longer sepals (3^1 mm vs. 2-3
mm).

Tabernaemontana leeuwenbergii is named after
Anthony (Toon) Leeuwenberg (WAG), in recognition
of his extensive taxonomic work in the tribe Taber-
naemontaneae, especially in the genus Tabernae¬

montana , and for his encouragement of my studies
on Apocynaceae.

Acknowledgments. 1 thank the directors and cu¬
rators of COAH, COL, CR, DUKE, F, HB, HUA,
INPA, K, MO, NY, RB, UB, UPS, and WAG for the
loan of material. The line drawings of Stemmadenia
simulans were prepared by Silvia Troyo.

Literature Cited

Leeuwenberg, A. J. M. 1994. Pp. 213^150 in A Revision
of Tabernaemontana 2. The New World Species and
Stemmadenia. Royal Botanic Gardens, Kew.

Miscellaneous Notes in Temnadenia and Laubertia (Apocynaceae)

J. Francisco Morales

Instituto Nacional de Biodiversidad (INBio), Apto 22-3100, Santo Domingo de Heredia,

Costa Rica

ABSTRACT. Tilt* new combination Temnadenia
odorifera (Vellozo) J. F. Morales is proposed here.
Laubertia gentlei Lundell is reduced to synonymy
of L. peninsularis Woodson.

While preparing the treatment of Apocynaceae
for Flora Mesoamericana and other projects from
South America, I determined the need for the fol¬
lowing nomenclatural changes.

Notes in Temxadenia

The name Temnadenia stellaris (Lindley) Miers
has been widely used in publications and identifi¬
cations since Woodson’s (1935a) monograph. How¬
ever, while preparing the reevaluation of the genus
Echites and Prestoma, 1 concluded that the original
illustration of Echites odorifera Vellozo included in
Florae Fluminensis (Vellozo, 1829), “referable to
some species of Temnadenia''’ fide Woodson
(1935a), clearly represents a typical specimen of T.
stellaris. Therefore, a new combination is necessary
and is here proposed.

Temnadenia odorifera (Vellozo) J. F. Morales,
comb. nov. Basionym: Echites odorifera Vello¬
zo, FI. Flumin. 104. 1829 [“1825”]. TYPE: t.
28 in Vellozo, FI. Flumin. leones 3, 1831
[“1827”].

Echites stellaris Lindley, Edwards’s Bot. Keg. 20: pi. 1664.
1835. Syn. nov. Temnadenia stellaris (Lindley) Miers.
Apocyn. S. Amer. 210. 1878. TYPE: “introduced
from Rio Janeiro to the Horticultural Society by the
Hon. Robert Gordon.”

For the complete synonymy see Woodson
(1935a).

Notes in Laubertia

Laubertia gentlei Lundell was distinguished from
L. peninsularis Woodson based on Hower size and
leaf indument (Lundell, 1976). However, careful
examination of the holotypes has shown that L. gen¬
tlei is the same species as L. peninsularis and is
therefore reduced to synonymy.

For the complete synonymy anil a key to the cur¬
rently accepted species of Laubertia, see Woodson
(1935b, 1938).

Laubertia peninsularis Woodson, Ann. Missouri
Bot. Card. 23: 374. 1936. TYPE: British Hon¬
duras [Belize]: Undesignated locality near Be-
lize-Guatemala boundary, Schipp s.n. (holo-
type, MO).

Laubertia gentlei Lundell, Wrightia 5: 256. 1976. Syn.
nov. TYPE: British Honduras [ Belize]: Toledo Dis¬
trict: Manga Camp, Edwards Koad beyond Columbia,
12 Apr. 1948, Gentle 6505 (holotype, TEX : isotypes,
F, MO; photograph, INB ex F).

Literature Cited

Lundell, C. L. 1976. Studies in American Plants XII.
Wrightia 5: 256.

Vellozo, J. 1829. florae Fluminensis. Rio de Janeiro.
Woodson. R. E. 1935a. Studies in the Apocynaceae. IV.
The American genera of Echitoideae XXL Temnadenia.
Ann. Missouri Bot. Card. 23: 253-260.

- . 1935b. Studies in the Apocynaceae. IV. The

American genera of Echitoideae XXVIII: Liubertia.
Ann. Missouri Bot. Card. 23: 253—260.

- . 1938. (Asclepiadales) Apocynaceae. In J. 11.

Barnhart (editor). North American Flora 29(2): 103-
192. New York.

Novon 9: 240. 1999.

A New Species of Votomita (Melastomataceae) from Venezuela,
with Thoughts on Ovule and Seed Number and Seed Size

Thomas Morley

Department of Plant Biology, University of Minnesota, St. Paul, Minnesota 55 108, U.S.A.

ABSTRACT. A new species of Votomita from south¬
ern Venezuela is described, illustrated, and distin¬
guished from its relatives on the basis of fruiting
material; certain useful floral characters can also
he determined from the specimen. The new species
illustrates a general condition in subfamily Meme-
cyleae, i.e., a relatively large number of ovules is
produced yet only one or a few large seeds are
formed. Ovule and seed numbers are discussed,
seed volumes are calculated for Mouriri and Votom¬
ita , and possible explanations for the differences
are explored.

Votomita ventuarensis Morley, sp. nov. TYPE:
Venezuela. Terr. Fed. Amazonas: Dept. Ata-
bapo, bosques medios y bajos inundables en
el rio Ventuari-Macabana, 4°15'N, 66°20'W,
140 msnm, Sep. 1989, Luz Delgado 594 (ho-
lotype, MO; isotype, PORT). Figure 1.

Arbor usque 15 m alia; pagina inferior costae mediae
anguste 2-alata ad angulos, minute puberula; cryptae sto-
matatae unaquaeque cavitatibus 1—5; epidermes folii sine
pigmento; hypodermis absens; stamina monadelpha; ovar¬
ium 4-loculare ovulis 22-25 axillaribus; semen unicum,
globosum.

Tree to 15 m high, glabrous except for the leaf
midrib; young twigs rounded. Petioles 4.5—6 mm
long; blades 7.9—12.4 cm long, 4.1—6 cm wide,
ovate-elliptic to elliptic, acute at base, abruptly
acuminate at the apex with an acuminum 0.7-1 cm
long; midrib low-rounded to plane or slightly
grooved adaxially when dry, prominent abaxially,
flat and winged on the edges, the undermidrih mi¬
nutely puberulent with hairs 2CM10 p.m long; lat¬
eral nerves when dry faintly visible or invisible
adaxially, faintly visible abaxially. Midrib xylem tu¬
bular; stomatal crypts 30—36 per mm2 near the mar¬
gin to 54 near the midrib, with (1— )2— 5 cavities
each, 50—60 p.m high; adaxial epidermis of uniform
thickness, 29-31 gm thick including the cuticle,
mostly one cell thick, occasionally two, mucilage
walls none, the cytoplasm unpigmented and clear
till stained, the inner walls straight and parallel
with the outer ones; cuticle thin; abaxial epidermis
also unpigmented; hypodermis none; foliar sclere-

ids all terminal on the veinlets, irregularly stellate,
often with an irregular horizontal central body 1^4
times as long as wide. Peduncles 1 per side at leaf¬
less nodes of twigs 3^4 mm thick below the leaf
zone, 10.5—14 mm long with 2 internodes, the lower
0.5—1 mm long, the upper 10—13 mm, 1-flowered;
bracts deciduous before fruit formation; pedicels
12-13 mm long in fruit; ovary loeules 4, placen-
tation axile, ovules 4—11 per placenta, 22—25 in all;
fruits yellow to orange, ellipsoid, crowned with the
calyx, 17-18 mm long including calyx by 12-14
mm diam. when dry, 20-21 mm long by 13—16 mm
when boiled, the fruiting calyx 6.4— 6.7 mm diam.,
3—3.3 mm deep, the lobes low-triangular, 0.8— 1.2
mm high, 4—5.2 mm wide; seed 1, spheroid, 9.7—
9.9 mm high, 8.8— 9.6 mm thick, with an irregularly
elliptic raised area (function unknown) ca. 4 mm
from the broken chalazal strand, the raised area 6
mm long, 4 mm wide, 0.5—1 mm high, with edges
that overhang 0—0.5 mm; the raised area presum¬
ably includes the rnieropyle. Petal scars on the fruit
rounded-triangular, 2.3-2. 7 mm wide, 0.9— 1.2 mm
long; stamen scars 8, the filament scars broad, thin,
and united, forming a continuous ring 3.2— 3.7 mm
in outside diam., 0.2-0. 5 mm thick, the stamens
thus monadelphous; thecae adaxial on the filaments
and placed low so that their bases leave imprints
0.9-1.2 mm wide, 0.8— 1.0 mm thick radially
around the style base, the total stamen thickness at
base 1.0— 1.3 mm.

Distribution. Known only from the type locality
east of San Fernando de Atabapo, northwest of cen¬
tral Amazonas, Venezuela.

Local name. Cometure tierra firmero.

Although the type specimen bears only ripe
fruits, the filament scars and anther imprints on the
fruits yield useful floral information, and with care¬
ful dissection of the 1-seeded fruit (when boiled)
all or most of the undeveloped ovules can be found
and the nature of the placentation determined, in
spite of the compaction and distortion that have
taken place.

The new species is unique in the genus in its
unpigmented epidermises. Further distinctions of
this plant from the other species follow, starting

Novon 9: 241-244. 1999.

242

Novon

Figure 1. Mouriri ventuarensis Morley. — A. Leaves. — B. Cleared portion of leaf Blade showing veins and terminal
sclereids. — C. Cross section of leaf blade showing upper epidermis, a sclereid. and stomata! crypts. — I). Cross section
of leal midrib. — L. fruit. — F. Apex of fruit showing the calyx lobes and the scars and imprints of the other floral
appendages: P, petal scar; S, scar of monadelphous stamens; T, imprints of anther thecae. — G. Seed: A, broken chalazal
strand shows halfway up; B, shape of elevated outgrowth is shown; the chalazal strand is at the top.

with those most easily and clearly distinguished; V.
pubescens Morley differs in the pubescent under¬
side of its lamina; V. guianensis Aublet, V. mon-
antha (Urban) Morley, and V. orinocensis Morley
have separate filaments and ordy 10 or fewer ovules
per ovary; V. monadelpha (Ducke) Morley has 48
ovules per ovary and two very different forms of
foliar sclereid; V. plerocarpa (Morley) Morley has 45
ovules per ovary, lateral nerves of the leaf moder¬
ately prominent, stomatal crypts 32—38 gm high,
an irregular upper epidermis in which the lower
walls of the cells are mostly strongly rounded and
variable in depth, and mucilage walls occasional in
the epidermis; V. orbinaxia Morley has a glabrous
unwinged midrib when viewed abaxially, separate

stamens, 36 ovules, and columnar sclereids. Votom-
ita cupuliformis Morley & Almeda differs in its
abaxially glabrous midrib, 16 ovules, simple sto¬
matal crypts, columnar foliar sclereids, and epider¬
mal cells with numerous mucilage walls. Votomita
roraimensis Morley departs in having petioles 1.5—
2.5 mm long, a seed lacking an elevated elliptic
disc, 15—16 ovules, and a double epidermis with a
deep inner layer and occasional mucilage walls.

Of the other species, the one most similar to Vo¬
tomita ventuarensis is V. plerocarpa, which agrees
in its abaxially puberulent midrib, the form of its
foliar sclereids and stomatal crypts, its monadel¬
phous stamens with low-placed thecae, and its axile
placentation. The features differentiating V. plero-

Volume 9, Number 2
1999

Morley

Votomita ventuarensis from Venezuela

243

carpa from V. ventuarensis are stated above. When
flowering plants of the new species are found, and
fruiting material of V. plerocarpa is collected, the
differences between the two doubtless will become
more apparent. As it is, the numerous anatomical
features that occur in the genus greatly facilitate
the distinctions between these two species as well
as between all the species of the genus. Long ex¬
perience has shown these features to be as reliable
as morphological ones.

Although only two fruits of the new species could
be dissected owing to scarcity of material, both
were one-seeded even though there were 20-25
ovules present; one-seeded fruits are also found in
the other species of the genus for which fruits are
available: V. guianensis (9—10 ovules), V. monantha
(5—8 ovules), V. pubescens (20—26 ovules), and V.
roraimensis (15—16 ovules). All seeds are relatively
large.

A similar pattern of ovule to seed number and
large seed size occurs in the closely related genus
Mouriri and in its Old World relatives. However, in
Mouriri many species have more than one seed.
Mouriri trunciflora Ducke has 40—80 ovules but
only 2-12 seeds; M. dimorphandra Morley, 42-71
ovules and 2—5 seeds; M. brachyanthera Ducke,
54—57 and 2—1; M. jicoides Morley, 50—60 and 2-
5; M. subumbellata Triana, 15—42 and 1—8; M. cras-
sifolia Sagot, 30—38 and 1—6; M. lunatanthera Mor¬
ley, 28—43 and 6; M. cauliflora Martius ex DC., 23-
40 and 5 (Morley, 1976). All the rest with known
flowers have 4—30 ovules and those with known
fruits have 1—5 seeds except for M. oligantha Pil-
ger, which has 9-12 ovules and 1-6 seeds; the
ovules are always 3—20 times as many as the seeds,
with the possible exception of M. oligantha (Mor¬
ley, 1976). It is believed that the wide range in
ovule numbers can be read as a form-series from
many to few.

The Old World genera Lijndenia, Memecylon,
Spathandra, and Warneckea are similar to Votomita
in ovule and seed numbers. The flowers produce
2-19 ovules (Bremer, 1981, 1982, 1983; Jaeques-
Felix, 1978, 1984, 1985) but only l(-2) seed(s).
The consistently single seed of Votomita and the
one or occasionally two seeds of the Old World gen¬
era agree with the interpretation of these on other
grounds as generally more specialized than Mouriri.

It would appear that the consistent and appar¬
ently inefficient pattern in the Memecyleae in
which one to many ovules do not function must
result from a hormonal cut-off from the undevel¬
oped ovules. The small number of ripened seeds,
which are relatively large, must be somehow ad¬
vantageous in relation to the final fruit size, the

nature of the fruit-eaters, and/or the conditions en¬
countered by the dormant or germinating seed. The
reduction in number of seeds per fruit must reflect
some major shift in the ecology of the group.

In all probability the Memecyleae (see Renner,
1993) are evolving from a group in which the fruits
have or had many small seeds with a matching
ecology, with conditions apparently wasteful of
ovules initially being produced as the ecological
adaptation changed and evolution proceeded in the
direction of few large seeds. Progressively fewer
ovules are produced in the more advanced mem¬
bers: as few as five in one species of Votomita, four
in a Mouriri, and two in some Lijndenias.

It was thought desirable to find a practical way
to describe seed size since sizes vary greatly in the
subfamily, particularly in Mouriri. Volume was
judged to be the critical parameter. To determine
the approximate volume the formula for the volume
of an ellipsoid was used: Y^tt X ( V2 length X Vz
width X Vi thickness of the seed). Even with some
divergence from a true ellipsoid this formula should
give a useful approximation of size.

In order to make a limited check on the accuracy
of the above method, displacement tests were made
wilh the relatively few seeds of Mouriri and Votom¬
ita available at the MIN herbarium. Seeds of 15
species were measured by placing them in 95%
EtOH in standard graduated cylinders of three sizes
and measuring the fluid rise. The accuracy of this
method is limited by the difficulty of getting an
exact reading of fluid level, especially when the
seed is a poor fit in the tube. In 11 of the 15 seeds
the figure calculated from measurements was be¬
tween 1 and 10% of the displacement figure, from
low to high. The exceptions were 11%, 14%, and
21% high, the 21% being a seed covered with ir¬
regular tubercules (the seeds were calipered). I
conclude that at least in the Memecyleae the mea¬
surements calculated by formula are accurate with¬
in 10% of the true volume in most cases and rarely
exceed a differential of 20%.

With one exception the largest seeds in Mouriri
occur in the two most primitive sections, Taphrox-
ylon and Abundiflos, where approximate seed vol¬
umes range from (34 — )777 to 3041 mm3, The ex¬
ception is M. megasperma Morley, which has a
volume of 3806 mm3, the largest known seed in the
genus; known only in fruit, the relations of this spe¬
cies are uncertain except that it is a specialized
plant and is not in either of the aforementioned
sections. Seed volumes in Mouriri other than in the
two most primitive sections or M. megasperma vary
from 39 to 1341 mm3. Only two species, M. arborea
Gardner and M. crassisepala Morley, have seeds ex-

244

Novon

feeding 780 nun1. The smallest seeds are those of
M. viridicosta Morley, 34 mm*, an anomalous mem¬
ber of the section Taphroxylon ; next smallest are
seeds ol M. helleri Britton, 39 mm3. Seed volumes
in the five species of Votomita known in fruit range
from 56 to 493 mm*. The four Old World genera
are estimated from illustrations or published di¬
mensions to have seed volumes of about 20—680
min'.

Since the largest seeds of Mnuriri occur mostly
in the most primitive sections it is possible that
large seed size played a part in the origin of the
subfamily. It has been suggested that larger seeds
favor survivability of the seeds or seedlings (Rich¬
ards, 1996; Howe, 1986), and it is stated that the
large-seeded local species in a study area at Los
Tuxtlas, Mexico, have higher establishment rates in
the closed canopy forests there than smaller-seeded
plants (Martinez-Ramos & Soto-Castro, 1993).
Howe (1993) made the general observation that
trees with specialized dispersal systems often have
large seeds. However, as van der Pijl pointed out
(1969), . .small and large seeds have many back¬

grounds!” This last point is illustrated in Mouriri,
where most of the species with large seeds occur
in the two most primitive groups, but one species,
M. megasperma, appears to have evolved large
seeds secondarily in an advanced group.

The general trend in the Memecyleae toward re¬
duction in number of seeds suggests either that a
one-seeded fruit makes the most desirable package
for convenient frugivor dispersal, or that the plant
resources would not permit ripening many large
seeds in the same fruit, or perhaps a combination
of the two. Unknown factors are probably involved.

The existence in the Memecyleae of plants with

many ovules, presumably a primitive condition,
suggests that any search for its ancestral connec¬
tions should be among groups that also have nu¬
merous ovules.

Literature Cited

Bremer, K. 1981. Seeds and embryos in Sri Lanka (Cey¬
lonese) species of Memecylon, with notes on Spathandra
(Melastomatacae). Nordic J. Bot. 1: 62-65.

- . 1982. Lijndenia, a re-established paleotropical

genus of the Melastomataceae— Memecyleae. Nordic J.
Bot. 2: 121-124.

- . 1983. Taxonomy of Memecylon (Melastomatacee)

in Borneo. Opera Bot. 69: 5—47.

Howe, H. F. 1986. Seed dispersal by fruit-eating birds and
mammals. Pp. 123—189 in I). R. Murray (editor), Seed
Dispersal. Academic Press, London.

- . 1993. P. 4 in T. 11. Fleming & A. Estrada (edi¬
tors), Frugivory and Seed Dispersal: Ecological and
Evolutionary Aspects. Kluwer Academic Publishers,
Dordrecht.

Jacques-F6lix, H. 1978. Les genres de Memecyleae (Me¬
lastomataceae) en Afrique, Madagascar et Mascareig-
nes. Adansonia 18: 221—235.

- . 1984. Les Memecyleae (Melastomataceae) de

Madagascar (part 1). Bull. Mus. Natl. Hist. Nat.. 4 s£r„

6, 1984, sect. B, Adansonia, no. 4: 383—45 1 .

- . 1985. Les Memecyleae (Melastomataceae) de

Madagascar (part 2). Bull. Mus. Natl. Hist. Nat., 4 s6r„

7, 1985, sect. B, Adansonia, no. I: 3—58.
Martinez-Ramos, M. & A. Soto-Castro. 1993. Seed rain

and advanced regeneration in a tropical rain forest. Pp.
299-318 in: T. II. Fleming <!4 A. Estrada (editors), Fru¬
givory and Seed Dispersal: Ecological and Evolutionary
Aspects. Kluwer Academic Publishers, Dordrecht.
Morley, T. 1976. Memecyleae (Melastomataceae). Flora
Neotropica Monograph 15: 1—295.

Renner, S. S. 1993. Phylogeny and classification of the
Melastomataceae and Memecylaceae. Nordic J. Bot. 13:
519-540.

Richards, P. W. 1996. The Tropical Rain Forest, ed. 2.

Cambridge Univ. Press, Cambridge.

Van der Pijl, L. 1969. Principles of Dispersal in Higher
Plants. Springer- Verlag, Berlin.

Senecio aetfatensis (Asteraceae: Senecioneae), a New Species from

Zimbabwe

Bertil Nordenstam

Department of Phanerogamic Botany, Swedish Museum of Natural History, SE-104 05 Stock¬
holm, Sweden, hertil.nordenstam@nrm.se

Abstract. A new species, Senecio aetfatensis, is
described from the Chimanimani Mountains of
eastern Zimbabwe.

During the AETFAT post-congress excursion in
the mountains of eastern Zimbabwe in February
1997, I was fortunate to find and collect an unde¬
scribed species of Senecio. It is a pleasure to name
the new species for the Association pour l’fitude
Taxonomique de la Flore d’Afrique Tropicale (AET¬
FAT), which held its XVth Congress in Harare on
3-7 February 1997.

Senecio aetfatensis B. Nordenstam, sp. nov.
TYPE: Zimbabwe. Chimanimani National
Park, along footpath from entrance office to
Mountain Hut, ca. 1600 m, 11 Feb. 1997, Nor¬
denstam 9292 (holotype, S; isotypes, K, MO,
SRGH). Figure 1.

Herba perennis ramosissima 0.6— 1.5 m alta et lata.
Caules et rami brevi-villosi. Folia alterna sessilia herba-
cea 3— 6(— 8) cm loiiga 1.5— 3.5 cm lata plana pinnatisecta
sparse hirsutula, basin versus angustata, basi semiam-
plexieaulia auriculata; lobis utrinque 3—5 oblongis pin-
natilobatis vel dentatis acutis. Capitula 3—10 corymbosa
radiata pedunculis gracilibus 1-5 cm longis. Involuerum
campanulatum; involucri bracteae 13—21 subuniseriatae
lineari-lanceolatae 4—5.5 mm longae 0.5— 1.5 mm latae
v hides glabrae nervis 1—2 resiniferis marginibus anguste
membranaceis apicibus acutis— aeuminatis puberulis. Ca-
lyculi bracteae 5—8 anguste triangulares acutae. Recep-
taculum planum nudum alveolatum. Flosculi radii pler-
umque 8 flavi. Flosculi disci numerosi, corolla flava
tubulosa superne dilatata quinquelobata. Antherae basi
breviter sagittatae, collum filamenti basi dilatatum. Sty 1 i
rami lineares. areis stigmatieibus scparatis. apicibus trun-
catis pilis eferrentibus brevibus. Cypsela anguste ellipti-
co-oblonga costata breviter villosa. Pappi setae numerosae
graciles minute barbellatae caducae.

Much-branched herb forming a bushy rounded
clump 0.6— 1.5 m high and wide. Stems and branch¬
es somewhat brittle, shortly villous. Leaves alter¬
nate, sessile, herbaceous, soft, doubly pinnatisect,
3— 6(— 8) cm long, 1.5— 3.5 cm wide, flat, shortly hir¬
sute, green above, paler below, with 4—5 lobes on
each side, narrowed toward the base and petiolar-
iform; leaf lobes 3—5 on each side, oblong, pin¬

nately lobed or dentate, acute; leaf base half-clasp¬
ing and auriculate with dissected or dentate small
ears. Capitula 3—10 in a terminal, laxly branched
corymbiform synflorescence; peduncles 1—5 cm
long, slender. Involucre campanulate; involucral
bracts 13—21, subuniseriate, linear-lanceolate, 4—

5.5 mm long, 0.5— 1.5 mm wide, green with narrow
membranous margins, 1-2-veined with resiniferous
veins, acute to acuminate, tips puberulous, other¬
wise glabrous. Calycular bracts 5—8, 1—2 mm long,
narrowly triangular, flat, acute. Receptacle flat, na¬
ked, alveolate. Ray florets (5— )8; corolla yellow;
tube 2—3 mm long, somewhat glandular-puberu-
lous; lamina elliptic-oblong, 4.5—8 mm long, 2.5—

3.5 mm wide, 4-veined, patent or rolled back. Disc
florets ca. 40—60; corolla yellow, tubular, widening
above to a narrowly campanulate limb 2—3 mm
long, 5-lobed, 3.5— 4.5 mm long; lobes deltoid or
triangular-ovate, 0.5— 0.7 mm long, minutely papil¬
late outside toward the apex, with a central resin¬
iferous vein. Anthers 1.3— 1.7 mm long including
appendage; apical appendage oblong-ovate, obtuse;
anther base shortly sagittate; filament collar basally
swollen (“balusterform”). Style branches linear,
0.7—1 mm long with stigmatic areas separated, apex
truncate with short sweeping-hairs. Cypsela nar¬
rowly elliptic-oblong, terete, 2—2.5 mm long, ca. 0.6
mm wide, with ca. 8 light brown ribs, white-villous
on and between ribs, hairs mucilaginous when
soaked. Pappus bristles numerous, erect, 3—4 mm
long, slender, minutely barbellate, white, caducous.

The single individual seen by me was growing
from a crevice on a vertical rock face with a western
aspect, and the specimens collected by Wild were
stated to grow among crags. The new species be¬
longs to Senecio sect. Senecio, but it is difficult to
point out any very close relatives. There are a num¬
ber of Senecio species with pinnatisect leaves in
southern Africa, but they are all clearly distinct.
Some are annuals with either purple florets (e.g., .S.
cakilefolius DC., S. eenii (S. Moore) Merxmiiller) or
yellow florets, but then with smaller capitula (e.g.,
S. cryphiactis 0. Hoffmann, S. piptocoma 0. Hoff¬
mann). Some forms of S. consanguineus DC. are

Novon 9: 245-247. 1999.

246

Novon

Figure 1. Senecio aetfatensis B. Nordenstam. — A. Flowering branch. — B. Capitulum. — C. Bay floret (most pappus
bristles removed). — D. Disc floret (most pappus bristles removed). — E. Corolla of disc floret, opened longitudinally.
— F. Cypsela, pappus removed. — G. Stamens. — H. Style branches from disc floret. Drawn by the author (based on
Nordenstam 9292 . S).

Volume 9, Number 2
1999

Nordenstam

Senecio aetfatensis from Zimbabwe

247

also superficially similar to the new species, hut
clearly distinct by the erect annual habit, longer
involucral bracts, and short rays. Among perennial
species with some resemblance, S. cineroscens Ai-
ton has leaves white-woolly below and larger flower
heads, S. poseidonis Hilliard & Burtt has purple or
white florets, and S. rhyncholaenus DC. has viscid
stems and leaves and discoid, white-flowered, nar¬
row capitula. The closest relative may be S. s ub-
rubrifolius 0. Hoffmann, a species known from the
Orange Free State, Transvaal, Natal, and Lesotho,
similar in habit and leaf shape, but different by the
turbinate, discoid flower heads, and the glandular
pubescence making the plant viscid.

The bushy habit, the pinnatisect flat leaves, and
the yellow, distinctly radiate capitula of S. aetfaten¬
sis are reminiscent of the genus Cineraria, which
explains the labeling of the paratype as ” Cineraria
sp.” However, this is a superficial resemblance, the
new species lacking the generic characters of Cin¬
eraria such as the flattened cypsela.

Paratypes. ZIMBABWE. Chimanimani Mts., Mt.
Peza, 6500 ft., 15 Oct. 1950, //. Wild 3621 (BR, S,
SRGH). [A pencilled note on the SRGH sheet suggests
that Gilliland 1890 may be the same species. The latter
collection was searched for but not found in several pos¬
sible herbaria (e.g., J, K, LI), PRE, SRGH).|

New Taxa of Sarcopera and Marcgraviastrum (Marcgraviaeeae)

from the Guayana Shield

Adrian C. de Roon

Herbarium Division, Department of Plant Ecology ami Evolutionary Biology, Utrecht Univer¬
sity, Heidelberglaan 2, NL-3584 CS Utrecht, The Netherlands

Stefan Dressier

Forschungsinstitut Senckenberg, Herbarium, Senckenberganlage 25,

D-60325 Frankfurt/Main, Germany

ABSTRACT. A new species, Sarcopera flammifera,
and a new subspecies, Sarcopera tepuiensis subsp.
coccinea, are described. Three new combinations,
Sarcopera aurantiaca, Sarcopera tepuiensis, and
Marcgraviastrum pendulum, are validated.

The planned treatments of Marcgraviaeeae in the
Flora oj the Venezuelan Guayana and Flora of the
Guianas series necessitate publication in ad¬
vance of some new taxa in the genera Marcgra¬
viastrum and Sarcopera, two genera recently val¬
idated by de Roon and S. Dressier (1997). A
number of new combinations and new taxa were
proposed by Bedell (1985) in her doctoral thesis,
but subsequently only some of them were validly
published (Bedell. 1988, 1993; de Roon &
Dressier, 1997).

Sarcopera Bedell is based on Norantea subg.
Pseudostachyum Delpino (Delpino, 1869). The
genus is characterized by a spicate inflorescence
with the nectariferous bracts inserted at the base
of the small flowers. In her thesis Bedell pro¬
posed three new species of Sarcopera from the
region, “S. atrovinosa, S. coccinea and S. flam-
mifera .” A new study of the available collections
led us to the conclusion that only S. flammifera
should be recognized as a new species. The pro¬
posed “S. coccinea" is here treated as a subspe¬
cies of S. tepuiensis.

Marcgraviastrum (Wittmack ex Szyszylowicz)
de Roon & S. Dressier is based on Norantea sub-
sect. Marcgraviastrum Wittmack ex Szyszylowicz
(Szyszylowicz, 1893). The genus is characterized
by an umbelliform inflorescence.

Two generic transfers proposed by Bedell are
herein validated and attributed to her, following an
agreement between her and the senior author.

Sarcopera aurantiaca (Spruce ex Gilg) de Roon
& S. Dressier, comb. nov. Basionym: Norantea
aurantiaca Spruce ex Gilg, Bot. Jahrb. Syst.
25, Beibl. 60: 32. 1898. TYPE: Brazil. Ama¬
zonas: prope Panure ad Rio Vaupes, Spruce
2719 (holotype, B destroyed, photographs F,
GH, MO, MICH, US; lectotype, selected here,
K; isolectotypes. BM, BR, C, E, G, GH, GOET,
LE, NY, OXF, P, TCD, W).

Wittmack (1878) cited Norantea aurantiaca
Spruce (“Msc. in coll, ad n. 2719”), a nomen nu¬
dum, as a synonym of Norantea anomala consid¬
ering the Spruce collection to be a “forma juveni¬
lis.” Though Wittmack in a note gave a short
description of Spruce’s collection, it was not his
intention to describe or publish a new species. Gilg
(1898) was the first to recognize and validly publish
Norantea aurantiaca Spruce as a separate species,
differing from Norantea anomala in the inflores¬
cence, the shape, the venation, and the hypophyl-
lous glands of the leaves and in partic ular the pres¬
ence of two large hypophyllous glands, one on each
side, in the apical part of the blade and some scat¬
tered very small ones. Thereupon Norantea auran¬
tiaca Spruce became validated, and consequently
Norantea aurantiaca Spruce ex Gilg is the correct
basionym of Sarcopera aurantiaca. Bedell (1985)
attributed Norantea aurantiaca Spruce to Gilg and
Werdermann (1925), overlooking the earlier publi¬
cation of Gilg (1898). Ferreira (1995) was also
wrong when she published Norantea aurantiaca
Spruce ex Ferreira. She erroneously considered
Norantea aurantica Spruce as used by Gilg (1898)
and by Gilg and Werdermann (1925) to be a nomen
nudum, since no Latin diagnosis was supplied.
This, however, was not required at that time (Art.
36.1 ICBN; Greuter et al., 1994). Because of its

Novon 9: 248-252. 1999.

Volume 9, Number 2
1999

de Roon & Dressier
Marcgraviaceae from Guayana Shield

249

spicate inflorescence this species clearly belongs in
Sarcopera.

Sarcopera flammifera de Roon & Bedell, sp.
nov. TYPE: Brazil. Roraima: Sierra Tepequem,
Maguire & Maguire 40104 (holotype, U; iso¬
types, MARY not seen, NY). Figure 1.

Frutex vimineus sarmentosus, plerumque scandens vel
liana. Folia ramorum floriferorum petiolis 3—10 mm longis;
lamina coriacea, elliptica, obovato-elliptica vel interdum
elliptico-oblonga, 5— 9(— 12) cm longa, 2—4. 5(— 6.5) cm lata,
apice obtusa vel rotundata vel raro perbrev iter acuminata,
basi acuta vel obtusa aut rotundata; glandulae hypophyl-
lae plerumque paucae, variabiles quoad magnitudinem et
numerum. nonnunquam nullae, interdum 1—3 conspicuae,
dispersae, et aliquot minutae secus marginem dispositae.
Inflorescentia spicata multiflora, rhachide (15-)2Q— 35(—
45) cm longa, badia; bracteae nectariferae cucullatae vel
cochleares, 10—15 mm longae. Flores sessiles; petala el¬
liptica vel elliptico-oblonga, 2.5—4 mm longa, 1.5— 2.5 mm
lata; stamina 7— 10(— 13); ovarium 3-4-loculare. Fructus
globosus, ad 1 cm diametro.

Scandent sprawling shrub or small liana; branch-
lets subterete, the older branches with a grayish
bark, often with striations of lenticels, the younger
ones reddish brown. Leaves coriaceous, dark green
above, dull and paler below, when dried grayish,
greenish gray, or dark brown colored above, light
to dark brown beneath; petiole 3-10 mm long, 1.5—
2 mm wide, flattened or canaliculate above; blade
elliptic to obovate, 5— 9(-12) X 2-4.5(-6.5) cm, ba-
sally acute to rounded, apically obtuse to rounded,
rarely acute, mucronate when young, retuse alter
loss of the deciduous mucro; margin revolute; hy-
pophyllous glands varying in number and size, of¬
ten 1—3 distinct small to medium-sized ones, 1-2
mm diam. with a light-colored circumvallation,
mostly only in the apical hall, sometimes very mi¬
nute and only visible with a hand lens or lacking
and only the two glands at the base of the midrib
present; midvein flat to slightly prominent above,
strongly prominent below, lateral veins usually
(slightly) prominent below, sometimes obscure. In¬
florescence (15— )20— 35(^45) cm long, with 100—
140 sessile flowers; rhachis 3—5 mm thick at the
base, sulcate when dried, dark reddish brown; fo-
liaceous bract ovate, 1 X 0.5 cm, with one pair of
hypophyllous glands; nectariferous bracts leathery
and succulent, deep red to maroon, ladle-shaped,
ea. 1—1.5 cm long, becoming progressively reduced
toward the base of the rachis, stalk flattened, wid¬
ened toward the cup, ca. 8—10 mm long, cup to ca.
5 mm long and ca. 4 mm wide. Flowers leathery,
deep red, 3—4 mm diam.; bracteoles triangular to
ovate, ca. 1 mm long; sepals sub- to semi-orbicular,
ca. 1 mm long, 1-2 mm wide; petals elliptic to

elliptic-oblong, 2.5-4 mm long, 1.5— 2.5 mm wide,
free or slightly connate at base, reflexed at anthesis;
stamens 7-10(-13), filaments flattened, 1.5-2 mm
long, connate at base and adnate to the petals, an¬
thers triangular-ovate, base cordate, ca. 1 mm long,
pollen magenta; ovary ovate-subglobose, ca. 1—1.5
mm long, 3-4-loculed. Fruit dark red to black-pur¬
ple, globose, 5—10 mm diam.; seeds few, reniform,
ca. 2 mm long, black, shining, reticulate.

Distribution. Sarcopera flammifera occurs in a
variety of habitats (forests, along rivers, rocky
slopes, and savannas at altitudes from 100 to 1800
in) in the Caura and Caronf basins of Bolivar state
in Venezuela and in northern Brazil.

Etymology. The name refers to the blazing in¬
florescence with its red-colored nectariferous bracts
giving the impression of little flames.

The species is distinguished from Sarcopera te-
puiensis, which is found on tepuis of Bolivar state
at altitudes above 1000 m, by its smaller and some¬
what less coriaceous leaves, and the lower number
of stamens. The species is also related to Sarcopera
aurantiaca from Colombia and Brazil, from which
it is mainly distinguished by the shape of the
leaves.

Remarks. Bedell (1985) described the pollen of
this species as pale yellow. Indeed in some collec¬
tions, e.g., Steyermark 90246 and 113260 , the pol¬
len seems to have this color, but in this and other
cases it is doubtful whether the pollen is mature.
In some other collections, e.g., Maguire & Maguire
40151 and Maguire et al. 53529 , there is no doubt
that the pollen is magenta-colored. Therefore we
assume that the pollen of Sarcopera flammifera has
no different color from that of all other species of
the genus.

The only known collection of the proposed “5.
atrovinosa ” (Bedell 1985), Agostini 402 (NY, U, US,
VEN), from pequena meseta del Norte de Serrania
Cararuban, SE de Canaima, Bolivar state, appears
to be no more than an individual of S. flammifera
with an unusual leaf shape.

Several collections, e.g., Holst & Liesner 3127
(MO, NY) and 3384 (MO) from the valley of Rio
Coro-Coro, W of Serrania de Yutaje, Maguire &
Maguire 35109 (NY, U) from Serrania de Yutaje,
Rio Manapiare, Cano Yutaje, Phelps & Hitchcock 3
(NY, VEN) and Phelps 114 (VEN) from Cerro Yavi,
all in northern Amazonas state, and Steyermark et
al. 109410 (NY, US, VEN) from selva de galeria
Rio Marajano, cumbre de Cerro Jaua in western
Bolivar state, collected at altitudes of 650-1800 m,
probably belong to Sarcopera flammifera , but are
different because of larger leaves and the presence

250

Novon

Volume 9, Number 2
1999

de Roon & Dressier
Marcgraviaceae from Guayana Shield

251

of rather numerous foliar sclereids. Bedell (1985)
wrongly placed some of these collections in her
“ Sarcopera coccinea ” (= Sarcopera tepuiensis
subsp. coccinea ) from which they are distinguished
by the shape of the leaves (obovate vs. oblanceo-
late) as well as by the number of the stamens (8-
12 vs. 15—30). Possibly these specimens belong to
an undescribed subspecies of Sarcopera flammi-
fera.

Specimens examined. VENEZUELA. Bolivar: pe-
quena meseta del Norte de Serranfa Cararuba, Agostini
402 (NY, U, US, VEN); Cerro Bolivar, Ciudad Piar, Aris-
teguieta 2184 (NY, VEN); 2 km de la margen izquierda
del medio Rio Paragua, 102 km SW de Ea Paragua, Ay-
mard 5967 (MO); cerca del Rfo Cushimf, afluente del Ere-
hato. Brewer 60 (VEN); Rfo Caronf, Raudales de Arepu-
elii, Cardona 1935 (US, VEN); Rfo Caronf, Cerro
Arepuchi, Cardona 1940 (NY. US. VEN); Rfo Carrao, Sal¬
to Hacha, Cardona 2842 (US); Cerro Guaiquinima, De-
lascio & Ortiz 14290 (VEN); afluentes mas occidental del
Rfo Carapo, Cerro Carapo, en el cumbre del Guaiquinima,
Duno & Brewer 331 (VEN); Parque Nacional “Canaima,”
10 km E of Cerro Venado, 25 km ENE of Canaima. Duno
de Stefano et al. 467 (VEN); Rfo Chicanan. 2 km S of Rfo
Chilian. Maguire et al. 53529 (NY. U, VEN); Rfo Paragua
near "Minas de Manaima,” Stergios 10297 (NY. VEN);
Sierra Ichun, N of Salto Marfa Espuma along Rfo lehun.
Steyermark 90246 (NY, U, US, VEN); Sabana de Cusimi.
near Rfo Cusimi, affluent of Rfo Erebato, Steyermark
93162 (NY, U. VEN); Cerro Guaiquinima, Salto del Rfo
Szczerbanari, Steyermark et al. 113260 (E, MO, U, US,
VEN). BRAZIL. Amazonas: Serra Araea, Rosa A Lira
2287 (U). Rorainia: Sierra Tepequem, Maguire & Ma¬
guire 40104 (NY. U); Sierra Tepequem, near E escarp¬
ment. Maguire & Maguire 40151 (NY, U); Sierra Tepe¬
quem, upper plateau and summit. Prance et al. 4417
(A AU. K. NY, U).

Sarcopera tepuiensis (de Roon) Bedell, comb,
nov. Basionym: Norantea tepuiensis de Roon,
Acta Bot. Venez. 2: 247. 1967. TYPE: Vene¬
zuela. Bolivar: Chimanta Massif, NW slopes of
Churi-tepui (Muru-tepui), along Rfo Sarren,
Wurdack 34353 (holotype, U; isotypes, MARY
not seen, NY, US).

Sarcopera tepuiensis subsp. coccinea de Roon
& S. Dressier, subsp. nov. TYPE: Venezuela.
Amazonas: Cerro Yapacana, Maguire et al.
30687 (holotype, U; isotype, NY).

A subsp. tepuiensis differ! foliis plerumque oblanceo-
latis glandulis hypophyllis parvis numcrosis biseriatis.

Liana to 30 m or erect to sprawling shrub or tree

to 10 m tall; older branches grayish, the younger
ones reddish brown. Leaves coriaceous, dark dull
green above, paler below, when dried greenish to
yellowish brown or dark brown above and light to
dark brown below; petiole rather stout, 0.5—2 cm
long, ca. 3 mm wide; blade obovate-oblong to ob-
lanceolate, 8-22 X 3.5— 7.5 cm, basally acute to
cuneate, apically obtuse or rounded, mucronate
when young, afterward often retuse, margins slight¬
ly revolute, hypophyllous glands small, to 0.5 mm
diam., with a minute pore, 2 at the base near the
midrib and 5—8 in an often irregular row at ca. 5
mm from the margin in the middle of the leaf and
closer to the margin toward the top. Inflorescence
30-50 cm long with 130-155 sessile flowers; rachis
stout, 5—7 mm diam. at the base, longitudinally
grooved and yellowish to reddish brown or dark
brown when dried; foliaceous bract elliptic, ca. 2.5
cm long, 0.5—1 cm wide, with 1—2 pairs of hypo¬
phyllous glands; nectariferous bracts succulent,
bright red, ladle- to helmet-shaped, fully developed
in the upper part of the rachis, becoming smaller
to wholly reduced to the base, 1—1.5 cm long, the
cup 3—5 mm long, ca. 5 mm wide, the stalk flat¬
tened, 5-10 mm long, inserted at the base of the
flowers. Flowers sessile, leatherv, bright red; brae-
teoles ± deltoid, 0.5—1 mm long, to 1.5 mm wide;
sepals broadly suborbicular, ca. 1 mm long, 1-2
mm wide; petals oblong, 3—5 mm long, ca. 2 mm
wide; stamens (15— )20— 25(— 30) in two or some¬
times three whorls; filaments flattened, connate for
Yi— Va of their length, ca. 2 mm long, in the outer
row longer than in the inner row, and adnate to the
petals; anthers triangular to ovate, cordate-sagittate
at base, ca. 1.5 mm long, pollen magenta; ovary
ovate or ovate-conical, 1—2 mm long, 1-1.5 mm
diam., 3-loculed with 6—8 ovules per locule. Fruit
dull red, ca. 1 cm diam.; seeds reniform, ca. 2 mm
long, shining, reddish black.

Distribution. Sarcopera tepuiensis subsp. cocci¬
nea occurs on several tepuis of Amazonas state in
Venezuela, at altitudes of 1000-1400 m.

Etymology. The name refers to the bright scar¬
let nectariferous bracts.

The differences between Sarcopera tepuiensis
subsp. tepuiensis and subspecies coccinea are main¬
ly in the shape of the leaves and the number of the
hypophyllous glands. In subspecies coccinea the
leaves are oblong-obovate to oblaneeolate with two

Eigure I. Sarcopera flammifera de Roon & Bedell. — A. Flowering branch. — B. Foliaceous bract. — C. Hypophyllous
gland. — D. Nectariferous bract. — E. Flower. — E. Fruit with persistent sepals. (A— I) from Maguire 40104\ E from
Steyermark 93 1 62: E from Prance et al. 4417.)

252

Novon

glands at the base and two rows of 5—8 small hy-
pophyllous glands along the margin, while subspe¬
cies tepuiensis has obovate to oblong-obovate leaves
with two glands at the base and usually only 1-2
hypophyllous glands per side. In our estimation
these differences are considered too variable to dis¬
tinguish the two taxa as separate species.

The two subspecies ol Sarcopera tepuiensis oc¬
cupy separate geographic areas: subspecies cocci-
nea is known from mountains in southern Amazon¬
as state in Venezuela and from Serra Araca in
adjacent Brazil; subspecies tepuiensis is known
from over 25 collections from the tepuis of eastern
Bolivar state in Venezuela and the Pakaraima Mts.
in adjacent Guyana. Most collections of both sub¬
species are from above 1000 m.

Paratypes. VENEZUELA. Amazonas: Cerro Duida,
Farinas et al. 488 (NY. U. VEN); Serranfa Paru, Cerro
Asisa, near laguna Asisa, Hoyos & Morillo 95 [herb.
96222/ (VEN); Cerro Marahuaca, above Rfo Yameduaka,
Liesner 17679 (MO); Cerro Marahuaca, N of Sima Camp,
Liesner 18477 (MO, VEN); Cerro Yapacana, Maguire et
al. 80680 (NY), 80687 (NY, U); Cerro de la Neblina, Rfo
Yatua. Cano Grande, Maguire et al. 42514 (MO, NY, U);
Cerro Yapacana, Steyermarh A Hunting 108167 (MO, NY,
U, VEN); Cerro Marahuaca, Sima Camp, Cano Negro.
Steyermark & Holst 180479 (VEN). BRAZIL. Amazonas:
Serra Araca, plateau of northern massif. Prance et al.
29080 (K, NY).

Marograviastrum pendulum (Lanjouw & van
Heerdt) Bedell, comb. nov. Basionym: Noran-
tea pendula Lanjouw & van Heerdt, Recueil
Trav. Bot. Need. 37: 282. 1940. TYPE: Suri¬
name. Emmagebergte, 700 m, B. W. 5682 (ho-
lotype, U).

The leaves with a dense tangle of foliar sclereids,
the umbelliform inflorescence, and the large flowers
with over 50 stamens are the convincing characters
to place this species in Marcgraviastrum.

Distribution. The species is known from moun¬
tains in Guyana, Suriname, French Guiana, and
eastern Venezuela. A representative collection from
the region is from Bolivar state: Alto Rfo Cuyuni,
Rfo Uiri-yuk, Maguire et al. 46979 (U, VEN).

Acknowledgments. We are indebted to the cu¬
rators of the herbaria of which material was avail¬
able for our study. We thank Hendrik Rypkema for
drawing the illustration, Lubbert Westra for helping
with the Latin diagnoses and the English text, and
Gea Zijlstra for her advice regarding nomenclature.
We are also grateful to Paul Berry for valuable
hints, which improved the manuscript.

Literature Cited

Redell, H. G. 1985. A Generic Revision of Marcgravi-
aceae I. The Norantea Complex. Ph.D. Dissertation
(ined.). University of Maryland, College Park.

- . 1988. Marcgraviaceae. In: R. A. Howard (editor),

Elora ol the Lesser Antilles 5: 300—310. Jamaica Plain,
Massachusetts.

- . 1993. In: L. Brako & J. L. Zarucchi, Catalogue

of the Elowering Plants and Gymnosperms of Peru. Ap¬
pendix 1. Monogr. Syst. Bot. Missouri Bot. Gard. 45:
1256.

Delpino, G. G. F. 1869. Ulteriori osservazioni e consider-
azioni sulla Dicogamia nel regno vegetale. Rivisla mon-
ografica della famiglia delle marcgraviacee precipua-
mente sotto l’aspetto della biologia ossia delle relazioni
di vita esteriore. Atti Soe. I tal. Sci. Nat. 12: 179-213.
Ferreira, G. L. 1995. Estudo Taxonomico das Especies
Brasileiras do Genero Norantea Aublet (Marcgravi¬
aceae). Arch. Jard. Bot. Rio de Janeiro 33(2): 9-53.
Gilg, E. 1898. Marcgraviaceae. In: Ign. Urban (editor),
Plantae novae americanae imprimis Glaziovianae. II.
Bot. Jahrb. Syst. 25. Beibl. 60: 25-35.

- & E. Werdermann. 1925. Marcgraviaceae. In: A.

Engler & K. Prantl (editors). Die natiirlichen Pflanzen-
familien, ed. 2, 21: 94-106. Wilhelm Engelmann, Leip¬
zig.

Greuter, W„ E. R. Barrie, H. M. Runlet, W. G. Chaloner,
V. Demoulin, D. L. Hawksworth, P. M. Jprgensen, D. H.
Nicolson, P. C. Silva, P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Roon, A. C. de & S. Dressier. 1997. New taxa of Norantea
Aublet s.l. (Marcgraviaceae) from Central America and
adjacent South America. Bot. Jahrb. Syst. 119: 327—
335.

Szyszylowicz, I. 1893. Marcgraviaceae. In: A. Engler & K.
Prantl (editors), Die natiirlichen Pflanzenfamilien 3(6):
157-164. Wilhelm Engelmann. Leipzig.

Wittmack, L. 1878. Marcgraviaceae. In: C. F. P. von Mar¬
tins (editor), Flora Brasiliensis 12(1): 213—258. Miin-
chen. Wien. Leipzig.

A New Species of Elvasici (Oehnaceae) from Mesoamerica with
Discussion of Suhgeneric Classification and Phytogeography

Claude Sastre

Museum National d’Histoire Naturelle, Laboratoire de Phanerogamic, 16 rue de Buffon, 75005

Paris, France

Caroline Whiteboard and Sandra Knapp

Department of Botany, The Natural History Museum, Cromwell Boad, London SW7 5BD,

United Kingdom

Abstract. A new species of Elvasia, E. bisepala,
is described from Belize and Honduras. Elvasia bi-
sepala possesses unique characters that help to
clarify subgeneric classification in the genus and
further expands the distribution of Elvasia from
South America well into Central America. The new
species is described and illustrated, and the sub¬
generic c lassification of Elvasia is reviewed and re¬
vised.

The genus Elvasia has previously been consid¬
ered to be exclusively South American and con¬
centrated in the Amazon basin and the Guianas
(Sastre & Lescure, 1978; Sastre, 1987), with a sin¬
gle species occurring in Bahia, Brazil (E. tricarpel-
lata Sastre), and another recently collected in Pan¬
ama and Costa Rica (E. elvasioides (Planchon) Gilg;
pers. obs.). Currently there are 11 (including the
new species described here) accepted species in
the genus, which is distinguished from the rest of
the Oehnaceae in the Neotropics by the uniovulate
carpels that are united in fruit. During a review of
material for the preparation of the Oehnaceae ac¬
count for Flora Mesoamericana, a distinctive new
species of Elvasia was found among recent collec¬
tions from Belize and Honduras. The discovery of
this taxon has clarified some previously ambiguous
points about the infrageneric classification and bio¬
geography of the genus.

Infrageneric classification in Elvasia has been
based primarily on either stamen number and ovary
morphology or fruit characteristics. Planchon
(1846) divided the genus into two subgenera: subg.
Euelvasia Planchon with 8 stamens and a 4-locular
ovary, and subg. Hostmannia Planchon, with 18—
20 stamens and a bilocular ovary. This same basic
division was followed in Engler (1876). Van Tiegh-
em (1902) elevated the genus to the rank of sub¬
family, and divided it into two tribes using fruit
characteristics, the Elvasieae with star-shaped

fruits and the Hostmannieae without star-shaped
fruits. Gilg (1893, 1925) more or less followed ear¬
lier floral-morphology-based schemes in dividing
the genus into two sections: sect. Euelvasia with 8
stamens (including E. calophyllea DC. & E. quin-
queloba Spruce ex Engler) and sect. Hostmannia
with 18—20 stamens (including E. hostmannia
(Planchon) Gilg & E. essequibensis Engler).

Revising the genus throughout its range, Dwyer
(1943) described three sections based on various
combinations of ovary locule and stamen number:
sect. Euelvasia with 7—10 stamens and a 4— 5-loc-
ular ovary (E. calophyllea, E. quinqueloba, and E.
canescens (Van Tieghem) Gilg), sect. Eussequibensa
Dwyer with 10—20 stamens and a 4— 5-loeular ovary
(E. essequibensis and E. brevipedicellata Ule), and
sect. Hostmannia with variable numbers of stamens
and a bilocular ovary (E. elvasioides (Planchon)
Gilg and E. caurensis Pittier). Cowan (1952) de¬
scribed E. sphaerocarpa, which he postulated was
closely related to E. essequibensis. Due to its glo¬
bose, unlobed fruit, he erected a new section for it,
section Conjunctipora Cowan. Additional new spe¬
cies (Cuatrecasas, 1950; Sastre & Lescure, 1978)
can be accommodated in Dwyer’s subgeneric clas¬
sification, with both E. macrostipularis Sastre &
Lescure (closely related to E. sphaerocarpa but with
only 2 carpels) and E. oligandra Cuatrecasas being
placed in section Hostmannia.

Placement of Elvasia tricarpellata Sastre (Sastre,
1981), with its 3-carpellate ovary and 10—15 sta¬
mens, in Dwyer's classification is problematic. With
the discovery of E. bisepala, described here, a sim¬
pler and less ambiguous subgeneric classification
of the genus Elvasia can be proposed with two sec¬
tions: section Elvasia (= Euelvasia ), characterized
by star-shaped fruits, and section Hostmannia
(Planchon) Gilg, characterized by globular fruits.
The component species in each section, with their

No VON 9: 253-256. 1999.

254

Novon

Table 1. Subgeneric classification and distribution of taxa in Elvasia (Ochnaceae).

Subgenus

Species

Distribution

Elvasia

E. calophyllea DC.

Colombia (Vaupes), Brazil (along the

Rio Negro), Guyana

E. quinqueloba Spruce ex Engler

Colombia (Rio Guainia), Brazil (along
the Rio Negro)

E. canescens (van Tieghem) Gilg

Venezuela (upper Rio Orinoco), Brazil
(along the Rio Negro and Rio Xingu)

E. essequibensis Engler

Guyana

Hostmannia (Planchon)

E. elvasioides (Planchon) Cilg

Venezuela (Amazonas, Bolivar), Suri-

Cilg

name, French Guiana, Brazil (Para),
Panama, Costa Rica

E. oligandra Cuatrecasas

Colombia (Norte de Santander)

E. macrostipularis Sastre & Lescure

French Guiana, Brazil (Amapd, Pard)

E. sphaerocarpa Cowan

Guyana

E. bisepala Sastre & Whitefoord

Belize, Honduras

prov. E. tricarpellata Sastre

Brazil (Bahia)

geographic distributions, are presented in Table 1.
Two species are not placed with certainty in the
current subgeneric classification: E. brevipedicella-
ta, of which the type at B was destroyed (the spe¬
cies is known only from photographs of this type),
and E. tricarpellata, which is placed provisionally
in section Hostmannia due to the overall similarity
of its young fruits to those of E. bisepala. It is clear
that section Elvasia is fundamentally Amazonian,
while section Hostmannia is more or less circum-
Amazonian and Central American.

Elvasia bisepala Sastre & Whitefoord, in Sastre,
Whitefoord & Knapp, sp. nov. TYPE: Belize.
Toledo District: Maya Mountains, lower part,
Richardson Creek (lower part), affluent of
Bladen Branch, 88°46-48'W, 16°33'N, short
semi-deciduous dry forest along stream, 100-
250 m, 2, 3, 11 Mar. 1987 (fl), Davidse &
Brant 32340 (holotype, BM; isotypes, BFD,
MO, P). Figure 1.

Ab E. tricarpellata Sastre foliarum laminis minoribus
(6—13 X 2.5—4 cm vs. 8-18 X 4—6 cm), inflorescentiis
minoribus (7—9 cm vs. 6—20 cm longiore) et sepalorum
numero 2 vs. 3-4 differt.

Trees 6-10 m tall, branches glabrous. Leaves 6—
13 X 2.5-4 cm, elliptic to narrowly elliptic, cori¬
aceous, principal vein prominent abaxially, sharply
keeled adaxially, the secondary veins fine and par¬
allel, perpendicular with the principal vein; base
cuneate or sometimes attenuate, margin with mi¬
nute, glandular, black teeth, appearing undulate
when dry, the apex acuminate; petiole 3-6 mm
long; stipules 1.5— 3.5 mm long, triangular, entire

or apparently trifid with a stouter central division,
appearing tattered and laciniate with age. Inflores¬
cence paniculate, terminal on short stems, 7-9 cm
long with 1—5 branches, the branches 2.5-4 cm
long, minutely pubescent; bracts similar to the stip¬
ules, 1-2.5 mm long, triangular, at branch points
and at the base of the flowers; pedicel 4-7 mm
long, articulate near the base, very shortly pedun¬
culate. Flowers solitary or occasionally clustered in
groups of 2-3, with the sepals 2 (rarely 3), ca. 4
mm diam., subcircular, coriaceous; petals 3-4(— 5),
5—5.5 X 4.5 mm, ovate to obovate, membranous, a
few somewhat coriaceous at the base; stamens 9-
14, the filaments ca. 1.5 mm long, the anthers ca.
2.5 mm long, dehiscence porose; ovary depressed-
globular, tricarpellate; style single, (2-)3^1 mm
long. Fruit 1— 2-loculate by abortion, subglobular,
8—10 mm diam., leathery (?), the style base persis¬
tent.

Distribution. In semi-deciduous forests in Be¬
lize and adjacent Honduras, 100-350 m.

Elvasia bisepala differs from E. tricarpellata in
its smaller leaf blades (6-13 X 2.5-4 cm vs. 8-18
X 4—6 cm), smaller inflorescences (7—9 cm vs. 6—
20 cm long), the number of sepals (2 (rarely 3) vs.
consistently 3^4) and its Central American distri¬
bution. Its globular fruit places Elvasia bisepala in
section Hostmannia (see above. Fig. 1).

Paratypes. BELIZE. Toledo District: Maya Moun¬
tains S., Bladen Nature reserve, Ek Xux canyon, 16°30'N,
88°55'W, forested canyon bottom with large exposed boul¬
ders, 320 m, 18 May 1996 (fl, young fr), Davidse 36088
(BM, MO, P), 18 May 1996 (fl), Davidse 3600 6 (BM, MO,
I’), 20 May 1996 (fr), Davidse 36164 (BM, MO, P). HON-

Volume 9, Number 2
1999

Sastre et al.
Elvasia bisepala

255

Figure 1. Elvasia bisepala Sastre & Whitefoord, from Davidse & Brant 32340 (BM), with details of flower and fruit.
Seale bar in upper left = 5 cm, all others = 5 mm.

256

Novon

DURAS. Atlantida: Pico Bonito, base of N, K of new
CURLA (Centro Universitario Regional del Littoral Atlan-
tieo), 15°42'N, 86°50'W, upland primary forest on slope,
300 m, 14 May 1993 (fr). Evans 1663 (MO).

Acknowledgments. We thank M. Tebbs for the
illustration and the curators of the herbaria men¬
tioned in the text for the loan of specimens.

Literature Cited

Cowan, R. S. 1952. Plant explorations of G. Wilson-
Browne, S.J., in British Guiana I. Kanaku Mountains.
Brittonia 7: 389^114.

Cuatrecasas, J. 1950. New and noteworthy Colombian
trees. Tropical Woods 96: 37-47.

I)w\ er, J. 1). 1943. The taxonomy of the monogeneric tribe
Elvasieae (Ochnaceae). Bull. Torrey Bot. Club 70: 42—19.

Engler, A. 1876. Ochnaceae. In C. P. E. von Martius, Flo¬
ra Brasiliensis 12(2): 297-306.

Gilg, E. 1893. Ochnaceae. In A. Engler, Die natiirlichen
Pflanzenfamilien 3(6): 131—153.

- . 1925. Ochnaceae. In A. Engler, Die natiirlichen

Pflanzenfamilien 2(21): 53—87.

Planchon, J. E. 1846. Sur le genre Godoya et ses ana¬
logues. London J. Bot. 5: 584—600, 644-656.

Saslre, C. 1981. Ochnac£es nouvelles du Bresil. Bull.
Jard. Nat. Belg. 51: 397-413.

- . 1987. Studies on the flora of the Guianas 30:

Considerations phytog6ographiques sur les 0chnac6es
Guyanaises. C.R. Soc. Biogeogr. 63: 89-97.

- & J. P. Lescure. 1978. Elvasia elvasioides (Och-

nac6es) et les espfeces affines. Caldasia 12: 131-144.
Tieghem, P. van. 1902. Sur les Oclmac£es. Ann. Sci. Nat..
Bot. 16: 161-416.

A New Species of Pilocarpus Vahl (Rutaceae) from Peru

Ijadislau A. Skorupa

Centro Nacional de Pesquisa de Recursos Geneticos e Biotecnologia, Embrapa/Cenargen, CP
02372, CEP 70.849-970, Brasilia, DF, Brazil

ABSTRACT. A new species of Pilocarpus Vahl (Ru¬
taceae) from Peru is described and illustrated as
Pilocarpus manuensis Skorupa. The new taxon is
known only from the type locality in Parque Na¬
cional Manu, Peru, and resembles Pilocarpus de-
merarae Sandwith from Guyana.

Pilocarpus is a Neotropical genus of arborescent
and shrubby plants, ranging from southern Mexico
to southern South America. Throughout its geo¬
graphic range it may be found in various habitats,
such as the moist Atlantic Forest or dry Caatinga
in Brazil. In a recent revision of the genus (Sko¬
rupa, 1996) 16 species were recognized, including
3 new species from Brazil (Skorupa, 1998) and 1
from Peru. The purpose of this paper is to describe
this new Peruvian species.

Pilocarpus manuensis Skorupa, sp. nov. TYPE:
Peru. Madre de Dios: Manu, Parque Nacional
Manu, Pakitsa Station, trails to N and NE of
camp, 11°56'S, 71°16'W, 350 m, 13 Sep.
1989, R. B. Foster & H. Beltran 13183 (ho-
lotype, NY; isotypes, F, MO). Figure 1A-C.

Folia imparipinnata; lamina elliptica vel anguste ellip-
tica, apice acuminato; fructus ex 1—3 coccis constans, coc¬
ci 12.0-12.5 X 11.0—12.0 mm; seminibus vinaceis quasi
nigris, ca. 10.0 X 8.0 mm, hilum ovatum et planum, ca.
3.0 X 1.8 mm. Flores non vidi.

Small tree, reported as 3.0 in tall; terminal
branchlets 7.0 mm diam., pale brown, glabrous.
Leaves imparipinnate, 3-jugate, alternate, ca. 45.0
X 42.0 cm, distance between pairs 9.0— 9.5 cm, the
rachis ca. 23.0 cm X 3.0— 3.5 mm, continuing be¬
yond the upper pair of leaflets ca. 4.0 cm, glabrous,
slightly canaliculate near the leaflets, slightly
winged, the wings 0. 3-0.5 mm broad, glabrous; pet¬
iole inserted at ca. 40°, 13.0—13.5 cm X ca. 3.5
mm, glabrous, slightly canaliculate near the first
pair of leaflets, slightly winged, the wings ca. 0.3
mm; lateral leaflets subopposite, the lowermost al¬
ternate; the blade shining green above, pale be¬
neath, chartaceous, slightly bullate among second¬
ary veins, glabrous, 18.3-21.0 X 7. 5-7. 8 cm,
narrowly elliptic to elliptic, acute to acuminate at

apex, acumen to 15.0 mm, base slightly asymmet¬
ric, attenuate, the margin revolute, entire or slightly
undulate; venation brochidodromous, the midvein
plane on adaxial surface and prominent on abaxial
surface, secondary veins 14—16 pairs, slightly
prominent on adaxial surface, prominent on abaxial
surface, angle of divergence ca. 65°; petiolule 2.0-
4.0 X 2.0-2. 5 mm, inserted at 50—60°, glabrous,
slightly winged, the wings ca. 0.3 mm broad, slight¬
ly canaliculate near wings; terminal leaflet blade
22.5 X 8.5 cm, narrowly elliptic, acuminate at
apex, asymmetric and attenuate at base. Raceme
terminal, erect, to 55.0 cm long and 3.0— 3.5 mm
wide in fruit, greenish brown, glabrous; bracts at
base 5.0— 6.0 mm long, triangular to oblong, pubes¬
cent near apex or glabrous; calyx 5-toothed, teeth
unequal, 0.2-0. 6 X 0. 5-1.0 mm, ovate to triangu¬
lar, rounded at apex, glabrous, ciliolate; bractlets
2, ovate with subacuminate apex, glabrous, cilio¬
late, alternate and inserted near middle of the ped¬
icel. Mericarps 1-3, 12.0-12.5 X 11.0-12.0 mm,
obovoid, slightly rounded at apex, pale brown be¬
coming dark, glabrous, conspicuous brown glands
becoming dark, line of dehiscence reaching to Va-
Vi of distance from apex, pedicels 6.5— 9.5 mm long
in fruit, glabrous; seeds ca. 10.0 X 8.0 mm, ob¬
ovoid, testa vinaceous to black, the hilum ca. 3.0
X 1.8 mm, ovate and plane. Flowers not seen.

Phenology. Fruiting specimens collected in
September.

Distribution. Known only from the type locality
in Parque Nacional Manu, southeastern Peru, in
forested high terrace and ravine slopes.

Despite the lack of flowering material, the new
taxon is distinguished easily from other species of
the genus by characters of its vegetative branches,
mericarps, and seeds. Pilocarpus manuensis resem¬
bles Pilocarpus demerarae Sandwith because of its
compound leaves with a rachis somewhat biart i-
culate near the apex and its narrowly elliptic to
elliptic leaflets with an acuminate apex and a bul¬
late blade. The two taxa may be separated, however,
by features of their mericarps and seeds. Pilocarpus
manuensis has obovoid mericarps with a short line
of dehiscence (Va to Vs of distance from apex).

No VON 9: 257-259. 1999.

258

Novon

Volume 9, Number 2
1999

Skorupa

Pilocarpus manuensis from Peru

259

whereas P. demerarae has ovoid to ellipsoid meri¬
carps, larger (14.5—16.0 X 12.5—13.5 mm) and with
a line of dehiscence more accentuated (% to % of
distance from apex) (Fig. ID). The larger (11.0—
12.0 X ea. 9.0 mm), ovoid to ellipsoid seeds of P.
demerarae also have a remarkable hilurn (5.0— 5.8
X 2.5— 3.0 mm) that is conspicuously recurved to¬
ward the apex (Fig. IE), whereas the seeds of P.
manuensis have a plane, not recurved hilum (Fig.

1C). The two species are allopatric in distribution,
the new one known only from Guyana.

Literature Cited

Skorupa, L. A. 1996. Revisao taxonomica de Pilocarpus
Vahl (Rutaceae). Ph.l). Thesis, University of Sao Paulo,
Sao Paulo.

- . 1998. Three New Species of Pilocarpus Vahl (Ru¬
taceae) from Brazil. Novon 8: 447—454.

Figure 1. A— C. Pilocarpus manuensis Skorupa ( R . H. Foster &: H. Beltran 13183, NY). — A. General aspect of the
branch. — B. Mericarps, lateral view. — C, and C_,. Seed, lateral and ventral view. I). E. Pilocarpus demerarae Sandwith
(D. B. Fanshawe 5610, K). — I). Mericarps, lateral view. — E, and E.,. Seed, lateral and ventral view.

New Names, a New Combination, and a New Species of Psychotria
(Rubiaceae: Psychotrieae) from Sao Paulo State, Brazil

Charlotte M. Taylor

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. The new species Psychotria leitana C.
M. Taylor is described, the combination Psychotria
subspathulata (Mueller Argoviensis) C. M. Taylor is
made based on Mapouria subspathulata Mueller
Argoviensis, and the new names Psychotria ararum
C. M. Taylor, for Palicourea glaziovii Standley, and
Psychotria setulifera C. M. Taylor, for Palicourea
hispidula Standley, are provided.

During study of Psychotria L. (Rubiaceae: Psy-
chotrieae) for the “Phanerogamic Flora of Sao Paulo
State” project, the following new species and no-
menclatural needs were discovered. The subgeneric
classification of Psychotria and its separation from
Palicourea Aublet have been outlined previously
(Taylor, 1996). Standley s reason for naming Pali¬
courea glaziovii Standley and Palicourea hispidula
Standley in Palicourea rather than Psychotria is not
clear, except he noted that the corollas in bud were
rather enlarged at the base. He also noted that no
other characteristics of the corollas of these species
were evident; with better material now available,
the corollas of both species can be seen to lack
both the swollen base and the internal ring of pu¬
bescence near this base that distinguish Palicou¬
rea. Within Psychotria , these species belong to sub¬
genus Heteropsychotria Steyermark.

Psychotria ararum C. M. Taylor, nom. nov. Re¬
placed name: Palicourea glaziovii Standley,
Publ. Field Columbian Mus., Bot. Ser. 8: 217.
1930, not Psychotria glaziovii Mueller Argo¬
viensis, in Martius, FI. bras. 6(5): 267. 1881.
TYPE: Brazil. Rio de Janeiro: Serra das Ar-
aras, 1888, A. Glaziou 17033 (holotype, B de¬
stroyed; isotypes, BR, P not seen, fragment F-
970563).

This species is now also known from the Atlantic
coastal forest in southeastern Sao Paulo State and
may be expected in the region between there and
the type locality.

Psychotria leitana C. M. Taylor, sp. nov. TYPE:
Brazil. Sao Paulo: mpio. de Ubatuba, Picin-
guaba, Trilha do Morro do Corsario, 26 Aug.
1991. R. Romero & N. Roque 354 (holotype,
IAC 34744; isotype, HRCB). Figure 1.

Haec species a Psychotria nemorosa Gardner stipulis
elobatis, pedunculo 4—20 mm longo. panicula 1—3.5 cm
longa. pedicellis sub anthesi usque ad 1 mm longis atque
tubo corollino 3.5-4 mm longo distinguitur.

Shrubs or small trees flowering at 1 m bill, to 3 m
tall; stems flattened becoming subterete, glabrous or
densely puberulous to short-pilosulous. Leaves with
blades elliptic, 7-13 cm long, 2.2-5.5 cm wide, at
apex acuminate with slender tips 1-1.5 cm long, at
base cuneate to obtuse, papyraceous to thinly so, on
adaxial face glabrous, on abaxial face densely puber¬
ulous to short-pilosulous; secondary veins 1—8 pairs,
looping to clearly interconnect near margins, with 2—
4 rather weak intersecondary veins present between
pairs of secondary veins, without domatia, on adaxial
lace venation plane, on abaxial face costa, secondary
veins, and reticulated minor venation all prominulous;
margins entire; petioles 5-12 mm long, glabrous to
densely puberulous or short-pilosulous; stipules per¬
sistent, glabrous or densely puberulous to short-pilo¬
sulous, united around the stem into a continuous
sheath, this entire, 0.8-2. 5 mm long, truncate to
broadly rounded, with age becoming indurate and oc¬
casionally splitting into two deltoid lobes. Inflores¬
cences densely puberulous to short-pilosulous, termi¬
nal, peduncles 4—20 mm long, sometimes articulated
with the articulation bearing a reduced leafless stipule
that terminates a basal intemode 3—5 mm long; pan¬
icles corymbiform-rounded to depressed-corymbiform
and nearly flat-topped, 1-3.5 X 2-8 cm, with sec¬
ondary axes 2-A pairs, often subverticillate or irreg¬
ularly distributed along the primary axis, dichoto-
mously branched 1—2 times with the resulting minor
axes becoming rather secund; bracts lacking or nar¬
rowly triangular, to 0.8 mm long, and rather irregu¬
larly distributed, never present at junction of second¬
ary axes; fl<ywers sessile or with pedicels to 1 mm long,
borne in open cymules of 5-11; hypanthium eupuli-
form to cylindrical, densely puberulous; calyx limb
densely puberulous, 0.5-0.8 mm long, lobed for 14-
14, lobes deltoid to rounded; corolla infundibuliform,
cream-colored, externally densely puberulous becom¬
ing glabrescent at anthesis, tube 3.5—4 mm long, ea.
1.2 mm diam. near middle, lobes 4—5. triangular, ca.
1.5 mm long, ac ute. Infructescences similar to inflo-

Novon 9; 260-262. 1999.

Volume 9, Number 2
1999

Taylor

Psychotria from Brazil

261

Figure 1. Psychotria leitana C. M. Taylor. — A. Flowering branch; based on de Assis et al. 348 (UEC). — B. Stipule.
— C. Flower. B, C. based on Tamashiro et al. 18738 (UEC).

rescences except pedicels becoming up to 2 mm long;
fruits broadly ovoid, 3.5-4 X 6 mm, becoming purple,
laterally flattened; pyrenes 2, with 3—5 planar angles
or weak, rounded, longitudinal ridges.

Southeastern Brazil (Sao Paulo State), in pluvial
Atlantic coastal forest. Collected in flower in Feb¬
ruary. March, August, and October, in fruit in Jan¬
uary, February, September, and November.

This new species belongs to subgenus Heterop-
sychotria and is distinguished by its unlobed stip¬
ule sheath, leaves with the venation reticulated and
prominulous on the abaxial surfaces, corymbiform
inflorescences with the branches becoming rather
secund especially as the fruits develop, sessile to
shortly pedicellate flowers, and funnelform corollas
with tubes 3.5—4 mm long. Psychotria leitana has
been confused with P. nemorosa Gardner; the latter
species can be distinguished by its stipules with a
truncate sheath bearing clearly developed triangu¬
lar lobes, inflorescences with peduncles 1.5—3 cm
long, panicles 4—10 X 4.5—15 cm, pedicels 1—5
mm long, and corollas with tubes 5—8 mm long.

The specific epithet honors the late Hermogenes
Leitao, a productive collector and student of Sao
Paulo State’s unusual flora.

Paratypes. BRAZIL. Sao Paulo: mpio. Bertioga, es-
trada Bertioga-Mogi das Cruzes, Tamashiro et al. 18738
(UFC); mpio. Iguape, Estayao Eeologica Jureia-ltatins,
Serra da Jureia. trilfia para a Figueira Grande, Anuncia^ao
& Rossi 374 (MO, SP); mpio. Iguape, Esta<;ao Eeologica
Jureia-ltatins, Serra da Jureia. trilha em diregao a Eazenda
do rio Verde, Anunciayao et al 633 (MO, SP); mpio. Uba-
tuba, Estrada de Itamambuca, Km 35 da Rodovia Rio/
Santos. ,4.s.s/x et al. 348 (HRCB); II ha de Sao Sebastiao,
23°18'S. 46°57'W, Collares 13 (RB); mpio. Ubatuba. Pi-

cinguaba, Estrada da Casa da Farinba, Furlan el al. 6 27
(HRCB). 779 (HRCB), Garcia et al. 140 (HRCB), Ribeiro
et al. 476 (1 AC). Romero et al. 93 (HRCB); mpio. Ubatuba,
Picinguaba, Trilha do Morro do Corsario em direyao ao
mangue, Ribeiro et al. 439 (IAC); mpio. Ubatuba. 20 Jan.
1997, Gajardo s.n. ( I AC-34737). 17 Feb. 1997. Gajardo
s.n. (IAC-34738).

Psychotria setulifera C. M. Taylor, nom. nov. Re¬
placed name: Palicourea hispidula Standley,
Field Mus. Nat. Hist., Bot. Ser. 11: 227. 1936,
not Psychotria hispidula Standley ex Steyer-
mark, Acta Bot. Venez. 4: 97. 1964. TYPE;
Brazil. Rio de Janeiro: Environs de Rio de Ja¬
neiro et D’Ouro Preto, 1883—1884, A. Glaziou
14900 [holotype, K not seen, photo (F neg.
#49916) F],

This species was apparently combined with Psy¬
chotria leiocarpa Chamisso & Schlechtendal by
Mueller. These species can be distinguished by
their pedicels, 1.5—7 mm long in P. leiocarpa vs.
2—10 mm long in P. setulifera. The minor venation
on the abaxial leaf surface is hardly evident and
plane in P. leiocarpa but is clearly evident, closely
reticulated, and prominulous in P. setulifera.

Psychotria suhspathulata (Mueller Argoviensis)
C. M. Taylor, comb. nov. Basionym: Mapouria
subspathulata Mueller Argoviensis, Flora
59(29): 465. 1876. TYPE: Brazil. Rio de Ja¬
neiro: prope Macahe, Riedel 635 (lectotype,
designated here, LE not seen; isotype, B de¬
stroyed, photo (Rockefeller neg. #437) MO).
SYNTYPE: Brazil. Sao Paulo: Ytii, Riedel s.n.
(not located).

262

Novon

As the genera of Psychotrieae are currently cir¬
cumscribed, Mapouria is a synonym of Psychotria
subg. Psychotria (Taylor, 1996). The lectotype des¬
ignated here is chosen over the other syntype be¬
cause of the wide availability of a photo and its
designation With a collection number.

Acknowledgments. 1 thank the curators ol F,
HRCB, IAC, SP, SPF, R, RB, and UEC for access
to specimens; the coordinator of the Rubiaceae for
the “Phanerogamic Flora of Sao Paulo State,” Sig-
rid Luiza Jung Menda^olli, the staff and the stu¬
dents at IAC, UEC, and SP. and especially Elisette
Araujo da Anunciagao (SP), Alice Correa (SP), and
Rosely Buzanelli Torres (IAC), for their generous

scientific collaboration and logistical help; R. E.
Gereau and J. H. Kirkbride, Jr., for very helpful
comments on the manuscript; and R. E. Gereau for
preparation of the Latin diagnosis. This work is part
of the “Phanerogamic Flora of Sao Paulo State”
project, under the coordination of Maria das Gramas
Lapa Wanderley of the Institute Botanico, Sao Pau¬
lo, and was very kindly supported in part by FA-
PESP (Fundagao de Amparo h Pesquisa do Estado
de Sao Paulo).

Literature Cited

Taylor, C. M. 1996. Overview of the Psychotrieae (Rubi¬
aceae) in the Neotropics. Opera Bot. Belg. 7: 261-270.

Morphology and Taxonomy of Arcytophyllum serpyllaceum
(Rubiaceae), a Transfer from Hedyotis

Edward E. Terrell

Norton-Brown Herbarium, H. J. Patterson Hall, University of Maryland, College Park,

Maryland 20742-5815, U.S.A.

Abstract. Morphological and distributional data
are presented for Hedyotis serpyllacea, native to
Guatemala and southern Mexico. The species is
transferred to Arcytophyllum, extending the range
of the genus northward from Costa Rica, Panama,
and South America.

The relationships of the Mexican and Guatema¬
lan species Hedyotis serpyllacea Schlechtendahl
(Rubiaceae; Hedyotideae) have been problematical
for some years. Its flowers, truits, and seeds are not
similar to those of Houstonia (Terrell, 1996), Old-
enlandia, or Hedyotis. In 1893 the species was giv¬
en the name Mallostoma shannonii Donnell Smith,
and later Standley transferred it to Arcytophyllum.
Morphological data demonstrate that Standley was
correct in placing it in Arcytophyllum; however, a
new combination is necessary.

Mena (1990), in a revision of Arcytophyllum
Willdenow ex Schultes & Schultes f., recognized 15
species native to higher elevations in Costa Rica,
Panama, and the Andes of Venezuela, Colombia,
Ecuador, Peru, and Bolivia. Eleven species are
erect subshrubs and four are prostrate, mat-form¬
ing, and suffruticose. Mena (1990) listed the name
Arcytophyllum shannonii (Donnell Smith) Standley
among “Excluded and Dubious Names” as a syn¬
onym for Hedyotis serpyllacea.

Hedyotis serpyllacea grows at high elevations
(1500 to 3500 m) in Guatemala and southern Mex¬
ico (Oaxaca, Veracruz, and Chiapas) and has sev¬
eral vegetative and reproductive morphological
similarities to species of Arcytophyllum. The plants
are prostrate, mat-forming, and suffruticose; the
stems have a “jointed” appearance; the stipules are
generally similar to those of Arcytophyllum ; the
leaves are small, rather thick, and leathery; the ca¬
lyx has intercalycine teeth; the corollas are rather
thick, purplish externally and white within; and the
capsules are broadly oblong, thick-walled, and tar¬
dily dehiscent. All of these characteristics occur
especially in the prostrate, suffruticose species of
Arcytophyllum, to which H. serpyllacea is most
closely allied. Seeds are compressed, rounded in

outline, and with a central punctiform hilum. The
seeds as shown by scanning electron microscopy
are very similar to those of Arcytophyllum muticum
(Weddell) Standley, a prostrate, suffruticose species
of Costa Rica, Panama, and South America.

Seeds of Hedyotideae have been found to be very
important in classification (Terrell, 1996). Seeds of
Houstonia are crateriform and the hilum is on a
hilar ridge (Terrell, 1996), whereas seeds of Hedy¬
otis serpyllacea are non-crateriform (without ventral
depressions or cavities) and lack hilar ridges. Old-
enlandia seeds are trigonous or conical and usually
much smaller than those of the other genera. Hed¬
yotis as presently recognized includes a heteroge¬
neous array of species, as previously pointed out
(Terrell, 1996). I have examined seeds of all of the
Western Hemisphere species of Hedyotis, as well as
those of many of the Asian species including the
type, H. fruticosa L., and all of these seeds differ
from those of Hedyotis serpyllacea.

A systematic treatment of Arcytophyllum serpyl¬
laceum (Hedyotis serpyllacea ) is presented here.
This extends the distribution of Arcytophyllum from
Panama and Costa Rica into Guatemala and south¬
ern Mexico, and also records the first collection of
H. serpyllacea in Oaxaca (first noticed by Robert
King, cited as Terrell & King 4441 ).

Arcytophyllum serpyllaceum (Schlechtendahl)
Terrell, comb. nov. Basionym: Hedyotis serpyl¬
lacea Schlechtendahl, Linnaea 9: 599. 1834.
Houstonia serpyllacea (Schlechtendahl) C. L.
Smith ex Greenman, Proc. Airier. Acad. Arts
32: 284. 1897. TYPE: Mexico. Veracruz: be¬
tween La Joya and San Salvador, June 1829,
Schiede 265 [Deppe’s name was not included
with Schiede’s] (holotype, HAL; isotypes, F,
HAL, MO; photos taken at B are at F, GH,
MO, NY, US; however, the B specimen was
destroyed during World War II).

Mallostoma shannonii Donnell Smith, Bot. Gaz. (Craw-
fordsville) 18: 203. 1893. Arcytophyllum shannonii
(Donnell Smilli) Standley, Contr. L'.S. Natl. Herb. 18:
128. 1916. TYPE: Guatemala. Chimaltenango:

Novon 9: 263-264. 1999.

264

Novon

“forming a carpet-like turf," alt. 9000 ft., Chichoy
[or Chicoy], Mar. 1892, W. C. Shannon s.n. (holotype,
US-48603; isotype, US-943475).

Perennial herb with woody rhizomes. Stems pros¬
trate, creeping, often matted, rooted at nodes, slender,
angulate, often woody at base, glabrous or pubescent
at nodes, outer layers of stem becoming loose and
broken, stems often appearing jointed. Stipules to 2
mm long, ovate, with short to long, narrow caudae,
these glabrous, pubescent, or ciliolate, 1— few marginal
teeth with or without reddish, stalked apical glands.
Leaves sessile or subsessile, 1-nerved, paler beneath,
thickened, somewhat leathery, 2-9 mm long, 1-5 mm
wide, ovate or elliptic, obtuse or acutish at apex,
rounded at base, glabrous, margins revolute, some¬
times scabrous. Flowers axillary, solitary, heterosty-
lous, pedicels to 7 mm long, slender to stout, some¬
times reflexed in fruiting stage. Hypanthium glabrous,
calyx lobes 4, 1. 0-3.8 mm long, 0.5-0.9(-1.5) mm
wide, lanceolate, ovate, oblaneeolate, or obovate, ob¬
tuse or acute at apex, sometimes with linear or awn¬
like intercalycine teeth to ca. 1 mm long. Corollas
6.0-10.0 mm long, funnelform, thickish, white within,
typically purple or reddish purple externally on lobes,
tube greenish externally; buds white or reddish pur¬
ple; tube 3—5 mm long, 2—3 mm wide at throat, gla¬
brous externally, pubescent distally within; lobes 4,
2.5— 4.8 mm long, 1.2— 1.7 mm wide, ovate or elliptic,
thickish, densely white-pubescent within (hairs to ca.
1 mm long). Pin flowers with styles slender, whitish,
stigma branches ca. 1-2.3 mm long, oblong or linear,
exserted to ca. 2 mm beyond corolla throat, anthers
1-2 mm long, whitish, oblong, included in distal V2—
% of corolla tube, on short filaments. Thrum flowers
with anthers 0.9— 1.8 mm long, subsessile, narrowly
oblong, slightly exserted on short filaments, stigma
branches ca. 1. 0-1.4 mm long, whitish, broadly ob¬
long, included in distal V2 of corolla tube, on short
styles. Capsules 1. 5-4.0 mm long, 2.0-3. 5 mm wide,
usually somewhat longer than wide, % to fully inferior,
broadly oblong or obovate, with several raised nerves,
thick-walled, apparently tardily dehiscent or indehis-
cent, apex rounded and with shallow central depres¬
sion. Seeds 12—29 per capsule, 0.6-1 .2 mm diam.,
black, strongly compressed, lenticular, in outline or¬
bicular, orbicular-polygonal, or broadly elliptic, dorsal
and ventral faces convex, flat, or obscurely ridged,
ventral face with hilum punetiform, ± centric, on flat
or minutely depressed surface, testa finely reticulate.
Flowering all year. Chromosome number not known.

Distribution. Guatemala; Mexico (Veracruz, Oa¬
xaca, Chiapas). Grassy places, meadows, pastures,
roadsides, open disturbed areas, stream banks, thin
forest, alt. ca. 1500-3500 m.

Additional specimens examined. GUATEMALA. Chi-
maltenango: Chichavec, Skutch 124 (US); Barranco de
la Sierra, SE of Patzum, Standley 6 1552 (MICH); near
Chocoyos, Williams 13142 (GH, MICH); Cerro Chichoy,
near Chichoy, Williams & Molina R. 15339 (GH). Gua¬
temala: San Jose Pinula. Salas 312 (US), lluehueteu-
ango: Sierra de los Cuchumatanes, km 311 on Ibita Nac.
9 N between Paquix and Chemal, Beaman 2983 (GH.
TEX, US); Chemal, Sierra Cuchumatanes, Molina R. &
Molina 26413 (ENCB. MICH); Ruta 9N ca. 10 mi. N of
Chiantla, Sanders 74109 (MICH). Quezaltenango: Cues-
ta de El Caracol, 5-8 km N of San Juan Ostuncalco, Wil¬
liams et al. 22771 (NY, US). Quiche: San Miguel Uspan-
tan, Apr. 1892, Heyde & Lux 3176 (GH. US).
Sacatepequez: loc. unknown, Standley 60766 (NY). San
Marcos: Volcan Tacana, E side of mountain at La Ha-
ciendita, Beaman 3197 (GH, US). Solnla: hwy. CA-1. 2.5
mi. E of intersection of hwys. 1 and CA-1, Denton 1794
(MICH). Totonicapan: Tecum Uman Ridge, 20 km E of
Totonicapan, Beaman 4167 (GH, TEX, US). MEXICO.
Chiapas: San Cristobal de las Casas, Alexander 1063
(MICH. NY), Breedlove 6044 (DS, ENCB, F, MICH),
Laughlin 9 (DS, EE, MSC), Paray 323 (ENCB), near same
locality. Nee & Mori 3487 (WIS); between San Cristdbal
de las Casas and Tenejapa Center, Breedlove 9239 (DS);
paraje of Yal Ichin, mpio. Chamula, Breedlove 9532 (DS,
F); Cerro Huitepec W of San Cristobal de las Casas,
Breedlove 25299 (DS); 3 mi. NW of San Cristobal de las
Casas, Carlson 1579 (F); loc. unknown, Ghiesbrecht 814
(GH, K, MO. NY); San Cristobal to Buenavista, Miranda
4990 (MEXU); Cerro del Boqueron, Purpus 7171 (BM. F.
GH, MO, NY, US). Oaxaca: El Tejocote, 52 km NW of
Oaxaca City, Terrell & King 4441 (US). Veracruz: Ea
Zimiento, Cofre de Perote, Balls 4629 (US); Ingenio El
Rosario, Mpio. Xico, Narave F. et al. 293 (XAE); road to
Las Minas, 3 km N of Cruz Blanca jet. with Hwy. 140,
Nee et al. 26053 (XAE). Terrell & Koch 5381 (CHAPA,
US); Las Vigas pedregal near La Joya, Sharp 45572
(MEXU, TENN, US); Jalapa. Smith 1486 (MO); just W of
La Joya. 10 km E of Las Vigas. Terrell & King 4455 (US),
Terrell & Koch 5380 (CHAPA, US); Zoatzingo, mpio. Al-
totonga, 2 km S of turn-off to Altotonga, Terrell & Koch
5384 (CHAPA, US); San Miguel El Soldado, mpio. San
Miguel. Ventura A. 648 (US).

Acknowledgments. I am grateful for field assis¬
tance by Robert King and Stephen Koch during two
collecting trips to Mexico. Curators of the herbaria
cited, especially HAL, were very helpful. Dan Nic-
olson provided nomenclatural information. Reviews
by David Lorence, Patricio Mena, and Charlotte
Taylor contributed important suggestions. I thank
James Reveal, the Department of Cell Biology and
Molecular Genetics, University of Maryland, and
the Maryland Agricultural Experiment Station for
support in publication.

Literature Cited

Mena V., P. 1990. A revision of the genus Arcytophyllum

(Rubiaceae: Hedyotideae). Mem. New York Rot. Card.

60: 1-26.

Terrell, E. E. 1996. Revision ol Houstonia (Rubiaceae—

Hedyotideae). Syst. Bot. Monogr. 48: 1-118.

Nomenclator and Review of Phyllostegia (Lamiaceae)

Warren L. Wagner

Department of Botany, MRC 166, National Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560, U.S.A. wagner.warren@nmnh.si.edu

Abstract. In 1987 Harold St. John published 93
names in Hawaiian Phyllostegia and 1 in Tonga;
only 3 ol them were accounted for in the most re¬
cent taxonomic revision of the Hawaiian species by
Wagner et al. in 1990. The types of these names
are analyzed, and the names are here incorporated
into the current classification. Only five of the new
Hawaiian species are here recognized as distinct
(P. kaalaensis, P. kahiliensis, P micrantha, P. pi-
losa, and P warshaueri), and one of the new com¬
binations is accepted here (P. velutina ), with the
remainder here placed into the synonymy of other
species. Five of the names were not validly pub¬
lished. The complete nomenclature of Phyllostegia
at the species level and below is presented, includ¬
ing 181 names. A total of 32 Hawaiian Phyllostegia
species are currently recognized, including the 5
recognized St. John species published in 1987 and
2 others, P haliakalae, resurrected from the syn¬
onymy of P. mollis, and P. renovans, discovered in
1989. The five Hawaiian species not treated in the
1990 revision are briefly characterized and dis¬
cussed. Notes adding information on taxonomy, dis¬
tribution, and population and conservation status
not included in the recent treatment also are given.
The genus also is known from Tahiti and Tonga,
each with one species; the nomenclature for these
two species is included, but they are not evaluated
in this paper.

In late 1987 and early 1988, as the final man¬
uscript for the Manual of the Flowering Plants of
Ilawai‘i (Wagner et ah, 1990) was about to be sent
to the publisher, St. John published an unfortunate,
hurried series of papers, mostly cited in Wagner et
al. (1990), in which he published about 880 man¬
uscript names. Over a period of more than 30 years,
he had developed lengthy manuscripts with accom¬
panying detailed illustrations for a number of gen¬
era in a style similar to his treatment of 0‘ahu Cyr-
tandra (St. John, 1966), but few of them had been
published. St. John published the new taxa and
combinations from all of these manuscripts in Phy-
tologia and in a series of privately published pa¬
pers, providing little more than a brief Latin diag¬
nosis and an abbreviated (and sometimes partly

erroneous) mention of the type. This has had seri¬
ous consequences: errors in citation of specimens,
the use of a name more than once in the same
genus (often even in the same paper), and the use
of the same specimen as the type of more than one
name. More importantly, he did not incorporate the
taxa into a carefully evaluated taxonomic system.
Now, after nearly 10 years, most of the approxi¬
mately 880 names (including combinations) have
been analyzed.

Th is is the first of a series of papers analyzing
the names published in 1987 and 1988 by St.
John, in which the nomenclature is evaluated and
the taxa are incorporated into current classifica¬
tions. The second paper will be on names in the
Cucurbitaceae (Wagner & Shannon, 1999). Addi¬
tional papers in this series will deal with names
in Stenogyne (Lamiaceae), Cyrtandra (Gesneri-
aceae), and the special problems of the not valid
and illegitimate names. St. John names in Cam-
panulaceae published in 1987 and 1988 have
been studied by T. Lammers (see Lammers, 1998,
and papers cited therein).

Although all names for Hawaiian angiosperms
published through 1986 and a few from 1987 and
1988 were included by Wagner et al. (1990) in the
Manual of the Flowering Plants of Hawai‘i, a rig¬
orous assessment of them has never been done. The
nomenclatural database, initiated in 1983 as part
of the Hawaiian Flora Project at Bishop Museum,
has been expanded to include all of these addition¬
al names by St. John. It will be available on the
WWW in the future.

In analyzing the nomenclature for the genus
Phyllostegia (Lamiaceae) a number of problems
were discovered. These included names not validly
published, errors in original publications, and typ-
ification problems. Only 3 of the 93 Hawaiian
names published by St. John (1987) were incor¬
porated into the taxonomy of the genus as treated
in the most recent revision (Wagner et al., 1990).
Presented here is a complete nomenclator of 181
names in Phyllostegia, including a number of not
validly published names. Since the emphasis of this
paper is on nomenclature rather than classification

Novon 9: 265-279. 1999.

266

Novon

or phylogeny of the group, the accepted names are
arranged alphabetically. Of the 93 Hawaiian taxa
(78 new species, 3 new varieties, and 12 new com¬
binations) described by St. John, 5 of the new spe¬
cies are here recognized as distinct ( P. kaalaensis,
P. kahiliensis, P. micrantha, P. pilosa, and P. war-
shaueri), with P. warshaueri being adopted in the
1990 revision by Wagner et al. In addition, one of
the new combinations is accepted here (P. velutina),
which was also adopted in the Wagner et al. revi¬
sion; the remainder are here placed into the syn¬
onymy of other species. Five of the names were not
validly published.

The complete nomenclature of Phyllostegia at
the species level and below is presented. A total of
32 Hawaiian Phyllostegia species are currently rec¬
ognized, including the 5 recognized St. John spe¬
cies published in 1987 and 2 others, P. haliakalae,
resurrected from the synonymy of P. mollis, and P
renovans, discovered in 1989. The five species (P.
kaalaensis, P. kahiliensis, P. micrantha, P. pilosa,
and P. renovans) not treated in the 1990 revision
are briefly characterized and discussed. Notes add¬
ing information on taxonomy, distribution, and con¬
servation status not included in the recent treat¬
ment also are given, especially for the 14 taxa listed
as endangered since 1987, and others I believe to
be at risk. The genus also occurs in Tahiti and Ton¬
ga, each with one species; the nomenclature for
these two species is included, but they are not eval¬
uated in this paper.

1. Phyllostegia amhigua (A. Gray) Hillebrand,
FI. Hawaiian Isl. 350. 1888. Phyllostegia brev-
idens A. Gray var.? ambigua A. Gray, Proc.
Ainer. Acad. Arts 5: 343. 1862. TYPE: Ha¬
waiian Islands [U.S.A.]. Maui: West Maui,
1840, U. S. Expl. Exped. s.n. (holotype, US-
57473; isotype, GH not seen).

Phyllostegia grandiflora (Gaudichaud) Bentham var. hir-
sutula Hillebrand, FI. Hawaiian Isl. 349. 1888.
Phyllostegia brevidens A. Gray var. hirsutula (Hille¬
brand) Sherff, Amer. J. Bot. 21: 699. 1934. TYPE:
Hawaiian Islands [U.S.A.]. Maui: West Maui, in for¬
ests, s.d., E. Bishop 76 (holotype, B presumably de¬
stroyed, photo E not seen). Maui: West Maui, mauka
of McGregor, rain forest, 9 Mar. 1952, 0. Degener.
Tam, Tousley A Barker 22024 (neotype, designated
by Sherff (1953: 18). E not seen; isoneotypes, BISH.
US). Sherff designated the Degener specimen as lec-
totype, but it must be taken as a neotype.
Phyllostegia brevidens A. (bay var. expansa Sherff, Bot.
Gaz. (Crawfordsville) 96: 139. 1934. TYPE: Hawai¬
ian Islands [U.S.A.]. Maui: West Maui. [Pu‘u Kukui|
Ml. Eke. 1851—1871. IT Hillebrand s.n. (holotype. B
presumably destroyed).

Phyllostegia brevidens A. Gray var. heterodoxa Sherff.
Amer. J. Bot. 21: 700. 1934. Phyllostegia heterodoxa

(Sherff) H. St. John, Phytologia 63: 175. 1987.
TYPE: Hawaiian Islands [U.S.A.]. Hawai'i: Ka'u,
Na'alehu, along 1 1 i lea Trail. 1(>(K)— 3(XM) ft.. 15 Jan.
1912. J. E Rock 10014 (holotype, B1SH-510654;
isotypes, E not seen, BISH [2]). Locality and date
from Rock's field notes.

Phyllostegia brevidens A. Gray var. pauciflora Sherff. Bot.
Gaz. (Crawfordsville) 96: 139. 1934. TYPE: Hawai¬
ian Islands [U.S.A.]. Maui: [East Maui], South Ha-
leakala. 1851—1871, W. Hillebrand s.n. (holotype, B
presumably destroyed).

Phyllostegia brevidens A. Gray var. pubescens Sherff. Bot.
Gaz. (Crawfordsville) 96: 138. 1934. TYPE: Hawai¬
ian Islands [U.S.A.]. Maui: |East Maui], Upper Kula,
s.d., ./. M. Lydgate s.n. (holotype, B presumably de¬
stroyed).

Phyllostegia brevidens A. Gray var. degeneri Sherff. Amer.
J. Bot. 28: 28. 1941.' TYPE: Hawaiian Islands
[U.S.A.]. Maui: |East Maui], Haleakala. NW side of
Ko'olau Gap, in fogswept jungle, 6000 ft., 20 Aug.
1939, 0. Degener. E. Ordonez A E C. Salueop 12464
(holotype, E not seen [mounted on 2 sheets]; iso¬
types, BISH. NY [microfiche]).

Phyllostegia degeneri Sherff. Brittonia 12: 172. 1960.
TYPE: Hawaiian Islands [U.S.A.]. Maui: |East
Maui], Haleakala, W side of Ko'olau Gap, in rich,
open forest. 28 May 1959, 0. Degener & W. Fleming
25235 (holotype, E not seen: isotypes, BISH [2|. E
not seen, NY [microfiche], US).

Phyllostegia longimontis II. St. John, Pacific Sci. 30: 29.
1976. TYPE: Hawaiian Islands [U.S.A. |. Hawai'i:
[Kona, 26— 29 Jan. 1779], I). Nelson s.n. (holotype,
BM -550453).

Phyllostegia alba II. St. John. Phytologia 63: 172. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Maui:
[West Maui, llonokahau Drainage Basin, summit of
Eeka, upper part of main ridge], 27 Sep. 1917, C.
N. Eorbes 363. M (holotype, BISH -53537; isotype,
BISH).

Phyllostegia chartacea H. St. John, Phytologia 63: 173.
1987. Syn. nov. TYPE: Hawaiian Islands | U.S.A . ].
Maui: East Maui, Waikamoi frail. May 1911. J. E.
Rock s.n. (holotype, BISH-53379).

Phyllostegia elliptiea H. St. John, Phytologia 63: 174.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: East Maui. Hana Forest Reserve, 5675 ft., 29
June 1973. B. Harrison 257 (holotype, BISH-
513265; isotype, BISH).

Phyllostegia hiati II. St. John. Phytologia 63: 175. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Maui:
East Maui. Ko’olau Gap, Ainahou, 6000 ft., 16 Feb.
1980, R. Hobdy 762 (holotype, B1SH-522877). R.
Hobdy (pers. comm.) indicated that the collection
number is 769, not 762 as indicated in protologue.
Phyllostegia konaensis H. St. John, Phytologia 63: 176.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Hanehane, Kona, a few miles SW of
Pu'uwa'awa'a. 16 June 1911, C. N. Eorbes 183a.H
(holotype, BISH-53371). St. John used a portion of
Eorbes I83.H, and added “a” to the collection num¬
ber; the other portion of this collection was pub¬
lished as the holotype of P decumbens and Stenogyne
adpressa. Neither of these names is validly published
under 1CBN Art. 34.2 (Greuter et al.. 1994).
Phyllostegia lebishopii H. St. John, Phytologia 63: 177.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: East Maui. Haleakala, Pu'u o Kakae, 47(X) ft..

Volume 9, Number 2
1999

Wagner

Phyllostegia Nomenclator

267

15 Apr. 1971. L. E. Bishop. fW] Gagne & S. L. Mont¬
gomery 0471 1 13 (holotype. BISH-573150).
Phyllostegia odorata H. St. John. Phytologia 63: 179.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.|.
Maui: West Maui. Kapilau Ridge, 28 Feb. 1980. B.
llobdy 775 (holotype. B IS H -522878).

Phyllostegia orientalis H. St. John. Phytologia 63: 179.
1987. Syn. nov. TYPE: Hawaiian Islands (U.S.A.].
Maui: East Maui, gulch next to base camp ridge,
USFWS Forest Bird Survey. 6300 ft.. 18 May 1985,
F R. Warshauer 2535 (holotype, BISH-51 1688).
Phyllostegia pilosieincta II. St. John. Phytologia 63: 179.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A. |.
Maui: East Maui, Haleakala, Upper Waikamoi, 14
Jan. 1976. E. Smith & J. Doris s.n. (holotype, BISH-
522875: isotype, BISH).

Phyllostegia pluriflora H. St. John, Phytologia 63: 180.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A. |.
Hawai’i: /. M. Tomich 4005 (holotype, BISH not
seen). The holotype has been missing since at least
October 1987. Tentative disposition, based on the
description.

Phyllostegia repanda H. St. John. Phytologia 63: 180.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai’i: South Kohala Dist., Kohala Mountains, SW
ol PuTi Poho'ula'ula. wet ‘ohi'a forest, USFWS Forest
Bird Survey Transect 80 (33/34). 4160 It.. 31 July
1979, L. Stemmermann & S. L. Montgomery 3976
(holotype, BISH-469848).

Phyllostegia suaveolens H. St. John, Phytologia 63: 181.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai’i: Kilauea Forest Reserve, near fence line ol
Meyer’s Ranch, 4300 ft., 15 Apr. 1980, J. Davis 276
(holotype, BISH-520029).

The range ol Phyllostegia ambigua is expanded
on the island ol Hawai‘i somewhat with the addition
of the types placed here to include Kilauea Vol¬
cano, the Kohala Mountains, and at least until 1911
Hualalai in addition to Mauna Loa. It also occurs
on East and West Maui.

2. Phyllostegia bracteata Sherff, Bot. Gaz.
(Crawfords ville) 96: 137. 1934. TYPE: Ha¬
waiian Islands [U.S.A.]. Maui: West Maui, Pu’u
Kukui, [upper] forest, 4000—5000 ft., 24—26
Sep. 1916. A. S. Hitchcock 14746 (holotype,
US-873135).

Phyllostegia bracteata historically occurred at a
number ol widely separated localities on East and
West Maui, but is now quite rare with apparently
only a lew populations still extant. It has only been
collected a lew times this century, and was most
recently collected in 1992 (Waikamoi, East Maui,
Wood et al. 1623, NY, PTBG, US). According to J.
Lau (pers. comm. 1998) there are several extant
populations, including one along the Waikamoi
Flume road, and another on the Waikamoi Flume,
Kipahulu Valley and the adjacent Manawainui. He
estimated there are most likely a number ol unre¬
ported populations since there is a lot ol wel lorest

habitat on East and West Maui, and it is in rela¬
tively good shape.

3. Phyllostegia brevidens A. Gray, Proc. Amer.

Acad. Arts 5: 343. 1862. TYPE: Hawaiian Is¬
lands [U.S.A.]. Hawai‘i: [in the forest] on
Mouna [Mauna] Kea, 3000 ft., 1840, U. S.
Expl. Exped. s.n. (holotype, US-57474; isotype,
GH not seen).

Phyllostegia polyantha H. Si. John, Phytologia 63: 180.
1987. Syn. nov. TYPE: Hawaiian Islands (U.S.A.].
Hawai'i: Kulani project land S of Mauna I.oa Boys’
School, moist gully, open to scattered Metrosideros
forest with open Cibotium understory, 5700 ft., 26
Feb. 1981. G. Clarke 574 (holotype, BISH-520031).

Phyllostegia brevidens was considered to be ex¬
tinct, because it was known from only two collec¬
tions made in the mid to late 1800s (Wagner et al.,
1990). It has been recently discovered on East
Maui in Kipahulu Valley, and rediscovered on Ha¬
wai’i (type of P. polyantha). It is apparently known
from only a few populations and probably should
be considered endangered. These collections, at
least one of which I earlier determined as P. am¬
bigua, did not completely fit the description of P
brevidens provided by Wagner et al. (1990). They
appear to be better accommodated within a some¬
what emended P. brevidens than in P. ambigua,
where they did not fit the description provided by
Wagner et al. (1990) in the more numerous flowers
per verticillaster (14—30), glabrate in the upper part
and densely hirsute on lower stems, smaller calyx
4—6 mm long, and the nearly obsolete calyx lobes
(or on Maui with lobes up to 1.8 mm long).

Specimens examined. HAWAIIAN ISLANDS | U.S.A.].
East Maui: Kipahulu Valley, lower clog leg, 730 m, 9 Mar.
1988. Wagner et al. 5897 (BISH. PTBG, US); Kipahulu
Valley, Palikea drainage, 2920-3050 ft.. 20 May 1994,
Wood et al. 3200 (Al), BISH, MO, NY, PTBG, US). Ha¬
waii: [North Hilo District], Ilakalau National Refuge,
Piha. KM) m in from Maulua middle fence, 5450 ft., 8
Oct. 1991. Perlman & Jeffery 12293 (PTBG); | South Hilo
District] Hilo, Hillebrand & Lydgate s.n. (BISH), llille-
brand s.n. (BISH); McCandless Ranch. Waiea Section,
rain forest. . . . growing from the floor of a collapsed lava
tube . . . [which] prevented cattle from reaching subject
plant. 16 Apr. 1985. Giffin s.n. (BISH).

4. Phyllostegia electra C. N. Forbes, Occas.

Pap. Bernice P. Bishop Mus. 6(3): 180. 1916.
TYPE: Hawaiian Islands [U.S.A.]. Kaua’i:
power line trail between Kapaka and Kualapa,
6—9 Aug. 1909, C. N. Forbes 143. K (holotype,
BISH-499098; isotypes, BISH [2], MO not
seen).

Phyllostegia mollis Bentham var. skottsbergii Sherff, Field

268

Novon

Mus. Nat. Hist.. Hot. Ser. 17: 577. 1939. TYPE: Ha¬
waiian Islands [U.S.A.]. Kaua'i: Koke'e, in dense
woods. 3600 ft., 8 Mar. 1926. G. P Wilder 451 (ho-
lotype, F-954570 [photo F not seen]).

Phyllostegia mollis Bentham var. resinosa Fosberg, Oceas.
Pap. Bernice P. Bishop Mus. 16(15): 345. 1942.
TYPE: Hawaiian Islands [U.S.A.]. Kaua'i: plateau at
head of Kalalau Valley, in wet forest, 375 m, 29 Dec.
1935. F. H. Fosberg 12696 (holotype, BISH not
seen). The holotype has been missing at least since
August 1997.

Phyllostegia mollis Bentham var .fagerlindii Sherff, Occas.
Pap. Bernice P. Bishop Mus. 20(1): 13. 1949. Phyl¬
lostegia fagerlindii (Sherff) H. St. John, Phytologia
63: 174. 1987. TYPE: Hawaiian Islands [U.S.A.].
Kaua'i: northernmost Kaua'i, Hanalei Valley, above
second house, 9 Mar. 1948, F Fagerlind & C.
Skottsberg 6533 (holotype, S | photo F not seen]).
Published as in GB; however, the photograph in the
original publication labeled “type" is of a specimen
at S.

Phyllostegia mollis Bentham var. hochreutineri Sherff. Bot.
Leafl. 8: 18. 1953. TYPE: Hawaiian Islands [U.S.A.].
Kaua'i: Waimea. near the Gay cottage [at Kaholu-
manu], 1400 m, 20 Apr. 1905, 11. P. G. Hochreutiner
3579 (holotype, G [photo F not seen]). Sherff erro¬
neously cited the island as Hawai'i.

Phyllostegia adenophora H. St. John, Phytologia 63: 172.
1987. Syn. nov. TYPE: Hawaiian Islands | U.S.A. |.
Kaua'i: Kau'umohua Bidge. 27 Jan. 1920,7. F. Rock-
17313 (holotype. BISH-53740).

Phyllostegia eapitata H. St. John, Phytologia 63: 173.
1987. Syn. nov. TY PE: Hawaiian Islands [U.S.A.].
Kaua'i: Alakai Swamp Trail, S of Lehuamakanoi,
3900 ft., 17 Apr. 1981, J. H. R. Plews 80 (holotype,
BISH-522872).

Phyllostegia fayi H. St. John, Phytologia 63: 174. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A. |. Kaua'i:
near crest of ridge separating Hali'i Stream and
Kamo'oloa Stream, 450 m, 28 June 1976, J. Fay 555
(holotype, PI'BG-15148; isotype, BISH). Although
not indicated in the publication, the holotype be¬
longs to PTBG; it was inadvertently mounted by
BISH while on loan.

Phyllostegia electra has been collected at many
localities over the past decade. It is not common
anywhere, but has a wide, scattered distribution in
mesie and wet forest over much of Kaua'i. It is
probably the most commonly collected species of
the genus on Kaua'i. It has a partly overlapping
distribution with P. helleri , with which it is some¬
times confused, but differs in its short-hispidulous
pubescence, the trichomes 0.1-0.3(-0.5) mm long,
and usually some of the trichomes are gland-tipped.
It is probably most closely related to P. parviflora.

5. Phyllostegia floribunda Bentham, Labiat.
Gen. Spec. 653. 1835. TYPE: Hawaiian Is¬
lands [U.S.A.]. Hawai'i: [probably 26—29 Jan.
1779], D. Nelson s.n. (holotype, BM-550454
[photo BISH not seen]; isotype, BM).

Phyllostegia floribunda Bentham var. forbesii Sherff, Amer.

J. Bot. 21: 699. 1934. Phyllostegia forbesii (Sherff)
H. St. John, Pacific Sci. 30: 25. 1976. TYPE: Ha¬
waiian Islands [U.S.A. |. Hawai'i: Kona,
Ka'alapu'uwale, 30 June 1911. C. N. Forbes 294. H
(holotype. MO-821348; isotypes, BISH [3]).
Phyllostegia villosa H. St. John, Pacific Sci. 30: 31. 1976
1 1977], TYPE: Hawaiian Islands |U.S.A.| Hawai'i:
upper edge of Waipio Valley, along trail from back
of Waimea, 4000 ft., June 1910, J. F. Rock 8352
(holotype, BISH-5 10103).

Phyllostegia axillaris H. St. John, Phytologia 63: 172.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Hilo Forest Reserve, Laupahoehoe section,
between Ka'awali'i and Kaiwilahilahi Streams, 3420
ft.. 2 Feb. 1982, S. Anderson 488 (holotype, BISH-
518753).

Phyllostegia floribunda is considered to be rare,
known from a few scattered sites; it was most re¬
cently collected in 1982 and again in 1991 behind
the Volcano Village dump ( Flynn 761, PTBG, and
Wood et al. 505, PTBG not seen), and in 1982 be¬
tween Ka'awali'i and Kaiwilahilahi Streams (type of
P. axillaris ). J. Lau (pers. comm. 1998) indicated
that extant populations are known from Puna Dis¬
trict (couple of sites in Pu'u Maka'ala; Kahaualea;
couple of places on the East Rift Zone; and Volcano
Dump), North Hilo District (several locations in
Laupahoehoe), and South Kona District (Kukuio-
pae).

This red-flowered species, along with P. knud-
senii and P wawrana, was placed in Phyllostegia
sect. Lateriflorae A. Gray by Sherff (1935). The
same delimitation was followed by Wagner et al.
(1990). Based on the ample material of the white-
flowered species now at hand I exclude them from
the section, because they have both axillary and
terminal inflorescences, which was the only char¬
acter to distinguish the section.

6a. Phyllostegia glabra (Gaudichaud) Bentham
var. glabra

Phyllostegia glabra (Gaudichaud) Bentham, Edwards’s
Bot. Reg. 15: 1292. 1830. [Epithet published as
“ glabrum .”] Prasium glabrum Gaudichaud, Voy. Ur¬
anic 452. 1829. TYPE: s.l., 550-735 m. 1819, C.
Gaudichaud- Beauprd s.n, (holotype, P not seen).
Phyllostegia chamissonis Bentham, Linnaea 6: 79. 1831.
TYPE: Hawaiian Islands [U.S.A.]. O'ahu: 1816-
1817, L. C. A. von Chamisso s.n. (holotype, LE not
seen).

Phyllostegia macraei Bentham, Prodr. 12: 554. 1848.
Phyllostegia glabra (Gaudichaud) Bentham var. ma¬
craei (Bentham) Sherff, Bot. Gaz. (Crawfordsville) 96:
136. 1934. TYPE: Hawaiian Islands [U.S.A.]. O'ahu:
9-26 May 1825. J. Macrae s.n. (lectotype,
designated by Sherff (1935: 31), K). Additional syn-
types, O'ahu: Beechey s.n,, Nelson s.n.

Phyllostegia glabra (Gaudichaud) Bentham f. fauriei 0.
Degener & 1. Degener, FI. Hawaiiensis, fam. 316.
Phyllostegia glabra. Part 2. I960. TYPE: Hawaiian

Volume 9, Number 2
1999

Wagner

Phyllostegia Nomenclator

269

Islands [U.S.A.]. kaua'i: Waimea, 1000 m. 1910, U.
Fnurie B96 (holotype, BISH-510030). The label may
not be correct as to island because this species is
otherwise unknown from kaua'i. Although the De-
geners did not indicate where the type was located,
the leaf description closely fits the I1ISH collection.
Part 2 was published on 8 Aug. 1960.

Phyllostegia rubritincta H. St. John, Phytologia 63: 181.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Pana’i: extreme head, Maunalei, on S slope of Lan-
ailiale, 10 Mar. 1915, G. C. Munro s.n. (holotype,
BISH -53448).

6I>. Phyllostegia glabra (Gaudichaud) Bentham
var. lanaiensis Sherff, Bot. Gaz. (Crawfords-
ville) 96: 136. 1934. Phyllostegia lanaiensis
(Sherff) H. St. John, Phytologia 63: 177. 1987.
TYPE: Hawaiian Islands [U.S.A.]. Lana’i: side
of mountain, 1864—1865, //. Mann & U. T.
Brigham 354 (holotype, F not seen; isotype,
BISH).

Phyllostegia glabra var. lanaiensis is listed as
endangered (Herhst, 1990, 1991). It is historically
known from Eana’ihale, but was definitely known
only from Kaiholena Gulch (Herhst, 1991), and was
last observed in the 1980s by R. Hobdy in a gulch
feeding into the back of Maunalei Valley, another
drainage from Lana’ihale (Herhst, 1990, 1991);
however, this observation appears to have actually
been of P glabra var. glabra (J. Lau. pers. comm.
1998).

7. Phyllostegia grandiflora (Gaudichaud) Ben¬

tham, Edwards’s Bot. Reg. 15: 1292. 1830.
Prasium grandiflorum Gaudichaud, Voy. Ur-
anie 453. 1829. TYPE: s.L, 530-735 m. 1819,
C. Gaudichaud- Be aupre s.n. (holotype, P not
seen).

Phyllostegia grandiflora (Gaudichaud) Bentham var. ma-
joricalycis H. St. John, Phytologia 63: 175. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. O'ahu:
s.l.. May 1825. J. Macrae s.n. (holotype, GH-1421).

8. Phyllostegia haliakalae Wawra, Flora 55:

530. 1872. TYPE: Hawaiian Islands [U.S.A.].
Maui: East Maui, foot of Haleakala, in dry ra¬
vine, 1869—1870, //. Wawra IB74 (holotype,
W not seen; isotype, W).

Phyllostegia imminutci (Sherff) H. St. John. Phytologia 63:
175. 1987. Syn. nov. Phyllostegia helleri Sherff var.
imminuta Sherff. Amer. J. Bot. 21: 699. 1934.
TYPE: Hawaiian Islands [U.S.A. |. Pana'i: kaiholena,
17 Mar. 1916, Mrs. C. Forbes s.n. (holotype, BISH-
510031).

Phyllostegia mollis Bentham var. micrantha Sherff, Amer.
J. Bot. 21: 700. 1934. TYPE: Hawaiian Islands
[U.S.A.]. Pana'i: Maunalei Valley, 25 June 1913. C.
A. Forbes 170. L (holotype, BISH-510099).

Phyllostegia deltoidea H. St. John, Phytologia 63: 174.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.|.
Moloka'i: gulch of Maunahui, 21 Mar. 1910. ./. F.
Hock 6127 (holotype, BISH-53705; isotype, GH).
Phyllostegia makawaoensis H. St. John, Phytologia 63:
177. 1987. Syn. nov. TYPE: Hawaiian Islands
[U.S.A.]. Maui: East Maui. Makawao, ,/. M. Lydgate
s.n. (holotype, BISH-53689). The type is possibly a
mixed collection; a densely villous sterile stem rep¬
resents either another species or a juvenile shoot.
The label on this collection is one made by Forbes
for Hillebrand and Pydgate material, which gives
both as collectors and was used on many collections
made exclusively by Pydgate.

Phyllostegia triquerta H. St. John. Phytologia 63: 182.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: East Maui. Haleakala, Ulapalakua, 7000 ft..
1864—1865, //. Mann & IP 7. Brigham 404 (holo¬
type. B1S1I -53687; isotype. GH).

This species was treated as Phyllostegia immi¬
nuta by Wagner et al. (1990), but examination of
the type of P. haliakalae showed it to be the same
taxon. It was historically known from Lana’i (last
collected in 1928 near Maunalei, Munro 247,
BISH), East Maui (Wagner et al., 1990), and from
Moloka’i ( Degener 5395 in 1928, BISH); it is ap¬
parently extinct as it has not been collected in 70
years.

9. Phyllostegia helleri Sherff. Bot. Gaz. (Craw-

fordsville) 96: 138. 1934. TYPE: Hawaiian Is¬
lands [U.S.A.]. Kaua’i: above Waimea, Kahol-
uamanoa [Kaholuamanu], 12 Oct. 1895, A. A.
Heller 2875 (holotype, US-368712; isotypes, F
not seen, GH not seen, MO, NY microfiche
seen). The date on the holotype and the GH
and NY sheets is 11—16 Oct., but both Heller
and Sherff cited 12 Oct. for this collection.

Phyllostegia olokeleensis H. St. John. Phytologia 63: 179.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
kaua'i: Olokele, J. M. Lydgate 15 (holotype, BISH-
53552).

Phyllostegia helleri was listed from only Koke’e
Plateau and Mount Kahili by Wagner et al. (1990).
The Mount Kahili collections have been excluded
here as P. kahiliensis. Phyllostegia helleri appears
to be closely related to P. wawrana, now that more
material is available of the latter species, but differs
in its shorter hirtellous pubescence, the trichomes
0.5—1 mm long, strictly terminal inflorescences,
shorter pedicels ( 1 — )3 — 5 mm long, and calyx veins
relatively inconspicuous. It is known from the
Koke’e Plateau and appears to be extinct, as it has
not been collected since 1916 (Hitchcock 15367,
US; Oct. 1916, Rock s.n., BISH).

10. Phyllostegia liillebrandii H. Mann ex Hil-

270

Novon

lebrand, FI. Hawaiian Isl. 350. 1888. TYPE:
Hawaiian Islands fU.S.A.]. Maui: [East Maui],
woods of Kula, July 1858, W. Hillebrand s.n.
(lectotype, designated by Sherff (1936: 30), B
[presumably destroyed, fragment BISH-
499095]; isolectotype, BISH). Additional syn-
type: woods of Ulupalakua, W. Hillebrand s.n.
(B presumably destroyed). Another sheet at
BISH states “Ulapalakua and Kula, East
Maui.” US has one sheet collected by Hille¬
brand without a specific locality.

Phyllostegia hillebrandii is known only from the
two collections cited above and is almost certainly
extinct (Wagner et al., 1990).

1 1 . Phyllostegia hirsuta Bentham, Edwards’s
Bot. Reg. 15: 1292. 1830. TYPE: Hawaiian
Islands [U.S.A.]. O’ahu: 9-26 May 1825, J.
Macrae s.n. (holotype, K). Published as “P.?
hirsuta."

Phyllostegia hirsuta Bentham var. laxior 0. Degener &
Sherff, in Degener, FI. Hawaiiensis, fam. 316. Phyl¬
lostegia hirsuta. 1934. Phyllostegia laxior (0. De¬
gener & Sherff) H. St. John, Phytologia 63: 177.
1987. TYPE: Hawaiian Islands [U.S.A.]. 0‘ahu:
summit of Pig-god Trail, Hau’ula, 31 May 1931, 0.
Degener. K. K. Parks. D. L. Topping & W. Bush 5363
(holotype, F not seen; isotype, BISH).

Phyllostegia ternaria H. St. John, Phytologia 63: 182.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.).
O’ahu: Wai’anae [Mountains], Kalua‘a-Lualualei
Ridge, Kalua’a, on N fare, well-shaded open but
steep terrain, 808 m, 7 Mar. 1982, J. Obata, P.
Kores, M. Toutv A K. Simens 441 (holotype, BISH-
520636; isotypes, BISH [3]).

Phyllostegia hirsuta, although historically known
from the central Wai’anae Mountains (from Ku-
kui’ula (Pahole) Gulch to North Palawai Gulch) and
nearly throughout the Ko’olau Mountains (from Pu-
pukea-Kahuku Trail to Palolo Valley), is now re¬
stricted to 10 populations in the Wai’anae Moun¬
tains in the southern part of its range from Makaha
and Wai’anae Kai to North Palawai Gulch, and from
only 6 populations in the Ko’olau Mountains from
Kawainui Gulch in Kawailoa Training Area to
South Kaukonahua drainage. It was listed as en¬
dangered in 1996 (Bruegmann, 1995c; Russell &
Bruegmann, 1996b). The total number of known
individuals is less than 200. J. Lau (pers. comm.
1998) gave a view seemingly at variance with this
assessment of less than 200 individuals. He said,
’’There are numerous populations in the Wai’anae
Mountains, with many plants. There are also many
populations in the Ko’olau Mountains, with many
plants, although it has not been seen in recent de¬
cades in the southern part of the mountain range.”

12. Phyllostegia hispida Hillebrand, FI. Hawai¬
ian Isl. 353. 1888. TYPE: Hawaiian Islands
[U.S.A.]. Moloka’i: heights of Mopulehu, July
1870, W. Hillebrand s.n. (holotype, B [presum¬
ably destroyed, fragment BISH-510033]; iso¬
types, MEL not seen [photo BISH], US [pos¬
sible]).

Phyllostegia hispida is known only from the
mountainous eastern portion of Moloka’i. It was last
collected in 1979 (Jacobi & Higashino 1561,
BISH). The current status of all populations is not
known, and J. Lau (pers. comm. 1998) considered
this species to be endangered. He reported that
only three solitary plants have been seen in recent
years, and one of them died about two years ago.
Both of the living plants are on the Pelekunu rim,
one in Kamakou Preserve, the other in Pu’u Ali’i
Natural Area Reserve. The dead plant was in Ka¬
makou Preserve in the Waikolu Drainage.

13. Phyllostegia kaalaensis H. St. John, Phyto¬
logia 63: 176. 1987. TYPE: Hawaiian Islands
[U.S.A.]. O’ahu: Wai’anae Mts., below Pu’u
Pane, Palikea Gulch, 45° slope, mesic summer
dry forest, 1900 ft., 24 Apr. 1977, J. Obata,
G. Carr & D. Palmer 77-315 (holotype, BISH-
522881; isotype, BISH).

Wagner et al. (1990) discussed the populations
here treated as Phyllostegia kaalaensis as possibly
distinct from P. glabra. Phyllostegia kaalaensis dif¬
fers from P. glabra in having leaves with dentate
margins and calyces and bracts sparsely ciliate. It
is listed as endangered (Bruegmann, 1995c; Rus¬
sell & Bruegmann, 1996b). It is known from only
five populations and about 50 individuals from
scattered areas in the Wai’anae Mountains (‘Eka-
hanui Gulch, Pahole Gulch, Palikea Gulch, and
Wai’anae Kai). In August and September 1998 J.
Lau (pers. comm. 1998) visited several populations
of P. kaalaensis at three locations (Wai’anae Kai,
Pahole Gulch, and ‘Ekahanui Gulch), with a total
of about 20 plants. The other populations (Palikea
Gulch, and additional drainages in Pahole and Ek¬
ahanui Gulches) have not been revisited in 10-20
years. Lau estimated the total number of plants still
alive may be under 100. The Wai’anae Kai popu¬
lation has gone way down in numbers since it was
first found seven years ago. The mesic habitat of
this species is very threatened.

14. Phyllostegia kahiliensis H. St. John, Phy¬
tologia 63: 176. 1987. TYPE: Hawaiian Is¬
lands [U.S.A.]. Kaua’i: steep, moist slope just
below the summit of Kahili Peak, to the south.

Volume 9, Number 2

1999

Wagner

Phyllostegia Nomenclator

271

860 m, 4 Jan. 1974, J. J. Fay 156 (holotype,
PTBG-5552; isotype, BISH).

The three collections of this species from Mount
Kahili at ea. 860 m were included under Phyllo¬
stegia helleri by Wagner et al. (1990), but they
pointed out that unlike all other populations of P.
helleri these three specimens (the type, Flynn et al.
2228 , PTBG, and Wagner et al. 521 7, BISH) have
retrorsely appressed pubescence and may represent
a distinct taxon. Phyllostegia helleri always has
spreading trichomes throughout its range. It seems
best to recognize the population with retrorsely ap¬
pressed pubescence as distinct. Similar pubes¬
cence differences are characteristic in a number of
other species in the genus, often making identifi¬
cations difficult. It also has pedicels only 1-2 mm
long, whereas P. helleri has pedicels usually
3— 5(— 9) mm long. Field or greenhouse study is
needed to determine whether there are other dif¬
ferences. Phyllostegia kahiliensis is apparently very
rare.

15, Phyllostegia knudsenii Hillebrand, FI. Ha¬
waiian Isl. 354. 1888. TYPE: Hawaiian Is¬
lands [U.S.A.]. Kaua‘i: woods of Waimea, s.d.,
V. Knudsen 190 (holotype, B [presumably de¬
stroyed, fragment BISH-49904]).

Phyllostegia knudsenii was known only from the
type collection from the woods of Waimea on Kaua'i
and was considered extinct by Wagner et al. (1990).
An individual of this species was located in the
Koai’e Canyon growing in diverse mesic forest and
was reported by Lorence et al. (1995). It was listed
as endangered based on this rediscovery (Brueg-
mann, 1995a; Russell & Bruegmann, 1996a). It is
currently known from two locations in Koai‘e Can¬
yon.

This species, along with P. wawrana and P. flo-
rihunda , was placed in Phyllostegia sect. Lateriflo-
rae A. Gray by Sherff (1935). The same delimita¬
tion was followed by Wagner et al. (1990). Based
on the ample material now at hand I exclude all
but the type, P. floribunda, from the section, be¬
cause the others have both axillary and terminal
inflorescences. With the exclusion of these two
white-flowered species the section can now also be
characterized by red corollas.

16, Phyllostegia lantanokles Sherff, Bot. Gaz.
(Crawfordsville) 96: 137. 1934. TYPE: Ha¬
waiian Islands [U.S.A.]. 0’ahu: Ka’ala Mts.,
1864-1865, H. Mann & W. T. Brigham s.n.
(holotype, CU [photo F not seen]).

Phyllostegia atomifera H. St. John, Phytologia 63: 172.

1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
O’ahu: Mt. Ka’ala. boggy area behind chopper land¬
ing pad in tangle of Rubus penetrans , May 1981. 8.
H. Gagne s.n. (holotype, BISH-5I2414).

Phyllostegia lantanoides Sherff var. konahuanuiana II. St.
John, Phytologia 63: 177. 1987. Syn. nov. TYPE:
Hawaiian Islands [U.S.A.]. O'ahn: Manoa-Palolo
ridge, 2450 ft.. 6 Jan. 1932, //. St. John 11506 (ho¬
lotype, BISH-404541; isotype, BISH).

Phyllostegia oahuensis H. St. John, Phytologia 63: 178.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
O’ahn: Ka’ala summit, boggy spot near TV station.
4025 ft., 12 Apr. 1981. R. H. Gagne s.n. (holotype,
BISH-440780).

1 7. Phyllostegia macrophylla (Gaudichaud)
Bentham, Edwards’s Bot. Reg. 15: 1292. 1830.
[Epithet published as “ macrophyllum ,”] Pras-
ium macrophyllum Gaudichaud, Voy. Uranie
453. 1829. TYPE: s.l., 550-735 m, 1819, C.
Gaudichaud- Beaupre s.n. (holotype, P). Based
on distribution of this species, Gaudichaud
probably collected the type on Hawai’i. St.
John treated this collection as mixed, using the
short infructescence as the holotype of P. in-
terrupta H. St. John.

Phyllostegia clavata Bentham, Edwards’s Bot. Beg. 15:
1292. 1830. TYPE: Hawaiian Islands [U.S.A.]. Ha¬
waii: “Owhyee, monte ignovomo,” 1825. J. Macrae
s.n. (holotype, K not seen).

Phyllostegia macrophylla (Gaudichaud) Bentham var. phy-
tolaccoides Sherff, Bot. Gaz. (Crawfordsville) 96:
137. 1934. Phyllostegia phytolaccoides (Sherff) H.
St. John. Phytologia 63: 179. 1987. TYPE: Hawaiian
Islands [U.S.A.]. Maui: [East Maui], E of Olinda,
verv wet forest along pipe line, 4000 ft., 1 Oct. 1916,
A. S. Hitchcock 14897 (holotype, US- 1172391).
Phyllostegia macrophylla (Gaudichaud) Bentham var. re-
myi Sherff, Bot. Gaz. (Crawfordsville) 96: 136. 1934.
Phyllostegia remyi (Sherff) H. St. John, Phytologia
63: 180. 1987. TYPE: Hawaiian Islands [U.S.A.].
Maui: s.l., 1851—1855, J. Rerny 386 (holotype, GH-
1425; isotype, P not seen).

Phyllostegia brevilobata H. St. John, Phytologia 63: 173.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: West Maui, Honokahau drainage basin, 25
Sep. -17 Oct. 1917, C. A. Forbes 502. M (holotype,
BISH-53669).

Phyllostegia cernua H. St. John. Phytologia 63: 173. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai’i:
North Hilo Dist., Laupahoehoe section of Hilo f orest
Beserve, 3100 ft.. 4 Mar. 1981. L. U. Cuddihy & J.
Davis 770 (holotype, BISH-513588; isotype, BISH).
Phyllostegia hualalaiensis H. St. John, Phytologia 63: 175.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai’i: Pu’u Hualalai, Pu’uwa’awa’a, 5500 ft., 29
Dec. 1931, //. St. John. J. W. Coulter. Hashimoto,
Lindsay & Mitchell 1 1355 (holotype, BISH-520688;
isotypes, BISH [6]).

Phyllostegia interrupta H. St. John, Phytologia 63: 176.
1987. Syn. nov. TYPE: Sandwich, C. Gaudichaud
s.n. (holotype, P). St. John treated the type of Pras-
ium macrophyllum as a mixed collection, using the

272

Novon

short infructescence as the holotype of P. interrupta
H. St. John.

Phyllostegia kohalaensis H. St. John, Phytologia 63: 176.
1687. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Mountains of Kohala, W of Honokanenui
Gorge, June 1910, J. P. Rock 8377 (holotype, BISH-
53655 (and -53656]; isotype, BISH). The isotype is
numbered 8378. Bock often numbered his dupli¬
cates individually: the status is based on locality and
date.

Phyllostegia lehuaensis H. St. John, Phytologia 63: 177.
1987. Syn. nov, TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Kona, Pulehua, Jan. 1912, J. F. Rock s.n.
(holotype, B1SH-92705; isotype, BISH).

Phyllostegia ovata H. St. John, Phytologia 63: 179. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i:
Honaunau Forest Beserve, ‘ohi'a-koa forest, 4050 ft.,
28 Mar. 1980, G. Clarke, J. Davis, L Cuddihy & C.
Corn 3 (holotype, BISH-520640; isotypes, BISH [3]).
Phyllostegia phytolaccoides (Sherff) H. St. John var. sale-
brosa H. St. John, Phytologia 63: 179. 1987. Syn.
nov. TYPE: Hawaiian Islands [U.S.A.]. Maui: F. R.
Warshauer 287b (holotype, BISH not seen). The ho¬
lotype has been missing since at least October 1987;
there is a specimen of P. macrophylla: Maui: Hana
District, along Koukouai Gulch, SW of Kipahulu
Valley, wet koa-'ohi'a forest, 22 July 1980, War¬
shauer & McEldowney 2867 (BISH-469843), which
may be the type if St. John transposed the number.
Phyllostegia plurinodosa H. St. John, Phytologia 63: 180.
1987. Syn. nov. TY PE: Hawaiian Islands [U.S.A.].
Hawai'i: Kohala Mts., USIWS Forest Bird Survey
Transect 77, Slat. 5 53/54, 2500 ft., 2 Aug. 1979, F.
R. Warshauer 2495 (holotype, BISH -520653).
Phyllostegia reflexa 11. St. John. Phytologia 63: 180. 1987.
Syn. nov. TY PE: Hawaiian Islands [U.S.A. |. Hawai'i:
Hilo Forest Reserve, Laupahoehoe Section, within
1(K)-2(X) ft. of W boundary of fencel ine [with | Wai-
punalei, 8 Sep. 1981, J. Davis 598 (holotype, BISH
[on loan from Hawaii State Division of Land and
Natural Resources]). Although published as at BISH.
the holotype is part of a Hawaii State Division of
Land and Natural Resources loan that will hopefully
be given to BISH for permanent storage following
recommendation 7A of the ICBN (Greuter et al„
1994).

Phyllostegia retrorsa H. St. John, Phytologia 63: 180.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Hilea forests, 8 May 1961. J. F. Rock s.n.
(holotype, BISH-53648).

Phyllostegia rubescens H. St. John, Phytologia 63: 181.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Hilo Forest Reserve, Laupahoehoe section.
W of Kilau Stream, 2750 ft.. 9 Feb. 1982, J. Davis
685 (holotype, BISH-520030).

lit. Phyllostegia inannii Sherff, Bot. Gaz. (Craw-
fordsville) 96: 136. 1934. Nom. nov. Replaced
name: Stenogyne parviflora H. Mann, Proc.
Amer. Acad. Arts 7: 193. 1867. TYPE: Ha¬
waiian Islands [U.S.A.]. Maui: East Maui, Ha-
leakala, [NW side], 6000-7000 ft., 1864-
1865, //. Mann & W. T. Brigham 407
(holotype, GH not seen; isotypes, BISH [2],
MO not seen, US).

Phyllostegia racemosa Bentham var. bryanii Sherff, Amer.
.1. Bot. 21: 699. 1934. Phyllostegia bryanii (Sherff)
H. St. John, Phytologia 63: 173. 1987. Syn. nov.
TYPE: Hawaiian Islands [U.S.A.]. Moloka'i: heights
back of Kamolo, 1851— 18< 1. W. Hillebrand s.n. (ho¬
lotype, B [presumably destroyed]; isotype, MEL not
seen [photo BISH]).

Phyllostegia niannii is listed as endangered
(Yoshioka et al., 1991; Herbst et al„ 1992). The
information on distribution and populations has not
changed since reported by Wagner et al. (1990). It
was last collected in 1979 on Moloka'i. Reports of
it from 1985 by Wagner et al. (1990) based on ob¬
servations rather than collections were apparently
a inisidentification of P. hispida.

19. Phyllostegia micrantha H. St. John, Phyto¬
logia 63: 178. 1987. TYPE: Hawaiian Islands
[U.S.A.]. O'ahu: Wai'anae Range, Popouwela,
27 Apr. 1910, C. N. Forbes 1589.0 (holotype,
BISH-53660).

When studying the type of this species for this
paper I was unable to place it in any other known
species of the genus. It has small, elliptic, hirsute
leaves 6—10 cm long, 6—8 flowers per verticillaster,
and the calyx is only 2. 5-2. 7 mm long, the smallest
in the genus. I therefore provisionally accept it as
a distinct species. No other collections are known
to me.

20. Phyllostegia mollis Bentham, Linnaea 6: 79.
1831. Phyllostegia parviflora (Gaudichaud)
Bentham var. mollis (Bentham) A. Gray, Proc.
Amer. Acad. Arts 5: 344. 1862. TYPE: Ha¬
waiian Islands [U.S.A.]. O'ahu: 1816-1817, L.

C. A. von Chamisso s.n. (holotype, LE not
seen).

Phyllostegia honolulensis Wawra, Flora 55: 531. 1872.
Phyllostegia parviflora (Gaudichaud) Bentham var.
honolulensis (Wawra) Sherff, Bot. Gaz. (Crawfords-
ville) 96: 138. 1934. TYPE: Hawaiian Islands
[U.S.A.]. O'ahu: dry woods near Honolulu, 1869-
1870, H. Wawra 1988 (holotype, W not seen (photo
BISH]).

Phyllostegia obatae II. St. John, Phytologia 63: 178. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. O'ahu:
Wai'anae Range, subgulch of Kalua'a Gulch, N of
the central ridge. 1900 ft.. 19 Mar. 1978, J. Obata,

D. Palmer & E. Funk 354 (holotype, BISH-520646;
isotypes. BISH [5 ]).

Phyllostegia swezeyi H. St. John, Phytologia 63: 181.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
O'ahu: Wai'anae Mountains, Pu'u Ka'ala. Mokuleia,
Kamananui. gulch E of Pu'u Kaupakuhale. 2<XK) ft.,
23 Oct. 1932, 0. Sweezy, F Williams, T. G. Yuncker
<& E. Y. Hosaka s.n. (holotype, BISH-40354).

Phyllostegia mollis as described by Wagner et al.

Volume 9, Number 2
1999

Wagner

Phyllostegia Nomenclator

273

(1990) was almost exclusively based on 0‘ahu
plants, but did include some of Sherffs descriptive
information because his P. mollis var. mollis was
delimited in the same way as P. mollis by Wagner
et al. I give the following diagnosis to clearly dif¬
ferentiate it from P. pilosa.

Stems densely short-hirtellous (trichomes some¬
what antrorse). Leaves flaccid, ovate to occasionally
elliptic-ovate, usually 10-24 cm long, 4-7 cm
wide, lower surface glandular-dotted and short-hir¬
tellous, densely so along veins, the trichomes erect,
ca. 0.3 mm long, petioles (3.5— )5— 7 cm long, flow¬
ers (2-4)6 per vertieillaster, in relatively open and
delicate inflorescences 8—17 cm long, pubescent
with spreading to erect trichomes, usually 0. 1-0.2
mm long, pedicels 2-3 mm long, calyx 3-4 mm
long, short-hispidulous, more densely so toward the
base, sometimes also more densely pubescent to¬
ward the apex, and conspicuously glandular-dotted,
the teeth linear-deltate, 0.7-1 mm long, apex acute;
corolla tube slender, slightly curved, 7.5—12 mm
long, short-villous and conspicuously glandular-
dotted, lower lip 6.5—9 mm long.

Phyllostegia mollis sensu Wagner et al. (1990)
was listed as an endangered species in 1991 (Ell-
shoff et al., 1990, 1991a), and thus also included
populations from Maui and Moloka'i giving P pilosa
protected status. Historically, P. mollis was known
from the central and southern Wai‘anae Mountains
from Mt. Ka4ala to Honouliuli and from the Ko‘olau
Mountains above Honolulu. It is now known to be
extant only in Kalua‘a Gulch and Pu‘u Kaua in the
Wai‘anae Mountains, and there are probably less
than 50 individuals remaining. J. Lau (pers. comm.
1998) indicated he knew of only two locations for
this species, with a total of about 20 plants. One
population is in Honouliuli Preserve, and the other
is in Schofield Barracks Military Reservation. He
also mentioned that W. Takeuchi collected quite a
few specimens in the late 1980s. In the label data
of some of the specimens from Honouliuli (e.g., Ek-
ahanui Gulch), he indicated that he saw fairly large
numbers of this species. These populations have
not been relocated.

21a. Phyllostegia parviflora (Gaudiehaud) Ben-
thain var. parviflora

Phyllostegia parviflora (Gaudiehaud) Bentham, Linnaea 6:
79. 1831. Prasiurn parviflorum Gaudiehaud, Voy. Ur-
anie 453. 1829. TYPE: s.l., 550—735 m. 1819. C.
Gaudiehaud- Beaupre s.n. (holotype, P not seen).
Phyllostegia parviflora (Gaudiehaud) Bentham var. canes-
cetis Sherff, Amer. J. But. 21: 700. 1934. TYPE: Ha¬
waiian Islands | U.S. A. |. Maui: West Maui, moun¬
tains, 1840, II. S. Expl. Exped. s.n. (holotype,
US-57495).

Phyllostegia parviflora (Gaudiehaud) Bentham var. major
Sherff, Amer. J. Hot. 21: 7(H). 1934. T\ PE: Hawaiian
Islands [U.S. A. |. O'ahu: Ko'olau Mts.. Punalu'u. 14—
21 Nov. 1908, J. E Rock 565 (holotype, BISH-
510100). Another sheet at B1SH is labeled as Rock
963, but, based on locality and date, it is a possible
isotype. Rock would often number duplicate speci¬
mens of collections made on the same day and at
the same place sequentially or in some cases, such
as this one, renumber the duplicates.

Phyllostegia manoana II. St. John, Phytologia 63: 177.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S. A.).
O'ahu: valley and ridge E of main Konahuinui trail.
16 Apr. 1909, C. N. Forbes 1506.0 (holotype, BISH-
53739).

Phyllostegia parviflora is listed as an endangered
species (Bruegmann, 1995d; Russell & Brueg-
mann, 1996c). Phyllostegia parviflora var. parviflo¬
ra was known historically from the Ko‘olau Moun¬
tains and a single collection from West Maui (type
of P. parviflora var. canescens) in 1840. It is cur¬
rently known from a single population of four plants
along North Kaukonahua Stream (Bruegmann,
1995d; Russell & Bruegmann, 1996c). It is also
known from a population of about 30 plants col¬
lected in 1995 from Pu‘u Pauao, between Poamoho
and Schofield Trail ( Perlman 14705 , PTBG, WU).

21b. Phyllostegia parviflora (Gaudiehaud) Ben¬
tham var. glabriuscula A. Gray, Proc. Amer.
Acad. Arts 5: 344. 1862. Phyllostegia gla¬
briuscula (A. Gray) H. St. John, Phytologia 63:
175. 1987. TYPE: Hawaiian Islands [U.S. A.].
Hawai‘i: Mount Kaah [Mauna Kea], 15-20
June 1825, J. Macrae s.n. (lectotype, here des¬
ignated, GH-1430). St. John designated the
other Macrae collection (also collected on
Mauna Kea, June 1825, GH-1431) as the ho¬
lotype of P. curta. He annotated this sheet as
the lectotype, but did not publish the leetotyp-
ification. St. John was correct in distinguishing
the two syntypes as separate collections, as the
one designated holotype of P. curta is in early
flowering and has very few glandular tri¬
chomes, whereas the other is in late flowering
and has many more glandular trichomes. They
do not, however, represent separate taxa in my
view.

Phyllostegia curta II. St. John, Phytologia 63: 174. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S. A.). Hawai'i:
Mount Kaah [Mauna Kea], 15— 20 June 1825,7. Ma¬
crae s.n. (holotype, GH-1431). The sheet designated
the holotype of P. curta by St. John is also one of
two syntypes (both Macrae collections from Mauna
Kea, June 1825) of P. parviflora var. glabriuscula.
Mounted on the same sheet is a vegetative specimen
of P. parviflora var. parviflora with a U. S. Expl. Ex¬
ped. label, collected on O'ahu.

274

Novon

Phyllostegia parviflora is listed as an endangered
species (Bruegmann, 1995d; Russell & Brueg-
mann, 1996c). Phyllostegia parviflora var. glabrius-
cula was known historically from a few collections
from Manua Kea and has not been collected since
the late 1800s and is presumed extinct.

21e. Phyllostegia parviflora (Gaudichaud) Ben-
tham var. lydgatei (Sheri!) W. L. Wagner, No¬
von 9: 282. 1999. Phyllostegia mollis Bentham
var. lydgatei Sherif, Amer. J. Bot. 21: 700.
1934. Phyllostegia lydgatei (Sherif) H. St.
John, Phytologia 63: 177. 1987. TYPE: Ha¬
waiian Islands [U.S.A.]. O'ahu: Wai'anae Mts.,
1869, J. M. Lydgate s.n. (holotype, B [presum¬
ably destroyed]). O'ahu: s.l., IE Hillebrand s.n.
(lectotype, designated by Wagner (1999), US-
809370).

Phyllostegia waianaeensis H. St. John, Phytologia 63: 182.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A. |.
O'ahu: Wai'anae Range, small gulch bottom inside
of South Palawai Gulch, 2300 ft.. 28 Mar. 1948, H.
1.. Wilbur 609 (holotype, BISH-53699).

In the revision by Wagner et al. (1990) the
Wai‘anae Mountain plants were thought to perhaps
represent a different variety. The situation was re¬
cently evaluated (Wagner, 1999), and the Wai'anae
Mountain populations were recognized as a distinct
variety, P. parviflora var. lydgatei. It is known at
730-825 m, from Napepeiauolelo Gulch and North
Palawai Gulch, Wai'anae Mountains, O'ahu. This
taxon appears to be restricted to north-facing slopes
anil is quite rare, with a current estimate of one
extant population in North Palawai Gulch with
about 20 known individuals. When Phyllostegia
parviflora was listed as endangered (Bruegmann,
1995d; Russell & Bruegmann, 1996c) these pop¬
ulations were included, although a formal name un¬
der P. parviflora was not yet available.

22. Phyllostegia pilosa H. St. John, Phytologia
63: 179. 1987. TYPE: Hawaiian Islands
[U.S.A.]. Maui: East Maui, Ukulele, Mar.
1912, J. F. Rock s.n. (holotype, BISH-53706;
isotype, BISH).

Phyllostegia brevis H. St. John, Phytologia 63: 173. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Maui:
H. W'awra 1915 (holotype, W not seen; isotype. W).
The label states only “Kauai” for the locality, but
Wawra cited this collection (Flora 55: 530. 1872) as
“Kauai, Wiilder des Haliakala.” This locality is on
East Maui.

Phyllostegia brighamii H. St. John, Phytologia 63: 173.
1987. Syn. nov. TYPE: Hawaiian Islands [ U.S.A . ].
Maui: West Maui, ridge above Kamalaea Bay, 1864—

1865, //. Mann A W. T. Brigham 405 (holotype, (ill;
isotype, BISH).

Phyllostegia kamokuensis H. St. John, Phytologia 63: 176.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Moloka'i: Kamoku Camp, 20 Mar. 1912. J. F. Bock
61 19 (holotype, BISH-53696; isotypes, BISH, NY).
Phyllostegia kauaiensis H. St. John, Phytologia 63: 176.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Kaua'i: H. Wawra 2310 (holotype, W). The label
stales only “Kauai” for the locality, but Wawra cited
this collection (Flora 55: 530. 1872) as “Kauai,
Walder des Haliakala.” This locality is on East
Maui.

Phyllostegia triangularis H. St. John, Phytologia 63: 182.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Moloka'i: above Kamolo near Pelekuna, Apr. 1910,
J. F. Bock 7023 (holotype, BISH-53691; isotype, NY).

This species was included within Phyllostegia
mollis by Wagner et al. (1990) following Sherif
(1935), but without seeing any specimens. Later, I
studied specimens from Moloka'i and Maui deter¬
mining them to represent a different species than
the O'ahu plants. I began using the name Phyllo¬
stegia haliakalae for it as this was the oldest name
in the synonymy of P. mollis in the 1990 treatment,
but omitted entirely in SherfTs revision (1935).
When I examined an isotype of Phyllostegia hal¬
iakalae I found that it represented an older name
for the species treated by Wagner et al. (1990) as
P. imminuta. Thus, I have here selected P pilosa
for this species among the names available. It dif¬
fers from P. mollis, which is restricted to O'ahu, in
several characters, most notably the greater number
of flowers per verticillaster, closer spacing of ver-
ticillasters, shorter pedicels, smaller flowers, small¬
er leaves, and shorter petioles. The following is a
brief diagnosis to contrast P. pilosa with P mollis,
as well as to distinguish it from other species in
the genus.

Stems moderately short-hirtellous, the trichomes
spreading, usually somewhat antrorsely so. Leaves
ovate to occasionally elliptic-ovate, up to 8-14 cm
long, 3-4. 5(— 7.5) cm wide, lower surface glandular-
dotted and short-hirtellous, densely so along veins,
the trichomes erect to somewhat crinkly, 0.2— 0.3
mm long, petioles 2.5-3.2(-4.5) cm long, flowers
6-12 per verticillaster, in relatively narrow, usually
simple inflorescences 7—12 cm long, pubescent
with spreading to erect trichomes usually 0. 1-0.2
mm long, pedicels 0.6-1. 8 mm long, calyx 3.2-3.6
mm long, short-hispidulous, more densely so to¬
ward the apex, and inconspicuously glandular-dot¬
ted, the teeth linear-deltate, 0.7-1. 6 mm long, apex
acute; corolla tube slender, slightly curved, 7.3-10
mm long, short-villous, the trichomes somewhat ap-
pressed, and inconspicuously glandular-dotted,
lower lip 3. 5-5. 5 mm long.

Volume 9, Number 2
1999

Wagner

Phyllostegia Nomenclator

275

Phyllostegia mollis sensu Wagner et al. (1990)
was listed as an endangered speeies in 1991 (Ell-
shoff et al., 1990, 1991a), and thus included pop¬
ulations from Maui and Molokai, giving P. pilosa
protected status. Ellshoff et al. (1991a) reported a
population of a few individuals of P. pilosa discov¬
ered by R. Hobdy and J. Lau in Waiopai Gulch on
East Maui; I have not been able to confirm their
identification. Other than this population the spe¬
cies was most recently collected in 1989 at Wai-
kamoi, TNCH [The Nature Conservancy of Hawaii]
preserve, in 1989 in Honomanu Gulch on East
Maui ( Perlman et al. 10773, PTBG), and in 1912
on Moloka‘i (Poholua, Forbes 102. Mo, BISH). J.
Lau (pers. comm. 1998) reported another popula¬
tion more than 4 km away, in the northwestern cor¬
ner of Waikamoi Preserve, near the Waikamoi
Flume road. He last saw this population in 1997,
when there were at least three plants still alive.

23. Phyllostegia racemosa Bentham, Edwards’s
Bot. Reg. 15: 1292. 1830. TYPE: Hawaiian
Islands [U.S.A.]. Hawaii: montem Kaah
[Mauna Kea], 15-20 June 1825, J. Macrae s.n.
(holotype, K).

Phyllostegia rhuakos H. St. John, Phytologia 63: 181. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawaii:
Kipukas. flow of 1823, level of Keawe Wye, 28 June
1915, C. N. Forbes 983. H (holotype, BISH-53747).

Phyllostegia racemosa is listed as endangered
(Bruegmann, 1995b, 1996). Historically, it was
known to usually grow epiphytically in mesic to wet
forests in the Hakalau and Saddle Road areas of
Mauna Kea and the Kulani/Keauhou and Kipuka
Ahiu areas of Mauna Loa. Currently only four pop¬
ulations are known from Hakalau, Kulani/Keauhou,
and in Hawaii Volcanoes National Park with at least
45 individuals (Bruegmann, 1996).

24. Phyllostegia renovans W. L. Wagner, No-
von 9: 280. 1999. TYPE: Hawaiian Islands
[U.S.A.]. Kaua‘i: Hanalei District, headwaters of
Wainiha River, NE fork, just SW of Mahinakehau
Ridge, 680—825 m, 29—30 Jan. 1993, D. H. Lor-
ence, S. Perlman & K. Wood 7315 (holotype, US-
3362751; isotypes, BISH, PTBG).

Phyllostegia renovans was discovered in 1989,
after the most recent revision of the genus was in
press, and was not described until 1999. It is
unique in the genus in that the stems resume veg¬
etative growth after flowering. A brief diagnosis is:
erect subshrub becoming scandent and the stems
up to 3-4 m long, short-hirsute throughout with
leaves narrowly to broadly ovate, flowers (4— )6(— 8)
per verticillaster, apparently the stem resuming
vegetative growth after flowering, calyx campanu-

late, 8-12 mm long, the lobes 4—7 mm long, mar¬
gins with 1 to several coarse teeth, apex acute; co¬
rolla white, ca. 19-22 mm long. It is relatively rare
with 23 small populations known only from three
adjacent valleys on the northeastern part of the is¬
land of Kaua‘i, Hawaiian Islands (Hanakoa, Lima-
huli, and Wainiha), at 680-1040 m; it grows along
watercourses and at the bases of waterfalls.

25. Phyllostegia rockii Sherff, Bot. Gaz. (Craw-
fordsville) 96: 139. 1934. TYPE: Hawaiian Is¬
lands [U.S.A.]. Maui: [East Maui], Hank of Ha-
leakala, 1220—1525 m, s.d., J. M. Lydgate s.n.
(holotype, B presumably destroyed). This name
is based on the unnamed Phyllostegia hispida
Hillebrand var. /J Hillebrand.

Phyllostegia rockii is known from only three col¬
lections from East Maui, the most recent of which
was made in 1912; it is almost certainly extinct
(Wagner et al., 1990).

26. Phyllostegia staehyoides A. Gray, Proc.
Amer. Acad. Arts 5: 344. 1862. TYPE: Ha¬
waiian Islands [U.S.A.]. Hawai‘i: [in the dis¬
trict of] Waimea, 1840, U. S. Expl. Exped. s.n.
(holotype, US-57485).

Phyllostegia staehyoides A. Gray var. hitchcockii Sherff,
Amer. J. Bot. 21: 700. 1934. Phyllostegia hitchcockii
(Sherff) H. St. John, Phytologia 63: 175. 1987.
TYPE: Hawaiian Islands [U.S.A.]. Moloka'i: Puko‘o,
Mr. Conradt’s place, rain forest, 8 Oct. 1916, A. S.
Hitchcock 15069 (holotype, US-874519).
Phyllostegia mollis Bentham var. glabrescens 0. Degener
& Sherff, in Degener, FI. Hawaiiensis, fain. 316.
Phyllostegia mollis. 1935. TYPE: Hawaiian Islands
[U.S.A.]. Molokai: W of Pepe‘opae [Bog], open rain
forest, 12 Apr. 1928, O. Degener 5413 (holotype, F
not seen; isotypes, BISH, NY). The label on the
BISH sheet indicates isotypes at MO, MICH, and B.
Phyllostegia ledyardii H. St. John, Pacific Sci. 30: 27.
1976. TYPE: Hawaiian Islands [U.S.A.]. Hawaii:
[slopes of Mauna Loa above Kealakekua], 26-29
Jan. 1779, D. Nelson s.n. (holotype, BM-550455).
Phyllostegia brevicalycis H. St. John, Phytologia 63: 172.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: West Maui, Wailuku Pali, 21 May 1920, C. N.
Forbes 2446. M (holotype, BISH-53760).

Phyllostegia rnolokaiensis H. St. John, Phytologia 63: 178.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Molokai: Nualele Valley, 18 Mar. 1952, 0. Degener
& A. Tam 22202 (holotvpe, W not seen [photo
BISH]; isotype. BISH).

Phyllostegia montana H. St. John, Phytologia 63: 178.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Molokai: Transect 1, Station 4, 2930 ft., 14 July
1982, F. K. Warshauer 2408 (holotype, BISH-
437828).

Phyllostegia occidentalis H. St. John, Phytologia 63: 178.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: West Maui, Ukumehame Gulch, 4500 ft., 28

276

Novon

Dec. 1955, II. St. John. R. Ihilmer, G. Pearsall & H.
A. Woolford 25696 (holotype, BISH-522873; isotype,
BISH [fragment)).

Phyllostegia olokuiensis H. St. John, Phytologia 63: 179.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.|.
Moloka'i: Wailau, Olokui, Wailele Stream, mossy
rain forest. 37(K) ft., 4 Pel). 1948, //. St. John & R.

I. . Wilbur 23246 (holotype, BISH -630988: isotype,
BISH).

Phyllostegia pilosula H. St. John, Phytologia 63: 180.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.],
Moloka'i: Moloka'i Dist., S of Papa'ala Pali, W of
kuana Ridge, 45(H) ft., 17 July 1979, F. R. War-
shauer & P. II. McEldowney 2438 (holotype, BISH-
522982).

Phyllostegia stachyoides is now quite rare and is
probably endangered, despite its three-island dis¬
tribution on eastern Moloka'i, West Maui, and Ha-
wai'i. The most recent collections are from Moloka'i
(Wawaia Gulch in 1997, Wood 6280, AD, BISH,
NY, PTBG, US) and collections from West Maui
(Honokohau Falls in 1996, Wood 5555, PTBG, US,
and in 1997, Perlman & Wood 15768, AD, BISH,
HAST, K, NY, US; Lihau in 1992, Welton & Haas
1571, BISH; Hanaula in 1985, Hobdy 2399, BISH;
Pu'u Kukui in 1996, Perlman et al. 15368, BISH,
PTBG, US). I am unaware of any recent collections
from the island of Hawai'i; however, J. Lau (pers.
comm. 1998) indicated that The Nature Conservan¬
cy records show the most recent collection was in
1978 (North Kona, Keauhou 1).

27. Phyllostegia tahitensis Nadeaud, Enum. PI.
Tahiti 56. 1873. TYPE: Society Islands. Tahiti:
Mount Marau, ravines, 1200 m, 26 May 1859,

J. Nadeaud 373 (holotype, P not seen; isotype,
G not seen).

This species is apparently known only from the
type collection and is almost certainly extinct.
Sherff (1935) provided a description.

28. Phyllostegia tongaensis H. St. John, Phyto¬
logia 63: 182. 1987. TYPE: Tonga. ‘Ata Island,
11 Nov. 1958, E. Soakai 226 (K not seen).

29. Phyllostegia variahilis Bitter, Abb. Natur-
wiss Vereine Bremen 16: 432. 1900. TYPE:
Laysan, near the shore, on W and E sides,
1896—1897, 11. 11. Schauinsland s.n. (holotype,
BREM not seen; isotypes, BISH [2]).

Phyllostegia arenicola It. St. John, Phytologia 63: 172.
1987. Syn. nov. TYPE: Midway Atoll. Eastern Is¬
land, central plain, 3 m, 23 Apr. 1923, E. L. Gaum
25 (holotype, BISH-53767).

Phyllostegia variahilis is extinct; it was last ob¬
served in 1961 (Wagner et al., 1990).

30. Phyllostegia velutina (Sherff) H. St. John,
Phytologia 63: 182. 1987. Phyllostegia macro-
phylla (Gaudichaud) Bentham var. velutina
Sherff, Bot. Gaz. (Crawfordsville) 96: 137.
1934. TY PE: Hawaiian Islands [U.S.A.]. Ha¬
wai'i: Mauna Kea, 1840, U. S. Expl. Exped.
s.n. (holotype, US-57471).

Phyllostegia velutina was raised to specific level
by St. John, and was accepted in the most recent
revision (Wagner et al., 1990). It is listed as en¬
dangered (Bruegmann, 1995b, 1996). Historically,
it occurred on the southern slopes of Hualalai and
the eastern, western, and southern slopes of Mauna
Loa. Currently, it is known from populations in
three areas (Pu'uwa'awa'a, Honuaula Forest Re¬
serve, and Kulani/Keauhou area) with estimates of
up to 116 individuals (Bruegmann, 1995b, 1996;
U.S. Fish and Wildlife Service, 1998). Another
population is presumably at Waiea Tract, but its
location and status are unknown (Bruegmann,
1995b, 1996). Among the remaining populations I
have seen specimens only from the Kulani popu¬
lation.

3 1 . Phyllostegia vestita Bentham, Edwards’s
Bot. Reg. 15: 1292. 1830. TYPE: Hawaiian
Islands [U.S.A.]. Hawai'i: Byron’s Bay [Hilo],
23 June— 16 July 1825, J. Macrae s.n. (holo¬
type, K).

Phyllostegia dentata Bentham. Edwards’s Bot. Reg. 15:
1292. 1830. TYPE: Hawaiian Islands [U.S.A.]. Ha¬
wai'i: Mount kaah [Mauna kea). 23-16 June-July
1825, ,/. Macrae s.n. (holotype. K; isotype, (ill not
seen).

Phyllostegia catenulata H. St. John. Phytologia 63: 173.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: Parker Ranch, Pa'auhau 2. J. F. Rock 3436
(holotype. BISH-53658).

Phyllostegia foliosa H. St. John, Phytologia 63: 175. 1987.
Syn. nov. TY PE: Hawaiian Islands [U.S.A. |. Hawai'i:
Alakahi. Kawainui. ditch trail, 13 July 1909. J. F.
Rock 4314 (holotype. BISH-512417).

Phyllostegia multiflora H. St. John, Phytologia 63: 178.
1987. Syn. nov. TYPE: Hawaiian Islands [ U.S. A. J.
Hawai'i: mountains above Waimea, June 1910, J. F.
Rock 8312 (holotype, BISH-53771: isotype, BISH).
Phyllostegia pubens H. St. John, Phytologia 63: 180. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A. |. Hawai'i:
South Kohala Dist., Kohala Mts. along kehena
Ditch, USFWS Forest Bird Survey transect 78(15),
wet 'ohi'a forest, 42(H) ft., 31 July 1979. F. R. War-
shauer & P. II. McEldowney 2477 (holotype. B1SH-
520644; isotype, BISH).

Phyllostegia serrata H. St. John, Phytologia 63: 181. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i:
South kohala Dist., kohala Mountains, along ke¬
hena Ditch, USEWS Forest Bird Survey Transect 78
(20/21), 43(H) ft., 2 Aug. 1979, F. R. Warsliauer &
P. II. McEldowney 2490 (holotype, BISH-469857).

Volume 9, Number 2
1999

Wagner

Phyllostegia Nomenclator

277

Phyllostegia sexiflora H. St. John, Phytologia 63: 181.
1687. Syn. nov. TYPE: Hawaiian Islands [U.S.A.j.
Hawai’i: South Hilo Dist., E slope ol Manna Kea
between Pu‘u Akala and Puaka’a near Honoli’i
Stream, 2580 ft., 13 Sep. 1977, F. R. Warshauer &
P. //. McEldowney 1405 (holotype, BISI 1-469847).

32. Phyllostegia waiineae Wawra, Flora 55:
531. 1872. TYPE: Hawaiian Islands [U.S.A.].
Kaua‘i: [Waimea District,] Halemanu, 1869-
1870, //. Wawra 2115 (holotype, W not seen;
isotypes W [2]).

Phyllostegia cordata H. St. John, Phytologia 63: 173.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Kaua‘i: Koke’e camp, 26 June 1926, collector un¬
known s.n. (holotype, BISH-51 3263). The collector
was almost certainly 0. Degener, as he was collect¬
ing at Halemanu on 26 and 27 June 1926 (Degener,
1969).

Phyllostegia hobdyi H. St. John, Phytologia 63: 175. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.). kaua’i:
below Waimea Canyon rim on W side, 2800 ft., 3
June 1969, R. Hobdy 109 (holotype, BISH-32511).

Phyllostegia waimeae is listed as endangered
(Ellshoff et al., 1991b; Merhroff et al., 1994). It was
historically known only from the western part of the
Waimea Canyon drainage, including Halemanu,
Ka'aha, Kaholuamanu, and Kawai’iki, but was un¬
doubtedly more widespread in the Koke‘e area. By
1969 it was known from a single individual in Wai¬
mea Canyon (Merhroff et al., 1994). Its current sta¬
tus is unknown, and it is perhaps extinct.

33. Phyllostegia warshaueri H. St. John, Phy¬
tologia 63: 182. 1987. TYPE: Hawaiian Is¬
lands [U.S.A.]. Hawai’i: 0‘okala Trail, Hama-
kua, 28 May 1979, F. R. Warshauer 4005 (leg.
M. Tamich) (holotype, BISH-520652).

Phyllostegia ambigua (A. Gray) Hillebrand var. longipes
Hillebrand. El. Hawaiian Isl. 350. 1888. Phylloste¬
gia brevidens A. Gray var. longipes (Hillebrand)
Sherff, Amer. J. Bot. 21: 699. 1934. TYPE: Hawaiian
Islands [U.S.A.]. Hawai’i: s.d., J. M. Lydgate s.n. (ho¬
lotype, B presumably destroyed; isotypes, BISH,
MEL not seen [photo BISH]). This plant is presently
known only from Hawai’i, and there is strong doubt
that the type was collected on Maui, although the
original label says “probably E. Maui.”

Phyllostegia warshaueri was raised to specific
level by St. John and was accepted in the most
recent revision (Wagner et al., 1990). It is listed as
endangered (Bruegmann, 1995b, 1996). Historical¬
ly, it occurred on the northern slopes of Mauna Kea
(Laupahoehoe) and from the Kohala Mountains.
Currently, it is known from two populations in the
Kohala Mountains, with estimates of 5 to 10 indi¬
viduals (Bruegmann, 1995b, 1996), and Hamakua

(O'okala Trail). Since listing, it has been rediscov¬
ered in Laupahoehoe near Kilau Stream (3 to 10
plants, Perlman et al. 141R5, 14691, both US).

34. Phyllostegia wawrana Sherff, Bot. Gaz.
(Crawfordsville) 96: 139. 1934. TYPE: Ha¬
waiian Islands [U.S.A.]. Kaua‘i: s.l., 23—29
Feb.— Mar. 1870, 11. Wawra 2060a (holotype,
W not seen; isotype, W).

Considered extinct by Wagner et al. (1990), this
Kaua‘i endemic species was known only from four
collections from Hanalei, Koke‘e Stream, and the
Koke‘e area and was last collected in 1926. Recent
collecting efforts by the National Tropical Botanical
Garden (Lorence et al., 1995) resulted in discovery
of five populations of Phyllostegia wawrana: a Mak-
aleha Plateau site with about 20—30 plants in two
populations in a lowland wet forest; a few individ¬
uals in Wainiha Valley, below Hinalele Falls in a
lowland wet forest; three plants in the upper Han-
akoa drainage in a montane wet forest with riparian
vegetation; a population of 5-6 individuals in the
back ol Honopu Valley growing along a stream bed
in Metrosideros—Dicranopteris montane wet forest;
and Nuololo, Kuia Natural Area Reserve, in a i Me¬
trosideros—Dicranopteris montane mesic forest with
a few individuals. It was listed as endangered based
on these rediscoveries (Bruegmann, 1995a; Russell
& Bruegmann, 1996a). Phyllostegia wawrana ap¬
pears to be closely related to P. helleri but differs
in the longer, stouter trichomes of the stem and
inflorescences (0.6-)0.8-1.3 mm long, longer ped¬
icels 9—11 mm long, conspicuous veins on the ca¬
lyx, and in having inflorescences axillary and some¬
times also terminal. This species, along with P.
knudsenii and P. floribunda, was placed in Phyl¬
lostegia sect. Lateriflorae A. Gray by Sherff (1935).
The same delimitation was followed by Wagner et
al. (1990). Based on the ample material now at
hand I exclude all but the type, P. floribunda, from
the section, because the others have both axillary
and terminal inflorescences. With the exclusion of
these two white-flowered species the section can
now also be characterized by red corollas.

Doubtful Names:

Phyllostegia decemiflorifer 11. St. John, Phytologia 63:
174. 1987. TYPE: Hawaiian Islands [U.S.A. |. Ha¬
wai’i: F. R. Warshauer 1649 (holotype, BISH not
seen). The holotype has been missing at least since
October 1987.

Phyllostegia glandulosa H. St. John. Phytologia 63: 175.
1987. St. John did not adequately designate a unique
collection as the type, as no specific locality, no date,
and no collector number were given, and Rock made
numerous collections in Kona. St. John only cited

278

Novon

“Hawai'i, Kona, J. F. Rock." No Rock collection of
Phyllostegia from Kona could be located. = Proba¬
bly P. slachyoides A. Gray based on the description.
Phyllostegia insignis H. St. John. Phytologia 63: 175.
1987. St. John did not adequately designate a unique
collection as tin* type, as no locality, no date, and no
collector number were given, lie only cited "Maui.
,4. Medeiros." = P. ambigua (A. Gray) Hillebrand
based on the description.

Phyllostegia kilaueaensis H. St. John, Phytologia 63: 176.
1987. TYPK: Hawaiian Islands [U.S.A.]. Hawai'i:
Kilauea Forest Reserve, F. H. Warshauer <A' R. L.
Stemmermann s.n. (holotype, BISH not seen). No
specimen matching the given information was found.
= Probably P. velutina (Sherfl) H. St. John based on
the description. St. John did not adequately desig¬
nate a unique collection as the type, as no specific
locality, no date, and no collector number were giv¬
en.

Phyllostegia secunda H. St. John, Phytologia 63: 181.
1987. TYPE: Hawaiian Islands [U.S.A.]. Kaua'i: Ka-
holuaniano, J. F. Rock s.n. (holotype, BISH not seen).
St. John did not adequately designate a unique col¬
lection as the type, as no specific locality, no date,
and no collector number were given, and Rock made
numerous collections in Kaholuamano. St. John only
cited “Kaua'i, Kaholuamano, J. F. Rock." No Rock
collection of Phyllostegia from Kaholuamano,
Kaua'i, could be located. The short description was
not adequate to determine the species.

Excluded Names:

Phyllostegia haplostachya A. ( I ray, Proc. Amer. Acad.
Arts 5: 345, 1862. TYPE: Hawaiian Islands [U.S.A.].
Maui: sand hills [ of the low isthmus], 1840, U. S.
Expl. Exped. s.n. (holotype, US-57480; isotype, Gil
[fragment] not seen). Sherfl (1935) cited 2 sheets at
US. = Haplostachys haplostachya (A. Gray) II. St.
John.

Phyllostegia haplostachya A. (fray var. leptostachya A.
Gray. Proc. Amer. Acad. Arts 5: 345. 1862. TYPE:
Hawaiian Islands [U.S.A.]. Kaua'i: mts. [on barren
ridges], 1840, U. S. Expl. Exped. s.n. (holotype, US-
57479; isotypes, GH not seen. K, NY not seen). =
Haplostachys haplostachya (A. (hay) H. St. John.
Phyllostegia leptostachys Bentham, Edwards’s Bot. Reg.
15: 1292. 1830. TYPE: Hawaiian Islands [U.S.A.].
Hawai'i: s.l., 1825, J ■ Macrae s.n. (holotype, K). =
Haplostachys haplostachya (A. Gray) H. St. John.
Phyllostegia linearifolia Drake. III. fl. ins. pacif. 47. t. 20.
1886. TYPK: "Ins. Sandvicensibus,” 1851-1855, J.
Ri‘my 396 (holotype, P not seen). = Haplostachys
linearifolia (Drake) Sherff.

Phyllostegia longiflora Gaum, Occas. Pap. Bernice P.
Bishop Mus. 9(5): 9. 1930. TYPE: Hawaiian Islands
[U.S.A.]. Maui: East Maui, Kula pipeline trail, edge
of gulch in deep forest, 7 Feb. 1929, II. Lyon & E.
L. Caum 150 (holotype, BISH-507697; isotypes,
BISH [2|). — Stenogyne kamehamehae Wawra.
Phyllostegia rosmarinifolia H. Mann, Boston Soc. Nat.

Hist. Mem. 1: 536. 1869. Nom. nud.

Phyllostegia truncata A. Gray, Proc. Amer. Acad. Arts 5:
345. 1862. TYPE: Hawaiian Islands [ U.S.A.]. Maui:
1851 — 1855. J. Remy 595 (holotype, GH not seen;
isotypes, BISH. fragment of B specimen at BISH.
P?). = Haplostachys truncata (A. Gray) Hillebrand.

Hybrids:

Phyllostegia cowanii H. St. John, Phytologia 63: 174.
1987. TYPE: Hawaiian Islands [U.S.A.]. O'ahu:
Ridge Trail to Palikea, Honouliuli Forest Reserve,
Wai'anae Range, 3000 ft.. 4 Feb. 1948. R. S. Cowan
808 (holotype, BISH-76011). = P. glabra (Gaudi-
chaud) Bentham var. glabra X P hirsuta Bentham.
Another collection (Wai'anae Mts., Honouliuli Pre¬
serve. below Pu'u Kaua, 2770 ft., 9 May 1996, Wood
5216, BISH, PTBC, US) also appears to represent a
hybrid of ibis combination. It differs, however, in its
closer resemblance to P. hirsuta, differing from it
primarily in its narrower leaves, less pubescent
leaves and stems with somewhat retrorse trichomes,
longer pedicels, and larger calyx. This gives the
plant a superficial appearance of P. parviflora, but it
does not have any glandular trichomes, and the non-
glandular ones are much longer than in P. parviflora.
Both putative parents occur in the general area.
Phyllostegia glabra was growing with Wood 5216
(according to his label), and I found collections
(. Perlman 5287. BISH, PTBG, US) of P. hirsuta grow¬
ing at least as close as the summit crest between
Pu'u Kaua and Pu'u Kanehoa.

Phyllostegia yamaguehii Hosaka & 0. Degener. Occas.
Pap. Bernice P. Bishop Mus. 14(3): 27. 1938. TYPE:
Hawaiian Islands [U.S.A.]. O’ahu: Ko'olau Range,
La’ie, Pu'u Kainapuaa, 2500 ft., 19 Dec. 1937, E.
Y. Hosaka & M. Yamaguchi 1924 (holotype, BISH-
510105; isotypes, BISH |2|, US). Probable hybrid of
P. glabra (Gaudichaud) Bentham var. glabra X P.
hirsuta Bentham.

Invalidi.y Published Names:

Phyllostegia decumbens H. St. John, Phytologia 63: 174.
1987. Stenogyne adpressa IT St. John. Phytologia 63:
159. 1987. Neither of these names are validly pub¬
lished under ICBN Art. 34.2 (Greuter et ah. 1994).
St. John used part of the same specimen, C. I\.
Forbes 188.11 (BISH-4081 12), as the type of both
names. = P. ambigua (A. (fray) Hillebrand.

Phyllostegia kauensis H. St. John, Phytologia 63: 176.
1987. Phyllostegia moniliformis H. St. John. Phyto¬
logia 63: 178. 1987. Neither of these names is val¬
idly published under IGBN Art. 34.2 (Greuter et ah,
1994). St. John used the same specimen, J. D. Jacobi
722 (BISH-426000), as the type for both names. =
P. macrophylla (Gaudichaud) Bentham.

Phyllostegia leukantha 11. St. John, Phytologia 63: 177.
1987. Phyllostegia longilubata H. St. John, Phyto¬
logia 63: 177. 1987. Neither of these names is val¬
idly published under IGBN Art. 34.2 (Greuter et ah,
1994). St. John used the same specimen, R. Hobdy
802 (BISH-522879), as the type for both of these
names. = P. ambigua (A. Gray) Hillebrand.

Phyllostegia parviflora (Gaudichaud) Bentham var. gau-
dichaudi A. Gray, Proc. Amer. Acad. Arts 5: 344.
1862. Not validly published, as this should have
been the autonymic variety (ICBN Art. 26.2; Greater
et ah, 1994). = P. parviflora (Gaudichaud) Bentham
var. parviflora.

Acknowledgments. I thank Robynn Shannon for

her assistance with compiling type information for

the database, reviewing my analyses of nomencla-

Volume 9, Number 2
1999

Wagner

Phyllostegia Nomenclator

279

ture, handling the loans of specimens, and word
processing the output from the database. I thank
Debbie Carino and Clyde Imada for providing in¬
formation from BISH collections for rare and en¬
dangered species, and Joel Lau for sharing his ex¬
tensive knowledge of populations of Hawaiian
plants. I appreciate the quick assistance of Tim
Flynn, Derral Herbst, Pat Holmgren, Chris Puttock,
Emily Wood, and Ken Wood with specimens. I ap¬
preciate the help of Werner Greuter, John McNeill,
and Dan Nicolson for helping me apply the ICBN
to the problem of simultaneous publication of two
species using tbe same specimen. I appreciate the
assistance from the staff of the Herbarium Pacifi-
cum and the National Tropical Botanical Garden
during research visits there. I thank the curators of
the following herbaria for the loans of specimens:
BISH, BM, CU, F, G, GH. K, NY, P, PTBG, and W.

Literature Cited

Bruegmann, M. M. 1995a. Endangered and threatened
wildlife and plants; Proposed endangered or threatened
status for nineteen plant species from the Island of Kau¬
ai. Hawaii. Fed. Beg. 60: 49359—49376.

- . 1995b. Endangered and threatened wildlife and

plants; Determination of endangered status for thirteen
plants from the island of Hawaii, state of Hawaii. Fed.
Beg. 60: 49377-49302.

- . 1995c. Endangered and threatened wildlife and

plants; Proposed endangered status lor twenty-five plant
species from the Island of Oahu, Hawaii. Fed. Beg. 60:
51398-51417.

- . 1995d. Endangered and threatened wildlife and

plants; Proposed endangered or threatened status for
fourteen plant taxa from the Hawaiian Islands. Fed.
Beg. 60: 51417-51432.

- . 1996. Endangered and threatened wildlife and

plants; Determination of endangered status for thirteen
plants from the island of Hawaii, state of Hawaii. Fed.
Beg. 61: 53137—53153.

Degener. 0. 1969. Flora Hawaiiensis, Field Work in the
Hawaiian Islands. Published privately, 8 pp. (D^ r3—

P„.,. . J-

Ellshoff, Z. F.. J. M. Yoshioka, J. F. Canfield. I). B.
Herbst & P. C. Welton. 1990. Endangered and threat¬
ened wildlife and plants; Proposed endangered status
for 26 plants from the Waianae Mountains, Island of
Oahu, HI. Fed. Beg. 55: 39664-39680.

- - - - & - 1991a. En¬
dangered and threatened wildlife plants; Determination
of endangered status for 26 plants from the Waianae
Mountains, Island of Oahu, Hawaii. Fed. Beg. 56:
55770—55786.

- . - . - . - & - . 1991b. En¬
dangered and threatened wildlife and plants; Proposed
endangered status for 23 plants from the Island of Kau¬
ai, Hawaii. Fed. Beg. 56: 55862—55885.

Greuter. W. F. B. Barrie, II. VI. Burdet, W. G. Chaloner,
V. Demoulin, I). I.. Hawksworth. P. M. Jorgensen. D. H.
Nicolson. P. G. Silva. P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Herbst, 1). B. 1990. Endangered and threatened wildlife
atul plants; Proposed endangered status for six plants
from the island of Lanai, HI. Fed. Beg. 55: 38236-
38242.

- . 1991. Endangered and threatened wildlife and

plants; Determination of endangered status for six
plants from the island of Lanai, Hawaii. Fed. Reg. 56:
47686-47694.

- . .1. F. Canfield, J. M. Yoshioka & Z. F. Ellshoff.

1992. Endangered and threatened wildlife and plants;
Determination of endangered or threatened status for 16
plants from the island of Moloka’i, Hawaii. Fed. Reg.
57: 46325—46340.

Lammers, T. 1998. New names and new combinations in
Campanulaceae. Novon 8: 31—35.

Lorence, I). H., T. W. Flynn & W. L. Wagner. 1995. Con¬
tributions to the flora of llawai'i. 111. New additions,
range extensions, and rediscoveries of flowering plants.
Bishop Mus. Occas. Pap. 41: 19—58.

Merhoff. L. A., \1. M. Bruegmann, D. B. Herbst. J. E.
Canfield. Z. F. Ellshoff & J. VI. Yoshioka. 1994. En¬
dangered and threatened wildlife and plants; Determi¬
nation of endangered status for 24 plants from the is¬
land of Kauai. HI. Fed. Beg. 59: 9304—9329.

Russell, C. & M. M. Bruegmann. 1996a. Endangered and
threatened wildlife and plants; Determination of endan¬
gered or threatened status for nineteen plant species
from the Island of Kauai, Hawaii. Fed. Reg. 61: 53070—
53089.

- & - . 1996b. Endangered and threatened

wildlife and plants; Determination of endangered status
for twenty-five plant species from the Island of Oahu.
Hawaii. Fed. Beg. 61: 53089—53108.

- & - . 1996c. Endangered and threatened

wildlife and plants; Determination of endangered or
threatened status for fourteen plant taxa from the Ha¬
waiian Islands. Fed. Reg. 61: 53108—53124.

St. John, H. 1966. Monograph of Cyrtnndra (Gesneri-
aceae) on Oahu, Hawaiian Islands. Bernice P. Bishop
Mus. Bull. 229: 1-465.

- . 1987. Diagnoses of Phyllostegia species (Labia-

tae): Hawaiian plant studies 140. Phytologia 63: 1 72 —
182

Sherff, F. E. 1935. Revision of Haplostachys, Phyllostegia.
and Stenogyne. Bernice P. Bishop Mus. Bull. 136: 1—
101.

- . 1953. Notes on miscellaneous dicotyledonous

plants. Bot. Leaflets 8: 1 3 — 2f».

U.S. Fish and Wildlife Service. 1998. Big Island II: Ad¬
dendum to the Recovery Plan for the Big Island Plant
Cluster. U.S. Fish and Wildlife Service, Portland,
Oregon.

Wagner, Wr. L. 1999. A new species of Hawaiian Phyllos-
tegia (Lamiaceae) from Kaua'i, and recognition of a
Wai'anae Mountain, 0‘ahu endangered variety of Phyl¬
lostegia parviflora. Novon 9: 280—283.

- & R. K. Shannon. 1999. Nomenclator of Hawaiian

Sicyos (Cucurbitaeeae). Novon 9.

- , I). B. Herbst & S. II. Sohmer. 1990. Manual of

the Flowering Plants of llawai'i, 2 Vols. Univ. Hawaii
Press & Bishop Museum, Honolulu (Bishop Museum
Special Publication 83).

Yoshioka, J. M„ Z. F. Ellshoff, J. E. Canfield & I). B.
Herbst. 1991. Endangered and threatened wildlife and
plants; Proposed endangered or threatened status for 16
plants from the island of Moloka'i. Hawaii. Fed. Reg.
56: 47718-47732.

A New Species of Hawaiian Phyilostegia (Lamiaceae) from Kaua‘i
and Recognition of a Wai‘anae Mountain, 0‘ahu, Endangered
Variety of Phyilostegia parviflora

Warren L. Wagner

Department of Botany, MRC 166, National Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560, U.S.A. wagner.warren@nmnh.si.edu

ABSTRACT. A rare new species of the Pacific ge¬
nus Phyilostegia , P. renovans, is described. It is
unique in the genus in that the stems resume veg¬
etative growth alter flowering. It is relatively rare,
with 23 small populations known only from three
adjacent valleys on the northeastern part of the is¬
land of Kaua‘i, Hawaiian Islands. A new combi¬
nation is also provided for the Wai‘anae Mountains,
(Palm, endangered endemic variety of Phyilostegia
parviflora now known from only one extant popu¬
lation of about 20 individuals.

The intensive collecting effort by the National
Tropical Botanical Garden collectors during the
past decade has shown that Kaua'i was one of the
least effectively collected Hawaiian Islands. Kaua'i
is among the most interesting because it has the
highest level of single-island endemism in the ar¬
chipelago (Wagner et ah, 1990; Sakai et al., 1995;
Wagner & Funk, 1995). Especially in need of in¬
tensive survey effort was the northeastern quadrant
of Kaua'i. A number of new species and many
range extensions have been made in the past sev¬
eral years of work in this region of the island. One
of these new species is a Phyilostegia with a unique
growth habit, in which the vegetative stem reini¬
tiates growth after flowering. Two additional basic
inflorescence patterns occur in the genus: short,
congested inflorescences on usually leafless lateral
branches in the axils of or below the lowermost
leaves (e.g., P. floribunda Bentham, P wawrana
Sherff); and leafy, terminal, racemose inflorescenc¬
es on the central stem and upper lateral branches
(most of the genus, including P. parviflora (Gaudi-
chaud) Bentham). The renewal of vegetative growth
can be viewed as an intermediate stage between the
usual type in Phyilostegia and the short, congested,
lateral type in two species of Phyilostegia. The lat¬
eral type in turn has evolved further into the axil¬
lary vertieillasters in Stenogyne , a closely related
Hawaiian genus of primarily bird-pollinated vines.

Phyilostegia renovans W. L. Wagner, sp. nov.
TYPE: Hawaiian Islands (U.S.A). Kaua'i: Han-
alei District, headwaters of Wainiha Kiver, NE
fork, just SW of Mahinakehau Ridge, lowland
wet forest with Metrosideros polymorpha Gau-
dichaud dominant, slopes moderate to steep,
occasional in clearings, wet soil, 680-825 m,
29—30 Jan. 1993, 1). //. Lorence, S. Perlman
& K. Wood 7315 (holotype, US-3362751; iso¬
types, BISH, PTBG). Figure 1.

Caulibus brevihirsutis, post anthesin recrescentibus;
fioribus (4)6 in quoque verticillastro; calyce 8— 12 mm lon-
gi. marginibus loborum uni-paucidentatis, corolla alba,
19—22 mm longa.

Erect subshrub when young, becoming scandent
and the stems up to 3-4 m long, short-hirsute.
Leaves narrowly ovate to ovate, sometimes broadly
so, 12.5—20 cm long, 5.0— 8.8 cm wide, spreading
hirsute on both surfaces, the hairs primarily con¬
centrated on the veins on the lower surface, the
veins usually red-tinged, margins coarsely dentate
or serrate-dentate, apex acuminate, base truncate,
broadly rounded or broadly cuneate, petioles red¬
dish green, 3.2—5 cm long, moderately to densely
hirsute, sometimes somewhat retrorsely so. Flowers
(4)6 per verticillaster, sometimes the tenninal one
with 8 flowers, in an open, unbranched, racemose
inflorescence 18-34 cm long, apparently the stem
resuming vegetative growth after flowering, pedicels
9-10 mm long, sometimes on a common peduncle
up to 4 mm long, moderately to densely short-hir¬
sute, bracts ovate, the lower ones 6-11 cm long,
the upper ones 2.5-3 cm long; calyx campanulate,
8-12 mm long, green, short-hirsute on both surfac¬
es, the hairs spreading to somewhat antrorsely ap-
pressed, the lobes 4-7 mm long, widely spreading,
the margins with 1— several coarse teeth, apex
acute; corolla white, ea. 19—22 mm long, short-hir¬
sute mostly on the upper side, upper lip 6—7 mm
long, the lower lip 9—12 mm long. Nutlets ca. 8-9
mm long, greenish black.

Novon 9: 280-283. 1999.

Volume 9, Number 2
1999

Wagner

Hawaiian Phyllostegia

281

Figure 1. Phyllostegia renovans W. L. Wagner ( Lorence et al. 7315). — a. Inflorescence showing the apex resuming
vegetative growth. — b. Flower. — c. Calyx with maturing fleshy nutlets. — d. Nutlet. — e. Pubescence of the inflores¬
cence. Scale bar: a— c = 2 cm; d = 1 cm; e = 2 mm. The original illustration is on indefinite loan to the Hunt Institute
for Botanical documentation and featured in their 8th International Exhibition.

Distribution. Known only from the upper parts
of three valleys on the northeastern side of Kaua‘i,
Hanakoa, Limahuli, and Wainiha, at 680—1040 m,
where it grows along the watercourses and at the
bases of waterfalls. A collection with only young
flower buds from Wai‘oli Valley ( Perlman et al.

13256, BISH, US) may also represent this species,
but is somewhat more pubescent; it should be re¬
collected in flower or fruit for verification of its sta¬
tus. Phyllostegia renovans occurs in scattered
patches; Wood (pers. comm., 1997) noted about 20
small populations in the area explored in upper Li-

282

Novon

mahuli Valley. The habitat is Metrosideros polymor-
pha Gaudichaud var. glaberrima (H. Leveille) H.
St. John wet forest near streams associated with An-
tidesma platyplyllum H. Mann, Boehmaria grandis
(Hooker & Arnott) A. Heller, Broussaisia arguta
Gaudichaud, Cheirodendron, Cibotium, Cyanea,
Cyrtandra, Dicranopteris linearis (N. L. Burmann)
Underwood, Diplazium sandwichianum (C. Presl)
Diels, Dubautia knudsenii Hillebrand, Freycinetia
arborea Gaudichaud, Gunnera kauaiensis Hock,
Hedyotis terminalis (Hooker & Arnott) W. L. Wag¬
ner & Herbst, Ilex anamola Hooker & Arnott, La -
bordia, Machaerina angustifolia (Gaudichaud) T.
Koyama, Peperomia, Perottetia sandwicensis A.
Gray, Pipturus, Psychotria, Sadler in, Sceavola, Sy-
zygium sandwicen.se (A. Gray) Niedenzu, and Tetra-
plasandra. Threats to these populations are intro¬
duced animals and plants such as Clidemia hirta
(L.) D. Don, Hedychium gardenarianum Ker Gaw-
ler, and Rubies argutus Link.

The red coloration in the leaves occurs at least
when plants are in sunny sites and apparently is
less intense in shaded sites. Lorence (pers. comm.,
1995) noted that the flowers are not fragrant in the
morning (10—11 a.m.), but have a light fragrance at
5 p.m., suggesting moth pollination.

Paratypes. HAWAIIAN ISLANDS (U.S.A.). Kaua‘i:
Hanalei District, upper Hanakoa Valley. Honoonapali Nat¬
ural Area Reserve, 3500 ft., Perlman & Hill 10830 (BISH.
PTBG, US); below and S of Pihea Peak, steep slopes with
hanging valleys, along drainage, 1060 m. Wood 1341 (AD,
BISH, F, K. MO, PTBG [2]); upper Hanakoa Valley, below
Pihea, on stream banks and slopes, 34(X) ft., Perlman
13008 (BISH, MO, NY, P, PTBG, US); Hanakoa drainage,
1 100-1170 m. Wood el al. 1750 (PTBG); Wainiha Valley,
below Hinalele Falls, 2500 ft.. Wood el al. 2255 (PTBG,
US), 21B0 ft., Perlman et al. 13215 (BISH, PTBG, US);
NF fork of Wainiha River, SW of Mahinakehau Ridge,
2340 ft., Perlman el al. 13302 (BISH. PTBG, US); upper
Limahuli Valley above falls, Hono'onapali, 2600-3200 ft..
Wood et al. 3265 (PTBG), 2560 ft., Perlman el al. 14256
(BISH, NY, PTBG, US); Limahuli Valley, back of valley
below Pali Ele'ele, along stream, 2800 ft., Perlman &
Wood 15617 (BISH, PTBG, US).

When the most recent revision of Phyllostegia
was nearing completion (Wagner et al., 1990), Steve
Perlman collected P. parviflora in the WaPanae
Mountains, 0‘ahu, between Palikea and Pohakea
Pass (Perlman 5679, BISH, PTBG). It differed from
populations of this species in the Ko‘olau Moun¬
tains, 0‘ahu, where it was historically most com¬
mon, and from Maui and Hawai‘i populations in its
shorter pedicels 2-5(-6) mm long, spreading hairs
on the petioles, slightly smaller flowers, and rela¬
tively few gland-tipped hairs in the inflorescence.
Wagner et al. (1990) thought that it represented a
new variety, but also that it may have already been

named as Phyllostegia mollis var. lydgatei Sherff.
At that time this name could not be definitely
placed under P. parviflora since the type material
is not extant. However, based on the description by
Sherff (1935) and examination of the Hillebrand
specimen at US annotated by Sherff, it appeared to
represent the WaPanae Mountains variety of P. par¬
viflora. I have studied the US sheet cited and de¬
termined that it apparently represents the only orig¬
inal material of Sherffs taxon extant and that it is
the same taxon as Perlman 5679 and other subse¬
quent collections from the WaPanae Mountains. I
therefore here make the new combination in P. par¬
viflora, lectotypifiy Sherffs name, and provide a full
description.

Phyllostegia parviflora (Gaudichaud) Bentham
var. lydgatei (Sherff) W. L. Wagner, comb. nov.
Basionym; Phyllostegia mollis Bentham var.
lydgatei Sherff, Amer. J. Bot. 21: 700. 1934.
Phyllostegia lydgatei (Sherff) H. St. John, Phy-
tologia 63: 177. 1987. TYPE: Hawaiian Is¬
lands (U.S.A.). 0‘ahu: WaPanae Mts., 1869, J.
M. Lydgate s.n. (B destroyed, photo F). Sherff
cited two other collections from the WaPanae
Mountains of 0‘ahu in addition to the type:
Makaleha: J. M. Lydgate s.n. (B destroyed);
and 0‘ahu: s.l., s.d., W. Hillebrand s.n. (lecto-
type, here designated, US-809370). The lec-
totype was annotated by Sherff.

Phyllostegia waiarmeensis H. St. John, Phytologia 63: 182.
1987. Syn. nov. TYPE: Hawaiian Islands (U.S.A.).
0‘ahu: WaPanae Range, small gulch bottom inside
of South Palawai Gulch, 2300 ft., 28 Mar. 1948. R.
L. Wilbur 609 (holotype, BISH-53699).

Shrubby perennial herbs; stems densely an-
trorsely strigose, becoming hirtellous and glandular
hirtellous toward and throughout the inflorescence
(hairs straight or slightly curved, 0. 1-0.6 mm long).
Leaves rugose, ovate to broadly ovate, (15— ) 19— 33
cm long, (6-)7.5— 15.3 cm wide, both surfaces con¬
spicuously (at least when dry) and usually densely
glandular-dotted, adaxial surface appressed long-
hirsute, abaxial surface hirsute to antrorselv stri¬
gose, densely so along veins, margins crenate, apex
acuminate, base truncate to rounded or subcordate,
petioles (5-)6— 13.5 cm long, densely hirtellous.
Flowers (4)6(10) per verticillaster, in leafless, usu¬
ally compound, racemose inflorescences usually
10-20 cm long, consisting of the terminal primary
stem and numerous secondary and tertiary lateral
branches immediately below, hirtellous and sparse¬
ly to moderately glandular hirtellous throughout,
pedicels 2-5(-6) mm long, bracts narrowly ovate to
elliptic-ovate, 6-9 mm long; calyx campanulate, 3—

Volume 9, Number 2
1999

Wagner

Hawaiian Phyllostegia

283

6 mm long, nerves conspicuous, hirtellous and
sparsely to moderately glandular hirtellous, and
glandular-dotted, the teeth narrowly deltate to lin-
ear-deltate or linear-elliptic, 1.5—2 mm long, apex
attenuate to acute; corolla 10—13 mm long, white,
purple-tinged at least on the upper lip, the tube
curved, ca. 8—10 mm long, glandular-dotted, short-
hirtellous, and glandular hirtellous, especially on
the upper side, the upper lip ca. 2-3 mm long, the
lower lip ca. 5—7 mm long. Nutlets ca. 3—6 mm
long.

Distribution. Occurring in diverse mesic forest
with Acacia hoa A. Gray, Antidesma, Claoxylon
sandwicense Muller Argoviensis, Cyanea membran-
aceae Rock, Melicope, Metrosideros polymorpha
Gaudichaud, Pittosporum, Pipturus albidus (Hooker
& Arnott) A. Gray, Pouteria sandwicensis (A. Gray)
Baehni & O. Degener, Pteralyxia macrocarpa (Hil-
lebrand) K. Schumann, Solanum sandwicense
Hooker & Arnott, Strebulus pendulinus (Endlicher)
F. Mueller, Urera glabra (Hooker & Arnott) Wed¬
dell, and U. kaalae Wawra, 730—825 m, from Na-
pepeiauolelo Gulch and North Palawai Gulch,
Wabanae Mountains, 0‘ahu. This taxon appears to
be restricted to north-facing slopes and is quite
rare, with a current estimate of one extant popu¬
lation in North Palawai Gulch with about 20 indi¬
viduals. Phyllostegia parviflora is U.S. federally
listed as endangered (Russell & Bruegmann, 1996).
The final rule included the Wai‘anae Mountains
populations here treated as P. parviflora var. lyd-
gatei as well as those restricted to P. parviflora var.
parviflora from the Ko‘olau Mountains (O'ahu),
West Maui, and Hawai‘i. Phyllostegia parviflora var.
parviflora from the Kcbolau Mountains, 0‘ahu, and
formerly Maui is apparently now restricted to one
locality and four individuals, while P parviflora var.
glabriuscula A. Gray from the island of Hawai‘i has
not been collected since the late 1800s and is also
apparently extinct.

Specimens examined. HAWAIIAN ISLANDS (U.S. A.).
0‘ahu: Wabanae Mountains, Napepeiauolelo Gulch. N-
facing slope. 24(X) ft., Perlman 5679 (BISH, PTBG[2],
US), 14 Mar. 1990, Obata s.n. (BISH), Perlman A Obata
10831 pro parte (BISH) [additional sheets under the same
label at PTBG, US are P. hirsuta Bentham. but Perlman
(pers. comm., 1997) indicated the material was not mixed
when collected and no P. hirsuta occurs at this locality];
Honouliuli, North Palawai Gulch, N-facing slope, two sub¬
gulches down the valley, 2400 ft., Perlman & Obata 6 162
(BISH. PTBG, US). 10 Feb. 1991, Obata A Liu s.n.
(PI'BG); Honouliuli, North Palawai Gulch, N-facing slope,
one subgulch down the valley, 2700 ft., Perlman & Obata
6165 (BISH. PTBG. US).

Acknowledgments. I thank Anna Asquith lor
her excellent illustration of Phyllostegia renovans ,
which was funded by the Smithsonian Walcott Fund
for Botanical Publication; Dan Nicolson for trans¬
lating the diagnosis into Latin; and Tim Flynn,
Dave Lorence, Ken Wood, and Steve Perlman for
help with a variety of details of the collections, hab¬
itats, and field characteristics of the plants. I ap¬
preciate the useful comments by Robynn Shannon
on a draft of this paper. I thank the curators and
staff at PTBG and BISH for loans of specimens and
assistance during my visits to their institutions.

Literature Cited

Russell, C. & M. Bruegmann. 1996. Endangered and
threatened wildlife and plants; Determination of endan¬
gered or threatened status for fourteen taxa from the
Hawaiian Islands. Fed. Reg. 61; 53108—53124.

Sakai, A. K., W. L. Wagner, D. M. Ferguson & 1). R.
Herbst. 1995. Biogeographical and ecological correlates
of dioecy in the Hawaiian angiosperm flora. Ecology 76:
2530-2543.

Sherff, E. E. 1935. Revision of Haplostachys, Phyllostegia,
and Stenogyne. Bernice P. Bishop Mus. Bull. 136: l-
101.

Wagner, W. L. & V. A. Funk (Editors). 1995. Hawaiian
Biogeography: Evolution on a Hot Spot Archipelago.
Smithsonian Institution Press, Washington, D.C.

- , D. R. Herbst & S. H. Sohmer. 1990. Manual of

the Flowering Plants of Hawabi, 2 Vols. Univ. Hawaii
Press & Bishop Museum, Honolulu (Bishop Museum
Special Publication 83).

An Autogamous Rainforest Species of Schiedea (Caryophyllaceae)
from East Maui, Hawaiian Islands

Warren L. Wagner

Department of Botany, MRC-166, National Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560, U.S.A. wagner.warren@nmnh.si.edu

Stephen G. Weller and Ann K. Sakai

Department of Ecology and Evolutionary Biology, University of California, Irvine, California
92717, U.S.A. sgweller@uci.edu, aksakai@uci.edu

Arthur C. Medeiros

Biological Resources Division, U.S. Geological Survey, Haleakala National Park, P.0. Box 369,

Makawao, Hawaii 96768, U.S.A.

ABSTRACT. A new autogamous species of Schie¬
dea is described and illustrated. It is known only
from cliff habitat in rainforest on a single ridge in
the Natural Area Reserve, Hanawi, East Maui.
With the addition of this species there are 28 spe¬
cies in this endemic Hawaiian genus. The new spe¬
cies appears to be most closely related to Schiedea
nuttallii, a species of mesic habitats on O'ahu, Mo-
loka'i, and Maui.

In June 1992 James Jacobi and Arthur Medeiros
were collecting in the wet forests of Hanawi. East
Maui, when they discovered a species of Schiedea
Chamisso & Schlechtendal they considered to be
probably new to science. It appeared similar to S.
nuttallii Hooker, but close study of the herbarium
specimens showed that it was a closely related new
species from a much wetter habitat than S. nuttallii.
In June 1995 Ken Wood and Steve Perlman revis¬
ited the population with Medeiros, W. Evanson, and
R. Hobdy. They obtained seeds lor greenhouse
propagation at the University of California, Irvine.
Studies ol plants grown in the greenhouse con¬
firmed that they represented a new species, here
named Schiedea jacobii. With the addition of this
species there are 28 species in this endemic Ha¬
waiian genus. Observations in the greenhouse also
showed that the new species is autogamous and that

it shares several characteristics with another rain¬
forest species in the genus, S. diffusa A. Gray.
These shared characters, related to the breeding
system and to seed dispersal, include anthers shed¬
ding pollen while the styles are elongated and re¬
ceptive, and seeds persistent on the placenta alter
capsule dehiscence. The latter feature may be as¬
sociated with germination of seeds in the capsule,
followed by dispersal of small seedlings, as noted
for S. diffusa (Wagner, unpublished obs.).

Schiedea jacobii W. L. Wagner, Weller & Med¬
eiros, sp. nov. TYPE: Cultivated from seeds at
University of California, Irvine greenhouse in
1997. Source: Hawaiian Islands. Maui: East
Maui, Hanawi, Kuhiwa unit, E of Hanawi
Stream, between Hanawai Stream and Kuhiwa
drainage [20°45'N, 156°7'40"W], 6340 ft., E-
facing slope, ca. 20 ft. below ridge top, in Me-
trosideros montane wet forest, 22 June 1995,
S. Perlman, K. Wood, A. Medeiros, R. Hobdy
& B. Evanson 14807 (holotype, US-3351775;
isotypes, BISH, CANB, CHR, K, MO, NY,
PTBG). Figure 1A-H.

Haec species S. nuttallii similis, seif ah ea caulibus
quadrangularibus, fforibus autogamis, sepalis sub angulo
ca. 100—135 e pedicello abeuntibus, tubo apieali 1.4— 1.8

l igure 1. Schiedea jacobii W. L Wagner, Weller & Medeiros. — A. Flowering stem. — B. Stem cross section. — C.
Portion ol inflorescence branch. — I). Flower, showing stamens ready to dehisce and styles elongated and receptive.
— E. Adaxial view of sepal with nectary and stamen. — F. Dehisced capsule surrounded by the sepals exposing the
mass of persistent seeds. — G. Seed, lateral view. — H. Seed, marginal view. Drawn from live greenhouse material of
the type.

Nov ON 9: 284-287. 1999.

Volume 9, Number 2
1999

Wagner et al.
Schiedea jacobii

285

286

Novon

mm longo nectarii omnis, stylibus plerumque 4, seminibus
cohaerentibus el persistentibus differt.

Perennial herbs or subshrubs, pale yellowish
green throughout; stems 4.0-7.6 dm long, ascend¬
ing to sprawling when longer, conspicuously qua¬
drangular, the angles weakly winged, glabrous
throughout. Leaves opposite; blade coriaceous, 4.5-
10.5 cm long, 1.4— 2.6 cm wide, lanceolate to ob¬
long-elliptic, only the midvein evident and slightly
eccentric, margin slightly thickened and weakly
revolute, apex acute to attenuate, base acute to ob¬
tuse; petiole 0.4-0.8 cm long, the base often flared
at juncture with stem and the leal pair slightly con¬
nate by petioles. Flowers perfect, facultatively au¬
togamous, in diaehasia on the main and upper lat¬
eral branches, these forming an open, panicle-like
inflorescence 40-50 cm long, the lateral branches
11-18 cm long, ascending, each with 25-35 flow¬
ers; bracts yellowish green, foliaceous, nearly as
large as the leaves in lower portions of central axis
and the lateral branches, those in upper part of
inflorescence and subtending flowers narrowly lan¬
ceolate to subulate, 3-11 mm long; pedicels 3-8
mm long at anthesis, elongating slightly in fruit, but
none over 9 mm long, conspicuously asymmetri¬
cally flattened. Sepals 3. 9—4. 2 mm long, lanceolate,
dull yellowish green, greener toward the midrib,
sometimes with purple toward apex, thick and
opaque, ascending, concave to shallowly navicular
toward apex, oriented at ca. 100° to 135° angle to
the pedicel, abaxially smooth and rounded, gla¬
brous, the margins weakly scarious, sparsely cili-
ate, the apex attenuate. Nectary base 0.4— 0.5 mm
long, somewhat broadened and flattened at base,
dark yellow, the nectary shaft 1.4— 1.8 mm long, the
apex deeply bifid. Stamens 10, the filaments weakly
dimorphic, the antisepalous whorl 3.2— 3.5 mm
long, the alternate set 2.8— 3.0 mm long, the anthers
attached near the middle, subequal, 0.3— 0.4 mm
long, yellow, dehiscing after flower opens. Styles
(3)4—5, the stigmas receptive when flower opens.
Capsule ovoid, 3. 7^4.2 mm long. Seeds ca. 16—22
per capsule, 0.7— 0.8 mm long, suborbicular, asym¬
metrical, compressed, brown, rugose, remaining co¬
herent as a unit and persistent on placenta after
dehiscence, seeds eventually dispersing (or per¬
haps germinating in situ in the wild).

Etymology. We are pleased to name this spe¬
cies in honor of its discoverer, James D. Jacobi,
Biological Resources Division, Pacific Islands Sci¬
ence Center, who has made significant contribu¬
tions to vegetation ecology and alien plant manage¬
ment in Hawaii.

Distribution and ecology. Known only from wet

cliffs, between Hanawai Stream and Kuhiwa drain¬
age, East Maui, at 1930 m. This area is within a
protected Hawaii State Natural Area Reserve. The
habitat is a Metrosideros Banks ex Gaertner mon¬
tane wet forest with Cheirodendron trigynum (Gau-
dichaud) A. Heller, Coprosma ochracea W. R. B.
Oliver, Dubautia dolosa 0. Degener & Sherff, D.
plantaginea Gaudiehaud, Melicope sp., Sadleria cy-
atheoides Kaulfuss, Broussaisia arguta Gaudi-
chaud, Elaphoglossum sp., Uncinia uncinata (L.f.)
Kukenthal, Deschampsia nubigena Hillebrand,
Myrsine lessertiana A. de Candolle, Dryopteris sub -
bipinnata W. H. Wagner & Hobdy, D. wallichiana
(Sprengel) Hylander, Labordia sp., Peperomia sp.,
Pipturus forbesii Krajina, Astelia menziesiana Small,
Carex alligata Boott, Rubus hawaiiensis A. Gray,
and Vaccinium sp. The only known population con¬
sisted of nine plants when first discovered in 1992,
although time for exploration was limited. In 1995
shoots representing only two to four plants could
be located with several hours of searching the area.
Two recently dead plants were located near the liv¬
ing ones and eight cotyledon-stage individuals,
which may be this species, were found growing
among bryophytes. All of the living plants in 1995
were contained within a 2-in2 area. A separate
small population located downhill in 1992 was not
relocated because of a small landslide/tree fall.
Other populations may occur in this rugged, largely
inaccessible region of East Maui, especially on wet
cliffs. Degradation of forests in the Hanawi area by
alien animals appears to be the major threat to
Schiedea jacobii. The only known population is cur¬
rently without any evidence of pigs because of Nat¬
ural Area Reserve System staff efforts.

Paralypes. HAWAIIAN ISLANDS. Maui: Last Maui.
Hanawi, Kuhiwa unit, K of Hanawi Stream, between Han¬
awai Stream and Kuhiwa drainage [20°45'N,
156°7'40"W], E-facing slope, ca. 6 m below ridge top, 22
June 1 995, Perlman et al. I4H07 (US); on rock face with
Deschampsia , 7 Apr. 1992, Medeiros & Jacobi 79H (I5ISH.
US).

Relationships. Schiedea jacobii, with its fleshy
stems, large leaves with one principal vein, open
inflorescence, and attenuate sepal apex, is clearly
a member of the S. nuttallii clade as defined by
Wagner et al. (1995) and Weller et al. (1995), which
consists of nine hermaproditic species of mesic to
wet forest. Schiedea jacobii appears to be most
closely related to S. nuttallii of mesic forests on
0‘ahu, Moloka‘i, and West Maui. This relationship
is suggested by the shape and size of the leaves,
similar shrubby habit, sepal size, and rugose seeds.
It differs from S. nuttallii most conspicuously in
characters related to the breeding system, espe-

Volume 9, Number 2
1999

Wagner et al.
Schiedea jacobii

287

cially those associated with autogamy. The most
significant are the shift in timing of receptivity of
the stigmas and the shorter staminal filaments. Sev¬
eral characters of S. jacobii are also found in S.
diffusa, but appear to have been independently de¬
rived based on the evolution of autogamy in a wet
forest environment. Re-analysis of the phylogeny of
Schiedea following inclusion of S. jacobii in the data
matrix may resolve the extent of convergence be¬
tween this species and S. diffusa (Wagner et al.,
unpublished).

Breeding system. Experimental crosses in the
greenhouse and observations of pollen tube growth
in styles indicate that S. jacobii is autogamous.
Emasculated flowers failed to produce capsules. In
unmanipulated flowers numerous pollen tubes were
observed in the styles (Weller, Sakai, Thai & Dieu,
unpublished data). These results, and the produc¬
tion of hybrids following interspecific crosses, in¬
dicate that the abundant capsule production of this
species in the absence of cross-pollination is likely
to result from self-fertilization rather than apomixis.
Autogamy is facilitated by the synchrony of pollen
dispersal and stigma receptivity, a feature not found
in most other species of Schiedea , which are strong¬
ly protandrous (Sakai et al., 1997). Schiedea dif¬
fusa, which occurs in very wet forests on Moloka‘i,
Maui, and Hawai‘i, is also autogamous. Our obser¬
vations suggest that in very wet habitats, autogamy
may limit exposure of pollen to rain, which may
either wash pollen from flowers or cause pollen to
burst. Additionally, visitation by insects capable of
pollination may be rare in very wet habitats.

Seed dispersal. Among species of Schiedea, re¬
tention of mature seeds on the placenta within de¬
hisced capsules is found only in S. jacobii. A re¬
lated species, S. diffusa, retains seeds within closed
capsules. Seedlings of S. diffusa have been ob¬
served germinating within disintegrating capsules

still attached to the plant. This trait may provide
an advantage during establishment in wet habitats.
Whether precocious germination occurs in S. ja¬
cobii is unknown, although seeds eventually dis¬
perse from capsules under greenhouse conditions.

Acknowledgments. Financial support for this
project was provided by a Smithsonian Institution
Scholarly Studies grant to Wagner, funds provided
by the Smithsonian Institution Mellon Fellowship,
grants from the National Science foundation (BSR
88-17616, BSR 89-18366, DEB 92-07724), and
the National Geographic Society to S. G. Weller and
A. K. Sakai, and a University of California Presi¬
dential Undergraduate Fellowship to Tuan Thai. We
thank Steve Perlman and Ken Wood for collecting
seeds for the greenhouse studies, Dan Nicolson for
translating the diagnosis into Uatin, Robynn Shan¬
non for commenting on and editing an earlier draft
manuscript, and Alice Tangerini for her superb il¬
lustration. We appreciate the helpful comments of
Roy Gereau and Richard Rabeler on the manu¬
script. We thank the Natural Area Reserve (State
of Hawaii) for permission and use of helicopters for
logistical support in gaining access to the Hanawi
NAR.

Literature Cited

Sakai. A. k.. S. G. Weller, M.-L. Chen. S.-Y. Chou & C.
Tasanont. 1997. Evolution of gynodioecy and mainte¬
nance of females: The role of inbreeding depression,
outcrossing rates, and resource allocation in Schiedea
adarnantis (Caryophyllaceae). Evolution 51: 724—736.
Wagner, W. L„ S. G. Weller & A. k. Sakai. 1995. Phy¬
logeny and biogeography in Schiedea and Alsinidendron
(Caryophyllaceae). Pp. 221—258 in W. L. Wagner A V.
A. Eunk (editors), Hawaiian Biogeography: Evolution on
a Hot Spot Archipelago. Smithsonian Institution Press,
Washington. D.C.

Weller. S. G.. W. L. Wagner & A. k. Sakai. 1995. A phy¬
logenetic analysis of Schiedea and Alsinidendron (Car¬
yophyllaceae: Alsinoideae): Implications for the evolu¬
tion of breeding systems. Syst. Bot. 20: 315—337.

Two New Combinations in Chinese Scrophulariaceae

Wu Zheng-yi

Kunming Institute of Botany, Chinese Academy of Sciences, Heilongtan, Kunming, Yunnan

650204, People’s Republic of China

Abstract. Two new combinations, Sopubia mat-
sumurae (T. Yamazaki) C. Y. Wu and Xizangia bart-
sioides (Handel-Mazzetti) C. Y. Wu & D. D. Tao,
are proposed. Sopubia lasiocarpa P. C. Tsoong is
reduced to synonymy of S. matsumurae, and Xiz¬
angia serrata D. Y. Hong is reduced to synonymy
of Xizangia bartsioides.

A re-examination of the generic dispositions of
certain Chinese species of Scrophulariaceae reveals
that the following two species need to he re-as¬
signed to their proper genera.

Sopubia matsumurae (T. Yamazaki) C. Y. Wu.
comb. nov. Basionym: Petitmenginia matsu¬
murae T. Yamazaki, J. Jap. Bot. 25: 214. 1950.

I Y PE: China. Jiangsu [as Kiangsu]: Nanjing
[as Nangking], Chungshan, 29 Aug. 1910, S.
Matsuda s.n. (holotype, TI).

Sopubia lasiocarpa I’. C. Tsoong, FI. Reipubl. Popularis
Sin. 07(2): 405. 1979. Syn. nov. TYPE: China. Un-
nan: Yizhang County, Changping, 340 m, Liang Bao-
han 83323 (holotype. PE).

Yamazaki (1980) reduced Sopubia lasiocarpa and
Petitmenginia matsumurae to synonymy of S. stricta
(Bent ham) G. Don. Sopubia stricta is a southern
Asiatic species that reaches its northernmost limits
near some islands along the Guangdong coast, and
no collections of it have been seen from mainland
China. On the other hand, S. matsumurae is dis¬
tributed in Hunan, Jiangsu, and Zhejiang provinc¬
es. Sopubia stricta differs from S. matsumurae in
having entire leaves, very short pedicels that are
sometimes lacking, a glabrous inner surface of the
calyx, lanceolate calyx lobes, and scabrous cap¬
sules a little longer than the calyx lobes and with
a flat apex. In contrast, S. matsumurae has leaves
with 1 or 2 teeth on each side, pedicels 1.5—2 mm
long, a scabrous calyx with triangular-lanceolate

lobes, and hairy capsules distinctly longer than the
calyx and with a curved apex.

Xizangia bartsioides (Handel-Mazzetti) C. Y. Wu
& D. D. Tao, comb. nov. Basionym: Pterygiella
bartsioides Handel-Mazzetti, Sitz. Akad. Wiss.
Wien 60: 186. 1923. TYPE: China. NW Yun¬
nan: “vallis Doyon-lumba ad fluvium Salween
sub jugo Pongatong, 28°9', s. micoschistaceo,”
3450 m, 4 Aug. 1916, Handel-Mazzetti 9671
(holotype, WU).

Xizangia serrata D. Y. Hong, Acta Phytotax. Sin. 24: 141.
1986. Syn. nov. TY PE: China. Xizang (Tibet): Bomi,
3100-3400 m. 17 Aug. 1983, B. S. Li, /. C. Ni &
S. /. Clieng 66 01 (holotype, PE).

The study of authentic material identified by
Hong De-yuan as Xizangia serrata and the rich col¬
lections of Pterygiella bartsioides from Yunnan re¬
veals that the two taxa are conspecific. I agree with
Hong (1986, 1998) that Xizangia D. Y. Hong is a
genus distinct from Pterygiella Oliver and Phther-
iospermum Bunge ex Fischer & C. A. Meyer. Xiz¬
angia can be readily distinguished from Pterygiella
by having serrate instead of entire leaves and by
lacking bracteoles instead of having two below the
calyx. Xizangia differs from Phtheriospermum by
having serrate instead of pinnatiseet leaves and an
unequally instead of equally parted calyx.

Acknowledgments. I am grateful to Peng Hua
for his help in drafting the manuscript and to Ihsan
Al-Shehbaz for his editorial assistance.

literature Cited

Hong, D. Y. 1986. A new genus of Scrophulariaceae from
Xizang, China. Acta Phytotax. Sin. 24: 139-142.

- . 1998. Xizangia. In: C. Y. Wu & P. H. Raven,

flora of China, Vol. 18: 97. Science Press, Beijing, and
Missouri Botanical Garden, St. Louis.

Yamazaki, T. 1980. New or noteworthy plants of Scrophu¬
lariaceae from Indo-China (4). J. Jap. Bot. 55: 1-13.

Novon 9: 288. 1999.

Volume 9, Number 2, pp. 127—288 of NOVON was published on 26 May 1999.

Volume 9
Number 3
1999

NOVON

Notes on Malesian Fabaceae (Leguminosae-Papilionoideae) 4.
Millettia chrysamaryssa , a New Species from Peninsular Malaysia

Frits Adema

Rijksherbarium/Hortus Botanicus, P.O. Box 9514, 2300 BA, Leiden, The Netherlands

Abstract. A new species, Millettia chrysamar¬
yssa Adema from Peninsular Malaysia, is de¬
scribed.

During the preparation of an account of Millettia
Wight & Arnott for Flora Malesiana, a new species
was found in the material identified as Millettia ser-
icea (Ventenat) Wight & Arnott.

Millettia sericea, although restricted to Sumatra,
Peninsular Malaysia, and Java, has always been
seen as a variable species. Various forms and va¬
rieties have been described (Miquel, 1855, 1860;
Prain, 1897; Dunn, 1912). H owever, these forms
differ only slightly among each other, and the need
to discriminate them has not been substantiated.
A significant group of plants, differing in indu¬
mentum, flowers, fruits, and seeds, has consis¬
tently been overlooked. These plants, all from
Peninsular Malaysia, are described here as a new
species.

Millettia chrysamaryssa Adema, sp. nov. TYPE:
Malaysia. Perak: Padang, Selim river, 12 Apr.
1978, Maxwell 78-107 (holotype, L: isotypes,
AA, SINU not seen). Figure 1.

Liana. Axes brunneo-velutini. Folia 5— 1 5-foliolata. Fo-
lioli elliptici ad anguste obovati raro anguste ovati infra
dense sericei aurei vel argentei nitentes. Stipellae desunt.
Pseudo-racemi axillares. Vexillum sine oallis. Stamina
monadelpha vel diadelpha teneslris basalibus. Legumina
applanata late ligulata extus brunneo-velutina. Semina ap-
planata oblonga.

Lianas. Twigs terete, 3-9 mm diam., dark brown
velvety. Leaves 5— 15-foliolate (mostly 13). Stipules
triangular, ca. 5X3 mm, caducous. Petioles 4—

10.5 cm long, terete, brown velvety; rachis 5—23.5
cm long, terete, brown velvety; pulvinus 7-20 mm
long. Stipels absent. Leaflets: terminal elliptic to
narrowly obovate, 7—19 X 2-5 cm, length:width ra¬
tio 2.5— 3.8, base cuneate to rounded, apex acumi¬
nate, acumen 4—20 mm long, above glabrous, below
densely silvery or golden sericeous, midrib slightly
sunken above, nerves flat or slightly sunken above,
7—11 per side, 5—25 mm apart; lateral mostly as
the terminal, narrowly ovate to narrowly obovate,
4-19 X 1—5 cm, length:width ratio 2.5— 1.2; pul¬
vinus 5—13 mm long. Inflorescences axillary, pseu¬
doracemes, often apieally clustered, or pseudopan¬
icles, 18—37 cm long, peduncle 3—13 cm long,
brown velvety. Bracts to the brachyblasts ovate,
2.5—3 X 1.2-2 mm, adaxially velvety, abaxially gla¬
brous. Brachyblasts 5— 10-flowered. Bracts to the
flowers broadly ovate, 1.7— 1.9 X 1.1— 1.2 mm,
adaxially velvety, abaxially glabrous. Pedicels 5-9
mm long. Bracteoles at the base of the calyx, ovate,
1.2— 1.9 X 0.6— 1.2 mm, adaxially velvety, abaxially
glabrous. Calyx 5—7 min long, tube 4—5 mm long,
adaxially velvety, abaxially glabrous, slightly
obliquely bilabiate, “lower lip” three lobes some¬
what longer than upper two lobes. Corolla papilio¬
naceous, white or pink, or yellowish white and vi¬
olet at the base, standard often with a light green
blotch abaxially near the base. Standard: ( law 2.5—
4 mm long, adaxially sericeous; blade orbicular,
10—13 X 9—14 mm, emarginate, adaxially seri¬
ceous, abaxially sericeous at the very apex, basally
with auricles ca. 1 mm long, callosities absent.
Wings: claw 4—5 mm long; blade narrowly ovate to
elliptic, 9—11 X 3.5— 4.5 mm, with some hairs at
the base and auricles, dorsal auricles 0.5-1 mm

Novon 9: 289-291. 1999.

290

Novon

Figure I. Millettia chrysamaryssa Adema. — a. Flowering branch. — b. Flower. — c. Fruit. — d. Seed, a, b from
Maxwell 7H-I07 ; c, d from FRI 2040.

long. Keel petals: claw 4—5 mm long, with very few
hairs; blade ± boat-shaped, 9—11.5 X 4.5—6 mm,
adaxially sericeous apically, with some hairs at the
base, abaxially sericeous along the ventral side in
apical part. Stamens mono- or diadelphous, with
basal lenestrae, 13—15 mm long, free for 3—8 mm,
glabrous; anthers 0.7-1. 1 X 0.4-0. 6 mm, glabrous.

Disc indistinct. Ovary 5—9 mm long, sericeous;
ovules 3—6; style 5—6 mm long, sericeous in lower
part; stigma terminal, capitate. Pod flat, broadly
strap-like, ca. 14 X 5 X 0.4 cm, densely dark
brown velvety adaxially. Seeds upright-oblong, ca.
24 X 16 X 2 mm, shiny black; hilum ca. 2 mm
long.

Volume 9, Number 3
1999

Adema

Millettia chrysamaryssa from Malaysia

291

Table 1. Differences

between Millettia chrysamaryssa Adema and M.

sericea (Ventenat) Wight & Arnott.

M. chrysamaryssa

M. sericea

Indumentum

velvety

short-sericeous

Stipules

5 X 3 mm

2—3.5 X 2-3 mm

Pedicels

5—9 mm long

4 mm long

Stamens

monodelphous or diadelphous

diadelphous

Free part of filaments

3-8 mm long

2—4 mm long

Anthers

0.7—1 . 1 X 0.4— 0.6 mm

().(>— 0.7 X 0.4— 0.6 mm

Pods

flat. 14 X 5 X 0.4 cm

thick, 5—7.5 X 1.5— 3.5 X 1-2 cm

Seeds

Hat. upright-oblong. 24 X l(i X 2 mm

thick, transverse-ellipsoid

25—45 X 10-25 X ‘I— 13 mm

Distribution. Malesia: Peninsular Malaysia.

Habitat and ecology. Secondary forest, forest
edge, roadside. Soil: sand, shale. Altitude: 30—1600
m. Flowering: March to November; fruiting: March
to July.

Note. Named chrysamaryssa (= a golden spar¬
kle), due to the manner in which the lower surfaces
of the leaflets reflect light.

Paratypes. MALAYSIA. Kedah: K of O. Bongsu,
Bongsu E.R.. 300 m, 20 Apr. 1070. PHI 14159 (M. Ev¬
erett) (A, K). Penang: Pulau Penang, Western Road Cem¬
etery, Apr. 1018, SF 3446 (Md. Man iff) (k. SI [\ (i); Tit i
Kerawang Hill, ea. 300 m, 06 Apr. 1068, Sidek bin kiah
246 (A, K. L. SING). Trengganu: summit leading eastern
ridge G. Mandi Angin, ea. 1600 m, 12 July 1968. FH!
12094 (T. C. \\ hitmore) (A, k. I. SING). Pahang: Eraser's
Hill. Gap Road. 100 m. 17 Apr. 1055, 1L Purseglove
4127 (A. k. L. LAE): Eraser’s Hill. 21 mi. near Gap on
the way to Raul). 28 June 1971. kb 4012 (L. E. Teo &
P.) (K, L, SING); Eraser's Hill, upon Selangor border,
1300—1450 m. 16-30 Sep. 1022. /. II. Hurkill & R. E.
Holltum 866 2 (k. SING): Balok E.R.. 03 Sep. 1000. FRI
2090 (k. M. koehumen) (k. L. SING). Selangor: Gading.
Mav 1801. D. ./. A. Hervey s.n. (k. SING); (fading E.R..
180 m. 20 July 1060. FRI !4485{\\. S. Lob) (k. L. SING);
47th mile (Jap Road. Rawang. 08 July 1066. FRI 2040
(k. M. Koehumen) (A. k. L. SING): Sedang experimental
plantation, Sep. 1925, Ixwibourne s.n. (SING); Kuala
Lumpur, near Universiti Malaya campus, ea. 30 m. 23
Nov. 1978. II. C. Stone 14002 (BO. k. L); l niversity cam¬
pus, near the entrance gate from Petaling Java. Jan. 1964,
Abd. Samat bin Abdulla 252 (k): kepong ERIM, 28 May
1963. Md Kasim bin Rajah 1514 (k); kepong, near
E.R.O’s house, II Oel. 1932. FRI 29102 (A. B. Walton)
(k); Melaka. 19 Mar. 1885. 1 1vins 1081 (SING); Bukit
kupayang, 07 July 1885, \lrins 1961 (SING). SINGA-
PORE. Singapore. 07 July 1061. hthb 410 (A. k).

Differences Between Millettia sericea and
Millettia ciir \ sam t r i .ss t

At first glance, Millettia sericea and lln1 new spe-
cies M. chrysamaryssa appear strikingly different
in the way the abaxial surfaces of the leaflets reflect

light: mostly silvery1 and more or less dull in M.
sericea , mostly golden and more shiny in M. chry¬
samaryssa. However, this character, which is rather
difficult to define, shows more overlap than origi¬
nally was thought. More convincing differences are
found in the pods and seeds: flat in M. chrysamar¬
yssa, thick in M. sericea (see Table 1).

In the key to the sections of Millettia in Dunn’s
(1912) monograph, Millettia chrysamaryssa keys
out as a member of section Macrospermae Dunn.
With this section it shares the following characters:
lianas; stipels absent; inflorescences pseudora¬
cemes; petals sericeous with standards lacking cal¬
losities; disc indistinct. However, M. chrysamaryssa
differs from the other species of section Macrosper¬
mae in its flat pods and seeds and its mono- or
diadelphous stamens. The brilliant sheen of the
lower leaflet surfaces of M. chrysamaryssa is rem¬
iniscent of the Indian species M. splendens Wight
& Arnott and M. rubiginosa Wight & Arnott, both
belonging to section Typicae Dunn. Sections Ma¬
crospermae and Typicae differ chiefly in the pres¬
ence or absence of stipels.

Acknowledgments. The loan of specimens from
the herbaria of A, BO, K, L, LAE, and SING is
gratefully acknowledged. J. F. Veldkamp kindly
translated the diagnosis into Latin. J. J. Wessendorp
prepared the drawing. J. F. Maxwell, the collector
of the type of the new species, was so kind to in¬
dicate where the duplicates are stored.

Literature Cited

Dunn, S. T. 1912. A revision of the genus Millettia, Wight
et Am. J. Linn. Sue. Bot. 41: 123-243.

Miquel, E. A. G. 1855. El. Ind. Bat. 1. Van der Post.
Amsterdam/Utrecht.

- . I860. FI. Ind. Bat. Suppl. I. Prod. El. Sumatra.

Van der Post, Amsterdam/Utrecht.

Prain, I). 1897. 26. Millettia W. & A. J. Asiat. Soe. Bengal
66: 86-94.

A New Species of Solarium subgenus Leptostemonum (Solanaceae)
from Chapada da Diamantina, Bahia, Brazil

Maria tie Fdtima Agra

Laboratorio de Tecnologia Farmaceutica, Universidade Federal da Parafba, Gaixa Postal 5009,
58.049-430 Joao Pessoa, Parafba, Brasi I . fatima.agra@openline.com.br

Abstract. A new species, Solarium diamanti¬
nense, is described and illustrated. It is endemic
to Brazil and restricted to the Chapada Diaman¬
tina in the State of Bahia, northeastern Brazil.
This species is closely related to the Brazilian
Solarium baturitense Huber, from which it differs
by its oblaneeolate leal blades, short calyx lobes,
anil dense, stalked, glandular-stellate pubes¬
cence.

Resumo. Uma nova especie, Solatium diamanti-
nense, € descrita e ilustrada e sua distribufeao e
sua afinidade sao discutidas. Especie endemica do
Brasil, restrita it Chapada da Diamantina, estado
da Bahia, nordeste do Brasil. Solatium diamanti-
nense possui afinidade com S. baturitense Huber,
da qual pode se distinguir pela lamina foliar ob-
lanceolada, calice com lobos curtos e a densa pu-
bescencia com tricomas estrelados e glandular-es-
trelados pedicelados.

Solarium L. is the largest and most complex ge¬
nus of the Solanaceae family, with about 1500 spe¬
cies and 5000 published epithets. It is distributed
mainly throughout the tropical and subtropical re¬
gions of the world and has its center of diversity
and distribution in South America.

Vi hile preparing a taxonomic revision of the So¬
larium erythrotrichum group sensu Whalen (1984)
of the subgenus Leptostemonum Dunal, a new spe¬
cies was identified on the basis of its distinctive
morphological characteristics and is herein named
Solan urn diamantinense.

Members of the subgenus Leptostemonum Dunal
Erythrotrichum group sensu Whalen (1984) form a
South American distinctive, diverse, and presum¬
ably monophyletic group that comprises species
possessing ferruginous to reddish glandular-stellate
pubescence, branched inflorescences, attenuate
and subulate anthers, somewhat accrescent calyces,
and glandular-pubescent berries with a leathery
epicarp at maturity.

Solatium diamantinense M. F. Agra, sp. nov.
TYPE: Brazil. Bahia: Mun. Seabra, a 2 km de
Seabra, na diregao de Campestre, altitude 930
m, vegetagao area de contato ecologico (caa-
tinga/floresta estacional), 3 Apr. 1998 (fl, fr),
M. F. Agra & A. M. Giulietti 5176 (holotype,
JPB; isotypes, JPB. JPB, HUEFS, M0, SPF,
UPS). Figure 1.

Frutex ad ramos juniores folios inflorescentiamque
trichomatibus stellatis glandulosis ferrugineis lomen-
tosus, ad caulem foliosque aculeis rufescentibus basi
dilatatis compressis apice leniter reeurvis armatus. Fo¬
lia alterna petiolata lanceolata apiculata basi acuta
marginibus integerrima utrinque tomentosa. Inflores-
centia ex cyma terminali hi- vel trichotoma multiflora
constans. Flos calyce breviusculo campanulato-scutel-
lari. laciuiis brevissimis tomentoso-ferrugineis; corolla
alba in lacinias lanceolatas intus glabras extns stellato-
tomentosas profunde part i ta. lobulis sub anthesi reflex-
is; antheris aequalibus 0.5-0. 6 cm longis lanceolalis
red is; ovario depresso-globoso piloso; stylo credo.
Bacca globosa glanduloso-pilosa diametro 1.5— 2.0 cm:
seminibus 4-5 mm longis, lenticularibus, subrenifor-
mibus, alutaceis.

Small erect, branched, perennial shrubs 0.4-2
m tall. Stems terete, reddish or ferruginous tomen-
tose with glandular-stellate trichomes and strongly
armed with reddish, laterally compressed prickles
that are enlarged at the base, 0.2-0. 3 cm wide,
0.2— 0.5 cm long, and slightly recurved at the
apex, longer and more numerous on the young
plants; bark of older stems glabrescent and dark
brown. Leaves alternate, usually unarmed in adult
plants; if armed, with prickles only on the midrib,
lateral veins, and petioles; blades entire, bicol¬
ored, subcoriaceous, oblaneeolate or elliptic, 4—
12 cm long, 2-3 cm wide, acute at apex, acute
and slightly attenuate at base, velutinous above
with glandular and eglandular, multiseriate-
stalked, ferruginous to reddish stellate hairs, 2-4
branches of the glandular-stellate hairs reflexed,
unarmed or sparsely armed below with small, fer¬
ruginous to brown prickles, ferruginous to yellow¬
ish tomentose with glandular-stellate hairs, the
multiseriate stalk 2— 3 times longer than in hairs
of upper surface; primary veins prominent, ferru-

Novon 9: 292-295. 1999.

Volume 9, Number 3
1999

Agra

Solanum diamantinense from Brazil

293

Figure 1. Solanum diamantinense M. F. Agra. — A. Flowering branch. — B. Calyx. - — C. Stamen, front view. — D.
Stamen, lateral view. — E. Stellate glandular hair from abaxial leaf surface. — F. Seed. — G. Stellate glandular hair
from adaxial leaf surface. — H. Gynoecium. — I. Flower.

ginous; petioles short, 1—2 cm long, unarmed or
sparsely armed with small reddish to brown prick¬
les, densely tomentose with reddish, stalked, stel¬
late hairs similar to those of the lower surface.
Inflorescences terminal, 4—10 cm long, remote from
the leaves, scorpioid, hi- to tri-furcate, more than
20-flowered; peduncle subterete, unarmed, un¬
branched, 2—10 cm long, 2-5 mm thick, ferrugi¬
nous to reddish tomentose, the hairs similar to the
branches and petiole; pedicels slender and artic¬
ulate at the base, 3—5 mm long, with hairs similar
to those of the peduncle. Flowers heterostylous,
the basal ones with elongate and functional styles
exceeding the anthers, the distal ones with re¬
duced gynoecia. Calyx unarmed, subcampanulate,
slightly zygomorphic, 5-lobed about !4 the length,
the lobes ca. 1 mm long, acute to apiculate, the

tube ca. 2—3 mm long, 4 mm wide in flower, 8—10
mm long and 5—7 mm wide in fruit, glabrous with¬
in, ferruginous-tomentose without, with glandular
stellate hairs similar to those of the petiole. Co¬
rolla white, long exserted from the calyx at anthe-
sis, stelliform, 2—2.5 cm across, deeply parted, the
tube 2—4 mm long, the lobes linear-lanceolate, re¬
flexed at anthesis, 8—10 mm long, 2—4 mm wide,
glabrous within, the midrib prominent, tomentose
without, with ferruginous glandular-stellate hairs
similar to those of abaxial surface of the leaf
blades. Anthers equal, connivent, 5—6 mm long,
linear-lanceolate, attenuate, subulate and introrse
at the apex, slightly asymmetric at the base; fila¬
ments glabrous, the free part 1—2 mm long. Ovary
depressed globose, ca. 2 mm long, densely villose
with stellate hairs, the long branches of the stellae

294

No von

Figure 2. Distribution of Solatium diamantinense VI. F.
Agra.

2—3 mm long. Style glabrous, straight or curved at
the apex, pubescent at base, exceeding the sta¬
mens, 8—10 mm long in functional flowers, 4—5
mm long in staminate flowers. Stigma clavate-
truncate, dark green, shiny, minutely papillose at
the tip. Fruits globose, held upright, 1.5— 2.0 cm
diam., with leathery epicarp, persistently ferrugi¬
nous pubescent with glandular-stellate hairs
mixed with short anti long unbranched gland-
tipped hairs, the calyx persistent and somewhat
accrescent. Seeds lenticular, subreniform, brown,
4—5 mm long, 3—3.5 mm wide, the testa with fine
reticulated ornamentation.

Distribution and habitat. Solanum diamanti¬
nense is endemic to Brazil and known only from
collections from the Morro do Chapeu and Seahra.
State of Bahia. It grows in secondary vegetation as¬
sociated with “campos gerais” at elevations of 900-
1200 rn. These areas are located in the mountain
zone known as the Chapada Diamantina (Fig. 2),
which is one of the richest centers of diversity of
plant life in Brazil and contains some of the rarest

species. It is characterized by a complex mosaic of
different types of vegetation (Harley, 1992). The
“campos gerais” is a feature of large tracts of up¬
land Bahia around Morro do Chapeu. It occurs on
deeper soils over flat or slightly undulating ground
and contains many typical cerrado species, the veg¬
etation being dominated by grasses and low shrubs
(Harley, 1995).

Affinities. Solanum diamantinense is similar
and probably closely related to S. baturitense Hub¬
er, another Brazilian species. It can he distin¬
guished from .S’, baturitense by its oblaneeolate to
elliptic leaves, very short calyx lobes, and dense,
stalked, glandular-stellate pubescence on all parts
ol the plant, especially both sides of the leaves.
Both species share branched inflorescences, lan¬
ceolate-subulate anthers, reddish to ferruginous pu¬
bescence, white flowers, and slightly accrescent ca¬
lyces. This set of morphological characteristics
makes S. diamantinense a very distinctive species
in the informal Erythrotrichum group of subgenus
Leptostemonum.

Paratypes. BRAZIL. Bahia: Mini. Seahra. a 12 km
ile Seahra. na direyao de Campestre, altitude 930 m,
vegetayao area de contain eeologico (eaatinga/floresta es-
tacional), 15 Nov. 1983 (fl, f r), ./. A. C. I.ima el al. 246
(HR B); Mun. Morro do Chapeu, lat. II°38'34"S,
40°55'45"W, vegetayao inistura (savana/estepe), latosso-
lo. 20 Aug. 1980 (II. fr). H. P. Bautista 355 (HUB); Mun.
Morro do Chapeu, lat. 1 1°31'S. long. 40°18'W, vegetayao
mistura (Savana/Kstepe). 5 Apr. 1984 (fl. fr), 0. A. Sal-
gado & II. B. Bautista 254 (HUB); Mun. Morro do Cha¬
peu, elev. ea. 1030 m, 3 km S de Morro do Chapeu.
eaminho de Utinga, aprox. 1 1°35'S. 41°1I'W. earrasco
on eaatinga de altitude, 28 Nov. 1992 (fl. fr). \1. M. Arbo
et al. 5358 (C'l’FS not seen. CH. JPB (2), MO. BB. SI’.
Sl’F); Mun. Morro do Chapeu. ea. 4 km SW of the town
of Morro do Chapeu on the road to lltinga. scrubby
“campos gerais” vegetation, alt. 1150 m I l°35'S,
trU'W. 2 June 1980 (fl). K. M. Ilarley et al. 22991
(CFPKC. F. KB): Mun. Seahra, 13 Ago. 1956 (II. fr). E.
Pereira 2162 (KB (3)); Mun. de Uagoinha, 18 km of La-
goinha, 5.5 km SW ol Delfino on side road to Minas do
Mimoso, disturbed woodland, secondary vegetation by
cultivation, and marsh by small stream, all. 950 m.
aprox. 10°20'S, 41°20W, 7 Mar. 1974 (fl). B. M. Harley
et al. 16943 (CFPKC. IPA. k. M. MO. KB. US); Mun.
Utinga, beira da estrada BA 142. proximo a Utinga
12°28'00"S. 41°27'(M)"W, all. 1000 m. 10 Mar. 1996 (fl).
K. Lima et al. 2215 (HUFFS); Mun. do Morro do Chapeu.
Cachoeira do Ferro Doido, 20 km SK da cidade, alt. ca.
900 m s.m., 19 Nov. 1986 (fl). L. P. Queiro z et al. 1305
(HUFFS); Mun. do Morro do Chapeu, 1-2 km sul da
cidade na estrada para Utinga. I I °33'00"S, 41°09'00"W,
16 Nov. 1984 (fl). L. R. Noblick 3497 (HUFFS. NY).

Acknowledgments. The author sincerely thanks
Lynn Bohs for valuable suggestions and revision of
the text; Ana Maria Giulietti for helpful comments
and suggestions; Simone Cabral for the illustra¬
tions; Raymond Harley for field observations; and

Volume 9, Number 3
1999

Agra

Solanum diamantinense from Brazil

295

John J. Pipoly III and Joseph Miller for revising the
English text.

Literature Cited

Harley, If. M. 1992. Plant diversity: Kew's role in north¬
east Brazil. Kew Mag. 9(3): 103—1 10.

- . 1995. Introduction. Pp. 1—42 in B. L. Stanaard

(editor). Flora ol the Pico das Almas. Royal Botanic
Gardens, Kew.

Whalen, M. I). 1984. Conspectus of species groups In
Solanum subgenus Ijeptostemonum. Gentes Herb. 1 2(4):
179-282.

Generic Placement of Species Excluded from Arabidopsis

(Brassicaceae)

Ihsan A. Al-Shehbaz

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Steve L. O' Kane, Jr.

Department of Biology, University of Northern Iowa, Cedar Falls, Iowa 50614-0421, U.S.A.

Robert A. Price

Department of Botany, University of Georgia, Athens, Georgia 30602, U.S.A.

ABSTRACT. All 59 binomials previously assigned
to Arabidopsis are critically evaluated and placed
in 14 genera, ol which Crucihimalaya, Olimarabi-
dopsis, and Pseudoarabidopsis are described as
new. Nine new combinations in Crucihimalaya,
three in Olimarabidopsis, and one in Pseudoarabi¬
dopsis, as well as 12 new synonyms, are proposed.
A key to all genera most commonly confused with
Arabidopsis, including the three new proposed here¬
in, is presented.

Generic delimitation in the Brassicaceae (Cru-
citerae) is one of the most difficult and often con¬
troversial aspects in the systematics of the family
(Al-Shehbaz, 1973, 1984; Rollins, 1993; Schulz,
1936). Perhaps the two principal reasons for this
are: (1) Convergence in basically every morpholog¬
ical character is so high that superficially very sim¬
ilar genera might well turn out to be remotely re¬
lated or unrelated upon critical examination of
so-called key generic characters and independent
assessment of phylogenetic relationship using mo¬
lecular comparisons (as in the genera herein seg¬
regated from Arabidopsis). (2) Although the family
exhibits tremendous diversity in fruit morphology,
other parts of the plant, especially the flowers, usu¬
ally do not show much diversity and, therefore,
then* are few morphological characters that can be
used to determine relationships. Characters of flow¬
ers and vegetative parts have often been ignored or
overlooked. Because fruit morphology has tradition¬
ally been used in taxonomic treatments of the Bras¬
sicaceae, the problem becomes more acute among
the numerous genera with relatively similar linear¬
shaped fruits. In cases like these, vegetative or flo¬
ral characters could easily be of greater signifi¬
cance than fruit or seed characters in delimiting
natural genera.

Given the great interest in Arabidopsis thaliana
(L.) Heynhold as a model experimental organism,
it is of particular value to define clear phylogenetic
groupings among its related genera. The limits of
Arabidopsis (DC.) Heynhold have been the subject
of continuous controversy, and many authors (e.g.,
Al-Shehbaz, 1988; Hedge (in Hedge & Rechinger),
1968; Price et ah, 1994) called for the need to
establish well-defined boundaries between the ge¬
nus and its relatives. Although O’Kane and Al-
Shehbaz (1997) retained only nine species in Ara¬
bidopsis, the generic placement of 50 of the 59
binomials previously assigned to Arabidopsis re¬
mained to be established. The present paper ad¬
dresses this problem, and keys for the determina¬
tion of taxa most often confused with Arabidopsis
are provided.

Molecular comparisons of both chloroplast DNA
(Price et ah, 1994, unpublished) and nuclear In¬
ternal Transcribed Spacer (ITS) regions (O'Kane et
ah, 1995, 1997, unpublished) have consistently
supported dividing the core group of the broadly
circumscribed Arabidopsis (e.g., table 1 of Price et
ah, 1994) into a small number ol well-separated
clades, most notably Arabidopsis sensu stricto (in¬
cluding Hylandra A. Love and Cardaminopsis (C.
A. Meyer) Hayek; see O’Kane & Al-Shehbaz, 1997;
Mummenhoff & Hurka, 1995), A. pumila (Stephan)
N. Busch and relatives (here newly described as
Olimarabidopsis), and .4. himalaica (Edgeworth) 0.
E. Schulz and relatives (here newly described as
Crucihimalaya). All of these genera belong to a ma¬
jor terminal elade, including a number of other
Eurasian and American genera such as Capsella
Medikus, Neslia Desvaux, Erysimum L., Malcolmia
R. Brown, and Halimolobos Tausch. Several other
species sometimes placed in Arabidopsis have been
found to belong to the distantly related genera Thel-

Novon 9: 296-307. 1999.

Volume 9, Number 3
1999

Al-Shehbaz et al.

Exclusion from Arabidopsis

297

lungiella 0. E. Schulz (Al-Shehbaz & O’Kane,
1995; Galloway et al., 1998) and Neotorularia
Hedge & J. Leonard (Al-Shehbaz & O’Kane, 1997),
or in one case to the genus Erysimum (Al-Shehbaz,
1994), which is closely related to Olimarabidopsis.
Although the taxonomy oi Arabidopsis has now been
worked out at the generic level (O’Kane & Al-Sheh¬
baz, 1997), our research indicates that Arabis is
polyphyletic and consists of at least three unrelated
clades.

Following up the suggestions of new phylogenet¬
ic groupings provided by molecular comparisons,
we have thoroughly reexamined the morphology of
the species previously placed in Arabidopsis in or¬
der to reassess morphological groupings of species
and to try to find morphological characters distin¬
guishing the groups indicated by molecular com¬
parisons. Over the last seven years, we have criti¬
cally examined more than 6000 specimens from
numerous herbaria. We have found that differences
in fruit morphology (terete vs. flattened) and seed
morphology (incumbent vs. accumbent cotyledons
and winged vs. unwinged seeds), which have been
previously used (e.g., Busch, 1909; Ball, 1993;
Jones, 1964; Mulligan, 1995; Rollins, 1993;
Schulz, 1936) to separate the traditionally circum¬
scribed genera Arabidopsis and Arabis, appear to be
very unreliable in the delimitation of natural ge¬
neric groups. Seven of the nine species of Arabi¬
dopsis sensu stricto have flattened fruits and ac¬
cumbent cotyledons, while two have terete fruits
and incumbent cotyledons (O’Kane & Al-Shehbaz,
1997). In contrast, differences in trichome branch¬
ing, flower color, and nature of the cauline leaf base
appear to be much more useful in defining natural
generic groups among species previously placed in
Arabidopsis sensu lato. Combinations of these char¬
acters, along with molecular phylogenetic data,
support the retention of 9 species in Arabidopsis
and the segregation of 13 species commonly placed

in the genus into three new, well-defined genera
herein proposed as Crucihimalaya, Olimarabidop¬
sis, and Pseudoarabidopsis. Based on ehloroplast
DNA sequencing (Price, unpublished), Olimarabi¬
dopsis is most closely related to Erysimum, and
both are readily separated from Arabidopsis by hav¬
ing yellow or orange (rarely cream or purplish) flow¬
ers and malpighiaceous and/or sessile stellate tri-
chomes ( Erysimum ) or submalpighiaceous and
subsessile stellate trichomes ( Olimarabidopsis ).

Only 3 of the 14 genera to which the 59 Arabi¬
dopsis binomials belong are not included in the fol¬
lowing key. These, Murbechiella Rothinaler, Sis-
ymbriopsis Botsehantsev & Tzvelev, and Robeschia
Hochstetter, are unrelated to Arabidopsis. Murbeck-
iella has auriculate cauline leaves, keeled valves,
veined septa, and winged seeds, whereas Arabidop¬
sis has petiolate cauline leaves, rounded or flat
valves, veinless septa, and wingless seeds. Sisym-
briopsis has pubescent, quadrangular fruits and
prominently 3-veined valves, whereas Arabidopsis
has glabrous, terete or flattened fruits and veinless
or obscurely 1-veined valves. Finally, Robeschia
has dendritic trichomes, 2-pinnatisect or 2-pinnate
leaves, much thickened fruiting pedicels as thick
as the fruit, and an obsolete style, whereas Arabi¬
dopsis has simple and stalked forked trichomes, un¬
divided to pinnatifid leaves, slender fruiting pedi¬
cels narrower than the fruit, and distinct styles.
These three genera have not yet been subjected to
molecular studies, but should be analyzed in the
near future.

The circumscriptions of Arabis and Halimolobos
in the following key follow that of Rollins (1993).
We are, however, aware that these genera, as pres¬
ently delimited, represent very heterogeneous as¬
semblages of species groups that will have to be
re-assigned to other genera, most of which have
already been proposed. We are currently working
on these groups.

ARTIEICIAL Kl'.'t TO THK GENERA WITH MEMBERS FORMERLY PLACED IN ARAHII)OrsiS SeNSI LaTO

la. Plants completely glabrous; leaves and stems glaucous; plants often restricted to strongly saline and/or

calcareous soil . Thellungiella

lb. Plants sparsely to densely hairy; leaves and stems not glaucous; plants usually on other soil types.

2a. Trichomes sessile and completely appressed, malpighiaceous and/or stellate with unbranched rigid

straight rays . Erysimum

2b. Trichomes short- or long-stalked, simple or branched, if stellate and sessile then rays slender and/or
branched.

3a. Scapose annuals without cauline leaves; fruiting pedicel nearly as thick as fruit . Drabopsis

3b. Nonseapose annuals, biennials, or perennials with few to many cauline leaves; very rarely peren¬
nials without cauline leaves; fruiting pedicels much narrower than fruit (except some Neotoru¬
laria).

4a. Fruits compressed; cotyledons accumbent.

5a. Cauline leaves short petiolate. neither auriculate nor sagittate at base; trichomes sim¬
ple and 2- or 3(or 4)-forked, never dendritic or stellate; fruit valves with a prominent
midvein; seeds usually wingless . Arabidopsis

298

Novon

5b. Cauline leaves auriculate to sagittate at base, rarely petiolate; at least some of leaf
trichomes dendritic or stellate; fruit valve without or with obscure midvein; seeds often

winged . Arabis

4b. Fruits usually terete or 4-angled; cotyledons incumbent.

6a. Inflorescence axis flexuous; leaves divided into 3 or 5 filiform to narrowly linear seg¬
ments . lanhedgea

6b. Inflorescence rachis not flexuous; leaves various but never divided into filiform or
narrowly linear segments.

7a. Flowers yellow.

8a. Cauline leaves petiolate; fruit apex strongly recurved or contorted; fruiting

pedicels stout, nearly as thick as fruit . Neotorularia

8b. Cauline leaves auriculate, rarely sessile; fruit apex neither recurved nor

contorted; fruiting pedicels slender, narrower than fruit .... Olimarabidopsis
7b. Flowers white, lavender, or purple, very rarely creamy white.

9a. Fruits glabrous.

I()a. Cauline leaves petiolate; branched trichomes forked, rays always sim¬
ple-

1 la. Fruiting pedicels slender, narrower than fruit; seeds mucilagi¬
nous when wetted; fruits straight . \rabidopsis

lib. Fruiting pedicels stout, nearly as thick as fruit; seeds not mu¬
cilaginous when wetted; fruits often twisted . Neotorularia

10b. Cauline leaves sessile, often auriculate, sagittate, or amplexicaul, if
short petiolate then plants eanescent; at least some of the trichomes
dendritic or stellate with some branched rays.

12a. All branched trichomes sessile; petals (6— )6.5— 8(— 9) mm; fruit

short stipitate . Pseudoarabidopsis

12b. At least some of the branched trichomes distinctly stalked; pet¬
als 1 .5 — 4( — S) mm; Iruit sessile.

13a. Seeds biseriate; North America . Halimolobos

13b. Seeds uniseriate; Asia . Crucihimalaya

9b. Fmit pubescent.

14a. Inflorescence bracteate at least on lower half . Crucihimalaya

14b. Inflorescence ebracteate.

15a. Fruit with submalpighiaceous or short-stalked to subsessile

stellate trichomes; septum lacking or perforated Olimarabidopsis

15b. Fruit with other triehome types; septum complete.

16a. Seeds not mucilaginous when wetted, uniseriate; fruit of¬
ten strongly torulose and/or twisted; stems often decum¬
bent; predominantly Old World . Neotorularia

16b. Seeds mucilaginous when wetted, mostly biseriate; fruit
often smooth; stems not decumbent; exclusively New
World . Halimolobos

Critcihiinalaya Al-Shehbaz, O'Kane & Price, gen.
nov. TYPE: Crucihimalaya himalaica (Edge-
worth) Al-Shehbaz, O’Kane & Price.

Folia eaulina sessilia vel subsessilia, auriculata vel raro
nonauriculata; pi 1 i ramosi stipilati stellati saepe pilis sim-
plicibus vel fureatis praesentibus; raeemi bracteati vel
ebraeteati, valde elongati; sepala oblonga, nonsaccata; pe-
tala alba vel purpurea; fructus lineares, teretes, saepe gla-
bri; stipitum nullum; septum completion; semina uniser-
iata vel raro biseriata, ( 1 5— )20— 60(— 75) per locula,
oblonga, mucilaginosa; cotyledones incumbentes.

Herbs annual or biennial, rarely perennial with
a eaudex. Trichomes stalked, 1- or 2-forked, often
mixed with simple and/or stellate ones, never den¬
dritic. Stems erect to ascending, simple or
branched basally and/or apically. Basal leaves ro-
sulate or not, simple, entire or dentate, rarely lvrate
or pinnately lobed. Cauline leaves sessile or sub-
sessile, auriculate, sagittate, or rarely cuneate into

a petiole-like base, entire, dentate, or rarely pin¬
nately lobed, rarely absent and plants scapose. In¬
florescences several- to many-flowered, corymbose
racemes, elongated considerably in fruit; rachis
straight; bracts present along entire inflorescence,
or restricted to lowermost flowers, or absent. Sepals
oblong, deciduous, erect, pubescent, base of inner
pair not saccate. Petals white, lavender, or purple,
spatulate. Stamens 6, slightly tetradynamous; an¬
thers ovate or oblong, sagittate at base, obtuse at
apex. Nectar glands confluent and subtending bases
of all stamens. Ovules (30— )40— 120(— 150) per ova¬
ry. Fruit dehiscent, linear, terete or somewhat 4-
angled to rarely compressed parallel to septum;
valves with a distinct midvein, glabrous or rarely
stellate hairy or puberulent, smooth or torulose; gy-
nophore absent; septum complete; style to 1 mm
long; stigma capitate, entire. Seeds uniseriate or
rarely biseriate, (15-)20-60(-75) per loeule, wing-

Volume 9, Number 3
1999

Al-Shehbaz et al.

Exclusion from Arabidopsis

299

less, oblong, plump; seed coat minutely reticulate,
mucilaginous when wetted; cotyledons incumbent.

Crucihimalaya, which means “cross Himalaya,”
is readily distinguished from Arabidopsis by having
at least some stalked stellate trichomes (these
sometimes mixed with simple or forked one s), ses¬
sile and auriculate to sagittate cauline leaves rarely
subsessile into a petiole-like base, and braeteate or
ebracteate inflorescences. Arabidopsis always has
stalked forked trichomes, petiolate cauline leaves,
and ebracteate inflorescences. Four species of Cru¬
cihimalaya (C. axillaris , C. himalaica, C. lasiocar-
pa, and C. stricta ) have inflorescences braeteate at
least on the lower half, and most of the remaining
species often have the lowermost flower braeteate.
However, this character can be very variable, and
the presence of stellate trichomes and/or auriculate
to sagittate cauline leaves should help m the sep¬
aration of such species from Arabidopsis.

Key to the Species oe Cm cihi\iaiay\

la. Fruit valves densely and coarsely stellate; pedi¬

cels pubescent all around; fruits often subap-
pressed to racliis . 4. C. lasiocarpa

lb. Fruit valves glabrous or very' rarelv puberulent;
pedicels glabrous adaxially, rarely (C. wallichii)
pubescent all around; fruits not appressed to ra-
cbis.

2a. Cauline leaves distinctly auriculate or am-
plexicaul at base.

3a. Inflorescences braeteate at least along
the lower portion; annuals or biennials

. 6. C. himalaica

3b. Inflorescences ebracteate; perennials.

4a. Fruits divaricate; plants biennial;
adaxial surface of lower cauline
leaves predominantly with forked

trichomes . 2. C. ovezinnikovii

4b. Fruits erect; plants perennial; ad¬
axial surface of lower cauline
leaves predominantly with stalked
stellate trichomes . . I . C. mollissima
2b. Cauline leaves sessile or subsessile into a
petiole-like base, neither auriculate nor arn-
plexieaul. sometimes absent.

5a. Plants scapose; cauline leaves absent or

only 1; Mongolia . 3. C. mongolica

5b. Plants with well-developed stems and
several cauline leaves: Himalaya and
southwest Asia.

6a. Lowermost flowers of main inflores¬
cence braeteate; basal leaves entire
to dentate, not canescent, withering
by flowering or fruiting.

7a. Stem leaves adaxially with
simple and forked trichomes,
ovate to elliptic or oblong,
rarely oblanceolate; plants 4—

15(— 20) cm tall; main inflores¬
cence braeteate nearly through¬
out . 7. C. axillaris

7b. Stem leaves adaxially with
stellate stalked trichomes, lin¬
ear-lanceolate; plants (18— )30 —

85( — 1 20) cm tall; only lower¬
most flowers of main inflores¬
cence braeteate .... 5. C. stricta

6b. Lowermost flowers of main inflores¬
cence ebracteate; basal leaves ly-
rate to pinnatifid, often canescent,
persistent in flower and fruit.

8a. Petals (2— )2.5— 3.5(— 4.5) mm
long; stem base and petioles of
basal leaves often with simple
trichomes (0.5-)0.8-l . 5( — 2 )
mm long; base of lower cauline
leaves usually minutely auricled;
plants from Nepal and Tibet
west into Iran .... 8. C. wallichii
8b. Petals 4.5— 5.5 mm long; stem
base and petioles of basal
leaves without simple tri¬
chomes; base of lower cauline
leaves not auricled; plants of
Sinai and western Saudi Ara¬
bia . 9. C. kneuckeri

1. Crucihimalaya iiiollissima (C. A. Meyer) Al-
Shehbaz, O’Kane & Price, comb. nov. Basio-
riym: Sisymbrium mollissimum C. A. Meyer, in
Ledebour, Icon. PI. FI. Ross. Impr. Altai. II-
lust. 19. 1831. TYPE: [Russia], Altai, “Hab.
in insulis atque ripis fl. Tschuja [Chuya Riv¬
er],” C. A. Meyer s.n. (holotype, LE; Visotypes,

P. W).

Distribution. Afghanistan, China, India, Kash¬
mir, Kazakstan, Mongolia, Pakistan, Russia (Sibe¬
ria), Tajikistan.

Reports of Crucihimalaya mollissima (as Arabi¬
dopsis) from Bhutan by Grierson (1984) and from
the Chinese provinces Sichuan and Yunnan by An
(1987) were based on misidentified plants of C.
himalaica. An also reported the species from Gan¬
su and Shaanxi, but we were unable to verify those
records.

Schulz (1924) divided the species into eight va¬
rieties, of which var. yunnanensis 0. E. Schulz is
based on an immature plant of a species of Arabis,
var. afghana 0. E. Schulz is Crucihimalaya wal-
lichii, and var. griffithiana (Boissier) 0. E. Schulz
is Arabis bijuga Watt. The remaining varieties were
based on minor continuous differences in the in¬
dumentum and leaf margin and, therefore, do not
merit recognition. None of Schulz’s eight varieties
were recognized in any subsequent taxonomic treat¬
ment.

300

Novon

2. Crueihiinalaya ovczinnikovii (Botschantsev)

Al-Shehhaz, O’Kane & Price, comb. nov. Ba-
sionym: Arabidopsis ovczinnikovii Botschan¬
tsev, in P. N. Ovczinnikov, FI. Tadzhitskoi SSR
5: 625. 1978. TYPE: Tajikistan. Badachschan,
fl. Gunt, canalis Chanif ca. urb. Chorog, 19
June 1966, R. Kamelin s.n. (bolotype, LE; iso¬
type, LE).

Distribution. Endemic to Tajikistan.

Tiie species is most closely related to and some¬
times difficult to separate from Crucihimalaya mol-
lissima. The most reliable characters that separate
the two species are listed in the key above.

3. Crucihiinalaya mongolica (Botschantsev) Al-

Shehbaz, O'Kane & Price, comb. nov. Basio-
nym: Arabis mongolica Botschantsev, Bot.
Zhurn. 60: 947. 1975. TYPE: Central Mon¬
golia. Gobi of Altai Mt., Bain Tzagan, 4 Aug.
1931, N. P. Ikonnikov-Galitzky & V. A. Ikon-
nikova-Galitzka 3805 (bolotype, LE).

Distribution. Endemic to Mongolia.
Crucihimalaya mongolica is most closely related
to C. mollissima ; it is readily distinguished from
this and the remaining species of the genus by the
scapose inflorescences and the lack of or presence
of a single non-auriculate cauline leaf. The species
has incumbent cotyledons, and Botschantsev’s
(1975) original placement of the species in Arabis
is erroneous because this genus always has accum-
bent cotyledons. MGsifek and Sojak (1995) were
correct in associating the species with C. mollissi¬
ma.

4. Crucihimalaya lasiocarpa (J. D. Hooker &

Thomson) Al-Shehbaz, O’Kane & Price, comb,
nov. Basionym: Sisymbrium lasiocarpum J. D.
Hooker & Thomson, J. Linn. Soc., Bot. 5: 162.
1861, not S. lasiocarpum F. Mueller, Fragm. 7:
20. 1869. Sisymbrium bhutanicum N. P. Ba-
lakrishnan, J. Bombay Nat. Hist. Soc. 67: 57.
1970. TYPE: Bhutan. Griffith s.n. (lectotype,
designated bv Jafri (1973), K; isolectotype,
BM).

Microsisymbrium duthiei 0. K. Schulz, Notizbl. Bot. Cart.
Berlin-Dahlem 9: 1089. 1927. Syn. nov. TYPE: In¬
dia. Kumaon | Uttar Pradesh]: Dhauli Valley. 2300-
2650 m, 7 Aug. 1886, ./. F. Duthie 5331 (holotype,

I ) I ); photo and fragments, B).

Sisymbrium monachorum W. W. Smith, Bee. Surv. Bot.
India 6: 35. 1913. Syn. nov. TYPE: Tibet. Gompa
Hill, Gayantse, 25 June 1907, //. M. Stewart s.n.
(holotype, CAL;? isotype, E).

Arabidopsis lasiocarpa (J. I). Hooker & Thomson) 0. E.
Schulz var. micrantha W. T. Wang, Bull. Bot. Res..

Harbin 8(3): 19. 1988. Syn. nov. TYPE: China. Vim-
nan: Deqin, 3000 m, July-Aug. 1935, Wang Cbi-wu
6 4727 (holotype, PE; isotype, PE).

Distribution. Bhutan, China, India, Nepal.

Balakrishnan (1970) proposed the name Sisym¬
brium bhutanicum to replace S. lasiocarpum J. D.
Hooker & Thomson because he erroneously be¬
lieved that ,S. lasiocarpum F. Mueller was the earlier
homonym.

Schulz's (1927) original description of Microsi¬
symbrium duthiei matches that of C. lasiocarpa in
every morphological detail, and it appears that
Schulz (1924) never examined any material of C.
lasiocarpa , a species he placed in Arabidopsis.
Schulz (1924, 1936) depended solely on the pres¬
ence vs. absence of seed mucilage to separate Mi¬
crosisymbrium from Arabidopsis , and it is likely that
he did not observe seed mucilage in the material
he described as M. duthiei.

5. Crucihimalaya stricta (Cambessedes) Al-
Shehbaz, O’Kane & Price, comb. nov. Basion¬
ym: Malcolmia stricta Cambessedes, in Jac-
quemont, Voy. Ind. Bot. 4: 16. 1844. TYPE:
NW Himalaya [India], Simla (as Semla), Jac-
quemont s.n. (holotype, P; isotype, K).

Arabidopsis stricta (Cambessedes) N. Busch var. bracteata
(). E. Schulz, Notizbl. Bot. Cart. Berlin-Dahlem 9:
1061. 1927. Syn. nov. TYPE: | India]. Kumaun, Deo-
pata, Naini Tal. 2300 m. 4 July 1885. J. F. Duthie
3835 (holotype, B).

Distribution. China, India, Kashmir, Nepal,
Pakistan.

An's (1987) report of Crucihimalaya stricta (as
Arabidopsis stricta ) from Sichuan is based on mis-
identifieation of plants of C. himalaica. The reports
in Schulz (1924), Jafri (1973), and Hajra et al.
(1993) of the species from Afghanistan were based
on collections (Aitchinson 210 and 251 , BM and K)
from Pakistan. A duplicate specimen at K of the
Jacquemont collection from India (no locality was
given, but it bears the numbers 2188 and “(786)”)
is definitely C. himalaica because it has large-au-
ricled leaves and bracts. It is not part of tin* type
collection of C. stricta, a species that never has
auriculate leaves or bracts.

Crucihimalaya stricta was not reported from Ne¬
pal (Hara, 1979), though Schulz (1927) cited one
collection (Duthie 5352) from Budhi village in
western Nepal. We have examined another collec¬
tion, Stain ton, Sykes tfr Williams 3365 (BM, E, G),
that was collected from Jagat and misidentified as
Arabidopsis mollissima.

Volume 9, Number 3

1999

Al-Shehbaz et al.

Exclusion from Arabidopsis

301

6. Crucihiinalaya himalaica (Edgeworth) Al-
Shehbaz, O'Kane & Price, comb. nov. Basion-
ym: Arabis himalaica Edgeworth, Trans. Linn.
Soc. 20: 31. 1846. TYPE: [India], “In glareosis
Himala, alt. ped. 10,000-11.000, Mana, ”
Edgeworth s.n. (holotype, K).

Arabis brevicaulis Jafri. Notes Hoy. Bot. Card. Kdinhurgh
22: 99. 1956. Syn. nov. TYPE: NW Himalaya. Ka¬
rakorum. Zangia Harar, Hunza Valley. 3600 m, 5
July 1939, R. S. Russell 1066 (holotype, BM).

Distribution. Afghanistan, Bhutan, China, In¬
dia, Kashmir, Nepal, Pakistan, Sikkim.

Arabis brevicaulis was transferred by Jafri (1973)
to Arabidopsis. He admitted that the species could
conceivably be placed in A. mollissima. The spec¬
imens cited by Jafri under .4. mollissima clearly
represent a mixture of that species and 4. hima¬
laica. One of the specimens, Bowes Lyon 849 (BM),
was annotated by Jafri as the holotype of Arabidop¬
sis chitralica Jafri, a name that was never pub¬
lished. The specimen was cited by Jafri (1973) as
4. mollissima , and it is most likely the same as the
type of Arabis brevicaulis.

On the basis of cited specimens, both Arabidop¬
sis himalaica and 4. mollissima were confused by
Schulz (1924), Hedge (in Hedge & Rechinger,
1968), and Jafri (1973). The principal character
used by these authors to separate these two species
is the presence in 4. himalaica of bracts along the
entire length of the inflorescences and the lack in
4. mollissima of bracts or their restriction to the
lowermost few flowers of the inflorescence. How¬
ever, this character shows tremendous variability in
the same population or even on the same plant. In
most plants of 4. himalaica the uppermost portion
of the inflorescence is ebracteate. For example, in
Lowndes 1166 (BM) one plant has the lowermost 24
flowers bracteate and another has only the lower¬
most 4 bracteate. In Stewart 1 7986 (US) one plant
has almost ebracteate inflorescences and another
has fully bracteate inflorescences. Plants of Cruci-
himalaya himalaica are annual or biennial with
coarse stellate and forked trichomes, at least ha-
sally bracteate inflorescences, slender fruits (0.4-
)0.5— 0.8(— 1) mm wide, and seeds 0.5— 0.8 mm long.
By contrast, plants of C. mollissima are perennial
with soft stellate trichomes, ebracteate inflores¬
cences, fruits 1—1.5 mm wide, and seeds 0.8— 1.1
mm long.

Plants from Chitral (Pakistan) treated by Jafri
(1956, 1973) as Arabis brevicaulis and Arabidopsis
brevicaulis, respectively, differ from typical C. him¬
alaica in being perennials with only the basal one
or two flowers bracteate, but they are perfectly at

home in the species in trichome tvpe, fruit width,
and seed length. They might be recognized at the
subspecific rank, but more material is needed for a
sound conclusion to be reached.

7. Crucihiinalaya axillaris (J. D. Hooker &

Thomson) Al-Shehbaz, O’Kane & Price, comb,
nov. Basionym: Sisymbrium axillare J. D.
Hooker & Thomson, J. Linn. Soc., Bot. 5: 162.
1861. TYPE: Sikkim. Alt. 8,000-10,000 ft.. J.
D. Hooker s.n. (lectotype, designated by Jafri
(1956), K; isolectotypes, B, G, W).

Microsisymbrium axillare (J. I). Hooker & Thomson) 0. E.
Schulz var. brevipedicellatum Jafri, Notes Roy. Bot.
Card. Edinburgh 22: 112. 1956. Syn. nov. TYPE:
India, babul. Keylang, 3600 m, 8 June 1889. G. Watt
2438 (holotype, E).

Microsisymbrium axillare var. dasycarpum 0. E. Schulz,
Pflanzenreich IV. 105(Heft 86): 160. 1924. Syn. nov.
TV PE: Bhutan. Griffith 1383 (holotype, CAL?; iso¬
types, K. P, W).

Microsisymbrium bracteosum Jafri, Notes Hoy. Bot. Card.
Edinburgh 22: 112. 1956. Syn. nov. TYPE: India.
Kuymaon. Bynas. ca. 2400 m. Apr. 1881, ./. R. Reid
s.n. (holotype, E).

Distribution. Bhutan, China, India, Kashmir.
Nepal, Sikkim.

Schulz (1924) recognized plants with puberulent
fruits as variety dasycarpum, but both glabrous and
puberulent fruits can be found in the same popu¬
lation. Robust plants of this species were described
by Jafri (1956) as Microsisymbrium bracteosum,
while those with short pedicels and ebracteate up¬
permost flowers were described as M. axillare var.
brevipedicellatum. However, these characters show
continuous variation and, therefore, these two taxa
do not merit recognition.

8, Crucihiinalaya wallichii (J. D. Hooker &

Thomson) Al-Shehbaz, O’Kane & Price, comb,
nov. Basionym: Sisymbrium wallichii J. D.
Hooker & Thomson, J. Linn. Soc., Bot. 5: 158.
1861. TYPE: [India]. Kumaon, Wallich 4784
(lectotype, designated by Jafri (1973), K; iso-
lectotype, BM).

Arabidopsis mollissima (C. A. Meyer) 0. E. Schulz var.
afghanica (). E. Schulz, Pflanzenreich IV. 105(Heft
86): 281. 1924. Syn. nov. TYPE: Afghanistan. Grif¬
fith 1470 (holotype. K).

Arabidopsis russelliana Jafri, Notes Hoy. Bot. Card. Ed¬
inburgh 22: 97. 1956. Syn. nov. TYPE: Karakorum,
Kero Eugma glacier. 3900 m, 27 July 1939, R. S.
Russell 1855 (holotype, BM).

Distribution. Afghanistan, Bhutan, China, In¬
dia, Iran, Kashmir, Kazakstan, Kyrgyzstan, Nepal,
Pakistan, Tajikistan, Turkmenistan, I zbekistan.

302

No von

Crucihimalaya wallichii is the most variable spe-
eies in the genus, especially in the density of in¬
dumentum, division of basal and eauline leaves,
length of fruiting pedicels, and length and degree
of compression of fruit. The continuous variation in
these characters has been adequately described by
Hedge (in Hedge & Rechinger, 1968), and il is
evident that no infraspecific taxa can be recog¬
nized. However, Jafri (1973) segregated two addi¬
tional species, as Arabidopsis taraxacifolia (T. An¬
derson) Jafri and A. russelliana Jafri, based
primarily on differences in the degrees of stem
branching, style length, and petal size. In our opin¬
ion, these highly variable characters are unreliable
and, therefore, these segregates do not merit rec¬
ognition at any rank.

Crucihimalaya wallichii is closely related to C.
kneuckeri, anil the two species can be separated
only by the characters in the key above.

Schulz (1927), in his original description of Mi¬
crosisymbrium flaccidum (). E. Schulz, cited two
syntypes, Duthie 11055 and Inayat 19172, both of
which are deposited at DD, with photos and frag¬
ments at B. The former collection was designated
by Jafri (1973) as the lectotype, and there is an
isolectotype at K. This is definitely an immature
plant of a species of Arabis. The second collection,
Inayat 19172, is a glabrescent form of Crucihi¬
malaya wallichii with slightly lyrate leaves and in¬
cumbent cotyledons.

9, Crucihimalaya kneuckeri (Bornmiiller) Al-
Shehbaz, O’Kane & Price, comb. nov. Basio-
nym: Sisymbrium kneuckeri Bornmiiller, in A.
Kneueker, Allgem. Bot. Zeitsehr. 9: 45. 1904.
TYPE: [Egypt |. Ml. Sinai, Dschebel Katharin,
5 Apr. 1902, A. Kneueker s.n. (holotype, JE;
isotype, B).

Distribution. Saudi Arabia, Egvpt (Sinai).

Schulz (1924) mixed the flower size of Crucihi¬
malaya kneuckeri with that of C. wallichii. The lat¬
ter has flowers as small as 2 mm long. More ma¬
terial of C. kneuckeri is needed to fully assess its
overall variation.

Oliinarabidopsis Al-Shehbaz, O’Kane & Price,
gen. nov. TYPE: Olimarabidopsis pumila (Ste¬
phan) Al-Shehbaz, O'Kane & Price.

Folia caulina sessilia. minute vel valde auriculata; pili
ramosi sessiles vel minute stipitati malpighiacei vel stel-
lati. ram is simplicibus; raeemi ebracteati. valde elongati:
sepala oblonga, nonsaceata; petala flava; fruetus lineares.
leretes, saepe stellati vel malpighiacei; stipitum nullum:
septum nullum vel perforatum vel completum; semina un-

iseriata. 9—30 per loeula. oblonga, mueilaginosa; eotyle-
dones incumbentes.

Her! )s annual. Trichomes short-stalked or ses¬
sile, malpighiaeeous and 3- or 4-rayed stellate.
Stems erect to ascending, sometimes decumbent,
simple or branched basally and/or apically. Basal
leaves not rosulate, simple, entire or very rarely
pinnately dissected. Cauline leaves sessile, minute¬
ly to conspicuously aurieulate, entire to dentate.
Inflorescences few- to several-flowered, ebracteate,
corymbose racemes, elongated considerably or
rarely slightly elongated in fruit; rachis straight. Se¬
pals oblong, deciduous, erect, glabrous or pubes¬
cent, base of inner pair not saccate. Petals yellow
or yellowish white, oblanceolate. Stamens 4 or 6
and only slightly tetradynamous; anthers oblong,
rounded at base, obtuse at apex. Nectar glands con¬
fluent and subtending bases of all stamens. Ovules
18-60 per ovary. Fruit dehiscent, linear, terete;
valves with a distinct midvein, pubescent with ex¬
clusively malpighiaeeous and/or short-stalked stel¬
late trichomes; gynophore absent; septum complete,
perforated, or reduced to a rim; style obsolete or
distinct and to 1 mm long; stigma capitate, entire.
Seeds uniseriate, wingless, oblong, plump; seed
coat minutely reticulate, slightly mucilaginous or
not mucilaginous when wetted; cotyledons incum¬
bent.

As delimited here, Olimarabidopsis, which
means “formerly Arabidopsis," consists of three
species all of which were previously placed in Ar¬
abidopsis. One of those species was described twice
under two genera (as Trichochiton umbrosum Bot-
schantsev & Vvedensky and 20 years later (see be¬
low) as Arabidopsis eseptata Hedge). The two ques¬
tions most relevant to Olimarabidopsis are: First, is
Trichochiton distinct from the closely related and
earlier published Cryptospora Karelin & Kirilow?
Second, is T. umbrosum correctly assigned to this
genus? All species of Cryptospora have 2-lobed
stigmas, large seeds 2.5— 4.5 mm long, indehiseent
fruits that break up at maturity into 1-seeded seg¬
ments, some simple trichomes on the leaves and/or
stems, white to lavender flowers, and non-auricu-
late cauline leaves (Botschantsev, 1963). The type
species of Trichochiton, T. inconspicuum Komarov,
has all of these characters. Therefore, we fully
agree with Schulz (1936) and Rechinger (in Hedge
& Rechinger, 1968) in reducing Trichochiton to
synonymy of Cryptospora and in recognizing its type
as C. inconspicua (Komarov) 0. E. Schulz.

The answer to the second question is no. Tri¬
chochiton umbrosum has entire stigmas, small seeds
to 1.3 mm long, readily dehiscent fruits, exclusively

Volume 9, Number 3
1999

Al-Shehbaz et al.

Exclusion from Arabidopsis

303

stellate and malpighiaceous (never simple) tri-
chomes, pale yellow flowers, and minutely aurieu-
late cauline leaves. In our opinion, these differenc¬
es are substantial, and T. umbrosum {— Arabidopsis
eseptata) is herein assigned to a new genus, Oli¬
marabidopsis, of three very closely related species.

Olimarabidopsis is readily distinguished from Ar¬
abidopsis by having yellow flowers, auriculate cau¬
line leaves, and pubescent fruits. In contrast, Ara¬
bidopsis has white to lavender flowers, short
petiolate stem leaves, and glabrous fruits. Olimar¬
abidopsis is distinguished from small-flowered spe¬
cies of the closely related genus Erysimum by its
auriculate cauline leaves and short-stalked
branched trichomes. All species of Erysimum have
petiolate cauline leaves and malpighiaceous and/or
sessile 3— 5-rayed stellate trichomes.

Kky to thk Speciks of Olimarabidopsis

la. Trichomes on fruit valve exclusively submalpigh-
iaceous; seeds 1-1.3 mm long; fruit attenuate to
apex; style obsolete; stamens 4 or rarely 6 ...

. 3. (). umbrosa

lb. At least some of the fruit trichomes 3- or4-rayed;
seeds less than 1 mm long; fruit cuneate to apex;
style distinct; stamens 6.

2a. Septum perforate; fruit 0.4 — 1( — 1 .5) cm long,

5— 18(— 20)-seeded . 2. <>. cabulica

2b. Septum complete; fruit 1.5— 3.2(-4) cm, (15—)

22 — 40( — 60)-seeded . I.O. pumila

1. Olimarabidopsis pumila (Stephan) Al-Sheh-
baz, O’Kane & Price, comb. nov. Basionym:
Sisymbrium pumilum Stephan, in Willdenow,
Sp. PI. ed. 4, 3(1): 507. 1800. TYPE: N Persia.
[Kizlar], Stephan s.n. (holotype, LE; isotype,
W ).

Distribution. Afghanistan, Armenia, Azerbai¬
jan, China, Georgia, India, Iran, Iraq, Israel, Jor¬
dan, Kazakstan, Kyrgyzstan, Lebanon, Oman, Pak¬
istan, Russia, Syria, Tajikistan, Turkmenistan,
Turkey, Uzbekistan.

Sisymbrium grijjithianum Boissier was trans¬
ferred to Microsisymbrium by Schulz (1924). An ex¬
amination of the type and numerous other collec¬
tions clearly shows that the species is a minor
variant of Olimarabidopsis pumila with strongly re¬
flexed instead of divaricate to ascending fruits. In
fact, plants with reflexed and divaricate to ascend¬
ing fruits are commonly found in the same popu¬
lation, as evidenced by Botschantsev 111 (LE) from
Uzbekistan, Hedge & Wendelbo 2901 (E) from Af¬
ghanistan, and Lammond 1026 (E, LE) from Paki¬
stan. It is surprising, therefore, to have the same
species cited by Schulz (1924) under two different
genera, as M. grijjithianum and Arabidopsis pumila.

We agree with Hedge (in Hedge & Rechinger,
1968) in reducing S. grijjithianum to synonymy of
Olimarabidopsis pumila (as A. pumila ).

2. Olimarabidopsis cabulica (J. I). Hooker &

Thomson) Al-Shehbaz, O'Kane & Price, comb,
nov. Basionym: Sisymbrium cabulicum J. D.
Hooker & Thomson, J. Linn. Soc., Bot. 5: 161.
1861. TYPE: Afghanistan. Griffith f 1465] (ho¬
lotype, K).

Arabidopsis korshinskyi Botschantsev. Novit. Syst. PI. Vase.
Acad. Sci. UKSS 1965: 272. 1965. Svn. nov. TY PE:
[Tajikistan|. Alaica Valley, near Katyn-Art. stony Ky-
zvlsu. 13 July 1895, ,S. Korshinsky 304 (holotype,
LE).

Distribution. Afghanistan, W China, Kyrgyzst¬
an, Tajikistan.

The type and other collections annotated by Bot¬
schantsev as Arabidopsis korshinskyi are indistin¬
guishable from the type of Sisymbrium cabulicum.
The latter was reduced by Jafri (1973) to synonymy
of Arabidopsis pumila. However, the perforate sep¬
tum, shorter fruits, and fewer seeds per locule (see
the key above) readily distinguish plants of the two
species herein placed in Olimarabidopsis. Further¬
more, 0. cabulica (reported as A. korshinskyi ) is a
hexaploid (2 n — 48), whereas O. pumila is tetra-
ploid (2n = 32) (Aryavand, 1983; Ginter & Ivanov,
1968; Polatschek, 1971).

An (1987) reported the species (as Arabidopsis
pumila (Stephan) N. Busch var. alpina (Korshinsky)
0. E. Schulz) from Xinjiang, but we have not seen
any material from China.

3. Olimarabidopsis umbrosa (Botschantsev &

Vvedensky) Al-Shehbaz, O’Kane & Price,
comb. nov. Basionym: Trichochiton umbrosum
Botschantsev & Vvedensky, Not. Syst. Herb.
Inst. Bot. & Zool. Acad. Sci. Uzbekistan 12:
10. 1948. TYPE: [Tajikistan]. Samarkand re¬
gion, shady places around Lake Koli-Kalart,
16 June 1916, Lipsky s.n. (holotype, TASH).

Distribution. Afghanistan, Tajikistan, Uzbeki¬
stan.

Arabidopsis eseptata Hedge was correctly re¬
duced to synonymy of Trichochiton umbrosum by
Pachomova (1974) and Junussov (1978), but Paeh-
omova erroneously cited the authorship of the spe¬
cies by considering it as a new combination based
on Komarov’s (1896) T. inconspicuum var. umbro¬
sum Komarov. Although they selected the same ep¬
ithet for the species, Botschantsev and Vvedensky
(1948) treated T. umbrosum as a new species and
cited a holotype collected 22 years after Komarov’s
publication.

304

Novon

Pseudoarabidopsis Al-Shehbaz, O’Kane & Price,
gen. nov. TYPE: Pseudoarabidopsis toxophylla
(Bieberstein) Al-Shehbaz, O'Kane & Price.

Folia caulina sessilia. sagittato-amplexicaulia vel raro
auriculata; pili ramosi sessiles stellati, ramis ramosis; ra-
ceini ebracteati, valde elongati; sepala oblonga, subsae-
cata; petala alba vel lavandula; fruetus lineares, teretes,
glabri; stipitum distincium; septum eompletum; semina
biseriata, 30—50 per locida, oblonga, mucilaginosa; coty-
ledones ineumbentes.

Herbs biennial or perennial. Trichomes sessile,
stellate, 4- or 5-rayed, with at least some rays
branched. Stems erect to ascending, simple or
branched basally. Basal leaves rosulate, often with¬
ered by fruiting time, simple, entire or dentate.
Cauline leaves sessile, deeply sagittate-amplexi-
caul or rarely auriculate, entire to dentate. Inflo¬
rescences several-flowered, ebracteate, corymbose
racemes, elongated considerably in fruit; rachis
straight. Sepals oblong, deciduous, erect, glabrous
or pubescent, base of inner slightly saccate. Petals
white to lavender, spatulate. Stamens 6, promi¬
nently tetradynamous; anthers oblong, sagittate at
base, obtuse at apex. Nectar glands confluent, sub¬
tending bases of all stamens, surrounding those of
lateral ones. Ovules 60—100 per ovary. Fruit dehis¬
cent, linear, terete; valves with obscure midvein,
glabrous; gynophore distinct; seplum complete;
style distinct and to 1 mm long; stigma capitate,
entire or slightly 2-lobed. Seeds biseriate, 30-50
per locule, wingless, oblong to ovoid, plump; seed
coat minutely reticulate, slightly mucilaginous
when wetted; cotyledons incumbent.

Pseudoarabidopsis is readily distinguished from
Arabidopsis and Erysimum by having sessile, sag-
ittate-amplexicaul cauline leaves, exclusively ses¬
sile, stellate trichomes with branched rays, and a
distinct gynophore. In Arabidopsis the leaves are
petiolate and never amplexicaul or sagittate, the tri¬
chomes are simple mixed with stalked forked ones,
and the fruits are sessile or subsessile. In its stel¬
late trichomes and sagittate-amplexicaul cauline
leaves, P. toxophylla resembles Capsella bursa-pas-
toris (L.) Medikus. Although drastically different in
fruit morphology, these two species show remark¬
able affinities on the basis of ITS results.

Pseudoarabidopsis toxophylla (Bieberstein) Al-
Shehbaz. O'Kane & Price, comb. nov. Basio-
nym: And) is toxophylla Bieberstein, FI. Taur.-
Cauc. 3: 448. 1819. TYPE: Not designated.
Busch (1939) selected the specimen collected
from Perekop and housed at LE as the type.
We have not seen this specimen.

Distribution. Afghanistan, western China, Ka¬
zakstan, Russia, Tajikistan.

Generic Placement of Species Previously
Inch ded in Arabidopsis

The following 59 binomials previously assigned
to Arabidopsis are assigned to 14 genera, including
Arabidopsis. Accepted taxa are in boldface, and
synonyms, excluded, or doubtful taxa are in italics.
Binomials marked with (*) are proposed in this pa¬
per.

Arabidopsis arenosa (L.) Lawalree, Bull. Soe.
Roy. Bot. Belg. 92: 242. 1960.

A. bactriana Ovczinnikov & Junussov, FI. Tadzhitskoi SSR
5: 626. 1978. No material has been seen, but ac¬
cording to the original description, the species can¬
not be assigned to Arabidopsis because it is a pul-
vinate. scapose perennial with cylindric fruits,
subbiseriate seeds, and leafless stems. It is likely
that the plant belongs to Crucihimalaya mongolica.
A. brevicaulis (Jafri) Jafri, FI. W. Pakistan 55: 272. 1973.
= Crucihimalaya hinialaiea (Edgeworth) Al-Sheh-
baz. O’Kane & Price (*).

A. bursifolia (DC.) Botschantsev, Not. Sysl. Herb. Inst. Bot.
Acad. Sci. URSS 19: 106. 1959. The systematic |x>-
sition ol this species is problematic and awaits further
study. Berkutenko (1988) suggested that the North
American (Greenland, Yukon, and Alaska) Arabidop¬
sis mollis (Hooker) 0. E. Schulz, which Boltins
(1943, 1952. 1993) treated as Halimolobos mollis
(Hooker) Hollins, is probably the same as Russian
(Far East) A. bursifolia (DC.) Botschantsev and that
A. tsehuktschorum (Jurtzev) Jurtzev, which is known
only from the type locality at Chegiton River (Chu¬
kotka), is probably an abnormal plant of A. bursifol¬
ia. Although A. bursifolia ami H. mollis may well
prove to be conspecifie, we have not yet conducted
extensive study on them, and our work on Halimo¬
lobos is still in progress. However, we are certain that
none of these species belongs to Arabidopsis.

A. campestris (). E. Schulz, Notizbl. Bot. Cart. Berlin-Dah-
lem 9: 1059. 1927. = Crucihimalaya wallichii (J.
D. Hooker & Thomson) Al-Shehbaz. O'Kane & Price
(*)•

A. cebennensis (DC.) O’Kane & Al-Shehbaz, No¬
von 7: 325. 1997.

A. croalica (Schott) O'Kane & Al-Shehbaz, Novon
7: 325. 1997.

A. dentata (Allioni) Dalla Torre, Alpenfl. 115. 1899. =
IVliirherkiella pinnalifida (Lamarck) Rothmaler,
Bot. Not. 1939: 469. 1939.

A. drassiana Naqshi & Javeid. J. Eeon. Taxon. Bot. 7: 624.
1986. No material has been seen. The species is
excluded from Arabidopsis because it has sessile am-
plexieaul leaves. Naqshi anil Javeid (1986) provided
an incomplete description and gave no mention of
the type of trichomes, which are very important in
the Arabidopsis complex, and whether the fruits are
terete or flattened. Because they indicated that the
linear fruits include only one or two seeds, it is pos¬
sible that the plant is of hybrid origin.

Volume 9, Number 3
1999

Al-Shehbaz et al.

Exclusion from Arabidopsis

305

A. erysimoides Hedge & Kit Tan. PI. Syst. Evol. 156: 202.
1087. = Erysimum hedgeunum Al-Shehbaz. No-
von 4: I. 1994.

A. eseptata Hedge, FI. Iran. 57: 334. 1968. = Olimara-
bidopsis umbrosa (Botschantsev & Vvedensky) Al-
Shehbaz, O’Kane & Priee (*).

A. gamosepala Hedge, FI. Iran. 57: 334. 1968. = Neo-
torularia gamosepala (Hedge) Al-Shehbaz X
O’Kane, Novon 7: 93. 1997.

A. glauca (Nuttall ex Torrey & A. Gray) Rydberg. FI.
Rocky Mt. 342. 1917. = Tludlungiella salsuginea
(Pallas) (). F. Schulz, Pflanzenr. IV. 105(Heft 86):
252. 1924.

A. griffithiana (Boissier) N. Busch, FI. Cauc. Crit. 3(4):
457. 1909. = ( Himarabidopsis puinila (Stephan)
Al-Shehbaz, O'Kane & Price (*).

A. halleri (L.) O'Kane & Al-Shehbaz, Novon 7:
325. 1997.

A. himalaica (Edgeworth) 0. F. Schulz, Pflanzenreich
IV.105(Heft 86): 283. 1924. = Crueibimalaya
himalaica (Edgeworth) Al-Shehbaz, O’Kane & Price
(*).

A. huetii (Boissier) N. Busch, Acta Hurt. Petrop. 28: 389.
1908. = Murbeekiella huetii (Boissier) Rothmaler,
Bot. Not. 1939: 472. 1939.

,4. kneuckeri (Bornmiiller) O. E. Schulz, Pflanzenr. IV.
105(Heft 86): 277. 1924. = Crueibimalaya
kneuckeri (Bornmiiller) Al-Shehbaz, O’Kane &
Price (*).

A. korshinskyi Botschantsev, Novit. Syst. PI. Vase. Acad.
Sci. URSS 1965: 272. 1965. = Olimarahidopsis
cabulica (J. I). Hooker & Thomson) Al-Shehbaz,
O’Kane & Price (*).

A. lasiocarpa (J. I). Hooker & Thomson) 0. E. Schulz,
Pflanzenr. IV. 105(Heft 86): 282. 1924. = Cruci-
bimalaya lasiocarpa (J. 1). Hooker & Thomson) Al-
Shehbaz, O'Kane & Price (*).

A. lyrata (L.) O’Kane & Al-Shehbaz, Novon 7:
325. 1997.

A. minutiflora (J. I). Hooker & Thomson) N. Busch, FI
Cauc. Crit. 3(4): 457. 1909 = lanbedgea ininuti-
llora (J. I). Hooker & Thomson) Al-Shehbaz &
O'Kane. Fdinb. .1. Bot. (1999 in press).

A. mollis (Hooker) O. F. Schulz, Bot. Jahrb. Syst. 66: 97.

1933. = llalimolobos mollis (Hooker) Rollins,
Rhodora 43: 480. 1941.

.4. mollissima (C. A. Meyer) N. Busch, FI. Sib. Or. Extr.
1: 136. 1913. = Crueibimalaya mollissima (C. A.
Meyer) Al-Shehbaz, O'Kane & Price (*).

A. monachorum (W. W. Smith) O. F. Schulz, Pflanzenr. IV.
105(Heft 86): 282. 1924. = Crueibimalaya lasio¬
carpa (Vi. Vi. Smith) Al-Shehbaz, O’Kane & Price
(*).

,4. mongolica (Botschantsev) Mesif'ek & Sojak. Folia Geo-
bot. Phytotax. 30: 448. 1995. = Crueiliimalaya
mongolica (Botschantsev) Al-Shehbaz, O’Kane &
Price (*).

A. multicaulis Pampanini, Sped. Ital. DeDilippi Himal.,
etc. 1913-1914, Ser. 2. 11 (Agg. FI. Carac.): 160.

1934. = Arabis tibetica J. I). Hooker & Thomson,
J. Finn. Soe.. Bot. 5: 143. 1861.

A. neglecta (Schultes) O’Kane & Al-Shehbaz, No¬
von 7: 326. 1997.

,4. novae-anglicae Britton, in Britton & Brown, Illus. FI.,
ed. 2, 2: 176. 1913. = Neotorularia bumilis (C.
A. Meyer) Hedge & J. Leonard. Bull. Jard. Bot. Nat.
Belg. 56: 394. 1986.

A. undo (Belanger) Bornmiiller, Beih. Bot. Zentralbl.
33(2): 275. 1915. = Drahopsis nuda (Belanger)
Stapf, Denkschr. Akad. Wiss. Wien. Math. -Nat. Kl.
51(2): 298. 1886.

A. ovezinnikovii Botschantsev, in P N. Ovczinnikov, FI.
Tadzhitskoi SSR 5: 625. 1978. = Crucihimalaya
ovezinnikovii (Botschantsev) Al-Shehbaz, O’Kane
& Price (*).

.4. parvula (Schrenk) 0. F. Schulz, Pflanzenr. IV. 105(Heft
86): 269. 1924. = Tbellmigiella parvula (Schrenk)
Al-Shehbaz & O'Kane. Novon 5: 309. 1995.

A. pedemoiilana (Boissier) O'Kane & Al-Sheh¬
baz, Novon 7: 326. 1997.

,4. pinnatifida (Lamarck) Ruprecht, Mem. Acad. Sci. St.
Petersb. Ser. 7. 15(2): 86. 1869. = Murbeckiella
pinnatifida (Lamarck) Rothmaler, Bot. Not. 1939:
469. 1939.

A. pumila (Stephan) N. Busch. FI. Cauc. Crit. 3(4): 457.
1909. = Olimarahidopsis puinila (Stephan) Al-
Shehbaz, O’Kane & Price (*).

A. qiranica /. X. An, FI. Xinjiang. 2(2): 376. 1995. =
Sisymhrinpsis mollipila (Maximowicz) Botsc hant¬
sev, Nov. Syst. PI. Vase. 3: 122. 1966.

A. riehardsonii (Rydberg) Rydberg, FI. Rocky Mt. 341.
1917. = Neotorularia bumilis (C. A. Meyer)
Hedge & .1. Leonard. Bull. Jard. Bot. Nat. Belg. 56:
394. 1986.

A. russeliana Jafri, Notes Roy. Bot. Card. Fdinb. 22: 97.
1956. = Crueibimalaya wallichii (J. 1). Hooker &
Thomson) Al-Shehbaz, O’Kane & Price (*).

A. salsuginea (Pallas) N. Busch, FI. Sib. 1: 136. 1913. =
Thellungiella salsuginea (Pallas) O. F. Schulz,
Pflanzenr. IV. 105(Heft 86): 252. 1924.

A. sarbalica Naqshi A Javeid. J. Fcon. Taxon. Bot. 7: 621.
1985 (1986). No material has been seen. The species
is excluded from Arabidopsis because I In* cauline
leaves are minutely auriculate. The original descrip¬
tion and illustration clearly support placing the spe¬
cies in Crucihimalaya wallichii. Naqshi and Javeid
(1986) separated A. sarbalica from A. wallichii be¬
cause the former has shorter fruits 2.5— 3.2 mm long,
but this fruit length falls within that of C. wallichii.

A. schimperi (Boissier) N. Busch, FI. Cauc. Crit. 3(4): 457.
1909. = Robescliia schimperi (Boissier) O. F.
Schulz, Pflanzenr. IV. 105(Heft 86): 360. 1924.

A. stenocarpa Rydberg. Torrea 7: 160. 1907. = Halimo-
lobos virgata (Nuttall ex Torrey & A. Cray) 0. E.
Schulz. Pflanzenr. IV. 105(1 left 86): 290. 1924.

A. stewartiana Jafri, Notes Roy. Bot. Card. Edinburgh 22:
96. 1956. We have not seen any material of this
species. Jafri (1956) distinguished it from Olimara-
bidopsis pumila (as A. pumila) on the basis of having
semi-amplexicaul instead of sagittate-amplexicaul
leaves and glabrous instead of pubescent fruits. Leaf
base is extremely variable in O. pumila, and gla¬
brous fruits, which are very rare in the species, can
be found in populations that have predominantly pu¬
bescent fruits. Because of the bright yellow flowers,
the species does not belong to Arabidopsis. The de¬
tailed original description of A. stewartiana leaves
no doubt that it is a minor variant of 0. pumila.

306

Novon

,4. stricta (Cambessfcdes) N. Busch, FI. Cauc. Grit . 3(4):
457. I'M)*). = Criiciliimuluya stricta (Cambes-
sfcdes) Al-Shehbaz, O’Kane & Price (*).

A. suecica (Fries) Norrlin, Meddel. Soc. Fauna FI.
Fenn. 2: 12. 1878.

4. laraxacifolia (T. Anderson) Jafri, FI. W. Pakistan, 55:
274. 1973. = Crueiliiinalaya wallichii (J. D.
Hooker & Thomson) Al-Shehbaz, O'Kane & Price
(*). For an excellent discussion on the variation in
this species and synonymy, see Hedge (in Hedge &
Rechinger, 1968).

A. tenuisiliqua (K. H. Rechinger & Kiiie) Jafri. FI. W. Pak¬
istan 55: 171. 1973. = Arabia tenuisiliqua K. H.
Rechinger & Kiiie. An/.. Math. -Mat. Kl. Oesterr.
Akad. Wiss. 7: 5. 1954. We have not seen the type
or any material of this species, and we are following
Hedge (in Hedge & Rechinger, 1968) in maintaining
it in Arabis, though the species is anomalous in that
genus because of its incumbent cotyledons. It is ex¬
cluded from Arabidopsis because it has stellate fruit
trichomes, and no species of this genus has stellate
hairs or pubescent fruits. Therefore. Jafri's (1973)
transfer of the species is unacceptable. It is likely
that the species is related to Crucihimalaya walli-
chii, from which it differs by having larger flowers
and smaller fruit.

A. thalianu (L.) Heynhold, in Hull & Heynhold,
Clav. Gen. FI. Sachsen 1: 538. 1842.

4. libetica Naqshi & Javeid, J. Econ. Taxon. Bot. 7: 621.
1985 (1986). We have not seen the type of this spe¬
cies, but the name is illegitimate because it is a later
homonym ol the following species. The plant is ex¬
cluded from Arabidopsis because it is a perennial
with stellate trichomes. Naqshi and Javeid (1986)
considered the species to be closely related to Cru¬
cihimalaya himalaica (as 4. himalaica ) and distin¬
guished solely on the basis of having non-auriculate,
distinctly veined cauline leaves. It remains to be
seen whether or not the two are conspecific.

4. libetica (J. I). Hooker & Thomson) Fan & C. H. An ex
K. C. Kuan. FI. Xizang. 2: 372. 1985. = Arabis
libetica J. I). Hooker & Thomson, J. Linn. Soc., Bot.
5: 143. 1861.

4. toxophylla (Bieberstein) N. Busch, FI. Cauc. Crit. 3(4):
457. 1909. = Pseudouruhidopsis toxophylla
(Bieberstein) Al-Shehbaz, O’Kane & Price (*).

4. trichocarpa B. F. Huang, in S. W. Liu, FI. Qinghaica 1:
509. I ‘>97. We have examined the holotype of this
species, and on the basis of its pubescent fruits and
bracteate lowermost portion of inflorescences, the
species is excluded from Arabidopsis. The plant is a
minor variant of Neotorularia humilis (C. A. Meyer)
Hedge & Leonard.

4. trichopoda (Turczaninow) Botsehantsev, Not. Syst.
Herb. Inst. Bot. Acad. Sci. IJRSS 18: 104. 1957.
Botsehantsev (1959) was correct in his placement of
this species in synonymy of Arabidopsis bursifolia.
See discussion under this species.

4. tschuktschorum (Jurtsev) Jurtsev, Bot. Zhurn. 60: 240.
1975. See discussion under Arabidopsis bursifolia.

A. luemurnica K. C. Kuan & C. H. An, Bull. Bot. Lab.
North-East Forest. Inst. 1980(8): 44. 1980. = Neo-

torularia humilis (C. A. Meyer) Hedge & J. Leo¬
nard, Bull. Jard. Bot. Nat. Belg. 56: 394. 1986.

4. verna (Koch) N. Busch, FI. Cauc. Grit . 3(4): 460. 1909.
= Drabopsis uuda (Belanger) Stapf, Denkschr.
Akad. Wiss. Wien, Math.-Nat. Kl. 51(2): 298. 1886.
4. virgata (Nultall ex Torrey & A. Cray) Rydberg, FI.
Rocky Ml. 342. 1917. = Ilaliniolobos virgata
(Nuttall ex Torrey & A. Cray) 0. E. Schulz, Pflan-
zenr. IV. 105(Heft 86): 290. 1924.

4. wallichii (J. 1). Hooker & Thomson) N. Busch, FI. Cauc.
Crit. 3(4): 457. 1909. = Crucihimalaya wallichii
(J. I). Hooker & Thomson) Al-Shehbaz, O’Kane &
Price (*).

4. yadungensis K. C. Kuan & C. H. An, FI. Xizang. 2:
375. 1985. An examination of the type collection
reveals that the species is definitely a member of the
Arabis hirsuta (L.) Scopoli complex. The immature
fruits are strongly compressed and subappressed to
rachis, and the cauline leaves are strongly auricu-
late, all of which are characters not found in any
Arabidopsis species. The species is under study by
the senior author in connection with the Flora of
China project.

Acknowledgments. Research and fieldwork
were supported by the National Science Foundation
(grant DEB-9208433) and the National Geographic
Society (grant 5068-93). The following people were
gracious hosts or invaluable field companions: Ab¬
dulla Abbas, An Zheng-xi, Katsutoshi Arai, Nog-
man Aralbaev, Isa 0. Baitulin, Alexandra Berku-
tenko, Deng Yanbin, Fu Dezhi, Gheorghe Dihoru,
Vladimir Dorofeev, Yang Guang, Qin Haining, Sun
Hang, Josef Holub, Peng Hua, Dilxat Kasm, H.
Kato, the late Sigizmund Kharkevich, Franta Kra-
hulec, Zbigniew Mirek, Hanna Kuciel, Lou Lianli,
Karol Marhold, Klaus Mummenhoff, Xia Nianhe,
Noriaki Murakami, Zvetlana Nesterova, Hideaki
Ohba, Nonna Pavlova, Ching-I Peng, Nina Proba-
tova, Boris Syomkin, Hirokazu Tsukaya, and Guilin
Zhu. Barbara Sehaal graciously made her labora¬
tory available for many of the molecular studies
leading to this paper. We are grateful to Henk van
der Werff for correcting the Latin, Zhu Guanghua
for help with Chinese literature, Tatiana Shulkina
for translating some Russian texts, and Michael G.
Gilbert and Anthony R. Brach for providing much-
needed literature. We thank the directors and cu¬
rators of the following herbaria (abbreviations fol¬
low Holmgren et al. (1990) and Fu et al. (1993))
for the loans and/or for allowing the study of their
holdings: A, AA, B, BH, BKL, BM, CAS, CDB1,
COLO, E, F, G, GH, HAST, HNWP, IBSC, K,
KATH, KUN, LE, M. MHA, MO, MW, NAS, NY,
NYS, P, PE, S, SHI, SZ, TAI, TI, TNS, UC, US, W,
WU, WUK, XJA, XJBI, XJU, and Z.

Literature Cited

Al-Shehb az, I. A. 1973. The biosytematics of the genus
Thelypodium (Cruciferae). Contr. Gray Herb. 204: 3—
148.

Volume 9, Number 3
1999

Al-Shehbaz et al.

Exclusion from Arabidopsis

307

- . 1984. The tribes of Crueiferae (Brassicaceae) in

the southeastern United States. J. Arnold Arbor. 65:
343-373.

- . 1988. The genera of Sisymbrieae (Crueiferae:

Brassicaceae) in the southeastern United States. J. Ar¬
nold Arbor. 69: 213—237.

- . 1994. Erysimum hedgeanum (Brassicaceae), a

new name replacing Arabidopsis erysimoides. Novon 4:
1-2.

- & S. I.. O'Kane, Jr. 1995. Placement of Arabidop¬
sis parvula in Thellungiella (Brassicaceae). Novon 5:
309-310.

- & - . 1997. Arabidopsis gamosepala and A.

tuemurnica belong to A leotorularia (Brassicaceae). No¬
von 7: 93—94.

An, Z. X. 1987. Arabidopsis. In: T. Y. Cheo (editor), Cru¬
eiferae. FI. Reipubl. Popularis Sin. 33: 280-288. Sci¬
ence Press, Beijing.

Aryavand, A. 1983. Contribution a l’etude eytotaxono-
mique des Crueiferes de l’lran. III. Bull. Soc. Neucha-
teloise Sci. Nat. 106: 123—130.

Balakrishnan, N. P. 1970. Nomenclatural notes on some flow¬
ering plants — II. J. Bombay Nat. Hist. Soc. 67: 57-66.
Ball. P W. 1993. Arabidopsis. Pp. 322—323 in T. G. Tutin.
N. A. Burges, A. O. Chater, J. B. Edmondson, V. H.
Heywood. I). M. Moore, I). H. Valentine, S. M. Walters
& D. A. Webb (editors), Flora Europaea, 2nd ed., vol.

1. Psilotaceae to Platanaceae. Cambridge Univ. Press,
Cambridge.

Berkutenko, A. N. 1988. Brassicaceae. Pp. 38-115 in: S.
S. Charkevicz [KharkevichJ (editor), Plantae Vasculares
Orientis Extremi Sovietiei, vol. 3. Academy of Sciences,
Leningrad.

Botschantsev, V. P. 1963. Generis Cryptosporae Kar. et
Kir. revisio. Notul. Syst. Herb. Inst. Bot. Acad. Sci.
URSS 22: 144-149.

- . 1959. De Cruciferis notae criticae, III. Notul.

Syst. Herb. Inst. Bot. Acad. Sci. URSS 19: 105—108.

- . 1975. The new Crueiferae from Mongolia. Bot.

Zhurn. 60: 947-948.

- & A. I. Vvedensky. 1948. Crueiferae novae ex

Asia Media. Notul. Syst. Herb. Inst. Bot. & Zool. Acad.
Sci. Uzbekistan 12: 3—12.

Busch, N. A. 1909. Arabidopsis. In: N. Kusnezow (editor),
El. Caucus. Cril. 3(4): 457—461.

- . 1939. Arabidopsis. In: V. I,. Komarov (editor), El.

URSS 8: 76—80. Academy ol Sciences Press, Lenin¬
grad.

Eu. I,. G., X. C. Zhan, H. N. Qin & J. S. Ma (Editors).
1993. Index Herbariorum Sinicorum. China Science &
Technology Press, Beijing.

Galloway, G. I... R. L. Malmberg & R. A. Price. 1998.
Phylogenetic utility ol the nuclear gene arginine decar¬
boxylase: An example from Brassicaceae. Molec. Biol.
Evol. 15: 1312-1320.

Ginter, T. N. & V. 1. Ivanov. 1968. Chromosome numbers
in Arabidopsis species. Arabidopsis Inf. Serv. 5: 23.
Grierson, A. J. C. 1984. Crueiferae. In: A. J. C. Grierson
& 1). G. Long, El. Bhutan 1(2): 416-445. Royal Botanic
Garden Edinburgh.

Hajra. P. K.. H. J. Chowdhery & G. C. Das. 1993. Sis-
ymbrieae. In: B. D. Sharma & N. P. Balakrishnan (ed¬
itors), El. India 2: 224—247. Botanical Survey of India,
Calcutta.

Hara, H. 1979. Crueiferae. In: H. Hara & E. II. J. Wil¬
liams, An enumeration of the flowering plants of Nepal

2. Bull. Brit. Mus. (Nat. Hist.), Bot. 810: 38-46.

Hedge, I. C. & K. H. Rechinger. 1968. Crueiferae. In: K.
H. Rechinger (editor), FI. Iran. 57: 1-372. Akademis-
che Druck- und Verlagsanstalt, Graz, Austria.

Holmgren, P. K., N. H. Holmgren & L. C. Barnett (Edi¬
tors). 1990. Index Herbariorum, 8th ed. New York Bo¬
tanical Garden, New York.

Jafri, S. M. H. 1956. Some Crueiferae of W. Pakistan,
Afghanistan and N.W. Himalaya. Notes Roy. Bot. Card.
Edinburgh 22: 95—119.

- . 1973. Brassicaceae. El. Pakistan 55: 1-308. Eer-

ozsons, Karachi.

Jones, B. M. G. 1964. Arabis. Pp. 290-294 in T. G. Tutin.
V. H. Heywood, N. A. Burges, D. II. Valentine, S. M.
Walters & D. A. Webb (editors), El. Europaea, ed. 1.
vol. 1, Cambridge Univ. Press, Cambridge, U.K.

Junussov, S. J. 1978. Trichochiton. In: P. O. Ovczinnikov
(editor), El. Tajikistan SSR 5: 164—166. Academy of
Sciences, Leningrad.

Komarov, V. 1896. Material to the flora of Turkestan
mountains: Basin of the River Zeravshan. Trav. Soc.
Nat. St. Petersb. 26: 30-162.

MftsiCek, J. & J. Sojak. 1995. Chromosome numbers of
Mongolian angiosperms II. Folia Geobot. Phytotax. 30:
445—453.

Mulligan, G. A. 1995. Synopsis of the genus Arabis (Bras¬
sicaceae) in Canada, Alaska and Greenland. Rhodora
97: 109-163.

Mummenhoff, K. & H. Hurka. 1995. Allopolyploid origin
of Arabidopsis suecica (Fries) Norrlin: Evidence from
chloroplast and nuclear genome markers. Bot. Acta
108; 449-456.

Naqshi, A. R. & G. N. Javeid. 1986 ["1985"]. The genus
Arabidopsis Heynh. (Brassicaceae) in J[ammu] &
K[ashmir[ State. J. Econ. Taxon. Bot. 7: 617—627.

O’Kane, S. L., Jr. & I. Al-Shehbaz. 1997. A synopsis of
Arabidopsis (Brassicaceae). Novon 7: 323-327.

- , - & B. A. Schaal. 1995. Phylogenetics of

Arabidopsis: Scope and content based on DNA sequenc¬
es of nuclear rDNA internal transcribed spacers. Amer.
.1. Bot. 82 (Suppl.): 154.

- , B. A. Schaal & I. Al-Shehbaz. 1997. The origins

of Arabidopsis suecica (Brassicaceae) as indicated by
nuclear rDNA sequences. Syst. Bot. 21: 559—566.

Pachomova, M. G. 1974. Trichochiton. In: A. 1. Vvedensky
(editor). Conspectus Florae Asiae Mediae 4: 152-153.
Academiae Scientiarum UzSSR, Taschkent.

Polatschek, A. 1971. Cytotaxonomische Beitrage zur Flora
Iranica. III. Ann. Naturhis. Mus. Wien 75: 173—182.

Price, R. A., J. I). Palmer & I. A. Al-Shehbaz. 1994.
Systematic relationships of Arabidopsis: A molecular
and morphological perspective. Pp. 7—19 in E. M. Mey-
erowitz & C. R. Somerville (editors), Arabidopsis. Cold
Spring Harbor Laboratory Press, New York.

Rollins, R. C. 1943. Generic revisions in the Crueiferae:
Halimolobos. Contr. Dudley Herb. 3: 241-265.

- . 1952. A note on Halimolobos. Rhodora 54: 161—

163.

- . 1993. The Crueiferae of Continental North Amer¬
ica. Stanford Univ. Press, Stanford.

Schulz, 0. E. 1924. Crueiferae— Sisymbrieae. In: A. Engler
(editor), Pflanzenreich IV. 105(Heft 86): 1-388.

- . 1927. Beitrage zur Kenntnis der Cruciferen des

nordwestlichen Himalay-Gebirges. Notizbl. Bot. Gait.
Berlin-Dahlem 9: 1057-1088.

- . 1936. Crueiferae. In: A. Engler & K. Prantl (ed¬
itors), Nat. Pflanzenfam., ed. 2. I7B: 227-658. Verlag
von Wilhelm Engelmann, Leipzig.

A Revision of Sisymbriopsis (Brassicaceae)

Ihsan A. Al-Shehbaz

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

An Zhengxi

Herbarium, Biological Centre, August 1st Agricultural University, Urumqi, Xinjiang 830052,

People’s Republic of China

Yung Guang

Piroebe Plants, Inc., 20542 McNeil Road, Pitt Meadows, B.C., Canada V3Y 1/1

ABSTRACT. The genus Sisymbriopsis is estab¬
lished, and a key, descriptions, and distributions of
the five species are given. The new combinations
5. pamirica, S. shuanghuica, and S. yechengnica
are proposed.

During work on the Brassicaceae (Cruciferae) lor
the Flora of China , it became evident that the
placement of many Chinese species fluctuated be¬
tween different genera, and no clear lines have
been drawn to delimit them. The present paper
deals with the placement of four such species, plus
a fifth endemic to Tajikistan.

In their original description of Sisymbriopsis , Bot-
schantsev and Tzvelev (1961) recognized a single
species, S. schugnana Botschantsev & Tzvelev,
which was mistakenly treated and illustrated by Vas-
silczenko (1939) as Torularia mollipila (Maximowicz)
0. E. Schulz. Botschantsev (1966) transferred the
latter species to Sisymbriopsis , and subsequent works
(e.g., Bondarenko, 1974; Czerepanov, 1995; Ikonni¬
kov, 1979; Junussov, 1978) recognized only these
two species. Although S. mollipila (Maximowicz)
Botschantsev was originally described from Tibet
(Maximowicz, 1889), Kuan’s (1985) account of the
family Cruciferae for the Flora of Xizang (Tibet) gave
no mention of the species under any genus. By con¬
trast, An (1987) reduced S. mollipila to synonymy of
T. mollipila. He (An, 1991) placed S. mollipila in
Neotorularia Hedge & J. Leonard (= Torularia O.
E. Schulz (1924), not Torularia Bonnemaison (1828)
of Rhodophyta) and recognized (An, 1995) some of
its variants as members of the genera Arabidopsis
(DC.) Heynhold and Microsisymbrium 0. E. Schulz.
Leonard (1986) excluded the species from Neotoru¬
laria and followed Botschantsev (1966) in assigning
it to Sisymbriopsis.

The important points that need clarification are
whether or not Sisymbriopsis should be recognized
and whether .S', mollipila should be kept in Neoto-

Novon 9: 308-312. 1999.

rularia. All species of Neotorularia have terete
fruits, valves with a prominent or obscure midvein,
typically incumbent cotyledons, and nectar glands
consisting of lateral ones only. By contrast, all spe¬
cies of Sisymbriopsis , including .S’, mollipila and the
others herein transferred to this genus, have flat¬
tened fruits somewhat rectangular in c ross section,
prominently 3-veined valves, obliquely aceumbent
cotyledons, and nectar glands consisting of median
nectaries confluent with the lateral ones. In our
opinion, these differences are of considerable sig¬
nificance, and they clearly support the recognition
of Sisymbriopsis as independent from its nearest rel¬
ative, Neotorularia.

Although Botschantsev and Tzvelev (1961) com¬
pared Sisymbriopsis with Sisymbrium , the- two gen¬
era do not appear to be closely related. The pres¬
ence in Sisymbriopsis of branched trichomes,
obliquely aceumbent cotyledons, and flattened
fruits somewhat rectangular in cross section readily
distinguish the genus from Sisymbrium , which has
simple trichomes, incumbent cotyledons, and terete
fruits. Although some of the southern African and
South American species of Sisymbrium have
branched trichomes (Marais, 1970; Romanczuk,
1982; Schulz, 1924), those species are anomalous
in the genus and probably ought to be placed in
other genera.

As delimited here, Sisymbriopsis includes five
species, of which one, S. schugnana , is endemic to
Tajikistan, three are endemic to China, and one
occurs in both countries.

Sisymbriopsis Botschantsev & Tzvelev, Not. Syst.
Herb. Inst. Bot. Acad. Sci. URSS 21: 143.
1961. TYPE: Sisymbriopsis schugnana Bot¬
schantsev A Tzvelev.

Herbs annual, biennial, or perennial. Trichomes
stalked 1- or 2 -forked to dendritic, sometimes

Volume 9, Number 3
1999

Al-Shehbaz et al.

Revision of Sisymbriopsis

309

crisped or flattened, rarely exclusively simple.
Stems erect to ascending, sometimes decumbent,
simple or branched basally and/or apically. Basal
leaves petiolate, pinnately lobed to coarsely den¬
tate, rarely subentire. Cauline leaves petiolate or
subsessile and not auriculate at base, entire, den¬
tate, to pinnatifid. Inflorescences several-flowered,
bracteate or ebracteate corymbose racemes, elon¬
gated slightly or considerably in fruit. Sepals ob¬
long, erect, glabrous or pubescent, base of inner
pair not saccate, margin membranous. Petals white
or lavender, obovate, spatulate, or oblanceolate;
claw usually obscurely distinct from blade. Stamens
6, slightly tetradynamous; filaments dilated or not
dilated at base; anthers ovate or oblong, sagittate
at base, obtuse or apiculate at apex. Nectar glands
confluent and subtending bases of all stamens; me¬
dian nectaries present; lateral nectaries annular.
Fruit linear, flattened and latiseptate, appressed to
raehis or not, sessile; valves papery, with a promi¬
nent midvein and 2 lateral veins and less promi¬
nent, anastomosing secondary veins, sparsely to
densely pubescent with branched trichomes, rarely
glabrescent or with simple hairs, somewhat rect¬
angular in cross section, torulose; septum complete;
style obsolete or short and less than 1 mm long;
stigma capitate, entire or 2-lobed. Seeds uniseriate,
wingless or distally with a small winglike append¬
age, oblong, slightly flattened; seed coat minutely
reticulate, not mucilaginous when wetted; cotyle¬
dons obliquely aecuinbent.

la. Trichomes simple, flattened; petals 6—10 X 3—5

mm . 5. .S'. yechengnica

lb. At least some trichomes branched, not flattened;
petals 2^1 X 0.6-1. 5 mm.

2a. Fruits secund; fruiting pedicels ascending to
divaricate, 5—11 mm long; Tajikistan ....

. 1 . S. schugnana

2h. Fruits not secund; fruiting pedicels erect,
appressed to raehis, 0.5— 5( — 6) mm long;
China, Kyrgyzstan, Tajikistan.

3a. Plants perennial; inflorescence bracte¬
ate; anthers not apiculate .

. 4. S. shuanghuica

3b. Plants annual; inflorescence ebracteate;
anthers apiculate.

4a. Fruit glabrescent or sparsely pu¬
bescent with crisped forked tri¬
chomes to 0.4 mm long; stigma
subentire to slightly 2-lobed; fruit¬
ing pedicel stout; sepals sparsely
pubescent with crisped forked tri¬
chomes . 2. S. mollipila

4b. Fruit densely villous with straight
forked trichomes to 1 mm; stigma
deeply 2-lobed; fruiting pedicels
slender; sepals densely villous with
straight simple trichomes ....
. 3. 5. pamirica

1. Sisymbriopsis schugnana Botschantsev &
Tzvelev, Not. Syst. Herb. Inst. Bot. Acad. Sci.
URSS 21; 145. 1961. TYPE: Tajikistan. W Pa¬
mir, valley Murgab, 5-6 km below fl. W
Pschart, ca. 3300 m, 22 June 1958, N. Tzvelev
256 (holotype, LE; isotype, LE).

Herbs biennial, 8—25 cm tall, densely canescent
with a mixture of simple and 1- or 2-forked or sub-
dendritic, short-stalked trichomes. Stems ascending
to decumbent, branched primarily from or near
base. Basal leaves rosulate, pinnatifid to pinnati-
sect; petiole (0.5— ) 1 — 2( — 3) cm long, not ciliate; leaf
blade oblong to oblanceolate in outline, (0.5 — ) 1 —
3.5 cm long, (2— )5— 10 mm wide, base cuneate,
margin obtusely lobed, the lobes entire or few
toothed, apex obtuse to rounded. Middle cauline
leaves subsessile to short petiolate, similar to basal
leaves but progressively reduced in size upward.
Inflorescence ebracteate, rarely lowermost few flow¬
ers bracteate, elongated considerably in fruit. Ped¬
icels in fruit ascending to divaricate-ascending, of¬
ten secund, straight, slender, 3 — 8( — 14) mm long,
sparsely pubescent with branched trichomes. Se¬
pals oblong, 2-3 X 1-1.5 mm, scarious at margin,
pubescent with branched trichomes. Petals white to
pink, spatulate to oblanceolate, 3-4 X 0.5-1. 5 mm,
attenuate to a clawlike base to 1.5 mm long, apex
obtuse. Filaments white to lavender, 2.5—3 mm
long; anthers ovate, 0.5-0. 6 mm long, apiculate.
Fruit (2-)2.5-3.5(-4) cm long, (1.5— )2-2.5 mm
wide, divaricate, secund; valves sparsely pubescent
with finely branched, crisped forked trichomes,
rarely glabrescent; style 0.2— 0.8 mm long; stigma
subentire to slightly 2-lobed. Seeds yellowish
brown, oblong, 1.4— 1.8 X 0.7-1 mm, slightly com¬
pressed, distal margin sometimes with a minute
wing; cotyledons obliquely accumbent. Flowering
June-August, fruiting July-September.

Distribution. Endemic to Tajikistan at 3600-
4500 m.

Specimens examined. TAJIKISTAN. Pamir: Gorno-
Badakhshan, Sareszskoe Lake, valley of River Ramanf, 5
km from mouth, Stanukovych et al. 9743 (LE); Gorno-
Badakhshan, River Topolovaya, Stanukovych et al. 8675
(LE); Gorno-Badakhshan, River Kazankul, 4 km from
mouth, Stanukovych et al. 9422 (LE).

2. Sisymbriopsis mollipila (Maximowicz) Bot¬
schantsev, Novit. Syst. PI. Vase. 3; 122. 1966.
Sisymbrium mollipilum Maximowicz, Fl. Tan-
gut. 1: 62. 1889. Neotorularia mollipila (Max¬
imowicz) Z. X. An, J. August 1st Agr. Coll.
14(2): 48. 1991. Stenophragma mollipilum
(Maximowicz) B. Fedtschenko, Rastit. Turkest.

310

Novon

457. 1915. Tondaria mollipila (Maximowiez)
0. E. Schulz, Pflanzenreich IV 105(Heft 86):
217. 1924. TYPE: China. Tibet: Nail Shan,
11-12,000 ft., 23 July 1879, N. M. Przewalski
s.n. (lectotype, designated by Vassilczenko
( 1939), LE; isolectotypes, LE, PE).

Arabidopsis qiranica X. X. An, FI. Xinjiang. 2(2): 376.
1995. Syn. nov. TYPE: China. Xinjiang: Oira Xian.
31(H) m. 22 July 1981. An Zhengxi 88 (holotype,
XJ A).

Microsisymbrium taxkorganicum /.. X. An, FI. Xinjiang.
2(2): 380. 1995. Syn. nov. TYPE: China. Xinjiang:
Taxkorgan Xian. 3520 m, 1 July 1978, Xinjiang Ex¬
pedition 958 (holotype, WUk).

Herbs annual or rarely biennial, (3 — ) 1 8 — 45( — 60)
cm tall, densely to rarely sparsely covered basally
with a mixture of simple and 1- or 2-forked, short-
stalked trichomes, sparsely to moderately pubes¬
cent apically with crisped forked trichomes or rare¬
ly glabrescent. Stems erect, simple or sometimes
branched at base. Basal leaves withered by fruiting
time; petiole (0.5 — ) 1 — 1 .7( — 3) cm long, strongly flat¬
tened at base, ciliate with simple trichomes to 1.5
mm long; leaf blade narrowly oblong to linear-lan¬
ceolate, ( 1 — )2 — 4( — 7) cm long, 2— 8( — 12) mm wide,
base cuneate, margin coarsely dentate to pinnatifid,
apex acuminate. Middle cauline leaves subsessile
to short petiolate, linear to linear-lanceolate, 1.5—
5(— 8) cm long, 1— 5(— 10) mm wide, progressively
reduced in size upward, margin entire to dentate,
apex acuminate. Inflorescence ebraeteate, elongat¬
ed considerably in fruit. Pedicels in fruit erect, ap-
pressed to rachis, straight, stout, (0.5— )1— 3(— 4) mm
long, sparsely pubescent with crisped, branched
trie homes. Sepals oblong, (1.5— )2— 2.5 X 0.8-1 mm,
scarious at margin, sparsely pubescent with crisped
branched trichomes. Petals white to lavender, ob-
lanceolate, (2 — )2.S — '4 X (0.6— )0.9— 1.2(— 1.5) mm,
attenuate to a clawlike base to 1.5 mm long, apex
obtuse. Filaments white to lavender. (1.5— )2— 3 mm
long; anthers oblong to subovate, (0.3— )0. 4— 0.6 mm
long, apiculate. Emit (1.5— )3— 5(— 7) cm long, 1—1.2
mm wide, erect, appressed to rachis; valves sparse¬
ly pubescent with finely branched, crisped forked
trichomes to 0.4 mm long, rarely glabrescent; style
obsolete to 0.6 mm long; stigma subentire to slight¬
ly 2-lobed. Seeds yellowish brown, oblong, 0.9— 1.4
X 0.6—1 mm, slightly compressed, distal margin
sometimes with a minute wing; cotyledons oblique¬
ly accumbent. Flowering June— August, fruiting
July— September.

Distribution and habitat. Afghanistan, China
(Oinghai, Xinjiang, Xizang), Kyrgyzstan, Tajikistan.
Exposed banks, gravelly steep slopes, flood land,
alluvial plains, sandy or gravelly beaches, mountain

valleys, grassy slopes, Juniperus woods; 2900—
4500 m.

Sisymbriopsis mollipila is reported here for the
first time from Afghanistan. It has been collected
extensively from Kyrgyzstan and Tajikistan, hut
only a few representative samples are c ited from
these two countries. It is the most widespread and
variable species of the genus. Leaf size and margin,
fruit length, and density of the indumentum on the
upper parts, especially the fruit valve, are the most
variable characters. However, the variation in these
characters is continuous and can be found within
the saint1 population.

The types of Arabidopsis qiranica and Microsi¬
symbrium taxkorganicum are short-fruited forms of
Sisymbriopsis mollipila, with the former being gla-
hrescent whereas the latter is moderately pubes¬
cent. Other features of both types are basically a
perfect match with plants of S. mollipila.

Sisymbriopsis mollipila is most closely related to
.S’, pamirica. Both species have erect fruits ap¬
pressed to the rachis, small llowers to 4 mm long,
and short pedicels rarely reaching 4 mm in length.
However, they c an he* readily distinguished by their
pedicels, stigmas, sepals, and type of trichomes on
the fruit valves. Sisymbriopsis mollipila has stout
pedicels, subentire to slightly 2-lobed stigmas, se¬
pals and fruit valves sparsely pubescent with
crisped, forked trichomes to 0.4 mm long, whereas
S. pamirica has slender fruiting pedicels, deeply 2-
lobed stigmas, sepals with straight simple tri¬
chomes, and fruit valves densely pubescent with
straight, forked trichomes to 1 mm long.

Specimens examined. AFGHANISTAN. Itadukh-
slian: Wakhan, NW Kol-e-Cliaqmaqtin, Anders 7567 ( M ).
CHINA. Qinghai: (Jagan Us river, C. T. Wit, S. K. Chen.
I)u Qing A S. /.. I At 75-207 (HNW I'): Delingha, Baishu
Shan. It. Z. Guo & H. K Wang 11616 (HNWI’); upper
reach of Nonihon He, Du Ding 240 (HNW I’); Qaiclarn,
Dangshe Shan, Qingliai-Gansu Expedition 584 (PE),
Qinghai-Gan.su Expedition 6 02 (PE); Qaiclam, Xilqou,
Qinghai-Gansu Expedition 641 (PE); Qaidam, Dulan,
Qingliai-Gansu Expedition 1681 (PE). Xinjiang: Yeeheng
Xian, Qinghai-Xizang Expedition 870862 (KUN), ). II.
Wu 1017 (HNW P); Ruoqiang Xian. Piaqiriketagor, ,S. G.
Wu, II. ( )h ha, V. II. Wu A Y. Eei 2152 (kllN. El); Ruo¬
qiang. N slope of Oimanlag Mt.. .S'. G. Wu, II. Ohba, Y. II.
Hi/ A F Eei 2686 (TI); Oira, Nur. .S. G. Wu, II. Ohba. ).
II. Wu A Y. Eei 1986 (kUN. Vl(). TI); (lira Xian. Nur.
Yamei, ,S. G. Wu, II. Oliha, V. //. Hi/ A Y. Eei 2520 (kUN.
MO. TI); Yuitian Xian. Pulu kanyan. .S'. G. Hi/, //. Ohba.
). II. Wu A Y. Eei 2779 (TI); Yulian Xian. Pulu Sanchak-
ou, S. G. Wu, II. Ohba, Y. II. Wu A >. Eei 2790 (HNW P.
kUN, TI); Pishati Xian, kaxtax Ruoqiang, S. G. Wu. II.
Oliha. Y. H. Wu A Y. Eei 2646 (1INW1’. MO, TI). K //.
Wu 2647 (HNW P); Taxkorgan Xian. II. S. Li, E. Zheng &
II. D. Zhang 1049 6 (PE). G. L. Zhou (MO); Taxkorgan
Xian. Hongqilapu, .S’. G. Wu, V. II. Wu A 1. Eei 4872
(kUN. TI); 106 km S of Taxkorgan. Kilim T-74 (XJA.

Volume 9, Number 3
1999

Al-Shehbaz et al.

Revision of Sisymbriopsis

311

XJBI). Xizang (Tibet): Burchan Budda, 14 Apr. 1884.
Przewalski s.n. (LE, P). KYRGYZSTAN. Semirechenskaya
Range, River Aksay, 25 July 1903, Lipsky s.n. (LE); Na-
rinsky Range, basin of River Aksay, Sovetkina & Uspen¬
skaya 1796 (LE); Prezwalski, River kujandi, 30 July
1913, Schischkin s.n. (LE); Tian Shan, bank of River
Ichiktor, 27 July 1960, Kozshevnikova s.n. (LE). TAJIKIS¬
TAN. Pamir & Tian Shan, 1906, Appelton s.n. (K); near
Bordaba. Alexeenko 2059 (LE).

3. Sisymbriopsis pamirica (Y. C. Lan & Z. X.

An) Al-Shehbaz, Z. X. An & G. Yang, comb,
nov. Basionym: Arabis pamirica Y. C. Lan &
Z. X. An, FI. Xinjiang. 2(2): 375. 1995. TYPE:
China. Xinjiang: Taxkorgan Xian, 3700 m, 5
July 1978, Xinjiang Expedition 1031 (h< do-
type, WUK).

Herbs annual or? perennial, ca. 30 cm tall,
densely villous throughout with straight, 1-forked,
stalked and simple trichomes. Stems erect,
branched at base and above. Basal leaves unknown;
lower and middle cauline leaves short petiolate,
lanceolate to oblong-lanceolate, 1-3 cm long, 4—10
mm wide, base cuneate, margin coarsely dentate to
pinnatifid, apex acuminate. Inflorescence ebracte-
ate, elongated considerably in fruit. Pedicels in
fruit erect, appressed to rachis, straight, slender, 2—
4 mm long, densely covered with simple and
branched trichomes. Sepals oblong, 1.5—2 X ca.
0.5 mm, scarious at margin, densely pubescent with
straight simple trichomes. Petals lavender, oblan-
ceolate, 2—2.5 X ca. 0.7 mm, attenuate to base,
apex obtuse. Filaments lavender, 1.2—2 mm long;
anthers suboblong, ca. 0.5 mm long, minutely apic-
ulate. Fruit 3-4.7 cm long, 1-1.1 mm wide, erect,
appressed to rachis; valves densely villous with
straight, stalked, forked trichomes to 1 mm long;
style obsolete; stigma deeply 2-lobed. Seeds yellow¬
ish brown, oblong, 0.8—1 X 0.6-0. 7 mm, slightly
compressed; cotyledons obliquely accumbent.
Flowering and fruiting July.

Distribution. Endemic to Xinjiang (China) and
known only from the type, which was collected at
ca. 3700 m.

The original description of Anibis pamirica (An,
1995) indicated that the type was collected in Au¬
gust rather than July, contrary to what the holotype
label shows.

4. Sisymbriopsis shuanghuiea (Kuan & Z. X.

An) Al-Shehbaz, Z. X. An & G. Yang, comb,
nov. Basionym: Torularia shuanghuiea Kuan &
Z. X. An, FI. Xizang. 2: 404. 1985. TYPE:
China. Xizang (Tibet): Shuanghu, 4950 m, 29
July 1976, Ling Kai-yong 9891 (holotype, PE;
isotypes, HNWP, KUN, PE).

Herbs perennial with a slender, simple or
branched caudex, (5—) 1 8 — 45(— 60) cm tall, canes-
cent, densely tomentose throughout with short-
stalked, dendritic and fewer 2-forked trichomes.
Stems erect to subdecumbent, branched basally
and above. Basal leaves not rosulate; petiole 5-10
mm long, not flattened, ciliate at base; leaf blade
spatulate to oblanceolate, 0.5-2 cm long, 1.5-5 mm
wide, base cuneate, margin coarsely dentate to re-
pand or entire, apex obtuse to subacute. Middle
cauline leaves subsessile to short petiolate, similar
to basal ones but progressively smaller upward. In¬
florescence bracteate at least along lower half, elon¬
gated slightly in fruit. Pedicels in fruit erect, ap¬
pressed to rachis, straight, slender, 3— 5( — 6) mm
long, densely tomentose as rest of plant. Sepals ob¬
long, 2-2.5 X ca. 1 mm, scarious at margin, dense¬
ly pubescent. Petals white to lavender, spatulate,
2. 5-3. 5 X 0.9-1. 5 mm, attenuate to a clawlike
base to 1 mm long, apex obtuse. Filaments laven¬
der, 2-3 mm long; anthers subovate, 0.4-0. 5 mm
long, not apiculate. Fruit 1.5—3 cm long, 1-1.3 mm
wide, erect, appressed to rachis; valves densely pu¬
bescent with finely branched, short-stalked den¬
dritic trichomes to 0.5 mm long; style to 0.5 mm
long; stigma subentire to slightly 2-lobed. Seeds
yellowish brown, oblong, 0.8-1 X 0.5-0. 7 mm,
slightly compressed; cotyledons obliquely accum¬
bent. Flowering and fruiting July.

Distribution and habitat. Endemic to Xizang,
China. Mountain slopes, sandy areas; 4800—4900 m.

Because of the limited collections available for
this study, little can be said about the variability of
Sisymbriopsis shuanghuiea. This species and S.
mollipila were placed in Neotorularia by An
(1991), but the presence in both species ol typically
flattened fruits, strongly 3-veined valves, obliquely
accumbent cotyledons, and median nectaries clear¬
ly support their placement in Sisymbriopsis. Sis¬
ymbriopsis shuanghuiea differs from the remaining
species of the genus on account of its perennial
habit and bracteate inflorescences. However, these
two characters alone do not carry much weight in
the delimitation of genera in the family.

Two sheets at PE were marked as the type of
Sisymbriopsis shuanghuiea. The one carrying the
PE number 1172825, which is the more complete,
is taken here as the holotype, and that numbered
1172826 is annotated as the isotype.

Specimens examined. CHINA. Xizang (Tibet): Shu¬
anghu, 4800 m. Anonymous 11990 (PE).

312

Novon

5. Sisymbriopsis yechengnica (Z. X. An) Al-
Shehbaz, Z. X. An & G. Yang, comb. nov. Ba-
sionyrn: Microsisymbrium yechengnicum Z. X.
An, Bull. But. Res. North-East Forest. Inst.
1(1-2): 99. 1981. TYPE: China. Xinjiang:
Yeeheng Xian, Kunlun Shan, 1 July 1974,
3000 m, Anonymous 3347 (holotype, HNWP;
isotype, PE).

Herbs annual or perennial, 15—50 cm tall, gla-
brescent to sparsely pubescent with flattened,
crisped, simple trichomes to 0.7 mm long. Stems
erect to subdecumbent, simple or many-branched
at base. Basal leaves withered by fruiting time, not
rosulate; petiole 0.5—3 cm long, not flattened at
base; leaf blade narrowly oblong to linear-lanceo¬
late, 2-4 cm long, 1-5 mm wide, base cuneate,
margin coarsely dentate to pinnatifid and lobes to
5X1 mm, rarely subentire, apex acuminate. Mid¬
dle cauline leaves subsessile to petiolate, similar to
basal leaves but progressively reduced in size up¬
ward, 3—7 cm X 2-6 mm, margin entire to dentate,
apex acuminate. Inflorescence ebracteate, elongat¬
ed considerably in fruit. Pedicels in fruit divaricate
to ascending, recurved to straight, slender, 0.7—
1.8(— 2.5) cm long, glabrescent or sparsely pubes¬
cent with crisped, simple trichomes. Sepals oblong,
3—4 X 1.5—2 mm, scarious at margin, glabrous.
Petals white to lavender, obovate, 6—10 X 3—5 mm,
cuneate to a claw 1-2 mm, apex obtuse. Filaments
white, 2.5— 3.5 mm; anthers narrowly oblong, 0.9—
1.1 mm long, not apieulate. Fruit (1.5— )2.5— 4 cm
long, 1-1.2 mm wide, spreading; valves glabres¬
cent, rarely sparsely pubescent with crisped simple
trichomes to 0.5 mm long; style 0.4— 0.6(— 0.9) mm
long; stigma entire. Seeds yellowish brown, oblong,
1. 2-1.7 X 0.8—1 mm, slightlv compressed, distal
margin with a minute wing; cotyledons obliquely
accumbent. Flowering June— July, fruiting July— Au¬
gust.

Distribution and habitat. Endemic to Xinjiang,
China. Grasslands, mountain slopes, sandy areas,
steep rocky river banks; 2500—3000 m.

Two collections, Wu et al. 1842 and 2410 , differ
from the type collection and other examined spec¬
imens in being coarse perennials and many-
branched at base (instead of smaller annuals and
unbranched at base). Both collections do not have
fully mature fruits, and they might represent an in¬
dependent infraspecific taxon of Sisymbriopsis yech-
engnica. However, in their leaves, flowers, tri¬
chomes, and pedieel morphology they are
indistinguishable from typical plants of the species.

Sisymbriopsis yechengnica is readily distin¬
guished from the other species ol the genus by its

simple, flattened, crisped trichomes and large pet¬
als 6—10 X 3—5 mm. All other species of Sisym¬
briopsis have branched trichomes and petals 2—1 X
0.6— 1.5 mm. However, in all other characters, es¬
pecially fruits, nectar glands, foliage, and cotyle¬
donary position, .S’. yechengnica is perfectly at home
in Sisymbriopsis.

Specimens examined. CHINA. Xinjiang: Pishan, Bu-
qun. .S'. G. Wu, II. Ohba, Y. II. Wu & 1. Fei 1842 (kUN.
MO, TI); Pishan Xian, Naoarbati, .S. G. Hi/. II. Ohba. V.
II. Wu & >; Fei 2410 (HNWP, kUN, MO, I I); 10 km N
of kalmakkuouk on Xinjiang— Xizang Hwy., Anonymous
523 (kl!N. PK); Yeeheng Xian. Qinghai-Xizang Expedi¬
tion 870754 (kl N).

Acknowledgments. We are grateful to Tatiana
Shulkina for her translation of some of the Russian
literature and to Vladimir Dorofeev for his review
of the manuscript.

I .iterature ( bled

An. Z. X. 1987. Sisymbrieae. In: T. Y. Cheo (editor). FI.
Reipubl. Popularis Sin. 33: 396—453. Science Press.
Beijing.

- . 1991. Emendation about plant names of Neoto-

rularia Hedge et J. Leonard of Chinese Crueiferae. J.
August 1st Agr. Coll. 14(2): 46-48.

- . 1995. Crueiferae. In: Z. M. Mao (editor), FI. Xin¬
jiang. 2(2): 38—229. 374-381. Xinjiang Science &
Technology & Hygiene Publishing House, IJrumqi.
Bondarenko, (). N. 1974. Sisymbriopsis. In: A. I. Vved¬
ensky (editor). Conspectus florae Asiae Mediae 4: 55.
Academiae Seientiarum UzSSIL Taschkent.
Botsehantsev, V. 1966. De Cruciferis notae eritieae, 5.
Nov. Syst. PI. Vase. 3: 122-139.

- X N. Tzvelev. 1961. Genus novum Sisymbriopsis

nob. e fain ilia Crueiferae. Not. Syst. Herb. Inst. Hot.
Acad. Sci. URSS 21: 143-145.

Czerepanov, S. k. 1995. Vascular Plants of Russia and
Adjacent States (the Former USSR). Cambridge Univ.
Press, Cambridge.

Ikonnikov, S. S. 1979. Definitorium Planarum Vascicular-
ium Badaehschaniae. Academy of Sciences, Leningrad.
Junussov, S. J. 1978. Sisymbriopsis. In: P. (). Ovezinnkov
(editor), FI. Tajikistan SSR 5: 32—34. Academy of Sci¬
ences. Leningrad.

kuan, k. C. 1985. Crueiferae. In: C. Y. Wu (editor), FI.

Xizang. 2: 323—14 I. Science Press. Reijing.

Leonard. J. 1986. Neotorularia Hedge & J. Leonard nom
generiqne nouveau de Crueiferae. Hull. Jard. Rot. Nall.
Belgique 56: 389-395.

Marais. \\. 1970. Crueiferae. In: I.. F. Codd. B. de Winter.
I). .1. B. killiek & II. B. Rycroft (editors), FI. Southern
Africa 13: 1-118. Government Printer, Pretoria.
Maximowicz, C. J. 1889. Flora Tangutica. Vol. I. Imperial
Academy of Sciences, St. Petersburg.

Romanczuk. M. (7 1982. FI geriero Sisymbrium (Crucifer-
ae) en la Argentina. Darwiniana 24: 75—156.

Schulz. (). F. 1924. Crueiferae-Sisymbrieae. In: A. Engler
(editor), Pflanzenr. IV 105(Heft 86): 1—388.
Vassilczenko. I. T. 1939. Torularia. In: V. L. komarov (ed¬
itor). FI. URSS 8: 59—69. Academiae Seientiarum
URSS. Moscow & Leningrad.

A New Species of Ruprechtia (Polygonaceae) from the
Venezuelan Guayana

Gerardo Aymard C.

UNELLEZ-Guanare, Programa de R. N. R., Herbario Universitario (PORT), Mesa de Cavacas,

Edo. Portuguesa, Venezuela 2.333

Paul E. Berry

Botany Department, University of Wisconsin, 132 Birge Hall, 430 Lincoln Drive,
Madison, Wisconsin 53706, U.S.A.

ABSTRACT. Ruprechtia howardiana, from decidu¬
ous forests in Bolivar state, Venezuela, is newly de¬
scribed and illustrated. It appears most closely re¬
lated to R. ramiflora (Jaequin) Meyer.

Ruprechtia is a genus of small trees or shrubs
from neotropical dry and riverine forests. The genus
has been closely allied with Triplaris, hut differs in
its 2- or 3-ffowered pistillate partial inflorescences,
a basal pedieel-like extension of the perianth tube
in fruit, and pedicellate male flowers with perianth
segments connate for less than a third of their
length (Brandbyge & 011gaard, 1984; Brandbyge,
1986). It also typically has fewer pairs of secondary
leal veins (less than 12) than Triplaris , smaller
leaves, and achenes loosely enclosed within the
fruiting perianth.

The species described below was treated by
Howard (1985: 507) as an “unplaced collection,” a
probable new species for which he wanted addi¬
tional specimens before describing. Although it is
still known only from staminate material, the spec¬
imens are distinct enough from other members of
the genus for us to describe it at this time, making
it available lor inclusion in the upcoming Polygon¬
aceae treatment for the Flora of the Venezuelan
Guayana.

Ruprechtia howardiana Aymard & P. E. Berry,
sp. nov. TYPE: Venezuela. Bolivar: middle Rio
Orinoco, Cerro Baraguan and vicinity, dry for¬
est, 100—330 m, 13 Jan. 1956, Wurdack &
Monachino 41230 (holotype, VEN; isotypes,
HB. MO, NY, US). Figure l.

Species nova quae R. ramiflorae (Jaequin) Meyer affinis
sed ah ea foliis subtus inflorescentiaque dense hirsutis,
costa supra depressa atque sepalis ovatis usque obovatis
differt.

Dioecious tree ca. 3 m tall; twigs with solid in¬

ternodes. Branchlets densely pubescent, becoming
glabrescent when mature, lenticellate; leafy inter¬
nodes 1.5 — 3( — 4) cm long. Leaves simple, alternate,
ovate, chartaceous, 7—16 X 4—8 cm, the base acute
or slightly unequal, the apex acuminate, margins
subrevolute, the midrib depressed above, glabrous
on the upper surface except the midrib and sec¬
ondary nerves covered by a gray appressed pubes¬
cence, brownish hirsute on the lower surface, more
densely so on the midrib and secondary nerves;
lateral nerves 10-15 on each side, tertiary venation
conspicuous on both sides; petioles 4—6 mm long,
densely yellow pubescent; stipules forming a short
(1 mm) tube around the stem and leaving a ring
around the stem. Inflorescence axillary racemes or
dense clusters of racemes 2-9 cm long; pedicels
1.5-3 mm long, sparsely appressed pubescent to
glabrous; staminate flowers white, sepals 3, ovate
to obovate, 1—2 mm long, membranaceous, glabrous
on both sides, ciliate at the margins; petals 3, sim¬
ilar to the sepals and alternate with them; stamens
7—9, filaments glabrous, 2—3 mm long; anthers gla¬
brous, ca. 0.8 mm long; oblong, pistillode absent;
pistillate flowers and fruit not seen.

Ruprechtia howardiana is most closely allied to
R. ramiflora from Venezuela, Colombia, Guyana,
and Brazil but differs in its inflorescence and lower
surface of the leaves densely hirsute, the midrib
impressed on the upper surface, and the sepals
ovate to obovate. Ruprechtia ramiflora occurs in a
wide variety of habitat types but has glabrous to
sparsely pubescent leaves anil inflorescences, the
midrib elevated on the upper leaf surface, and the
sepals spathulate.

This species is named after Richard Alden How¬
ard (1917 — ) for his previous work on Triplaris, Coc-
coloba, and other members of the Polygonaceae.

Acknowledgments. We are grateful to Roy Ger-

Novon 9: 313-314. 1999.

314

No von

Figure 1. Ruprechiia howardiana Aymard & P. E. Berry. — A. Flowering braneh. — IE Detail of male flower. Based
on Wurdark A' Monnrhino 11230.

eau lor revising the Latin diagnosis and to Bruno
Manara for preparing the illustration.

Literature Cited

Brandbyge, .1. 1986. A revision of the genus Triplaris (Po-
lygonaceae). Nordic J. Hot. 6: 545—570.

- & B. 011gaard. 1984. Inflorescence structure and

generic delimitation of Triplaris and Ruprechiia (Poly-
gonaceae). Nordic J. Bot. 4: 765—769.

Howard, If. 1985. The “ Triplaris scandens (Veil. Cone.)
Cocucci" complex (Polygonaceae). J. Arnold Arbor. 66:
508-508.

Rudgea tayloriae (Rubiaceae), an Unusual New Species from the
Eastern Slopes of the Venezuelan Andes

Gerardo Aymard C.

UNELLEZ-Guanare, Programa de Recursos Naturales Renovables, Herbario Universitario
(PORT), Mesa de Cavacas, Edo. Portuguesa 3323, Venezuela

L. J. Dorr

Department of Botany, MRC-166, National Museum of Natural History, Smithsonian Institution,

Washington, D.C. 20560-0166, U.S.A.

Nidia Cuello

UNELLEZ-Guanare, Programa de Recursos Naturales Renovables, Herbario Universitario
(PORT), Mesa de Cavacas, Edo. Portuguesa 3323, Venezuela

Abstract. Rudgea tayloriae (Rubiaceae), a new
species from the eastern slopes of the Venezuelan
Andes, is described and illustrated, and its rela¬
tionships are discussed.

Rudgea Salisbury is a neotropical genus of ea.
150-170 species (Dwyer, 1980; Taylor et ah, 1991;
Taylor, 1996), most of which are South American.
Anderssons (1992) checklist of neotropical Rubi¬
aceae included 214 species names, but his objec¬
tive in compiling the list was not to make taxonomic
decisions, and a number of the names he cited are
clearly synonyms. Steyermark (1974) recognized 28
species ot Rudgea in his treatment of the Venezue¬
lan Rubiaceae. Subsequently, Steyermark (1981,
1988) described an additional four species oi Rud¬
gea from Venezuela.

The valvate corolla lobes, 2-celled ovary, and
fleshy drupe with two pyrenes place Rudgea in the
subfamily Rubioideae tribe Psychotrieae A. Richard
ex Dumortier (Robbrecht, 1988; Taylor, 1996).
Rudgea is similar to Palicourea Aublet and Psy-
ehotria L, but it usually can be separated from
these closely related genera by the pectinate or fim¬
briate appendages on its stipules (Steyermark,
1974; Standley & Williams, 1975; Dwyer, 1980;
Taylor, 1996). This stipule character, however, can
present practical problems in identifying material
since the stipules frequently are caducous, as has
been noted earlier by Taylor et al. (1991).

Rudgea tayloriae Aymard, Dorr & Cuello, sp. nov.
TYPE: Venezuela. Trujillo: Dtto. Bocono,
Parque Nacional Guaramacal, selvas nubladas
de la vertiente sur, aprox. 9°12'45"N,

70°09'51"W, 2300 m, 3-5 Jan. 1996 (fl, fr), N.
Cuello, R. Gonzalez, C. Orda & A. Gonzalez
1350 (holotype, PORT; isotypes, MO, US,
VEN). Figure 1.

Species nova quae a Rudgea hospe Standley & Steyer¬
mark floribus sessilibus, corollis purpureis, tubo 3-4 mm
longo, lobis 1—1.5 mm longis, filamentis ca. I mm longis.
papillatis, basi tubi affixis, anlheris ca. 2.5—3 mm longis.
stigmate furcato differt.

Small trees to 8 m tall, branches and branchlets
glabrous throughout. Leaves opposite, decussate;
blades elliptic to ovate-elliptic, 5—11 cm long, 2.5-
7 cm wide, acuminate at apex with the tip up to 2-
2.3 mm long, acute at base, margin entire, glabrous
and well reticulated on upper surface, glabrous un¬
derneath, except along the midrib and lateral
nerves, which are sparsely appressed-pubescent to
puberulent when mature, lateral nerves 7-10 pairs,
convergent and linking toward margin; petioles 5-
13 mm long, ca. 1 mm thick, subcanaliculate,
sparsely appressed-pubescent at first, glabrous
when mature. Stipules interpetiolar, ciliate, cadu¬
cous, briefly persistent only on uppermost 1—2
nodes, united around the stem into a cupuliform
structure, 1.5-4 mm long, rounded to truncate api-
cally, sparsely appressed-pubescent at first, gla¬
brous when mature, with 6—8 projections ca. 1 mm
long. Inflorescences terminal, spiciform to subcap-
itate, 1.5— 2.5 cm long (or unexpanded?), ap¬
pressed-pubescent with white trichomes; bracts
subtending the flowers subulate, 4-8 mm long, apex
with the tip 0.5-2 mm long, shortly appressed-pu¬
bescent externally, glabrous internally. Flowers ses¬
sile, calyx ca. 2 mm long, cupuliform, puberulent
to densely appressed-pubescent at the base exter-

Novon 9: 315-317. 1999.

316

Novon

Eigure 1. Rudgea tayloriae Aymard, Dorr & Cuello. — A. Habit. — B. Stipule. — C. Inflorescence (immature). — I).
Corolla (immature). — E. Corolla, opened to show anthers (immature). — E. Style and stigma. — C. Calvx. — H. Lon¬
gitudinal section through calyx and hypanthium. — I. Emit. A, C— I drawn from Cuello ft til. 1350 (PORT) and It from
Cut’ll o ft al. 1310 (US).

Volume 9, Number 3
1999

Aymard et al.

Rudgea tayloriae from Venezuela

317

nally, puberulent internally, 5-lobed, the lobes ca.
0.5 mm long, triangular, persistent in fruit; corolla
purple, evidently funnelform. but shortly rounded
or constricted at the apex (mature flowers not seen),
the tube 3—4 mm long, short appressed-pubescent
to puberulent externally, glabrous internally, 5-
lobed, the lobes 1—1.5 mm long, pubescent as the
tube; stamens 5, inserted at the base of the tube
(in immature flowers?), filaments ca. 1 mm long,
glabrous, papillate, anthers 2.5—3 mm long, oblong,
glabrous; style ca. 2 mm long, barbate apically,
stigma bifurcated, barbate basally, 1 mm long. Fruit
ovoid, black, 9—10 mm long, 6—7 mm wide, sparse¬
ly pilose to glabrous when mature, subtended bv
one bract.

Distribution and habitat. Rudgea tayloriae is
known only from the type locality on the eastern
slopes of the Andes in Trujillo state, Venezuela,
where it has been collected in moist cloud forest
from 1950 to 2300 m. Flowers and fruit are known
only from January.

Rudgea tayloriae appears to be most similar to
R. hospes Standley & Steyerinark, which also is en¬
demic to the Venezuelan Andes. The lateral
branches of the inflorescences of both species are
not developed, and the lateral nerves on the upper
leaf surfaces are convergent and linking toward the
leaf margins. Rudgea tayloriae differs from R. hos¬
pes in having sessile (vs. shortly pedicellate with
pedicels 1-2.5 mm long) flowers, purple (vs. white)
corollas, smaller (3-4 mm vs. ca. 7 mm long) co¬
rolla tubes, smaller (1—1.5 mm vs. ca. 5.5 mm long)
corolla lobes, shorter (ca. 1 mm vs. 2.8—3 mm long)
filaments, and shorter (2.5—3 mm vs. ca. 5 mm long)
anthers. Also, the filaments of R. tayloriae are pa¬
pillate (vs. smooth), evidently inserted at the base
(vs. middle) of the corolla tube, and the stigmas are
bifid (vs. undivided).

The appendages on the corolla lobes of Rudgea
tayloriae are distinctive and further serve to distin¬
guish this new species from R. hospes, which lacks
appendages on the corolla lobes. The purple corol¬
las of R. tayloriae are remarkable in that the ma¬
jority of species in the genus seem to have white

corollas. In addition, the black fruits are unusual
in the genus, which usually has white fruits at ma¬
turity (Taylor, 1996).

We take pleasure in naming this new species af¬
ter Charlotte M. Taylor (b. 1955), who initially drew
the attention of the first author to its distinctive
characters, and who has made significant contri¬
butions to our knowledge of neotropical Rubiaceae.

Paratypes. VENEZUELA. Trujillo: Dtto. Bocond,
Parque Nacional Guaramacal, selvas nubladas de la ver-
liente snr, 9°12'28"N. 70°09'41"W, 1950 m, 19-22 Dec.
1995 (st), Cuello et al. 1310 (PORT. US), 1312 (MO.
PORT, US), aprox. 9°12'45"N, 70°09'51"W, 2300 m, 3-5
Jan. 1996 (st), Cuello et al. 1331 (MO. PORT, US), aprox.
9°13'N, 70°07’W, 2100 m, 20-22 Dec. 1996 (st), Licata
et al. 673 (PORT US).

Acknowledgments. We thank Roy Gereau (MO)
for helping with the Latin description, Cathy Pas-
quale for illustrating the new species, Frederic
Piesschaert (LV) for critically reviewing the man¬
uscript, INPARQUES for granting permission to
work in Guaramacal, and the Wildlife Conservation
Fund of the National Museum of Natural History
(supported by Molson Breweries U.S.A., Inc.) for
funding our fieldwork.

Literature Cited

Andersson, L. 1992. A provisional checklist of neotropical
Rubiaceae. Scripta Bot. Belg. 1: 3—199.

Dwyer. .1. I). 1980. Rubiaceae — Part II. In: R. E. Wood-
son, Jr. & R. W. Schery (editors), Flora of Panama, Part
IX. Ann. Missouri Bot. Card. 67: 257-522.

Robbrecht, E. 1988. Tropical woody Rubiaceae. Opera
Bot. Belg. 1: 1-271.

Standley, P. C. & L. 0. Williams. 1975. Rubiaceae. Flora
of Guatemala, part. XI, nos. 1—3. Fieldiana, Bot. 24(11,
1-3): 1-274.

Steyermark, J. A. 1974. Rubiaceae. In: T. Passer (editor),
Flora de Venezuela 9(2): 603—1101.

- . 1981. New species of Rubiaceae from Venezuela.

Pittieria 9: 5-16, tt. 1-3.

- . 1988. Flora of the Venezuelan Guayana — IV.

Ann. Missouri Bot. Card. 75: 31 1—351.

Taylor, C. M. 1996. Overview of the Psychotrieae (Rubi¬
aceae) in the Neotropics. In: E. Robbrecht, C. Puff &
E. Smets (editors). Second International Rubiaceae
Conference Proceedings. Opera Bot. Belg. 7: 261-270.

- , J. Gomez-Laurito & J. Dwyer. 1991. Rudgea

monofruetus (Rubiaceae), an unusual new species from
Costa Rica. Novon 1: 50—51.

Cirsium ochrocentrum subsp. martinii (Asteraceae), a New
Subspecies of the Santa Fe Thistle from the Gila River Region of

Arizona and New Mexico

Patricia Barlow-hick

Department of Biology, University of New Mexico, Albuquerque, Newr Mexico 87131, U.S.A.

Abstract. A morphological study of Cirsium
ochrocentrum supports separation of the phenotype
with red corollas of the upper Gila River drainage
from typical materials with purple or white corol¬
las from surrounding regions. This taxon shows in¬
tergradation and overlap of character distributions
with typical populations and is recognized as a
subspecies. Cirsium ochrocentrum subsp. martinii
differs from the typical variety by its red corolla
color, short style tips and spines, and generally
shorter decurrence.

The American thistle genus Cirsium, with over
140 species (Ownbey et al., 1975), is an important
part of the North American flora. Although Amer¬
ican thistles have been falsely maligned as aggres¬
sive weeds because of their Eurasian congeners
(Lamp & McCarty, 1981), western thistles provide
a vital, and sometimes obligate, food source for
many animal species including goldfinches, hum¬
mingbirds, and crescent butterflies (Martin et al.,
1951; Tilden & Smith, 1986). Despite the ecolog¬
ical importance of these plants, taxonomists have
largely neglected thistles. Current thistle taxonomy
has both a confusing number of synonyms and a
multitude of undescribed phenotypes that merit
taxonomic recognition. In this report, I describe one
such unrecognized group.

Ownbey and Hsi (1963) suggested that the red-
corolla phenotype of the Santa Fe Thistle, Cirsium
ochrocentrum, might warrant recognition at the sub¬
species or species level. Hsi (1960) proposed a new
taxon based on materials from Arizona with red co¬
rollas but did not validate the name with a pub¬
lished description. Barlow (1992) studied variation
within the C. ochrocentrum group and showed that
tin* red-flowered phenotype, which is found in the
Gila River drainage of southeastern Arizona and
southwestern New Mexico, was separable from the
normally purple-flowered populations of the Rio
Grande, Great Plains, Trans-Pecos, and Colorado
Plateau. A summary of that analysis and the re¬
sulting description is presented here.

Mt/mons AND M VI'KRIALS

Twelve morphological characters were measured
on field-collected specimens from 56 populations
(Fig. 1). Voucher specimens are deposited at UNM.
The morphological characters used in this analysis
were terminal leaf spine length, decurrence length,
leaf length, corolla tube length, corolla throat
length, corolla lobe length, style tip length, anther
appendage tip length, anther appendage body
length, pappus length, phyllary length, and phyllary
spine length. Each population was represented by
1 to 18 specimens, and each floral character was
represented by the average of 10 measurements per
plant. This data set anil the characters are fully
described in Barlow (1992).

A canonical discriminant analysis (CDA), using
population averages, was used to graphically por¬
tray the distinctiveness of the red-corolla popu¬
lations of the Gila River drainage from the typical
Rio Grande valley purple-corolla populations. A
third group, consisting of specimens from geo¬
graphically intermediate populations with inter¬
mediate colors, was included to both verify their
intermediacy in other character states and to offset
the polarization ol a CDA with only two taxa. In¬
dividuals were assigned to one of these three
groups for this analysis. Discriminant analysis
gives a linear combination of characters that best
describes the classification. Ninety-five percent
probability ellipses for the variance of each pop¬
ulation were calculated. In a second analysis, 65
purple-corolla specimens from the upper Rio
Grande Valley near the type locality of the typical
subspecies were compared with 49 red-flowered
specimens from Catron County, New Mexico, in a
Kruskal- Wallis non-parametric comparison of the
means. Statistical analysis was performed with
NEAR 1 WAY, CANDISC, and IML procedures of
the Statistical Analysis System (SAS Institute,
1989). Two hundred twelve herbarium specimens
of Cirsium ochrocentrum were examined.

Nov ON 9: 318-322. 1999.

Volume 9, Number 3
1999

Barlow-lrick

Cirsium ochrocentrum subsp. martinii

319

Figure I. Distribution of populations under study.

Results

The CDA procedure separated the red (Gila) and
purple (Rio Grande) phenotypes except for a small
intersection (Fig. 2). These two color phenotypes
separated along the first canonical axis, with the
geographical intermediates placed near their inter¬
section. The canonical correlations of the first axis
are highly significant at P < 0.0001, while those
of the second axis are not significant (Table 1).

The first canonical axis explains 92.7% of the to¬
tal variance between phenotypes. This axis contains
a linear combination of all quantitative characters
with length of both anther appendage variables,
phyllary spine length, and style tip length being the
most prominent factors (Table 2). The Kruskal-Wallis
comparison of means for the Gila and Rio Grande
phenotypes results in significant differences in 10
characters although there is extensive overlap in the
distribution of characters (Table 3). Pappus length
and phyllary length are not significantly different be¬
tween these phenotypes.

Discussion

These analyses support the recognition of Cir¬
sium ochrocentrum subsp. martinii by showing that
using quantitative morphological characters alone,
the two taxa can be separated with very little over¬
lap when geographically intermediate populations
are excluded. The small degree of overlap between
the taxa, despite the power of CDA to divide
groups, indicates that they are not divided by any
discontinuous character distributions, as evi¬
denced in Table 3. Although these taxa have dif¬
ferent corolla colors and morphological features,
they apparently intergrade in the intermediate
populations along the narrow zone of contact be¬
tween their geographical ranges. Some authors
might rank these two taxa as separate species;
however, a conservative approach, ranking them
as subspecies, better reflects the close relationship
and high degree of morphological similarity, rel¬
ative to more differentiated congeners. The sub¬
specific rank denotes morphologically distinguish¬
able, geographically separate groups within
species (Davis, 1978).

320

Novon

95% Confidence Ellipses

. Limits for purple phenotype

Cirsium ochrocentrum A. Gray subsp. martiiiii
P. Barlow, subsp. nov. TYPE: U.S.A. New Mex¬
ico: Catron County, TBS, R18W sec. 10,
108.203°W, 33.768°N, 5800 ft. elev., 5 mi. E
of the San Francisco River bridge at Reserve,
New Mexico, along N.M. Highway 12, 28 July
1997; Barlow 97-8 (holotype, US; isotypes,
ASU, MO, NMC, UNM).

Subspecies haec a subspecie typica differt corollis pu-
niceis, styli apice supra uodum breviore (saepe minus
quam 5 mm longo), et anllierarum appemlicum eorporibus
plerumque plus quam 2 mm longis (1.5— 2.8 mm).

Perennial caulescent herbs to 90 cm. Stems
moderately branched, striate-ribbed with a dense,
lanate covering ol white appressed hairs. Leaves
lanceolate, ascending or sometimes deflexed in age;

Table 1. Significance of canonical correlations be¬
tween Gila, Santa Fe, and intermediate phenotypes.

Test of HO: The canonical correlations in the current row
and all that follow are zero
Likelihood Num

ratio Approx F OF Den OF Pr > F

1 0.14728959 5.0845 24 70 0.0001

2 0.75470889 1.1523 II 39 0.3502

basal leaves much larger than upper leaves. Leal
surface densely white tomentose on abaxial side;
adaxial surface loosely arachnoid. Cauline leaves
mostly less than 10 mm in decurrence; basal leaves
not decurrent, petiolate. Leaf blades pinnatifid
shallowlv or nearly to the midrib; pinnae cleft into
4 or 5 angular-tipped lobes each terminating in a
yellow-colored spine (2— )5(— 12) mm long. Heads
solitary, borne in lealy panicles. Involucre globose
in bud, campanulate at maturity, 25—45 mm high.

Table 2. Total canonical structure between Gila, Santa
Fe, and intermediate phenotypes.

Character

CAM

CAN2

Corolla tube length

0.206

0.210

Corolla throat length

-0.378

-0.016

Corolla lobe length

-0.432

0.172

Style tip length

0.546

-0.01 1

Anther appendage tip length

0.620

0.184

Anther appendage body length

-0.693

-0.103

Pappus length

-0.091

0.470

Pin Man length

0.214

0.223

Phvllarv spine length

0.560

0.436

Terminal leaf spine length

0.444

0.471

Decurrence length ol leaf base

0.423

0.201

Leaf length

-0.293

0.223

Volume 9, Number 3
1999

Barlow-lrick

Cirsium ochrocentrum subsp. martinii

321

Table 3. Comparison of means between subspec
at 3rd leaf below capitulum. All variables differed s
comparison of means.

ies. All measurements in mm. All vegetativ<
significantly (P < 0.01) between subspecies

t* characters measured
in the Kruskal- Wallis

Subspecies

Character

N

ochrocentrum

N

martinii

Corolla tube length

65

18.5 (1 1-25)

35

16.4 (8-22)

Corolla throat length

65

9.2 (6-12)

35

11.6 (7-17)

Corolla lobe length

65

10.2 (6-12)

35

11.1 (7-15)

Style tip length

64

5.6 (3-8)

33

4.8 (2-7)

Anther appendage tip length

65

1.2 (0.8-2)

34

1.1 (0.7-2)

Anther appendage body length

65

1 .6 (0.8-2)

34

2.2 (1-3)

Middle phyllary spine length

65

6.3 (2-12)

35

4.8 (2-12)

Leaf decurrence

49

11.2 (1-30)

23

7.7 (3-36)

Leaf length

48

47.9 (20-133)

24

68.4 (15-134)

Leaf spine length

45

8.0 (2-13)

24

6.5 (0.8-12)

22-45 mm broad. Phyllaries with white glandular
ridge when Iresh, obscure in dried specimens; mar¬
gins with scattered arachnoid hairs; tips sometimes
red; phyllary spines reflexed at maturity, spines
stramineous, 3-12 mm long. Corolla carmine red;
lobe and throat of similar length, 7—15 mm; upper
tube diameter nearly equal to the throat diameter.
Anther appendage free tip opaque, red or pink, ros¬
trate to aristate, 0.8— 1.8 mm long. Style tip red or
pink, about (3— )5( — 7) mm long; ring usually mi¬
nutely purberulent or trichomes not apparent; style
scape longitudinally corrugate below ring. Cypselas
compressed or somewhat trigonal, evenly brown or
with yellow and reddish streaks, 6.5 mm long. Pap¬
pus white, 25—43 mm long; bristles sparsely plu¬
mose toward tip, densely plumose toward base, api¬
ces slightly expanded.

The distribution ol this subspecies largely coin¬
cides with the Apachian floristic areas as defined
by McLaughlin (1986). The Apachian floristic area
includes much of southeastern Arizona and south¬
western New Mexico and has close affinities with
the flora of the Sierra Madre Occidentale in Mexico.
The division between these subspecies closely cor¬
responds to the division between the Gila River
Drainage and that of the Rio Grande and Colorado
River drainages.

The type material for C. ochrocentrum (Gray,
1849) was originally collected from hillsides around
Santa Fe, New Mexico. Cirsium ochrocentrum
subsp. ochrocentrum, the “Santa Fe” phenotype,
from the New Mexico Rio Grande river valley and
points eastward, is a relatively homogeneous mor¬
phological group. The more western subspecies
martinii differs from the typical subspecies in dis¬
tribution, corolla color, and continuous morpholog¬
ical characters. These taxa separate in phenetic

analysis along a discriminant function axis primar¬
ily composed of anther appendage tip and body
length, phyllary spine length, and style tip length.
Morphological intermediates between these two
forms are observed in a narrow zone along the Con¬
tinental Divide. In these areas, populations may
have intermediate flower colors to the red or purple
phenotypes or mixed populations of both colors. Al¬
though these intermediate populations were vari¬
able in many characteristics, there is no decrease
in pollen stainability or other indications of loss of
fertility in these areas (Barlow, 1992). The subspe¬
cies rank is appropriate to separate these taxa for
the following reasons: (1) there is intergradation be¬
tween the Gila phenotype and more typical Santa
Fe thistles in the zone of contact between them;
and (2) these taxa have overlapping distributions of
characters, which only difler in population means.
It is also of note that in populations of what super¬
ficially appears to be the typical subspecies in Jeff
Davis County, Texas, specimens have extremely
short style tips, and may be another segregate of
Cirsium ochrocentrum. Although Hsi (1960) dis¬
cussed this red-corolla form of C. ochrocentrum, my
analysis does not support recognition of a taxon
based on cypsela as a diagnostic characteristic (Ta¬
ble 3).

The subspecies martinii may be distinguished in
the field from subspecies ochrocentrum by its
bright-red florets, short style tips and spines, and
less prominent decurrence of leaf bases. It differs
from Cirsium arizonicum, another species with red
corollas, by its globose involucres and its large ca-
pitula with more than 120 florets per head as com¬
pared to fewer than 60 florets in the cylindrical
capitula of the Arizona thistle.

This subspecific epithet commemorates William

322

Novon

C. Martin, who inspired a generation of students
with his love of plants and his dedication to botan¬
ical education.

Paratypes. U.S.A. Arizona: Cochise County, 6 mi. E
of Portal, Cazier 1 14971 (API/); Huaclnica Mts., Ijemmon
2790 (US); Miller Canyon ltd. in Huaehuca Mts.. McGill
& Irlito lUOlii (ASU). Graham County, Ft. Grant. Shreve
5213 (ARIZ). Mojave County, 3 mi. NW of Ml. Dellen-
haugh on Shivvvits Plateau. Holland 745 (UNI.V); T34N,
R12W, S32 along roadside, Coombs & Handy 2597
(ARIZ). Navajo County, Heber, Lehto 1904 (ASU); Fort
Apache, Whitehead A' Holies 1020 (ARIZ). Pima County,
just N of Ophir Gulch, 4 mi. N & 2 mi. W of Sonoita,
Tranuinto t-10 (ARIZ). Santa Cruz County, along Rt. 82,
1.8 mi. S of junction of Rt. 83, McGill & Pinkava 0730
(ASU); Ted knipe Property, ca. 3 mi. W of Sonoita on Rt.
82, Heeres & Irlito 07501 (ASU). Yavapai County, Mingus
Mt., Mingus Water Shed, station If., Demaree 40001
(ASC); I mi. from Perkinsville. Grue 72A024 (ASC);
1.5 mi. S of Dewey on AZ 69, Keil K 1 1440 (ASU, WTS).
New Mexico: Catron County, cliff, Casletter 0362 (UNM);
road cut across rock outcrop, 3 mi. W of Reserve, Weber
& Salamun 12001 (COIX), UT); Lower Plaza. Frisco, Woo-
ton 1900 (NMC). Grant County, Gila NF. T21S, R16W, 8
mi. S of Silver City, Gentry & Jensen 2274 (BRY, COLO.
ARIZ); 15 mi. S of Red Rock. Maquire, Ranchards &
Moeller 11424 (ARIZ). Hidalgo County, 5 mi. S of Haeh-
ita. Casletter 10370 (UNM); I mi. N of Rodeo, Cazier 431
(ASU); Arroyos W of Hermanos. Clark 10740 (UNM).
Luna County, Cook. A. T. Hyatt 1907 (NMC).

Acknowledgments. 1 thank Timothy K. Lowrey,
Anne Evans, Laura Hyatt, and Christopher Frazier
for helpful criticism of this manuscript. Special
thanks go to Leslie Marcus lor help with probability
ellipses. Jane Mygatt, Audrie Clifford, Julie Saiki,
Karen Lightfoot, Peggy Norton, Tina Carlson, Alley
Helmer, Mike Clifford, Abe Plotsky, Clifford
Qualls, and Bruce Milne all helped with the orig¬
inal study of this problem, which was in part fund¬
ed by the Graduate Student Association of the Uni¬
versity of New Mexico. ARIZ, ASC, ASU, BRY,
COLO, NMC, UNLV, UNM, US, UT, and WTS gra¬

ciously provided specimens associated with this

study.

Literature Cited

Barlow, P. 1992. Biosystematic Study of the Genus Cir-
sium [Asteraceae): Cirsium ochrocentrum and its Red-
flowered Phenotype from the Region of the Mogollon
Rim. Master’s Thesis, University of New Mexico. Al¬
buquerque.

Davis, P. H. 1978. The moving staircase: A discussion on
taxonomic affinity. Notes Roy. Rot. Gard. Edinburgh 36:
325-340.

Gray, A. 1849. Plantae Fendlerianae Novi-Mexicanae: An
account of a collection of plants made chiefly in the
vicinity of Santa Ee, New Mexico, by Augustus Fendler;
with descriptions of the new species, critical remarks,
and characters of other undescribed or little known
plants from surrounding regions. Mem. Amer. Acad.
Arts n. ser. 4: I 10.

Hsi, Y. 1960. Taxonomy, Distribution and Relationships
of the Species of Cirsium Belonging to the Series Un-
dulata. Ph.D. Dissertation. University of Minnesota,
Saint Paul.

Lamp, W. (). & M. k. McCarty. 1981. Biology and ecology
of the Platte thistle ( Cirsium eanescens). Weed Sci. 29:
686—692.

Martin, A. C.. II. S. Zim & A. k. Nelson. 1951. American
Wildlife and Plants, a Guide to Wildlife Food Habits;
the Use of Trees, Shrubs, Weeds, and Herbs by Birds
and Mammals of the United States. Dover Publications,
New York.

McLaughlin, S. P. 1986. Florisitic analysis of the south¬
western United Slates. Great Basin Naturalist 46: 4(>-
65.

Ownbey. G. B. & Y. Ilsi. 1963. Chromosome numbers in
some North American species of the genus Cirsium.
Rhodora 65: 339-354.

- , P. II. Raven & I). W. kyhos. 1975. Chromosome

numbers in some North American species of the genus
Cirsium. 111. Western United States, Mexico, and Gua¬
temala. Brittonia 27: 297—304.

SAS Institute, Inc. 1989. SAS/STAT User’s Guide, Version
6, Ed. 4, Vol. 2. SAS Institute, Cary, North Carolina.

Tilden, J. W. & A. C. Smith. 1986. A Field Guide to
Western Butterflies. Houghton Mifflin, Boston.

A New Species of Monochilus (Lamiaceae) from Brazil

Philip D. Cantino

Department of Environmental and Plant Biology, Ohio University,
Athens, Ohio 45701-2979, U.S.A.

ABSTRACT. A new species, Monochilus obovutus P.
D. Cantino, is described from Goias, Brazil. It dif¬
fers from the only previously known species of
Monochilus in leaf shape, calyx size, and corolla
shape.

The genus Monochilus Fischer & C. A. Meyer
has traditionally been placed in the Verbenaceae
but was transferred to Lamiaceae by Cantino et al.
(1992), along with many other genera formerly clas¬
sified as Verbenaceae. Monochilus is unusual in
having alternate to subverticillate leaves and a ra¬
cemose inflorescence, whereas most Lamiaceae
have opposite leaves and a cymose inflorescence
(often a thyrse). Racemes and alternate leaves also
occur frequently but not universally in Amasonia
L. f., probably the closest relative of Monochilus.
The two genera are distinguished by corolla shape
(unilabiate in Monochilus vs. subactinomorphic to
weakly zygomorphic in Amasonia) and the absence
in Monochilus of the brightly colored calyx and flo¬
ral bracts that characterize Amasonia.

Until now, Monochilus was thought to be mono-
typic, comprising only M. gloxiniifolius Fischer &
C. A. Meyer, which is endemic to the states of Rio
de Janeiro and Espfrito Santo in coastal southeast¬
ern Brazil. The new species described here, Mon¬
ochilus obovatus, is known only from two flowering
specimens collected in gallery forest in Goias cer-
rado, about 600 miles northwest of the nearest
known locality of M. gloxiniifolius. The first spec¬
imen (US) was encountered in the process of pre¬
paring a description of the genus tor Kubitzki’s
(1990—) Families and Genera of Vascular Plants. A
duplicate (MBM) was subsequently borrowed from
the collector, G. Hatschbach, but attempts to locate
other specimens in Brazilian herbaria were unsuc¬
cessful.

Monochilus obovatus has the diagnostic features
of the genus (i.e., alternate leaves, racemose inflo¬
rescence, unilabiate corolla) but differs from M.
gloxiniifolius in four characters. The leaves of M.
gloxiniifolius are ovate to elliptical, with a petiole
1—3 cm long; those of M. obovatus are narrowly
obovate to oblanceolate and either sessile (US spec¬
imen) or tapering gradually to a petiole up to 1.5

cm long (most leaves on the MBM specimen). The
flowering calyx of M. gloxiniifolius is 5—6 mm long,
vs. 8.5—11 mm long in M. obovatus. The corolla
limb of M. gloxiniifolius is shorter than the tube,
whereas the corolla limb in M. obovatus is much
longer than the tube. Although few specimens of
either species exist, the four distinctions noted
here, together with the 600-mile separation of their
ranges, strongly suggest that they represent distinct
species.

Calyx shape is variable in Monochilus obovatus.
In the MBM specimen, the calyx is subactino¬
morphic as in M. gloxiniifolius, whereas the sinus
between the lower two calyx lobes in the US spec¬
imen is much deeper than the other four sinuses.
If additional collections are made in the future, ex¬
amination of this character in a larger sample of
both species would be worthwhile.

Monochilus obovatus R D. Cantino, sp. nov.
TYPE: Brazil. Goias: Sao Joao da Alian^a, ro-
dovia para Vaozinho, mata de galeria (zona de
cerrado), 9 Feb. 1994, G. & M. Hatschbach
60244 & J. M. Silva (holotype, US #3290138;
isotype, MBM #165645).

A Monochilo gloxiniifolio foliis anguste obovatis, calyce
8.5—1 1 mm longo, et limbo eorollae tubo longiore differt.

Perennial herb, pubescent with ± straight, mul¬
ticellular, nonglandular hairs, more densely so up¬
ward. Leaves alternate to subverticillate, clustered
below inflorescence, sessile or tapering to a petiole;
petiole (if present) up to 1.5 cm long; blade 7-11
cm long, 2.5-4 cm wide, narrowly obovate to ob¬
lanceolate, cuneate at base, broadly acute to obtuse
at apex, denticulate; upper surface evenly pubes¬
cent with ± straight hairs; lower surface densely
pubescent on principal veins, sparsely so else¬
where. Inflorescence a terminal raceme, hirsute;
bracts narrowly oblanceolate, 5—8 mm long; brac-
teoles paired in upper half of pedicel, linear. Calyx
(in flower) 8.5—11 mm long, exteriorly hirsute, sub¬
actinomorphic to zygomorphic, sinus between lower
2 lobes sometimes much deeper than other sinuses;
tube campanulate; lobes 5, ovate to broadly lan¬
ceolate, acute to attenuate. Corolla white, unilabi-

Novon 9: 323-324. 1999.

324

Novon

ate (upper lip only); tube cylindrical, deeply cleft
on lower side, 5 mm long; limb 5-lobed, 8—9 mm
long, 4 outer lobes oblong to ovate, median lobe
obovate and larger than others. Stamens 4, slightly
didynamous, attached near top of corolla tube,
slightly exserted from tube; filaments pubescent at
base; anther thecae parallel, not confluent at de¬
hiscence. Ovary unlobed; style apex with two very
short, equal to subequal lobes. Fruit unknown.

Kky to thk Speciks ok Monochhus

la. leaves ovate to elliptical, petiole 1—3 cm long;

flowering calyx 5—6 mm long; corolla limb short¬
er than tube; coastal southeastern Brazil .

. M. gloxiniifolius

lb. Leaves narrowly obovate to oblanceolate, sessile

or with petiole up to 1.5 cm long; flowering calyx
8.5—1 1 mm long; corolla limb much longer than
lube; Goias . M. obovatus

Acknowledgments. I am grateful to Gert Hatsch-

bach, the collector of the type specimens, for send¬
ing me material for examination, and to Silvia Ziller
for helping me to make contact with Dr. Hatsch-
bach. This research was carried out at t lie Harvard
University Herbaria, with travel support from the
Mercer Fund of the Arnold Arboretum and Ohio
University Research Committee grant #9957. 1
thank Michael Donoghue and Emily Wood for fa¬
cilitating my research at Harvard, and the U.S. Na¬
tional Herbarium and the Museu Botanico Munic¬
ipal in Curitiba, Brazil, for lending herbarium
material.

I .iterature Cited

Cantino, P. I).. R. M. Harley & S. J. Wagstaff. 1992. Gen¬
era of Labiatae: Status and classification. Pp. 51 1-522
in R. M. Harley <!4 T Reynolds (editors). Advances in
Labiate Science. Royal Botanic Gardens, kew.
kubitzki, k. (Editor). 1990-. The Families and G< mera of
Vascular Plants. Springer- Verlag, Berlin.

A New Subgenus and Fifteen New Species of Echeandia
(Anthericaceae) from Mexico and the United States

Robert William Cruden

Department of Biological Sciences, University of Iowa, Iowa City, Iowa 52242, U.S.A.

ABSTRACT. Two subgenera in Echeandia (Anther¬
icaceae), i.e., subg. Echeandia and subg. Mscavea,
are recognized based on time of flower opening,
tepal shape, flower color, and habitat differences.
Fifteen species are proposed. One, E. texensis, is
based on four collections made near Brownsville,
Texas, and is surely rare, possibly extirpated. The
recognition of five species clarifies the distribution¬
al range of E. reflexa , a common species of eastern
Mexico. Seven species are endemic to Guerrero and
adjacent areas of Miehoacan, Mexico, Morelos, and/
or Oaxaca. A key to the species of subgenus Msca¬
vea and white-flowered species of subgenus
Echeandia is provided.

Echeandia, as described by Ortega (1800: 135,
tab. 18), included New World species with connate
anthers and scaled filaments. These traits effective¬
ly separated Echeandia and Old World Anthericum,
which had free anthers and smooth or papillate fil¬
aments. With few exceptions. New World species
with free anthers were included in Anthericum for
the next 180 years. However, during that time, spe¬
cies with smooth filaments were included in
Echeandia (see Weatherby, 1910), and species with
scaled filaments were described in Anthericum
(Baker, 1876; Greenman, 1898). Thus, for the past
century or more. New World Anthericum and
Echeandia were separated simply on the nature of
their anthers, and the unique traits they shared
were ignored. Because most New World species de¬
scribed in Anthericum and Echeandia share traits
that are unique in Anthericaceae, including a corm
and scaled filaments, they are now included in
Echeandia (see Cruden, 1987, 1994), which, with
78 described species, is the second largest genus
in Anthericaceae (see Chase et al., 1996; Dahlgren
et al., 1985: 186). Excluded from Echeandia are
approximately 20 white-flowered South American
species with smooth filaments and free anthers that
probably constitute an undescribed genus.

W ithin Echeandia there are two distinct lineages,
and each includes species with free anthers and
others with connate anthers. The two lineages share
at least two synapomorphies. First, all Echeandia

have a corm that produces new storage roots and
inflorescences yearly. Other Anthericaceae are rhi-
zomatous, and only the terminal segment produces
leaves and inflorescences. There is no evidence
that a segment produces inflorescences in succes¬
sive years. Second, most genera in Anthericaceae
s. str. have smooth and/or papillate filaments, e.g.,
Chlorophytum s.l. (Kativu & Nordal, 1993) and
Hagenbachia (Cruden, 1987), whereas most
Echeandia (63/78) have scaled filaments and those
with smooth filaments occur in derived groups.

In addition, anthers adapted for pollination by
vibratory bees occur in both lineages and are
unique in Anthericaceae s. str. In the white-flow¬
ered lineage the anthers in 21 of 25 species are
connate and form a cone, and in the other 4 species
the individual anthers are adapted for buzz polli¬
nation (see below). In the yellow-flowered lineage,
19 of 53 species have anther cones anil in 4 ad¬
ditional species the individual anthers are adapted
for buzz pollination. The filaments of all these spe¬
cies are inserted dorsally within a basipetally open¬
ing pocket, which effectively holds the anther on
the same axis as the filament and probably provides
the structural stability needed to support a bee
while it vibrates an anther or anther cone. It is
possible that connate anthers constitute a parallel
apomorphy. In other Echeandia and most other An¬
thericaceae, the anthers are free, usually versatile,
dorsifixed, and dehisce laterally. In most Chloro¬
phytum, however, the filaments are inserted more
or less basally in a deep pit, which restricts the
movement of the anthers.

Finally, relative to Echeandia' s putative ancestor,
both lineages combine a mixture of primitive and
derived traits. Based on other Anthericaceae, that
ancestor had white flowers, which probably opened
relatively early in the morning, broadly elliptical
tepals, smooth filaments, and versatile anthers that
dehisced laterally. Thus, yellow flowers are apo-
morphic in one lineage, as are narrowly elliptical
inner tepals and late opening flowers in the other.

Thus, the available data are consistent with a
common ancestor diverging into two lineages,
which are formally recognized below as subgenera.

Novon 9: 325-338. 1999.

326

Novon

The two lineages share several synapomorphies and
are separated hy a number of apomorphies. Further,
the two lineages tend to occur in different climatic
zones (see below). I include the two lineages in
Echeandia as subgenera because the differences
that separate them do not appear to be as great as
those separating other genera in Anthericaceae.
However, a more detailed examination of the rela¬
tionships among genera in Anthericaceae using
other techniques might lead to a different conclu¬
sion.

Subgenus Mscavea is proposed to include spe¬
cies in the white-flowered lineage, whose flowers
open late in the morning or in the afternoon, and
have narrow inner tepals. This subgenus honors
Marion S. Cave, eminent embryologist and student
of the Liliaceae s. lat., who stimulated my interest
in Mexican Liliaceae in general and Echeandia in
particular.

Echeandia subg. Mscavea Cruden, subg. nov.

I f PE: Echeandia mcvaughii Cruden, Contr.

Univ. Michigan Herb. 16: 129. 1987. [Illus¬
trated in McVaugh, 1989: 187.]

Ah Echeandia subg. Echeandia tepalis interioribus an-
gustc ellipticis albis vel cremis raro flavis, floribus aper-
ientibus ante vel post meridiem differ!.

The two subgenera differ in time of flower open¬
ing, tepal shape, width, and color, capsule shape,
and altitudinal distribution. The flowers of species
in Echeandia subg. Echeandia open shortly before
or after sunup. All but 1 of S3 species have broadly
elliptical inner tepals. In the 40 species for which
I have reasonable data, the minimum width of the
inner tepal in 27 species was 4.5 mm, and the me¬
dian width exceeded 4.5 mm in 36 species. Most
of the species (44/53) are yellow-flowered, and six
include white-Howered populations. Nine species
are white-flowered. In contrast, in subgenus Msca¬
vea, the flowers open late in the morning or the
afternoon, 24 of 25 species have narrowly elliptical
inner tepals, and, in 22 of 23 species for which I
have reasonable data, the maximum width of the
inner tepal was 4.5 mm. Most of the species have
white flowers (22/25), two have cream-colored flow¬
ers, one has orange flowers, and one, possibly two,
species include orange- or yellow-flowered popu¬
lations.

Capsule shape differs significantly between the
two subgenera (^ = 13.91, p»0.001; analysis with
a contingency table and Chi-square test). For those
species in subgenus Echeandia for which there are
data, 26 have oblong capsules (length at least twice
the width) and 16 have subglobose to broadly ob-

lable I. Distribution of species in subgenus Echean¬
dia and subgenus Mscavea as a function of elevation. A
species was assigned to a given category if most of its
elevational range was included in that category. The range
of two species in Echeandia extended from slightly below
800 m to 2400 m. They were counted as being half in one
category and half in the other category.

Below

8(H)-

15(H)-

above

8(H) in

1500 m

24(H) in

24(H) m

Subg. Echeandia

(>

8

30

9

Subg. Mscavea

8

12

5

0

long capsules, whereas just 2 species in subgenus
Mscavea have oblong capsules, and 17 have subglo¬
bose to broadly oblong capsules.

Relative to species in subgenus Mscavea, those
in subgenus Echeandia tend to occur at higher el¬
evations (Table 1; = 19.84, pCO.001). Most spe¬

cies in subgenus Echeandia (>70%) occur com¬
pletely or primarily above 1500 m, and nine occur
above 2400 m, whereas 80% of the species in sub¬
genus Mscavea occur completely or primarily below
1500 m, and only four reach 2400 m. It is not
obvious that a disproportionate number of species
in subgenus Mscavea relative to subgenus Echean¬
dia occur below 800 m (^ = 4.93, p<0.05). In
essence, most species in subgenus Mscavea occur
in drier habitats with subtropical to warm-temper¬
ate climates, whereas most species in subgenus
Echeandia occur in relatively mesic habitats w itb
warm- to cold-temperate climates.

The occurrence of connate anthers is not equiv¬
alent in the two subgenera. Only 19 of 53 species
in subgenus Echeandia have connate anthers, com¬
pared to 21 of 25 species in subgenus Mscavea.
Because most species in subgenus Mscavea have
connate anthers, one might anticipate that species
with connate anthers occur at lower elevations.
However, there is no relationship between occur¬
rence of connate anthers and elevation in either
subgenus Echeandia (^ = 1.18, p>0.5) or sub¬
genus Mscavea (^ = 0.09, p>0.975).

Among those species with connate anthers, tin*
two subgenera differ with respect to the insertion
of the filaments, which is either dorsal and hidden
within a pocket or between the anther sacs and
essentially basal. In most species of subgenus
Mscavea (12/20 examined), the distance from the
lower edge of the pocket to the base of the anther
is shorter than in 18 of 19 species in subgenus
Echeandia examined (0. 1-0.6 mm vs. 0. 6-2.0 mm;

= 13.137, pCO.001). There is no correlation
between anther length and distance from the lower

Volume 9, Number 3
1999

Cruden

Echeandia

327

edge of the pocket to the base of the anther in
either subgenus Echeandia (r = 0.289, n = 19,
p»0.05) or subgenus Mscavea (r = 0.237, n — 20,
p»0.05). Finally, in six species of subgenus Msca¬
vea but just one of subgenus Echeandia, insertion
of the filaments is basal or essentially so.

Among species with free anthers, the anthers of
all four species in subgenus Mscavea dehisce api-
cally compared to 4 of 33 species in subgenus
Echeandia. In three species of subgenus Mscavea,
the line of dehiscence is fused Vi—2/) the length of
the anther, and dehiscence is through a relatively
large teardrop-shaped opening. In one species of
subgenus Mscavea and the four species of subgenus
Echeandia, the line of dehiscence is completely
open, and dehiscence is through a much smaller
apical opening.

Kf.Y TO MEXICAN SlItGENlS MSCUEA AND
WWTK-FI.OVt K HKD SlJBGENUS ECHEANDIA

In any couplet, a clause preceded by “if" takes prece¬
dence over other clauses. Leaf shape follows that proposed
by the Systematies Association Committee (1962). The
names of species described in this paper are not followed
by the author’s name. The subgenus of each species is
given: (E) = subg. Echeandia, (M) = subg. Mscavea. See

also keys in McVaugh (1989) and Cruden (1994).

1. Anthers free . 2

1. Anthers connate, forming a cone . 3

2(1). Filaments smooth . 4

2. Filaments scaled . 7

3(1). Filaments smooth . 20

3. Filaments scaled . 26

Authors Free, Filaments Smooth

4(2). Sea[>es with scattered enations or glabrous . . 5

4. Scapes hirsute or scabrous, at least in lower

half, frequently throughout . 6

3(4). Anthers not versatile, line of dehis¬
cence partially fused, bases flared;
tepals narrowly elliptic, 6—8 mm
long (below 1200 m, Jalisco) ....

. E. pihuamensis Cruden (M)

5. Anthers versatile, dehiscence later¬

al, bases straight; tepals broadly el¬
liptic, 8 — 1 3 mm long (above 1600
m, Durango and San Luis Potosf to

Mexico) .

E. durangensis (Greenman) Cruden (E)
6(4). Basal leaves narrowly linear, veins papillate
below, margins ciliate, if tips strongly re¬
curved (1800—1900 m, Oaxaca. Puebla) . .

. E. i larva Cruden (F)

6. Basal leaves narrowly ovate to narrowly el¬

liptic, flat, veins smooth below, margins en¬
tire or nearly so. tips straight (mostly above
2000 m, San Luis Potosf to Veracruz and

Puebla W to Michoacan) .

. E. nana (Baker) Cruden (E)

Anthers Free, Filaments Sealed
7(2). Eastern Mexico (Tamaulipas, San Luis Po¬
tosf, Veracruz, Tabasco, Campeche) below
5(X) m . 8

7. Elsewhere in Mexico and/or at higher ele¬
vations . 9

8(7). Flowers white (Tamaulipas) . . .

. E. tamaulipensis (E)

8. Flowers yellow-orange (Campeche)

. E. campechiana Cruden (M)

9(7). Basal leaves 2—6 mm wide, 5—15 cm long,
margin long ciliate, veins on lower surface
papillate and/or bearing long hairs; storage
roots 4-10 cm from the corm (Oaxaca) . . .

. E. confertifiora Cruden (E)

9. If most basal leaves wider or longer, margin
entire to ciliate, and/or the veins smooth be¬
low; if storage roots enlarged 1—2 cm from

the corm . 10

10(9). Anthers dehisce apically; anther
walls not reflexed but may be open
along line of dehiscence and the
wall flared at apex; most more than

3 mm long . II

10. Anthers dehisce laterally; anther
walls reflexed, if anthers twisted or
obviously versatile; most less than 3
mm long (sometimes longer in E.
gentryi (M) and E. durangensis (E)) 14

11(10). Cauline leaves 7—11, becoming white when
dry; storage roots close to the corm; basal
leaves more than 20 mm wide (Jalisco) . . .

. E. robusta Cruden (M)

II. Cauline leaves 1—6, if more, then storage
roots 4—10 cm from the corm; if basal leaves

less than 20 mm wide . 12

12(11). Tepals 7—8 mm long; line of dehis¬
cence partially fused (Oaxaca. Pueb¬
la) . E. mirandae Cruden (M)

12. Tepals 11—21 mm long; line of de¬
hiscence completely open . 13

13(12). Ovaries 3-4 mm long (Puebla) .

. E. michoacensis (Poellnitz) Cruden (E)

13. Ovaries 2—3 mm long (Chiapas) .

. E. matudae Cruden (E)

14(10). Scapes pubescent, sometimes weak¬

ly so or only near the base . 15

14. Scapes glabrous . 16

15(14). Storage roots enlarged 1—4 cm from corm;
margins of basal leaves ciliate to long ciliate;
scape scabrescent, at least toward the base
(Chihuahua ami Nayarit SE to Michoacan)

. E. scabrella (Bentham) Cruden (E)

15. Storage roots enlarged 3-8 cm from corm;
margins of basal leaves denticulate; shoot
with minute scattered hairs in lower half (Oa¬
xaca) . E. sp. (E)

16(14). Filaments 9—15 mm long; most basal
leaves 1—6 mm wide; if style scaled
(Guerrero, Mexico, Morelos) . . .

.... E. tenuis (Weatherby) Cruden (E)
16. Filaments less than 8 mm long; or if
most basal leaves more than 6 mm

wide . 17

17(16). Basal leaves 1-7 mm wide, narrowly linear
to linear, margins denticulate; most storage
roots enlarged 3—8 cm from corm; capsules
subglobose to broadly oblong (length < twice

the width) (Oaxaca, Puebla) .

. E. graminea Martins & Galeotti (E)

328

Novon

17. II most l>asal leaves 7 or more mm wide;
storage roots enlarged 1—3 cm from corm; if
capsules oblong (length at least twice the

width) . 18

18(17). Basal leaves 1-6 mm wide, narrowly
linear; filaments inserted in a pock¬
et, not versatile; ovaries 1.8— 2.5 mm
long (Jalisco. Nayarit. Sinaloa. Du¬
rango) . E. gentryi Cruden (E)

18. II most basal leaves 7 mm or more

wide; ovaries 2.5-4 mm long .... 19

19(18). Leaves broadly linear to narrowly ovate; an¬
thers versatile (above 16(H) m, mts. of Du¬
rango and San Luis Potosf S to Mexico and
Hidalgo) . . E. durangerisis (Greenman) Cniden (E)

19. Leaves linear, frequently falcate; anthers not

versatile due to reflexed anther walls. (Gu¬
anajuato) . E. sanmiguelensis (E)

Anthers ('.minute. Filaments Smooth

20(3). Basal leaves abruptly expanded near the
middle into a broad blade, appearing petio-
late (Guerrero) . E. pseudopet iolata ( \1 )

20. Width of basal leaves not changing abruptly

. . 21

21(20). Scapes scabrous beyond first cauline

leal and/or il basal margins of leaves

eiliate to long eiliate . 22

2 1 . Scapes glabrous or rarely scabres-

eent near the base . 25

22(21). Scapes usually more than I m high, cauline
leaves 6—15; hasal leaves 15—40 mm wide
(rarely narrower), margin entire to short fil¬
iate; tepals 9—13 mm long (Guerrero, More¬

los) . E. mexiae (M)

22. Scapes rarely 1 m high; if cauline leaves 0-
4. il most basal leaves less than 10 mm wide,
their margins eiliate to long eiliate, and/or

tepals less than 9 mm long . 23

23(22). Basal leaves 2-1 I mm wide, most
less than 10 mm wide, flat, margin
entire to denticulate; tepals 8.5—

10.5 mm long; anthers ( 1 .8 — )2— 3
times the length of the filaments;
scapes 15—40 cm high (below 500

m. Jalisco, Sinaloa) .

. E. sinaloensis Cruden (M)

23. If most of the basal leaves more than
10 mm wide, falcate, their tips
strongly recurved, or the margin cil¬
ia! e to long eiliate; if tepals less than
8 mm long; anthers 1 — 1 ,6(— 2) times
the length ol the filaments; if scape

more than 50 cm high . 24

24(23). Basal leaves (>—12 mm wide, very narrowly
linear to verv narrowly elliptic, frequently fal¬
cate and/or with strongly recurved tips, fre¬
quently undulate, margins densely eiliate to
long eiliate (rarely denticulate or short eiliate)
(Jalisco, Golima SE to Oaxaca, Ghiapas, and
Guerrero to Veracruz) . . E. parviflora Baker (M)

24. Basal leaves (8—) 13— 28 111111 wide, flat, nar¬
rowly ovate to narrowly elliptical, margins
entire to minutely denticulate (below 8(H) m,

Guerrero, Mexico) . E. atoyacana (M)

25(21). Capsules globose, 2.5—4 mm long;
basal leaves 4—8, (> — 1 5 mm wide;

cauline leaves 2-6 (Jalisco, Nayarit)

. E. parvicapsulata Cruden (M)

25. Capsules broadly oblong, 6-10 mm
long; basal leaves 7—13, if most
more than 15 mm wide; cauline
leaves (4— )5 — 10 (Sonora, Chihua¬
hua, to Guerrero) .

. . . E. ramosissima (Fresl) Cruden (M)
Anthers Connate, Filaments Sealed

26(3). Eastern Mexico (Tamaulipas, San Luis Po¬
tosf, Veracruz, Campeche, Yucatan) below

1600 m; il flowers cream-colored . 27

26. Elsewhere in Mexico or above 1800 m ... 29

27(26). Flowers cream-colored (Campeche,
Quitana Boo. Yucatan. Oaxaca) . . .

. E. luteola Cruden (M)

27. Flowers white . 28

28(27). Anthers 7-9 mm long, twice the length of the
filaments; leaf margins eiliate to long eiliate

(Tamaulipas, San Luis Potosf) .

. E. macrophylla Rose (M)

28. Anthers 4—7 mm long, barely longer than the
filaments; leaf margins denticulate to eiliate

(Veracruz) . E. albiflora (Schlechtendal &

Chamisso) Martins & Galeotti (M)
29(26). Apexes of anther cones 1 .2-3.5 mm
wide; anther cones 2.5— 4.0 mm wide
and 4-5(-7) mm long; basal leaves
2—7 mm wide, veins on lower sur¬
face noticeably papillate (Oaxaca)

. E. conzattii Cruden (E)

29. II apexes ol anther cones less than
I mm wide, or anther cones less
than 2 mm wide; if basal leaves

smooth below and/or most more than
10 mm wide; if storage roots en¬
larged 1—3 cm from corm . 30

30(29). Most I. asal leaves less than 5 mm wide ... 31
30. Most hasal leaves more than 5 mm wide . . 34

31(30). Storage roots enlarged 3 cm or more
from corm; anthers 1-1.5 times the

length of the filaments . 32

3 1 . Storage roots enlarged 0.5— 2.0 cm
from corm; if anthers 1.8-2. 2 times

the length ol the anthers . 33

32(31). Tepals broadly elliptic, 5-7 rnm wide, 1 1.5—

15 mm long; filaments (4 — )5 — 7 mm long
(Guerrero, Mexico, Morelos, Oaxaca) . . .

. E. luxe ana (E)

32. lepals narrowly elliptic, 1.5—3 mm wide,
7.5—12.5 mm long; filaments 2.5—4 mm long

. E. parviflora Baker (M)

33(31). Anthers 1 — 1.6 times the length of
the filaments; capsules 7-11.5 mm
long (above 800 m, Nayarit) . . .

. E. flexuosa Greenman (E)

33. Anthers 1.8-2. 2 times the length of
the filaments; capsules 5—7 mm
long, (below 500 m, Nayarit, Sina¬

loa) . E. nayaritensis (M)

34(30). Scapes pubescent, at least toward the base.

and sometimes sparsely so . 35

34. Scapes glabrous . 39

35(34). Leaf margins long eiliate, hairs usu¬
ally 0.2— 0.5 mm long; scapes usu¬
ally hirsute, hairs long and flat; has¬
al leaves (10— )15— 40 mm wide

Volume 9, Number 3
1999

Cruden

Echeandia

329

(Guerrero. Mexico. E Michoacan)

. E. hirticaulis (M)

35. Leaf margins entire to ciliate. rarely
long ciliate; scapes mostly scabres-
cent; if most basal leaves 2—8 mm

wide . 36

36(35). Leaf margins ciliate to long ciliate, some¬
times densely so; basal leaves 2-8 mm wide,
frequently falcate; anthers 3—7 mm long,
most 1.2— 1.8 times the length of the fila-
ments; tepals 7.5—12 mm long; basal leaves

1—8 mm wide . E. parviflora Baker (M)

36. Leaf margins entire to short ciliate, rarely
ciliate; if most basal leaves more than 10 mm
wide; if anthers 7.5-12 mm long and/or 1.8-
2.2 times the length of the filaments; if tepals

more than 12 mm long . 37

37(36). Storage roots enlarged 4—8 cm from
conn; if lower leaf surface minutely
papillate between the veins; cap¬
sules 7—9 mm long (Guerrero) . . .

. E. hintonii (M)

37. Storage roots enlarged 1—2 cm from
conn; lower leaf surface smooth be¬
tween the veins; il capsules shorter

or longer . 38

38(37). Anther cones 6—1 1 mm long, usually widest
in the middle; if anthers 1.8— 2.2 times the
length of the filaments; capsules 4—7 mm
long; basal leaves 5—10. frequently falcate,
if most more than 15 mm wide and/or less
than 25 cm long (Jalisco. Nayarit) ....

. E. mcvaughii Cruden (M)

38. Anther cones 5.5-7 mm long, 1-1.4 times
the length of the filaments, usually widest to¬
ward the base; capsules 9-12 mm long; basal
leaves 3—5, narrowly elliptic, straight, 6—12
mm w ide and more than 30 cm long (Oaxaca)

. E. oaxacana Cruden (M)

39(34). Most storage roots enlarged 3—8 cm
from corm; anther cones widest to¬
ward the base, tapered gradually to
the apex, usually 1-1.7 times the

length of the filaments . 40

39. Most storage roots enlarged 1—3 cm
from corm; if anther cones widest in
middle and/or 1.8—3 times the

length of the filaments . 41

40(39). Tepals 9—13 mm long; anther cones 4—8 mm
long; cauline leaves 1—5; il underside of bas¬
al leaves minutely papillate between veins

(Guerrero) . E. hintonii (M)

40. Tepals 13—20 mm long; anther cones 8-12
mm long; cauline leaves 4—10; basal leaves

smooth below (Guerrero. Morelos) .

. E. elegans (M)

41(39). Tepals 9—11 mm long; anther cones
5.5—8 mm long, usually widest near
the base and gradually tapered to
the apex; basal leaves narrowly lin¬
ear. 30—45 cm long, 5—14 mm wide,

not falcate (Michoacan) .

. E. coalcomanensis Cruden (M)

41. Tepals 10-19 mm long; anther cones
T» — 1 2 mm long, if cones widest in
middle; if most basal leaves falcate.

less than 20 cm long, and/or more

than 15 mm wide . 42

42(41). Basal leaves 4—8 mm wide, narrowly linear,
straight; ovaries 1-2 mm long (below' 600 m,

Nayarit, Sinaloa) . E. nayaritensis (M)

42. Basal leaves 10—30 mm wide, most narrowly
obovate or narrowly elliptic, frequently fal¬
cate; ovaries 2—3 mm long (above 1200 m,
Jalisco, Nayarit) . . . . E. mcvaughii Cruden (M)

Fifteen species are proposed below: eight in sub¬
genus Echeandia and seven in subgenus Mscavea.

The species in subgenus Echeandia with connate
anthers comprise the taxonomically most difficult
group in the genus and include E. reflexa (Cavan-
illes) Rose, the type species of the genus. Because
this species is quite variable and has a broad geo¬
graphical range, and given the lack of recent sys¬
tematic analysis, it is not surprising that specimens
with connate anthers from eastern and central Mex¬
ico, as well as elsewhere, were included, indiscrim¬
inately, in E. reflexa. I propose E. pseudoreflexa and
E. texensis to include such material from Chiapas
(Cruden, 1994) and Texas (see Johnston, 1990), re¬
spectively; E. palmeri is proposed to include spec¬
imens from Durango and Chihuahua. Only E. pseu¬
doreflexa appears to be closely related to E. reflexa
(see below). The description of two additional spe¬
cies should clarify the distribution of Echeandia in
eastern and central Mexico. Echeandia falcata is
proposed to include plants from Queretaro and Gu¬
anajuato with relatively short, broad anther cones,
and E. taxcana is proposed to included narrow¬
leaved plants with strongly tapered anther cones
from Oaxaca, Guerrero, Morelos, and Mexico.

With the recognition of these species, E. reflexa
is a more coherent, albeit variable, entity found pri¬
marily between 200 and 3000 m on the eastern
slopes of the Sierra Madre Oriental and adjacent
parts of the Sierra Volcanica Transversal from cen¬
tral Nuevo Leon and southern Tamaulipas to south¬
ern Puebla, adjacent Oaxaca, and Veracruz. Popu¬
lations from low elevations are usually robust, with
highly branched inflorescences, and long, broad
straight leaves. At mid elevations, the plants are
less robust with fewer branches and narrower,
shorter leaves; and at high elevations, the leaves
are usually narrow and frequently falcate. Speci¬
mens from higher elevations in Puebla and adjacent
Veracruz and Oaxaca have falcate, undulate basal
leaves and resemble small specimens of E. pani-
culata Rose. Flowers of E. reflexa have a long (6—
9 mm), slender, strongly tapered anther cone, which
usually has a small, minutely lobed apex.

330

Novon

1. Echeandia pseudoreflexa Cruden, sp. nov.
TYPE: Mexico. Chiapas: road to Venustiano
Carranza, ea. 3 km SW ol Santa Rosalia, hill¬
side with Convolvulus, Calliandra, Lantana,
Dahlia, 1200 in, 3 Sep. 1971, Cruden 1936
(holotype, UC; isotypes, MEXU, MO).

Radicibus penariis a eormo 1—3 cm. foliis basal i bus
anguste linearil)us vel angusle ellipticis 7—12 mm latis
marginibus dentieulatis ad breviciliatis, scapis glabris, te-
palis flavis, lilamentis squamosis, antheris eonnatis 5—7.5
nim luugis. <onis antherarum ad apicem valde contractis.

Storage roots enlarged 1—3 cm from corm. Basal
leaves 5—8, narrowly linear to narrowly elliptical,
34—64 cm long, 7—12 mm wide, denticulate to short
ciliate. Scapes glabrous, 0.6— 1.5 in high, 2-6-
branched, major axis with 12—20 flower-bearing
nodes. Cauline leaves (0— )2-6, lowest 17 — 24( — 32)
cm long, usually exceeding the next node. Flowers
yellow, nutant. Tepals 11—16 mm long, inner broad¬
ly elliptic 5 mm wide, outer elliptic 2.5 mm wide
(one (lower examined). Filaments scaled, narrowly
elavate, 5—7 mm long, insertion dorsal in a pocket.

Anthers connate, 5—7.5 mm long, equal to or slight¬
ly longer than the filaments; the cones strongly ta¬
pered, apex barely emarginate to shallowly and
broadly lobed 0.5-0.8(-1.2) mm wide. Ovaries 2.5-
4 mm long. Capsules oblong to narrowly oblong, 9—
14.5 mm long, 4.5— 5.5 mm wide. In flower August-
September, December. Chromosome number n —
16 ( Cruden 1936).

With the exception of material collected near La
Cienega de Leon, this species is known only from
central Chiapas between 1100 and 1800 m (Fig. 1).
It occurs in forests dominated by some combination
of [line, oak, and Liquidambar.

Material of E. pseudoreflexa was included in E.
reflexa in Flora Mesoamericana (Cruden, 1994) be¬
cause of its robust shoots and leaves and strongly
tapered anther cone. It is distinguished from most
populations of E. reflexa by storage roots that are
enlarged 1—3 cm from the corm, narrower basal
leaves, and the bases of the previous year’s leaves
forming a fibrous collar. The storage roots of E. re-

Volume 9, Number 3
1999

Cruden

Echeandia

331

flexa are usually enlarged 3-6 cm from the corm,
and the previous year’s leaf bases remain more or
less intact. The long, lower cauline leaf distinguish¬
es this species from many specimens of E. formosa
(Weatherby) Cruden.

Paratypes. MEXICO. Chiapas: Ruta 190, 58 km N
El Jocote, 7 mi. from La Trinitaria, 1350 m. 19 Sep. 1988.
Stevens & Martinez 25 767 (MO); Mpio. Cintalapa, near
Oaxaca-Chiapas border, near La Cienega de Leon, 1080—
1230 m, 1 Dee. 1980, Breedlove & Almeda 48171 (CAS);
Mpio. Pueblo Nuevo Solistahuaoan, 3 km NW Pueblo
Nuevo Solistahuaean. 5800 ft.. 28 Aug. 1970, Zwill 298
(l)S); Mpio. Tenejapa, barrio Tuk, paraje Matsab. 7500 ft..
30 Sep. 1965, Breedlove 12552 (DS): Mpio. Teopisea,
southern limits of Teopisea, 5900 ft.. 19 Aug. 1966,
Breedlove 15049 (I)S. E).

2. Echeandia texensis Cruden, sp. nov. TYPE:

U.S.A. Texas: Green Island, 23—29 June 1922,

Tharp 1129 (holotype, TEX; isotype, US).

Radieibus penariis a cormo 2—4 cm, seapis glabris. fo¬
lds basalibus 4—9 linearibus ad elliptieis 10—20 mm latis
marginibus denticulatis, tepalis flavis, filamentis squa-
mosis. antheris connatis 4—5.5 mm longis, oonis anthera-
rum 2—3 mm latis plerumque filamentis brevioribus.

Storage roots enlarged 2-4 cm from corm. Basal
leaves 4-9, 33—60 cm long, 10-20 mm wide, linear
to narrowly elliptic, sometimes weakly falcate, mi¬
nutely denticulate to denticulate. Scapes glabrous,
56—105 cm high, unbranched. Cauline leaves 4—5,
(4.3— ) 10-20 cm long, lowest frequently exceeding
the next node. Flowers yellow, nutant. Tepals 12-
19 mm long, inner broadly elliptic, 6-8.5 mm wide,
outer elliptic, 2.5— 3.5 mm wide. Filaments scaled,
narrowly clavate, 5—9 mm long, insertion dorsal in
a pocket. Anthers connate, 4-5.5 mm long, the
cones 2-3 mm wide, usually shorter than the fila¬
ments, weakly to strongly tapered, apex 1—1.5 mm
wide, deeply and broadly lobed. Ovaries 2—3.5 mm
long. Capsules oblong, 9-13 mm long, 4—6 mm
wide. In Hower June, October— November.

This species is based on lour collections made
near Brownsville, Texas (Fig. 1), prior to 1952.
These suggest a rather limited distributional range,
and the lack ol recent collections suggests this spe¬
cies is rare and probably endangered, if not already
extirpated. The limited data suggest it is a species
of clay dunes and llanos.

Specimens of E. texensis were included in E. re-
flexa because they have long broad basal leaves,
overlapping cauline leaves, and oblong capsules.
The short, broad, and deeply lobed anther cones
are similar to those of E. conzattii Cruden, E. pal-
meri Cruden, E. flexuosa Greenman, and E.falcata
Cruden, all of which have narrower basal leaves,

fewer cauline leaves that only occasionally overlap,
and anthers that are usually equal to, or longer
than, the filaments. Echeandia texensis is easily dis¬
tinguished from E. reflexa by the short, usually
broad anther cone with a broadly lobed apex.

Paratypes. U.S.A. Texas: Cameron County, vicinity
Brownsville, 10 m, 25 Oct. 1945. Runyon 4082 (TEX);
Brownsville, Runyon 4036 (TEX); Loma del Cayo. 29 Nov.
1951, Johnston & Tharp 38b-l (TEX).

3. Echeandia palineri Cruden, sp. nov. TYPE:
Mexico. Durango: Santiago Papasquiaro, Apr.
& Aug. 1896, Palmer 410 (holotype, US; iso¬
types, F, GH, MO, NY).

Radieibus penariis a cormo 2-A cm, foliis basalibus
anguste linearibus 1—7.5 mm latis marginibus denticulatis
ad ciliatis, seapis glabris, tepalis flavis. filamentis clavatis
squamosis, conis antherarum 3.5—7 mm longis ad apicem
infirme contractis.

Storage roots 1.5—3 cm long, enlarged 2—1 cm
from corm. Basal leaves 5—9, narrowly linear, (15—)
24—40 cm long, (1.5— )3— 7.5 mm wide, margins
denticulate to ciliate, veins forming the midrib
sometimes papillate. Scapes glabrous, (35— )58-74
cm high, bearing 0-2 branches. Cauline leaves 1—
2, lowest 3.5— 7.6(-13) cm long, rarely reaching
next node. Flowers yellow, probably nutant. Tepals
13—19 mm long, inner broadly elliptic, 6—8 mm
wide, outer elliptic, 3-4 mm wide. Filaments
scaled, narrowly clavate, 4—6 mm long, insertion
dorsal in a pocket. Anthers connate, (3.5— )4— 7 mm
long, the cones widest at base, tapered to apex,

1.5— 3.5 mm wide, apex (0.7— )1. 1—2.1 mm wide,
broadly sometimes coarsely toothed. Ovaries 2.5—
5.5 mm long. Capsules oblong, (9— )11— 15 mm long,

4.5— 6 mm wide. In flower August— September.

The broad, weakly tapered anther cone, narrow
leaves, and storage roots enlarged 2 or more cm
from the corm are similar to those of E. conzattii,
E. flexuosa, and E. falcata. The veins on the un¬
derside of the basal leaves of E. conzattii are usu¬
ally papillate, and it has white flowers. In general,
E. flexuosa and E. falcata have more cauline anil
basal leaves than E. palmeri; the basal leaves of
many E. flexuosa plants are narrower and those of
E. falcata are falcate.

The specific epithet honors Edward Palmer, not¬
ed collector of Mexican plants.

Paratypes. MEXICO. Chihuahua: hills near Chihua¬
hua, 6 Sep. 1886. Pringle 1109 (MEXU, MO. NY). So¬
nora: Y4cora, 3 Sep. 1970, Pennington 103 (TEX); old
road to Maycoba, 0.5 mi. E Arroyo Yeeora. 1550 m, 7
Sep. 1995, Fishbein 2478 (MO).

332

No von

4. Echeandia f'alcata Cruden, sp. nov. TYPE:
Mexico. Guanajuato: Ruta 49, near Km 32, ea.
7 km N San Miguel de Allende, grassy hillside
with Opuntia , Ipomoea, 1880 m, 4 Aug. 1968,
Cruden 1459 (holotype, UC; isotypes, ENCB,
F, GH. K. MEXU, MO, NY, US).

Radicibus penariis a cormo 3—5 cm, loliis basalilms 2—
1 2(— 1 4) mm latis linearibus ad anguste ovatis plerumque
falcatis marginibus ciliatis, scapis glabris vel ad basim
infinite pubescentibus, tepalis flavis, filamentis clavatis
squamosis, antheris connatis 4—6.5 mm latis, conis an-
therarum ad apicem infirme contractis.

Storage roots enlarged 3-5 cm from the corm,
usually long-tapered. Basal leaves (S — )7— 1 1( — 1 4),
linear to narrowly ovate, weakly to strongly falcate,
occasionally straight. 10-52 cm long, 2— 12(— 1 4)
mm wide, margin short ciliate; veins on lower sur¬
face smooth, rarely papillate. Scapes glabrous or
weakly scabrous, 15—85 cm high, bearing 0-1 (-2)
branches. Cauline leaves 2-4, the lowest 1— 7.5(—
10) cm long. Flowers yellow, nutant. Tepals 13-17
mm long, inner broadly elliptic, 6—8.5 mm wide,
outer elliptic, 2.5—5 mm wide. Filaments scaled,
clavate to narrowly clavate, 4—6 mm long, insertion
dorsal in a pocket. Anthers connate, 4—6.5 mm
long, the cones weakly tapered, ( 1 . 5—) 1 .8 — 2 . 5 mm
wide, (0.8— )l-2(— 2.8) mm wide at apex. Ovaries
2.5^1 mm long. Capsules broadly oblong to oblong,
(5 — ) 10 — 13( — 16) mm long, 5.5— 6.5 mm wide. In
flower (late June— (August— September. Chromosome
number n = 16 (Cruden 145 6, 1459 , 1803).

This tetraploid relative of E. flexuosa is endemic
to Guanajuato and Queretaro (Fig. 1). Two speci¬
mens from Guanajuato have long, non-falcate
leaves, but the rest have shorter, usually narrow
and strongly falcate leaves. The large flowers with
weakly tapered, usually broad anther cones are
similar to those of E. flexuosa from western Mi-
choac&n and eastern Jalisco, and distinguish this
species from both E. reflexa and E. mexicana Cru¬
den of adjacent Hidalgo. The falcate leaves, clavate
filaments, and somewhat larger capsules distin¬
guish E. falcata from E. flexuosa.

Paralypes. MEXICO. Guanajuato: Guanajuato,
1880, Duges s.n. (GH); Ruta 51, between San Miguel de
Allende and Comonfort, 2100 m, 5 July 1971, Genelle &
Fleming 898 (DUKE, MO); 18 km S San Miguel de Al¬
lende. 17(M) m, 4 Aug. 1968, Cruden 1456 (ENCR, E, E.
MEXU, MICH. NY, UC, WIS); ca. 8 km N San Miguel de
Allende, 1880 m. 23 Aug. 1970. Cruden 1806 (GH. UC)-
Queretaro: E San Miguel de Allende. 15 km E Cuana-
juato-tjueretaro state line. 2125 m, 23 Aug. 1970, Cruden
1803 (UC); Hacienda Ciervo. between San Juan del Rio
and Cadereyta. 20 Aug. 1905, Rose. Painter & Rose 9647
(US); Del Ciervo, 20 Aug. 1905. Altamirano 1597 (US);
road toward Tequisquiapan, 21 Aug. 1903, Rose , Painter

& Rose 9703 (US); Tequisquiapan. 18 Sep. 1957, Parax
2500 (ENCB); near San Juan del Rio. 28 Aug. llX)5. Rose.
Painter & Rose 9554 (NY. US); 15 km from Higuerillas
toward Bernal, 20°54'N, 99°54'W, 2100 m. 25 June 1972.
Wendt. Chian g & Johnston 8100 (TEX).

5. Echeandia taxacana Cruden, sp. nov. TYPE:
Mexico. Mexico: Ruta 55, 6.5 km NW Guer¬
rero State Line, 1485 m, 1 Aug. 1970, Cruden
1721 (holotype, UC; isotypes, F, GH, K,
MEXU, MO, US).

Radicibus penariis 1 cm longis a cormo 1.5— 5.5 cm.
foliis basalibus peranguste linearibus 1—4 mm latis mar¬
ginibus denticulatis ad ciliatis. scapis glabris vel ad basim
scabridis. tepalis flavis vel albis, filamentis clavatis squa¬
mosis, antheris connatis 5-7.5 mm longis, conis anthera-
rum ad apicem valde contractis.

Storage roots enlarged 1.5— 5.5 cm from corm, ca.
1 cm long. Basal leaves 5-14, very narrowly linear,
(10—) 15-55 cm long, l-3(-4) mm wide, margin
denticulate to ciliate. Scape glabrous or scabres-
cent toward the base, 35—85 cm high, bearing 0-1
branches. Cauline leaves 1— 3(-5). lowest 1 .6-5.5
cm long. Flowers yellow or white, nutant. Tepals
11.5—15 mm long, inner broadly elliptic, 5—7 mm
wide, outer narrowly elliptic, 2-3 mm wide. Fila¬
ments scaled, clavate, (4— )5— 7 mm long, insertion
dorsal in a pocket. Anthers connate, (5— )6— 7.5 mm
long, the cones usually longer than the filaments,
strongly tapered, (1.7— )2-2.5 mm wide, apex shal¬
lowly lobed, 0.5— 0.9 mm wide. Ovaries 2^4 mm
long. Capsules broadly oblong, 7.5-9 mm long, 4-
4.5 mm wide. In flower July— October. Chromosome
number n — 8 ( Cruden 1721. 1953).

The yellow-flowered populations are found in a
variety of habitats, including thorn scrub, oak
woodlands, and pine-oak woods in Oaxaca, Guer¬
rero. Morelos, and Mexico between 1400 and 2400
m (Fig. 1). The white-flowered plants are essentially
like the yellow-flowered ones, but may be from
more mesic habitats. The sampled populations are
southeast of the distributional range of the yellow-
flowered populations between 2000 and 2600 m.

The numerous narrow leaves, slender inflores¬
cences, and broadly oblong capsules are reminis¬
cent of E. flexuosa, from which it is distinguished
by the strongly tapered anther cone and small stor¬
age roots that usually develop some distance from
the corm. The narrow leaves distinguish this spe¬
cies from E. reflexa, and it usually has more basal
leaves than E. smithii Cruden.

The specific epithet honors the city of Taxco de
Alearon, which sits at the northwestern end of dis¬
tributional range of this species.

Volume 9, Number 3
1999

Cruden

Echeandia

333

Paratypes. MEXICO. Guerrero: Taxco Hoad, 18 July
1938, Kenoyer A519 (F); 19 mi. from Taxco de Alarcon,
3800 ft., 3 July 1941, Chute M-I08 (MICH). Morelos:
1905, Rose, Painter & Rose 8609 (US). Oaxaca: Ruta
190, Km 56, ca. 16 km USE Tamazulapan, 2350 m, 15
Aug. 1970. Cruden 1773 (UC), 7 Sep. 1971. Cruden 1953
(ENCB. K. MEXU, NY. TEX, UC, WIS); Ruta 190. Km
406-407, 26.5 km SE Huajuapan de Leon, 2270 m. 30
July 1968, Cruden 1444 (UC).

White-flowered specimens examined. M EX ICO. Oa¬
xaca: Mpio. San Juan Mixtepec, 16 km S San Juan Mix-
tepec. Monte de Tesoro. 17°14'N. 97°51'W, 26(H) m. 6
Oct. 1988, Santiago 880 (MEXU); 6.7 km W of Miahu-
atlan on road to Piedra Larga (Coatlanes), 25 Oct. 1982,
Torres, Marline: & Villasenor 1716 (MEXU); Tlaxiaco, San
Pedro Molinos. Km 64 on carretera Tlaxiaco— Putla,
1 7°14'34"N. 97°43'10"W. 2000 m. 3 Aug. 1994, Panero
& Calzada 4439 (TEX).

6. Echeandia attenuala Cruden, sp. nov. TYPE:
Mexico. Sinaloa: Ruta 40, Km 204, ca. 5 km
W El Palmito, seepage area with Begonia , Bi-
dens, Dahlia , Castilleja, Cosmos, 2000 m, 18
Sep. 1971, Cruden 1997 (holotype, UC; iso¬
types, ENCB, F, GH, K, MEXU, MO).

Radicibus penariis a cormo 1—2 cm, 3-4 cm longis,
foliis basalibus anguste linearibus attenuatis 1 — 5( — 7) mm
latis, scapis glabris vel ad basim scabris 30—90 cm altis,
foliis caulibus 2—6 longiattenuatis, tepalis (lavis ellipticis
valde reflexis, (damentis squamosis unumquidque in mar-
supio insertis 4—6.5 mm longis, antheris libris non ver-
satilis dehiscens lateralibus.

Storage roots enlarged 1—2 cm from the corm, 3—
4 cm long. Basal leaves narrowly linear, attenuate,
23-53 cm long, 1 — 5( — 7) mm wide, margin dentic¬
ulate to serrulate. Scapes glabrous or scabrescent
toward the base, 30—90 cm high, bearing 0-4
branches. Cauline leaves (2-)3— 5(— 6), long atten¬
uate, lower to 17.5 cm long. Flowers yellow, cer-
nuous-declinate. Tepals strongly reflexed, (6.5— )8—
13 mm long, inner elliptic, 3.5—5 mm wide, outer
narrowly elliptic, 2—2.5 mm wide. Filaments
scaled, linear, 4—6.5 mm long, insertion dorsal in
a pocket. Anthers free, 2.5—3.25 mm long, not ver¬
satile, dehiscence lateral. Ovaries 1 .5—3 mm long.
Capsule (only one seen) 6.5 mm long, 5 mm wide.
In flower late August— September. Chromosome
number n — 8 ( Cruden 1827).

Echeandia attenuata is known only from seepage
areas between 2000 and 2200 m along the Duran-
go-Sinaloa border (Fig. 1). It is distinguished by the
long-attenuate leaves, both basal and cauline,
smallish, cemuous-declinate flowers with relatively
narrow and strongly reflexed tepals, and filaments
that are inserted in a pocket, which forces the an¬
thers to extend along the same axis as the filaments.
The narrow, strongly reflexed tepals and nonversa-

tile anthers distinguish it from E. flavescens (Schul¬
tes & Schultes f.) Cruden, a common species of
desert grasslands and conifer forests of interior cen¬
tral and northern Mexico.

Paratypes. MEXICO. Durango: 2 mi. NE El Parafso,
Sinaloa on road between Villa Union and El Salto. 7000
ft.. 28 Sep. 1953, Ownbey & Ownbey 1931 (MICH). Sin¬
aloa: Ruta 40. K 204, 5 km W El Palmito, 2000 m, 27
Aug. 1970. Cruden 1827 ( NY, US).

7, Echeandia sanmiguelensis Cruden, sp. nov.

TYPE: Mexico. Guanajuato: Ruta 49, Km 30—
31, ca. 9 km N of San Miguel de Allende,
grazed field along road with Acacia, Opuntia,
Agave, 1950 m, 4 Aug. 1968, Cruden 1460
(holotype, UC; isotypes, K, MEXU, MO).

Radicibus penariis a cormo 2—3 cm, scapis glabris 45—
75 cm altis, foliis basalibus anguste ovatis falcatis 7-13
mm latis, tepalis albis, filamentis squamosis 5—6 mm lon¬
gis, antheris libris non versatilis.

Storage roots enlarged 2—3 cm from corm, 3—5
cm long. Basal leaves linear, attenuate, falcate, 5—

8, 17-33 cm long, 7-13 mm wide, margin ciliate.
Scapes glabrous, 45—75 cm high, bearing 0-2(^4)
branches. Cauline leaves 0-2, when present, lower
aristate, 18-67 mm long. Flowers white, cernuous-
declinate. Tepals 11—15 mm long, inner broadly el¬
liptic, 5.5—7 mm wide, outer elliptic, 3-4.5 mm
wide. Filaments scaled, linear, 5—6 mm long, in¬
sertion dorsal in a pit. Anthers free, 1.5— 2.5 mm
long, not versatile due to the reflexed anther walls,
dehiscence lateral. Ovaries 2.5— 3.5 mm long.
Styles bent upward below the stigma. Capsules ob¬
long, (8.5— )10 — 16 mm long, 3.5— 6.5 mm wide. In
flower late July-August.

I bis species is known from only three collections
possibly representing just one or two populations
near San Miguel de Allende (Fig. 1). The two Cru¬
den collections may be from a single large popu¬
lation.

The falcate leaves, white, declinate flowers, and
storage roots enlarged close to the corm suggest a
relationship with E. scabrella (Bentham) Cruden,
which has smaller flowers, subglobose capsules,
and scabrescent scapes. The few aristate cauline
leaves, if present, broader lanceolate leaves, and
longer capsules differentiate it from E. tamaulipen-
sis Cruden. The white flowers distinguish this spe¬
cies from E. flavescens, which occurs in similar
habitats. In addition, lew individuals of the latter
species have storage roots enlarged close to the
corm in combination with 5—8 basal leaves that
tend to be falcate.

334

Novon

Paratypes. MEXICO. Guanajuato: Ruta 49. Km 31-
32. ea. 8 km IN of San Miguel de Allende, 1950 m. 23
Aug. 1970. Cruder i 1805 (ENCB. E. GH. MICH. NY. UC.
US); San Miguel de Allende, 6500 ft.. 26 July 1950, Spiv¬
ey 177 (UC).

8. Eeheantlia tainaulipensis Cruden, sp. nov.
TYPE: Mexico. Tamaulipas: Papolote de la
Mirandena, 3 mi. SSW of headquarters, Loreto
Ranch, 24°20'N, 98°W, 16 Sep. 1960, John¬
ston & Crutchfield 5550 (holotype, MEXU; iso¬
types, LL, MICH, TEX, UC).

Radicibus penariis a corino (1— )3-6 cm clavatis, foliis
basalibus linearibus 4—12 mm latis. seapis glabris vel sca-
bris, tepalis albis, filamentis pauei squamosis 4.5— 6.5 mm
longis, antheris libris non versatilis dehiscens lateralilnis.

Storage roots enlarged ( 1 — )3 — 6 cm from corm,
long tapered, 3—9 cm long. Basal leaves 3-7, linear,
20-65 cm long, denticulate to short ciliate, 4-12
mm wide. Scape glabrous or scabrescent, 40—60 cm
high, bearing 0— 2 (—3) branches. Cauline leaves 1—
4, reduced to brac ts. Flowers white, probably cer-
nuous-declinate. Tepals 10-15 mm long, inner
broadly elliptic, 4.5—5 mm wide, outer narrowly el¬
liptic to elliptic, 1.5-3. 5 mm wide. Filaments weak¬
ly scaled, linear, 4.5— 6.5 mm long, insertion usu¬
ally dorsal in a deep pit or pocket, if in a shallow
pocket, the reflexed anther walls hold the anther on
same radius as the filaments. Anthers free, 1.7-2. 5
mm long, not versatile, dehiscence lateral. Ovaries
2.5— 3.5 mm long. Capsules broadly oblong to ob¬
long, 8—11 mm long, 5—5.5 mm wide. In flower Sep¬
tember— October.

This white-flowered species is known only from
the coastal plain of central and southern Tamauli¬
pas (fig. 1). It might be confused with either E.
flavescens or E. chandleri (Greenman & Thompson)
M. C. Johnston, which both have yellow flowers.
The former occurs at much higher elevations and
has versatile anthers, while the latter has somewhat
larger flowers and strongly sc aled filaments. The
Mexican collections of E. chandleri are from higher
elevations (900—1000 m) or from northern Tamau¬
lipas.

Paratypes. MEXICO. Tamaulipas: S Lomas del Real.
7 mi. N of Altamira. 27 Oct. 1959, Johnston <& Graham
4540 (MICH, TEX); Sierra de Tamaulipas, ea. 40 km
N N W Aldama, ea. 23°14'N, 98°10'W, above Juan Tomas,
E Las Yueas, 13 Oct. 1957. Dressier 2384 (GH, MO); 15
mi. from Tampico on the Mante highway, 27 Sep. 1959.
Graham A- Johnston 4087 (MEXU, MICH. TEX).

The following species are included in subgenus
Mscavea.

9. Eeheamiia atoyaeana Cruden, sp. nov. TYPE:

Mexico. Guerrero: rocky outcrop with Matalea,
Commelina , Cnidosculus, Ruta 95, Km 103-
104, ea. 31 Km S bridge over Rio Papagayo,
300 m, 4 Aug. 1970, Cruden 1734 (holotype,
UC; isotypes, F, GH, K, MEXU, MO).

Radicibus penariis a cormo 3-8 cm, foliis basalibus 6-
12 ovalis 8—28 mm latis marginibus integris ad denticu-
latis, seapis scabris ad 1 m altis, foliis caulibus 2-4, te¬
palis albis, filamentis linearibus glabris flexus intro super
ovarium, antheris connatis.

Storage roots 3—8 cm from corm, 1—2 cm long.
Basal leaves 6—12, narrowly ovate to narrowly el¬
liptic, (12-)23-52 cm long, (8-)13-28 mm wide,
margins entire to denticulate or serrulate. Scapes
scabrous, 33—97 cm high, bearing (0 — ) 1 — 4(— 9)
branches. Cauline leaves 2^4, reduced to bracts.
Flowers white, nutant, opening in early afternoon.
Tepals narrowly elliptic, 6.5-9 mm long to 1 mm
wide. Filaments smooth, linear, bent or pinched in
above ovary, 2. 5-3.3 mm long, insertion basal or
nearly so. Anthers connate, 3—5 mm long, the cones
strongly tapered. Ovaries 1-2 mm long. Capsules
globose, 4—6.5 mm long. In flower July— September.
Chromosome number n = 16 ( Cruden 1732, 1733,
1734).

This tetraploid species is known only from Guer¬
rero and adjacent Mexico between 250 and 650 m
(Fig. 2). The combination of small flowers, short,
strongly tapered anther cones, small capsules,
broad basal leaves, and pubescent scapes is unique
in subgenus Mscavea.

The specific epithet recognizes the town of Atoy-
ac de Alvarez anil the Rfo Atoyac in whose valley
I first collected this species.

Paratypes. MEXICO. Guerrero: ea. 9.5 km N Atoy¬
ac, road to San Vincente anil I’uerlo de Gallo, 300 m, 3
Aug. 1970. Cruden 1732 (ENCB. K. UC): ea. 6 km N
Atoyac, 260 m, 3 Aug. 1970, Cruden 1733 (MEXU. MO,
NY, UC, US, Vi IS); Dist. Caleana, Atoyac, 25—300 m, 10
Aug. 1939. Hinton et al. 14526 (ARIZ, (ill. 1.1. (2), MO,
NY, US); 23 km N Ruta 2(H) (coast road) on road to
Ciudad Altamirano, 500 m, 26 Sep. 1983, Anderson
12816 (MICH). Mexico: Palmar, Hist. Temascaltepec,
[650 m|. 21 July 1934, Hinton et al. 6321 (GH, K. NY,
US).

10. Eeheantlia elegans Cruden, sp. nov. TYPE:
Mexico. Guerrero: oak woodland, with Cosmos,
Sessilanthera, Commelina, Calochortus, Agave,
ca. 15 km SW Xochipala, road to Puerto de
Gallo, 1950 m, 21 Sep. 1973, Cruden 2096
(holotype, UC; isotypes, ENCB, F, GH, K,
MEXU, MO, NY, US).

Radicibus penariis a cormo 2-6 cm, foliis basalibus 5—
8 anguste linearibus ad anguste obovatis 7-18 mm latis

Volume 9, Number 3
1999

Cruder)

Echeandia

335

t rated.

marginibus intergris ad breviciliatis, scapis glabris glau-
cis, foliis caulibus 4—10, tepalis albis 13—20 mm longis,
filamentis clavatis squamosis, antheris connatis 8—11.5
mm longis.

Storage roots 2—6 cm from conn, 1.5— 2.5 cm
long. Basal leaves 5—8, narrowly linear to narrowly
obovate, 20—70 cm long, 7-14(-18) mm wide, mar¬
gins entire to short ciliate. Scape glabrous, glau¬
cous, 0.7— 1.5 m high, bearing 0—9 branches. Cau-
line leaves 4—10, lowest 9—30 cm long, others
reduced upward. Flowers white, nutant, opening in
late morning. Tepals narrowly elliptic, 13-20 mm
long, inner 3.5—5 mm wide, outer 1.5—3 mm wide.
Filaments scaled, clavate, 3.5— 7.5 mm long, inser¬
tion dorsal in a pocket. Anthers connate, 8—11.5
mm long, 1.5—2 times the length ol the filaments,
the cones usually strongly tapered, rarely weakly
so. Ovaries 2.5—4 mm long. Capsules broadly ob¬
long, 8—9.5 mm long, 5—5.5 mm wide. In flower
September— October. Chromosome number n = 8
o Cruden 2096, 21 73).

This is a species of oak woodlands between 1900
and 2300 m in Guerrero and adjacent Morelos (Fig.

2). The available material includes two types. The
specimens from Temisco have 5—9 branches, anther
cones that are weakly tapered, and may have large
storage roots. The material from central Guerrero
has 0—3 branches, strongly tapered anther cones,
and quite slender storage roots. Vegetatively, this
species is difficult to distinguish from E. hintonii.
Relative to the latter species, E. elegans has longer
tepals and anthers, an anther cone that is usually
twice the length of the filaments, and smooth lower
leaf surfaces.

Paratypes. MEXICO. Guerrero: 14 rni. from Mex 95
on road to Chichihualco, 26 Oct. 1970. Graham 1226
(MICH). Morelos: Kuta 95 (libre) in Temisco. ca. 9 km
S Cuernavaca, 22 Sep. 1974, Cruden 2173 (GH, k.
MEXU, MICH. UC).

I 1 . Echeandia hintonii Cruden, sp. nov. TYPE:
Mexico. Guerrero: on steep, damp bank with
Blechnum, Adiantum, Oxalis, Cuphea, in pine-
oak woods, Ruta 95, near Agua de Obispo, ca.
39 km S Chilpancingo, 780 in, 2 Aug. 1970,
Cruden 1729 (holotype, UC; isotypes, ENCB,
F, GH, K, MEXU, MO, NY, US, WIS).

336

Novon

Radicibus penariis a comm 4—8 cm, foliis basalibus 3—
5(— 8) anguste linearibus ad anguste obovatis 5—20 mm
latis, scapis glabris vel scabris, foliis caulibus 1 — 4( — 5),
Icpalis all)is 10—13 mm longis, filamenlis clavatis s<jna-
mosis, anthcris connatis 4.5-8 mm longis. conis anthera-
rum ad apicem valde contract is.

Storage roots 4—8 cm from corm, 1—2 cm long.
Basal leaves 3— 5(— 8), narrowly linear to narrowly
obovate, in many minutely papillate between the
veins on lower surface, 13-56 cm long, 5-20 mm
wide, margin entire to denticulate. Scapes glabrous
or scabrescent, 37—111 cm high, bearing 0—4
branches. Cauline leaves 1— 4(— 5), the lower 8.5—
28 cm long. Flowers white, nutant, opening in late
morning. Tepals narrowly elliptic, 10-13 mm long,
inner 3.5— 4.5 mm wide, outer 1.5—2 mm wide. Fil¬
aments scaled, clavate, sometimes narrowly so,
3.5—5 mm long, insertion basal or dorsal in a pock¬
et. Anthers connate, 4.5-8 mm long, the cones
strongly tapered. 1—1.7 times the length of the fil¬
aments. Ovaries 2—4 mm long. Capsules broadly
oblong, 7-9 mm long, 4—5 min wide. In flower late
June— October. Chromosome number n = 8 (Cruden
1723 , 1729, 2102).

Among species in subgenus Mscavea with narrow
leaves and scaled filaments, this species is distin¬
guished by its slender, well-removed storage roots.
All collections are from oak and pine-oak woods in
central and north-central Guerrero between 750
and 2350 m (Fig. 2). Specimens with pubescent
scapes are a little larger and have lewer cauline
leaves and tiny papillae on the lower surfaces of
the basal and cauline leaves. There is no geograph¬
ic pattern to plants with glabrous and pubescent
scapes, and in the vicinity of Agua de Obispo the
leaves of specimens with both types of scapes wen-
papillate; thus, there is little reason to recognize
the variation formally.

This species honors George B. Hinton whose col¬
lections from Guerrero have added considerably to
our understanding of the region’s Echeandia.

Paratypes. MEXICO. Guerrero: Pasion. Dist. Mon¬
tes de Oca. 1000 in. 5 Oct. 1937. Hinton et al. 10703
(OH. 14.. NY, US); Ruta 95. ca. 17 km S Chilpaneingo.
ea. 1250 m, 2 Aug. 1970. Cruden 1723 (ENCB, (41. k.
UC); 61 mi. NK Acapulco. Aug. 1965. Irving 230 (TEX);
ca. 29 km SW of Xoohipala, on road to Puerto ilc Oallo.
2225 m. 22 Sep. 1973, Cruden 2102 (MEXU. UC); 65.3
km from the "desviaeion” of the Acapulco— Mexico high¬
way, 2350 m, 21 Oct. 1986. Romo 321 (MEXU); Agua de
Obispo, IKK) m. 10 Oct. 1986. Palomino. Romo & Kenton
303 (MEXU); Taxeo, 12 July 1932. Abbott 239 ((41). 23
July 1 936, Abbott 1 00 (GH); N Taxeo, 25 June 1935. Chirk
7233 (NY).

12, Echeandia hirticaulis Cruden, sp. nov.
TYPE: Mexico. Mexico: Ruta 130, ca. 2.5 km
NE Temascaltepec, oak woods with Tigridia,
Sphenostigma, Cuphea, Eryngium, 1750 m, 11
Sep. 1971, Cruden 1967 (holotype, UC; iso¬
types, ENCB. GH. k. MEXU. MO).

Radicibus penariis a cormo 2—4 cm. foliis basalibus
anguste ovatis (8—) 15— 42 mm latis marginibus longicilia-
lis. scapis plerumque hirsutis in dimidio inferiore, foliis
caulibus 3—4, tepalis albis, (ilamentis squamosis, anlheris
connatis.

Storage roots 2 — 1 cm from corm, 1—2 cm long.
Basal leaves 4—8, narrowly ovate, (17 — )25 — 55 cm
long, (8-)15-42 mm wide, margins long eiliate and
usually undulate. Scapes usually hirsute on the
lower half, hairs long and flat, 0.2— 0.5 mm long,
rarely with just a few hairs above the lowest cauline
leal, 0.6— 1.3 m high, (3— )5— 9-branched, which in
large individuals bear secondary branches. Cauline
leaves 3-4, lowest 12-54 cm long, others reduced
to bracts, long eiliate. Flowers white, nutant, open¬
ing in late morning. Tepals narrowly elliptic, 12-
17 mm long, 1-2 mm wide. Filaments scaled, cla¬
vate, (3— )3.5— 5.5 mm long, insertion dorsal in a
pocket. Anthers connate, (6— )7. 5— 10.5 mm long,
the cones strongly tapered, usually twice the length
of the filaments. Ovaries 1-3 mm long. Capsules
broadly oblong, 6—8 mm long, 4—5 mm wide. In
flower August— mid October. Chromosome number
n — 8 (Cruden 1967 , 1970).

This species is found between 1000 and 2000 m
in openings in oak and pine-oak woods in a small
region in western Mexico, eastern Michoaean, and
northern Guerrero (Fig. 2). The closest relative of
this species may be E. macrophylla Rose, which is
found between 1000 and 1500 m in San Luis Potosf
and Tamaulipas. Other robust species in tin- Sierra
Volcanica Transversal have straight, smooth fila¬
ments, numerous cauline leaves, and/or flowers that
open in the afternoon. The specific epithet de¬
scribes the hirsute scapes.

Paratypes. MEXICO. Guerrero: Manchdn. Dist.
Mina. 27 Sep. 1936. Hinton et al. 9390 ((41. k. l.L, NY.
US). Mexico: between Xitdeuaro and I ^as Anonas. Dist.
Zitacuaro. 1600 m. 23 Aug. 1938. Hinton et id. 13140
((41. k. I,U, NT US); Ixtapan, Dist. Temascaltepec. I(HK)
m. 5 Aug. 1932. Hinton 1203 (GH. k. MICH, NY. UC.
US); Carboneras. Dist. Temascaltepec, 2030 m, 27 Sep.
1932. Hinton 1352 (k); 4.5 mi. NE Temascaltepec, 1850
m. II Oct. 1966. Anderson A' Uiskowski 3934 (ENCB.
MICH); 5 km S Temascaltepec, KXX) m, 3 Sep. I <>65, Roe.
Roe, Mon A Rzedowski 1707 (Vi IS); road to Tingambato,
E San Nicholas. 16(H) in, 20 Sep. 1974, Cruden 21 09
(UC). Michoaean: 7 mi. E Hidalgo. 20 Aug. 1953, Man¬
ning & Manning 531039 ((ill); Ruta 15. near km 116,

Volume 9, Number 3
1999

Cruden

Echeandia

337

ca. 9.5 km S Tuxpan, 2000 m, 11 Sep. 1971, Cruden 1970
(ENCB. GH. K. UC).

13. Echeandia mexiae Cruden, sp. nov. TYPE:
Mexico. Guerrero: tropical deciduous with pal¬
metto palm. Salvia, Dahlia, Bursera, 4.5 km
SW Xochipala, 1280 m, 21 Sep. 1973, Cruden
2094 (holotype, UC, isotypes, F, GH, K,
MEXU, MO, US).

Radicibus penariis a cormo 3—9 cm, foliis basalibus 3—
5 anguste obovatis (10— )15— 40 mm lads marginibus brev-
iciliatis, scapis scabris 1—1.3 m aids, foliis caulibus 6—
15, tepalis albis, filamentis linearibus glabris, antheris
connatis.

Storage roots enlarged 3-9 cm from the corm,
1.5—3 cm long. Basal leaves 3—5, narrowly obovate,
37^15 cm long, (10— ) 15-40 mm wide, margin short
ciliate. Scape scabrous throughout, 1—1.3 m high,
(3— )5— 14-branched, some nodes with two branches
and some branches bearing a secondary branch.
Cauline leaves 6—15, reduced to bracts. Flowers
white, nutant, opening in late afternoon. Tepals nar¬
rowly elliptic, 9-13 mm long, inner 2-3 mm wide,
outer 1-2 mm wide. Filaments smooth, linear, 2-
3.5 mm long, insertion basal or nearly so. Anthers
connate, 4.5— 6.5 mm long, the cones strongly ta¬
pered. Ovaries 1.7-2. 5 mm long. Capsules broadly
oblong, 5-8 mm long, 4—5 mm wide. In flower Au-
gust-October. Chromosome number n — 8 ( Cruden
2086, 2087, 2088, 2094, 2172).

All collections of this species were made in
north-central Guerrero and adjacent Morelos (Fig.
2) between 400 ami 1500 m in relatively dry hab¬
itats, including deciduous thorn scrub. Several col¬
lections are from limestone-derived soils. The nu¬
merous cauline leaves and smooth filaments
suggest a relationship with the widespread E. ra-
mosissima (Presl) Cruden of western Mexico.
Scapes of the latter are usually glabrous, rarely sca-
brescent near the base, and the basal leaves are
usually narrower.

The specific epithet honors Ynes Mexia, who col¬
lected extensively in Mexico and South America in
the 1920s and 1930s.

Paratypes. MEXICO. Guerrero: Canon de la Mano
Negra, near Iguala, 11 Aug. 1905, Rose, Painter & Rose
9328 (US); Ruta 951), km 63-64. ca. 12 km S Morelos
state line, 1070 m, 20 Sep. 1973, Cruden 2086 (ENCR.
GH. k. TEX. UC, WIS); Ruta 51, ca. 22 km A Iguala,
915 m, 20 Sep. 1973, Cruden 2087 (ENCB, F, CH. k.
UC); ca. 35 kin W Iguala. 1500 m, 21 Sep. 1973, Cruden
2088 (UC); ca. 36 km W of Iguala, 1000 m, 30 Sep. 1982.
litis . Benz & Burd 28634 (IA. WIS); Ruta 55, km 118—

I 19, ca. 30 km N Taxco, 1150 m, 21 Sep. 1974. Cruden
2172 (MEXU. MO. NY. UC); Temisco, top of barranca
Limo, 380-400 m, 30 Oct. 1937. Mexia 8711 (AR1Z. F.

CH. k. MO. NY, U, UC, US). Morelos: vicinity of Cuer¬
navaca, 1905, Lemmon & Lemmon s.n. (UC).

14. Echeandia nayaritensis Cruden, sp. nov.
TYPE: Mexico. Nayarit: Mpio. Ruiz, 1—3 km
W El Venado, road from Ruiz to Jesus Marfa,
60 m, 9 Aug. 1980, Breedlove A Almeda
45284 (holotype, CAS).

Radicibus penariis a cormo 0.5—2 cm, foliis basalibus
anguste linearibus, scapis glabris, tepalis albis 12—19 mm
longis, filamentis fusiformis squamosis, antheris connatis
6—1 1 mm longis, conis antheraruni filamentis duplo lon-
gioribus.

Storage roots enlarged 0.5—2 cm from corm, 1—
4 cm long. Basal leaves 3-10, narrowly linear, 16—
38 cm long, (2— )4— 8 mm wide, entire, sometimes
falcate. Scapes glabrous, 41-58 cm high, bearing
0-3 branches. Cauline leaves 1-2, bract-like.
Flowers white, nutant. Tepals narrowly elliptic, 12—
19 mm long, inner 2—3 mm wide, outer 1-2 mm
wide. Filaments scaled, fusiform, 3.0— 4.5 mm long,
insertion basal or nearly so. Ant Iters connate, (6-
)7.5 — 11 mm long, the cones weakly tapered, widest
in the middle or parallel-sided in lower half, twice
the length of the filaments. Ovaries 1—2 mm long.
Capsules broadly oblong, 5—7 mm long, 4.5— 5.5
mm wide. In flower July— August.

This little-collected species is endemic to the
coastal lowlands of northern Nayarit and adjacent
Sinaloa, where it occurs in open areas and palm
savanna below 400 m. The long tepals, long an¬
thers, and storage roots enlarged and clustered
close to the corm suggest a relationship with E.
mevaughii Cruden, which occurs at higher eleva¬
tions in Nayarit, Jalisco, and western Michoaean.
The scape of the latter species is usually pubes¬
cent, and the basal leaves are usually more than
10 mm wide and falcate.

Paratypes. MEXICO. Nayarit: 26.8 mi. S of Sinaloa
state line, 200 ft.. 24 July 1975, Dunn. LeDoux & Wallace
21850 (EL). Sinaloa: Ua Jarretadera, ca. 149 km N Tepic,
19 Aug. 1988. Walker s.n. (k. MEXU. MO, UC).

15. Echeandia pseudopetiolata Cruden, sp.
nov. TYPE: Mexico. Guerrero: Ruta 95, Km
44—45, near El Rincon, ca. 44.5 km S Chil-
pancingo, 600 m, 2 Aug. 1970, Cruden 1731
(holotype, UC).

Radicibus penariis a cormo 1—3.5 cm, foliis basalibus
4—6, 43—56 cm longis dimidio inferiore anguste linearis
3—5 mm latis dimidio superiore magnopere expansis ovatis
ad ellipticis (2 I -)26— 35( — 45) mm latis, scapis glabris, fo¬
liis caulibus 2^4, tepalis albis. filamentis linearibus pau-
cisquamosis, antheris connatis.

338

Novon

Storage roots 1—3.5 cm from corm, 1—1.5 cm
long. Basal leaves 4—6, oblanceolate, 43—56 cm
long, lower half narrowly linear, 3—5 mm wide, ex¬
panded in the upper half into a broad, ovate to
elliptical blade, (21— )26— 35(— 45) mm wide, mar¬
gins short ciliate to c iliate. Scape glabrous, 88—106
cm high, up to 7 branches. Cauline leaves 2^4, the
lowest 12.5—36 cm long, resembling the basal
leaves, but without the “petiole.” Flowers white,
nutant, opening in the afternoon. Tepals narrowly
elliptic, 6—9 mm long, 1—1.5 mm wide. Filaments
weakly scaled, linear, sometimes wrinkled and/or
twisted, 3—4 mm long, insertion dorsal in a pocket.
Anthers connate, 4—5.5 mm long, the cones strong¬
ly tapered. Ovaries 1—2 mm long. Capsules not
seen. In llower late July— August. Chromosome
number n — 8 ( Cruden 1731).

This species is known from a few collections in
south-central Guerrero between 600 and 1100 m
along the highway from Acapulco to Iguala (Fig. 2).
It occurs in pine-oak and oak woods, as well as
disturbed roadsides. The specific epithet describes
the distinctive pseudopetiolate basal leaves.

Paratypes. MKX1CO. Guerrero: Agua de Obispo,
1100 m. 12 Aug. 1002, Kruse 796 (FNCB); 1 ( )( X ) m. 8
Aug. 1072. lioege 2441 (MKXU); Hwy. to Acapulco. Km
338, 3 km beyond Acahui/.otlu, 3000 It.. 20 Aug. 1048,
Moore & Wood 4675 (K, GH, MICH. UC. US).

Acknowledgments. I thank Diana Horton. Fred¬
erick H. Utech, and Wendy B. Zomlefer for helpful
and constructive comments on the manuscript. Sal¬
ly and Tim Walker provided material of Echeandia
nayaritensis. Curators of the following herbaria
made material available for study: ARIZ, BRIT,

CAS, DS, DUKE. ENCB, F, GH, IA, K, LL, MEXU,
MICH, M0, NY, PAUH, TEX, UC, US, WIS.

I iterature Cited

Baker, J. G. 1876. Revision of t He genera and species of
Antherieeae and Eriospermeae. J. Linn. Soc., But. 15:
253—363.

(’.base, M. \\.. P. J. Rudall & J. G. Conran. 1006. New
circumscriptions and a new family of asparagoid lilies:
Genera formerly included in Anthericaceae. Kew Bull.
51: 667-680.

Cruden. R. W. 1087. Hagenbachia, a misplaced genus of
New World Liliaceae. Nordic J. Bot. 7: 255—260.

- . 1094. Echeandia. I'p. 27—30 in: G. Davidse, M.

Sousa & A. (). (dialer (editors), Flora Mesoamericana,
Vol. (>, Alismataceae a Cyperaceae. Universidad Na-
eional Autonoma de Mexico, Mexico, D.F.; Missouri Bo¬
tanical Garden, St. Louis; The Natural History Museum,
I .ondon.

Dablgren. R. M. T.. II. T. Clifford & P. F. Yeo. 1985. The
Families of the Monocotyledons. Springer- Verlag, Ber¬
lin. Heidelberg, Germany.

Greenmail. .1. M. 1898, Diagnoses of new and critical Mex¬
ican Phanerogams. Daedalus. Proe. Amer. Acad. Arts
33: 471-489.

Johnston, M. C. 1990. The Vascular Plants of Texas, 2d
Fd. Published Privately, Austin. Texas.

Kativu. S. & I. Nordal. 1993. New combinations of African
species in genus Chlorophytum (Anthericaceae). Nordic
J. Bot. 13: 59—65.

McVaugh, R. 1989. f lora Novo-Galieiana, Vol. 15. Bro-
meliaeeae to Dioscoreaceae. Univ. Michigan Herbari¬
um. Ann Arbor.

Ortega. C. G. 1800. Novarum, aut rariorum plantarum hor-
ti reg. botan. inatrit. diseriptionum decades, . . . Matriti,
Fx Typographia Ibarriana.

Systematics Association Committee for Descriptive Ter¬
minology. 1962. Terminology of simple symmetrical
plane shapes. Taxon I I: 145—156. 245—247.

Wealherbv, C. A. 1910. A preliminary synopsis of the ge¬
nus Echeandia. Daedalus. Proe. Amer. Acad. Arts 45:
387-394.

Austral Hepaticae 29. More New Taxa and Combinations in
Telaranea (Lepidoziaceae) and a New Name for
Frullania caledonica (Schuster) Schuster (Frullaniaceae) from

New Caledonia

John J. Engel and G. L. Smith Merrill

Department of Botany, The Field Museum, Chicago, Illinois 60605-2496, U.S.A.

Abstract. Telaranea consobrina, T. fragilis, and
T. palmata, new species from Australasia, are de¬
scribed and illustrated. Telaranea clatritexta (Ste-
phani) Engel & Merrill and T. quadricilia (Ste-
phani) Engel & Merrill of Australasia and T.
fernandeziensis (Stephani) Engel & Merrill from
Juan Fernandez are new combinations. Frullania
neocaledonica of New Caledonia is proposed as a
new name for Frullania caledonica (Schuster)
Schuster, an illegitimate name.

The following new taxa and new combinations,
together with those previously published (Engel &
Merrill, 1995), are the result of a phylogenetic
study of the genus Telaranea, which is currently in
preparation by the authors. The names are here
published separately to make them immediately
available for use. Studies on Australasian Hepati¬
cae have also revealed a nomenclatural change re¬
quired in the genus Frullania.

Telaranea clatritexta (Stephani) Engel & Merrill,
comb. nov. Basionym: Lepidozia clatritexta
Stephani, Spec. Hep. 3: 583. 1909. TYPE:
Australia. Western Australia: Swan River,
Drummond, ex Herb. Kew (G).

Lepidozia complanata Herzog, Memoranda Soc. Fauna FI.
Fenn. 27(1950-1951): 92. fig. 39. 1952. Syn. nov.
Telaranea complanata (Herzog) Engel & Merrill.
Phytologia 79: 251. 1995 [June 1996J. TYPE: Aus¬
tralia. Western Australia: without specific loc., Goe¬
bel (holotype, not seen).

In a previous paper (Engel & Merrill, 1995), we
published the combination Telaranea complanata
(Herzog) Engel & Merrill. Further studies have re¬
vealed that Lepidozia clatritexta Stephani is the
same taxon, and a transfer to Telaranea is therefore
required.

Telaranea consobrina Engel & Merrill, sp. nov.
TYPE: Tasmania. Eastern slope of Black Bluff
just below summit, S of Burnie, 1250 m, 21
Mar. 1977, Engel 15799 (holotype, F; isotype,
HO). Figure 1.

Species nova T. meridianae affinis disco folii basin ver¬
sus 8 cellulis lato, lobis angustioribus, amphigastriis la-
tioribus brevioribusque, hyalodermate conspicuo differt.

Plants soft and flexuous, ascending to suberect,
pale green, medium in size, to 6 mm wide, includ¬
ing branches. Branching rather regularly 1 -pinnate,
the branches of the Frullania- type, at times becom¬
ing flagelliform; first branch underleaf undivided
(very rarely bilobed), broadly acuminate to lance¬
olate. Ventral-intercalary branches not seen. Stems
with cortical cells distinctly differentiated, thin-
walled, in 13—14 rows; cortical cells in section
slightly to distinctly larger than the numerous (49-
53) medullary cells. Leaves on main shoot oblique¬
ly spreading, contiguous, the disc plane or weakly
convex, the lobes ventrally decurved, the insertion
distinctly incubous, the disc broader than high;
leaves 475-600 gm wide X 440-525 gm long, the
leaves 4(5)-lobed to 0.4— 0.5, the lobes straight to
moderately divergent. Lobes narrowly acute to acu¬
minate, 4—5 cells wide at base, often 4 cells wide
in basal sector, then biseriate for 2-3 tiers, termi¬
nating in a short uniseriate row of 2—3 cells (or
sporadically a single cell or 2 laterally juxtaposed
cells); lobe cells ± isodiametric to short rectangu¬
lar, thin-walled. Leaf disc somewhat asymmetrically
cuneate, the disc (5)6 — 7(8) cells high (from median
sinus base to leaf base), 13-16 cells wide in distal
portion narrowing to 8(9) cells wide in basal por¬
tion. Cells of disc thin-walled, median cells short
rectangular, 24-38(42) gim wide X 42^19 gun long,
the cells in ± irregular rows; cuticle smooth. Un¬
derleaves somewhat smaller than leaves, 1.7— 2.3 X
stem width, 4(5)-lobed to 0.4—0.45, the lobes nar¬
rowly acute, 3-4 cells wide at base and biseriate
for 1—2 tiers, ending in a uniseriate row of 2—3
short cells; disc symmetrically broadly cuneate
(wider than high), 4—5 cells high (median sinus),
14—18 cells wide. Asexual reproduction lacking.

Androecia and gynoecia not seen.

Taxonomy. This species strongly resembles T.
meridiana of New Zealand, with broad, almost lon-

Novon 9: 339-344. 1999.

340

Novon

Figure I. Telaranea consobrina Kngel & Merrill. — 1. Leaf, in situ , dorsal view (sp = slime papilla). — 2. Leaf. — 3.
Leaf (sp = slime papilla). — 4. Leaf. — 5. Leaf lobe. — 6. Sector of main shoot with Frullania- type branch (= KB;
KBU — first branch underleaf), ventral view. — 7. Underleaf; note small-celled rhizoid initials in distal part of disc and
in basal portion of a lobe. — 8. First branch underleaves. — 9. Stem, cross section. (All from holotype, Engel 15799.)

Volume 9, Number 3
1999

Engel & Merrill
Austral Hepaticae 29

341

gitudinally inserted leaves and ventrally decurved
leal lobes. It dillers most notably in the shape of
the leal disc, which narrows to 8 cells wide at the
insertion (Fig. 1: 1 — 4) vs. 14-16 cells wide in T.
meridiana, and in the narrower lobes, which are
only 4-5 cells wide at base, and as many as 6(8)
cells wide in T. meridiana. The underleaves of the
Tasmanian plants are larger, roughly 2X the width
of the stem (Fig. 1: 6) (vs. 1.2 X the stem diameter
in T. meridiana ), and somewhat shorter. As in other
Telaranea species, the stem cortical cells are few
in number and form a conspicuous hyaloderm (Fig.
1:9).

The epithet consobrina (“cousin”) refers to the
evident relationship of this Tasmanian species to T.
meridiana (Hodgson) Hodgson of New Zealand.

Distribution and ecology. The species is known
from two stations in Tasmania, occurring in the
crevice ol a dripping cliff face in an area with al¬
pine vegetation (type locality) and in a creek bed
within an area having a mosaic of cushion plants,
Diselma, scattered pools, and small streams.

Additional specimen examined. TASMANIA. Cradle
Mtn.-Lake St. Clair Natl. Park, Plateau Creek area, be¬
tween Cradle Plateau and Marions Lookout, NNW ol Cra¬
dle Mtn., 1250 m, Engel 13960 (F).

Telaranea fernandeziensis (Stephani) Engel &
Merrill, comb. nov. Basionym: Lepidozia fer¬
nandeziensis Stephani, Kongl. Svenska Veten-
skapsakad. Handl. 46(9): 63. fig. 24e. 1911.
TYPE: Juan Fernandez. Mas a Tierra, El
Yunque, 24 Aug. 1908, Skottsberg (lectotype,
designated by Solari (1987), S).

We have examined the lectotype of Lepidozia fer¬
nandeziensis, and find that this plant belongs to the
complex ol Telaranea species that includes T. tetra-
dactyla (Hooker f. & Taylor) Hodgson.

Telaranea fragdlis Engel & Merrill, sp. nov.
TYPE: New Zealand. North Island: North
Auckland Prov., NE Waitakere Ranges, Swan¬
son University Reserve, Tram Valley Road, 95
m, Engel 20465 (holotype, F; isotype, CHR).
Figure 2.

Ex affinitate T. centipedis foliis fragilibus, disco lateri-
bus parallelis ubique 8 cellulis lato, et serie uniseriata
lobi cellulis angustioribus longioribusque distinguenda.

Plants delicate, flexuous, prostrate, glaucous,
whitish to bluish green, dull and distinctly water
repellent; shoots medium in size, to ca. 1 cm wide,
including branches. Branching loosely and irregu¬
larly 1-pinnate, the branches ol the Frullania- type,
rarely becoming flagelliform; first branch underleaf

undivided, subulate. Ventral-intercalary branches
present. Stems with cortical cells distinctly differ¬
entiated, in 9-12 rows, thin-walled; medullary cells
ca. 16. Leaves fragile, erose-truncate (the lobes all
or mostly broken oil, and often the distal tiers of
disc cells missing), widely spreading, at times near¬
ly at right angles to stem, distant to loosely imbri¬
cate, plane, strongly horizontally oriented, the in¬
sertion distinctly incubous; leaves 260-290 gm
wide, the fragmented leaf 280-390 gun long (in¬
cluding basal cell of lobe), 465-505 gun long with
lobes, leaves subsymmetric, 4-lobed to ca. 0.4, the
lobes ± parallel with disc margins or only slightly
divergent. Lobes (when present) subcaudate, 2 cells
broad at base, terminating in a uniseriate row of 6-
7 cells; lobe cells thin-walled, the distal cells of
uniseriate row slender, ± elongate. Leaf disc ±
symmetrically quadrate to subrectangular, 5—6(7)
cells long (from median sinus base to leal base),
mostly 8 cells wide throughout; margins ± straight
to weakly arched. Cells of disc in regular longitu¬
dinal rows, thin-walled but firm to moderately
thick-walled; median disc cells large, subquadrate,
33^5 gun wide, 40-52 gun long; cuticle a finely
granular and faintly striate coating. Underleaves
much smaller than leaves, 4-lobed to 0.75-0.85,
the lobes ciliiform, the cells of uniseriate row 3-4,
elongated, thin-walled; disc abbreviated, 2 cells
high, 8 cells wide. Asexual reproduction evidently
by fragmentation of leaf lobes and disc.

Androecia and gynoecia not seen.

Taxonomy. Hodgson (1956: 606) observed that
the leaves of this species “present a curious ap¬
pearance, in that the apices are either crenulate
with protruding cells of the discus, or with 1 or 2
segments showing as reduced to one roundly quad¬
rate cell, sitting as it were on 2 terminal cells of
the lengthwise rows,” which aptly describes the
leaves of this species (Fig. 2: 7, 10). It is likely that
detached lobe and disc cells function as gemmae.
Hodgson (1956) cited several populations with
these characteristics, and included them under the
name Lepidozia centipes Taylor.

Telaranea fragilis resembles T. centipes in being
dull, glaucous, and water-repellent.

Distribution and ecology. Known only from a
few scattered sites in North Auckland Province,
New Zealand. The species occurs on moist, clayey
banks or at times over rock in lowland forests. The
type occurred on a vertical clayey bank above a
small stream in an old Kunzea forest with Agathis
and (common) Phyllocladus trichomanoides Don.

Additional specimen examined. NEW ZE ALAND.

342

Novon

Figure 2. Telaranea elegans (Colenso) Kngel & Merrill (1-6; see Fngel & Merrill, 1995) and T. fragilis Engel &
Merrill (7—13). — 1. 2. Leaves. — 3. Distal portion of leal. — 4. Leaf lobe. — 5. Underleaf. —6. Stem, surface view. —
7. Sector of main shoot, dorsal view. — 8. Leaf lobes (cuticle shown in part). — 9. Underleaf. — 10. Sector of branch
showing (top leaf) fragmenting of leaf tip forming a gemmae like body (at arrow) and (lower leaf) a disc in process of
partially breaking away (fracture at arrow). — 1 I. Half-leaf. — 12. Branch base (FB = Frullania- type branch; FBU =
first branch underleaf), the underleaf of main shoot shown iu part. — 13. First branch underleaf shown in Figure 12.
(Figs. 1, 4—6, from type of T. elegans, Winkelmann s.n., WELT; 2, 3, from Engel 17216, Australia. New South Wales,
Morton Natl. Park. Fit/.roy Falls; 7—13, from type of T. fragilis, Engel 20465.)

Volume 9, Number 3
1999

Engel & Merrill
Austral Hepaticae 29

343

Figure 3. Telaranea palmata Engel & Merrill. — 1. Portion of main shoot (the terminal branches not shown, but note
half leaves), dorsal view. — 2. 3. Leaves. — 4, 5. Leaf lobes. — 6. Underleaf. — 7. Antheridial stalk. (Figs. 1—6, from
type. Engel 16251: 7, from Engel 19365. Tasmania. Cradle Mtn.-Lake St. Clair Natl. Park, between Ml. Rufus and \lt.
Hugel.)

344

Novon

North Island: North Auckland Prov.. Waipoua Forest. Al¬
lison H7II (CHR).

Telaranea quadricilia (Stephani) Engel & Merrill,

comb. nov. Basionym: Lepidozia quadricilia
Stephani, Species Hepaticarum 6: 338. 1922.
TYPE: Australia. New South Wales, Clarence,
Verreaux (G).

Telaranea quadricilia is similar to 7. lindenbergii,
but differs by the ineubously inserted leaves, the
longer leal lobes, and the papillose cuticle.

Telaranea palmata Engel & Merrill, sp. nov.
TYPE: Tasmania. Ridge SE of Black Bluff
near junction of access road to plateau area
and road to Devonport gold mines, S-faeing
slope, S of Burnie, 1000 m, Engel 1625] (ho-
lotype, F; isotype, HO). Figure 3.

Telaranea patentissimae affinis foliis palmato-lobatis.
insertione foliorum transversa, disco hreviore, serie unise-
riate loborum longiore, et cellulis disci atque loborum iu-
crassatis diversa.

Plants with a rather stiff and wiry appearance,
yellowish green; shoots small, to 0.6 cm wide, in¬
cluding branches. Branching somewhat irregularly
and loosely l(2)-pinnate, the branches remaining
rather short, the branches of the Frullania- tvpe. oc¬
casionally to frequently flagelliforin; first branch
underleal undivided and subulate. Ventral-interca¬
lary branches common. Stems with cortical cells
markedly differentiated, in 12 rows; cortical cells
in section larger than the numerous (ca. 50) med¬
ullary cells. Leaves on main shoot rigid, the disc
widely spreading to squarrose, distant to loosely im¬
bricate, moderately concave to hand-like, the lobes
erect and incurved, at times subfalcate, the inser¬
tion transverse to weakly incubous; leaves 480—
665(700) /am wide (measured between tips of lobes)
X 400-510 /am long, moderately asymmetric, ±
equally palmately 4-lobed to ca. 0.6. the lobes often
widely divergent. Lobes attenuate to subcaudate,
2^1 cells wide at base (when 4 cells wide some¬
times with an additional 3— 4-seriate tier), then bis-
eriate for 1(2) tiers, terminating in a uniseriate row
of 4—6 cells; lobe cells ± firm, often distinctly
thick-walled. Leaf disc symmetrically to somewhat
asymmetrically short euneate, 4 (rarely 5) cells high
(from median sinus base to leaf base), 13—16 cells
wide in distal portion, narrowing to 8 cells wide at
tin- insertion. Cells of disc moderately to distinctly
thick-walled, the cells in ± regular tiers, the me¬
dian and basal disc cells 31—36 /am wide X 39-
48 fim long; cuticle smooth. Underleaves much
smaller than leaves, 4-lobed to ca. 0.5 or a little

more, the lobes divergent, ciliiform, straight, basi¬
cally 2 cells wide at the base, the cells of uniseriate
portion 3(4), slightly elongated; disc symmetrically
subquadrate to weakly euneate, 3(4) cells high (me¬
dian sinus); disc 8-11 cells wide in distal portion,
8 cells wide at base. Asexual reproduction lacking.
Androecia and gvnoecia not seen.

Taxonomy. This species differs from most Te¬
laranea species in the transverse insertion of the
leaves (Fig. 3: 1). The leaves of this species are
palmately lobed (to 0.6) like those of T. praenitens
(Lehmann & Lindenberg) Hodgson, hence the
name. However, the cuticle of T. palmata is smooth
vs. distinctly striate papillose in T. praenitens.

Distribution and ecology. Endemic to Tasman¬
ia, and, lor the most part, a subalpine-alpine spe¬
cies, occurring above 1000 m in protected, moist
niches. The type is from 1000 m elevation in a
mosaic of Gymnoschoemus (button grass), subalpine
shrubs, Nothofagus cunninghamii (Hooker) Oer¬
sted, and rocky outcrops.

Additional specimens examined. TASMANIA. Mt.
Field Nall. Park. Tarn Shelf . below and E of Rodway
Range, 1270 in. Engel 14358 (F); Cradle Min. -bake St.
Clair Nall. Park, Lake Si. Clair area, between Ml. Rufus
and Ml. Hugel, 1120—1130 m. Engel 19363 (F).

Frullania neocaledonica Engel, nom. nov. Re¬
placed name: Frullania caledonica (Schuster)
Schuster. Hep. Anthoc. No. Amer. 5: 34. 1992;
non Gottsche ex Stephani, Hedwigia 33: 156.
1894. Neohattoria caledonica Schuster, J. Hat-
tori Bot. Lab. 33: 291. 1970. TYPE: New Cal¬
edonia. Mt. Mou, Compton.

Acknowledgments. The senior author gratefully
acknowledges support of the National Science
Foundation (grants BMS76-03616 and DEB-
8109680), which funded fieldwork in New Zealand,
Tasmania, and Australia, and the National Geo¬
graphic Society, which supported field studies in
New Zealand (grants 5375-94 and 5795-96). The
senior author thanks Zorica Dabich for preparing
the habit figures of the illustrations.

Literature Cited

Engel. J. .1. & G. L. Smith Merrill. 1995 [1996|. Austral
Hepatieae 23. New laxa and new combinations in Te¬
laranea Spruce ex Schiffn. (Lepidoziaceae). Phytologia
79: 250-253.

Hodgson, E. A. 1956. New Zealand Hepatieae (Liver¬
worts) — IX. A review of llie New Zealand species of the
genus Lepidozia. Trans. Roy. Soc. New Zealand 83:
589-620.

Solari, S. S. 1987. lepidozia disticha vs. L. fernandeziensis
(Hepatieae). Bol. Soc. Argent. Rot. 25: 143—147. fig. 1.

Simaba docensis , a New Brazilian Species of Simaroubaceae

Edivani Villaron Franceschinelli

Departamento de Botanica, Institute) de Ciencias Biologicas, Universidade Federal de
Minas Gerais, 31.270-010, Belo Horizonte, MG, Brazil

Kikyo Yamamoto

Departamento de Botanica, Institute de Biologia, Universidade Estadual de Campinas,

13.083-970, Campinas, SP, Brazil

Abstract. Simaba docensis, a new species of Si¬
maroubaceae, is described and illustrated. Its geo¬
graphical distribution seems to be restricted to
semideciduous forest in the region of the Rio Doce
basin, Minas Gerais, southeastern Brazil. Accord¬
ingly, it is likely to be a rare endemic and endan¬
gered species. It is very closely related to the only
other Amazonian species of section Floribundae
Engler, S. paraensis Ducke, from which it is basi¬
cally distinguished by the number and shape of leaf¬
lets, leal indument, and flower and fruit size.

We describe here a new species of Simaba Au-
blet (Simaroubaceae), .S', docensis. Its geographical
distribution is apparently restricted to southeastern
Minas Gerais, Brazil, in the semideciduous forest
of the Rio Doce basin. The closest species to S.
docensis is S. paraensis Ducke.

The South American genus Simaba has 25 spe¬
cies, divided into three sections: Tenuiflorae, Flo¬
ribundae, and Grandiflorae (Engler, 1874; Cron-
quist, 1944). They are trees or shrubs with
pinnately compound leaves. The main diagnostic-
characters used to distinguish the three sections of
Simaba are listed in Table 1. Section Tenuiflorae is
an Amazonian taxon, whereas the other two sec¬
tions are mostly extra- Amazonian. Simaba paraen¬
sis is noteworthy bec ause it is the only Amazonian
species of section Floribundae. Furthermore, it can
have small flowers, as found in section Tenuiflorae,
and it has stamen appendages as occur in section
Grandiflorae. However, S. paraensis has large pa¬
niculate inflorescences and other flower characters
similar to those found in Floribundae, which places
it in this section. The new species resembles S.
paraensis in its general appearance and because it
has androecium characteristics of the section Gran¬
diflorae. However, 5. docensis fits better the char¬
acteristics of section Floribundae than S. paraensis
because of its larger flowers.

Couepia monteclarensis Prance, another endemic

species of the Rio Doce basin, is also taxonomically
very close to two Amazonian congeneric species: C.
sandwithii Prance and C. bernardii Prance (Prance,
1989). Other field collectors (J. A. Lombardi & K.
Yamamoto, pers. comm. 1999) have noticed the oc¬
currence of Amazonian species affinities in the
semideciduous forest of the Rio Doce basin. The
conservation and detailed studies of this forest can
help us to understand the evolution and history of
the South American forests.

Simaba docensis Franceschinelli & K. Yamamoto,
sp. nov. TYPE: Brazil. Minas Gerais: Marli-
eria, Parque Estadual do Rio Doce, on pen¬
insula at Lake Helvecio, 24 Nov. 1975 (fl),
Heringer & Eiten 15182 (holotype, MO; iso¬
type, US). Figure 1.

Ad sectionem Floribundae Engler referenda, S. paraen¬
sis Ducke androecio similis, sed foliis 7— 10 jugatis, foliolis
subtus puberulis vel pi losis, floribus majoribus 8—10 mm
longis dilfert.

Tree 10—30 m tall. Branches glabrous with gray
bark, strongly grooved and furrowed longitudinally,
with prominent transversal scars due to leaf and
bud abscissions. Leaves with 6—21 leaflets, usually
7—10 jugate, at apex of the branches; raehis 7.5—
20 cm long, cylindric, puberulous. Leaflets sessile
or subsessile, opposite or subopposite, interjuga 2—
1.5 cm long. Leaflet 2.5-5 X 1.1-1. 7 cm, blade
oblong to obovate or elliptic, chartaceous, apex
usually acute, sometimes obtuse to rounded, with
gland at the tip, lateral leaflet base usually oblique
and terminally cuneate, puberulent beneath with
prominent central vein, glabrescent and slightly
wrinkled above with prominent and puberulent
central vein; lateral veins not visible. Inflorescence
paniculate, 8-22 cm long, with femigineous to-
mentose axes, secondary axes 1.5—11 cm long, usu¬
ally forming a right angle with the primary axis,
subtended by bracts narrowly oblong to obovate
with gland on the tip, 2.0— 5.0 cm long, flowers at

Nov ON 9: 345-348. 1999.

346

Novon

Table 1 . Main diagnostic

characters used to distinguisl

h the three sections ol Sinuiba.

( characters

Sect. Tcnuiflorac

Sect. FlorUnindaae

Sect. Grandiflorac

Inflorescence type

raceme or small panicu¬
late

large paniculate

large lew-branched panic¬
ulate or pseudoraceme

Mower length

3.5—7 mm

(5.5— )7— 15 mm

15—35 mm

Calyx

united at the base to the
middle length

united to the middle
length, rarely at the
base and eupuliform
(.S', ins ignis)

eupuliform

Stamen appendage in indu-

glabreseent to hairy, free

hairy, free or united, but

hairy on their edges.

ment

from each other or

tml in a stamen tube

hairs intertwined form-

slight!) united, hut not
in a stamen tube

(except .S', paracusis
and .S', docensis)

mg a stamen tube

Stamen appendage length

from ¥i to Vt of the f i hi-

from Yi to -Va of the fila-

almost the same length as

ment length, covering

ment length, covering

the filament, covering

the basal part of the

the ovary and part of

the ovary and almost

ovary to the whole ova-

the style

the whole style

( Geographical distribution

Amazonian region

mostly extra- Amazonian

mostly extra- Amazonian

the apex of t he axes. Flowers 8—10 mm long, sub¬
tended by 3 pilose, greenish braeteoles with gland
near tip. Sepals light green, triangular, united only
basallv, apex aeute, 2—3 X 1.2— 1.7 mm. Petals light
green or slightly ferrugineous externally, tomentose
above and beneath, 8—10 X 2-3 mm, oblong with
apex rounded to aeute. Stamens 5.5—6 mm long,
filaments 5—5.5 mm long; appendage 4.2^4. 7 mm
long, adaxiallv pilose, completely united to fila¬
ment, with two apical teeth; anthers 0.5— 0.8 X 0.2
mm, oblong, white. Gynophores 1 .5—2 mm high, pi¬
lose; ovaries 5-carpellate, 1.5-1. 7 mm high, pilose,
orange; styles 3-3.3 mm long; stigmas minute, 4—
5-lobed. Drupes 2—2.2 X 1.5— 1.6 cm, ellipsoid,
densely ferrugineous pilose-villous.

The specific epithet “docensis" refers to tin* col¬

lection site of the type specimen, the Parque Es-
tadual do Rio Doee. The only two collections known
for this species were made in this region. We have
tried to re-collect this species, but we could only
find one tree at the same locality where the type
specimen was collected. This suggests that S. do¬
censis may be a rare species restricted to the Rio
Doee Basin forest. Its geographical distribution is
the simplest character that separates S. docensis
from .S'. paraensLs. All characters used to distinguish
S. docensis from .S', paraensis are presented in Table
2. Besides geographical distribution, leaflet apex
shape, and secondary vein visibility, sepal and gy-
nophore size and fruit color and size are the most
helpful characters that separate these two species.

I’amlypes. BRAZIL. Vliuas Gerais: Mimictpio de

Table 2. Main diagnostic characters used to distinguish Simaha docensis from .S. paracusis.

Characters

.S. docensis

S. paraensis

( Geographical distribution

Minas (Gerais (Zona da Mata). Brazil

Amazonian Region of Brazil, Bolivia, and Vene¬
zuela

1 .eaflet number per leaf

(>-21

6-13

Leaflet size

2.5— 5.0 X 11 — 1.7 cm

3.5-9 X 1 .3-3.3 cm

Leaflet indument

puberulent beneath, glabreseent

glabrous on both sides, sometimes with puberu-

above

lent veins

Leaflet apex shape

usually acute, obtuse to rounded

retuse or obtuse to emarginate

Leaflet vein evidence

secondary veins not visible

secondary veins visible

Sepal size

2.0— 3.0 X 1.2-1. 7 mm

0.8— 1.5 X 1.0— 1.5 mm

Petal size

8.0-10.0 X 2.0— 3.0 mm

5.5— 9.0 X 1 .5-3.0 mm

( Gynophore size

1 .5—2.0 mm high

0.7-1 .7 mm high

Drupe size

2. 0-2. 2 X 1 .5-1.6 mm

1 .5—1.8 X 1. 1-1.3 cm

1 )rupe indument and color

densely ferrugineous, pilose-villous

puberulent and yellow

5 mm

Volume 9, Number 3
1999

Franceschinelli & Yamamoto
Simaba docensis from Brazil

347

Figure 1. Simaba docensis Franceschinelli & K. Yamamoto. — A. Inflorescence. — B. Vegetative branch. — C. Flower.
— I). Stamen. — E. Ovary and gynophore. A-E from E. P. Heringer & Eilen 15182 (MO). Drawn by Edivani V.
Franceschinelli.

348

Novon

Marlieria. Parque Kstadual do Kio Doce. Heringer & Eitcn
15194 (NY, US).

Acknowledgments. We thank the curators of
MO, NY, and US for making the materials of Sim-
aba available to us for study. We are grateful to
Tarciso S. Filgueiras for help with the Latin diag¬
nosis and to Esmeralda Z. Borghi for help with the
illustrations. This publication is part of a taxonomic
study of the genera Simaba, Simarouba, and Quas¬
sia, carried out in the Depto. de Botanica, IB, UN¬

ICAMP, supported by a scholarship to E. V. Fran-
ceschinelli (Bolsa de Aperfei^oamento da FAPESP,
processo n° 83/2526-2).

Literature Cited

Cronquist, A. 1944. Studies in the Simaroubaceae 111. The
genus Simaba. Lloydia 7: 81-92.

Kngler. A. 1874. Simaroubaceae. Pp. 197—246. tabs. 40—
46 in C. F. P. Martius & I. Urban. Flora Brasiliensis,
Vol. 12(2). Frid. Fleisher in comm.. Lipsiae.

Prance, G. 1989. Chrysobalanaceae. Flora Neotropica.
The New York Botanical Garden, Bronx.

Two New Species of Clusiella (Clusiaceae) with a
Synopsis of the Genus

Barry E. Hammel

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

ABSTRACT. Two new species of Clusiella , one from
Costa Rica and Panama (C. isthmensis ) and one
from Brazil (C. impressinervis ), are described, one
old species from Colombia ( C . cordifolia Cuatre-
casas) is newly placed in synonymy, and all eight
species in the genus are brought together in a key.
A description of the genus and brief descriptions
of the species are given, specimens are cited, and
distributions are mapped for all species. In view of
its stipuliform structures, contorted petals, psilate
pollen exine, baccate fruits, exarillate seeds, and
embryos with well-developed cotyledons, Clusiella'’ s
presumed close relationship to Clusia is chal¬
lenged.

Taxonomic History

When Planchon and Triana (1860) published
Clusiella (based on C. elegans ) the genus was con¬
sidered monotypic and known only from pistillate
flowers. It remained that way for nearly 100 years.
However, Planchon and Triana appended to their
treatment of Clusiella a full description of staminate
material of a species now known as C. axillaris
(Engler) Cuatrecasas. Planchon and Triana implied
that the collections on which that description was
based ( Spruce 2159 and 2854) might belong to a
second species of Clusiella , but they gave it no spe¬
cies name and presented the description under the
heading “ Asthotheca , Miers,” with the prophetic in¬
junction to see it “without prejudice as to place¬
ment among the other Clusioids” (Planchon & Tri¬
ana, 1860, ser. 4, vol. 14: 254). Although all
genera, except this one, are specifically designated
as such in their large publication, Planchon and
Triana did effectively publish the genus. The spell¬
ing “ Asthotheca," however, must certainly have
been a typographical error tor Astrotheca ; on page
240 of the same publication they referred to the
“genre manuscrit Astrotheca , Miers. “ Bentham and
Hooker (1862) repeated that error when they re¬
duced “ Asthotheca ” to a section of the genus Clu¬
sia. Meanwhile, Engler (1888) published Clusia ax¬
illaris Engler based on the same two Spruce
collections. Vesque (1892, 1893), using the correct
spelling, and aware that “the discovery of its female

flowers and of the male flowers of Clusiella might
one day permit the combination of these two gen¬
era” (Vesque 1892: 16), assigned two species to
Astrotheca, A. cuspidata Vesque and A. sulphurea
(Poeppig) Vesque. The former was clearly a super¬
fluous epithet because Vesque cited the earlier Clu¬
sia axillaris Engler and both Spruce collections.
Astrotheca sulfurea is not a Clusiella (Cuatrecasas,
1950). Cuatrecasas (1949, 1950) finally brought As¬
trotheca (as to A. cuspidata Vesque) to rest in syn¬
onymy under Clusiella and described five new spe¬
cies.

Key and Descriptions

Clusiella Planchon & Triana, Ann. Sci. Nat. Bot.
ser. 4. 14: 253. 1860. TYPE: Clusiella elegans
Planchon & Triana.

Astrotheca Miers ex Planchon & Triana, Ann. Sci. Nat.
Bot. ser. 4. 14: 254. 1860 (as "Asthotheca." a typo¬
graphical error). No species indicated, but based on
Spruce 2159 & 2854. from Brazil. LECTOTYPE,
here designated: Astrotheca cuspidata Vesque (=
Clusiella axillaris (Engler) Cuatrecasas).

Clusia subg. Criuva sect. Astrotheca (Miers ex Planchon
& Triana) Bentham & Hooker, Gen. PI. 1: 170. 1862
(as "Asthotheca").

Dioecious, epiphytic shrubs; resin (latex) clear;
internodes lenticellate, bearing several pairs of ca¬
ducous or rarely persistent bud scales usually clus¬
tered just above the node (one pair more distant).
Minute, interpetiolar stipuliform structures present,
these broadly triangular, searious, drying dark, ca¬
ducous. Petioles short, somewhat channeled by the
narrowly decurrent leaf blade. Leaves opposite, the
blades elliptic, sometimes cordate, ± coriaceous,
occasionally glaucous in dried material; major lat¬
eral veins often very indistinct until dried, merging
to a submarginal vein, the intersecondaries often
almost as prominent as the secondaries; translueid
(sometimes only with intense light) resin dots and
dashes present, occasionally drying dark and sur-
ficially visible but then usually only on the lower
surface, resin canals sometimes also visible on thin,
new leaves. Inflorescences terminal or evicted by
the growth of an axillary shoot and then appearing

Non on 9: 349-359. 1999.

350

Novon

axillary, at only one side of the node, short, 1— sev¬
eral-flowered, ± dichotomously branching eymes
with numerous small, decussate bracts dispersed on
the rachis. Flowers with 5 small imbricate sepals
and 5 larger contorted petals, the petals white (ma¬
roon or wine-red), often with yellow or pink mark¬
ings at the base within. Staminate flowers with the
filaments connate for most of their length into a
narrow column but free for a short distance apically
and forming a ± globose capitulum, the column
surrounded by a collar of clavate, resiniferous sta-
minodia or resin globules; anthers ± globose and
only slightly wider than the filaments (Figs. 1, 2);
pollen spheroidal-oblate, tricolporate, psilate (Fig.
3). Pistillate flowers with the ovary surrounded by
a resiniferous collar of staminodes that grade acrop-

etally from resin glands to more clearly stamen-like
structures; stigmas sessile, 5 to ca. 20, circular and
distinct (when few) or packed into a disk and in¬
distinctly triangular to rectangular (when many);
ovules sometimes with a long funicle. Fruit a ±
globose, many-seeded berry with about as many
locules as stigmas; seeds elongate-ovoid, small (ca.
1 X 0.5 mm) with a clear, gelatinous outer integ¬
ument attached only at the hilum. the exotegmen
pitted (Figs. 4, 5), with sinuous anticlinal cell walls,
lignified; endosperm white, embryo with cotyledons
ca. % the length of the seed.

Distribution. Clusiella is restricted to the Neo¬
tropics, where species occur in Costa Rica, Pana¬
ma, Colombia, Venezuela, Ecuador, northern Peru,
and northern Brazil (Fig. 6).

KkY TO THK SPECIES OF Cl.VSih.UA

la. Midrib impressed above; resin dots conspicuous (with hand lens) and prominent on both leaf surfaces ....

. C. impressinervis sp. nov.

lb. Midrib prominent above; resin dots obscure and visible mostly only on lower leaf surface.

2a. Leaves amplexicaul; petioles 2 mm or shorter . C. amplexicaulis

2b. Leaves not amplexicaul; petioles 2 mm or longer.

3a. Petioles 8— 10 mm; lateral veins mostly ca. 1 mm distant . C. albiflora

3b. Petioles mostly 5(-8) mm or less; lateral veins mostly ca. 2 mm or more distant.

4a. Fruits ± narrowly ovoid (ca. 5 mm diam. or less) with 5-7 distinct stigmas; leaves small. 4.5-

5.5(— 10) X 1 — 1.5(— 4) cm . C. elegans

4b. Fruits globose, larger and with It) or more connivent stigmas; leaves larger.

5a. Flowers wine-red [buds ca. 5 mm diam.; fruits ca. 1.5—2 cm diam.; twigs tan or white; leaf

acumen ca. 2 cm; Valle del Cauca, Colombia, and Esmeraldas, Ecuador] . C. pendula

5b. Flowers white.

6a. Leaves drying dark brownish gray, never glaucous; twigs black; flower buds ca. 10—15
mm diam.; fruits, ca. 4 cm diam.; Amazonia (Brazil, Colombia, Peru, Venezuela) ....

. C. axillaris

6b. Leaves drying lighter, usually glaucous; twigs mostly brown to white; (lower buds less
than 10 mm diam.; fruits ca. 10 mm diam.

7a. Fruit wall thick (ca. 0.5-1 mm) and appearing woody In dried material; leaves
strongly coriaceous, drying glaucous above and below, the apex usually abruptly

acuminate, the petiole ca. 5 mm; Colombia . C. macropetala

7b. Fruit wall thin (ca. 0.2 mm), leathery; leaves membranaceous, rarely drying glaucous
below, the apex gradually acuminate, the petiole 2-3 mm; Costa Rica and Panama
. C. isthmensis sp. nov.

Clusiella albiflora Cuatrecasas, Anales Inst. Biol.
Univ. Nac. Mexico 20: 111. 1949. TYPE: Co¬
lombia. Valle del Cauca: Rfo Yurumanguf,
Pena de Candelario, Costa Pacifica, 10 m,
Feb. 1944 (stam. fl), Cuatrecasas 16123 (ho-
lotype, F; isotype, US).

Bud scales sometimes persistent, to ca. 4 cm
long, their scars up to 3.5 cm above the node. Pet¬
ioles 8—10 mm; leaf blades elliptic, 8—12 X 4—6
cm, the base acute, the apex acuminate to a 1.5—
2.5-cm acumen, drying glaucous above, dark tan
above and below, lateral veins ca. 1 — 2( — 4) mm
apart, merging to a distinct submarginal vein ca.
1—1.5 mm from the margin, the intersecondaries
prominent and easily confused with the secondar¬

ies. Inflorescences branched several times from
above a peduncle ca. 5 mm long, 1-several-flow-
ered; flower buds ca. 8 mm diam.; petals ca. 10 X
5 cm. Staminate flowers with the staminal column,
including capitulum, very short, ca. 3 mm long.
Fruits globose, ca. 10 mm diam., the stigmatic disk
with ca. 6 closely packed, ± triangular stigmas,
fruit wall rather thick, ca. 0.05 mm.

Distribution. Colombia, from near sea level.

This species is still known only from the type
and one paratype. It has longer petioles and closer
lateral veins than the other species, as well as a
very long acuminate leaf tip and a very short sta¬
minal column. In leaf shape and color, though not
in size, it appears most similar to C. elegans.

Volume 9, Number 3
1999

Hammel

Clusiella

351

I' inures 1-5. Clusiella (1—3, C. elegans Planchon & Triana; Hammel & Kress 11287: 4, 5, C. isthmensis Hammel,
Hammel & Trainer 12721). — 1. Slaminate bud with petals removed, showing undehiseed anthers and staminal column
subtended by whorl of resinous staminodia. — 2. Mature staminate flower. — 3. Pollen grain. — 4. Seed with integument
laid back to show pitted exotegmen. — 5. Seed with the clear integument partially removed. Scale bar: for Figs. 1, 4
& 5 = 3 mm; for Fig. 2 = 6 mm; for Fig. 3 = 10 gm.

352

No von

O C. albiflora
V C. amplexicaulis
^ C. axillaris
& C. elegans
■ C. impressinervis
• C. isthmensis
+ C. macropetala
l/C. pendula

Figure 6. Distribution of Clusiella. (Occurrence of C. axillaris with an arrow pertains to off-map collection in Paid
state of Brazil.)

Additional specimen examined. COLOMBIA. Cauca:
Rfo Micay. Guayabal, Costa Pacifica. 5—20 in. Feb. 1943
(fr). Cuatrecasas 14132 (F. US).

Clusiella amplexicaulis Cuatrecasas, Revista
Acad. Colomb. Ci. Exact. 8: 61. 1950. TYPE:
Colombia. Valle del Cauca: Rfo Cajambre,
Barco, .5—80 m, Apr. 1944 (stain, fl), Cuatre¬
casas 1718H (holotype, F).

But! scales up to 7 mm above the node. Petioles
indistinct, ca. 1-2 mm; leaf blades ovate, 13-15 X
7—9 cm, the base cordate and ± amplexicaul, the
apex abruptly acuminate, drying slightly glaucous
above, light tan on both surfaces; lateral veins (3— )5—
10 mm apart, merging to a distinct submarginal
vein ca. 2 mm Irom margin, the intersecondaries
faint but occasionally reaching to the submarginal
vein; translueid resin dashes faint. Inflorescence
with ca. 1—3 flowers (in bud). Fruits unknown.

Distribution. Colombia, from near sea level.

This species is still known only from the type
collection, material with immature staminate flow¬
ers only. The large, amplexicaul leaves with short
petioles and distant primary lateral veins distin¬
guish it from all other species. In leaf color and
texture, C. amplexicaulis is similar to C. elegans.

Clusiella axillaris (Engler) Cuatrecasas, Revista
Acad. Colomb. Ci. Exact. 8: 61. 1950. Clusia
axillaris Engler, in Mart., Fl. Brasil. 12: 413.
1888. Astrotheca cuspidata Vesque, Epharm.
Genitalia Foliaque Clusiearum et Moronobear-
um. 3: 16, lab. 80 & 81. 1892. TYPE: Brazil.
Amazonas: Sao Gabriel de Cachoeira: Rio Ne¬
gro, Jan. -Aug. 1852 (stam. fl). Spruce 2159
(lectotype, here designated, GH; isolectotype,
GOET! photo A, MO).

Bud scales sometimes persistent, 5—20 mm long,
the scars up to 20 mm above the node. Petioles 4—
9 mm, deeply channeled; leaf blades elliptic, 10-

Volume 9, Number 3
1999

Hammel

Clusiella

353

13(— 14) X 4— 6(— 7) cm, the base acute to rounded,
the apex acute to acuminate to a 5-20-mm acumen,
drying somewhat shiny, dark olive gray above, tan
below, lateral veins 2.5—5 mm apart, merging to an
indistinct submarginal vein 1 mm or less from the
margin, intersecondaries ca. 1-2 mm apart; translu-
cid resin dots visible only with intense light or on
thin, new leaves. Inflorescence usually 1 -flowered;
flower buds to 1.5 cm diam.; petals to 2 X 1.3 cm.
Staminate flowers with the staminal column, in¬
cluding capitulum, ca. 5-8 mm tall. Fruit to ca. 4
cm diam., the stigmatic disk with ca. 20 closely
packed rectangular stigmas, fruit wall rather thick,
ca. 0.5 mm.

Distribution. Amazonian Brazil, Colombia, and
Venezuela, from ca. 100 to 750 m.

This species is easily recognized by virtue of its
black twigs, dark-drying leaves, large flowers, and
large fruits. It is distinctive also for its Amazonian
distribution. It is probably related to C. impressi-
nervis and C. pendula.

Additional specimens examined. BRAZIL. Acre: E
slope of Serra da Moa, Apr. 1971 (stam. ff). Prance et al.
12584 (MO); Rio Negro, Rio Cauaburf, 750 m. Nov. 1965
(stam. ff), Maguire et al. 60422 (F, GH). Amazonas: Ma-
raa, Rio Japura, Dec. 1982 (pisl. fl). Plowman et al. 12252
(F); Sao Paulo de Oliven^'a, Apr. 1944 (fl, fr), Ducke 1626
(A, F, MO); basin o( Rio Solimoes, Dec. 1936 (stam. fl),
Krukoff 8626 (A. F. MO); basin of creek Belem, (stam. fl),
Krukofj 8933 (A, F, MO); Rio Solimoes, Igarape Jandiaru-
ba, Jan. 1969 (fl), Frdes 23906 (GH). Para: Road BR 22.
Camparema to Maranhao, km 66, vicinity of Piriroro, Nov.
1965 (pist. fl). Prance & Pennington 1971 (GH). COLOM¬
BIA. Amazonas: Leticia, Tarapaca, Parque Nacional Nat¬
ural Amacayacu, 200 m, Nov. 1991, Pipoly et al. 16055
(MO), 1(H) m, July 1992, Hildas et al. 5471 (MO). Guain-
fa: San Felipe Neri, on Rfo Negro, across from San Carlos
de Rfo Negro, Venezuela, 120 m. May 1979 (stam. fl),
Clark & Clark 7161 (MO). Vaupes: Rfo Apaporis, Sora-
jama (above mouth of Rfo Kananari) and vicinity, 275 m,
(pist. fl), Schultes <&• Cabrera 16089 (GH); Mitru & vicinity,
along Rfo Vaupes about 6 km below Mini at Tukanare,
Aug. 1976 (fr). Zarucchi et al. 1883 (GH). PERU. Lor¬
eto: Maynas, Mishana. Rfo Nanay between Iquitos and
Santa Marfa de Nanay, 150 m. Mar. 1979 (stam. fl). Gentry
& Aronson 25305 (F, MO), 130 m. Nov. 1978. Diaz et al.
1984 (MO); Alipahuayo (Estacion IIAP). Nov. 1984 (pist.
fl), Vdsquez et al. 5887 (F, MO); Requena, Jenaro Herrera,
Rfo Ucayali, May 1982 (fr), Encarnacidn 26132 (MO); Sa-
puena. Jenaro Herrera, 1 70 m, Nov. 1987 (stam. fl), Vds¬
quez el al. 10051 (F. MO). VENEZUELA. Territorio
Federal Amazonas: Dept. Atures, Rfo Autana, 90—110
m, Nov. 1984 (fr), Guanchez 3249 (MO); Dept. Rfo Negro,
middle part of Rfo Barfa, 80 m, July 1984 (fr), Davidse
27549 (MO); 3 km NE of San Carlos de Rfo Negro. 120
m, Apr. 1979 (stam. fl), Liesner 6777 (MO); Rfos Paci-
moni-Yasya, Casiquiare, I (K) — 1 40 m, Jan. 1954 (stam. fl),
Maguire. Wurdack A Hunting 37439 (GH); Rfo Negro, Pi-
cora de Cocuf, Dec. 1947 (stam.? fl), Schultes & Lopez
9436 (GH).

Clusiella elegans Planchon & Triana, Ann. Sci.
Nat. Bot. 13 (ser. 4): 254. 1860. TYPE: Co¬
lombia. Choco: 2000 m, 1866 (fr), Triana s.n.
(holotype, P; isotypes, F-fragment, G [photo F,
MO], W [photo MO]).

Bud scales inconspicuous, their scars clustered
at or just above the nodes. Petioles l-2(-3) mm;
leal blades elliptic, 4-4.5(-10) X 1— 1 ,5( — 4.5) cm,
the base acute (rarely rounded), the apex acuminate
to a l-1.5(-2)-cm acumen, drying glaucous and ol¬
ive green above, shiny and tan below, lateral veins
2.5-5 mm apart, merging to an indistinct submar¬
ginal vein ca. 1 mm Irom the margin, intersecon¬
daries laint, ca. 1—2 mm apart; resin dots and dash¬
es translucid or sometimes dark. Inflorescences
unbranched or often with up to 6 branches; flower
buds 2— 4(— 6) mm diam.; petals ca. 4X3 mm.
Staminate flowers poorly known, the staminal col¬
umn, including capitulum, ca. 2(— 6?) mm tall.
Fruits ovoid, 4-8(— 12) X 3— 5(— 10) mm, somewhat
apiculate; stigmas 5—7, ± distinct, circular.

Distribution. Panama and Colombia from near
sea level to 2000 m.

Originally described from pistillate material, this
species is still known from very few staminate col¬
lections. Only three such collections are known:
one from the disjunct population in Panama ( Hum¬
mel & Kress 11287) and two from Colombia ( Gar -
cia-Barriga 11130 and Luteyn et al. 10682). In
general, the plants are much more delicate than the
other species, but C. elegans is particularly distin¬
guished by its small, ovoid fruits with 5—7, more or
less distinct stigmas. As discussed below, the large-
leaved forms of C. elegans look much like C. tst li¬
me ns is and also like the small-leaved forms of C.
macropetala. Indeed, the glaucous upper surface of
the leaves, as well as their color and venation, sug¬
gests that the three species are closely related. Be-
tancur 802 (cited below) has leaves that are large
for the species and somewhat glaucous below as in
C. macropetala. The fruits of that collection are also
large for the species, but in concordance with
placement in C. elegans they have a thin and leath¬
ery, rather than thick and bony, fruit wall, and the
stigmas are few and more or less distinct.

Additional specimens examined. PANAMA. Code:
between Llano Grande and Coclecito, 600 m, July 1979
(fr), Antonio 1396 (MO). Dec. 1979 (fr), Antonio 3050
(MO), Mar. 1982 (stain, fl), Hammel & Kress 1 1287
(DUKE). COLOMBIA. Antioquia: Amalfi a Rumazdn.
1550 m, Sep. 1988 (fr), Betancur et al. 802 (MO). Choco:
between La Oveja and Quibdo, Apr. 1931 (fr). Archer 1753
(US); 15 km E of Quibdd, 75 m. Apr. 1931 (fr). Archer
2214 (US); Quibdo, Carretera Yute— 1 .loro, 70 m, Sep.
1976 (fr), Forero & Jaramillo 2721 (MO). 80 m, June

354

Novon

1983 (fr), Forero et al. 9700 (MO). 100 in, Aug. 1976 (fr).
Gentry & Fallen 17817 (pro parte, mixed collection with
C. macropetala, MO), Nov. 1088 (fr), Ramirez I (MO);
highway Bolivar-Quilxld, km 52-70, 5(H)— 600 m, July
1944 Garcia- Barriga 11130 (US); km 175—176, 456 m,
Mar. 1084 (Ir), Juncosa 2578 (MO); 37—40 km W of Id
Carmen, 671—1360 m. May 1084 (stain, fl). Luteyn et al.
10032 (MO); road to Cabf S\A of Tutuendo, 1(H) m. 10
Jan. 1070 (Ir). Gentry & Renteria 24450 (MO); Rfo Ser¬
rano, 4—6 km arriba de Guayabal, 50 m, Apr. 1975 (fr),
Forero et al. 1362 (MO). Valle del Cauca: Buenaventura,
Bajo Calima, 50 m, July 1088 (fr), Croat 69479. 69515
(MO), Dec. 1081 (fr). Gentry 35589 (MO), ca. 100 m. July
1084 (fr). Gentry & Monsalve 48413 (MO), Feb. 1989 (11),
Gentry et al. 65553 (MO), June 1987 (fr), Monsalve 1509
(MO), Oct. 1087 (fr). Monsalve 1982 (MO), 50-100 m,
Feb. 1984 (fr), Juncosa 2125 (MO); Buenaventura to Cali,
1(H) m. June 1944 (fr), Killip & Cuatrecasas 38926 (F.
(ill. US); Punta Arenas, Buenaventura Bay, near sea level,
June 1044 (fr). Killi/) & Cuatrecasas 38647 (US); Ouebra-
da de Aguadulce, 0-10 m, Feb. 1046 (pist.? fl), Cuatre¬
casas 19996 (F, US).

Clusiella impressinervis Hammel, sp. nov. TYPE:

Brazil. Amazonas: Tonantms, Mar. 1944 (stam.

fl), Ducke 1625 (holotype, US; isotypes. A, F).

Figure 7.

A speciebus aliis Clusiellae in costa supra impressa et
punctis resiniferis utrinque foliorum manifestis, differt.
Clusiella axillaris fortasse affinis, sed foliis et Horibus mi-
noribus.

Bud scales olten persistent, several sets liome
just above the node and one pair borne further
above. Petioles 2-3 mm, deeply channeled; leaf
blades elliptic to obovate, widest at or above the
middle, (7 — )8 — 1 9 X (2.5— )3— 3.5 cm, the base acute,
the apex abruptly acuminate to a 1.5-cm acumen,
drying shiny, light green on both surfaces; lateral
veins mostly 1 ,5( — 3) mm apart, merging to a very
indistinct submarginal vein mostly hidden along
the recurved margin; resin dots very conspicuous
on both surfaces. Inflorescences apparently 1-flow-
ered; flower buds 7—8 mm diam.; petals ca. 10 mm
long X 7 mm wide, while. Staminate flowers with
the staminal column, including capitulum, ca. 4
mm tall, filaments ± free apically but congested.
Pistillate flowers and fruits unknown.

Distribution. Brazil and Venezuela, from ca. 50
to 1220 m.

Since Clusiella impressinervis is so poorly known
it is difficult to relate to the other species. However,
by virtue of its non-glaueous leaves, its deeply
channeled petioles, its persistent bud scales, and
its Amazonian distribution, the species may be re¬
lated to C. axillaris. Leaves of that species also
occasionally have resin dots visible on the upper
surface. The new species differs from C. axillaris
by its smaller leaves and flowers and from all spe¬

cies not only by its impressed midrib but also be¬
cause its leaves dry light green on both surfaces.
More material, especially fruiting, is needed to
clarify its relationships.

Paratypes. VENEZUELA. Bolivar: Cerro Venamo,
1220 m. Jan. 1964 (11 bud), Steyermark et al. 92790 (F.
US).

Steyermark et al. 92790 will key to C. impressi-
nervis by virtue of its impressed midrib and resin
dots on both leaf surfaces. It also has deeply
grooved petioles and somewhat persistent bud
scales. However, it differs from the type by its ovate
leaves, ca. 6—7 X 3 cm, which are widest below
the middle, have a rounded base, and dry dull on
both surfaces, with upper surface olive gray and the
lower reddish tan, and by its shorter petioles only
ca. 1-2 mm long. It is also found at a much higher
elevation and is the easternmost collection known
of the genus. By virtue of sharing unusual vegeta¬
tive characters with C. impressinervis, this specimen
is tentatively included here. The description of the
species is based on the type collection alone.

Clusiella isthmensis Hammel, sp. nov. TYPE:
Costa Bica. Limon: Cerro Coronel, 5 m, Sep.
1986 (H, fr), Davidse & Herrera 31421 (holo¬
type, CR; isotypes, F, MO) Figure 8.

Clusiella macropetala Cuatrecasas affinis sed in exo-
carpio coriaceo (non osseo) folia membranaciore, et petio-
lio breviore differt.

Bud scales mostly deciduous, their scars con¬
gested at or near the nodes, rarely persistent ( Ken¬
nedy et al. 2394) with a pair up to 3.5 cm long
appearing above the node. Petiole 2—3 mm, some¬
what channeled; leaf blades elliptic-lanceolate,
widest below the middle (6-)8-14 X (2— )3— 4.5 cm,
the base acute to rounded, the apex gradually acu¬
minate to a 1— 2-cm acumen, in dried material
brown to glaucous and gray above, tan and some¬
times glaucous below, lateral veins 2-3(— 5) mm
apart, merging to an indistinct submarginal vein
mostly less than 1 mm from and hidden along the
recurved margin; translucid resin dots and dashes
sometimes also surfieially visible on lower surface.
Inflorescences mostly 1(— 3)-flowered; petals ca. 10
X 6 mm, white. Staminate flowers with the staminal
column, including capitulum, ca. 6 mm tall, the
filaments free for a short distance apically. Pistillate
flowers with apparently 10 stigmas, but these form¬
ing a disk and difficult to distinguish; ovary 10-
celled. Fruits ca. 1 cm diam. at maturity, translu¬
cent white, soft.

Distribution. Costa Rica and Panama, from
near sea level to 1500 m.

Volume 9, Number 3
1999

Hammel

Clusiella

355

Although this species was previously identified
as Clusiella elegans (D'Arey, 1980; Hammel, 1986),
it is more closely related to C. macropetala. As here
understood, C. elegans is generally a much more
delicate plant with smaller leaves, flowers, and
fruits than both C. isthmensis and C. macropetala.
Most importantly, the ca. 10 stigmas of the latter
two species are more or less fused into a stigmatic

cap, whereas those of C. elegans are fewer (ca. 5—
7) and distinct. In addition to the characters men¬
tioned in the diagnosis, C. isthmensis can he dis¬
tinguished from C. macropetala by its generally
smaller, more gradually acuminate leaf blades and
by its shorter, thinner petioles. Nevertheless, in leal
size, shape, and color, both C. macropetala and C.
elegans appear to overlap somewhat with C. isth-

356

Novon

Figure 8. Clusiella isthmensis Hanimel; Davidse & Herrera 31421. — A. Habit. — B. Pistillate flower. — C. Base of
leaf, abaxial view. — 1). Base of leaf, adaxial view.

mensis. The only other species that might he con¬
fused with C. isthmensis is C. pendula Cuatrecasas.
However, the leaves of this species dry darker with
more prominent lateral veins, are shiny rather than
glaucous, and the flowers are red or maroon rather
than white. One collection, Foster et al. 14661 ,
from lowland Bocas del Toro province, Panama, is

problematic. The leaf size, shape, and coloration,
as well as flower color and fruit size, all correspond
with C. isthmensis, hut the fruits have fewer and
nearly separate stigmas, more like those of C. ele-
gans.

Original field notes for the type of this species
indicate 11 duplicates. These were distributed as

Volume 9, Number 3
1999

Hammel

Clusiella

357

C. elegans, and their whereabouts, except as indi¬
cated above, are uncertain.

Paratypes. COSTA MCA. Heredia: I, a Selva Biolog¬
ical Station just E of the juncture of the Sarapiquf and
Puerto Viejo Rivers, ca. 100 m, (all collections from La
Selva are pistillate), Apr. 1981, Folsom 9673 (DUKE),
Apr. 1981, Folsom 9893 (DUKE, F. MO), Mar. 1980.
Homme! 8252 (DUKE, F, MO), May 1980, Hammel 8625
(DUKE). June 1980. Hammel 8934 (DUKE), July 1980.
Hammel 9286 (DUKE), June 1982. Hammel & Trainer
12721 (DUKE), Aug. 1981, Damon Smith 109 (DUKE.
MO), Oct. 1981. Damon Smith 457 (DUKE, F. MO);
Parque Nac. Braulio Carrillo, between Rio Peje and Rio
Sardinal, Atlantic slope of Volcan Barva, 1200-1400 m,
Nov. 1985 (fr), Grayum & Herrera 7867 (CR. MO); Esta-
cion el Ceibo. 520 m. Dec. 1992 (fr), Boyle 1448 (CR);
along W fork of Rio Sardinal, Atlantic slope of Volcan
Barva, 670 m, Dec. 1987 (fr). Grayum 8501 (F, INB. MO).
Linion: ca. 2 km S of Rio Colorado along new road and
ca. 1 km E, 5 m, Mar. 1987 (fl, fr), Stevens el at. 25074
(MO); Canton de Talamanca, entre Sukiit y Amubri, 700—
900 m. July 1989 (stam. fl), Hammel et al. 17625 (CR);
Alto Uren, 1190 m, July 1989 (stam. fl). Herrera 3368
(INB. MO); Alto Lari, 1300 in. Mar. 1992 (stam. fl). Her¬
rera 5190 (INB. MO); 1500 m. Mar. 1992 (stam. fl). Her¬
rera 5422 (INB, MO). PANAMA. Code: Continental di¬
vide above El Cope, 750 m, Feb. 1982 (stam. fl), Knapp
& Dressier 3410 (MO). Darien: Rio Tuquesa headwaters,
ca. 2 km from continental divide, ca. 400 m, Aug. 1974
(fl, fr), Croat 27213 (MO). Panama: El Llano— Cartf road,
ca. 350 m. Mar. 1973 (stam. fl), Croat 22894 (MO), Eeb.
1973, Kennedy et al. 2394 (MO). Feb. 1973, Kennedy et
al. 2397 A (MO), Mar. 1973 (fl. fr), Liesner 658 (MO). Mar.
1973 (stam. fl), Liesner 1208 (MO), Nov. 1985 (fl), Mc¬
Pherson 7585 (MO), Feb. 1975, Mori et al. 4701 (MO),
Eeb. 1975 (fl, fr), Mori et al. 4705 (MO). San Bias Com-
arca: El Llano— Cartf road, 13.8—15.8 km N of Interamer-
ican Highway, 325 m, Aug. 1984 (stam. fl), de Severs et
al. 3747 (MO); 19.1 km N of Interamerican Highway, 350
m. Nov. 1984 (fr), de Severs 4300 (MO). Veraguas: At¬
lantic slope NW of Santa Fe, 450—550 m. Dec. 1974
(stam. fl), Mori et al. 3875 (MO).

Clusiella macropetala Cuatrecasas, Anales Inst.
Biol. Univ. Nac. Mexico: 110. 1949. TYPE:
Colombia. Valle del Cauca: Rio Cajambre, La
Trojita, 5—50 in, 29 Feb. 1944 (stam. fl), Cua¬
trecasas 16578 (holotype, F; isotype, US).

Bud scales deciduous, the scars clustered at or
just above the nodes. Petioles 3-10 mm, shallowly
channeled; leal blades elliptic, ( 1 0 — ) 1 3 — 1 9 X (4—)6-
8 cm, the base acute to rounded, the apex acute to
acuminate to a 5— 15-mm acumen, in dried material
glaucous on both surfaces, pale gray-green above,
tan below, distinctly coriaceous; lateral veins 3 — 5(—
7) mm apart, merging to a submarginal vein ca. 1
mm from margin, the intersecondaries often ±
prominent; translucid resin dots and dashes visible
only under intense light. Inflorescences un¬
branched or branched several times, with ca. 1-6
flowers; flower buds ca. 7—10 mm diam.; petals 10—

17 X 6-10 mm. Staminate flowers with a staminal
column, including capitulum, 5—10 mm tall. Fruits
8—12 mm diam., the stigmatic crest with ca. 10
closely packed stigmas, fruit wall rather thick, ca.
0.7 mm.

Distribution. Colombia, from near sea level to
700 m.

As noted above, the leaves of C. macropetala are
glaucous on both surfaces and the fruits have a
thick, several-layered fruit wall that is bony when
dry. Smaller-leaved and smaller-flowered forms may
otherwise be confused with C. elegans and C. isth-
mensis (q.v.), but C. macropetala is probably more
closely related to the latter.

Additional specimens examined. COLOMBIA. Antio-
quia: Anorf, Valle rle Anon, entre Dos Bocas y Norf, 700
m, June 1971 (stam. fl), Soejarto 2869 (E, MO, (ill). Cho-
cd: Quibd6— Istmina road to Lloro, ca. 100 m. June 1983
(fl, fr), Forero et al. 9562 (MO), June 1983 (stam. fl). For-
ero et al. 9591 (MO). June 1983 (fl. fr), Forero et al. 9596
(MO), June 1982 (stam. fl). Gentry & Brand 36855 (MO),
June 1982 (fl, fr). Gentry & Brand 36958 (MO). June 1982
(fl. fr). Gentry & Brand 36963 (MO), Aug. 1976 (stam. fl).
Gentry & Fallen 17817 (pro parte, mixed collection with
C. elegans, MO); San Jos6 del Palmar, hoya del Rio San
Juan, alrededores de Docordo, 0 m. Mar. 1979 (stam.? fl),
Forero et al. 4328A (MO); Tutuendo— Quibdd road to Tu-
bado, 90 m. Jan. 1979 (fr). Gentry & Renteria 24507 (MO).
Valle del Cauca: Bajo Calima, Buenaventura, 100 m.
Jan. 1988 (stam. fl). Monsalve 2074 (MO); Rib Cajambre,
5—80 in. May 1944 (stam. fl), Cuatrecasas 17497 (E. US).

Clusiella pendula Cuatrecasas, Anales Inst. Biol.
Univ. Nac. Mexico 20: 109. 1949. TYPE: Co¬
lombia. Valle del Cauca: Bajo Calima, Bue¬
naventura, Bahia de Buenaventura, 0—10 m,
20 Feb. 1946 (stam. fl), Cuatrecasas 19877
(holotype, F; isotype, US).

Clusiella cordifolia Cuatrecasas, Revista Acad. Colomb.
Ci. Exact. 8: 61. 1950. Syn. nov. TYPE: Colombia.
Valle del Cauca: Bajo Calima, Buenaventura, coastal
thickets, Killip 11731 (holotype, US; isotype, GH).

Bud scales deciduous, the scars clustered at or
just above the nodes. Petioles ca. 3—5 mm, shal¬
lowly channeled; leaf blades elliptic to cordate,
(9.5-)11.5-16(-19) X (3.5-)4— 6.5(-8) cm, the base
acute or rounded to sometimes cordate, the tip acu¬
minate to a ca. 2-cm acumen, in dried material
often shiny (not glaucous) on both surfaces and
dark olive above, dark brown below; lateral veins
2.5—4 mm apart, merging to a submarginal vein ca.
1 mm from margin, intersecondaries often also
quite prominent; resin dots and dashes translucid.
Inflorescence usually much-branched and with sev¬
eral flowers; flower buds ca. 5 mm diam.; petals ca.
10 X 5 mm, purplish brown (maroon) to red. Sta-

358

Novon

minate flowers with a staminal column, including
capitulum ca. 3 mm tall. Fruits globose or slightly
ovoid, ca. 15—20 X 10—15 mm, green(?) at matu¬
rity; stigmas ca. 8 or 9, closely packed and difficult
to distinguish on lruits.

Distribution. Coastal Colombia and Ecuador
from near sea level to ca. 800 m.

This species is easy to distinguish by virtue of
its dark red instead of white or pinkish flowers and
especially by its dark-drying, shiny leaves with
prominent lateral veins. Based on the shape and
color of its leaves, it is most similar to the Ama¬
zonian C. axillaris.

The many new collections from Colombia, Valle
Dept., Buenaventura in Bajo Calima, near the type
locality of both C. pendula and C. cordifolia, allow
the conclusion that the latter should be considered
a synonym of the former. The leaf color, texture,
and venation of the type of C. cordifolia coincide
with those of C. pendula. Although the type spec¬
imens differ in shape of the leaf base, it is now
obvious that this can vary from acute to rounded to
slightly cordate even on a single specimen.

A label or numbering mix-up may have occurred
with the type specimen of C. cordifolia. As noted
by Cuatrecasas in the original description of the
species, the label reads “Tree, fruit blue.” However,
not only are trees and blue fruits otherwise un¬
known in the genus, the specimen is of a flowering
staminate individual.

Additional specimens examined. COLOMBIA. Valle
<lel Cauca: Bajo Calima, Buenaventura, Bahia de Bue¬
naventura. Apr. 1939 (stain. (1), Kill ip 34966 (F, US);
coastal thickets, Oct. 1922 (stain, fl). Killip 1 1731 (GH,
US); el Forge, sea level, June 1944 (fl, f'r). Killip & Cua¬
trecasas 38962 (F, US); Carretera Hans, km 22. 50 m, Dec.
1937 (fl). Gentry et al. 59534 (MO); Coneesion Carton de
Colombia, 50 — 1<M) m, July 1937 (stain. 7 fl). Faber-Lan-
gendoen & Renteria 1209 (MO), Dec. 1931 (sterile). Gen¬
try 35301 (MO), Feb. 1933 (stain. II). Gentry et al. 40394
(MO), Mar. 193b (slam. (I). Gentry et al. 53618 (MO), Apr.
1937 (slam, fl). Gentry et al. 56701 (A. MO). Nov. 1986
(stain, fl). Monsalve 1206 (MO). Dec. 1986 (stain, fl). Mon¬
salve 1405 (MO), Apr. 1987 (stain, fl). Monsalve 1457
(MO), Sep. 1987 (fr). Monsalve 1823 (MO). Oct. 1937
(stain, fl). Monsalve 1892 (MO), Oct. 1987 (fr). Monsalve
1933 (MO), Oct. 1987 (slant, fl). Monsalve 1980 (MO),
Jan. 1988 (stain, fl). Monsalve 2083 (MO), June 1982
(stain, fl). Murphy 558 (MO); Coneesion Pulpapel, 100 nt,
Aug. 1984 (fr). Monsalve 324 (GH); Quebrada Algeria,
Bahia Malaga, 50 nt, Dec. 1985 (pist. fl). Gentry et al.
53334 (MO); Bfo Calima. 1 -a Ksperanza, 5—10 m. Mar.
1944 (slant, fl), Cuatrecasas 16761 (F, US); road to Juan-
ehaco Palnteras, 1(H) m. July 1984 (pist. fl). Gentry et id.
47819 (MO). July 1984 (frj. Gentry et al. 48291 (MO).
ECUADOR. Esmeraldas: San Uorenzo, 22 km from Lila
on road to San Lorenzo, 800 nt. May 1990 (fl). Gentry et
al. 69990 (MO); Alto Tambo, 15 km oeste de Lita, 4(H)
m. Sep. I 'XX) (fr), Rubio ( >18 (MO); Carretera Lita— Alto

Tambo— La Punta, 400 m, Feb. 1991. Gudina A Moran
1263 (MO).

Relationships ok the Genus

Discussing the placement of Clusiella within the
family, Planchon and Triana (1860) pointed out that
the contorted aestivaton of the petals was charac¬
teristic of the tribe Moronobeae (subfamily Moron-
oboideae of Engler, 1925). However, because of the
sessile, largely cupuliform stigmas and non-ascen¬
dent ovules of Clusiella, they preferred to place it
in their tribe Clusieae, although they were not cer¬
tain of their disposition because of the lack of fruits
and staminate flowers. Engler (1925) also felt that
these incompletely known plants probably belonged
next to Clusia in subfamily Clusioideae. Subse¬
quent published accounts have made no explicit
statements about possible relationships of Clusiella,
although most imply a close relationship with Clu¬
sia (but see Hammel, 1984, unpublished thesis).

Numerous additional characters, including inter-
petiolar stipuliform structures, bud scales, resin
gland-dots in the leaves, psilate pollen exine, bac¬
cate fruits, small fovelate seeds that lack an aril,
and an embryo with unusually large cotyledons, are
now known for Clusiella and make this genus very
unusual, if not misplaced, among the clusioid gen¬
era. Stipuliform structures are uncommon in Clu-
siaeeae, but ones similar to those found in Clusiella
also occur in the Moronoboideae (e.g., Moronobea,
Symphonia ) and in Garcinia s.l. Bud scales also
occur in Moronobea and Symphonia and in some
related genera, while some species of Garcinia have
one or two pairs of bud scales, and bud scales occur
in some of the Calophylloideae. Contorted petal
aestivation does not otherwise occur within the
whole subfamily Clusioideae (Engler, 1925) but is
most likely the primitive state for the family (P.
Stevens, pers. comm. 1987). The psilate pollen ex¬
ine of Clusiella also appears to be otherwise un¬
known within the family except for Symphonia and
Moronbea (cf. Seetharam, 1983). It should be noted
that Seetharam's study and light micrographs of
“ Clusiella elegans Klug 1950," describing it as
having “perreticulate” exine, pertain to a species
of Quapoya (i.e., Clusia), not Clusiella. The small,
foveolate seeds and embryo with well-developed
cotyledons of Clusiella are most like those in the
Hypericoideae and basal Calophylloideae, e.g.,
Marila. Were it not for its epiphytic habit, dioecy,
and resiniferous androecium (all reminiscent of
Clusia), but particularly because ol its small fove¬
olate seeds and embryo structure, Clusiella would
seem to be closely related to Symphonia. More de-

Volume 9, Number 3
1999

Hammel

Clusiella

359

tailed, comparative studies of pollen and the foliar
resin system in particular are needed.

Characters such as those noted above in a sense
pull Clusiella strongly outward and downward from
Clusia. Work in progress toward an understanding
of relationships among genera in the whole family
in fact suggests that Clusiella may be sister to a
large clade including both the Clusioideae and Mo-
ronoboideae (P. Stevens, pers. comm. 1987). The
diminutive Clusiella, distinctively derived in many
ways, and long masquerading as Clusia' s little cous¬
in, seems to be an epiphytic refugium for a number
of relatively primitive characters in the Clusioi-
deae-Moronobiodeae, and could be called, more ac¬
curately, Clusia's great aunt.

Acknowledgments. This work was supported, in
part, by funds from the National Science Founda¬
tion through grants (e.g., DEB-9300814) to the au¬
thor and co-PI M. H. Grayum for the Manual to the
Plants of Costa Rica. An early draft of the manu¬
script benefitted greatly through correspondence
with Peter Stevens. I thank John Myers for the line
drawings of the two new species.

Literature Cited

Bentham, G. & J. D. Hooker. 1862. Genera Plantarum, ad
exemplaria imprimis in Herbariis Kewensibus servata.
Codicote, Herts, Wheldon and Wesley, New York.

Cuatrecasas, J. 1949. Gutiferas nuevas o poco conocidas
en Colombia. Anales Inst. Biol. Univ. Nac. Mexico 20:
91-1 12.

- . 1950. Notas a la flora de Colombia, X. Revista

Acad. Colomb. Ci. Exact. 8: 33—64.

D’Arcy, W. G. 1980. Guttiferae. In Flora of Panama. Ann.
Missouri Bot. Gard. 67: 969—1043.

Engler, A. 1888. In Martius, FI. Brasil. 12: 413.

- . 1925. Guttiferae. In A. Engler & k. Prantl, Die

naturlichen Pflanzenfamilien 21: 154—237.

Hammel, B. E. 1984. Systematic Treatments of the Cy-
clanthaceae, Marantaceae, Clusiaceae, Cecropiaceae,
Lauraceae, and Moraceae for the Flora of a Wet Low¬
land Tropical Forest, Finca La Selva, Costa Rica. Ph.D.
Dissertation. Duke University, Durham, North Carolina.

- . 1986. The vascular flora of La Selva Biological

Station, Costa Rica. Guttiferae. Selbyana 9: 203-217.

Planchon, J. & J. Triana. 1860. Memoire sur la famille
des guttiferes. Ann. Sci. Nat. Bot. ser. 4, 13: 306—376;
14: 226-367; 15: 240-319.

Seetharam, Y. N. 1983. Clusiaceae: Palynology and Sys-
tematics. Ph.D. Dissertation, Bangalore University.

Vesque, J. 1892. Epharmosis III. Paris.

- . 1893. Guttiferae. In A. & C. De Candolle, Mon-

ographiae Phanerogamarum. 8. Paris.

Synopsis of Chrysochlamys (Clusiaceae: Clusioideae: Clusieae)

in Mesoamerica

Barry E. Hammel

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166, U.S.A.

ABSTRACT. Delimitation of genera in the Tovomita
group of tribe Clusieae has fluctuated, in Central
America, with almost every new floristic work. Nev¬
ertheless, with some aberrations, the theme of a
unified Chrysochlamys/Tovomitopsis has persisted
and is here supported by concurring with Bentham
and Hooker’s placement ol Tovomitopsis in synon¬
ymy ol Chrysochlamys. The strictly South American
Balboa is also placed in synonymy of Chrysochla¬
mys-, this action necessitates a new name (here pro¬
vided) for its only species. A key to the 13 species
of Chrysochlamys recognized for Mexico and Cen¬
tral America is provided. An annotated list of all
Mesoamerican Chrysochlamys, including brief de¬
scriptions, taxonomic and nomenclatural notes, five
new combinations, and two new species, elucidates
the generic concept and provides a rationalized no¬
menclature for use in numerous regional floras. To¬
vomita croatii Maguire, a somewhat problematic
species, transferred to Tovomitopsis by D’Arcy, is
included in the key to Chrysochlamys and fullv dis¬
cussed.

Taxonomic History

The Tovomita group of the tribe Clusieae (Clu¬
siaceae, subfamily Clusioideae) comprises primar¬
ily arborescent, terrestrial genera with succulent-
capsuled fruits and 1 (rarely 2) large arillate seed(s)
per locule. Three genera therein, Tovomita, Chry¬
sochlamys, and Tovomitopsis, have suffered a par¬
ticularly tortuous taxonomic history. Planchon and
Triana (1860) made the genus Tovomitopis to ac¬
commodate Sprengel’s Bertolonia, a later homonym;
Bertolonia Kaddi (Melastomataceae) is now con¬
served over Bertolonia Spin (Myoporaceae) (Farr et
al., 1979). Planchon and Triana distinguished 7b-
vomitopsis from Chrysochlamys, claiming it had an
arilloid (originating from the micropyle) rather than
the true aril (originating from the funicule) they
claimed for Chrysochlamys. According to them, the
so-called aril of Tovomita is neither, but rather the
outer integument of the seed. These differences
were further supported by floral structure: Tovomi¬
topis, calyx 4-merous; Chrysochlamys, calyx 5-mer-
ous; Tovomita, calyx 2- or 4-parted, the outer two

sepals overtopping the bud. Planchon and Triana
also described Balboa in the same publication with
Tovomitopsis, but they placed it in their subtribe
“Euclusiae” rather than Tovomiteae. Bentham and
Hooker (1862: 167-176) recognized Balboa, Chry¬
sochlamys, and Tovomita (as related) but treated To¬
vomitopsis as a synonym of Chrysochlamys. How¬
ever, they transferred no species. Subsequently,
Balboa was apparently forgotten in discussions re¬
lating to the Tovomita group. Hemsley (1879: 87)
transferred all Central American species of Tovom¬
itopsis to Chrysochlamys. Cuatrecasas (1950) also
declared unjustified the separation between Chry¬
sochlamys and Tovomitopsis and transferred two
(South American) species to Chrysochlamys. Wil¬
liams (1959) subsumed Tovomitopsis within Tovom¬
ita, but kept Chrysochlamys separate. Maguire
(1977) tentatively recognized all three genera.
D'Arcy (1980) expressed the same taxonomic opin¬
ion begun by Bentham and Hooker and put into
practice by Hemsley and Cuatrecasas, i.e., a uni¬
fied Chrysochlamys/Tovomitopsis, but used the later
name, Tovomitopsis. D'Arcy also elevated Engler’s
section Dystovomita ol Tovomita to the genus Dys-
tovomita.

Most recently, an attempt has been made to cir¬
cumscribe Chrysochlamys as comprising species
with succulent petals, eehinate pollen exine, and
black seeds with white arils, rather than membra¬
naceous petals, psilate-foveolate pollen, anil green
seeds with orange arils (cf. Hammel, 1984, 1986).
Ignoring whether or not these characters support a
monophyletic group or even if they coincide, ty-
pologically, with Chrysochlamys, neither such a
separation, nor one based on the considerations of
Planchon and Triana is practical. Dystovomita, with
axillary inflorescences (i.e.. mostly lateral and be¬
low the leaves), large, adaxial petiolar pits and rel¬
atively small seeds, is easily recognized and may
even lie closer to Clusia (P. Stevens, in prep.). To¬
vomita, with terminal inflorescences, leaves clus¬
tered at the ends of side shoots, leaves sometimes
with petiolar pits, outer two sepals fused in bud or
at least overtopping the bud, styles often evident,
fruits usually with the inner wall and placenta dark

Novon 9: 360-374. 1999.

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

361

reddish wine, can also be recognized, even with
most herbarium material. The older name, Chry¬
sochlamys, takes precedence over Tovomitopsis for
the remaining species in the group, i.e., those with
terminal inflorescences, leaves not clustered on
side shoots, lacking adaxial petiolar pits, sepals
(usually) not overtopping the bud, styles lacking,
and the inner lruit wall and placenta mostly white
or pink. Balboa is a synonym of Chrysochlamys.

The above paragraph can be summarized by the
following key:

la. Inflorescences apparently axillary; leaves with
large, adaxial, margined petiolar pits; seeds ca.

6 mm . Dystovomita

lb. Inflorescences usually terminal; leaves mostly
lacking petiolar pits; seeds ca. I cm or more.

2a. Leaves clustered at the ends of side shoots;

outer two sepals fused in bud or at least
overtopping the bud; styles often evident; in¬
ner fruit wall and placenta usually dark red¬

dish wine . Tovomila

2b. Leaves ± evenly dispersed along the

branches; outer two sepals never fused in
bud. nor both overtopping the bud; inner
fruit wall and placenta mostly white or pink
. Chrysochlamys

Taxonomic Treatment

Chrysochlamys Poeppig, Nov. Gen. Sp. PI. 3: 13,
t. 211. 1842. TYPE: Chrysochlamys multiflora
Poeppig.

Balboa Planchon & Triana, Ann. Sci. Nat. Bot., ser. 4, 13:
315. 14: 252. I860. TYPE: Balboa membranacea
Planchon & Triana.

Tovomitopsis Planchon & Triana, Ann. Sci. Nat. Hot., s£r.
4. 14: 261. I860. TYPE: Tovomitopsis paniculata
(Sprengel) Planchon & Triana. Basionym: Bertolonia
paniculata Sprengel.

Dioecious, terrestrial trees or shrubs, sometimes
epipetric rheophytes, the branching decussate and
the leaves dispersed ± evenly along the branches;
resin mostly clear, occasionally white, rarely yel¬
low; internodes mostly without bud-scale scars. Pet¬
iolar pits and interpetiolar stipuliform structures
absent. Leaves opposite, the blades elliptic; major
lateral veins mostly prominent, resin canals some¬

times visible in dried material. Inflorescences ter¬
minal (rarely left pseudo-axillary by the dominant
growth of an axillary, vegetative bud), many-flow¬
ered, panicles. Flowers 4- or 5-merous, corolla aes¬
tivation imbricate, petals thin, membranaceous, to
thick and succulent, pink, or yellow to white. Sta-
minate flowers with numerous (rarely few) ± sep¬
arate stamens, or the filaments fused basally into a
short column, or rarely with fertile stamens sur¬
rounding a central sterile capitulum, or very rarely
the anthers all sessile on a resiniferous capitulum.
Pistillate flowers with numerous separate stamino-
dia, or staminodia mounted on a ring or rarely just
a ring surrounding the ovary; ovary with mostly 4
or 5 stigmas, rarely 3. Fruits globose to ovoid, usu¬
ally terete, rarely winged or star-shaped in cross
section, succulent capsules mostly with 1 seed ma¬
turing per locule, sometimes only 1 seed per fruit;
seeds green to nearly black with an orange or white
aril.

Distribution and ecology. Chrysochlamys is
known from Mexico to Bolivia and Brazil, where it
occurs in wet to very wet primary forest from near
sea level to at least 2300 m.

Chrysochlamys, in total, contains perhaps as
many as 50 species, most of them South American,
and is in much need of revision. At least two Cen¬
tral American species, C. glauca (Oersted, Plan¬
chon & Triana) Hemsley and C. nicaraguensis (Oer¬
sted, Planchon & Triana) Hemsley, have been
erroneously reported from South America, and nu¬
merous South American names (e.g., C. membran¬
acea Planchon & Triana, C. micrantha Engler, C.
myrcioides Planchon & Triana) have been loosely
(and erroneously) applied to Central American spe¬
cies, as well as within South America. The follow¬
ing key, distinguishing 13 species of Chrysochlamys
for Mexico and Central America, and the subse¬
quent annotated list, including citation of represen¬
tative specimens and descriptions of two new spe¬
cies from Costa Rica and Panama, aim to clarify
the species and names for the region. One or two
possible new species, from among Panamanian ma¬
terial at hand, await resolution and are not included
here.

Kky to the Species oe Chrysochiamys (and one Tovomita) in Mesoamerica

la. Internodes with a pair of bud scales (or their scars), usually just above the subtending node; primary lateral

veins close (ca. 3 mm apart), ± prominent but difficult to distinguish from the intersecondaries; inflorescences
often appearing axillary; anthers sessile on a capitulum; stigmas 3; far eastern Panama to Ecuador .

. C. tenuifolia

lb. Internodes without bud-scales; primary lateral veins distant (usually 1 cm or more), these and the interse¬
condaries obscure or prominent; inflorescences unambiguously terminal; anthers with filaments; stigmas (4)5
or 6; Guatemala to northern South America.

362

Novon

2a.

21).

I. eaves mostly < 2.5 cm wide, often drying tan: lateral veins very obscure in dry material.

da. I' lower buds 6—9 mm long; stamens 30-40; inflorescence and infructescence stout, fruits pyriform,

but not rostrate, mostly 5-seeded; epipetric rheophytes . C. angustifolia

3b. flower buds 3^4 mm long; stamens ca. 20; inflorescence and infructescence delicate, usually
nodding or pendulous; fruits narrowly obovoid or pyriform and usually rostrate, mostly 1- or 2(3)-

seeded; small shrubs in forest and sometimes along streams . C. tenuis sp. nov.

Leaves mostly over 3 cm wide, drying green, gray-green, reddish tan, to tan; lateral veins easily seen
in ( I ry material.

4a. Leaf blades nearly sessile, 20—60 cm long; small shrubs < 2 m . Tovomita croatii

4b. Leaf blades distinctly petiolate, shorter; shrubs to trees, often over 3 m.

5a. Lateral leaf veins mostly < 9 pairs; calyx deciduous in fruit.

6a. Flower buds < 4 mm long; stamens ea. 20—30; inflorescence delicate, often pendulous;
lateral leaf veins mostly ca. 1 cm apart; leaves usually dull, drying gray-green, often

with obvious resin canals; Costa Rica and Panama . C. glauca

6b. f lower buds > 5 mm long; stamens 35 or more; inflorescence stout, erect; lateral leaf
veins > 1 cm apart; leaves shiny.

7a. Leaves without obvious intersecondary veins or resin canals, the primary lateral
veins usually 2—3 cm apart, the leaves drying gray-green; flower buds ca. 5—7

mm long; stamens 35-40; Nicaragua, Costa Rica, Panama . C. nicaraguensis

7b. Leaves with obvious intersecondarv veins and often the resin canals also obvious,
the primary lateral veins mostly 1-2 cm apart, the leaves drying reddish tan;
flower buds ea. 5 mm long; stamens ca. 50; Mexico, Guatemala, Belize, Honduras
. C. guatemaltecana

5b. Lateral leaf veins mostly > 10 pairs; calyx persistent or not.

8a. Submarginal leaf vein usually distinct, at least above mid-leaf; calyx persistent in
fruit.

9a. Petals succulent; filaments fused basally, the free portion shorter than the fused
portion; staminodia connate, forming a thin ring around the base of the ovary;

fruits terete; central to eastern Panama . C. eclipes

9b. Petals membranaceous; filaments free to the base; staminodia distinct; Costa
Rica to central Panama.

10a. Fruits star-shaped in cross section; anthers usually 2—3 times longer than
wide; outer two sepals ± equal, shorter than the bud; flower buds ea. 7

mm long; inflorescence glabrous or pubescent . C. grandifolia

10b. fruits terete; anthers about as wide as long; outer sepals very unequal, the
larger ± equal to the bud; flower buds ca. 5 mm long; inflorescence gla¬
brous . C. skutchii sp. nov.

8b. Submarginal leaf vein lacking or indistinct; calyx deciduous or persistent.

1 la. Calyx deciduous in fruit; plants from the mountains (mostly > 1000 m) of Costa
Rica and western Panama; lateral leaf veins 9-15 distinct pairs; leaves drying
tan to pink or green.

12a. Fruits ± pyriform, ca. 2—2.5 cm long, often greenish yellow at maturity
and the stigmas set to one side; stamens > 50; relatively large (ca. 15 m)
trees with ± coriacious, tan- to pink-drying leaves; 700-2000 m . . C. allenii
12b. Fruits usually obovoid-mammillate, often 4 cm or more long, mostly red at
maturity, the stigmas centered; stamens mostly ca. 15—20; shrubs or small
trees (ca. 7 m or loss) with membranaceous, mostly green-drying leaves;

800-1500 m . C. psycho! rii folia

lib. Calyx persistent in fruit; plants from lowland Costa Rica (30—900 m) or from
central and eastern Panama; lateral leaf veins 10—17 pairs, sometimes nearly
indistinguishable from the intersecondaries; leaves usually dark, reddish brown
to tan-drying.

13a. fruits strongly alate; leaves narrowly elliptic, often falcate. ± coriaceous,
the apex gradually acuminate, the resin canals very' obvious on dry mate¬
rial; outer pair of sepals ± equal and nearly as long as the bud; resin
canals of petals obvious; resin of stems white; lateral leaf veins 10—13: far

eastern Panama . C. membrillensis

13b. fnnts not alate; leaves oblong to elliptic, not falcate, membranaceous, the
apex abruptly acute to acuminate, the resin canals obscure; outer pair of
sepals very unequal, the smaller much shorter than the bud; resin canals
of petals obscure or lacking; resin of stem clear or white.

14a. Stamens ca. 50; leaves broadly elliptic, mostly 8—10 cm wide; lateral
leaf veins mostly 11-17 distinct pairs; resin of stems clear; Costa

Rica . C. skuteliii

14b. Stamens 20—30; leaves elliptic to oblong, mostly 5-6(— 9) cm wide;
lateral leaf veins mostly 8— 10 somewhat indistinct pairs, often diffi¬
cult to distinguish from the intersecondaries; resin of stems white;
Costa Rica and Panama . C. silvicola

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

363

Chrysochlamys allenii (Maguire) Hammel, comb,
nov. Basionym: Tovomitopsis allenii Maguire,
Phytologia 38: 214. 1978. TYPE: Panama. Bo-
cas del Toro: Robalo Trail, northern slopes of
Cerro Horqueta, 2000-2300 m, Allen 5014
(holotype, NY; isotype, MO).

Tovomitopsis centistaminibus D’Arcy, Ann. Missouri Bot.
Card. 67: 1030. 1980. Syn. nov. TYPE: Panama.
Chiriquf: 8 km N of Camp El Volcan, 2000 m. Little
6070 (holotype, MO; isotype, US).

Trees 5—25 m, sometimes with stilt roots, occa¬
sionally reported as epiphytic; resin clear to
creamy. Leaf blades elliptic to slightly obovate,
widest just above the middle, 11—16 X 4-6 cm, the
apex rounded to acuminate, the base acuminate,
coriaceous; lateral veins ca. 0.8—1 cm apart, ca. 9—
12 pairs, reaching to the margin; petiole 1-2 cm.
Inflorescence 8-12 cm, erect, puberulent; flower
buds ca. 5—7 mm; stamens 50 or more, anthers of
both stamens and staminodia about as wide as long,
< 1 mm, filaments free; stigmas large (ca. 2 mm),
distinctly spreading. Fruits pyriform, 2-2.5 cm,
yellow-green, the stigmas often mounted slightly to
one side; calyx deciduous in fruit.

Distribution. Very wet cloud forest, (700—)
1450-2300 m; continuous from the Cordillera de
Guanacaste in northwestern Costa Rica to Chiriquf
province of western Panama and disjunct from Da¬
rien province in far eastern Panama.

A larger tree than any other Chrysochlamys in
Central America and growing at higher elevations,
C. allenii is further distinguished by its coriaceous
leaves, numerous stamens (more than 50), often
very large stigmas, and pyriform, often greenish
yellow fruits.

Although D’Arcy (1980) compared Tovomitopsis
centistaminibus with T. psychotriifolia, the type
specimen’s high-elevation habitat, large stature, co¬
riaceous leaves, anil numerous stamens conspire to
alienate it from the latter species, and instead, to
bring it into synonymy with C. allenii. This species,
however, can at times be confused with C. psycho¬
triifolia, which usually grows at somewhat lower el¬
evations, has fewer (and often longer) stamens, and
larger fruits. Material at hand, especially that from
far eastern Panama, suggests that C. allenii may be
closely related to, perhaps not distinct from, C. col-
ombiana Cuatrecasas.

Representative specimens. COSTA K1CA. Guana-
caste: Volcan Cacao, 1 100 m, Nov. 1990 (fr), Espinoza 24
(INB). Puntarenas: Monteverde Reserve, 1500 in. June
1985 (11, fr). Hammel 12856 (MO); cloud forest above Wil¬
son’s finca S of San Vito de Java, 1600 m, Aug. 1967 (fr),
Raven 21845 (MO). Heredia: 12 km above San Rafael,

1800 m, July 1967 (fl). Lent 1141 (MO). San Jose: La
Palma, 1550 m, Aug. 1898 (fl). Tonduz 7378 (MO). PAN¬
AMA. C hiriquf: 8 km N camp El Volcan, 2000 m. Mar.
1943 (fr). Little 6068 (MO); about 2 km NE of Guadelupe,
2 km N of Cerro Punta, Dec. 1971 (fr), Wilbur et al. 15331
(DUKE, MO); Bajo Mono— Robalo trail. 1600-2300 m.
July 1947 (fl), Allen 4839 (MO). Hoeas del Toro: vicinity
of Cerro Colorado, 1700 m, Apr. 1986 (fr), McPherson
8925 (MO); vicinity of Fortuna Dam, 1200 m, Apr. 1986
(fl), McPherson 9049 (MO). Darien: Cerro Sapo, 800—
10(H) m. Mar. 1986 (11, fr), Hammel et al. 14877 (MO).

Chrysochlamys angustifolia (Maguire) Hammel,
comb. nov. Basionym: Tovomitopsis angustifol-
ia Maguire, Phytologia 36: 399. 1977. TYPE:
Panama. Veraguas: Mori & Kallunki 5357 (ho¬
lotype, M0; isotype, NY).

Small shrubs to 2 m, often on rocks along
streams; resin milky. Leaf blades narrowly elliptic
7—13 X l-2(— 2.5) cm, the apex acute to acuminate,
base acute, membranaceous, drying light gray-
green or tannish green; lateral veins obscure, ca.
0.5—1 cm apart, ca. 5 or 6 pairs, reaching to the
margin; petiole ca. 0.5-1 cm. Inflorescence ca. 6
cm, erect, glabrous; flower buds 6-9 mm; stamens
30^40, anthers of stamens and staminodia ± tri¬
angular, slightly longer than wide, ca. 1 mm, fila¬
ments free; stigmas spreading to often erect. Fruits
pyriform, 1.5—2 cm, red or greenish red; calyx de¬
ciduous in fruit.

Distribution. Very wet, mostly lowland forest,
usually along streams, 100—1000 m; Bocas del Toro
to Veraguas and Code provinces of central Panama.

This distinctive, small, narrow-leaved, epipetric
rheophyte with flower buds over 6 mm long is en¬
demic to a few lowland or mid-elevation sites in
west-central Panama. Its habitat, larger flowers, and
stouter fruiting rachis easily distinguish it from the
similarly narrow-leaved new species, C. tenuis, de¬
scribed below. Among Central American species,
C. angustifolia is one of the few with copious milky
resin rather than the more common clear to tardily
creamy resin.

Representative specimens. PANAMA. Koras del
Toro: Upper Rfo San Pedro, 5(H) m, July 1979 (fr), Gor¬
don 59Da (MO). Veraguas: vicinity of Santa Fe, dirt road
past Escuela Alto de Piedra, 490 m, June 1987 (11, fr),
Croat 66871 (MO). Coele: along Rfo San Juan near fork
with Rfo Tife, 400 m. June 1978 (fl, fr), Hammel 3329
(MO).

Chrysochlamys eclipes L. O. Williams, Trop.
Woods 111: 15. 1959. TYPE: Panama. Canal
Zone: Barro Colorado Island, Standley 31459
(holotype, US).

Shrubs or trees 2—10 m; resin (dear. Leaf blades

364

Novon

elliptic to oblong, 12-26(-29) X 3—12 cm, the apex
acute, base acute, membranaceous, drying dark
reddish tan or rarely greenish tan; lateral veins ob¬
vious, 1—1.5 cm apart, 10—12 pairs, loop-connected
to a submarginal vein ca. 1—3 mm from the margin;
petiole ca. 1—3 cm. Inflorescence ca. 7-21 cm,
erect or pendulous, minutely puberulous; flower
buds ca. 4—5 mm; stamens ca. 25—30, anthers of
stamens and staminodia wider than long, < 1 mm,
filaments basally fused into a succulent torus, the
free part of filaments < 1 mm, much shorter than
the fused part; stigmas rounded, pressed against the
ovary. Fruits mostly globose, 2—3 cm, green to
greenish red or red; seeds black with a white aril;
calyx persistent in fruit.

Distribution. Wet to very wet forest, from near
sea level to 1600 m; central to eastern Panama.

A distinctive, usually dark-reddish-brown-dry-
ing, dull-leaved, Panamanian endemic (from the
Canal to San Bias and Darien) with succulent and
yellow rather than membranaceous white or pink
petals, stamens with short, lree filaments mounted
on a succulent, fused basal portion, white rather
than orange arils, and black rather than green
seeds, not at all closely related to C. nicaraguensis,
under which it was synonymized by D’Arcy (1980).
This is the only species, described from the region,
with succulent, rather than membranaceous, petals.
Even in fruit it can be identified by the combination
of large, dark-drying leaves and fused staminodia.
The petals dry black or very dark burgundy.

Some of this material with particularly large in¬
florescences from far eastern Panama has been
identified as C. floribunda Cuatrecasas and C. we-
berbaueri Engler. However, the abundant material
at hand suggests that probably only one species is
involved. Pending study of type material of the old¬
er, South American names, I here use C. eclipes for
all Panamanian collections.

Representative specimens. PANAMA. Colon: end of
Rfo Boquer6n road, 5(H)— 650 m, Apr. 1078 (fr). Hammel
2445 (MO); Rfo Ouanrhe, Feb. 1074 ((I). Roster 2780
(MO); Santa Rita Ridge Road, 450 m. Mar. 1975 (fr). Mori
A Kail un hi 5040 (MO), (.anal Zone: Harm Colorado Is¬
land. Mar. 1969 (lr). ('.root 8820 (MO); Pipeline Road.
Sep. 1071. (II). Gentry 1946 (MO). Panama: Cerro Jefe
region, 750 in, Jan. 1086 (fr). McPherson A Merello 8124
(MO); Corgas Memorial Labs yellow lever research camp,
25 km NE of Cerro Azul on Rfo Piedras. 550 in. Nov.
1071 (fr), Mon A Kallunki 3371 (M0); Llano— Carti road.
350 m. Aug. 1986 (II). McPherson 0002 (MO). Sail Bias:
Cangandi, 30 m, Feb. 1086 (lr), de Movers A Herrera 7180
(MO). Darien: Cana— Cuasi trail. Chepijana. 1600 m. Mar.
1940 (lr). Terry A- Terry 1002 (MO); Parque Nacional del
Darien, ridge between Rfo Topalisa and Rfo Pucuro, ca.
17 km E of Pucuro, 850 m, Oct. 1087 (II). de Movers el
(d. 8280 (MO); slopes ol Cerro Mali. 1300—1400 in, Oct.

1087 (II & lr). Cuadros et al. 2027 (MO); S of El Real on
trail up Cerro Pirrc. 550—1030 m. Mar. 1085 (fr). Mc¬
Pherson 7045 (MO): trail to Cerro Pirrc, 20-600 m. Aug.
1086 (fl). ,/. F. McDonagh et <tl. 482 (MO).

Chrysochlamys glauca (Oersted, Planehnn & Tri-
ana) Hemsley, Biol. Cent.-Amer., Bot. 1: 87.
1879. Tovomitopsis glauca Oersted, Planchon
& Triana, in Planchon & Triana, Ann. Sci. Nat.
Bot., ser. 4, 14: 264. 1860. Tovomita glauca
(Oersted, Planchon & Triana) L. 0. Williams,
Fieldiana Bot. 29: 357. 1961. TYPE; Costa
Rica. [San Jose]: sur le inont Jaris, [ca. 850
m]. Oersted 3 [3589] (holotype, C).

Tovomita gracilis L. O. W illiams, Fieldiana Bot. 20: 357.
1061. TYPE: Costa Rica. Alajuela: San Pedro de
San Ramon, 1050 m, Hrenes 4217 (lcctolypc, des¬
ignated by Hammel (1986), F).

Tovomitopsis multiflora Standley, Field Mus. Nat. Hist.
Rot. Ser. 4: 234. 1020. TYPE: Panama. Chiriquf:
Progreso. Cooper A Slater 171 (holotype. F-573,086;
isotype. NY).

Shrubs or rarely trees 1— 6(— 14?) m; resin- clear
to slightly milky. Leaf blades elliptic, 7—20 X 2—7
cm, the apex and the base acute to acuminate, dry¬
ing gray-green; lateral veins 0.5— 1.5 cm apart, 6—
10 pairs, fading toward the margin; petiole 1-2.5
cm. Inflorescence 2—12 cm, often pendulous, gla¬
brous; flower buds ca. 3^4 mm, outer 2 sepals
equal to markedly unequal, shorter than the inner
ones; stamens 20—30, anthers of stamens and sta¬
minodia ovoid to ± globose, < 1 mm. Fruits pyr¬
iform to globose, 1—2 cm, greenish cream to nearly
red. often two-toned; calyx deciduous in lruit.

Distribution. Wet to very wet forest, from near
sea level to 1400 in; Costa Rica and Panama.

Even segregating the material described below
as Chrysochlamys tenuis , C. glauca remains a rath¬
er variable species, characterized by its usually
dull, few-veined leaves and small flower buds.
Some of that variation may be worthy of taxonomic
recognition. Many ol the collections from the Osa
peninsula in Costa Rica (cf. Aguilar 431 , Kernan
713) have leaves with especially obvious and nu¬
merous latex canals and flower buds near the lower
end of the size range for the species, giving such
material an aspect predictive of its origin. Allen
(1956: 345), in fact, distinguished two species from
the Osa region within the C. glauca complex. For
one (vouehered by Skutch 5286, F, not seen), he
used the name C. costaricana, stating that it was
“nearlv identical” to C. glauca, “differing princi¬
pally in the smaller flowers.” Although C. costari¬
cana (discussed below under C. psychotriifolia ) is
not its correct name, the species distinguished by

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

365

Allen may deserve recognition. Populations at the
other end of Costa Rica, in the Cordillera de Guan-
acaste, have almost shiny leaves with relatively ob¬
scure lateral veins. The plants also appear to flower
more on lateral shoots than material from else¬
where. The anthers are slightly longer and the stig¬
mas are larger than those of C. glauca from else¬
where, usually at least Vi (vs. no more than V3-V4)
the length of the ovary in newly opened flowers.
Several collections from Panama, Panama province
and San Bias (de Nevers 7223, Herrera et al. 1137,
Herrera et al. 1145), tentatively placed here, have
unusually dense and much-branched resin lines
and fruits with persistent sepals. These may rep¬
resent a distinct species, but flowering material is
needed.

Representative specimens. COSTA RICA. Guana-
caste: Volcan Cacao, Rancho Harold. 700—1200 m, July
1986 (fr). /. & A. Chacon 2154 (CR); Parque Nacional
Guanacaste, Estacion Cacao. 1100 m, Nov. 1990 (fr), Cha¬
vez 368 (INB). Alajuela: Rfo Calaratas, San Ramon. 600
m. Jan. 1984 (fl), Gdmez-Laurito 9700 (CR). Union: Can¬
ton de Pococf, Barra del Colorado. 15—20 m. Jan. 1995
(H). Araya 713 (CR. INB. MO). Puntarenas: airport 4 mi.
W ol Rincon tie Osa, 40 m, Aug. 1967 (fr). Raven 21564
(CR); wooded slopes above Golfito, 0—80 m. Jan. 1967 (fl).
Burger & Mata 4775 (CR. F); Las Cruces. Jardin Botanico
Wilson. 1 1(X) m. Sep. 1993 (fr), G. Vargas et al. 1369
(CR). PANAMA. Boeas del Toro: Rfo Teribe, alrede-
dores de la estacion Teribe II del 1RHE, Oct. 1984. Car-
rasquilla 2026 (MO). Chiriquf: E of Canas Gordas near
Costa Rican border on road to Volcan. 8(H) — 1200 m, Feb.
1973 (fl. fr), Liesner 268 (MO); road from Nueva California
to Rfo Serano, 1400 m, Apr. 1979 (fl), D'Arcy et al. 13057
(MO). Panama: Llano— Cartf road. 400 m. jan. 1974 ((I).
Nee & Dressier 9371 (MO). San Bias: Cordillera tit' San
Bias, cabecera del Rio Piriati. 460—799 m, Feb. 1992 (fr).
Herrera et al. 1137 (MO).

Authorship of the four Central American Tovom-
itopsis names published in Planchon and Triana
(1860) has been variously obfuscated in numerous
publications, including my own. Oersted alone nev¬
er published these names, as sometimes implied
(cf. Williams, 1959), nor is there any basis for the
even more common “Oersted ex Planchon & Tri¬
ana” (cf. Standley, 1937; Croat, 1978; D’Arcy,
1980), let alone the terse permutation of the latter
to just “Planchon & Triana” (e.g., Hammel, 1986).
Since Anders Orsted (1812—1872) was professor of
botany at the University of Copenhagen until 1862
(http://www.nathimus.ku.dk/bot/orsted.htm) during
the time Planchon and Triana were working in Paris
on their Guttiferae memoir, nothing suggests other
than the three, “Oersted, Planchon & Triana,” as
co-authors, as originally indicated for all four spe¬
cies in Planchon and Triana (1860).

The citation of two collection (?) numbers for the
types of these same four species results from ex¬

amination of the actual specimens. The single-digit
numbers are cited in the original publication (e.g..
Oersted 4), and are also written on or in packets
with the specimens. The four-digit numbers printed
on the labels (e.g., Plantae Centroamericanae Oer¬
sted No. 3585) are not mentioned in the publica¬
tion. Neither series appears to come from normal
consecutive field-note numbers; earlier dates do not
correspond, consistently, to lower numbers within
the series.

Chrysochlamys grandifolia (L. 0. Williams)
Hammel, comb. nov. Basionym: Tovomita
grandifolia L. O. Williams, Trop. Woods 112:
96. I960, new name for T. macrophylla L. O.
Williams, Trop. Woods 111: 16. 1959, non T.
macrophylla (Poeppig) Walpers, 1842. TYPE:
Costa Rica. Puntarenas: Esquinas forest, be¬
tween Rfo Esquinas and Palmar Sur, 30 m,
Allen 5968 (holotype, US; isotypes, EAP, F).

Shrubs or trees 2—10 m; resin clear to yellowish.
Leaf blades elliptic, 18—35 X 10—18 cm, the apex
acute to rarely rounded, the base acute to acumi¬
nate, drying reddish; lateral veins 1-2 cm apart,
14—18 pairs, looping to form a ± distinct submar¬
ginal vein; petiole (2— )3-4 cm. Inflorescence ea. 6—
10 cm, erect, glabrous; flower buds ca. 7 mm, outer
2 sepals ± equal to each other and shorter than
the inner ones; stamens ca. 50, anthers of stamens
and staminodia longer than wide, ca. 2 mm. Fruits
ovoid, strongly 5- or 6-ribbed (star-shaped in cross
section), greenish cream to nearly red, often two-
colored; calyx persistent in fruit.

Distribution. Very wet forest, 10— 600(— 1300)
m; Costa Rica to central Panama.

This distinctive Central American species, clear¬
ly delimited by its protologue, is characterized by
strongly ribbed fruits and monomorphic stamens
with anthers up to 2-3 times longer than wide. Cos¬
ta Rican material from the Pacific slope has gla¬
brous inflorescences and anthers nearly 2 mm long,
whereas that of the Caribbean slope (also occurring
at higher elevation than that of the Pacific) and the
Panamanian collections have markedly puberulent
inflorescences and smaller anthers.

This species was originally published as Tovom¬
ita macrophylla L. O. Williams, a homonym of the
earlier T. macrophylla (Poeppig) Walpers; Williams,
therefore, made a new name, T. grandifolia, for the
same species. Coincidentally, an earlier Chrysoch¬
lamys macrophylla Pax would prevent the use of
that combination for the present species, in any
case. Nothing, however, prevents the use of the

366

Novon

combination C. grandifolia (L. 0. Williams) Ham-
mel.

Although D'Arcy (1980) created the combination
Tovomitopsis myreioides (Planchon & Triana)
D’Arcy for this material (with Tovomita grandifolia
as a synonym), the South American C. myreioides
Planchon & Triana is a quite different and poorly
known species. It has very short petioles anil the
dimorphic stamens of section Heterandra “externa
fertilia, libera, interna sterilia in corpus centrale
concreta” (Planchon & Triana, 1860, vol. 14: 260)
with nearly globose anthers. Furthermore, I know
of no such rib-fruited Chrysochlamys species from
South America. Chrysochlamys membrillensis (see
below) of far eastern Panama also has somewhat
ribbed fruits, but is otherwise quite different from
the present species.

Representative specimens. COSTA RICA. Cartago:
Tuirialba, 2 km al oeste tie Grano de Oro, 1200 m, July
1995 (H). Cascante 6 39 ((41). Limon: Cord. Talamanea,
eutre Cerro Cbimii y Cerro Matama. 1200 m, Apr. 1985
(11. fr), Gomez it Herrera 23536 (CR); fila entre cuenca
superior del Rib Xichiari y cuenca superior del Rio Royei.
1300 m. Aug. 1995 (fl), Herrera 3427 (CR). San Jose:
ea. 20 km S of San Isidro de Fl General, El Pilar de
Cajon, 620 m, Nov. 1988 (lr), Zamora et al. 1521 (CR.
MO). Piinturcnus: Canton de Osa, Fila al Oeste de Ran¬
cho Ouemado. 800 m, Aug. 1991 (fr), ./. Marin 127 (CR,
MO); Sierpe. eabeeeras de Quebrada Guerra, 200 m, June
1990 (fl), Herrera 4240 (CR. 1NB. MO); Corcovado Na¬
tional Park, trail from Los Chiles to base of hills, 20—100
m, July 1977 (11), Liesner 3175 (CR, MO); Los Planes (La
Gloria). I(M) m, Feb. 1991 (fr), R. Gonzalez II (CR, F,
INR. MO); Canton de Golfito, Refugio Nacional Golfito, S
end of Fila Gamha, 160—200 m. Dee. 1988 (fr), Grayum
et al. 0130 (CR, MO). PANAMA. Itoeas del Toro: road
from Fortuna Dam to Chiriquf Grande, 400 m, Apr. 1986
(11), McPherson 0020 (MO). Cliirupu': Camino haeia la
finea Landau, NF del campamento de Fortuna (Hornito),
silio de presa, 1 KM) m, Aug. 1976 (fr), Correa et al. 2330
(MO). Veraguas: 5 mi. W of Santa Fe on road past Fs-
euela Agrfcola Alto Piedra, 8(M)-I200 m. Mar. 1978 (fr).
Groat 23151 (MO). Code: Fl CopG Atlantic slope F of
sawmill, 'MM) in, Feb. 1979 (fr). Ilammel 6271 (MO). Co¬
lon: Santa Rita Ridge Road. <>(M) m. Feb. 1980 (fl). An¬
tonio 3756 (MO). Panama: between peaks on Cerro Trin¬
idad. May 1968 (II), Kirkbride & Duke 1647 (MO); Cerro
Campana, near tower. Mar. 1977 (fr), Folsom et al. 2311
(MO).

Chrysochlamys guateinaltecana Donnell-Smith,
Bot. Gaz. 13: 26. 1888. Tovomitopsis guate-
rnaltecana (Donnell-Smith) Standley & Stey-
ermark, Publ. Field Mils. Nat. Hist., Bot. Ser.
23: 66. 1944. T\ PE: Guatemala. Alta Verapaz:
von Tuerckheim 989 (holotype, US; isotype,
NY).

Trees or shrubs 3.5—15 m; resin (dear to pinkish.
Leaf blades elliptic, 12—25 X 5— 9(— 10) cm, the

apex acute, the base ± acuminate; drying tan to
grayish tan and usually with rather obvious resin
lines; lateral veins 1—2 cm apart, ca. 8—12 pairs,
fading to the margin; petiole 1.5-2 cm. Inflores¬
cence 7—15 cm, erect, glabrous; flower buds ca. 5
mm. the outer 2 sepals ± equal, < Vi the length
of the bud; stamens ca. 50, anthers of stamens and
staminodia about as long as wide, < 0.5 mm. Fruits
pyriform, 2.5 -4 cm, red tinged with green; calyx
deciduous in fruit.

Distribution. Wet forest; 100—1450 m. Mexico
to Belize.

This name has been considered a synonym of
Chrysochlamys nicaraguensis (cf. Standley & Wil¬
liams, 1961) and C. psychotriifolia (cf. Hammel,
1984). However, a clearer picture of variation with¬
in the genus throughout the region now supports its
reinstatement. The species is characterized by more
or less shiny, usually reddish-brown-drying leaves
with rather few (6—9) and distant (1—2 cm) major
lateral veins and usually quite prominent interse¬
condaries and obvious resin lines. The rather large
flower buds (5—7 mm) have visible resin lines. It
has more stamens (ca. 50) than either C. nicara¬
guensis (35—40) or C. psychotriifolia (ca. 20). The
fruits are often quite elongate, up to 4 cm, as in C.
psychotriifolia , and the sepals are early deciduous.

Representative specimens. MEXICO. Oaxaca: 18 km
al este de Cuahutemoc, por la carretera Cuahtemoc— Sar-
abia. Dec. 1974 (fr), Vazquez et al. 1432 (MO); Mpio. Ma¬
fias Romero, 8.3 km al S de Esmeralda, luego 2.3 km al
F, 1 10 m. May 1981 (fl, fr), Wendt et al. 3267 (MO); Mpio.
Sta. Marfa Chimalpa, Rio Verde, 360-430 m, July 1986
(fl), Galetti et id. 210 (MO); Mpio. Minatillan. lomas al S
del Poblado II. ca. 30 km al E del Campamento La La¬
guna, 180 m, July 1980 (fl. fr), Wendt et al. 2642 (MO).
GUATEMALA. Alta Verapaz: 7 mi. up road to Oxec, 7(X)
ill, July 1977 (fl), Croat 41631 (MO). Izabal: Puerto Men¬
dez. km 7, Sep. 1970 (fl), Contreras 10237 (MO). BELIZE.
Toledo: Central camp, Edwards Road beyond Columbia,
June 1951 (fl). Gentle 7354 (MO); SW Maya Mountains,
Columbia River Forest Reserve, 750 m, Apr. 1 *>92 (fl),
Holst 4407 (MO); vicinity of San Jos£ Mayan Indian vil¬
lage. June 1973 (11. fr), Croat 24450 (MO). HONDURAS.
Copan: Parque Nacional Cerro Azul, 1440 m, Feb. 1992
(lr). Hawkins & Mejia 134 (MO). Olanclio: Refugio de
Vida Silvestre La Muralla. 14(M) m, June 1992 (fl), Mejia
164 (MO).

Chrysochlamys menihrillensis (D'Arcy) Ham¬
mel, comb. nov. Basionym: Tovom itopsis mem -
brillensis D'Arcy, Ann. Missouri Bot. Gard. 67:
1036. 1980. TYPE: Panama. Darien: Upper
Rfo Membrillo, 100—800 m, Duke 10872 (ho¬
lotype, MO).

Shrubs or trees to ca. 8 m; resin clear. Leaf
blades narrowly elliptic, 16—25 X 4—6 cm, the apex

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

367

acuminate, the base acute and slightly decurrent,
drying reddish to grayish tan and with numerous,
closely spaced, often-branched, resin lines; lateral
veins 1-2.5 cm apart, ca. 10 pairs, fading toward
the margin and very indistinctly loop-connected;
petiole 1—3.5 cm. Inflorescence ca. 10 cm, erect,
glabrous; flower buds ca. 5-6 mm, the outer 2 se¬
pals ± equal and as long as or shorter than the
others; stamens ca. 40, anthers of stamens and sta-
minodia slightly longer than wide, ca. 1 mm. Fruits
ovoid-pyriform, ca. 3 cm, strongly alate but the
wings thin, pink; calyx ± persistent in fruit.

Distribution. Very wet primary forest, 100-850
m. Known for sure from just five collections, only
lrom Panama, low to middle elevations of Colon
province and extreme eastern Darien province. This
species must certainly also occur in nearby Colom¬
bia, but no such specimens have been seen.

This Panamanian endemic has narrowly elliptic
and often falcate, coriaceous leaves, with very vis¬
ible resin canals. The lateral leaf veins are without
prominent intersecondaries, and the fruits are ba¬
sically pyriform and winged. Although the winged
fruits are very characteristic of this species, they
are not obvious in young pistillate material or even
more mature, dried material.

Specimens examined. PANAMA. Colon: Santa Rita
Ridge Road, 6(X)— 7(X) m, Apr. 1981 (fl). Sytsma et al.
4176 (MO); 250-400 m. May 1982 (fl). Knapp & Schmal-
zel 5264 (MO). Darien: Parque Nacional del Darien, be¬
tween Rio Topalisa y Rio Pucuro, 750-850 m, Oct. 1987
(fl. fr), Hammel et al. 16282 (MO); Upper Rio Membrillo.
100-800 m. Apr. 1967 (fr). Duke 10863 (MO), Apr. 1967
(fr). Duke 10901 (MO).

Chrysochlamys nicaraguensis (Oersted, Plan-
chon & Triana) Hemsley, Biol. Cent.-Amer.,
Bot. 1; 87. 1879. Tovomitopsis nicaraguensis
Oersted, Planchon & Triana, in Planchon &
Triana, Ann. Sci. Nat. Bot., ser. 4. 14: 266.
1860. Tovomita nicaraguensis (Oersted, Plan¬
chon & Triana) L. 0. Williams, Trop. Woods
111: 18. 1959. TYPE: Nicaragua. [Rfo San
Juan]: along the Rfo San Juan, fca. 10-50 m].
Oersted 2 [3587] (holotype, C).

Chrysochlamys standleyana L. O. Williams, Trop. Woods
111: 16. 1959, new name for C. pauciflora Standley.
Ceiba 3: 214. 1953, non C. pauciflora Steyermark,
1952. Tovomitopsis standleyana (L. O. Williams)
D’Arcy, Ann. Missouri Bot. Card. 67: 1040. 1980.
TYPE: Panama. Boeas del Toro: Chiriqui Lagoon,
non Wedel 1019 (holotype, E; isotype, MO).

Shrubs or trees, 3-8 m; resin clear. Leaf blades
elliptic to obovate-oblong, 14-25(-27) X 6-10(-
12.5) cm, the apex acute to rounded and apiculate.

the base acute to acuminate, drying gray or tannish
green, ± glossy; lateral veins 1.5-3 cm apart, (5-)
6— 8( — 10) pairs, fading to the margin; petiole 1.5—4
cm. Inflorescence 3—19 cm, usually erect; flower
buds 5—7 mm, outer pair of sepals ± equal to each
other and shorter than the inner; stamens 35-40,
anthers of stamens and staminodia about as wide
as long, < 1 mm. Fruits obovoid to globose, 1.5-3
cm, greenish pink to red; calyx deciduous in fruit.

Distribution. Very wet forest of the Caribbean
lowlands, from near sea level to 600(-1200) m;
Nicaragua, Costa Rica, Panama.

Chrysochlamys nicaraguensis, often erroneously
interpreted as a wide-ranging species (cf. discus¬
sion under C. eclipes), is still only known for certain
from the very wet Caribbean lowlands of Nicaragua,
Costa Rica, and adjacent Panama. It is distinctive
for its very glossy and distantly veined (2—3 cm)
leaves almost totally lacking visible intersecondar¬
ies and resin canals. For discussion of authorship
and type specimen citation of this species, see un¬
der C. glauca.

Representative specimens. NICARAGUA. Rio San
Juan: Campamento La Lupe, ca. 15.5 km al Noreste del
poblado El Castillo. 100 m, Dec. 1991 (fr), Grijalva 5633
(CR). COSTA RICA. Alajuela: Canton de San Ramon,
camino al Volcdn Muerto, 1 1 (X) — 1 200 m. May 1993 (fl).
h. Araya et al. 294 (CR, INB. MO). Heredia: Canton de
Sarapiquf, La Selva Research Station, 1 (X) m, Feb. 1981
(fr), Folsom 8775 (CR. DUKE). Limbti: Cerro Coronel,
60— 100 m. Mar. 1987 (fl), Stevens et al. 24851 (CR. MO);
Parque Nacional Tortuguero, ca. 3 km al Sur del pueblo,
3 m, Nov. 1988 (fr), Robles 2175 (CR, MO); Canton de
Talamanca, Aguas arriba de Rfo Banano, Fila Asuncion,
4(H) m. Feb. 1995 (fr), Rodrigue z 580 (CR, INB. MO); bills
between headwaters of Quebrada Mata de Lirnon and up¬
per branches of Quebrada Tigre (Sixaola region), 25—30
m. Nov. 1984 (fr). Grayum et al. 4468 (CR. MO). PAN¬
AMA. Boeas del Toro: Fish Creek Mts., vicinity of Chi¬
riqui Lagoon, Apr. 1941 (fl). von Wedel 2300 (MO); Pen¬
insula Valiente, on E side of Lagoon of Chiriquf, 25 m,
Aug. 1987 (fl), McPherson 11453 (MO). Veraguas: valley
of Rfo Dos Boeas along road between Escuela Agricola
Alto Piedra and Calovebora, 450—550 m, 31 Aug. (fl),
Croat 27581 (MO). Code: Alto Calvario, N of El Cope,
700—900 m. May 1977 (fr), Folsom 3274 (MO); region N
of El Valle, 1000 m. Sep. 1946 (fl), Allen 3683 (MO).
Panama: Cerro Campana, 800 m, Aug. 1982 (fl), Ham¬
ilton & D'Arcy 662 (MO); Finca El Indio at headwaters of
Rfo Indio, slopes of Cerro Jefe, 750 m, Nov. 1979 (fl),
Antonio 2425 (MO). Colon: ca. 2—3 m up Rfo Guanche,
10-20 m, Jan. 1973 (fr), Kennedy A- Foster 2193 (MO).
San Bias: Cerro Habu, 500—850 m. Dec. 1980 (fr), Syts¬
ma et al. 2676 (MO).

Chrysochlamys psychotriifolia (Oersted, Plan¬
chon & Triana) Hemsley, Biol. Cent.-Amer.,
Bot. 1: 87. 1879. Tovomitopsis psychotriifolia
Oersted, Planchon & Triana, in Planchon &

368

Novon

Triana, Ann. Sci. Nat. Bot., ser. 4. 14: 263.
1860. TYPE: Costa Rica. [Cartago]: Turrialha,
[ca. 650 m], Oersted 4 [3585] (holotype, C).

Tovomitopsis eostaricnna Oersted. Planchon & Triana. in
Planohon & Triana, Ann. Sci. Nat. Bot., s£r. 4. 14:
265. I860. Chrysochlamys costaricana (Oersted,
Planchon & Triana) Hemsley. Biol. Cent.-Amer.. Hot .
1: 87. 1879. T\ PE: Costa Rica. [Cartago]: Turrialha
[ca. 650 m]. Oersted I [8588\ (holotype, C).
Tovomitopsis faucis D'Arcy, Ann. Missouri Bot. Card. 67:
1031. 1080. Syn. nov. TYPE: Panama. Chiriqui: for¬
ests around El Boquete. 1000—1300 m, Pittier 3039
(holotype, US).

Shrubs or trees, 3—10 in; resin c lear or slightly
creamy. Leal blades oblong-elliptic, 12-21 X 3—
8(-9) cm, the apex acute, the base acute to acu¬
minate, dull, drying gray-green; lateral veins mostly
1( — 1.5) cm apart, 10-15 distinct pairs, fading to
the margin or somewhat loop-connected toward the
leal apex; petiole 1-2 cm. Inflorescence 6-16 cm,
erect, minutely pubescent or glabrous; flower buds
ca. 5—7 mm, outer 2 sepals ± equal to each other
and shorter than the inner; stamens 15—20, anthers
of stamens and staminodia about as long as wide,
< 1 mm; stigmas ± reflexed. Fruits oblong and
often apiculate, 3—5.5 cm, greenish yellow-pink to
red; calyx deciduous in fruit.

Distribution. Very wet montane lorest, 600—
1500 m; Costa Rica and Panama.

Chrysochlamys psyehotriifolia, well known in
Costa Rica, is a species with numerous, rather
close, lateral veins, lew (ca. 20) monomorphic sta¬
mens, and usually large (3—5 cm), obovate-mam-
millate Iruits. Tovomitopsis costaricana, T. faucis,
and what D'Arcy (1980) called T. membranacea
(Planchon & Triana) D'Arcy were based on very
scant material that is, nevertheless, now seen to tall
easily within the variation of C. psyehotriifolia. For
discussion of authorship and type specimen citation
of this species, see under C. glauca. The South
American Chrysochlamys membranacea Planchon
& Triana, by its type and numerous more recent
collections, is clearly a different species, belonging
to Planchon and Triana’s section Heterandra, with
dimorphic stamens in the staminate flowers and the
staminodia connate into a ring or collar in the pis¬
tillate ones.

Representative specimens. COSTA RICA. Guana-
caste: Parque Naeional Rincon dc La Vieja, 950-1 100
m. Nov. 1987 (fr). Herrera 1446 ((41); Canton de Tilaran,
1—2 km W of Lago Cote, 800—900 m, Aug. 1993 (fl). Il<i-
ber & /uchowski 11626 (CR, MO). Puntarenas: Reserva
Monteverde, Vert. Pacific, 1520—1580 in. Sep. 1976 (fl).
Dryer 757 (CR). Alajuela: I .a Pal ma de San Ramon. 1250
m. Nov. 1925 (fl). Hrenes 4600 (385) ((41)- Heredia: llor-
quetas de Sarapiquf, Reserva Rara Avis, 700 m, July 1995

(fr). Marten 1008 (CR). San Jose: Braulio Carrillo Na-
tional Park, vicinity of Estacion Carrillo, 600 m. July 1985
(fr), Hammel & Trainer 14276 ((41. MO). Cartago: Can¬
ton de Turrialha, Valle del Reventazon. Grano de Oro.

1 100 m. Sep. 1993 (fl), Campos & Campos 185 (CR, INB.
MO). PANAMA. Boras del Toro: road to Chiriqui
Grande N ol Fortuna Dam, 650—700 m, June 1987 (fl),
McPherson 1 1 180 (MO). Chiriqui: in forest along road to
I,a Fortuna, 1400 in, Feb. 1979 (fl), Hammel 6282 (MO);
ridge road at Cerro Colorado, 1450—1750 m, Aug. 1977
(fl, fr), Folsom et al. 4882 (MO): Veraguas: above Santa
Fe beyond Escuela Agricola lnteramericana, Apr. 1976
(fr). Croat 84216 (MO).

Chrysochlamys silvicola (Hammel) Hammel,
comb. nov. Basionym: Tovomitopsis silvicola
Hammel, Selbyana 9: 113. 1986. TYPE: Costa
Rica. Heredia: La Selva Biological Station [ca.
100 m], Hammel 9404 (holotype, DUKE; iso¬
types, CR, F, MO).

Shrubs or trees 3—6 m; resin white. Leal blades
oblong-elliptic (13— )16— 22 X 4— 7(— 8) cm, the apex
acute to acuminate, the base acute, both surfaces
dull, drying reddish tan; major lateral veins 0.5—2
cm apart, ca. 10—16 pairs, difficult to distinguish
from the intersecondary veins, fading to the margin
where indistinctly loop-connected; petiole 1-2.5
cm. Inflorescence 5-10 cm, ± pendent, minutely
pubescent or glabrous; flower buds ca. 4—5 mm,
outer 2 sepals often markedly unequal with the
larger one ± overtopping the bud; stamens 20—30,
anthers ol stamens and staminodia about as long as
wide, < 1 mm; stigmas ± reflexed. Fruits ovoid to
globose, 1.5— 2.5 cm, red throughout or whitish at
base; calyx ± persistent in lruit.

Distribution. Very wet, primary lowland lorest,
Caribbean slope, from near sea level to 700(-900)
m; Costa Rica and Panama.

This species, particularly as to its type, can be
characterized by its primary lateral veins and in¬
tersecondaries being almost equally salient. Nev¬
ertheless, a few collections from Costa Rica includ¬
ed here (Heredia, Canton de Sarapiquf, Rara Avis,
400-700 m. Marten 948 (CR), G. Vargas 1650
(CR), O. Vargas 2 (CR), 0. Vargas 4 (CR)) have
fewer and less prominent intersecondaries. Material
from Panama appears to grade imperceptibly into
an extreme with the intersecondaries very obscure.
All the material has dull, yellowish to reddish-tan¬
drying leaves, the outer pair ol sepals often very
unequal, the petals usually without obvious resin
canals, and stamens about 20—30. At least in Costa
Rica, the resin of the twigs is bright white, rather
than clear or tardily creamy as in most other spe¬
cies in Central America. For its reddish-tan-drying
leaves, this species sometimes resembles the more

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

369

northern C. guatemaltecana, which has more sta¬
mens and shiny leaves with fewer, more distant
veins.

Representative specimens. COSTA RICA. Alajuela:
Canton de San Carlos, (near) Boca Tapada, Finca La Mon-
alisa, 100 m. May 1995 (fl), Rodriguez A Estrada 702
(CR. INB. MO). Heredia: between Rio Peje & Rio Sar-
dinalito, Caribbean slope of Volcan Barva, 700—750 m,
Apr. 1986 (fr), Grayum 6695 (MO); Finca La Selva, Puerto
Viejo de Sarapiquf (ca. 100 m), July 1979 (fl, fr), Grayum
1902 (CR. DUKE). San Jose: Parque Nacional Braulio
Carrillo. Carillo, 6(X) m. May 1982 (fl), Gdmez-Laurito
8440 (CR). Limon: Ref. Nac. Barra del Colorado, forests
and pastures between Chirripocito and Rfo Sardina, 12 m,
Apr. 1990 (fr), Grayum 9794 (CR, F, MO); La Aurora,
Guacima. ca. 5 km al E del pueblo, 100 m. Aug. 1988
(11), Robles et al. 2087 (CR, F, MO); Canton de Talamanca,
Alto Uren. 7(X)-9(X) m. Sep. 1994 (fr). Gallardo A l^epi:
813 (CR. F. INB, MO). PANAMA. Veraguas: NW of San¬
ta Fe, 1 km from Escuela Agrfcola Alto de Piedra, Feb.
1975 (fl), Mori A Kallunki 4794 (MO), 900-1200 m. May
1981 (fr), Sytsma A Andersson 4640 (MO). Code: El
Cope. 1200 m. June 1979 (fr), Antonio 1189 (MO); N of
Penonome on road to Coclesito, 100—170 m, Eeb. 1978
(fl). Hammel 1681 (MO). Panama: between Cerro Azul
and Cerro Jefe. July 1969 (fl), Foster 1157 (MO); El Lla¬
no— Cartf Road, 400—450 m, Jan. 1974 (fl). Nee A Dressier
9352 (MO), 350-400 m, Feb. 1973 (fr), Kennedy 2543
(MO). San Bias: R fo Diablo y vecindad de Duque Sui,
80-110 m. July 1992 (fl). Herrera et al. 1174 (MO). Da¬
rien: vicinity of upper gold mining camp of Tyler Kit-
tredge on headwaters of Rio Tuquesa, Aug. 1974 (fr),
Croat 27209 (MO).

Chrysochlamys skutchii Hammel, sp. nov. TYPE:
Costa Rica. San Jose: basin of El General,
675-900 m. Mar. 1940 (fl), Alexander Skutch
4759 (holotype, CR; isotype, MO). Figure 1.

Altitudo altior C. allenii affinis a qua differt foliis ma-
joribus, tenuioribus, par extimus sepala inaequalia, calyx
ad maturitatem fruetus persistens et inflorescentia glabra.

Trees or shrubs 2-10 m; resin clear to creamy
(red?). Leaf blades membranaceous, elliptic, (17-)
21-35 X 6—12 cm, the apex and base acute to
acuminate; lateral veins ca. 1-2 cm apart, ca. 10-
17 pairs, mostly reaching to the margin or the api¬
cal veins looping into a ± distinct, submarginal
vein; petiole 2^4 cm. Inflorescence 7-15 cm, erect,
glabrous; flower buds ca. 5 mm, outer pair of sepals
very unequal, one < Vi the length of the bud, the
other ± equal to the bud; petals membranaceous,
lacking distinctly contrasting resin canals; stamens
ca. 50, varying in length centripetally, 1-4 mm, the
anthers nearly as wide as long, ca. 0.3 X 0.4 mm;
stigmas sessile, spreading or slightly elevated.
Fruits ± pyriform or obovate, 2.5^4 cm, pale green-
pink to red; seeds ca. 1 cm, the aril orange; calyx
± persistent in fruit.

Distribution. This species is known only from

a few collections from relatively low 100— 750(—
900) m elevations in wet forest on the Pacific slope
of southern (and eastern) Costa Rica, from the re¬
gions of Acosta (Fila Bustamante), Valle del Gen¬
eral, the Osa Peninsula and vicinity.

Flowering or sterile collections of this species
may be difficult to separate from Chrysochlamys
grandifolia, a species most readily distinguished by
its strongly ribbed fruits. The two grow sympatri-
cally and have been confused on Costa Rica’s Osa
Peninsula. However, the leaves of C. grandifolia
usually have a more distinct submarginal vein, it
has fewer stamens, and the anthers are often 2—3
times longer than wide rather than about as long as
wide. In many details, e.g., fruit shape and color,
number and size of stamens, and leaf coloration, C.
skutchii comes much closer to C. allenii. Chrysoch¬
lamys skutchii is distinguished from C. allenii by
its larger, thinner leaves, glabrous inflorescence,
unequal outer sepal pair, persistent sepals in fruit,
and lower-elevation habitat. The epithet is in honor
of the first collector of the species, one of Costa
Rica’s premier natural historians, Alexander
Skutch.

Paratypes. COSTA RICA. Puntarenas: Canton de
Golfito, La Gamba, “Bosque de los Austriacos,” 300 m.
Dee. 1993 (fl), Huber A Weissenhofer 68 (MO), (fl), Huber
A Weissenhofer 90 (CR). (st.). Huber A Weissenhofer 94
(CR); Canton de Osa, Parque Nat ional Corcovado, Cerro
Brujo, 600 m. Jan. 1991 (fl). Castro 268 (MO); Cerro Rin¬
con. 745 in, Aug. 1990 (fr). Morales et al. 20 (CR. MO);
Cerro Mueller. 150—650 in. May 1988 (fr). Kernan A Phil¬
lips 474 (CR, MO): Estacion Los Patos, 250 m. May 1995
(fr), Moraga 191 (CR, M0), 500 m, July 1988 (fr), kernan
A Phillips 662 (CR); Los Planes, La Gloria, 170 m. Feb.
1991 (fr), Induni 266 (CR. INB. MO), 100 ni, Feb. 1991
(fr), R. Gonzalez 32 (CR); Rincon. Fila Casa Loma, 50-
400 m. July 1990 (fr). Chacon 941 (CR. MO); Camino de
A It ura, 100-300 m. Aug. 1967 (fr). Raven 21512 (MO);
Aguabuena, 3.5 km W of Rinc6n, 450 m, Nov. 1992 (fl),
Thomsen 188 (fill); Rancho Quemado, 500 m, Nov. 1991
(fl). Marin et al. 256 (MO). 300-400 in. May 1988 (fr).
Hammel et al. 16898 (CR. MO), 200-350 m. Apr. 1993
(fr). Aguilar 1771 (CR, INB. MO).

Chrysochlamys temiifolia Cuatrecasas, Revista
Acad. Colomb. Ci. Exact. 29: 59. 1950. TYPE:
Ecuador. Rfo San Miguel o Sucumbfos, 250 m,
Cuatrecasas 10880 (holotype, F; isotype,
COL).

Shrubs or small trees to 2.5—7 m, most twigs with
[laired bud-scale scars just above the node; resin
yellow to orange. Leaf blades elliptic, 8— 14(— 24) X
3-4.5(-9) cm, the apex acuminate, the base acute
to obtuse, drying gray-green; major lateral veins ca.
3(— 20) mm apart, ca. ( 12— )16 prominent pairs but
with several nearly equally prominent intersecon-

370

Novon

Figure I. Ch rysochlarnys skutchii Hammel (A. B, Skulch 4759: C, Aguilar 1771). — A. Flowering Franc
Staminale flower. — C. Fruit.

daries, fading to the margin or looping to an ob¬
scure submarginal vein; petiole 0.6-1. 2 cm. Inflo¬
rescence appearing axillary or terminal, 1—3 cm,
erect, glabrous; flower buds 2-3 mm, the outer 2
sepals Vi tin* length of the bud; stamens oa. 60,
apparently without filaments, anthers of stamens
and staminodia broader than long, < 0.5 mm, sur¬
mounting a resiniferous capitulum. Fruits ellipsoid-

falcate, bean-shaped, red. 1.5— 2.5 cm, 3-valved but
usually just 1 -seeded, the seed black or brown with
an orange aril; sepals persistent in fruit.

Distribution. Very wet forest; 300-1400 m.
Eastern Panama, Colombia, Ecuador.

Because of its apparently axillary inflorescences,
bud-scale scars, and resin-producing staminate

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

371

flowers, C. tenuifolia's position in Chrysochlamys is
somewhat dubious. It has many close and very
prominent lateral veins, and small, essentially ter¬
minal inflorescences that are often left in an ap¬
parently axillary position (only one per node) by ihe
dominant axillary vegetative shoot. It also some¬
times has unambiguously terminal inflorescences,
and has an androecium with the stamens more or
less eonnivent into an apparently resiniferous ca-
pitulum, much as in the resin-producing flowers of
some species of Clusia. The staminodia of the pis¬
tillate flowers are connate into a ring or collar
around the base of the ovary as in Chrysochlamys
memhranacea. The small, narrowly ellipsoid-fal¬
cate, capsulate fruits mostly have only three sessile
stigmas, and usually one seed with a markedly cel¬
lular aril. In leal venation and bud scales it is very
similar to Symphonia, but for numerous reasons
would be even more out of place there than in
Chrysochlamys. The species must certainly belong
to the Tovomita group, but further study is in order.
Although D’Arey (1980) created the combination
Tovomitopsis micrantha (Engler) D’Arcy for this ma¬
terial, the Peruvian Chrysochlamys micrantha En¬
gler, by its original description and type photo,
bears no resemblance to it.

Representative specimens. PANAMA. San Bias: forest
SE of Puerto Obaldia, Aug. P>71 (fr). Croat 16754 (MO);
mountains above Puerto Obaldia, Aug. 1971 (fr). Gentry
1487 (MO). Darien: Cerro Mali on Colombian border,
1400-1500 m, Jan. 1975 (fr). Gentry & Mori 18806 (MO);
Parque Nacional del Darien, ridge between Rfo Topalisa
and Rfo Pucuro, ca. 17 km E of Pueuro, 600—850 m, Oct.
1987 (fr), de Nevers el al. 8843 (MO), (fl). Hammel et al.
16379 (MO), 700-1400 m. Oct. 1987 (fl. fr). Hammel et
al. 16357 (MO); Rfo Tacarcuna, vicinity of old Tacareuna
village, 580 m. Jan. 1975 (fl). Gentry & Mori 13586 (MO).

Chrysochlamys tennis Hammel, sp. nov. TYPE:
Costa Rica. Limon: Cordillera de Talamanca,
Fila Matama, 09°47'18"N, 83°08'45"W, 1350
m, Apr. 1989 (staminate fl), G. Herrera & ,4.
Chacon 2673 (holotype, CR; isotypes, MO (2)).
Figure 2.

Chrysochlamys glauca affinis a qua differt foliis angus-
tioribus et plerumque in sicco ferruginis vice glaucorum,
venatione magis inconspicuo. inflorescentia tenuiore.

Shrubs or small trees 2—5 m; resin clear to tar¬
dily creamy. Petiole 0.5—2 cm. Leaf blades narrowly
elliptic, 5—13 X 1— 2.5(— 3.5) cm, the apex acumi¬
nate, the base narrowly acute and shortly decurrent,
both surfaces dull, drying reddish tan to gray-green;
lateral veins ca. 0.5 cm apart, 6—8 pairs, very in¬
distinct, often difficult to distinguish from interse¬
condaries, fading to the margin or looping to form

an obscure submarginal vein; resin canals usually
very obscure, ± parallel to the midrib or at least
somewhere crossing the lateral veins. Inflorescence
paniculate, very lax to slightly congested, ± glo¬
bose, ca. 2 — 5( — 8) cm long, 3- to ca. 30-Howered.
the rachis very slender, often deflexed or pendu¬
lous, glabrous; flower buds 3-4 mm long, usuallv
conspicuously striate with resin canals; outer 2 se¬
pals ± equal and shorter than the bud, sometimes
very narrowly triangular, rarely (in the type and one
nearby paratype, Herrera 2495) the third and fourth
sepals long-acuminate with the filiform tips over¬
topping the rest of the bud by ca. 2 mm; petals
membranaceous; stamens ca. 20, free, 2-2.5 mm
long, the anthers ± globose, ca. 0.3 mm long; stig¬
mas 5. Fruits narrowly obovoid to pyriform, usually
rostrate, 2—3 cm long, reddish, often tinged with
white or pale yellow, 1-3-seeded; seeds with an
orange aril; calyx deciduous in fruit.

Distribution. This species is known from scat¬
tered localities in very wet forest on the Atlantic
slope of Costa Rica and Panama, and from a few
collections from wet forest on the Burica Peninsula,
shared by both countries along their Pacific border.
Although most material is from below 800 m, the
type is from nearly 1400 m.

The type and all other Costa Rican, Atlantic
slope collections, as well as one Panamanian col¬
lection from Bocas del Toro ( Kirkhride & Duke
621), one from Veraguas (McPherson 10724), and
two from the Rfo Blanco de Norte area of the At¬
lantic lowlands of Code, have particularly long,
narrow leaves without obvious intersecondaries and
relatively few-flowered, very lax inflorescences with
the outer sepals slightly to very narrowly triangular.
The Burica Peninsula outliers and all other collec¬
tions from Panama tend to have somewhat broader
and shorter leaves with more obvious intersecon¬
daries, multi-flowered, more congested inflores¬
cences, and broader outer sepals.

Chrysochlamys tenuis is similar to C. glauca in
its dull, few-veined leaves and small flower buds.
The Rfo Guanche and Cerro Campana populations
cited here were discussed by D’Arcy (1980; 1034)
under that species as “suggestive of a distinct va¬
riety or even species.” Rather than further blurring
what then seemed only suggestive evidence, more
recent collections, particularly those from Atlantic
Costa Rica, strengthen the case for recognizing this
entity as a distinct species. Other than the narrower
and usually reddish tan vs. gray-green-drying
leaves, more obscure venation, and smaller, more
delicate inflorescences, C. tenuis is distinguished
from C. glauca by the resin canals of the leaves.

372

Novon

H igure 2. Chrysochlamys tenuis Hammel (A-C, Herrera & Cluicdn 26 72. I). Herrocal et al. 74). — A. Flowering branch.
— IF Flower bud. — C. Staminate flower. — I). Fruit.

which, when visible, are more or less parallel to the
midrib or at least somewhere crossing the lateral
veins, rather than parallel to the lateral veins, as
in C. glauca and most other species. The narrowly
obovoid fruits of C. tenuis are usually only 1- or 2-
seeded, whereas the more or less globose or broadly
obovoid fruits of C. glauca are normally 5-seeded.
Mori et al. 3950 recorded the seeds as “steel blue.”
The new species also superficially resembles the
Panamanian C. angustifolia , which, however, is an

epipetric rheophyte with milky resin, stouter inflo¬
rescence axes, and much larger flower buds.

Original field notes for the type of this species
indicated four duplicates. These were earlier dis¬
tributed as either Tovomitopsis angustifolia or T.
glauca, and their whereabouts, except as indicated
above, are uncertain.

Paratypes. COSTA RICA. Limon: Cordillera de Tal-
amanca, Fila Matama, Cerro Muchilla, 850 m, Apr. 1080
(Ir). Robles & Cluteon 267R (CR. MO); Reserva Biol. Ilitoy

Volume 9, Number 3
1999

Hammel

Synopsis of Chrysochlamys

373

Cerere, Cerro Bitarkara, 800 m, Feb. 1089 (fl bud), Her¬
rera & Soli's 2495 (CR, MO); Rio Tercero, 000 m. May
1985 (fr), Berrocal el al. 74 (CR); Suretka, 200 m. July
1982 (fr), Gomez- Laurito BB39 (CR). Punlurenas: Burica
Peninsula, Punta Burica, M. M. Chavarria 306 et al. (CR.
INB). PANAMA. Chiriqui: Burica Peninsula, 8—11 mi.
W of Puerto Armuelles, 200—500 m, Feb. 1973 (fl, fr),
Croat 21995 (MO), Mar. 1973 (fl, fr), Croat 22472 (MO);

l, a Fortuna hydroelectric [iroject, 12(H) m. Mar. 1978 (fl.
fr). Hammel 2181 (MO). Boeas (lei Toro: (Juebrada Hu¬
ron on Cerro Bonyic, 170-400 m, Apr. 1968 (fr), Kirkbride
& Duke 621 (MO); Chiriquicito— Calderas trail. Apr. 1968
(fr), Kirkbride & Duke 765 (MO). Veraguas: Escuela
Agrfcola Alto de Piedra, road to Calovebora, 450—550 m,
Aug. 1974 (fr), Croat 27665 (MO), Nov. 1974 (fr). Mori &
Kallunki 3224 (MO), Dec. 1974 (fr), Mori et al. 3950
(MO); slopes of Cerro Tute, 600-750 m. Mar. 1987 (fl),
McPherson 10724 (MO). Code: above Fl Cope, ca. 700—
9<K) m, Apr. 1977 (fr). Folsom & Robinson 2371 (MO).
Apr. 1977 (fr). Folsom & Kauke 2652 (MO). Oct. 1977
(fr). Folsom et al. 5707 (MO), Nov. 1977 (fr), Folsom &
Collins 6516 (MO), Jan. 1978 (fr), Hammel 953 (MO),
Apr. 1978 (fr), Hammel 2432 (MO), June 1978 (fr), Ham¬
mel 3587 (MO), Jan. 1989 (fr). McPherson 13627 (MO);
Atlantic slope from Fl Cope to Rio Blanco del Norte, ca.
150 m, Feb. 1980 (fl. fr), Antonio 3653 (MO). 350 m. Feb.
1982 (fl), Knapp 3675 (MO); near Fl Valle, ca. 7(M)-10(M)

m, Aug. 1967 (fr), Duke & Dwyer 13936 (MO), Jan. 1972
(fr), Gentry & Dwyer 3654 (MO), Dec. 1972 (fr). Gentry
6848 (MO), Sep. 1978 (fr), Hammel 4794 (MO), May 1967
(fl), Lewis et al. 1769 (MO), 1785 (MO), Nov. 1974 (fr),
Mori & Kallunki 2960 (MO), June 1975 (fl), Mori et al.
6565 (MO). Panama: Cerro Campana, ca. 900 m. May
1971 (fl), Croat 14681 (MO). 14753 (MO), Sep. 1982 (fr).
D’Arcy 15071 (MO), Mar. 1969 (fl). Porter et al. 4259
(MO); Cerro Trinidad, 8(H)- l(XX) m, Oct. 1946 (fr). Allen
3773 (MO); El l.lano— Carti Road, Feb. 1977 (fl, fr), Foster
2770 (MO). Colon: Santa Rita Ridge, Jan. 1968 (fl).
Dwyer 8564 (MO); Rfo Fscandaloso near Cerro Bruja, 10—
2(H) m. Apr. 1978 (fl), Hammel 2672 (MO), 2713 (MO),
Mar. 1982 (fl), Knapp & Kress 4283 (MO), (fr), Knapp et
al. 4461 (MO); Rio Guanche near Portobelo, ca. 0-100
m. May 1984 (fl. fr). Churchill 5119 (MO), Oct. 1978 (fr).
Dec. 1973 (fl). Gentry 8764 (MO), Oct. 1978 (fr). Hammel
et al. 4932 (MO), July 1971 (fl), Kennedy 1174 (MO). Jan.
1973 (fr), Kennedy & Foster 2132 (MO), Aug. 1981 (fr),
Knapp 1024 (MO), Mar. 1 675 (fr). Mori & Kallunki 5194
(MO), Mar. 1975 (fl), 5212 (MO). June 1975 (fl). Mori et
al. 6443 (MO). Aug. 1975 (fr). Mori <V Witherspoon 7955
(MO), Dec. 1973 (fr). Nee & Gentry 8719 (MO). July 1976
(fr), Sullivan 155 (MO). Oct. 1980 (fr). Sytsma 1645 (MO).

Extralimital Species

Although the strictly South American Balboa
does not occur within the geographic focus of this
paper, taxonomic clarity and logical consistency
prescribe its mention here as an additional syno¬
nym of Chrysochlamys. Planchon and Triana de¬
scribed the genus and its single species, B. mern-
branacea, based only on staminate material and
placed it in their subtribe “Euclusieae” rather than
Tovomiteae. Examination of type material and more
recent collections, of both staminate and pistillate
plants (e.g., Boyle A Boyle 3303 (MO), Croat 73004

(MO), Gentry et al. 14655 (MO), Hammel & Trainer
15831 (MO), Hoover 1287 (MO), Hoover et al. 3136
(MO), Romero-Castaheda 2739 (COL), Rubio &
Quelal 1339 (MO)) support the view that B. mem-
branacea is a species of Chrysochlamys:

Chrysochlamys halbou Hammel, nom. nov. Bas-
ionyrn: Balboa membranacea Planchon &
Triana, Ann. Sci. Nat. Bot., ser. 4. 14: 252.
1860, non C. membranacea Planchon & Tri¬
ana, 1860. TYPE: Colombia. Between Tuquer-
res and Barbacoa, 1176 m, Apr. 1853, Triana
s.n. (holotype, P; isotype, COL).

Excluded Species

Chrysochlamys clusiifolia subsp. panamensis Maguire,
Phytologia 36; 395. 1977. TYPE; Panama. Darien:
Cerro Mai, Serranfa del Darien, Gentry & Mori
13769 (holotype, MO) = Dystovornita paniculata (J.
1). Smith) Hammel, Ann. Missouri Bot. Card. 76:
927. 1989.

D’Arcy (1980) already showed this to belong in
his newly elevated genus, Dystovornita.

Tovomitopsis membranifolia (Standley) D’Arcy, Ann. Mis¬
souri Bot. Card. 67: 1035. 1980. Tovomita mem¬
branifolia Standley, Field Mus. Nat. Hist. Bot. Ser.
22: 160. 1940. TYPE: Panama. Darien: Cana— Cuasf
trail. Chepigana District, Terry & Terry 1478 (holo¬
type. F; isotype, MO) = Tovomita stylosa Hernsley,
Biol. Cent.-Amer. 1; 88. 1879.

Hammel (1986) already showed this to be a syn¬
onym of Tovomita stylosa.

Tovomita croatii Maguire. Phytologia 36: 404. 1977. To¬
vomitopsis croatii (Maguire) D’Arcy, Ann. Missouri
Bot. Card. 67: 1031. 1980. TYPE: Panama. Vera-
guas: road to Calovebora, Croat 14268A (holotype,
MO).

Shrubs 1—2 m. resin clear. Leal blades obovate,
widest above the middle, 20—60 X 7-25 cm, the
apex acuminate, the base narrowing to an acute or
cordulate base; lateral veins ca. 1—2 cm apart, ca.
16-22 pairs, looping to form a distinct submarginal
vein 2—5 mm from the margin; petiole lacking. In¬
florescence 10—26 cm, usually pendulous, glabrous
or puberulent; flower buds ca. 6 mm, the outer se¬
pals often overtopping the inner; stamens 25—30,
anthers of both stamens and staminodia about twice
as long as wide, ca. 1.5 mm; stigmas concave and
spreading. Fruits pyriform, green to pink, ca. 2 cm;
seeds ca. 1 cm; calyx ± persistent in fruit.

Distribution. Very wet, mostly lowland forest,
30— 350(— 1500) m; Costa Rica, Panama, Ecuador.

This rare, somewhat problematic species might
appear to walk the generic borderline between

374

Novon

Chrysochlamys and Tovomita. The outer pair of se¬
pals, although thinly membranaceous, exceed the
hud as in Tovomita, yet styles are lacking and the
lmits apparently do not have the dark maroon pla¬
centa and inner fruit wall typical of Tovomita. Nev¬
ertheless, two Costa Rican collections ( Grayum
■ ii 74, Grayum & Schatz 5276) clearly indicate the
presence ol side shoots with clustered leaves. With
present knowledge, placement of this species in To¬
vomita, particularly because of its branching pat¬
tern, is more justifiable than placement in Chry¬
sochlamys. As a species, it is very distinctive, not
only for the nearly sessile and sometimes very large
leaves, hut also for the numerous (16-22) lateral
veins that are loop-connected to a distinct submar¬
ginal vein. Probably because of the rarity of this
species, its short stature, and large leaves, the typ¬
ical Tovomita branching pattern of T. croatii has
gone unobserved and unrecorded.

Representative specimens. COSTA RICA. Linidii: Re-
serva Biol. Hitoy Cerere, 90—200 m, July 1985 (fr), Gra¬
yum & Hummel 5763 (CR. MO); Canton de Talamanca,
Snkut. 850 m. July 1989 (fr). Herrera 3168 (CR. MO);
Finea Anai, W of Mata de Union (Sixaola region), 30 m.
May 1985 (II). Grayum & Schatz 5276 (CR. MO). PAN¬
AMA. Chiriqui: l’ortuna Dam Area. 1 l(M) rn. May 1984
(fl). Churchill 5243 (MO). Code: Kl Valle. Cerro Gaital.
' MIO-IOOO m. July 1968 (II. fr). Dwyer & Correa 8913
(MO). KCLJADOR. Pichincha: Quito a Sto. Domingo de
los Colorados. km 94, 1300-1500 m, Dec. 1986 (fr). Za k
1529 (MO).

Acknowledgments. This work was supported, in
part, by funds from the National Science Founda¬
tion through a grant (DEB-9300814) to the author
and co-PI M. H. Grayum for the Manual to the
Plants of Costa Rica. I thank Mike Grayum for his

always useful comments on an early draft of the
manuscript and Sylvia Troyo for the line drawings
of the two new species.

I .iterature Cited

Allen, P, H. 1956. The Rain Forests of Golfo Dulce. Univ.
Florida Press, Gainesville.

Bentham, G. & J. I). Hooker. 1862. Genera Plantarum,
Vdl. 1 (1). A. Black. London.

Croat. T. B. 1978. Flora ol Barro Colorado Island. Stanford
Univ. Press, Stanford, California.

Cuatreeasas, J. 1950. Notas a la flora de Colombia, X.

Revista Acad. Colomb. Ci. Exact. 8: 33—64.

D'Arov. W (7 1980. Guttiferae. In: Flora ol Panama. Ann.

Missouri Bot. Card. 67: 969—1043.

Farr. E. R.. J. A. Leussink & E. A. Stafleu. 1979. Index
Nominum Genericorum (Plantarum), Vol. I. Bolin.
Sclieltema Holkema, Utrecht dr. Vi. Junk b. v.. Pub¬
lishers. The Hague.

Haminel. B. E. 1984. Systematic Treatments ol the Cv-
clanthaceae. Marantaceae. Clusiaceae. Cecropiaeeae,
Lauraceae, and Moraceae for the Flora of a Wet Low¬
land Tropical Forest, Finca La Selva, Costa Rica. Un¬
published I’ll. D. Dissertation, Duke University. Dur¬
ham, North ( larolina.

- . 1986. New species ol Clusiaceae from Central

America with notes on Clusia and synonymy in the tribe
Clusieae. Selbyana 9; 1 12—120.

Hemsley, W. B. 1879. Biolgia centrali-americana. Botany.

Vol. 1. Porter and Bulau, London.

Maguire, B. 1977. Notes on the Clusiaceae — Chief!) of
Panama. I. Phytologia 36; 391—407.

Planchon. J. & J. Triana. I860. Memoire sur la famille
des guttiferes. Ann. Sci. Nat. Bot. ser. 4. 13: 306-376;
14: 226-367; 15: 240-319.

Poeppig, E. F. 1842. Nova genera ac species plantarum.
Vol. 3. Hofmeister, Leipzig.

Standley, P. (.. 1937. Guttiferae. In: Flora ol Costa Rica.
Field Mus. Nal. Hist.. Bot. Ser. 18: 703-712.

- A L. (). Williams. 1961. Guttiferae. In: Flora of

Guatemala. Eieldiana: Bot. 24, 7: 36—61.

Williams, L. 0. 1959. Guttiferae from middle America.
Tropical Woods 11: 15-18.

A New Name for a Common Ecuadorian and Peruvian Wild

Tomato Species

Sandra Knapp

Department of Botany, The Natural History Museum, Cromwell Road, London SW7 5BD,

United Kingdom

David M. Spooner

USDA-ARS, Department of Horticulture, University of Wisconsin, 1575 Linden Drive,
Madison, Wisconsin 53706-1590, U.S.A.

ABSTRACT. The new name Solarium habrochaites
S. Knapp & D. M. Spooner is coined to replace the
homonym Solarium agrimoniifolium (Dunal) J. F.
Macbride, which has been used for the wild tpmato
species known in Lycopersicon as Lycopersicon hir-
sutum Dunal.

While preparing the Solanaceae for the Cata¬
logue of the Vascular Plants of Ecuador (Jorgensen
& Leon-Yanez, 1999), we discovered that a new
name was necessary for the common wild tomato
species previously known as Lycopersicon hirsutum.
We accept species of tomatoes in the genus Sola¬
num following molecular and morphological studies
of Child (1990), Spooner et al. (1993), Bohs and
Olmstead (1997), and Olmstead and Palmer (1997).
Spooner et al. (1993) provided a compilation of
names and combinations in Solanum for all nine
taxa of tomatoes recognized by Rick et al. (1990).
The epithet hirsutum is not available for this taxon
in Solanum, as it is occupied by Solanum hirsutum
Dunal (1816), and the pre-existing combination
used by Spooner et al. (1993), Solanum agrimoni¬
ifolium (Dunal) J. F. Macbride, for L. hirsutum is a
later homonym of an epithet already published in
Solanum. A new epithet in Solanum is therefore
necessary for this taxon.

Solanum habrochaites S. Knapp & D. M. Spoo¬
ner, nom. nov. Replaced name: Lycopersicon
hirsutum Dunal, Solanorum synopsis 4. 1816,
non L. hirsutum Dunal, in DC., Prodromus
13(1): 24. 1852. TYPE: Ecuador. Loja: “Hab.
in agris Peruvianis prope Loxa,” Humboldt &
Bonpland s.n. (holotype, P). Locality for type
collection fide Kunth (Humboldt et al., 1818),
“Crescit prope Loxa Novo-Granatensium, alt.
1060 hex.”

Solarium agrimoniifolium (Dunal) J. F. Macbride, Field

Mus. Nat. Hist., Bot. Ser. 13: 159. 1962. Horn, illeg.,
non Rydberg, Bull. Torr. Bot. Club 51: 154. 1924.
lycopersicon agrimoniifolium Dunal, in DC.. Prod¬
romus 13(1): 24. 1852; “ agrimoniaefolium TYPE:
Peru. Sin. loe., Pavon s.n. (holotype, G not seen).

Both Dunal’s and Rydberg’s spellings of the ep¬
ithet agrimoniifolium are errors to be corrected un¬
der Article 60.8 of the Code (Greuter et al., 1994).
The new specific epithet is derived from the Greek,
habro = soft, delicate, and chaites = haired, thus
retaining as closely as possible Dunal’s original in¬
tention.

Dunal (1816) stated that he saw the specimen he
named Lycopersicon hirsutum in the Humboldt and
Bonpland herbarium (“v.s.h. H et B.”), but a spec¬
imen of this taxon could not be found in P-Bonpl.
Kunth (Humboldt et al., 1818) did not see speci¬
mens of this taxon while writing his account of
Humboldt and Bonpland’s plants, “Specimena hu-
jus et praecedentis speciei haud vidi.” No speci¬
men identifiable as this taxon and collected by
Humboldt and Bonpland is present in the Willd-
enow herbarium (microfiche edition IDC). Dunal
had intended to publish an illustrated edition of his
Solanorum synopsis (1816), and to that end had a
series of drawings of the taxa described therein pre¬
pared by the artist Node-Veran, to which he re¬
ferred in Solanorum synopsis (Dunal, 1816). These
drawings are held in the collections at Montpellier
(MPU), and in the set is a plate of Lycopersicon
hirsutum (Fig. IB). A specimen in the general her¬
barium at P collected by Humboldt and Bonpland
and annotated by Dunal that is a perfect match for
this drawing is clearly the holotype for this name
(Fig. 1). On the specimen are pinned several labels
with different descriptions, one of them (far left
lower corner) is that of Kunth (Humboldt et al.,
1818). On the label in the upper left is written
“Solanum dentatum” and a draft description in
Dunal’s hand. The description, however, does not

Novon 9: 375-376. 1999.

376

No von

Figure 1. — A. Holotype specimen of Solarium habrochailes S. Knapp & D. M. Spooner (Lycopersicon hirsutum Dunal)

at P. — IF Original drawing of Lycopersicon hirsutum Dunal by Node-Veran for Dunal, Solanaceae n° 83 (copyright
University Montpellier II. reproduced with permission: National collections of the herbarium of the Institut de Botanique
(MPU)).

match that ot L. hirsutum in Solanorum synopsis,
nor does it match that of L. dentatum Dunal, also
described in the same work (a synonym of 5. pe-
ruvianum L.). To preserve nomenclatural stability,
we prefer to coin a new name rather than use an
epithet that perhaps was not intended by Dunal for
this particular plant.

Acknowledgments. We thank R. K. Brummit of
the Royal Botanical Gardens, Kew, for valuable ad¬
vice and a thorough and careful review, and J.
Mathez of the Institut de Botanique, Montpellier,
for permission to reproduce the unpublished Node-
Veran drawing.

Literature Cited

Bolls, L. & K. C. Olmstead. 1997. Phylogenetic relation¬
ships in Solatium (Solanaceae) based on ndhV sequenc¬
es. Syst. Bot. 22: 5-17.

Child, A. 1990. A synopsis of Solarium subgenus Potatoe

(G. Don) (D’Arcy) ( Tuberarium (Dun.) Bitter (s.l.)). Eed-
des Bepert. 101: 209—235.

Dunal, M.-F. 1816. Solanorum synposis. Montpellier.
Greuter. W.. F. B. Barrie, H. M. Burdet, W. G. Chaloner,
V. Demoulin, I). L. Hawksworth, P. M. Jprgensen, D. H.
Nicolson, P. C. Silva, P. Trehane & J. McNeill. 1994.
International Code ol Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Humboldt, F. II. A. von, A. Bonpland & C. S. Kunth.

1818. Solaneae. Nov. gen. sp. 3: 1—64.

Jorgensen. P. M. & S. Le6n-Yanez (Editors). 1999. Cata¬
logue ol the Vascular Plants of Ecuador. Monogr. Syst.
Bot. Missouri Bot. Card, in press.

Olmstead, B. G. & J. I). Palmer. 1997. Implications for
the phylogeny, classification, and biogeography ol So¬
larium from cpDNA restriction site variation. Syst. Bot.
22: 19-29.

Biek, C. M., II. Laterrot & J. Philouze. 1990. A revised
key for the Lycopersicon species. Tomato Genet. Coop.
Rep. 40: 31.

Spooner, D. M., (7 J. Anderson & B. K. Jansen. 1993.
Chloroplast DNA evidence for the interrelationships of
tomatoes, potatoes, and pepinos (Solanaceae). Amer. J.
Bot. 80: 676—688.

A New Species of Deciduous Azalea (Rhododendron section
Pentanthera ; Ericaceae) from South Carolina

Kathleen A. Kron

Department of Biology, Wake Forest University, Winston-Salem,

North Carolina 27109-7325, U.S.A.

Mike Creel

Department of Natural Resources, 1000 Assembly Street, P.0. Box 167, Columbia,

South Carolina 29202, U.S.A.

Abstract. A new species of deciduous azalea.
Rhododendron eastmanii, has been discovered in
South Carolina, U.S.A. This azalea can he distin¬
guished from the other members of Rhododendron
sect. Pentanthera by its distinctive morphology,
flowering time, and fragrance. It is presently known
only from Orangeburg and Richland counties.

A new species of Rhododendron for South Car¬
olina is recognized based on the diagnosability cri¬
terion of Davis and Nixon (1992). The deciduous
leaves, narrow corolla tube, and 5 declinate sta¬
mens strongly exserted from the corolla place this
species in Rhododendron sect. Pentanthera G. Don.
Within the section, 14 species are currently rec¬
ognized in North America (Kron, 1993). Known
from only two localities in South Carolina. R. east¬
manii presently consists of approximately 500 in¬
dividuals at each site.

Rhododendron eastmanii Kron & Creel, sp. nov.
TYPE: U.S.A. South Carolina: Orangeburg
Co., Santee State Park along Limestone Sink
hiking trail, 17 May 1993, Kron 3023 (holo-
type, WFU; isotypes, MO, USCH, WFU). Fig¬
ures 1, 2.

Species haec a Rhododendron arborescens (Pursli) Tor¬
re y differt corolla supra macula aurea, perulis genimis
marginibus glandulosis, ramunculis dense pubeseentibus,
et floribus fragrantibus.

Shrub or small tree to 5 m tall, non-rhizomatous;
young twigs usually reddish brown, densely covered
with unicellular and multicellular eglandular hairs.
Vegetative bud scales glabrous abaxially; margin
unicellular-ciliate. Leaf blade membranaceous,
ovate or obovate to elliptic, 4. 3-7.1 X 1 .8-2.9 cm;
base acute to oblique; apex acute to obtuse, mu-
cronate; adaxial surface sparsely to densely covered
with unicellular hairs, the midvein densely covered

with unicellular hairs, multicellular eglandular
hairs sparsely scattered on adaxial surface; abaxial
surface moderately to densely covered with unicel¬
lular hairs and multicellular eglandular hairs, the
midvein densely covered with unicellular hairs and
multicellular eglandular hairs, 2° veins occasion¬
ally with multicellular eglandular hairs; margin en¬
tire, ciliate with multicellular eglandular hairs; pet¬
ioles 0.25-0.70 cm long, densely covered with
unicellular hairs and sparsely to densely covered
with multicellular eglandular hairs. Flower bud
scales chestnut brown; abaxial surface glabrous;
margin unicellular-ciliate near apex, glandular
along the lower % of margins. Flowers appearing
after the leaves have expanded; inflorescence a
shortened raceme of 5—9 flowers. Pedicels 0.5— 1.1
cm long, densely covered with unicellular hairs and
sparsely to densely covered with multicellular
eglandular hairs, occasionally with multicellular
glandular hairs. Sepals 0.05—0.1 cm long, often
varying in length on the same flower; margins se¬
tose with multicellular eglandular hairs; abaxial
surface sparsely to densely covered with unicellular
hairs and multicellular eglandular hairs, occasion¬
ally with a few weakly glandular multicellular hairs.
Corolla white with a yellow blotch on the upper
corolla lobe and pink-tinged lobes on newly opened
flowers, fragrance strong, fresh, and sweet, the tube
longer than the limb and gradually expanding into
it; upper corolla lobe 0.9-1. 7 X 0.8-1. 6 cm; lateral
lobes 1. 1-2.5 X 0. 5-1.1 cm; corolla tube 1.3-2. 5
cm long, 0.25—0.35 cm wide at base; outer surface
of corolla densely covered with unicellular hairs
and sparsely to densely covered with multicellular
gland-tipped hairs, the glands usually weakly de¬
veloped; inner surface of corolla densely covered
with unicellular hairs. Stamens 4.5— 6.7 cm long,
with dense flattened unicellular hairs on proximal
2.3— 2.7 cm of filament, exserted 2.2— 3.5 cm be-

Novon 9: 377-380. 1999.

378

Novon

Figure 1. Rhododendron eastmanii Kron & Creel. Photograph hy VI. Creel.

yond the throat of corolla. Style (3.7— )5. 4— 6.5 cm
long, exserted ( 1 .6— )2.S — 4.2 cm beyond the throat
of corolla, with dense unicellular hairs on the prox¬
imal 0.4— 2.0 cm; stigma 0.15—0.25 cm wide. Ovary
0.25—0.35 cm long, 0.15—0.20 cm wide at the base,
densely covered with unicellular and multicellular
eglandular hairs. Capsules 1.2-1. 9 cm long, 0.4-
0.8 cm wide at base, sparsely to densely covered
with unicellular hairs and multicellular eglandular
hairs. Seeds pale to dark chestnut brown, elliptic
to fusiform, 2.5— 4.0 mm long; testa expanded and
dorsiventrally flattened, surrounding the body, the
cells elongate with transverse end-walls.

Etymology. The specific epithet honors Charles
Eastman of Columbia, South Carolina, who first dis¬
covered this species while bird-watching in Santee
State Park.

Distribution and ecology. Orangeburg and
Richland counties, South Carolina. On north-facing
slopes with well-drained, nearly neutral soils, often
near limestone. Associated species; Rhododendron
canescens (Michaux) Sweet, R. periclymenoides (Mi-
chaux) Shinners, Stewartia malacodendron L., Sty-
rax grandifolia Aiton, Oxydendrum arboreum (L.)
DC., Kalmia latifolia L., Fagus grandifolia Ehr-

hart, Quercus spp., Liriodendron tulipifera L., and
Symplocus tinctoria (L.) L'Heritier. Flowering in
early May.

Rhododendron eastmanii is similar in general ap¬
pearance to R. occidentale (Torrey & A. Gray) A.
Gray, from which it can be distinguished by its lack
of glandular hairs on the leaves and petioles (Fig.
2; Kron, 1993). Rhododendron eastmanii is geo¬
graphically isolated from R. occidentale, which oc¬
curs from southern California to southern Oregon.
Further investigation of the potential relationships
of R. eastmanii and R. occidentale would be an in¬
teresting evolutionary study. Among the other
white-flowered azaleas of the southeastern United
States (i.e., R. alabamense Rehder, R. arborescens
(Pursh) Torrey, R. atlanticum (Ashe) Rehder, and
R. viscosum (L.) Torrey), Rhododendron eastmanii
can be distinguished from R. alabamense by the
flowers opening alter the leaves have expanded,
and from R. arborescens, R. atlanticum, and R. vis¬
cosum hy the presence of a yellow blotch on ihe
upper corolla lobe (Fig. 1, Table 1). In addition, the
densely hairy young twigs found in plants of R.
eastmanii are lacking in R. arborescens. The weakly
glandular corolla of R. eastmanii is also different

Volume 9, Number 3
1999

Kron & Creel 379

Rhododendron eastmanii from S. Carolina

Figure 2. A— F. Rhododendron eastmanii {Kron 3023). — A. Flowering branch. — B. Flower. C. Inflorescence hud
srale. — |). Close-up of inflorescence bud-scale margin show ing glands along lower %. — F,. Leaf. — F. Capsule.

380

Novon

Table 1. Disti

inguishing features of the white-flowered species i

of Rhododendron sect.

Pentanthera.

Time of dowering relative

Inflorescence bud

1 Yellow blotch on upper Glandular hairs on

Species

to leaf expansion

scale margin

corolla lobe

leaf margins and petiole

R. eastmanii

after

glandular

present

absent

R. alabamense

before

non-glandular

present

absent

R. arborescens

after

non-glandular

absent

absent

R. atlanticum

before

non-glandular

absent

absent

R. occidentale

variable

variable

present

present

R. viscosum

after

variable

absent

absent

from the very sticky, glandular corolla of R. viscos¬
um and R. atlanticum (Kron, 1993).

Reports of R. alabamense from South Carolina
(Radford et al.. 1981) may actually refer to R. east-
manii. A specimen (i.e., Abies 25522, NCU) iden¬
tified as R. alabamense was seen l>v the authors. It
lacks corolla color information, hut in other re¬
spects fits R. eastmanii. This specimen also has
fully expanded leaves at the time of flowering — a
characteristic of R. eastmanii hut not R. alaba¬
mense. The authors and other field botanists have
noted that R. eastmanii flowers when no other na¬
tive azaleas in the area flower, thus potentially sug¬
gesting the reproductive isolation of this species
from other native azaleas.

Preliminary cladistic analysis of morphological
characters (Kron, 1993) for species within Rhodo¬
dendron sect. Pentanthera place R. eastmanii as
closely related to R. calendulaceum (Michaux) Tor-
rey and R. cumberlandense E. Braun. Although the
latter two species possess orange to red corollas, all
three share the character of glandular inflorescence
hud scales (Fig. 2). Further, more detailed study is
necessary to confirm this relationship.

The discovery of a new species of deciduous aza¬

lea in South Carolina is unexpected and significant.
Although the site of the holotype is within Santee
State Park, the only other known population is pres¬
ently threatened by urban development (Richland
Co.). To discourage inappropriate collection by aza¬
lea enthusiasts, MC has developed a collection of
rooted cuttings that will be made available to the
public in the near future.

Paratypes. U.S.A. South Carolina: Richland Co., 1
May 1997 Pittman 05109701 (USCH). Orangeburg Co.,
Santee State Park. 17 May 1993, Kron 5021. 3022 (A KIJ).

Acknowledgments. We thank Berl Pittman for
providing additional specimens of R. eastmanii , R.
Fuller for the illustration, and the Research and
Publication Fund of the Wake Forest Graduate
School for support.

Literature Cited

Davis, J. I. & K. C. Nixon. 1992. Populations, genetic
variation, and the delimitation of phylogenetic species.
Syst. Biol. 41: 421-435.

Kron, K. A. 1993. A revision of Rhododendron section
Pentanthera. Edinburgh J. Hot. 50: 249—364.

Radford. A. E„ H. E. Aides & C. R. Bell. 1981. Manual
of the Vascular Flora of the Carolinas. Univ. North Car¬
olina Press, Chapel Hill.

A New Lobelia from Mexico, with Additional New Combinations in

World Campanulaceae

Thomas G. hammers

Department of Botany, Field Museum of Natural History, Chicago, Illinois 60605-2496, U.S.A.
Current address: Department of Biology and Microbiology, University of Wisconsin Oshkosh,
Oshkosh, Wisconsin 54901, U.S.A. lannners@uwosh.edu

ABSTRACT. Lobelia guerrerensis is a new species
endemic to pine, pine-oak, and mixed deciduous
lorests of the Sierra Madre del Sur in Guerrero,
Mexico. Though collected several times over the
past 60 years, the specimens were consistently mis-
identified as the narrow-leaved phase of L. laxiflo-
ra. Detailed phenetic analyses, however, have
showed them to be morphologically quite distinct.
In addition, several new combinations needed tor a
forthcoming checklist of world Campanulaceae are
published.

Work is drawing to a close on a checklist of Cam¬
panulaceae s.l. currently recognized in the world’s
taxonomic literature. Four precursor papers effect¬
ing requisite nomenclatural changes (Lammers,
1995, 1998a, 1999) and describing recently de¬
tected species (Lammers, 1998b) have already ap¬
peared. This is the fifth and (I hope) final install¬
ment in that series.

It should be noted that some of the names vali¬
dated here stem from recent studies conducted un¬
der my direction by two students, Michael James
Eakes of the University of Illinois and Kari Ellen
Morris of the University of Illinois-Chicago. It will
not be possible to publish their complete results
prior to publication of the checklist. Therefore, no¬
menclatural innovations dictated by their studies
are extracted and published here in advance, so
that the resulting names will be available for use
in the checklist and elsewhere.

Actions undertaken here are of three types: (1)
description of a new species of Lobelia L. from
Guerrero, Mexico (with a change in rank for a re¬
lated taxon); (2) new combinations in Campanula
L., Centropogon C. Presl, and Lobelia, dictated by
the principle of priority; and (3) new combinations
for species and subspecies of Codonopsis Wallich,
Cyanea Gaudichaud, Cyclocodon Griffith, and De-
lissea Gaudichaud, which do not have a name at
that rank in the genus or species to which they are
assigned.

New Species

Unpublished morphometric studies in Lobelia
sect. Homochilus A. DC. by Eakes (cf. Hamlin,
1995; Eakes & Lammers, 1996) disclosed the exis¬
tence of a discrete group of specimens from the Mex¬
ican state of Guerrero, which did not conform to any
previously recognized taxon. These specimens had
been collected over the past 60 years by different
botanists and widely distributed to the world’s her¬
baria. In every instance, however, they were identi¬
fied as the narrow-leaved phase of L. laxiflora Kunth,
a species widespread in Mexico and Central Amer¬
ica (see below). It was only with detailed phenetic
analyses that the distinctness of these plants was
revealed. Because these plants were as distinct mor¬
phologically as any species in the section, they are
here described as a new species:

Lobelia guerrerensis Eakes & Lammers, sp. nov.
TYPE: Mexico. Guerrero: municipio de Atoyac
de Alvarez, a 6 km al SO de Puerto del Gallo,
bosque mesofilo de montana, 2320 m, 29 Mar.
1983, Soto & Martinez 5136 (holotype,
MEXU). Figure 1.

A lobelia laxiflora f’oliis cernuis cum marginibus inte-
gris vel subintegris et minute revolutis, corollae tubo flavo
vel luteo-aurantiaco et lobis rubris vel aurantiacis, tubo
antherarum dense villoso a basi ad summum tecto cum
trichomatibus isabellinis 2—4 mm longis, antheris ventral-
ibus apice caespitosis cum trichomatibus isabellinis 1.5—2
mm longis, capsulis 9—11 mm diametro, et seminibus 0.7—
0.8 mm longis 0.3— 0.4 mm latis differt.

Stems 1—2 m tall, woody or sufiruticose, branched
or unbranched, erect or ascending, glabrous. Leaves
sessile, glabrous; lamina linear or lanceolate, 7—18 cm
long, 0.3-1. 1 cm wide, cemuous, chartaceous; margin
entire or sometimes with a few distant minute callose
teeth, minutely revolute; apex narrowly acuminate;
base cuneate or attenuate. Flowers solitary in the axils
of the upper leaves; pedicels 60-130 mm long, ca. 1
mm diam., straight or slightly incurved, glabrous, bi-
bracteolate in the lower half; bracteoles 0.5—3 mm
long, linear. Hypanthium depressed hemispheric, 4—

Novon 9: 381-389. 1999.

382

Novon

figure 1. htbelia guerrerensis Hakes & Hammers (from the holotype arul Reveal et al. 4337 , TEX). — A. Upper portion
of stem. — B. Flower.

Volume 9, Number 3
1999

Lammers
New Lobelia

383

5 mm long, 9-11 mm diam., glabrous; base rounded
or truncate. Calyx lobes narrowly triangular or trian¬
gular, 2.5-5 mm long, 1-2 mm wide, glabrous; margin
entire; apex acuminate. Corolla yellow or yellowish
orange on tube, grading to orange or red on the lobes,
35-44 mm long, glabrous: tube 19-25 mm long, 7—
10 mm diam. at base, gradually tapering to 6—8 mm
at mouth, laterally fenestrate, dorsaUy cleft nearly to
base; dorsal lobes linear, 14—22 mm long, 1.5-3 mm
wide, recurved, the apex acute; ventral lobes 12-21
mm long, straight or slightly deflexed, connate for
most of their length, forming a trilid lip, these seg¬
ments triangular, 4—8 mm long, 1.5— 2.5 mm wide,
acute at apex. Staminal column somewhat exserted;
filament tube 29-33 mm long, 1.3— 1.7 mm diam.,
pale yellow, glabrous; anther tube 2.5—4 mm diam.,
the surface from apex to base densely villous with
dirty yellow trichomes 2—4 mm long; dorsal anthers
8.5-9 mm long; ventral anthers 7-7.6 mm long, with
tufts of dirty yellow trichomes 1.5—2 mm long at apex.
Capsule broadly ovoid, 8-10 mm long, 9—11 mm
diam. Seeds honey-colored, ellipsoid or cylindrical,
0. 7-0.8 mm long, 0.3-0.4 mm wide, 0. 1-0.2 mm
thick, the surface faintly striate.

Distribution , habitat, and phenology. Endemic to
Guerrero, Mexico, where the plants grow on steep
wooded slopes and moist embankments in pine, pine-
oak, and mixed deciduous forests of the Sierra Madre
del Sur, at elevations of 1890-2750 m. Flowering be¬
gins in early October and continues through February.

Etymology. This species takes its name from
the state of Guerrero.

Relationships. As noted above, specimens of L.
guerrerensis were identified previously as the narrow¬
leaved phase of L. laxiflora (see below). The two taxa
are similar in habit, general size and shape of their
leaves, inflorescence structure, floral structure, and
seed morphology. However, the new species differs
in its cemuous (vs. flat) leaves with the margin mi¬
nutely revolute (vs. flat) and entire or subentire (vs.
denticulate, serrulate, serrate, or doubly serrate);
longer, denser, dirty yellow (vs. white) pubescence
on the anther tube; broader capsules; and slightly
larger seeds. Perhaps most conspicuously, the prig-
mentation pattern of the corolla is the reverse of that
seen in L. laxiflora. In that species, the tube is or¬
ange or red and the lobes yellow or yellowish orange,
while L. guerrerensis has just the opposite: yellow or
yellowish orange tube with orange or red lobes.

Paratypes. MEXICO. Guerrero: Mazatlan. Berlin 53
(ENCB); 1-3 km NW of Puerto El Gallo, Breedlove 36053
(CAS, MICH); 37.9 km NE de El Parafso, Cowan 4961
(TEX); 6.5 km W de Puerto del Gallo por camino a Par¬
afso, Hernandez & Tenorio 858 (MICH); second ridge W
of Petlacala, Mexia 9049 (ARIZ. B, CAS, G. K. MO, NY.

W); 3 mi. SW of Puerto del Gallo, Reveal et al. 4337
(CAS, GH. K. MICH, MO. NY. TEX); Cerro A (quit ran
cerca Tuxpan, Schwabe in 1977 (B).

Key to the Species of Lobelia sect. Homochiuis

la. Leaves broadly and conspicuously decurrent on

stem, their margins coarsely dentate or lobed;
pedicels shorter than the flowers; calyx lobe mar¬
gins fimbriate-dentate; corollas monochromatic,
purple or red-purple (Peru) .

. L. decurrens Cavanilles

lb. Leaves not decurrent on stem, their margins sub¬
entire or minutely toothed; pedicels equaling or
longer than the flowers; calyx lobe margins entire
or minutely toothed; corollas usually bicolored,
yellow, orange, or red.

2a. Lamina broadest at or near middle, 3.5—8
cm wide, on stout petioles 10—30 mm long;
stems, (lowers, and ventral snrlace ol lamina

white-tomentose (Oaxaca) .

. L. ghiesbreghtii Decaisne

2b. Lamina usually broadest below middle, 0.2-
5 cm wide, sessile or on slender petioles up
to 7 mm long; stems, flowers, and ventral
surface of lamina glabrous or pubescent, but
never white-tomentose.

3a. Pedicels 85—210 mm long; calyx lobes
6—18 mm long; dorsal corolla lobes 20-
33 mm long, the ventral 18—26 mm
long: dorsal anthers 9-12 mm long, the
ventral 7.5— 9.5 mm long; capsules 10—

15 mm long; seeds finely reticulate

(Guerrero to Guatemala) .

. L. aguana F. E. Wimmer

3b. Pedicels 20—130 mm long; calyx lobes
1—6 mm long; dorsal corolla lobes 10—

22 mm long, the ventral 10—21 mm
long; dorsal anthers 6—9 mm long, the
ventral 4.5— 7.6 mm long; capsule 6-12
mm long; seeds faintly striate.

4a. Lamina cernuous, the margin entire
or nearly so and minutely revolute;
corolla yellow or yellowish orange
on tube, red or orange on lobes: an¬
ther tube with dense dirty yellow
trichomes 2—4 nun long from apex
to base; ventral anthers with a tuft
of dirty yellow trichomes 1.5—2 mm
long at apex; capsules 9-1 1 mm
diam.; seeds 0.7— 0.8 mm long. 0.3—

0.4 mm wide (Guerrero) .

. ... L. guerrerensis Lakes & Lammers
4b. Lamina plane, the margin denticu¬
late. serrulate, serrate, or doubly
serrate and flat; corolla red or or¬
ange on tube, orange or yellow on
lobes; anther tube with sparse to
moderately dense white trichomes
0.5—1 mm long on dorsal surface
toward apex; ventral anthers with a
tuft of white trichomes 0.7— 1.3 mm
long at apex; capsules 7—9 mm
diam.; seeds 0.5— 0.6 mm long, 0.2—

0.3 mm wide (Arizona to Colombia)
. L. laxiflora kunth

384

Novon

Lobelia laxiflora, the species most similar to L
guerrerensis, is a geographically widespread and high¬
ly polymorphic species in which several varieties and
forms have been recognized. The most commonly fol¬
lowed infraspecific classification is that of MeVaugh
(1943), who divided L laxiflora into four varieties:
var. angustifolia A. DC., var. laxiflora, var. nelsonii
(Femald) MeVaugh, and var. stricta (Planchon & Oer¬
sted) MeVaugh. However, Eakes’s morphometric stud¬
ies failed to support this classification. Instead, his
analyses suggested that L laxiflora could only be di¬
vided meaningfully into two sets of populations. The
one set corresponded to variety angustifolia, while the
other comprised specimens referable to the three oth¬
er varieties. The two groups were largely parapatric
and connected by occasional morphological interme¬
diates in the contact zone.

Situations such as this are best expressed no-
menclaturally by use of subspecific rank (cf. Lam-
iners, 1988, 1990, 1995). Thus, L. laxiflora is here
divided into subspecies angustifolia and autonymic
subspecies laxiflora. The requisite change is ef¬
fected below, followed by a key for their identifi¬
cation. Complete synonymy is given for subspecies
angustifolia-, all other names included under L. lax¬
iflora by MeVaugh (1943) are synonyms of the au¬
tonymic subspecies.

An added benefit of the use of subspecific rank
in this case is that workers who find it useful to do
so may continue to recognize MeVaugh s (1943) three
other varieties. Though Eakess analyses did not sup¬
port such a classification, subspecies laxiflora could
be subdivided into variety laxiflora, variety nelsonii,
and variety stricta by those who believe that these
taxa can be distinguished in their region.

Lobelia laxiflora subsp. angustifolia (A. DC.)
Eakes & Lammers, stat. nov. Basionym: Lo¬
belia laxiflora var. angustifolia A. DC., in DC.,
Prodr. 7: 383. 1839. Lobelia persicifolia var.
angustifolia (A. DC.) Vatke, Linnaea 38: 723.
1874. Lobelia laxiflora f. angustifolia (A. DC.)
Voss, in Siebert & Voss, Vilm. Blumengiirtn.
(ed. 3) 1: 576. 1894. Lobelia angustifolia (A.
DC.) Urbina, Cat. PI. Mexican. 201. 1897; non
Chamisso, Linnaea 8: 219. 1833; nec Ben-
tham, in Endlicher, Enum. PI. 74. 1837.
TYPE: Mexico. Mexico: Toluca, Apr. 1834,
Andrieux 267 (lectotype, here designated, K;
photograph. F; isolectotypes, K. W). De Can¬
dolle based the name on this collection plus
three icones, two published descriptions, and
living plants seen in cultivation. Of these, the
specimen can best be identified with the de¬
scription. However, the collection is not rep¬
resented in De Candolle’s own herbarium, and
a duplicate is designated as the lectotype.

Lobelia dracunculoid.es Willdenow ex Schultes, in Roemer
& Schultes, Syst. Veg. 5: 56. 1816. TYPE: “America
merirlionalis," leg. Humboldt & Bonpland, Herb.
Willdenow 3989 (holotype, R-A |IDC-microfiche!]).
Rapuntium kunthianum C. I’resl. Prodr. Monogr. Eobel.
27. 1836. lobelia persicifolia var. amygdalina Vatke,
l.innea 38: 723. 1874. Roth validated by indirect
reference to: Lobelia persicifolia Cavanilles sensu
kunth, in HRK, Nov. Gen. Sp. 3: 310 (quarto), 242
(folio). 1819; non Cavanilles, Icon. 6: 12. 1801; nec
Lamarck, Encycl. 3: 584. 1792. TYPE: Mexico.
Mexico: prope Zumpango, Humboldt & Bonpland
s.n. (lectotype, here selected, P-Bonpl. [IDC-micro-
fiche!]: isolectotypes, R. R-A [IDC-microfiche!]).
Though there is no evidence that Presl saw any of
these sheets. Article 7.7 of the Code (Creuter et al.,
1994) dictates that in cases such as this, the type
be designated in the context of the validating de¬
scription. This is not an avowed substitute for ille¬
gitimate L. persicifolia Cavanilles; on the same page,
Presl recognized that species under the replacement
name Rapuntium cavanillesianum (Schultes) C.
Presl.

lobelia cavanillesii var. lutea F. Haage & k. Schmidt, Gar-
tenflora 52: 577. 1903. Lobelia laxiflora f. lutea (F.
Haage & k. Schmidt) F. F. A i miner, Pflanzenr.
IV.276b: 682. 1953. TYPE: not located. Name based
on yellow-flowered plants raised from normally pig¬
mented progenitors at a commercial nursery in Er¬
furt. Germany.

lobelia nelsonii var. fragilis R. E. Robinson & Fernald,
Proc. Amer. Acad. Arts 43: 27. 1907. Lobelia laxi¬
flora f. fragilis (R. E. Robinson & Fernald) F. E.
Wimmer, Pflanzenr. IV.276b: 682. 1953. TYPE:
Mexico. Morelos: Parque Station, rocky hills, 7500
ft., 13 Feb. 1907, Pringle 10360 (holotype, GH; pho¬
tograph, F; isotypes, R. RM. E, F. G[3], GOET, k.
MEXU, MIN. MO, NY, PH, US. A).
lobelia laxiflora var. brevipes F. E. Wimmer, Pflanzenr.
lV.276b: 683. 1953. TY PE: Mexico. Raja California:
Cape region, Jan.-Mar. 1901, Purpus 234 (holotype,
ATJ; isotypes, E, k, MO. US).

kKY to the Subspecies of Lobelia iaxifwra

la. Lamina 1—5 cm wide, 2-8 times longer than
wide, the base rounded, obtuse, or cuneate . . .

. L. laxiflora subsp. laxiflora

lb. Lamina 0.2— 1.4 cm wide, 12—18 times longer

than wide, the base attenuate .

. L. laxiflora subsp. angustifolia

Ebooks in Priority

While preparing the checklist of world Campan-
ulaceae, I discovered three species that were
known by illegitimate names, because an epithet
with priority had been overlooked. Although Article
14.1 permits conservation of names threatened in
this fashion, I do not consider such action appro¬
priate in these cases, as the changes proposed here
will likely prove “disadvantageous” to few if any
workers. As was the case with Campanula immo-
desta Lammers and Wahlenbergia brehmeri Lam¬
mers (Lammers, 1998a), these three species belong

Volume 9, Number 3
1999

Lammers
New Lobelia

385

to large genera and are not particularly well known.
Few botanists would recognize them off-hand or be
able to distinguish them from congeners without re¬
course to keys and comparative material. As such,
they scarcely constitute “useful exceptions,” and I
judge it best to allow the ndes of priority to operate
unhindered.

The Flora of Turkey includes a species of peren¬
nial Campanula known in the literature (e.g., Dam-
boldt, 1978; Greuter et ah, 1984; Heller & Heyn,
1993) as C. latiloba A. DC. From its description,
C. grandis Fischer & C. A. Meyer clearly refers to
the same species and has long been treated as its
synonym (Boissier, 1875; Beddome, 1907; Bailey
& Lawrence, 1953; Lewis & Lynch, 1989). The two
names were both published during 1839. However,
De Candolle’s binomial was published in volume 7,
part 2, of his lather’s Prodromus, which appeared
in late December, while Fischer and Meyer's ap¬
peared in the fifth annual Index Seminum of the
botanical garden at St. Petersburg, issued the pre¬
vious January (Stafleu & Cowan, 1976). Thus, as
pointed out by Crook (1951), Campanula grandis
had more than 10 months’ priority over C. latiloba
and is the legitimate name of the species. A new
combination is effected here so that a recently de¬
scribed subspecies has a name under the legitimate
species name:

Campanula grandis subsp. rizeensis (A. Giiner)
Lammers, comb. nov. Basionym: Campanula
latiloba subsp. rizeensis A. Giiner, Candollea
39: 348. 1984. TYPE: Turkey. Rize:

£amlihem§in gevresi, sarp kayahk, ca. 250 m,
21 May 1981. A. Giiner 3547 (holotvpe, HL B
not seen; isotype, ANK not seen).

In tht' Cordillera Central of Colombia is a robust
species of Centropogon that was called C. willden-
owianus (C. Presl) F. E. Wimmer by Wimmer
(1943) and McVaugh (1949). However, the basio¬
nym of that name was an avowed substitute for a
legitimate name, and a later homonym besides.
Therefore, it must be replaced.

This species was first described under the name
Lobelia ayavacensis Schultes in the fifth volume of
the Systema Vegetabilium, issued in December
1819. The description was based on a sheet in the
Willdenow herbarium at B, which was given the
serial number 4003. This sheet bears no informa¬
tion on locality or collector, only the binomial and
the initial “W.” for Willdenow. In the protologue,
however, Schultes indicated it was collected in
tropical America by Humboldt and Bonpland. This
sheet is a mixed collection. In the center is a fertile

stem apex belonging to the Colombian species.
Flanking it are a detached leaf (left) and flower
(right) belonging to a superficially similar but en¬
tirely different species, Siphocampylus umbellatus
(Kunth) G. Don (based on Lobelia umbellata
Kunth), endemic to the states of Minas Gerais and
Rio de Janeiro in Brazil.

The source of this extraneous material is appar¬
ently the type of L. umbellata ( Humboldt & Bonp¬
land s.n., P-Bonpl. [IDC-microfiche!]). That sheet
is likewise a mixed collection, bearing not only a
fertile stem apex of Brazilian S. umbellatus, but also
a detached flower and leaf of the Colombian spe¬
cies. Thus it appears that portions of two discrete
specimens were inadvertently interchanged. Per¬
haps this occurred when Humboldt took his collec¬
tion to Berlin to sort and arrange it with Willdenow
(cf. McVaugh, 1955). Whatever its origin, this mix¬
ing of material resulted in taxonomic confusion.
Kunth (1820, 1823) treated L. ayavacensis as a syn¬
onym of L. umbellata (which had a month’s priori¬
ty), as did Don (1834), who transferred the species
to Siphocampylus Pohl.

Presl (1836) was the first to become aware of the
mix-up and attempt to remedy it. He recognized L.
umbellata and immediately following it, a second
species, L. willdenowiana C. Presl. This name was
not accompanied by a description or diagnosis,
merely the parenthetical notation “L. ayavacensis
Willd. herb. n. 4003, specim. medium."

In light of modern practice, Presl accomplished
three things with this brief citation. First, as defined
by Article 32.5 of the Code, he provided an indirect
reference to a validly published diagnosis, that of
L. ayavacensis. Second, he designated the portion
of the original material of that name that best
matched the diagnosis to be the leetotype. Third,
he proposed the name L. willdenowiana as an
avowed substitute (nomen novum) lor L. ayavacen¬
sis. It was this last move that caused problems.
Though it was common in the nineteenth century
to simply abandon names based on mixed collec¬
tions, the Code now requires (Art. 9.10) that such
a name remain attached to that part of the type that
corresponds most nearly with the original diagnosis.
Because the portion explicitly cited by Presl better
matches the diagnosis of L. ayavacensis, he should
have retained that name. Furthermore, the new
name he proposed violates Article 53.1, as it is a
later homonym of L. willdenowiana Schultes, a syn¬
onym of Lysipomia aretioides Kunth (Wimmer,
1953). For these reasons, new combinations are
necessary for both the species proper and for the
heteronymic subspecies recognized by McVaugh
(1949):

386

Novon

Centropogon ayavacensis (Willdenow ex Schul¬
tes) Lammers, comb. nov. Basionym: Lobelia
ayavacensis Willdenow ex Schultes, in Roemer
& Schultes, Syst. Veg. 5: 37. 1819. Lobelia
willdenoiviana C. Presl, Prodr. Monogr. Lobel.
39. 1836, nom. super!!.; non Schultes, in Roe¬
mer & Schultes, Syst. Veg. 5: 634. 1819. Siph-
ocampylus umbellatus var. willdenowianus (C.
Presl) A. DC., in DC., Prodr. 7: 406. 1839.
Centropogon willdenowianus (C. Presl) F. E.
Wimmer, Repert. Spec. Nov. Regni Veg. 22:
204. 1926. TYPE: "In America merid.,” leg.
Humboldt & Bonpland, Herb. Willdenow 4003
(leetotype, selected by Presl [1836: 39], stem
portion in middle, B-W [IDC-microfiche!]).

Sipliocampylus stellalus (Reason, Hull. Toney Bot. Club
52: 67. 1925. T\ I'!,: Colombia. Tolima: between
Voleancito and Rohle. 27 Jan. 1853. Ilolton s.n. (Iio-
loty |>e, IN i ).

Centropogon ayavacensis subsp. cylindricus

(Gleason) Lammers, comb. nov. Basionym:
Siphocampylus cylindricus Gleason, Bull. Tor-
rey Bot. Club 32: 66. 1925. Centropogon cy¬
lindricus (Gleason) F. E. Wimmer. Pflanzenr.
I \ .2 7 61 1 : 257 (1943). Centropogon willdenow¬
ianus subsp. cylindricus (Gleason) McVaugh,
Brittonia 6: 480. 1949. TYPE: Colombia. To¬
lima: Rosalito to Murillo, Pennell 3142 (holo-
type, NY).

In southern and central Africa is a small short¬
lived species of Lobelia that was called L. depressa
L. f. by Wimmer (1953). However, the type of that
name (Linne filius 11177. S-LINN not seen) is not
referable to the genus Lobelia but rather to Mon-
opsis Salisbury. Timlin (1983, 1984) identified the
specimen as 47. simplex (L.) E. E. Wimmer, while
Phillipson (1986) treated it as 47. debilis (I,, f.) C.
Presl var. depressa (L. f.) Phillipson.

The earliest name actually referable to this Lo¬
belia was Mezleria depressa A. DC. However, that
name was illegitimate, as it was a later homonym
of 47. depressa (L. f.) C. Presl. Though De Candolle
had cited Presl as the authority for the binomial,
he specifically excluded its basionym, L. depressa,
making it a new name based on its own type (Art.
48.1). The earliest legitimate name referable to this
Lobelia is Mezleria dregeana Sonder, which had
been proposed as an explicit replacement for ille¬
gitimate 47. depressa. However, use of its epithet in
Lobelia is precluded by the existence of L. dre¬
geana (C. Presl) A. DC. Finally, L sonderi Zahl-
bruckner was published so that 47. dregeana would
have a name in the genus Lobelia.

Timlin (1983, 1984) enlarged the circumscrip¬
tion of this species considerably. Among the many
new synonyms were several that had priority over
L. sonderi, the earliest among them L. angolensis
and L. lythroides. Thulin took up the former for the
enlarged species.

In doing so, he (like Zahlbruckner before him)
unfortunately overlooked an earlier legitimate epi¬
thet. When Kuntze took up the generic name Dort-
tnanna Hill for the species then assigned to Lobe¬
lia, he published D. sonderiana Kuntze as an
avowed substitute for 47. dregeana. This replace¬
ment name was necessitated by his simultaneous
validation of D. dregeana (C. Presl) Kuntze for L.
dregeana. Thus, sonderiana is the earliest epithet
referable to this species, and the new combination
required by the Code is effected here:

Lobelia sonderiana (Kuntze) Lammers, comb,
nov. Basionym: Dortmanna sonderiana Kuntze,
Revis. Gen. PL 2: 972. 1891. Replaced name:
Mezleria dregeana Sonder, in Harvey & Sen¬
der. 11. Cap. 3: 533. 1865; non Dortmanna
dregeana (C. Presl) Kuntze, Revis. Gen. PL 2:
972. 1891; nee Lobelia dregeana ((7 Presl) A.
DC., in I)C.. Prodr. 7: 731. 1839. Replaced
name: Mezleria depressa A. DC., in DC., Prodr.
7: 350. 1839; non Mezleria depressa (L. 1.) C.
Presl, Prodr. Monogr. Lobel. 7. 1836; nee Lo¬
belia depressa L. f., Suppl. PL 395. 1782. Lo¬
belia sonderi Zahlbruckner, Ann. K. K. Na-
turhist. Hofmus. 18: 404. 1903, nom. superfl.
Lobelia depressa var. dregeana (Sonder) F. E.
Wimmer, Notizbl. Bot. Cart. Berlin-Dahlem
15: 633. 1941. TYPE: South Alriea. Cape
Province: Witbergen, Drege s.n . (holotype, G-
DC [IDC-microfiche!]).

hibelm lythroides Diels. Hot. Jahrh. Syst. 25: 113. 1899.
IS PL: South Africa. Transvaal: Pretoria, Wilms 333
(syntype, li not seen) and Wilms 336 (syntype, 8 not
seen).

lobelia angolensis Engler & Diels, Bot. Jahrh. Syst. 26:
114. 1899. TAPE: Angola. Huila, Welwitsch 1146
(syntype, R not seen) and Anilines 94 (syntype, B not
seen).

[Additional heterotypic synonyms cited by Thu¬
lin (1983, 1984). |

As an aside, it should be noted that L. depressa
is the type of Mezleria C. Presl (Pfeiffer, 1874: 298).
This generic name typically has been used for a
subgenus of Lobelia (e.g., Wimmer, 1953; Murata,
1995), but can no longer serve that purpose be¬
cause, as noted above, the type of L. depressa is
referable to tin1 genus Monopsis. As a result, Mezle¬
ria is a synonym of Monopsis and cannot be used

Volume 9, Number 3
1999

Lammers
New Lobelia

387

for any taxon in Lobelia. When circumscribed fol¬
lowing Wimmer (1953), the subgenus has no name
available; under the broader circumscription of
Murata (1995), it may be called Lobelia subg. Iso-
lobus (A. DC.) Y. S. Lian.

New Assignments

Continued work on the checklist disclosed sev¬
eral more species (or subspecies) that do not have
a legitimate name in the genus (or species) to which
they should be assigned.

Recent studies in my lab (cl. Morris & Lammers,
1997a, 1997b) have supported Grey-Wilson’s
(1990) suggestion that Leptocodon (Hooker f.) Le-
maire be subsumed into Codonopsis. Unpublished
cladistic analyses of morphological and palynolog-
ical data by Morris indicate that the two species
that comprise Leptocodon are embedded well within
the structure of Codonopsis. The type of Leptocodon
already has a name available in Codonopsis. Here,
a new combination is proposed for the other spe¬
cies, to complete the merger ol the two genera:

Codonopsis hirsuta (D. Y. Hong) K. E. Morris &
Lammers, comb. nov. Basionym: Leptocodon
hirsutus D. Y. Hong, Acta Phytotax. Sin. 18:
246. 1980. TYPE: China. Xizang: Zayu, Shang
Zayu Zhong Xiang A Zha, mixed forest on
slope, 2500 m, Qinghai-Xizang Exped. Team
73-1014 (holotype, PE not seen).

These same studies also support the recognition
of Campanumoea Blume sect. Cyclocodon (Griffith)
C. B. Clarke as a distinct genus, a conclusion
reached independently by Hong (1998). In order to
preserve the subspecific classification proposed by
Moeliono (1960), a new combination under Cyclo-
codon is required:

Cyclocodon lancifolius subsp. celebicus (Blume)
K. E. Morris & Lammers, comb. nov. Basio¬
nym: Campanumoea celebica Blume, Bijdr.
727. 1825. Campanula celebica (Blume) D.
Dietrich, Syn. PL 1: 758. 1839. Codonopsis ce¬
lebica (Blume) Miquel, FI. Ned. Ind. 2: 566.
1857. Codonopsis lancifolia subsp. celebica
(Blume) Moeliono, in Steenis, FI. Males, (ser.
1) 6(1): 121. 1960. Cyclocodon celebicus (Blu¬
me) D. Y. Hong, Acta Phytotax. Sin. 36: 109.
1998. TYPE: Indonesia. Sulawesi: sine loc.,
Reinwardt s.n. (holotype, L not seen).

Codonopsis leucoearpa Miquel, FI. Ned. Ind. 2: 565. 1857.
TYPE: Indonesia. Sumatra: in de kloof van den Sin-
galang, Teijsmann s.n. (holotype, U not seen).

As noted previously (Lammers, 1998a), my treat¬
ment of the Hawaiian Lobelioideae (Lammers,
1990) for the Manual of the Flowering Plants of
Hawaii occasionally synonymized taxa that sub¬
sequent studies suggest should be recognized at
some level. Case in point: Cyanea baldwinii C. N.
Forbes & G. C. Munro from Lana‘i. Impressed by
the overall similarity of its leaves and flowers to
those of C. lobata H. Mann from northern West
Maui, I treated the former as a mere synonym of
the latter. More detailed study, however, revealed
characters by which the two could be distinguished,
and it is consistent with my practice elsewhere to
treat them as allopatric subspecies:

Cyanea lobata subsp. baldwinii (C. N. Forbes &
G. C. Munro) Lammers, comb, et stat. nov.
Basionym: Cyanea baldwinii C. N. Forbes &
G. C. Munro, Occas. Pap. Bernice Pauahi
Bishop Mus. 7: 43. 1920. Delissea baldwinii
(C. N. Forbes & G. C. Munro) H. St. John,
Phytologia 63: 81. 1987. TYPE: Hawaiian Is¬
lands. Lana‘i: Lana‘ihale, but one tree seen,
near top of ridge close to trail, just before as¬
cending the strip part of trail, growing amongst
thick scrub, 3000 ft., Sep. 1919, Munro 674
(holotype, BISH; isotypes, BISH[2], NSW, NY,
UC).

Key to the Subspecies of Cyanea lobata

la. Petiole V6— Va as long as the lamina; calyx lobes

10—15 mm long, 2.4—4 mm wide, narrowly ob¬
long or narrowly triangular; dorsal anthers 12-
13.5 mm long . C. lobata subsp. baldwinii

lb. Petiole V3—V2 as long as the lamina; calyx lobes

16-22 mm long. 3—7.5 mm wide, elliptic or ob¬
long; dorsal anthers 9—11.5 mm long .

. C. lobata subsp. lobata

Previously (Lammers, 1988, 1990), I treated all
populations of Delissea with a palm-like habit as a
single species, D. undulata Gaudichaud, divided
into three allopatric subspecies: subsp. niihauensis
(H. St. John) Lammers from the westernmost island,
Ni‘ihau; subsp. kauaiensis Lammers from nearby
Kaua‘i; and subsp. undulata from the easternmost
islands of West Maui and Hawaii. This classifica¬
tion was not consistent with my treatment of similar
complexes among the Hawaiian Lobelioideae, as
the morphological gap separating the two western
subspecies from the eastern subspecies was quite
wide when compared to that which usually sepa¬
rates conspecific subspecies. Furthermore, in no
other instance have I recognized conspecific sub¬
species that are separated geographically by several
islands. For these reasons, I now recognize two spe-

388

Novon

t ies of palmiform Delissea: D. undulata on Hawai'i,
and D. niihauensis H. St. John with allopatric sub¬
species on Ni‘ihau and Kaua‘i. A new combination
is needed for the last:

Delissea iiiihauensis subsp. kanaiensis (Lam-
mers) Lammers, comb. nov. Basionym: Delis¬
sea undulata subsp. kauaiensis Lammers, Syst.
Bot. 13: 505. 1988. TYPE: Hawaiian Islands.
Kaua‘i: on the Hanapepe and Wahiawa water¬
shed. 24 June 1895, Heller 2430 (holotype,
MSC; isotypes, A. AC. BM, E. F[2], G, GH. K.
MIN, NY, P. UC, US).

Kky to Pai mikorm Deussha

la. Lamina elliptic or lanceolate, 3-4 times longer
than wide, the base euneate, tin1 margin undulate
and denticulate; peduncles 15-50 mm long (Ha¬
wai'i) . I). undulata

lh. Lamina ovate. 1.3—2 times longer than wide, the
base cordate or truncate, rarely euneate, the mar¬
gin flat and crenate or serrate; peduncles 5—20

mm long . I). niihauensis

2a. Lamina 5—1 I cm long, 3.5—7 cm wide, base
cordate, apex acute, margin callose-crenate
(Ni'ihau) ... I). niihauensis subsp. niihauensis
2b. Lamina (7-) LI— 18 cm long, (3 — )7 — 10 cm
wide, base truncate or rarely euneate, apex
acuminate, margin coarsely eallose-serrate
(kaua'i) . I). niihauensis subsp. kauaiensis

Acknowledgments. I thank my students, Mi¬
chael James Eakes and Kari Ellen Morris, lor per¬
mission to publish their nomenelatural innovations
here; Field Museum Scientific Illustrator Zorica
Dabich lor skillfully executing Figure 1; Fred Bar¬
rie (MO) lor reviewing the manuscript prior to sub¬
mission; and the administrators and staffs of the
many herbaria from which specimens were seen: A,
AC, ARIZ, B, BISH, BM, CAS. ENCB, E, G, GH.
GOET, k, MEXU, MICH, MIN, MO. MSC, NSW.
NY, P, PH, TEX, UC, US, and W.

Literature Cited

Hailey, L. II. & C. II. M. Lawrence. 1953. The Garden of
Bellflowers in North America. MacMillan. New York.
Beddome, R. H. 1907. An annotated list of the species of
Campanula. J. Roy. Hurt. Soc. 32: 196-221.

Boissier, E. 1875. Flora Orientalis, Vol. 3. H. Georg. Ge¬
neva.

Crook, H. C. 1951. Campanulas: Their Cultivation and
Classification. Country Life. London.

Damboldt. J. 1978. Campanulaceae. Pp. 2-89 in I’. Davis
(editor). Flora of Turkey and the East Aegean Islands,
Vol. 6. Univ. Press, Edinburgh.

Don, G. 1834. A ( General History of the Dichlamydeous
Plants, Vol. 3. J. G. & K Rivington, London.

Eakes, VI. J. <X T. G. Lammers. 1996. Phenetic analysis
of the Ixibelia laxiflora complex (Campanulaceae: Lob-

elioideae) reveals a previously undescribed species in
Mexico Amer. J. Rot. 83 (6, Suppl.): 154. | Abstract. |

Grenier. W., F. R. Barrie. H. M. Runlet. W. G. Chaloner,
V. Demoulin, D. L. Hawksworth, P. M. Jprgensen, I). H.
Nicolson, P. C. Silva. P. Trehane & J. McNeill. 1994.
International Code ol Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

- . H. M. Runlet & G. Long. 1984. Med-Checklist,

Vol. 1 . ( Conservatoire et Jardin Rotaniques, Geneva.

Grey-W ilson, C. 1 990. A survey of Codonopsis in culti¬
vation. Plantsman 12: 65—99.

Hamlin, J. 1995. Sophomore’s discovery. In the Field
66(5): 3.

Heller. D. & (.. (,. Heyn. 1993. Conspectus Florae Orien¬
talis, Vol. 8. Israel Academy of Sciences and Humani¬
ties, Jerusalem.

Hong, I). Y. 1998. The restoration of the genus Cyclocodon
(Campanulaceae) and its evidence from pollen and
seed-coat. Acta Phytotax. Sin. 36: I (K> — 1 It).

Kunth. C. S. 1820. Index plant arum nostrarum, a Leh-
manno, Romerio et Sclmltesio sub aliis nominibus vul-
gatarum. Pp. 451— 456 in A. Humboldt, A. Ronpland &
C. S. Kunth, Nova Genera el Species Plantarum (quarto
ed.). Librariae Graeco-Latini-Germanicae, Paris.

- . 1823. Synopsis Plantarum, Vol. 2. F. G. I.evrault,

Paris.

Lammers, 1. G. 1988. New laxa, new names, and new
combinations in the Hawaiian Lobelioideae (Campan¬
ulaceae). Syst. Rot. 13: 496-508.

- . 1990. Campanulaceae. Pp. 420-489 in W. L.

Wagner. I). R. Herbst & S. H. Sohmer, Manual of the
Flowering Plants of Hawai'i. Univ. Hawaii Press and
Bishop Museum Press, Honolulu.

- . 1995. Transfer of the southern African species of

Lightfootia, nom. illeg., to Wahlenbergia (Campanula¬
ceae, (Campanu loideae). Taxon 44: 333—339.

- . 1998a. New names and new combinations in

(Campanulaceae. Novon 8: 31—35.

- . 1998b. Review ol the neotropical endemics Bur-

meistera , Centropogon, and Siphocampylus (Campanu¬
laceae: Lobelioideae), with description of 18 new spe¬
cies and a new section. Brittonia 50: 233-262.

- . 1999. Nomenelatural consequences of the syn-

onyinization of Hypsela renijormis (Campanulaceae:
Lobelioideae). Novon 9: 73—76.

Lewis. P. & M. Lynch. 1989. Campanulas. Christopher
Helm. London.

McVaugh, R. 1943. Campanulaceae (Lobelioideae). N.
Amer. FI. 32 A: 1-134.

- . 1949. Studies in South American Lobelioideae

(Campanulaceae) with special reference to Colombian
species. Brittonia 6: 450—493.

- . 1955. rhe American collections of Humboldt and

Ronpland, as described in the Systema V egetabiliurn of
Roemer and Schultes. Taxon 4: 78—86.

Moeliono, R. 1960. Codonopsis. Pp. 118—121 in C. G. G.
J. van Steenis (editor). Flora Malesiana (ser. 1), Vol.
6(1). Noordhoff-Kolff. Djakarta.

Morris, K. E. & T. G. Lammers. 1997a. Circumscription
of Codonopsis and the allied genera Campanumoea and
Leptocodon (Campanulaceae: Campanuloideae). I. Pal-
ynological data. Rot. Bull. Acad. Sin. 38: 277-284.

- & - . 1997b. Circumscription of Codonopsis

(Campanulaceae: Campanuloideae) and allied genera.
Amer. J. Rot. 84 (6. Suppl.): 218. [Abstract.]

Murata, J. 1995. A revision of infrageneric classification
of Ixibelia (Campanulaceae-Lobelioideae) with special

Volume 9, Number 3
1999

Lammers
New Lobelia

389

reference to seed coat morphology. J. Fac. Sci. Univ.
Tokyo (sect. 3) 15: 349—371.

Pfeiffer. L. 1874. Nomenclator Botanicus, Vol. 2. T. Fi¬
scher. Kassel.

Phillipson, P. B. 1986. Taxonomy of Monopsis (Lobeli-
aceae): M. simplex, M. debilis and a new species. Bot.
J. Finn. Soc. 93: 329-341.

Presl, C. 1836. Prodromus Monographiae Lobeliacearum.
Theophilus Haase, Prague.

Stafleu, F. A. & R. S. Cowan. 1976. Taxonomic Literature,
2nd ed., Vol. 1. Bohn. Seheltema & Holkema, Utrecht.

Thulin. M. 1983. Lobeliaceae. Pp. 1 16-157 in E. Launert
(editor). Flora Zambesiaea, Vol. 7(1). Flora Zambesiaca
Managing Committee, London.

- . 1984. Lobeliaceae. In R. M. Polhill (editor). Flo¬
ra of Tropical East Africa. A. A. Balkema, Rotterdam.

Wimmer, F. E. 1943. Campanulaceae-Lobelioideae, l.
Teil. Pp. i-viii, 1—260 in R. Mansfeld (ed.). Das Pflan-
zenreich, IV.276b. Akademie-Verlag, Berlin.

- . 1953. Campanulaceae-Lobelioideae, 11. Teil. Pp.

i-viii. 261—814 in R. Mansfeld (editor). Das Pflanzen-
reieh, !V.276b. Akademie-Verlag, Berlin.

Seven New Species of Pilea Lindley (Urticaceae)
from Mesoameriea

A. K. Monro

Department of Botany, Natural History Museum, Cromwell Road,
London SWT 5BD, United Kingdom

ABSTRACT. Seven new species of Piled from Me-
soamerica are described and illustrated: Pilea pter-
idophylla A. K. Monro, P. plumulosa A. K. Monro,
P tripartita A. K. Monro, P. rostulata A. K. Monro,
P. quadrata A. K. Monro, P tutensis A. K. Monro,
and P. magnicarpa A. K. Monro. Their affinities are
discussed and positions within Weddell’s and Kil-
lip's subdivisions of the genus indicated.

Pilea is a genus mainly composed of small
shade-loving herbs and is distributed throughout
the tropics, subtropics, and temperate regions (with
the exception of Australia. New Zealand, and Eu¬
rope), with the greatest diversity in the Greater An¬
tilles. Pilea can be distinguished from other Me-
soamerican Urticaceae by opposite leaves and a
single ligulate intrapetiolar stipule in each axil.
Cystoliths, found on the stem, leaves, and frequent¬
ly inflorescence, vary greatly within and between
species, but are distinctive in frequently appearing
branched, forming “V,” “Y,” and “X” shapes.

The last major worldwide taxonomic treatment of
the genus Pilea was H. A. Weddell’s treatment of
1869, in which 159 species were recognized (24 of
which were described as new). Since then, over 500
new species names have been published, and it is
likely that there are over 500 to 600 “good” species
worldwide (Adams, 1970; Burger, 1977). Major
contributions to the taxonomy of neotropical Pilea
have been made by E. P. Killip in his revision of
the Andean species (Killip, 1936, 1939), P. C.
Standley and J. A. Steyermark (1952) in their treat¬
ment for the Flora of Guatemala , G. D. Adams
(1972) in his treatment for the Flowering Plants of
Jamiaca, and W. C. Burger (1977) in his treatment
for Flora Costaricensis. Weddell (1869) subdivided
the genus, on the basis of leaf morphology, into
three principal species groups: Integrifoliae, with
leaves of equal length at each node and entire mar¬
gins; Heterophyllae, with leaves of unequal length
at each node; and Dentatae, with leaves of equal
length at each node and toothed margins. Killip
(1936) subdivided Pilea into 12 species groups,
based largely on those of Weddell’s conspecti of

Novon 9: 390-400. 1999.

1856 and 1869. In the absence of recent phyloge¬
netic analysis, the new species described here are
best placed in three of the subgeneric groups cited
in ibis article: Fallaces, Molles, and Centradeno-
ideae. These three subgeneric groups can be rec¬
ognized on the following suites of characters: sub¬
genus Fallaces with toothed pinnately nerved
leaves, where no pair of the secondary nerves are
more prominent than the others; subgenus Molles
with pubescent, toothed, 3-nerved or pinnately
nerved leaves (where pinnately nerved a basal pair
of secondary veins is markedly more prominent
than the others), of equal size at each node, aver¬
aging more than 2 cm in length and compact sta-
minate inflorescences where the peduncles are lon¬
ger than the panicle branches; subgenus
Centradenoideae with 3-nerved or pinnately nerved
leaves (where pinnately nerved a basal pair of sec¬
ondary veins is much more prominent than the oth¬
ers), of unequal size at each node, the smaller leaf
at each node symmetrical, with stipules less than
3 cm long and pedunculate pistillate inflorescenc¬
es.

In preparing revisionary accounts of Pilea for
Flora Mesoamericana, over 1700 collections from
Central and South America were examined. During
the course of this work, the following seven new
species were identified.

Pilea pteridophylla A. K. Monro, sp. nov. TYPE:
Mexico. Tabasco: Sierra el Madrigal, Teapa,
Tabasco, a 600 m al E del edifieio principal
del Centro Regional Tropical Puyacatengo de
la Universidad Autonoma de Chapingo, 5 June
1991, A. M. Hanan A. 438 (holotype, MEXU).
Figure 1A, B.

Species foliis pectinatis a congeneribus diversa.

Herb to 35 cm, perennial; epipetric, monoecious.
Stems erect, sparsely branched, rooting at the base;
internodes 2—8 X 1.5— 2.5 mm, weakly striate, dry¬
ing dark green, glabrous, cystoliths not visible.
Blades of leaves at the same node of equal length,
petiolate, 45—90 X 7—12 mm, linear-lanceolate.

Volume 9, Number 3
1999

Monro

New Pilea from Mesoamerica

391

Figure 1. A. B. Pilea pteridophylla A. k. Monro. — A. Habit, plant with pistillate inflorescences. — B. Fruit. C. 1).
Pilea plumulosa A. k. Monro. — C. Habit, plant with infructescences. — D. Staminate llower prior to anthesis. A and
B based on Hanan 438. C and I) based on Kirkbride & Duke 944.

pectinate above the basal Vs to '/,,, chartaceous; ad-
axial surface drying dark green, glabrous, cystoliths
fusiform, V-, Y-, and X-shaped; abaxial surface
drying dark gray-green, glabrous, cystoliths fusi¬

form, V-, Y-, and X-shaped, punctate glandular;
base asymmetrical, decurrent, attenuate, rounded;
apex symmetrical, weakly cuspidate; primary ve¬
nation pinnate, secondary veins 19—28 pairs, 75—

392

Novon

90° to the midrib; petioles equal, 0.8— 1.3 X ea. 0.8
mm, glabrous; stipules 0. 8-1.0 X 1.0-1. 3 mm, del-
tate, membranous, gray, persistent. Inflorescences
7-24 per stem, unisexual, pistillate inflorescences
preceding and concurrent with staminate inflores¬
cences; peduncle and pedicels subtended by sti-
puliform bracts, peduncular bract 0.5-0. 8 mm, del-
tate, pedicellular bracts ea. 0.4 mm, narrow debate;
intact staminate inflorescences not seen; flowers in
bud immediately prior to anthesis 1.0— 1.3 X 0.8-
1.3 mm, cream, apically green; tepals 4, ca. 1.3
mm, ovate, glabrous; subapical appendages ca. 0.3
mm, ridge-like, glabrous; stamens 4, filaments ca.
1.3 mm, anthers ca. 0.4 X ca. 0.5 mm; pistillate
inflorescences I or 2 per axil, 1.5— 2.0 mm, bearing
5—35 flowers in a loose panicle, peduncle Va to Vi
inflorescence length, ca. 0.4 mm diam., glabrous;
pedicels ca. 0.3 X ca. 0. 1 mm to inconspicuous,
glabrous; flowers 0. 5-0.7 X ca. 0.3 mm, glabrous;
tepals 3, unequal, the central one ca. 0.5 mm, ob¬
long with a dorsal thickening, the two lateral ca.
0.5 mm, ovate, lacking a dorsal thickening. Infruc-
tescences ca. 2—3 mm, peduncle 0.5 mm diam.; fruit
0. 7-0.8 X ca. 0.4 mm, asymmetrical, subcom¬
pressed ovoid, orange-brown, the margin broaden¬
ing and flattened toward the apex.

Distribution. Known only from the type collec¬
tion, growing in evergreen forest in Tabasco, south¬
ern Mexico.

Discussion. This species falls within Weddell’s
Dentatae and Killip’s Fallaces species groups. It is
unmistakable on account of its pectinate leaves.
These resemble the fronds of a fern in color and
texture as well as shape (Fig. 1A). It is the only
species of Piled with pectinate leaves known from
the Americas.

Pilea plumidosa A. K. Monro, sp. nov. TYPE:
Panama. Bocas del Toro-Chiriquf border: elfin
forest at continental divide on Chiriquisito—
Caldera trail, 20 Apr. 1968, J. //. Kirkbride &
J. A. Duke 944 (holotype, MO (sheet no.
2605434]; isotypes, MO [sheet no. 2090983],
NY). Figure 1C, D.

Species P. killipianae Staiulley & Steyermark affinis,
sed floribus staminaliluis minimis, seminibus majoribus,
differt.

Shrublet to 25 cm, perennial; epipetrie, monoe¬
cious. Stems erect, branched, rooting at the base;
internodes 3.0-10.5 X 1-2 mm, weakly striate,
drying gray-brown, glabrous, eystoliths elliptic, be¬
coming obscure with age. Blades of leaves at the
same node of equal length, petiolate, 12-38 X 1.5-
4.0 mm, linear-lanceolate, chartaeeous; adaxial

surface drying dark brown-green, glabrous, cysto-
liths fusiform; abaxial surface drying yellow-green,
glabrous, eystoliths orbicular, punctate glandular;
base asymmetrical, subcordate and euneate, acute;
margin crenate with basal Va to Vi entire; apex sym¬
metrical, acute; primary venation pinnate, occa¬
sionally appearing 3-nerved, where appearing 3-
nerved then lateral nerves much less prominent
than the midrib, running Vi to Va the length of the
leal, secondary veins 4-13 pairs, 30-45° to the
midrib; petioles unequal in pair by ratio 1:2. 0-4. 5,
the longer 2.3— 4.5 X 0.5— 0.8 mm, the shorter 1.0—
1.8 X 0.5 mm, glabrous; stipules 0. 5-0.8 X 0.5
mm, ovate, membranous, brown, persistent. Inflo¬
rescences 6—28 per stem, unisexual, pistillate inflo¬
rescences preceding and concurrent with staminate
inflorescences; peduncle and pedicels subtended
by stipuliform bracts, peduncular bract ca. 0.5 mm,
deltate, ovate, pedicellular bracts 0.4— 0.8 mm,
ovate; staminate inflorescences 1 or 2 per axil, 3.5-
7.0 mm, bearing 6-10 flowers in a single compact
head; peduncle Vi to % inflorescence length, ca. 0.3
mm diam., glabrous; pedicels 0.8-1. 3 X ca. 0.1
mm, glabrous; flowers in bud immediately prior to
anthesis 1.0— 1.3 X 1.0— 1.3 mm, cream; tepals 4,
ca. 1.5 mm, ovate, elliptic, glabrous; subapical ap¬
pendages ca. 0.3 mm, ridge-like, glabrous; stamens
4, filaments ca. 2 mm, anthers ca. 0.5 X ca. 0.5
mm; pistillate inflorescences 1 or 2 per axil. ca. 3
mm, bearing 5-8 flowers in a single compact head,
peduncle Vi inflorescence length, ca. 0.1 mm diam.,
glabrous; pedicels 0.3— 0.5 X ca. 0.1 mm, glabrous;
flowers ca. 0.8 X ca. 0.5 mm, glabrous; tepals 3,
unequal, the central one ca. 0.8 mm, oblong-oh-
ovate, with a dorsal thickening, the lateral two ca.

O. 7 mm, ovate, without a dorsal thickening. Infruc-
tescences 3. 0-4. 5 mm, peduncle ca. 0.1 mm diam.;
fruit ca. 1.5 X ca. 1.0 mm, asymmetrical, subcom¬
pressed elliptic, pale brown, the margin narrow,
thickened.

Distribution. Known only from the type collec¬
tion made in the western end of the Cordillera Cen¬
tral in Panama, where it was growing on rocky out¬
crops in elfin forest.

Discussion. This species falls within Weddell's
Dentatae and Killip’s Fallaces species groups. A
sterile collection from Chiriquf, Dwyer & Lallathin
H769 (MO), collected at 4500—5500 m, resembles

P. plumulosa. Pilea plumulosa most closely resem¬
bles P. killipiana from Chiapas (Mexico) and Gua¬
temala, differing as summarized below:

Pilea killipiana: sparsely branched or
unbranched, the leaf-bearing part of the stem her-

Volume 9, Number 3
1999

Monro

New Pilea from Mesoamerica

393

baceous; staminate flowers in bud immediately pri¬
or to anthesis 1.8— 2.0 mm; fruit ca. 0.8 mm.

Pilea plumulosa: extensively branched, the leaf¬
bearing part of the stem woody; staminate flowers
in bud immediately prior to anthesis 1.0— 1.3 mm;
fruit ca. 1.5 mm.

From the available collections, the distributions
of the two species do not appear to overlap: P. kil-
lipiana is known only from Chiapas in southern
Mexico and the states of the Peten and Alta Vera-
paz in northern Guatemala, while P. plumulosa is
known only from western Panama.

This species is named for the feather-like outline
of the leaves.

Pilea tripartita A. K. Monro, sp. nov. TYPE: Costa
Rica. San Jose-Cartago border: summit of the
Interamerican Highway at 3200-3300 m alti¬
tude near La Asuncion, 21 Nov. 1969, W. C.
Burger & R. L. Liesner 6330 (holotype, MO;
isotypes, NY, US). Figure 2A, B.

Species P. cornuto-cucullatae Cufodontis similis, sed
floribus staminalibus tripartitis, seminibus majoribus, dif-
fert.

Herb to 25 cm, perennial; terrestrial, epiphytic
or epipetric; monoecious. Stem erect or repent,
sparsely branched, rooting at base and adventi¬
tiously where repent; internodes 12—60 X 1.5— 2.0
mm, weakly striate, drying green to pale brown,
pubescent, the hairs to 0.8 mm, upright or weakly
appressed, curved, cystoliths fusiform. Blades of
leaves at the same node of equal length, petiolate,
12—40 X 10—20 mm, elliptic to ovate, chartaceous;
adaxial surface drying dark green, glabrous to
sparsely pubescent on the midrib, where pubescent
the hairs to 0.8 mm, weakly appressed, curved, cys¬
toliths fusiform; abaxial surface drying olive green,
glabrous to pubescent on the primary and second¬
ary veins, where pubescent the hairs to 0.8 mm,
appressed, curved, cystoliths fusiform or minute
pusticulate, eglandular; base asymmetrical, weakly
decurrent to cuneate; margin serrate; apex acute;
primary venation 3-nerved, the two lateral nerves
stopping short of the leal apex, secondary veins 4—
6 pairs, 30—45° to the midrib; petioles equal or un¬
equal in pair by ratio of 1:1.5— 3.0, 10^40 X 0.5
mm, pubescent, the hairs to 0.8 mm, weakly ap¬
pressed, weakly curved; stipules 4—6 X 2-3 mm,
oblong, obovate, membranous, red-brown, persis¬
tent. Inflorescences 1—6 per stem, unisexual, pistil¬
late inflorescences preceding staminate inflores¬
cences; peduncle and pedicels subtended by a
stipuliform bract, peduncular bract 1.0— 1.3 mm,
ovate, pedicellular bracts 0.8— 1.3 mm, deflate; sta¬

minate inflorescence 1 per axil, 5—45 mm, bearing
5—9 flowers in a single compact head; peduncle %
to % inflorescence length, 0.5 mm diam., glabrous;
pedicels 1-2 X 0.5 mm, glabrous; flowers in bud
immediately prior to anthesis 2. 5-3. 5 X 1. 5-2.0
mm; tepals 1.0— 1.5 mm, 3, ovate, glabrous; sub-
apical appendages 1. 5-3.0 mm, narrow ovate, gla¬
brous; stamens 3, filaments 2.0— 2.5 mm, anthers

O. 5 X 1.0 mm; pistillate inflorescences 1 or 2 per
axil, 10—25 mm, bearing 8—20 flowers in a single
compact head; peduncle % to % inflorescence
length, 0.5 mm diam., glabrous; pedicels 0.5— 0.8
X 0.5 mm, glabrous; flowers 1.0— 1.5 X 0.8— 1.0
mm, glabrous; tepals 3, unequal, the central one

l. 0-1. 5 mm, ovate, the lateral two 0.8 mm, lance¬
olate. Infructescences 12—50 mm, peduncle 0.5 mm
diam.; fruit 2-3 X 1.5— 2.0 mm, asymmetrical, com¬
pressed ovoid, glabrous, the margin narrow, thick¬
ened.

Distribution. Central Costa Rica at 2500—3200

m, in forest on, or close to riverbanks and by the
side of forest tracks, growing on tree trunks and
rocks.

Discussion. This species falls within Weddell’s
Dentatae and Killip's Molles species groups. Burger
(1977: 254) equated this species with Pilea cor-
nuto-cucullata, another Costa Rican species. Al¬
though no representative collections are cited in
Flora Costaricensis, collections determined by Bur¬
ger as P. cornuto-cucullata constitute the type col¬
lection for P. tripartita as well as one of the para-
type collections ( Burger & Liesner 6509).

Pilea tripartita A. K. Monro most closely resem¬
bles P. cornuto-cucullata , differing as summarized
below:

P. cornuto-cucullata: staminate flowers 4-parted,
subapical appendage 0.5— 1.0 mm; fruit ca. 1.5 mm.

P. tripartita: staminate flowers 3-parted, subapi¬
cal appendage 1.5— 3.0 mm; fruit 2-3 mm.

To a lesser extent, Pilea tripartita also resembles

P. fallax Weddell from western Venezuela, Colom¬
bia, and Ecuador, distinguished as outlined below:

P. fallax: stipules auriculate, gray to gray-brown
in color; leaves pinnately nerved (no secondary
nerves more prominent than the others); pistillate
inflorescences bearing 3—7 flowers.

P. tripartita: stipules oblong to obovate, red-
brown in color; leaves 3-plinerved; pistillate inflo¬
rescences bearing 8—20 flowers.

Paratypes. COSTA K1CA. Limon: canton de Tala-
manca, sendero entre Cerro Durika y sabanas de Durika,
siguiendo el cauce de la quebrada intermitente,
9°24'30"N, 83°18'35"W. 25(H) m, 21 Oct. 1989, Herrera
3752 (BM, MO not seen). San Jose-Cartago border:
about 22 km SE of Empalme, along the Interamerican

394

No von

1 mm

figure 2. A. B. Pilea tripartita A. k. Monro. — A. Habit, plant with staminate inflorescences and infructescences. —
15. Staminate flower prior to anthesis. C— K. I ‘lira rostulata A. k. Monro. — C. Habit, plant with infructescences and
stoloniferons stems. — I). Pistillate flower. — h. fruit. A and B based on Burger & Liesner 6330. C and K based on
Antonio 1237. 1) based on Mori et al. 6449.

Volume 9, Number 3
1999

Monro

New Pilea from Mesoamerica

395

Highway at 2500-2600 m, 9°40'N, 83°50'W, 27 Nov.
1969, Burger & Liesner 6 509 (F. MEXU).

Pilea rostulata A. K. Monro, sp. nov. TYPE: Pan¬
ama. Colon: along the Rio Guanche, elev. 400
m, 5 July 1979, Antonio 1237 (holotype, BM;
isotype, MO). Figure 2C.

Species P involucratae (Sims) Urban similis, sed cau-
libus stoloniferis, foliis basi cordatis, floribus pistillatis
dorsaliter appendiculatis, differt.

Herb to 10 cm, perennial; epiphytic or epipetric,
dioecious or monoecious. Stems stoloniferous from
rootstock, otherwise unbranched, erect, rooting at
base and adventitiously; internodes 2-7 X 1.0-1. 5
mm, weakly striate, drying dark brown, pubescent,
the hairs to 1 mm, upright or weakly appressed,
straight or weakly curved, cystoliths elliptic. Blades
of leaves at the same node of equal length, petiolate,
17-120 X 10—35 mm, obovate, elliptic, smaller
leaves frequently suborbicular, subehartaeeous to
chartaceous; adaxial surface drying dark green to
brown-green, pubescent, the hairs to 1.5 mm, up¬
right, weakly curved or straight, cystoliths fusiform,
Y-shaped; abaxial surface drying gray-green, pu¬
bescent as the adaxial surface, cystoliths disc¬
shaped, fusiform, V-shaped, punctate glandular;
base asymmetrical, cordate; margin serrate, basal

14 entire; apex symmetrical, obtuse to acute; pri¬
mary venation 3-plinerved, the two lateral nerves
stopping short of the leaf apex, secondary veins 6—
10 pairs, 45—65° to the midrib; petioles equal, 2—

15 X 0.8— 1.0 mm, pubescent, the hairs to 1.5 mm,
upright, weakly curved or straight; stipules 4—9 X
2.5— 3.5 mm, ovate-oblong, obovate, membranous,
pale brown to dark brown, persistent. Inflorescences
1 or 2 per stem, unisexual, rarely bisexual, pedun¬
cle and pedicels subtended by stipuliform bracts,
peduncular bract 1.0 mm, narrowly ovate, pedicel-
lular bracts 0.5— 0.8 mm, broadly ovate; staminate
inflorescences solitary, 37-55 mm, bearing 12-20
dowers in a compact head; peduncle % to % indo-
rescence length, 0.8— 1.3 mm diam., glabrous; ped¬
icels 0.8— 2.0 X 0.3 mm, glabrous; dowers in bud
immediately prior to anthesis 2.8— 3.0 X 1.3— 1.5
mm, pale green; tepals (2)4, 2.0— 2.8 mm, ovate,
glabrous; subapical appendages 1.0— 1.5 mm, nar¬
rowly ovate, glabrous; stamens (2)4, dlaments 2.0—
2.5 mm, anthers 0.5 X 1.0 mm; pistillate indores-
cences 1 or 2, 12-45 mm, bearing 100—250 dowers
in a loose panicle or 1-3 compact panicles, rarely
with a few staminate dowers; peduncle % to % in-
dorescence length, 1.0 mm diam., glabrous; pedi¬
cels 0.5— 1.0 X 0.1— 0.2 mm, glabrous; dowers 0.8—
1.0 X 0.4— 0.5 mm, glabrous; tepals 3, unequal,
glabrous, the central one 0.8— 1.0 mm, obovate, with

a dorsal subapical appendage 0.5— 1.0 mm long, the
lateral two 0.8 mm, asymmetrically ovate, lacking
any appendages. Infructescences 23—60 mm, pedun¬
cle 0.5— 1.0 mm diam.; fruit 1.3 X 0.8— 1.0 mm,
symmetrical or asymmetrical, compressed elliptic,
orange-brown to dark brown, the margin erenulate,
submarginal ridge present.

Distribution. Known only from the Caribbean
side of central Panama where it is found growing
in damp shady places on, or close to, riverbanks,
at an altitude of 0—400 m.

Discussion. This species tails within Weddells
Dentatae and Killip’s Malles species groups. Pilea
rostulata most closely resembles P. involucrata,
which is known from Costa Rica to Colombia and
Venezuela and the Windward Isles. It differs as
summarized below:

P. involucrata: terrestrial, stolons absent; leaf
margin crenate, basal Vi to Vi entire; leaf base ob¬
tuse or attenuate or decurrent; tepals of the pistil¬
late dowers lacking a subapical appendage.

P. rostulata: epiphytic or epipetric, stolons pre¬
sent; leaf margin serrate, basal Va entire; leal base
cordate; broader tepal of the pistillate dowers with
a dorsal subapical appendage ca. 1 mm in length.

Paratypes. PANAMA. Colon: near Rio Guanche, 17
July 1971. Dressier 4056 (BM, MO); Rfo Guanche, ca. 2.5
km upriver from bridge on road to Portobelo. 3 June 1975,
Mori et al. 6449 (BM. MO); Rfo Guanche, 2.5 km up¬
stream from the bridge on the road to Portobelo, 27 Aug.
1975, Mori & Witherspoon 7952a (MO); walking upstream
from bridge over Rio Guanche, 0—50 It., 27 May 1980,
Antonio 4778 (MO); walking upstream from bridge over
Rio Guanche, 0-50 ft., 27 May 1980. Antonio 4 779 (MO);
Rio Guanche above bridge on Portobelo Road. ca. 3—5 km
above bridge, 50—200 m, 8 July 1976, Croat 36941 (MO);
to Rfo Iguanita, 390 m, 7 Apr. 1977. D'Arcy 11254 (MO);
S approach to Cerro Bruja from Rfo Escandaloso, ridge
top. 20 May 1978. Hammel 3216 (MO).

Pilea quadrata A. K. Monro, sp. nov. TYPE: Pan¬
ama. Colon: along Rfo Escandaloso near mina
boqueron Numero 2, 250 m elev., 14 July
1979, T. Antonio 1345 (holotype, MO). Figure
3A-C.

Species a P ecboliophylla Donnell Smith inflorescentiis
staminalis brevibus, inflorescentiis pistillatis multo pau-
cifloris, differt.

Herb to 25 cm, perennial; terrestrial, epiphytic
or epipetric, monoecious. Stems erect, repent,
sparsely branched, rooting at base and adventi¬
tiously where repent; internodes 7—15 X 0.5— 0.8
mm, weakly striate, square in cross section, drying
gray-green, glabrous, cystoliths elliptic. Blades of
leaves at the same node of unequal length by ratio
1:4.5—13.0, the major leaves petiolate, 23—51 X

396

Novon

Figure 3. A— C. Pilea quadrata A. K. Monro. — A. Habit, plant with staminate inflorescences. — B. Staminate inflo¬
rescence. — C. Infructescence. I)-G. Pilea tutensis A. K. Monro. — I). Habit, plant with staminate inflorescence and
infructescences. — F. Infructescence. — F. Staminate inflorescence. — G. Fruit. A based on Liesner 1195. B and C
based on Antonio 1345. D-G based on Witherspoon et al. 8860.

11—18 mm, ovate, elliptic, rhomboid, subcharta-
ceous; adaxial surface drying dark gray-green, gla¬
brous, cystoliths fusiform, V-shaped, occasionally
Y-shaped; abaxial surface drying pale brown-green
or pale gray-green, glabrous, cystoliths fusiform, V-

shaped, disc-shaped, occasionally Y-shaped, punc¬
tate glandular; base asymmetrical, decurrent and
obtuse, decurrent and acute; margin serrate, basal
Vi to % entire; apex symmetrical, acute to weakly
acuminate; primary venation 3-nerved or 3-pli-

Volume 9, Number 3
1999

Monro

New Pilea from Mesoamerica

397

nerved, the two lateral nerves visible for Vi to % of
the leaf length, secondary veins 6—9 pairs, 65—80°
to the midrib, tertiary venation clearly visible; the
minor leaves subsessile to sessile, 4—5 X 2 mm,
elliptic, ovate, obovate, the leaf base strongly asym¬
metrical, decurrent and obtuse or cuneate, the leaf
apex asymmetrical, acute to obtuse, otherwise as
major leaves; petioles of the major leaves 1.5— 2.5
X 0.5-0. 8 mm, glabrous; stipules 0. 5-0.8 X 0.8
mm, ovate, membranous, brown, persistent. Inflo¬
rescences 2—30 per stem, unisexual, pistillate inflo¬
rescences concurrent with staminate inflorescences,
peduncle and pedicels subtended by stipuliform
bracts, peduncular bract 0.3— 0.5 mm, ovate, ob¬
scure, pedicellular bracts 0.3-0. 5 mm, narrowly
ovate; staminate inflorescences solitary, 1. 5-2.0
mm, bearing 5—12 flowers in a compact head, ap¬
pearing fasciculate; peduncle !4 inflorescence
length, glabrous; pedicels 0.8-1. 5 mm, glabrous;
flowers in bud immediately prior to anthesis 0.8-
1.0 X 0.8— 1.0 mm, cream; tepals 4, 1 mm, ovate,
glabrous; subapical appendages 0.3— 0.5 mm, eor-
niculate, glabrous; stamens 4, filaments 1 mm, an¬
thers 0.5 X 0.8 mm; pistillate inflorescence soli¬
tary, 1.0— 2.5 mm, bearing 5-8 flowers in a compact
head; peduncle Va to Vi inflorescence length. 0.3
mm diam., glabrous; pedicels 0.3 X 0.3 mm to in¬
conspicuous, glabrous; flowers ca. 0.5 X ea. 0.3
mm, glabrous; tepals 3, unequal, the central one
0.5 mm, ovate to oblong, the lateral two 0.3— 0.5
mm, asymmetrically ovate. Infructescences ca. 2.5
mm, peduncle ca. 0.3 mm diam.; fruit ca. 1.0 X
0.8 mm, symmetrical, compressed, elliptic;, the
margin not thickened.

Distribution. Central Panama, forest understory
and riverbanks at 200-500 m.

Discussion. This species falls within Weddell’s
Heterophyllae and Killip’s Centradenioideae species
groups. Pilea quadrata most closely resembles P.
ecboliophylla, which is found throughout Meso¬
america. They differ as summarized below:

P. ecboliophylla: stem an irregular circle in cross
section; staminate inflorescences 6—25 mm, bearing
24—60 flowers, appearing paniculate; pistillate in¬
florescences bearing 12—200 flowers.

P. quadrata: stem square to rhombic in cross sec¬
tion; staminate inflorescences 1.5— 2.0 mm, bearing
5-12 flowers, appearing fasciculate; pistillate inflo¬
rescences bearing 5—8 flowers.

Paratypes. PANAMA. Panama: along El Llano— Car-
tf — ' Tupile road. 12 mi. above Pan-American Hwy., 2(H)—
250 m, 26—27 Mar. 1973, Liesner 1195 (GH, MO. US); El
Llano to Carti Road, 13.7 km N of Pan-American High¬
way, San Bias border, 1 June 1977. Folsom 3498 (MO);
El Llano to Carti Road. 12 km N of Pan-American High¬

way at El Llano, ca. 400 m, 11 Mar. 1974, Nee 10452
(US).

Pilea tutensis A. K. Monro, sp. nov. TYPE: Pan¬
ama. Veraguas: Cerro Tute, slopes up to 4000
ft., trail past agricultural school near Santa Fe,
17 Sep. 1979, T. Antonio 1845 (holotype,
MEXU; isotypes, MO, NY). Figure 3D-G.

Species P. tilaranae W. Burger similis, sed floribus
staminalis plurimis, pedicellis florium staminalium lon-
gioribus, petiolio foliorum minorum longioris.

Herb to 35 cm, perennial; terrestrial, epiphytic,
monoecious. Stem erect, repent, sparsely branched,
rooting at base and adventitiously where repent; in¬
ternodes 6—17 X 0.8— 1.0 mm, weakly striate, dry¬
ing dark brown, glabrous; cystoliths elliptic. Blades
of leaves at a node of unequal length by ratio 1:2—
4; the major leaves petiolate, 19—30 X 11-18 mm,
elliptic, obovate, chartaceous; adaxial surface dry¬
ing dark brown to green-brown, glabrous, cystoliths
fusiform; abaxial surface drying brown, glabrous,
cystoliths fusiform, punctate glandular; base asym¬
metrical or symmetrical, cuneate, acute, weakly de¬
current; margin erenate, basal Va to V) entire; apex
symmetrical, obtuse, weakly acuminate; primary
venation 3-plinerved, the two lateral nerves visible
for % of the leaf length, secondary veins 4—10 pairs,
65—90° to the midrib; the minor leaves petiolate,
6.0— 9.5 X 3—7 mm; suborbicular to broad elliptic,
otherwise as major leaves; petioles unequal by ratio
1:3.5— 6, of the major leal 4—12 X 0.8 mm, of the
minor leaf 1—2 X 0.5 mm, glabrous; stipules 0.5—
0.8 X 0. 5-0.8 mm, broadly ovate to deflate, mem¬
branous, brown, persistent. Inflorescences 6—18 per
stem, unisexual, pistillate inflorescences preceding
or concurrent with staminate inflorescences, pedun¬
cle and pedicels subtended by stipuliform bracts,
peduncular bract 0.30—0.8 mm, ovate to narrow
ovate, pedicellular bracts 0.5 mm, narrowly deflate;
staminate inflorescences solitary, 17—22 mm, bear¬
ing 50—80 flowers in a single compact head; pe¬
duncle Va inflorescence length, 0.5 mm diam.,
glabrous; pedicels 0.5— 1.0 X 0.1 mm, glabrous;
flowers in bud immediately prior to anthesis 1.8—
2.0 X 1.0 mm, cream, apically green; tepals 4, 2
mm, ovate, glabrous; subapical appendages ca. 0.8
mm, corniculate; stamens 4, filaments ca. 1.5 mm,
anthers ca. 0.5 X ca. 1.0 mm; pistillate inflores¬
cences solitary, 2^4 mm, bearing flowers in a semi¬
compact to compact head; peduncle Vs to Vi inflo¬
rescence length, 0.1— 0.3 mm diam., glabrous;
pedicels 0.5— 0.8 mm, glabrous; flowers 1.0— 1.3 X
0. 5-0.8 mm, glabrous; tepals 3, unequal, the cen¬
tral one 0.8— 1.0 mm, oblong, the lateral two 0.5-
0.8 mm, asymmetrically ovate. Infructescences 4—15

398

Novon

Figure 4. Pilea magnicarpa A. K. Monro. — A. Habit, plant with staminate inflorescences and infructescenoes (in-
fructescences toward the base). — B. Pistillate flower. A and B based on Hammel 4014.

mm, peduncle 0.5 mm diam.; fruit 1.8— 2.5 X 1.0—
1.5 mm, asymmetrical, compressed ovoid, the mar¬
gin dorsally thickened and ventrally flattened.

Distribution. Cerro Tute of the Cordillera Cen¬
tral iti Central Panama, in forested and open areas
at 900-1600 m.

Discussion. This species falls within Weddell’s
Heterophyllae and Fillip's Centradenioideae species
groups. Pilea tutensis most closely resembles P. ti-
larana from Costa Rica, differing as indicated be¬
low:

P. tilarana : minor leaf subsessile with petioles
less than 0.3 mm; staminate inflorescences with ca.
15 flowers, staminate pedicel 1.3— 1.5 mm.

P. tutensis: minor leaf petiole 1-2 mm; staminate
inflorescences with 50^80 flowers, staminate pedi¬
cel 0.5-1. 0 mm.

Paratypes. PANAMA. Veraguas: vicinity of Escuela
Agrfcola Alto Piedra near Santa F£, 0.3 mi. beyond fork
in road near the school, toward Atlantic slope, along trail
to top of Cerro Tute. 3500-4100 ft., 30 Nov. 1979, Antonio
2976 (MO); vicinity of Fscuela Agrfcola, Alto Piedra near
Santa Fe .3 mi. beyond the fork in the road near the
school, toward Atlantic slope along trail to top of Cerro
Tute, 3200-3400 ft., 26 Jan. 1980, Antonio 3486 (MO):
along steep trail to summit of Cerro Tute ca. 3 km above
Fscuela Agrfcola Alto Piedra near Santa Fe. 3000— 3100
ft.. 4 Jan. 1 981 . Sytsma & Antonio 3059 (MO); Cerro Tute,
F slopes, I km beyond Escuela Agrfcola Alto Piedra

above Santa Fe, low cloud forest, 900-1200 m, 14 May
1981, (MO); trail up E side of Cerro Tute, to 1200 m. 25
Oct. 1975, Witherspoon et al. 8860 (MO, NY); trial on
ridge to summit of Cerro lute. Cordillera de lute, 1 km
past Escuela Agrfcola Altos de Piedras, W of Santa Ff\
8°36’N, 81°06’W, 15 Dec. 1981. Knapp A: Sytsma 2600
(MO); Cerro lute, W-siopes, forest, 850 m, 23 Oct. 1980,
Maas & Dressier 5030 (F); forested mountains W of Alto
de Piedras W of Santa Fe. 3200—5600 ft.. 8 Sep. 1978.
Hammel 4592 (MO).

Pilea magnicarpa A. K. Monro, sp. nov. TYPE:
Panama. Code: El Cope, on slope and ridge
W of sawmill, 5 Apr. 1978, B. Hammel 2424
(holotype, MO; isotype, NY). Figure 4A, B.

Species P. ecboliophyllae Donnell Smith similis, sed in-
florescentiis pistillatis longioribus, floribus pistillatis ina-
joribus, seminibus majoribus, differt.

Herb to 60 cm, perennial; terrestrial, epiphytic
or epipetric; monoecious. Stems erect, repent,
sparsely branched, rooting at base and adventi¬
tiously where repent; internodes 7-30 X 1.5— 3.0
mm, striate, drying dark brown to almost black, gla¬
brous, cystoliths pusticulate. Blades of leaves at the
same node of unequal length by ratio of 1:3-13,
the major leaves petiolate, 32-170 X 7-65 mm,
obovate, oblong, elliptic, chartaceous; adaxial sur¬
face drying green, dark green, dark brown, gla¬
brous, cystoliths fusiform; abaxial surface drying

Volume 9, Number 3
1999

Monro

New Pilea from Mesoamerica

399

brown to pale green, glabrous, eystoliths fusiform,
eglandular, occasionally punctate glandular; base
asymmetrical, acute, cuneate, decurrent, subcor-
date; margin serrate, basal Vs to % entire; apex sym¬
metrical, acuminate; primary venation 3-plinerved,
the two lateral nerves stopping short of the leal
apex, secondary veins 8—23 pairs, 80—90° to the
midrib; the minor leaves petiolate to subsessile, 7—
23 X 4.5— 7.0 mm, ovate, elliptic, or obovate, the
margin entire or apieally serrate, the apex asym¬
metrical, acute, obtuse, occasionally cuspidate, oth¬
erwise as major leaves; petioles unequal by ratio of
1:4—10, of the major leaves 2-10 X 1.0 mm, ot the
minor leaves 0.3— 1.0 X 0.3— 0.5 mm to inconspic¬
uous, glabrous; stipules 0.5— 1.5 X 1.0— 2.5 mm,
debate, membranous, brown to gray-brown, gla¬
brous, persistent. Inflorescences 3—18 per stem, uni¬
sexual, pistillate inflorescences preceding or con¬
current with staminate inflorescences, peduncle
and pedicels subtended by stipuliform bracts, pe¬
duncular bract 0.5— 1.5 mm, ovate, pedicellular
bracts 0.5— 0.8 mm, narrowly ovate; staminate inflo¬
rescences solitary, 5—35 mm, bearing 1 1-50 flowers
in a loose panicle; peduncle Vs to Vz inflorescence
length, 0.5— 0.8 mm diam., glabrous; pedicels 1-2
X 0.3 mm, glabrous; flowers in bud immediately
prior to anthesis 1. 5-2.0 X 1.0— 1.5 mm, cream,
apieally green; tepals 4, 2.0— 2.5 mm, obovate, gla¬
brous; subapieal appendages 0.5— 1.0 mm, cornic-
ulate, glabrous; stamens 4, filaments ca. 1.0 mm,
anthers ca. 0.5 X ca. 1.0 mm; pistillate inflores¬
cences 1 or 2 per axil, 7—30 mm, bearing 4—30
flowers in a semi-compact head; peduncle % inflo¬
rescence length, ca. 0.5 mm diam., glabrous; ped¬
icels 0.5— 0.8 X 0. 3-0.5 mm, glabrous; flowers 1.0—
1.8 X 0.5-0. 8 mm, glabrous; tepals 3, unequal, the
central one 1-2 mm, oblong to spatulate, the apex
reflexed, the lateral two 0.8— 1.5 mm, ovate, the
apex not reflexed. Infructescences 15—30 mm, pe¬
duncle 0.5 mm diam.; fruit 2.0-2. 5 X 1.5 mm,
compressed ovoid, the margin apieally and ven-
trally flattened.

Distribution. Forest at an altitude of 350—1400
m in Code, Panama, and San Bias departments of
central Panama.

Discussion. This species falls within Weddells
Heterophyllae and Killip's Centradenioideae species
groups. Pilea magnicarpa most closely resembles
P. ecboliophylla, which is found throughout Me¬
soamerica; it may be distinguished from the latter
as summarized below:

P. ecboliophylla-. V- and Y-shaped eystoliths pre¬
sent on adaxial leaf surface, abaxial surface minute
punctate glandular; pistillate inflorescences 1^4

mm; infructescences 2.5— 6.0 mm, fruits 0. 8-1.0
mm.

P. magnicarpa : V- or Y-shaped eystoliths absent,
abaxial leaf surface eglandular, rarely punctate
glandular; pistillate inflorescences 7-30 mm; in¬
fructescences 15—30 mm, fruit 2. 0-2. 5 mm.

To a lesser extent this species also resembles P.
chiriquina Killip from Costa Rica and Panama, but
may be distinguished as outlined below:

P. chiriquina: staminate flowers 15—200 per in¬
florescence, borne in 1-4 compact beads; pistillate
inflorescences 2.0— 8.5 mm.

P. magnicarpa : staminate flowers 1 1 —10 per in¬
florescence, borne in a loose panicle; pistillate in¬
florescences 7—30 mm.

Paratypes. PANAMA. Code: El Cope, along gravel
road to the right before sawmill, 2400 It.. 18 Oct. 1979,
Antonio 2180 (MO); road to Fortuna Dam site N of Gual-
aca. 22.7 mi. beyond the bridge over Rfo Estf, 1400 m,
22 Nov. 1979, Antonio 2767 (MO); Alto Calvario cloud
forest. 5.3 m above El Cope, above sawmill, on continental
divide. 9.30 m, 6 Dee. 1979, Antonio 3035 (MO); vicinity
El Cope, 5—6 mi. N of El Cope, along trail which leads
into the lowlands from old Riviera saw works area,
08°38'N, 08°35'W, (>00-4500 m, 8 July 1994. Croat A' Zim
77220 (I5M. MO not seen): 7 km N of El Cope de Vera-
guas, circa Rivera sawmill. Alto Calvario, ca. 900 m, con¬
tinental divide at 1.3(H) m. I 1 Jan. 1977. Folsom 1212
(MO); lumber camp at Alto Calvario, 7 km N of El Cope,
ca. 900 m, 14 Jan. 1977, Folsom 1260 (MO); summit. Alto
Calvario, 7+ km N of El Cope, 850 m. 19 May 1977,
Folsom & Hutton 3233 (MENU, MO not seen); near con¬
tinental divide along lumbering road 8.4 km above El
Cope (1 km beyond sawmill). 9(H) m. 19 Jan. 1978. Hum¬
mel 950 (MO); near continental divide along lumbering
road, 2.2 km beyond sawmill in forest along lumber road
above El Cope, 9(H) m, 20 Jan. 1978, Hummel 1031 (MO);
along road from La Pindeda to El Cope by way of Piedras
Gordas, sawmill above El Cope cloud forest. 3000 ft.. 20
Apr. 1978, Hummel 2613 (MO); continental divide on
road to Coclecito, in patch of forest near road, 1500 ft.,
20 June 1978, Hummel 3431 (MO); continental divide N
of Penonome on road to Coclesito small patch of forest at
roadside, 16(H) ft.. 25—26 July 1978. Hummel 4014 (MO);
sawmill above El Cope, in forest along stream E ol saw¬
mill. Atlantic drainage, 2300 ft., 27 July 1978. Hummel
4138 (MO). Panama: Llano— Cartf road. 1 mi. past saw¬
mill on dirt road, 1 I Nov. 1979, Antonio 2550 (BM. MO).
San Bias: El I Jano— Cartf road, 12 mi. from Pan American
Highway, along deeply shaded large stream, 350— 4(H) m,
10 May 1981, Sytsma & Andersson 4495 (MO).

Acknowledgments. I thank Helen Greenop for
the illustrations, Dennis Adams (BM), William Bur¬
ger (F), Sandy Knapp (BM), Amy Pool (MO), and
Karen Sidwell (BM) for help with the preparation
of the manuscript, and Norman Robson (BM) for
help with the Latin diagnoses. I also thank the cu¬
rators at C, F, GH. MEXU, MO, NY, P, and US for
the loan of specimens.

400

Novon

Literature Cited

Adams, C. I). 1970. Notes on Jamaican Flowering Plants
I. Mitt. Hot. Staatssamml. Miinohen H: 99—1 10.

- . 1972. Flowering Plants of Jamaica. Univ. of the

West Indies, Mona.

Hurger, W. 1977. Pile<i. In: flora Costaricensis. Fieldiana,
Hot. 40: 246-272.

Killip, K. P. 1936. New species of Pilea from the Andes.
Contr. U.S. Natl. Herb. 26(8): 367-394.

- . 1939. The Andean species of Pilea. Contr. U.S.

Natl. Herb. 26(10): 475-530.

Standley. P. C. A J. (7 Steyermark. 1952. Pilea. In: Flora
of Guatemala. Fieldiana, Hot. 24(3): 410—122.

Weddell, H. A. Monographic de la famille ties Urticaeees.
Arch. Mus. Hist. Nat. 9: 1-400. 1856; 401-591. 1857.

- . 1869. Pp. 104—163 in A. I)e Candolle, Prodro-

mus, systematis naturalis regni vegetabilis. 16(1). Victor
Masson et fils, Paris.

Seis Nuevas Especies de Vriesea sect. Xiphion
(Bromeliaceae: Tillandsoideae) para Costa Rica

J. Francisco Morales

Instituto Nacional de Biodiversidad (INBio), Apto. 22-3100,
Santo Domingo de Heredia, Costa Rica

Resumen. Se describen y se ilustran seis nuevas
especies de Vriesea sect. Xiphion para Costa Rica:
V. barii, V. haberii, V. osaensis, V. simulans, V. ti-
quirensis y V. vulcanicola. Se discuten sus relacio-
nes dentro del genero. Todas son endemieas por el
momento.

ABSTRACT. Six new species of Vriesea from Costa
Rica, V. barii, V. haberii, V. osaensis, V. simulans,
V. tiquirensis, and V. vulcanicola, are described and
illustrated. In addition, their relationships are dis¬
cussed. All are endemic as far as is known.

El genero Vriesea (Bromeliaceae: Tillandsoideae)
comprende alrededor de 250 especies distribuidas
en las regiones humedas del neotropico (Utley &
Ltlev, 1994). A nivel de Mesoamerica, Costa Rica
y Panama presentan un centro de diversidad bas-
tante alto, donde aproximadamente 60 especies son
encontradas, 43 de ellas endemieas y confinadas a
la Cordillera de Talamanca y sus estribaciones.

Como resultado de la elaboracion del tratamiento
de Bromeliaceae para el Proyecto Manual de las
Plantas de Costa Rica, fueron encontradas seis
nuevas especies de Vriesea subgenero Vriesea sec-
cion Xiphion. Esta seccion se caracteriza por sus
flores con los estambres incluidos y bracteas ma-
yormente verde palido o marron (Smith & Downs,
1977).

A pesar de recientes tratamientos taxonomicos
en los cuales se da reconocimiento generieo a var-
ias secciones dentro de Vriesea (Grant, 1995; Grant
& Zijlstra, 1998), yo prehero seguir la elasificacion
establecida por Smith y Downs (1977) hasta que se
proponga una evaluaeion integral de Vriesea y los
generos mas cercanamente relacionados, Tillandsia
y Guzmania, debido a la problematica delimitacion
entre estos y algunas de sus secciones.

Vriesea barii J. F. Morales, sp. nov. TIPO: Costa
Rica. Cartago: Canton de El Guarco, Valle del
Reventazon, La Sierra, cabeceras Rio Estrella,
1800-2100 m, 18 abr. 1994, Morales 3329
(holotipo, INB 2 laminas). Figura 1.

A Vriesea macrantha Mez & Werckle, cui affinis, brac-

teis florigeris et inflorescentia permajoribus differt; a V.
bicolor I.. 8. Smith bracteis florigeris rugosis non secundis
differt.

Epfhtas, raramente terrestres; plantas en flora-
cion 60—70 cm alto. Hojas rosuladas, moderamente
extendidas, (39— )48— 65 cm largo, punteado-lepi-
dotas, especialmente en la cara abaxial, cartaceas,
verdes a moradas abaxialmente. Vainas de las hojas
anchamente elipticas, 13—17 X 7—8.5 cm, canela-
marron, a veces con una mancha cafe oscuro cerca
de la base, moderadamente punteado-lepidotas, es-
pecialmente en la cara abaxial. Laminas de las ho¬
jas liguladas, 3.5—5 cm ancho, acuminadas. Escapo
erecto, 18.5—21 cm X 5—8 mm, marron claro. Brac¬
teas del escapo erectas, 6.5—9 cm largo, mas largas
que los entrenudos, densamente imbricadas, las
distales cafe-marron, las proximales verde-canela.
Inflorescencia simple, erecta, 30^13.5 X 4—5.5 cm,
distiea (aun en fructificacion), multiflora. Bracteas
primarias como las bracteas del escapo, mas largas
que los entrenudos, marron oscuro. Bracteas flora-
les 5—6.5 cm largo, anchamente ovadas, obtusas a
agudas, imbricadas y coriaceas, ecarinadas, rugo-
sas, marron oscuro. Pedicelos 7—9 X 5—6 mm. Se-
palos 2.7^1 cm largo, obovados, subcoriaceos, ea-
nela. Petalos blancos con los hordes morado, 6-7
cm largo, angostamente obovados, obtusos, con tins
escamas basales. Frutos capsnlas eilfndricas, 2.5—
3.2 cm, mds cortas o igualando los sepalos, ocultas
dentro de las bracteas florales; coma de la semilla
crema a blanco-crema.

Distribucion. Vriesea barii esta restringida a la
Cordillera de Talamanca en Costa Rica, entre 1700
y 2100 m, donde se encuentra en areas alteradas
(pastizales) y en el dosel superior de bosques pri¬
maries o intervenidos.

Fenologia. Florece de abril a junio; fructifica
de septiembre a diciembre.

Esta especie se encuentra bastante relacionada
con el eomplejo de especies de Vriesea gladioliflora
(H. Wendland) Antoine; sin embargo, es facilmente
separable de esta especie por el color castano os¬
curo de sus bracteas, capsulas muy pequenas, to-
talmente ocultas dentro de las bracteas al madurar

Novon 9: 401-406. 1999.

402

Novon

6 cm 3 cm

Figura I. A— I). I riesea barii J. F. Morales (Morales 3329,
INB). — A. Hoja. — B. Inflorescencia. — C. Petalo mos-
trando las escamas basales. — 1). Sepalos y frulos.

y distribucion altitudinal. Generalmente, V. gladio-
liflora es muy comun en zonas bajas, donde se le
encuentra desde nivel del mar hasta los 1300 m;
sin embargo, esta especie prdcticamente desapa-
rece sobre los 1500 m y a pesar del continuo tra-
bajo de campo que he realizado en los ultimos
afios, nunca he encontrado poblaciones sobre esa
faja altitudinal. Algunas veces, especfmenes ile V.
ampin L. B. Smith sobre los 2000 m han sido iden-
tificados erroneamente como V. gladioliflora.

Vriesea barii se distingue de V. macrantha Mez
A W erekle por sus brdeteas e inflorescencia mds
larga. Asimismo, puede ser confundida con V. In¬
color L. B. Smith, pero se distingue por sus bracteas
no unilaterales y conspicuamente rugosas. De V.
ampin L. B. Smith se reconoce por sus inflorescen-
cias dfsticas, aun en fructificacidn, erectas y nunca
arqueadas y por los Irutos totalmente ocultos dentro
de las bracteas florales.

El epiteto honra a Alvaro “Bari” Fernandez,
qui£n fue el primero en descubrir esta especie,
como reconocimiento por su trabajo como recolec¬
tor de especfmenes cientfficos y por la amistad per¬
sonal que nos une.

Paratipos. COSTA HICA. ( art ago: Cordillera de Ta-
lamanea, Fila Alto lndias, cerca del cruce a Casarnata,
Morales 4907 (INB. SKL); Cordilleru de Talamanca. Vuel-
ta Ventolera, 2 km S de Casarnata, Morales et al. 5387

3 cm

Figura 2. A— C. Vriesea haberii J. F. Morales (Haber A
Zuchowski 8947, INB). — A. Hoja. — B. Inflorescencia.
— C. Sepalos y fnitos.

(INB). I,iin6n: Canton de Talamanca, Cordillera de Ta¬
lamanca, Rfo < ioen. l.jamis a San Jose Cabecar, Ferndn-
de: 1052 (INB).

Vriesea haberii J. F. Morales, sp. nov. TIPO: Costa
Rica. Guanacaste: Canton de Abangares, 5 km
N Monteverde, western margin of Reserva Bio-
logica Monteverde, junction of road to San Ge¬
rardo valley with Rfo Negro, 1550 m, 23 Dec.
1988, Haber & Zuchowski 8947 (holotipo, INB
2 laminas). Figura 2.

Volume 9, Number 3
1999

Morales

Vriesea sect. Xiphion

403

A Vriesea acuminata Mez & Werckle, cui affinis, brac-
teis florigeris rugosis papillosis differt; a V. burgeri 1.. B.
Smith bracteis florigeris rugosis papillosis et non secundis
differt.

Epffitas; plantas en floracion basta 65 cm alto.
Hojas rosuladas, moderamente extendidas, (25-)34—
42 cm largo, esparcida a moderadamente punteado-
lepidotas, especialmente en la cara abaxial, carta-
ceas, verdes a variegadas con morado. Vainas de
las hojas anchamente elfpticas, 7—9 X 7-8.5 cm,
canela, a veces con una leve banda basal cafe os-
curo, moderadamente punteado-lepidotas, especial¬
mente en la cara abaxial. Laminas de las hojas li-
guladas, 6—6.7 cm ancho, acuminadas. Escapo
erecto, 70-76 cm X 5 mm, matron claro. Bracteas
del escapo erectas, 4.5— 5.5 cm largo, mas largas
que los entrenudos, imbricadas, matron, densamen-
te rugoso-papilosas, opacas. Inflorescencia simple,
ereeta, 17.5-19.5 cm largo, dfstica, con 15-18 flo-
res. Bracteas primarias como las bracteas del es¬
capo, mas largas que los entrenudos, matron os-
curo. Bracteas florales 3.5— 4.1 cm largo,
anchamente elfpticas, agudas, algo imbricadas an¬
tes de la antesis, patentes a levemente ascendentes
en fructificacibn y subcoriaceas, ecarinadas, con-
spicuamente rugoso-papilosas, matron. Pedicelos
3—4 X 3 mm. Sepalos 1.4— 1.6 cm largo, ancha¬
mente ovados, subcoriaceos, matron. Corola des-
conocida. Frutos inmaduros capsulas 2.5— 3.1 cm
largo, moderadamente acuminadas hacia el apice;
coma de la semi 11a canela.

Distribucion. Costa Rica, en bosques muy hu-
tnedos de la Cordillera de Tilaran (Valle de San
Gerardo), 1500-1700 m.

Fenologia. Se encuentra en fructificacibn entre
noviembre y ertero.

Vriesea haberii se encuentra algo relacionada con
V. acuminata Mez & Werckle, pero es facilmente
separable por sus bracteas florales y bracteas del
escapo conspicuamente rugoso-papilosas, matron
opaco. La coleccion tipo fue previamente identifi-
cada como V. burgeri L. B. Smith; sin embargo, di-
fiere de esta especie por sus bracteas florales no
secundas en la antesis, dfsticas a ligeramente se-
cundas en Iructifieacion y bracteas del escapo ru-
goso-puneticuladas.

Vriesea osaensis J. F. Morales, sp. nov. TIPO: Cos¬
ta R iea. Puntarenas: Canton de Golfito,
Jimenez, Dos Brazos de Rfo Tigre, Cerro Mue¬
ller, cuenca superior del Rfo Rincon, 744 tn,
26 ago. 1990, Herrera 4144 (holotipo, INB).
Figura 3.

A Vriesea riridiflora (Regel) Wittmack ex Mez. cui af-

Figura 3. A, B. Vriesea osaensis J. F. Morales (Herrera
4144, INB). —A. Habito. — B. Sepalos.

finis, bracteis florigeris et sepalis majoribus, inflorescentia
non secunda differt.

Epffitas; plantas en floracion 40-60 cm alto. Ho¬
jas rosuladas, moderamente extendidas, 27-73 cm
largo, moderada a densamente lepidotas, especial¬
mente en la eara abaxial, cartaceas, verdes. Vainas
de las hojas anchamente elfpticas, 4.5— 7.5 X 3—5
cm, canela a concoloras con la lamina, lepidotas,
especialmente en la cara abaxial. Laminas de las
hojas liguladas, (1.4— )2. 5-3.8 cm ancho, acumi¬
nadas a subredondeadas y apiculadas, glabradas a
esparcidamente lepidotas. Escapo erecto y arquea-
do, 29-40 cm X 4—6 mm, canela. Bracteas del es¬
capo erectas, S.S-^.S cm largo, ligeramente mas
largas que los entrenudos, apenas imbricadas, ver¬
des. Inflorescencia simple, ereeta, (8—) 12-22 X ca.
2 cm. dfstica (aun en fructificacibn), con 4-7 flores.
Bracteas primarias como las bracteas del escapo,
mas largas que los entrenudos, verdes. Bracteas flo¬
rales 4—5.3 cm largo, anchamente elfpticas, agudas,
densamente imbricadas y subcoriaceas, ecarinadas,
lisas a levemente rugulosas, verdes. Pedicelos 2^1
X 2-3 mm. Sepalos 3-3.4 cm largo, elfpticos, sub-
cartaceos, verdes. Petalos crema, 4. 7-5.2 cm largo,
angostamente obovados, obtusos, con dos escamas
basales. Frutos capsulas cilfndricas y obtusas api-
calmente, 3.4— 3.7 cm, ocultas dentro de las brac¬
teas florales a ligeramente visibles; coma de la se-
milla blanco-crema.

404

Novon

Distribution. Costa Rica, en bosques muy hu-
medos de la Peninsula de Osa, Golfito y en las fajas
costenas del canton de Puriscal, Provincia de San
Jose, entre 0 y 750 m.

Fenologta. Se eneuentra en floracion entre
agosto y septiembre; fructifiea de octubre a marzo.

Vriesea osaensis ha sido anteriormente identifi-
cada como V. viridiflora (Regel) Wittmack ex Mez;
sin embargo, difiere de esta especie por sus brac-
teas y sepalos conspicuamente mds largos e inflo-
rescencia dfstiea, no unilateral. En adicion, V. vi¬
ridiflora se eneuentra en bosques nubosos muy
humedos entre los 900 y 1800 m.

Pardtipos. COSTA RICA. Puntarenas: Parque Na-
eional Corcovado, eslacidn Los Patos, Aguilar 2227 (INB);
Canton de Golfito, Hfo Fsquinas, Induni 194 (INB): Re-
serva Forestal Golfo Dulce, Albergue Cerro de Oro,
Sendero I Patos, Mora 82 (INB, SKI.); Reserva Forestal
Golfo Dulce, Bahfa Rincon. Rincon. Morales 1915 (IN It);
Canton dc Osa, Rancho Quemado, Morales & Hamrnel
8188 (INB). San Jose: Canton dc Puriscal, Zona Protec-
tora La Cangreja. Mastatal dc Puriscal, Morales 741 (INB).

Vriesea simulans J. F. Morales, sp. nov. TIPO:
Costa Rica. Alajuela: Canton de San Ramon,
Reserva Forestal San Ramon, 1000 m, 17 abr.
1994, Morales & Lepiz 2704 (holdtipo, INB;
isotipos, MO, SEL). Figura 4.

A Vriesea kupperiana Sucssenguth ct L werekleana Mez,
cui affinis, bractcis florigcris majoribus imbricatis non se-
cundis differt.

Epffitas, terrestres o rupfcolas; plantas en flora-
ci6n hasta 200 cm alto. Hojas rosuladas, extendi-
das, 91—105 cm largo, punteado-lepidotas, espe-
cialmente en la cara abaxial, cartdceas, verdes.
Vainas de las hojas anchamente elfpticas, 20-25 X
1 1.5—15 cm, marron oscuro, densamente punteado-
lepidotas, especialmente en la cara abaxial. Lami-
nas de las hojas liguladas, 9-10 cm ancho, aeu-
minadas. Escape erecto, algunas veces levemente
arqueado, 60—110 cm X 19—26 mm, canela-mar-
r6n. Brdcteas del escape erectas, 10—13 cm largo,
rods largas que los entrenudos, densamente imbri-
cadas, basalmente marron oscuro, apicalmente ver-
<le-canela. Inflorescencia compuesta, bipinada con
5—7 ramos laterales, erecta, 65—80 X 60 cm en
diametro, todas las ramas laterales emergiendo
agrupadas en aproximadamente 15-20 cm, multi-
flora. Brdcteas primarias como las brdcteas del es¬
cape, mds cortas que la base esteril de las ramas
laterales. Ramas laterales arqueadas y algo pen-
dulosas distalmente, 30-65 cm largo, con abun-
dante secrecion acuosa, bracteas florales y flores
dfsticas. Brdcteas florales 3. 6-4.4 cm largo, ancha¬
mente elfpticas a anchamente ovadas, obtusas a

C. Sepalos.

agudas, densamente imbricadas y coridceas, las
proximales carinadas, el resto ecarinadas, lisas a
rugulosas, verdes en la antesis, tornando luego a
pardo-canela. Pedicelos 7-9 X 3.5—5 mm. Sdpalos
2.2— 2.5 cm largo, angostamente obovados, coria-
ceos, verdes a verde-canela. Corola y frutos des-
conocidos.

Distribution. Costa Rica, conocida solamente
en la Reserva Forestal de San Ramon y en sectores
aledanos, en la vertiente atlantica de la Cordillera
de Tilaran, entre 800 y 1000 m.

Fenologta. Florece de abril a julio.

Vriesea simulans se eneuentra cercanamente re-
lacionada con V. kupperiana Suessenguth y V.
werekleana Mez; sin embargo, se distingue de am-
bas por sus bracteas florales mas largas, pardo-ca¬
nela, en lugar de crema-canela, conspicuamente
imbricadas, antes y despues de la antesis, ramas
laterales arqueados y algo pendulosos y por sus
brdcteas y flores dfsticas.

Volume 9, Number 3
1999

Morales

Vriesea sect. Xiphion

405

Pardtipos. COSTA RICA. Alajuel a: carretera San
Ramon-Bajo Rodriguez. Luther et al. 2801 (INB. SKL).

Vriesea tiquirensis J. F. Morales, sp. nov. TIPO:
Costa Rica. San Jose: Canton de Acosta, Valle
del Candelaria, Hacienda Tiquires, cuenea
alta del Rio Tiquires, 1200 m, 22 abr. 1995,
Morales et al. 3984 (holotipo, INB; isotipo,
SEL).

A Vriesea tonduziana I .. B. Smith, cui affinis. liracteis
florigeris membranaceis fragilis differt.

Epifitas o terrestres; plantas en floracion hasta
65 cm alto. Hojas rosuladas, extendidas, (29-)34—
76 cm largo, punteado-lepidotas, especialmente en
la cara abaxial, cartaceas, verdes a moradas abaxi-
almente. Vainas de las hojas anchamente elipticas,
16—19 X 5.8—9 cm, canela a canela-marron, den-
samente punteado-lepidotas. Laminas de las hojas
liguladas, 3.5— 5(— 6.2) cm ancho, acuminadas. Es-
capo erecto a erecto y algo arqueado, 30-47 can X
7-9 mm, verde-canela a marron claro. Bracteas del
escapo erectas, 6—7 cm largo, mas largas que los
entrenudos, densamente imbricadas, verde-canela
a canela-marron. Inflorescencia simple, ereeta a li-
geramente erecto-arqueada, 19-30 X 3^1 cm, Ho-
res y bracteas Horales dispuestas unilateralmente,
con 8—15 flores. Bracteas primarias como las brae-
teas del escapo, mas largas que los entrenudos.
Bracteas Horales 3.9— 5.5 cm largo, anchamente
ovadas, agudas a redondeadas, imbricadas, deli-
cadamente membranaceas, ecarinadas, conspieua-
mente rugosas, canela-marron, desintegrandose ra-
pidamente despues de la antesis y no persistentes
en fructificacion. Pedicelos 6-9 X 4 mm. Sepalos
2.5— 4.1 cm largo, angostamente oblongo-obovados,
cartaceos, verdes a verde-canela. Petalos verde-
blanco a verde-crema. Frutos desconocidos.

Distribucidn. Costa Rica en bosques muy lui-
medos y areas alteradas de la Fila Aguabuena (Ti¬
quires) y la Fila San Jeronimo, en el Valle del Rio
Candelaria en el canton de Acosta, provincia de
San Jose, 900—1300 m.

Fenologia. Florece de abril a mayo.

Esta especie se encuentra cercanamente relacio-
nada con Vriesea tonduziana L. B. Smith, pero se
distingue por sus bracteas Horales membranaceas y
fragiles, descomponiendose rapidamente despues
de la antesis, no persistentes en fructificacion.

Pardtipos. COSTA RICA. Sail Jose: Canton de Acos¬
ta, Vila Bustamante, cabeceras Quebrada San Jeronimo.
Morales 4462 (INB).

6 cm

Figura 5. A— D. Vriesea vulcanicola J. F. Morales (Mo¬
rales et al. 2237, INB). — A. liabito. — B. Petalo mos-
trando las escamas basales. — C. Sepalos. — I). Frutos.

Vriesea vulcanicola J. F. Morales, sp. nov. TIPO:
Costa Rica. Guanacaste: Parque Nacional Rin¬
con de La Vieja, Volcan Rincon de la Vieja,
bosque enano en la falda SW, 9 ene. 1994,
Morales et al. 2237 (holotipo, INB; isotipo,
MO). Figura 5.

A Vriesea brunei Mez & Werckle, eui affinis, bracteis
et sepalis rugosis differt.

Terrestres; plantas en Horacion hasta 45 cm alto.
Hojas rosuladas, moderadamente extendidas, 27—
34 cm largo, densamente a moderadamente pun¬
teado-lepidotas, cartaceas, verdes a moradas abaxi-
almente. Vainas de las hojas anchamente elipticas,
9—1 1 X 9—10 cm, canela a canela-marron, densa¬
mente punteado-lepidotas. Laminas de las hojas li¬
guladas, 5—6 cm ancho, subredondeadas y apicu-
ladas. Escapo erecto, 33—37 cm X 7—9 mm, canela
a verde-canela. Bracteas del escapo erectas, 5.5—
6.5 cm largo, mas largas que los entrenudos, den-

406

Novon

samente imbricadas, canela a verde-canela, cons-
picuamente y densamente rugosas. Inflorescencia
simple, erecta, 18-22 X 4.5-6 cm, dfstica (aun en
fructificaci6n), con 10—13 flores. Bracteas primarias
como las brdcteas del escape, mas largas que los
entrenudos. Brdeteas Horales 5.2— 5.9 cm largo, an-
chamente ovadas, agudas y conspicuamente api-
culadas, imbricadas y coriaceas, ecarinadas, den¬
samente rugosas, verde-marron a verde-rojizo en la
antesis, canela-marrdn con la edad. Flores dfsticas
(raramente muy ligeramente unilaterales). Pedice-
los 2-4 X 4-5 mm. Sepalos 3—3.4 cm largo, an-
chamente ovados a oblongo-obovados, coriaceos,
canela-marrdn. Petalos crema, 5.5—6 cm largo, an-
gostamente obovados, obtusos, con dos escamas ba-
sales. Frutos capsulas cilfndricas, 3.8— 4.1 cm; se-
rnillas desconocidas.

DLstribucion. Costa Rica, donde es conocida
solamente en los bosques enanos y ventosos de la
cima del Volcan Rincon de la Vieja, Cordillera de
Guanacaste, entre 1700 y 1800 m.

Fenologia. Se encuentra en floracidn entre di-
ciembre y lebrero; fructifica de marzo a agosto.

Vriesea vulcanicola se encuentra cercanamente
relacionada con V. brunei Mez & Werckle, de la
que se dilerencia por sus brdcteas y sepalos cons-
picuamente rugosos. Este cardcter puede ser con-
siderado poco consistente dentro del genero; sin

embargo, intensivos estudios de campo en pobla-
ciones de V. brunei en la localidad tipo (Aguabue-
na, Valle del Candelaria) me han demostrado que
esta especie siempre tiene las brdeteas de la inflo¬
rescencia y escapo lisas, asf como en otros sitios
aparte de la localidad tipo (estribaciones de la Cor¬
dillera de Talamanea).

Ademas se puede confundir con V. tonduziana
L. B. Smith, pero V. vulcanicola se reconoce facil-
inente por su inflorescencia dfstica (no unilateral)
y las bracteas Horales engrosadas en un conspicuo
apfculo.

Pardtipos. COSTA RICA. Guanacaste: Parque Na-
cional Rincon de la Vieja, Sector Las Pailas, Volcan Rin¬
con de la Vieja, 15 ago. 1996, Morales 5701 (INR, MO,
SKI.).

l.iteratura Citada

Grant, J. R. 1995. The resurrection of Alcantarea and We-
rauhia , a new genus. 'Prop. Sublrop. Pflanzeimelt 91:
7—57.

- iV G. Zijlstra. 1998. An annotated catalogue of the

generic names of the Bromeliaceae. Selbyana 19: 91 —

121.

Smith. I.. R. & R. J. Downs. 1977. Tillandsioideae (Bro-
meliaeeae). FI. Neotrop. Monogr. 14: 1068—1275.

Utley, J. F. & K. R. Utley. 1994. Vriesea Lindley. Pp. 122—
135 in: G. Davidse. M. Sousa S. & A. O. Chater (edi¬
tors). Flora Vlesoamerieana. Vol. 6. Alismataceae a
Cyperaceae. Universidad Nacional Autdnoma de Mexi¬
co. Mexico D.F.; Missouri Botanical Garden, St. Ix>uis;
I he Natural History Museum, London.

A New Species of Villarsia (Menyanthaceae) from South Africa

Robert Ornduff

Department of Integrative Biology, University of California, Berkeley,
California 94720-3140, U.S.A.

ABSTRACT. A second African species of Villarsia,
V. goldblattiana, is described from South Africa.
The genus otherwise is Australian (13 species) and
southeastern Asian (1 species). Villarsia goldblat¬
tiana differs from V. capensis in its larger ovule
number, glabrous seeds, longer calyx lobes, taller
inflorescences, and apparent requirement of fire to
induce flowering. It is known only from the south¬
ern Cape Peninsula. Both South African Villarsia
species are tetraploid with 2n = 36. An emended
description of V. capensis is given, and the origin
and location of its holotype and of the holotypes of
the synonyms Menyanthes ovata L. f. and M. ca¬
pensis Thunberg are discussed.

The only species of Villarsia (Menyanthaceae)
currently recognized in Africa is V. capensis (Hout-
tuyn) Merrill of Western Cape Province (Marais &
Verdoorn, 1963; Dyer, 1975). Elsewhere, there are
13 species of the genus in Australia (Aston, 1969;
Ornduff, 1990) and an additional one in southeast¬
ern Asia (Ornduff, 1994). As traditionally circum¬
scribed, Villarsia capensis exhibits interpopulation
variability in habitat, leaf size, and floral charac¬
ters. I have suggested that further study of the
South African plants may result in the recognition
of more than one species (Ornduff, 1974). Recently,
I examined specimens of South African Villarsia
borrowed from BOL and NBG and concluded that
a second species of Villarsia occurs in South Africa.
I am naming it Villarsia goldblattiana in recogni¬
tion of Peter Goldblatt’s significant contributions to
southern African botany, particularly the system-
atics and reproductive biology of the Iridaceae, and
in appreciation for his considerable assistance in
my study of Villarsia in South Africa.

Villarsia goldblattiana Ornduff, sp. nov. TYPE:
South Africa. Western Cape Province: growing
abundantly in shallow standing water of marsh,
N of road to Olifantsbos, 1.7 mi. W of its junc¬
tion with main road to Cape Point Nature Re¬
serve, 5 Feb. 1971, Orndujf 7099 (holotype,
sheet I, NBG; isotype, sheet II, NBG). Figure 1.

Villarsiae capensi affinis set! differ! in characteribus

pluribus: inflorescentia 45—84 cm elata. 1 ,8-vel 2.5-plo
longior quam folium longissimum; calycis lobi 6—14 mm
longi, 2—3 mm late basi; ovaria ovulis 3—8; semina glabra.

Erect, tufted, non-st oloniferous perennial. Basal
leaves erect, evergreen, with petioles 12-32 cm
long; blades fleshy, oblong to ovate, 4—9 cm long,
1.5—8 cm wide, attenuate to cordate at base, entire.
Inflorescence an open panicle with 16-55 flowers,
45—85 cm tall, the tallest 1.8— 2.5 times the length
of the longest leaves, with cauline leaves gradually
reduced upward to scales. Flowers distylous, open
for only one day, on pedicels 1—2 cm long. Corolla
yellow, the 5 obovate lobes about twice the length
of the calyx lobes, the margins fimbriate, with hairs
on parts of the upper surface. Calyx lobes 5, ±
lanceolate, 6-14 mm long, 2—3 mm wide at the
base. Ovules 3-8 per ovary. Capsules ovoid, equal¬
ing or slightly longer than the calyx lobes. Seeds
glabrous, yellow, ± orbicular, biconvex, 1.5—2 mm
diam. Wet soil and marshes. Cape Peninsula. Near
sea level to 600 m. Flowers October to January,
reportedly after fire.

Paratypes. SOUTH AFRICA. Western Cape Prov¬
ince: “. . . in summo monte ‘Steenberg’ alt. circa 800
pcd..” Dec. 1903, L Kensit s.n. (BOL); Muizenberg Pla¬
teau, Nov. 1913, E. Kensit s.n. (BOL); Noord Hock Mts.,
alt. 2000 ft.. 2 Jan. 1945, Compton 16898 (BOL, NBG);
Cape Point. Patrys Vlei, 24 Nov. 1946, Compton 18826
(NBG); Smitswinkel Flats, Cape Peninsula, 29 Oct. 1942.
Compton 14007 (NBG).

Villarsia goldblattiana has 3—8 ovules per ovary,
glabrous seeds, calyx lobes that are 6—14 mm long,
and inflorescences that are 45—85 cm tall and 1.8—
2.5 times the length of the longest leaves. It grows
from near sea level to 600 m in elevation and is
known only from the southern Cape Peninsula,
where it has been collected on the slopes above
Noordhoek and the Steenberg plateau south to the
Smitswinkel Flats in tin* Cape Point Nature Re¬
serve. It is reported to flower after a fire (Fraser &
McMahon, 1994; P. Salter, pers. comm. 1997). I
suspect that the primary effect of fire is to increase
light intensity by temporarily reducing the shading
that is caused by an overstory of fynbos shrubs. I
revisited the type locality of this species in early
September 1998, but was unable to find any plants

Novon 9: 407-409. 1999.

408

Novon

Figure I . Holotype of Villarsia goldblattiana. The seale
is 5 cm.

at the site. Villarsia capensis has 1—2 ovules per
ovary, (usually) papi 1 late seeds (illustrated in
Chuang & Omduff, 1992), calyx lobes that are 4—
6 mm long, and inflorescences that are 12^15 cm
tall and 0.9-2. 1 times the length of the longest
leaves. It grows from 200 to 1520 m in elevation
and ranges from the Humansdorp region to the
mountains east of Paarl and Stellenbosch and west¬
ward from there to the Cold Bokkeveld. It is also
common on Table Mountain. It does not require fire
to flower. It and V. goldblattiana are tetraploid with
n = 18 (Ornduff, 1974). A photograph of V. capensis
can be found in Burman and Bean (1985), and a
painting of V goldblattiana has been published by
Fraser and McMahon (1994, as V. capensis).

An emended description of V. capensis follows:

Villarsia capensis (Houttuyn) Merrill

Erect, tufted, non-stoloniferous perennial. Basal
leaves erect, evergreen, with petioles 2—25 cm long;
blades fleshy, oblong to ovate, 1-9 cm long, 0.6-
6.5 cm wide, rounded to cordate at base, usually
with widely spaced shallow teeth or sometimes en¬
tire. Inflorescence a congested to open panicle with

5-75 flowers, 12-45 cm tall, the tallest 0.9-2. 1
times the length of the longest leaves, with cauline
leaves gradually reduced upward to scales. Flowers
distylous, open for only one day, on pedicels 0.2—
2.5 cm long. Corolla yellow, the 5 linear to narrowly
obovate lobes about twice the length of the calyx
lobes, the margins fimbriate, usually hairy on parts
of the upper surface. Calyx lobes 5, ± lanceolate,
4-6 mm long, 1.5-2. 5 mm wide at the base. Ovules
1-2 per ovary. Capsules ovoid, equaling or slightly
longer than the calyx lobes. Seeds usually papillate,
yellow or brown, ± orbicular, biconvex, 1.5-2 mm
diam. Wet soil, marshes, and stream margins, 200—
1520 m. Flowers September to January.

The epithet “ capensis ” traces back to Renealmia
capensis Houttuyn (1777). Houttuyn’s work was dis¬
cussed extensively by Merrill (1938), who wrote (p.
310), “All attempts to locate a Houttuyn herbarium
have failed, anti the probability is that most of his
actual types are no longer extant.” Marais and Ver-
doorn (1963) thus considered Houttuyn’s illustra¬
tion to be the type, but the floral details in that
illustration are crude and inaccurate. In 1998, G.
Thi jsse (L) wrote me, “According to an unpublished
mss. the main set of Houttuyn’s herbarium is at
Geneva. Also the holotype of Renealmia capensis is
at G.” The types of Menyanthes ovata L. f. (1782,
not 1781 fide Manitz, 1976) and M. capensis Thun-
berg (1794) both are Thunberg collections. Photo¬
graphs of these types and of the holotype of R. ca¬
pensis in the Burman herbarium at G indicate that
these three specimens are almost certainly dupli¬
cates of the same collection, and an additional
Thunberg specimen in the Vahl Herbarium at C
very likely is also from the same collection. Thun¬
berg (1823) stated that M. capensis “erescit in
Krum-rivier [= Krom River] et in summo Tuff el-
berg.” His specimens, however, are clearly not from
Table Mountain (Taffelberg), since Table Mountain
plants are uniformly smaller, and have elliptic or
oblong rather than ovate leaves. The Thunberg
specimens are larger, with inflorescences to 35 cm
tall; these are 1.5— 1.7 times the length of the lon¬
gest leaves, which are ovate. They resemble plants
collected at Kirby ( Fourcade 5901 , BOL), “Ralets”
(= Ratels) Bosch flats ( Fourcade 596, BOL), and
Wagenbooms River, Long Kloof (//. Bolus 2404,
BOL), all at the eastern end of the range of V. ca¬
pensis (Fourcade, 1940). Krom River originates at
the eastern end of the Tsitsikamma Mountains and
empties into St. Francis Bay in the general region
of the three collections cited above. Thunberg vis¬
ited this area between September 1772 and early
January 1773 and again between September 1773

Volume 9, Number 3
1999

Ornduff

Villarsia goldblattiana from South Africa

409

and January 1774 (Karsten, 1939). The distribution
of this larger form (to which the holotype of V. ca-
pensis belongs) is disjunct. It has been collected at
the eastern end of the range of the species, but also
in the Ceres area ( Guthrie 2191, NBG, and Comp¬
ton 10090, NBG, the latter with the comment that
it “looks very different from the Table Mountain
swamp-plant!”) and the Badsberg ( Walters 1557,
BOL). The smaller, more widespread form also has
been collected near the last locality in Bain’s Kloof
(Schlechter 9174, BOL; Leighton 2153, BOL; Orn¬
duff 7267, LC). No taxonomic separation of these
two forms is suggested at present.

Acknowledgments. I am indebted to Peter Gold-
blatt (MO) lor his considerable assistance in car¬
rying out this study and his comments on an earlier
version of this paper, John Manning (NBG) for as¬
sistance in the field and for help in identifying held
localities, Helen I. Aston (MEL) for helpful com¬
ments on an earlier draft of this paper, the staff of
BOL. and NBG for lending specimens, G. Thijsse
(L) for assistance in locating the holotype of R. ca-
pensis, F. Jacquemoud (G) for providing a photo¬
graph of that holotype, and B. Hansen (C) for send¬
ing a photocopy of the Thunberg specimen in the
Vahl Herbarium.

I .ilerature Cited

Aston, H. I. 1969. The genus Villarsia (Menyanthaceae)

in Australia. Muelleria 2: 3—63.

Burman, I,. & A. Bean. 1985. Hottentots Holland to Her¬

manns (South African Wild Flower Guide 5). Botanical
Society of South Africa, Claremont.

Chuang. T. I. & R. Ornduff. 1992. Seed morphology and
systematics of Menyanthaceae. Amer. J. Bot. 79: 1396-
1406.

Dyer, R. A. 1975. The Genera of Southern African Flow¬
ering Plants, vol. 1. Department of Agricultural Tech¬
nical Services, Pretoria.

Fourcade, H. G. 1940. Check-list of the flowering plants
of the divisions of George, Knysna, Humansdorp, and
Uniondale. Botanical Survey Memoir No. 20: 1-127.

1 raser. M. & I.. McMahon. 1994. Between Two Shores.
David Philip Publishers, Claremont.

Houttuyn. M. 1777. Natuurlyke historic of uitvoerige bes-
chryving der dieren. planten en mineraalen. ... II. 8.
I)e Erven van F. Houttuyn. Amsterdam.

Karsten, M. C. 1939. Carl Peter Thunberg. An early in¬
vestigator of Cape botany. III. Travels. J. S. African Bot.
5: 105—155.

Linnaeus f., C. 1782 (not 1781). Supplementum Plantar-
urn . . . Impensis orphanotrophei, Braunschweig.

Manitz, H. 1976. Friedrich Ehrhart und die Publikation
des “Supplementum Plantarum” von Linnefilius. Taxon
25: 305-322.

Marais, W. & 1. C. Verdoorn. 1963. Gentianaceae. Pp.
171-243 in R. A. Dyer, L. E. Codd & H. B. Rycroft
(editors). Flora ol Southern Africa, vol. 26. Government
Printer. Pretoria.

Merrill. F. I). 1938. A critical consideration of Houttuyn’s
new genera and new species of plants, 1773-1783. J.
Arnold Arbor. 19: 291-375.

Ornduff, R. 1974. Cytotaxonomic observations on Villarsia
(Menyanthaceae). Austral. J. Bot. 22: 513—516.

- . 1990. A new species of Villarsia (Menyantha¬
ceae) from the Porongurup Range. Western Australia.
Syst. Bot. 15: 216—220.

- . 1994. Villarsia cambodiana (Menyanthaceae) in

southeastern Asia. Nordic J. Bot. 14: 647—648.

Thunberg, C. P. 1794. Prodromus plantarum capen-
sium. . . , pars prior. J. Edman. Uppsala.

- . 1823. Flora Capensis . . . J. G. Cotta, Stuttgart.

A New Species of Ceratozamia (Zamiaceae, Cycadales) from

Chiapas, Mexico

Miguel A. Perez- Far re r a

Escuela tie Biologfa, Universidad de Ciencias y Artes del Estado de Chiapas,
Calzada Samuel Leon Brindis S/N, C.P. 29,000, Tuxtla Gutierrez, Chiapas, Mexico

Andrew P. Vovides

Institute de Ecologfa, A.C., Apartado Postal 63, 91000, Xalapa, Veracruz, Mexico

Carlos Iglesias

Institute de Ecologfa, A.C., Apartado Postal 63, 91000, Xalapa, Veracruz, Mexico

ABSTRACT. A new species of cycad, Ceratozamia
alvarezii (Zamiaceae) from Chiapas, is described
and illustrated. This species has affinities with C.
matudai from Chiapas and C. sabatoi from Quere-
taro anil Hidalgo. It differs from these species with
regartl to trunk, leaf habit, and male and female
cones. It differs from C. norstogii, also from Chia¬
pas, in that the latter is a much larger unbranched
plant with a spirally twisting leaf rachis.

Rksumkn. Se describe e ilustra una nueva espe-
cie, Ceratozamia alvarezii (Zamiaceae). Esta espe-
cie presenta afinidad con C. matudai y C. sabatoi
de Queretaro e Hidalgo. Esta especie difiere de las
anteriores por la hoja, habito de los estrobilos y
morfologfa del tronco. Tambien difiere de C. nors¬
togii dado t|ue la ultima es una planta mas grande,
no ramificada y la raquis torcida espiralmente.

During the course of botanical explorations in
one of the natural areas of Chiapas that has recently
been decreed as a Biosphere Reserve, we collected
a species of Ceratozamia with a unique combina¬
tion of trunk, leaf, and cone characters.

We believe that the new species is close to Ce¬
ratozamia matudai Lundell. Specimens of the new
plant and C. matudai were cultivated under the
same conditions lor comparison purposes for a pe¬
riod of two years, in which time new leaf flushes
and cones occurred. The plants under cultivation
continued to present the same characteristics as
those of the natural habitat. The cones of Cerato¬
zamia alvarezii differed from those of C. matudai
as well as those of C. norstogii Stevenson, which
also is from Chiapas; C. norstogii also differs in its
much larger habit, and in having an erect non¬
branching trunk and leaves with a spirally twisted
rachis.

Ceratozamia alvarezii Perez-Farrera, Vovides &
Iglesias, sp. nov. TYPE: Mexico. Chiapas: Si¬
erra Madre of Chiapas, 4 Mar. 1996, M. A.
Perez-Farrera 889 6 (holotype, CHIP; iso¬
types, F, MEXU, MO). Figure 1.

Planta trunco subgloboso ad cylindricum semihypogaeo
ad hypogaeum, ramoso, 10-50 cm alto; cataphyllis lanatis
triangularibus, stipulatis. Folia pinnata; petiolo 14^12 cm
longo; rhachidi 25—66 cm longa; foliolis oppositis ad sub-
opposita, 24— 62-jugis, lineari-lanceolatis. Strobilus mas¬
culines lineari-cylindricus I 1-31 cm longus pedunculo
tomentoso 4—5 cm longo insidens; strobilus feminineus
14.5—19 cm longus pedunculo tomentoso 4. 5-6. 5 cm lon¬
go insidens; seminibus 1.7-2. 5 cm longis.

Trunk subglobose, 10-50 cm long, partially sub¬
terranean, becoming cylindrical with age, branch¬
ing freely, 8.9-17.5 cm diam. and protected by per¬
sistent petiole bases. Cataphylls stipulate, lanulose,
2.1—5 cm long, 1.5—3 cm wide, brown. Leaves 4—
18, pinnate, spirally arranged forming an open
crown, 54-109 cm long, 31.5-61 cm wide. Petiole
14-42 cm long, rachis 25—66 cm long, both as¬
cending, tomentose, especially at the base, armed
with short stout prickles 2-5 mm, which decrease
toward the apical part of the rachis, occasionally
mildly twisted. Leaflets 24—62, linear-lanceolate,
opposite to subopposite, fiat, coriaceous, margins
entire, pubescent when immature, glabrous with
age, adaxial surface dark green, abaxial surface
light green, 16—32.5 cm long, 4—9 mm wide, ve¬
nation visible on ihe abaxial surface, number of
veins 5-9, distance between veins 0.05—1 mm. Mi-
crostrobilus cylindrical to conical, light green to ol¬
ive green upon emergence, light yellow to creamy
yellow when mature, 11-31 cm long, 2. 5-4. 5 cm
diam., peduncle tomentose, 4—5 cm long, 1—1.8 cm
diam. Microsporophylls numerous, inserted spirally
with vertical rows, cuneiform, 1.4— 1.7 cm long,

Novon 9: 410-413. 1999.

Volume 9, Number 3
1999

Perez-Farrera et al.

Ceratozamia alvarezii from Chiapas

411

0.6— 0.9 cm wide, bicornate in the distal part, with
the fertile portion covering Vfc— % of the abaxial sur¬
face excluding the horns. Microsporangia numerous
in sori of 3^1, longitudinally dehiscent. Megastro-
bilus cylindrical to barrel-shaped, olive green upon
emergence, brown ochre when mature, 14.5—19 cm
long, 5.7-10.5 cm diam., peduncle tomentose, 4.5-
6.5 cm long, 1.8-2. 2 cm diam. Megasporophylls
numerous, peltate, inserted spirally on the cone
axis forming apparent vertical rows, 2. 8—4.5 cm
long, 1.5-2. 9 cm wide, distal end hexagonal, bi¬
cornate, with brown ochre tomentum on the lohu-
late part around the horns. Seed ovate, sareotesta
white when immature and creamy yellow when ma¬
ture, sclerotesta smooth and beige in color, 1.7— 2.5
cm long, 1. 7-2.0 cm diam. with 7—9 radial ridges.
Chromosome number 2 n = 16.

Paratype. MEXICO. Chiapas: Sierra Madre <le Chia¬
pas, 2 Oct. 1996, Perez-Farrera 1260 9 (X AC).

Other specimens examined. MEXICO. Chiapas: Sierra
Madre de Chiapas, Perez-Farrera 64, 67 (CHIP), Caslillo-
Hernandez 624. 445 (CHIP). Breedlove 70956. 60509
(CAS).

Ceratozamia alvarezii principally occurs in oak
forest as described by Rzedowski (1978) at an al¬
titude of 950 m. This forest has an overstory of
Quercus magnifolia Nee, Quercus elliptica Nee, and
Pinus ooacarpa Sehiede above a shrubby layer
dominated by Calliandra houstoniana (Mill) Stand-
ley and Canavalia hirsuta (Martens & Galeolti)
Standley. The herbaceous layer consists mainly of
Anthurium cerrobaulense Matuda, Lisinas procerri-
ma (Hackel) Hitchcock, Elaphoglossum sartorii
(Liebmann) Mickel, Polypodium furfuraceum
Schlechtendal & Chamisso, and Polypodium san-
tae-rosae (Maxon) C. Christensen.

Discussion

The specific epithet was chosen to honor the late
Miguel Alvarez del Toro, in recognition of his tire¬
less pioneering efforts in conservation, which led
to the establishment of the first biological reserves
in the state of Chiapas.

Specific locality information for this species is
being purposely omitted to avoid potential illegal
collection and the possible decimation of its pop¬
ulations. We have only located two populations of
this species in a single geographical area in the
Sierra Madre de Chiapas. Even though leaf char¬
acteristics occasionally display morphological sim¬
ilarities between Ceratozamia matudai and C. al¬
varezii (both have flat linear and linear-lanceolate
leaflets), the two species differ with respect to the
growth habit and the female cones. The peduncle

of the female cone of C. alvarezii is thick and erect,
while that of C. matudai is long, thin, and decum¬
bent. Ceratozamia sabatoi Vovides, Vazquez Torres,
Schutzman & Iglesias (Vovides et al., 1993) differs
from C. alvarezii by having wider, longer leaflets
and blue-green to blue-brown megastrobili. Cera¬
tozamia norstogii is a much larger plant with an
unbranching trunk and a spirally twisted leaf ra-
chis. The consistent differences in cone and vege¬
tative habit led us to consider C. alvarezii as a sep¬
arate species. The chromosome count for the new
species (2 n — 16) appears consistent with members
of this genus investigated so far. We consider C.
alvarezii to be endangered, since the oak forests in
which it occurs are being transformed for agricul¬
tural expansion.

The following key separates Ceratozamia alva¬
rezii from C. matudai, C. sabatoi, and C. norstogii.

I)i \<;nostic Kky

la. Leaf rachis spirally twisted . C. norstogii

II). Rachis not twisted.

2a. Leaflet articulation yellow . C. matudai

2b. Leaflet articulation green.

3a. Leaflet lanceolate, narrowly obovate to
subulate; leaflet veins 9 to 14 ....

. C. sabatoi

3b. Leaflet linear lanceolate; leaflet veins 5

to 9 . C. alvarezii

Due to the morphological characteristics that the
new species presents, coriaceous linear-lanceolate
leaflets with long-attenuate tips, we include it with¬
in the second group within Ceratozamia described
by Stevenson et al. (1986), which includes C. mexi-
cana Brongniart, C. zarogozae Medellin, C. matu¬
dai, C. kuesteriana Regel, C. robusta Miquel, C.
sabatoi, and, in our opinion, C. whitelockiana
Chemnick & Gregory.

Acknowledgments. The principal author wish¬
es especially to thank Angel Hernandez and his
son Jairo Hernandez Leon who helped and guided
us in our travels in the mountains of the Sierra
Madre de Chiapas. The authors also thank Jesus
de la Cruz Rodriguez, Rigoberto Hernandez Jo-
napa, Armando Dominguez Cruz, M.V.Z. Carlos
Tejeda Cruz, and Nevin Coutino for assistance
with fieldwork and Oscar Farrera Sarmiento for the
identification of some of the herbs. Finally, the
authors express their gratitude for the grant sup¬
port of CONABIO project No. FB177A/C120/94,
Fondo Mexicano para la Conservacion de la Na-
turaleza project No. B2-134, Fauna and Flora In-
ternacional project No. 96/64/15, and MAB-Trop-
ical Ecosystem Directorate, all of which partially

412

Novon

5 cm

Volume 9, Number 3
1999

Perez-Farrera et al.

Ceratozamia alvarezii from Chiapas

413

financed exploration and monitoring studies of the
cycads of the Sierra Madre de Chiapas. The sec¬
ond author acknowledges CONACYT project No.
1837P-N9507 for travel to Chiapas and partici¬
pation in some of the explorations. The authors
also thank Edmundo Saavedra for the excellent
botanical illustrations and the reviewers of Novon,
G. J. Breckon and D. W. Stevenson, for useful
comments and suggestions for improving the man¬
uscript.

Literature Cited

Rzedowski, J. 1978. La vegetacion de Mexico. Limusa,
Mexico City.

Stevenson, I).. S. Sabato & M. Vazquez Torres. 1986. A
new species of Ceratozamia (Zamiaceae) from Veracruz.
Mexico, with comments on species relationships, habi¬
tats, and vegetative morphology in Ceratozamia. Brit-
tonia 38: 17—26.

Vovides, A. P., M. Vazquez Lorres, B. Schutzman & C.
Iglesias. 1993. A new species of Ceratozamia (Zami¬
aceae) from (jueretaro and Hidalgo, Mexico. Novon 3:
502-506.

f igure 1. Ceratozamia alvarezii Perez-Farrera, Vovides & Iglesias. — a. Adult plant. — b. Detail of petioles. — e— e.
Leaf details. — f. Cataphyll. — g. Microstrobilus. — h. Detail of mierosporophyll and microsporangia. — i. Megastrobilus.
— j. Megasporophyll. — k. Seed. — 1. Seed with selerotesta removed. Based on living collection (from type locality) at
Clavijero Botanic Garden, Xalapa; accession number 96-012.01. Female cone, Perez Farrera 1261) (XAL); male cone,
Perez-Farrera s.n. (XAL).

Wimmerella, a New South African Genus of
I .obel ioideae (Campanulaceae)

Luis Serra and Manuel B. Crespo

Departamento de Ciencias Ambientales y Reeursos Naturales (Botanica),
Universidad de Alicante, Apartado 99, E-03080 Alicante, Spain, crespo@carn.ua.es

Thomas G. hammers

Department of Botany, Field Museum of Natural History, Chicago, Illinois 60605-2496, U.S.A.
Current address: Department of Biolog)' and Microbiology, University of Wisconsin Oshkosh,
Oshkosh, Wisconsin 54901, U.S.A. lammers@uwosh.edu

ABSTRACT. The 10 South African species cur¬
rently assigned to Laurentia (an otherwise Mediter¬
ranean genus) are segregated here as the genus
Wimmerella on the basis of their basal (vs. medial)
bracteoles; larger flowers and fruits; and subglobose
(vs. ellipsoid) seeds lacking a strophiole, which are
sulcate with flattened (vs. keeled) walls.

Laurentia Adanson, as circumscribed by Wim-
mer (1953, 1968), was a genus of oddly discontin¬
uous distribution. Its 27 species were restricted to
either the Mediterranean region (3 spp.). South Af¬
rica (10 spp.), Australia (10 spp.), western North
America (3 spp.), or the West Indies (1 sp., spread
in historic times throughout much of the tropics).
Phytogeographic coherence of the infrageneric taxa
recognized in Wimmer’s treatment was no better.
The Mediterranean and South African species and
two of the species from North America were as¬
signed to section Laurentia ; the remaining North
American endemic to section Palmerella (A. Gray)
E. Wimmer; and the Australian and West Indian
species to section Isotoma (R. Brown) Endlicher.

However, this very broadly construed Laurentia
was something of an innovation with Wimmer.
Many of the species had been assigned to smaller
genera by earlier authors, and that is the classifi¬
cation preferred by most current workers (e.g.,
MeVaugh, 1940a, 1943; Mason, 1957; Melville,
1960; McComb, 1970; Elliot & Jones, 1990; Chap¬
man, 1991; Skog, 1991; Lammers, 1993; Morin,
1993; Kartesz, 1994). Here, the Australian species
comprise Isotoma (R. Brown) Lindley, the sole West
Indian species is segregated as Hippobroma G.
Don, and the North American species are assigned
to Porterella Torrey and Palmerella A. Gray. As
noted by Lammers (1997), this leaves only the
Mediterranean and South African species in Lau¬

rentia. The present paper executes the final dis¬
memberment of Laurentia, by segregating the spe¬
cies of these two regions into separate genera.

Evidence supporting the generic distinctness of
the Mediterranean and South African species was
provided in detail by Serra and Crespo (1997) and
Crespo et al. (1998). These authors placed partic¬
ular emphasis on differences in seed morphology:
seeds of the South African species are subglobose,
lack a strophiole, and are sulcate with flattened
walls; those of the Mediterranean species are ellip¬
soid, strophiolate, and sulcate with keeled walls.
The extreme value of seed features in Lobelioideae
was first stressed by MeVaugh (1936, 1940b) and
recently expanded upon by Murata (1992, 1995).
The two groups of species were also distinguished
by bracteole position (basal in South Africa, medial
in the Mediterranean) and by the larger flowers and
fruits of the former. Furthermore, in the Mediter¬
ranean species, the plants are erect and the flowers
solitary in an axillary position (appearing terminal
in rosulate species). In contrast, the South African
species have decumbent stems with solitary axillary
flowers; or if the stems are erect, then the flowers
are borne in a 2— 15-flowered terminal raceme. As
these differences are consonant with differences
used to distinguish genera in the subfamily, Serra
and Crespo (1997) and Crespo et al. (1998) rec¬
ognized the Mediterranean and South African spe¬
cies as two distinct genera.

But what names should these genera bear? Mei-
kle (1979) published a brief note contending that
the name Laurentia was a superfluous renaming of
Lobelia. In its place, he adopted the name Solenop-
sis C. Presl. Recently, however, Lammers (1997)
presented evidence that Laurentia was not illegiti¬
mate, and formally proposed that the name be con¬
served to ensure stability. The type of Laurentia is

Novon 9: 414-418. 1999.

Volume 9, Number 3
1999

Serra et al.

Wimmerella from South Africa

415

Lobelia laurentia L. [ — Solenopsis laurentia (L.) C.
Presl, = Laurentia gasparrinii (Tineo) Strobl], while
the lectotype of Solenopsis (designated by Pfeiffer,
1874: 1192) is Lobelia minuta L. [= Solenopsis
minuta (L.) C. Presl, = Laurentia minuta (L.) A.
DC.]. Both species belong to the Mediterranean ge¬
nus, which will thus be called either Laurentia or
Solenopsis , depending on the outcome of Lammers’s
(1997) proposal.

For the South African genus, Serra and Crespo
(1997) and Crespo et al. (1998) took up the name
Enchysia C. Presl, which had been cited as a syn¬
onym ol Laurentia by Wimmer (1953). When pub¬
lished by Presl (1836: 40), Enchysia included six
species. One of these, E. scapigera (R. Brown) C.
Presl, is now treated (Elliot & Jones, 1990; Chap¬
man, 1991) as a member of the Australian genus
Isotoma [i.e., /. scapigera (R. Brown) G. Don], while
three others (E. baueri C. Presl, E. gaudichaudii C.
Presl, and E. lessonii C. Presl) are all synonyms of
the related /. fluviatilis (R. Brown) F. Mueller ex
Bentham (McComb, 1970). The last two of the six

original species were from South Africa, but are not
referable to Laurentia sensu Wimmer (1953, 1968).
Enchysia repens (Thunberg) C. Presl (based on Lo¬
belia repens Thunberg) is a synonym of Lobelia an-
ceps L. f. (Thulin, 1983), while Enchysia erinoides
(L.) C. Presl (based on Lobelia erinoides L.) is a
synonym of Lobelia erinus L. (Thulin et al., 1986).
Because it was this last species, L. erinoides , that
was designated as the lectotype of Enchysia bv
Pfeiffer (1874: 1199), that generic name is properly
a synonym of Lobelia L. and cannot be used for the
South Alriean species formerly assigned to Lauren¬
tia. Even if Pfeiffers choice were to be overturned,
there is no way to lectotypify this name that would
permit its use for South African Laurentia ; if it is
not a synonym of Lobelia , it will be a synonym of
Isotoma.

A thorough survey of all generic names referable
to Lobelioideae shows that none can be typified on
the basis of one of the South African species of
Laurentia. Therefore, a new genus is erected to ac¬
commodate these species. The key below may be
used to distinguish the new genus from its allies.

Key to the Genera Included in Laurentia s.e.

la. Corolla scarcely zygomorphic; filament tube fully adnate to the corolla.

2a. Calyx lobes 2—10 mm long; corolla 8—30 mm long; filament tube 3—20 mm long; anther tube black, the

ventral two setose at apex (Australia) . Isotoma

2b. Calyx lobes 10-22 mm long; corolla 60-150 mm long; filament tube 55-145 mm long; anther tube white,

all five setose at apex (pantropical weed, originally endemic to West Indies) . Hippobroma

lb. Corolla distinctly bilabiate, the dorsal lip 2-lobed, the ventral 3-lobed; filament tube free from corolla, or only
the two dorsal filaments adnate.

3a. Corolla 20-30 mm long; filament tube 9.5-17 mm long, the two dorsal filaments adnate to the corolla

(western North America) . Palmerella

3b. Corolla 3—20 mm long; filament tube 1.5—7 mm long, free from corolla.

4a. Stems fleshy; calyx lobes 3 — 8( — 1 1 ) mm long; capsules 5— 10(— 16) mm long; seeds ca. 1 mm long

(western North America) . Porterella

4b. Stems herbaceous; calyx lobes 1-4 mm long; capsules 1-6 mm long; seeds 0.3-0.5 mm long.

5a. Stems erect, the flowers solitary and axillary or terminal; pedicels with 1—3 bracteoles near the
middle; filament tube 1.5-2. 5 mm long; capsules 1-3 mm long; seeds ellipsoid, strophiolate,

sulcate with keeled walls (Mediterranean region) . Solenopsis (= Ixiurentia , nom. cons, prop.)

5b. Stems decumbent, the flowers solitary and axillary, or if erect, the flowers 2-15 in a terminal
raceme; pedicels bibracteolate at base; filaments 2—6 mm long; capsules 2.5—6 mm long; seeds
subglobose, lacking a strophiole, sulcate with flattened walls (South Africa) . Wimmerella

Wimmerella L. Serra, M. B. Crespo & Lammers,
gen. nov. TYPE: Wimmerella secunda (L. f.) L.
Serra, M. B. Crespo & Lammers.

A Solenopsis (= Laurentia s. str.) bracteolis basalibus,
floribus capsulisque majoribus, atque seminibus subglo-
bosis strophiolo destitutis testa sulcata cum parietibus ap-
planatis differ!.

Annual or perennial plants up to 30(^40) cm
high. Stems decumbent or erect, simple or
branched, glabrous or pubescent. Leaves cauline,
alternate, in some species only rosulate, sessile or
petiolate; lamina orbicular to linear, entire or with

5—9 teeth or lobes, 3-40 mm long, 2-12 mm wide.
Flowers solitary in leaf axils or (in plants with erect
stems) 2—15 in terminal racemes up to 25 cm long;
bracts linear to lanceolate, 5-30 mm long, 1—3 mm
wide; pedicels (2— )5— 20 mm long in fruit. Calyx
lobes oblong, lanceolate or linear-lanceolate, entire,
1—3 mm long, glabrous or pubescent. Corolla white
or blue (sometimes with a white patch in the throat),
3—18 mm long, bilabiate with 2 linear dorsal lobes
and 3 oblong ventral lobes or nearly regular with 5
subequal linear-oblong lobes, glabrous; tube fun¬
nel-shaped or tubular. Staminal column free from

416

Novon

the corolla; filaments 2-6 mm long; anther tube
0.8-1 mm long, light black, glabrous or short pu¬
bescent distally, bisetose at the apex of the ventral
pair. Ovary inferior, bilocular; placentation axile.
Capsule apically dehiscing via two valves, 2.5—6
mm long, the calyx persistent. Seeds subglobose,
lacking a strophiole, light brown, smooth and lus¬
trous, 0.4— 0.5 mm long, sulcate with flattened
walls.

Wimmerella is endemic to South Africa. Popu¬
lations occupy diverse habitats from the coastal re¬
gions to the central plateaus, at elevations from sea
level up to 1800 m. After study of over 100 spec¬
imens (see Acknowledgments for herbaria), it was
decided to follow the treatment of Wiminer (1953,
1968) and Welman (1993), in which 10 species are
recognized. All are confined to the Cape Province
with the exception of W. arabidea , which also oc¬
curs in Natal (Welman, 1993).

The names of two species originally described

by Wimmer (1968) and recognized by Welman
(1993) were not validly published. Wimmer failed
to designate a type as defined by Article 8.1 of the
Code (Greuter et al., 1994), an oversight perhaps
attributable to the fact that his manuscript was not
published until seven years after his death. Two
duplicates of one gathering were cited for each
name without singling one out as the holotype, thus
violating Article 37.3. These two species are here
named by designating a holotype and providing an
explicit reference to Wimmer’s (1968) effectively
published Latin diagnosis, as allowed by Article
32.4. [See note added in proof.]

Etymology. The genus is named for Franz El-
fried Wimmer (1881—1961), the Austrian bota¬
nist and Roman Catholic priest who was the fore¬
most twentieth century student of the
Lobelioideae (Rechinger, 1961). The honorific is
rendered as a diminuitive to avoid homonymy
with Wimmeria Schlechtendal (Celastraceae) and
to parallel the related genera Palmerella and
Porterella.

Key to the Species ok Wimmkreua

la.

lb.

Leaves suborbicular or corclate-reniform.

2a. Stems erect or decumbent, glabrous; pedicels 2—7 mm long; corolla 3 mm long . W. frontidenlata

2b. Stems decumbent or prostrate, pubescent; pedicels 8—18 mm long; corolla 3—8 mm long.

3a. Leaves very sparsely pubescent with trichomes ca. 0.4 mm long, the margin 5—9-toothed (rarely

subentire) . lb pygmaea

3b. Leaves densely pubescent with trichomes ca. 0.2 mm long, the margin 5— 9-lobed.

4a. Pedicels 8—1 I mm long, pubescent throughout; corolla 3—4 mm long . lb hederacea

4b. Pedicels 12—18 mm long, glabrous or pubescent only at the base; corolla 6— 8 mm long .

. tb hedyotidea

Leaves lanceolate, linear-lanceolate, lanceolate-spatulate, or oblong-spatulate.

5a. Leaves pubescent (rarely glabrous); bracts oblong or linear, longer than the pedicel; corolla white, 4—7

mm long . lb secunda

51). Leaves glabrous; bracts linear, up to half the length of the pedicel; corolla blue, sometimes with a white
patch in the throat, 6-18 mm long.

6a. Corolla 18 mm long . lb longitubus

6b. Corolla 6— 12 mm long.

7a. Leaves oblong-spatulate, incised-dentate; corolla I 1 — 12 mm long . 14! giflbergensis

7b. Leaves linear-lanceolate or lanceolate-spatulate, entire or denticulate; corolla 6—10 mm long.

8a. Leaves I inear- lanceolate; corolla 6 mm long; capsule 2.5 mm long . lb marine

8b. Leaves lanceolate-spatulate; corolla (6— )7— 10 mm long; capsule 4 mm long.

9a. Leaves 5—10 mm long; inflorescence 2— 5-flowered . 14. bifida

9b. Leaves 10—30 mm long; inflorescence 5— 1 5-flowered . lb arabidea

Wiiiimerella arabidea (C. Presl) L. Serra, M. B.
Crespo & Lammers, comb. nov. Basionym:
Rapuntium arabideum C. Presl, Prodr. Mon-
ogr. Lobel. 18. 1836. Laurentia arabidea (C.
Presl) A. DC., in DC., Prodr. 7: 410. 1839.
Lobelia arabidea (C. Presl) Steudel, Nomencl.
Bot. (ed. 2) 2: 59. 1841. TYPE: South Africa.
Cape Province: “Cap B. Sp. in parte inferiore
occidentali,” Anonymous s.n. (lectotype, des¬
ignated by Wimmer (1968: 853), PR not
seen).

Wimmerella bifida (Thunberg) L. Serra, M. B.
Crespo & Lammers, comb. nov. Basionym:
Lobelia bifida Thunberg, Prodr. FI. Cap. 40.
1794. Rapuntium bifidum (Thunberg) C.
Presl, Prodr. Monogr. Lobel. 30: 1836. Lau¬
rentia bifida (Thunberg) Sonder, in Harvey &
Sonder, FI. Cap. 3: 552. 1865. TYPE: South
Africa. Cape Province: “Am Rande des Ber-
ges Bockland,” Thunberg s.n. (lectotype, des¬
ignated by Wimmer (1953: 394), UPS not
seen).

Volume 9, Number 3
1999

Serra et al.

Wimmerella from South Africa

417

Wimmerella frontidenlata L. Serra, M. B. Cres¬
po & Lammers, sp. nov. TYPE: South Africa.
Cape Province: Swellendam Div., Anysberg,
sheltered S side rocks, 1600 m, 21 May 1950,
Esterhuysen 17070 (holotype, PRE; isotype,
BOL). [Validated by reference to the effective¬
ly published Latin description of “Laurent ia
front identata" E. Wimmer, Pflanzenr. IV.276c:
855. 1968, nora. invalid.]

Wimmerella giftbergensis (E. Phillips) L. Serra,
M. B. Crespo & Lammers, comb. nov. Basio-
nym: Lobelia giftbergensis E. Phillips, Ann. S.
African Mus. 9: 121. 1913. Laurentia giftber¬
gensis (E. Phillips) E. Wimmer, Reperl. Spec.
Nov. Regni Veg. 38: 77. 1935. TYPE: South
Africa. Cape Province: Van Rhynsdorp Div.,
Giftberg Range, 1000-2000 ft., Sep. 1911,
Phillips 7599 (holotype, K).

Wimmerella liederacea (Sonder) L. Serra. M. B.
Crespo & Lammers, comb. nov. Basionym:
Laurentia hederacea Sonder, in Harvey & Son¬
der, FI. Cap. 3: 553. 1865. TYPE: South Af¬
rica. Cape Province: “C. B. S., Hab. Eastern
frontier,” Hutton s.n. (holotype, K).

Wimmerella herlyotidea (Schlechter) L. Serra,
M. B. Crespo & Lammers, comb. nov. Basio¬
nym: Laurentia hedyotidea Schlechter, Bot.
Jahrb. Syst. 27: 197. 1899. TYPE: South Af¬
rica. Cape Province: in regioni austro-occiden-
tali, in saxosis montium pone Bainskloof, in
ditione Ceres, alt. ca. 4000 ped., 11 Nov.
1896, Schlechter 9104 (lectotype, designated
by W immer (1953: 396), B not seen: isolec-
totype, K).

Wimmerella longitubus (E. Wimmer) L. Serra,
M. B. Crespo & Lammers, comb. nov. Basio¬
nym: Laurentia longitubus E. Wimmer, Report.
Spec. Nov. Regni Veg. 22: 193. 1926. TYPE:
South Africa. Cape Province: Cape of Good
Hope, in humidis Langebergen prope Rivers-
dale, Schlechter 1902 (holotype, W U not seen).

Wimmerella mariae L. Serra, M. B. Crespo &
Lammers, sp. nov. TYPE: South Africa. Cape
Province: Worcester Div., plateau between Ma-
troosberg and Sonklip Peak, in dry mud on
floor of small pans, ca. 1800 m, 17 Jan. 1959,
Esterhuysen 28132 (holotype, BOL; isotype, W
not seen). [Validated by reference to the effec¬
tively published Latin description of “Lauren¬
tia mariae” E. Wimmer, Pflanzenr. IV.276c:
854. 1968, nom. invalid.]

Wimmerella pyginaea (Thunberg) L. Serra, M. B.
Crespo & Lammers, comb. nov. Basionym: Lo¬
belia pygmaea Thunberg, Prodr. FI. Cap. 40.
1794. Rapuntium pygmaeum (Thunberg) C.
Presl, Prodr. Monogr. Lobel. 22. 1836. Lauren¬
tia pygmaea (Thunberg) Sonder, in Harvey &
Sonder, FI. Cap. 3: 553. 1865. TYPE: South
Africa. Cape Province: “Kapland, ostliehes
Gebiet, in Querspalten des Berges Ribeck-
Kastel,” Thunberg s.n. (lectotype, designated
by Wimmer (1953: 397), S not seen).

W immerella seeimda (L. f.) L. Serra, M. B. Cres¬
po & Lammers, comb. nov. Basionym: Lobelia
secunda L. f., Suppl. PI. 395. 1782. Enchysia
secunda (L. f.) Sonder, in Harvey & Sonder,
FI. Cap. 3: 551. 1865. Laurentia secunda (L.
f.) Kuntze, Revis. Gen. PI. 3(2): 188. 1898.
TYPE: South Africa. Cape Province: “Cap,”
Herb. Linn. 1051.17 (lectotype, designated by
Crespo et al. (1996: 119), LINN).

Acknowledgments. We are grateful to the staffs
of BM. BOL, K. LINN, and PRE for their kind as¬
sistance. This study was partly supported by grants
to Serra from Bancaja Foundation (Spain) and to
Crespo from the Spanish government’s Ministerio
de Educaeion y Ciencia (PR 95/011).

Literature Cited

Chapman, A. I). 1991. Australian Plant Name Index D-
J. Australian Government Publishing Serv ice. Canberra.
Crespo, M. 8., I.. Serra & A. Juan. 1998. Solenopsis (Lob-
eliaeeae): A genus endemic in the Mediterranean Re¬
gion. PI. Svst. Evol. 210: 21 1—229.

- , - & N. Turland. 1996. Lectotypification ol

four names in Lobelia (Lobel iaceae). Taxon 45: 117—
120.

Elliot, W. R. & I). L. Jones. 1990. Encyclopedia of Aus¬
tralian Plants Suitable for Cultivation. Vol. 5. Lothian,
Melbourne.

Greuter, W. F. R. Barrie, H. M. Runlet. W. G. Ghaloner.
V. Demoulin, D. L. Hawksworth, P. M. J0rgensen. 1). H.
Nieolsou. P. C. Silva, P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Kartesz. J. T. 1994. A Synonymized Checklist of the Vas¬
cular Flora of the United States, Canada, and Green¬
land, 2nd ed., Vol. I. Timber Press, Portland. Oregon.
Lammers, T. G. 1993. Chromosome numbers of Campan-
ulaceae. III. Review and integration ol data lor subfam¬
ily Lobelioideae. Amer. J. Bot. 80: 660-675.

- . 1 997. Proposal to conserve the name iMiirentia

(Campanulaceae) with a conserved type. Taxon 46:

795-799.

Mason. H. L. 1957. A Flora of tin1 Marshes of California.

Univ. California Press, Berkeley.

McComb, J. A. 1970. A revision of the species Isotoma
fluviatilis. Contr. New South Wales Natl. Herb. 4: 1 ( —
111.

McVaugh, R. 1936. Stuilies in the taxonomy and distri-

418

Novon

Inition of tin' eastern North American species of Lobelia.
Rhodora 38: 241-263, 276-298, 305-329, 346-362.

- . 1940a. A revision of “Laurentia” and allied gen¬
era in North America. Bull. Torrey Hot. Club 67: 778—
798.

- . 1940b. A key to the North American species of

lobelia (sect. Hemipogon). Amer. Midland Naturalist
24: 681-702.

- . 1943. Campanulaceae (Lobelioideae). N. Amer.

FI. 32 A: 1-134.

Meikle, R. I). 1979. Some notes on Laurentia Adanson
(Campanulaceae). Kew Bull. 34: 373-375.

Melville. R. [960. Contributions to the flora of Australia:
VI. The pollination mechanism of Isotoma axillaris
Lindl. and the generic status of Isotoma Lindl. Kew
Bull. 14: 277-279.

Morin, N. 1993. Porterella. P. 468 in J. C. Hickman (ed¬
itor), The Jepson Manual: Higher Plants of California.
Univ. California Press, Berkeley.

Murata, J. 1992. Systematic implication of seed coat mor¬
phology in Lobelia (Campanulaceae -Lobelioideae). J.
Fae. Sci. Univ. Tokyo, Sect. 3, Bot. 15: 155—172.

- . 1995. A revision of infrageneric classification of

Ijobelia (Campamdaeeae-Lobelioideae) with special ref¬
erence to seed coat morphology. J. Fac. Sci. Univ. To¬
kyo, Sect. 3, Bot. 15: 349—371.

Pfeiffer. L. K. G. 1874. Nomenclator Botanicus, Vol. 2.
Theodore Fischer, Kassel.

Presl, C. B. 1836. Prodromus Monographiae Lobeliacear-
um. Theophilus Haase, Prague.

Rechinger, K. H. 1961. Franz Elfried Wimmer 1881 —
1961. Taxon 10: 239-240.

Serra. L. & M. B. Crespo. 1997. An outline revision of

the subtribe Siphocampylinae (Lobeliaeeae). Lagascalia
19: 881-888.

Skog. L. 1991. Lobeliaeeae. Pp. 129-131 in I). II. Nic-
olson, flora of Dominica, Part 2: Dicotyledonae. Smith¬
sonian Contr. Bot. 77: 1-274.
rhulin, M. 1983. Lobeliaeeae. Pp. Ilf)— 157 in K. Launert
(editor), f lora Zambesiaca, Vol. 7(1). Managing Com¬
mittee, London.

- . P. B. Phillipson & I). (). Wijnands. 1986. Typi-

fication of Lobelia erinus L. and Lobelia erinoides L.
Taxon 35: 725—729.

Welman. W. G. 1993. Lobeliaeeae. Pp. 693-698 in T. 11.
Arnold & B. C. I)e Wet (editors.), Plants of Southern
Africa: Names and Distribution. National Botanical In¬
stitute, Pretoria.

Wimmer, f. E. 1953. Campamdaeeae-Lobelioideae 11.
Teil. Pp. i— viii. 261—814 in II. Stubbe & K. I. Noack
(editors). Das Pflanzenreich IV.276b. Akademie-Verlag,
Berlin.

- . 1968. Campanulaceae-Lobelioideae supplemen-

tum et Campanulaceae-Cyphioideae. Pp. i— x, 815—1024
in H. Stubbe (editor). Das Pflanzenreich IV.276e. Aka¬
demie-Verlag. Berlin.

Note added in proof. Because Article 37.3 was
amended by adoption of Proposition 80 (Barrie, Taxon 47:
{581—889. 1998) at the August 1999 International Botan¬
ical Congress, the names Laurentia frontidentata E. Wim¬
mer and L. marine E. Wimmer are validly published. They
are the basionyms of the new combinations Wimmerella
frontidentata (E. Wimmer) L. Serra, M. B. Crespo & ham¬
mers and W. marine (E. Wimmer) L. Serra, M. B. Crespo
& hammers, respectively.

Alloplectus martinianus, a New Species of
Gesneriaceae from Ecuador

James F. Smith

Department of Biology, 1910 University Drive, Boise, Idaho 83725, U.S.A.

ABSTRACT. Alloplectus martinianus, a new species
of Gesneriaceae from Ecuador, is described. The
species is distinctive for its long narrow leaves,
densely villous short corollas, and unusual reticu¬
lations on the adaxial epidermis of the leaf.

This paper describes a new species of Alloplectus
from Ecuador that is readily distinguished because
of the combined character states of long narrow
leaves, short pedicels, densely villous short corol¬
las, and unique reticulations on the adaxial epi¬
dermis of the leaves.

Alloplectus Martins has a history of taxonomic
confusion with Columnea L. (Wiehler, 1983), which
lias not been resolved with modern molecular tech¬
niques (Smith & Sytsma, 1994a, b, c; Smith & Car-
roll, 1997). These two genera are distinguished by
the dehiscent fleshy fruits in Alloplectus and inde-
hiscent berries in Columnea (Wiehler, 1973, 1983).
This diagnostic fruit character is useful on fresh
and live material, but plants without fruits are fre¬
quently difficult or impossible to identify. In addi¬
tion, herbarium specimens with only immature
dried fruits are difficult to determine as being de¬
hiscent or not. Such confusion has been the case
lor the species described here. Specimens of this
species have been attributed to Columnea by the
author and by other specialists in the Gesneriaceae.
The occurrence of fruits in one specimen ( Gentry
12590) shows the dehiscence typical of Alloplectus
and has permitted the proper placement of this spe¬
cies.

Alloplectus martinianus J. F. Smith, sp. nov.
TYPE: Ecuador. Napo: road from Baeza to
Tena, 5 km S of Baeza, mountain rain forest,
26 June 1983, Bohlin A’ Bohlin 698 (holotype,
US; isotype, GB). Figure 1.

Differt ah Alloplecto boliviano (Britton) Wiehler pedi-
cellis brevioribus, foliis angustioribus.

Epiphytic or terrestrial, woody-based herbs with
erect stems to 1.5 m long, 3—8 mm diam. (Fig. 1A),
red-brown, appressed trichomes sericeous, denser
toward apex. Internodes 0.4— 5.0 cm long, leaf scars
raised from the nodes. Leaves opposite, subequal.

blades lanceolate, 4.0—13.9 X 1. 1-3.3 cm, apices
acute, bases cuneate, margin entire or serrulate,
adaxially green, pannose with reticulations on epi¬
dermis (Fig. IF) (reticulations may not be promi¬
nent on all leaves), abaxially green to rose, densely
tomentose with yellow trichomes, pubescence of
veins abaxially colored pink or red-purple. Petioles
0. 5-2.4 cm, tomentose. Inflorescences of 2—5 flow¬
ers in axils of both leaf pairs (Fig. 1A), floral bracts
lanceolate, 3.0 X 1-1.3 mm, apex acute, margin
entire, tomentose. Pedicels 0.5— 0.9 cm, erect in
leaf axil, tomentose. Calyces campanulate, clasping
corolla at base, lobes elliptic, equal in size, 0.8—
1.8 X 0.20-0.50 cm, apex acute, margins entire or
serrulate at apex, green, exterior and interior to¬
mentose. Corollas yellow, gibbous at base, slightly
ventricose (Fig. IB), 1.3-1. 7 cm long, 1.5— 3.0 mm
at base, 3.5— 3.8 mm at widest point, 2. 5-2. 6 mm
before limb, exterior densely villous with golden-
yellow trichomes, interior mostly glabrous but with
white capitate trichomes dorsally and distally (Fig.
1C). Lobes equal, 2.0 X 2.0 mm, semi-orbicular.
Stamens included, filaments white, glabrous, ad-
nate to corolla base for 2.5 mm then connate for
3.5 mm at base, anthers quadrate, 1.0— 2.0 X 1.0—
2.0 mm. Nectaries of 2 dorsal fused glands. Ovaries
tomentose, 3.0 mm long, conical, styles reddish,
glabrous to slightly pubescent at base, stigmas sto-
matomorphic, papillose. Fruits globose, 6-10 mm
diam., red, villous. Seeds fusiform, red-brown to al¬
most black, twisted, striate, 0.8 mm long.

Distribution. Known only from 1200 to 2500 m
elevation in Ecuador from a small geographic area
in Napo province; flowering March to December
and fruiting in May to November.

Alloplectus martinianus is superficially and mor¬
phologically similar to Columnea formosa (C. V.
Morton) C. V. Morton, to which specimens were ear¬
lier attributed. It differs from the latter species in
several respects. First of all, the fruits are appar¬
ently dehiscent, although this character state was
not immediately obvious from the initial specimens
examined. However, it was clear from the fruits of
Gentry 12590 that this is a species of Alloplectus.
Additionally, the pedicels of the new species are

No VON 9: 419-421. 1999.

420

Novon

Figure 1. Alloplectus rnartinianus J. F Smith. — A. Habit. — B. Flower with ealyx cut open to show corolla base. — C.
Corolla interior with stamens. The bases ol the stamens were torn in the specimen this illustration was based on. and
therefore do not reflect the lull degree of connation seen in other flowers (see text for full description). — I), Gynoecium
with calyx and nectaries. — E. Uniseriate trichome. — F. Adaxial leaf surface. From Bohlin & liohlin 6 9B.

Volume 9, Number 3
1999

Smith

Alloplectus martinianus from Ecuador

421

conspicuously shorter than in C. formosa and other
species of Columned , and the corolla is smaller and
shorter than in most species of Alloplectus. Allo¬
plectus martinianus is also superficially similar to
Alloplectus bolivianos (Britton) Wiehl. However, the
latter species has longer pedicels, corollas, and pet¬
als and broader leaves. Additionally, A. bolivianus
is generally much less pubescent than A. martini¬
anus. Lastly, the reticulate epidermal pattern read¬
ily seen on many specimens at 10X magnification
on the adaxial surface of the leaves is distinctive
and not known from any other species of Alloplec¬
tus. It should be noted that this feature is not al¬
ways present on all leaves of a collection. The pat¬
tern may be influenced by the age and condition of
the leaf at the time of collection, as well as how it
is dried. However, all specimens examined to date
show at least a portion of one leal from each spec¬
imen with the reticulations. The specific epithet
honors the author’s friend and companion, Steven
G. Martin, of Boise, Idaho.

Paratypes. ECUADOR. Napo: Cosanga, 20 km S of
Baeza, 20 Sep. 1977. Maas, Berg & ter Welle 3008 (QCA.
SEI„ U); Rfo Panteor. SW of Borja, 22 Sep. 1980. Holm-
Aielsen et al. 26710 (K); Quijos. Sierra Azul. 2 May 1992.
Alvarez. Mosquera, Proafio & Uteri 396 (QCN E. US); 16.5
km NNE of Santa Rosa on road from Baeza to Eago Agrio,
19 Oct. 1971, MacBryde 747 (US); junction Baeza-Eago
Agrio road with Rfo Azuela, 27 Mar. 1972. Dwyer &
MacBryde 9636 (US): 28 km E of El Chaco, Quito— Eago
Agrio road. 7 Nov. 1974. Gentry 12590 (NY. US); 2.5 km
N of Cosanga, 18 Feb. 1978. Kirkbride & Chamba 4168
(NY, US); Baeza, 9 May 1990, Palacios & Freire 5002

(QCNE, US); 17 km from Cosanga, 24 Aug. 1990, Jar-
amillo, Grijalva A- Grijalva 12142 (QUA); 1 1 km E of El
Chaco. 22 Mar. 1989, Huttel 1634 (OCA, QCNE); Santa
Lucia tie Bermejo, 15 Dec. 1998, Alvarez , Brito, Romero
& Guerrero 969.4 (QCNE); environs of Baeza, Jan. 1979.
Besse. Tan & Halton 1156 (SEE).

Acknowledgments. I thank Lewis Eslick for the
illustration of this species, Steven G. Martin for his
support, Laurence Skog for comments on the man¬
uscript and for forwarding several specimens for my
study, and an anonymous reviewer for comments on
the manuscript. Funding for this research was made
available through NSF grant DEB-9317775 and a
faculty research associates program grant from Boi¬
se State University.

Literature Cited

J. E. Smith & C. E. Carroll. 1997. A cladistic analysis of
the tribe Episcieae (Gesneriaceae) based on ndh E se¬
quences: Origin of morphological characters. Svst. Bot.
22; 718-724.

- & K. J. Sytsma. 1994a. Evolution in the Andean

epiphytic genus Columnea (Gesneriaceae): Part I. Mor¬
phological variation. Syst. Bot. 19: 220—235.

- & - . 1994b. Evolution in the Andean epi-

phytic genus Columnea (Gesneriaceae): Part 11. Chlo-
roplast DNA restriction site variation. Syst. Bot. 19:
3 1 7-336.

- & - . 1994c. Molecules and morphology:

Congruence of data in Columnea (Gesneriaceae). PI.
Syst. Evol. 194: 37—52.

Wiehler, H. 1973. One hundred transfers from Alloplectus
and Columnea (Gesneriaceae). Phytologia 27: 309—328.
- . 1983. A synopsis ol the neotropical Gesneri¬
aceae. Selbyana 6: 1—219.

A New Combination in the Cactaceae, Opuntia lutea

James C. Solomon

Missouri Botanical Garden, P.O. Box 299. St. Louis, Missouri 63166-0299, U.S.A.

ABSTRACT. A new combination in the Cactaceae,
Opuntia lutea (Rose) Solomon, is needed to bring
one of the species treated in the Flora de Nicara¬
gua into conformity with the generic concepts uti¬
lized in the Flora.

Historically, the genus Nopalea Salm-Dyck has
been separated from Opuntia Miller subg. Opuntia
based on floral morphology. In Nopalea, flowers
have a more or less long-exserted style and stamens
and the inner tepals are relatively short, erect, and
imbricate, instead of spreading as in other species
of Opuntia. This floral syndrome is an adaptation
to hummingbird pollination (Porseh, 1938) and is
not considered by many authors to be sufficient for
generic recognition (Benson, 1982; Hunt, 1967;
Hunt & Taylor. 1986, 1990). Thus, Nopalea would
be considered a subgroup within Opuntia. Because
the Opuntioideae, as a whole, have been inade¬
quately studied and little is known about evolu¬
tionary relationships within the subfamily, it re¬
mains unclear what the relationships are between

the species that have traditionally been included in
Nopalea and the rest of the subfamily. In the Flora
de Nicaragua, Opuntia is treated in its broad sense,
including Nopalea, making necessary the following
new combination:

Opuntia lutea (Rose) Solomon, comb. nov. Bas-
ionym: Nopalea lutea Rose, Contr. U.S. Natl.
Herb. 12: 405, t. 58. 1909. TYPE: Guatemala.
Zacapa: El Rancho, 300 m, 28 Dec. 1907, W.
A. Kellerman 7046 (holotype, US-535175; iso¬
types, MO-911486, US-535186).

Literature Cited

Benson, L. 1982. The Caeti of the United States and Can¬
ada. Stanford Univ. Press, Stanford.

Hunt. I). 1967. Cactaceae. Pp. 427-467 in J. Hutchinson,
The Genera of Flowering Plants, vol. 2. Clarendon
Press, Oxford.

- & N. P. Taylor. 1986. The genera of the Cacta¬
ceae: Towards a new consensus. Bradleya 4: 65—78.
- & - . 1 990. The genera of Cactaceae: Pro¬
gress towards consensus. Bradleya 8: 85-107.

Porseh. (). 1938. Das Bestaubungsleben dec Kakteenbliite
1. Cactaceae 1938(1): 1-80.

Novon 9: 422. 1999.

A N ew Combination in Ampelocissus (Vitaceae), a Victim of
Historic Deforestation in Nicaragua

Warren Douglas Stevens and Amy Pool

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63133-0299, U.S.A.

ABSTRACT. A neotype is selected for lit is javalen-
sis Seemann, and V. javalensis is transferred to Am-
pelocissus. Ampelocissus costaricensis Lundell is re¬
duced to synonymy under Ampelocissus javalensis
(Seemann) W. D. Stevens & A. Pool, comb. nov.
Changes in vegetation of the Javali gold and silver
mines region (Santo Domingo/La Libertad) of Chon-
tales, Nicaragua, the original published locality of
V. javalensis, are discussed.

In 1869, Berthold Seemann described Vitis ja¬
valensis Seemann based on material observed in the
Department of Chontales, Nicaragua in the vicinity
of the Javali gold and silver mines. Living material
was gathered and delivered to horticulturist Wil¬
liam Bull of Kings Road, Chelsea, England (See¬
mann, 1869).

J. E. Planchon transferred Vitis javalensis to Cis¬
sus as Cissus javalensis (Seemann) Planchon, where
it was included in his monograph of Ampelideae
under “ Cissus . — Species pleraque dubiae” as a
“nomen tantum” (Planchon, 1887). Bull's retail list
of new Plants, ann. 1870, p. 5, was cited in refer¬
ence. In J. A. Lombardis manuscript FI. Neotrop.
Monogr. lor Vitaceae-Generos Ampelocissus, Ampe-
lopsis e Cissus (in press), Lombardi treats Cissus
javalensis (Seemann) Planchon under “Nomes du-
vidosos e taxa excluidos.” A holotype, Seemann
s.n., assumed to be at BM, was not located.

In our efforts to prepare a written flora for Nic¬
aragua, we are attempting to account for all names
based on Nicaraguan material. Herbarium material
of Vitis javalensis collected by Seemann was sought
at the British Museum, where the primary set of
Seemann’s Nicaraguan material is housed, without
success. Illustrations of Vitis or Cissus javalensis
published by Seemann or Bull were also sought,
again without success.

Seemann described Vitis javalensis as a species
with simple leaves with acute apices, cordate bases,
mucronately-dentate margins, “beautifully” veluti-
nous adaxial surfaces and purplish abaxial surfac¬
es, and bright scarlet flowers. This description does
not correspond to any Central American species of
Cissus or Vitis. However Seemann’s description is

consistent with Lundell's description (Lundell,
1937) and specimens of Ampelocissus costaricensis
Lundell, with the exception that the hairs of the
leaves of the latter are better described as arach¬
noid or floccose.

The genus Ampelocissus is not otherwise known
from Nicaragua. Ampelocissus costaricensis is a spe¬
cies apparently currently restricted to wet, ever¬
green forests below 500 m in the Provinces of Li-
mon, Costa Rica, and Bocas del Toro, Panama. The
habitat of this region is similar to the presumed
habitat of the Javali mines region of Chontales, Nic¬
aragua, at the time when Seemann made his col¬
lections, 1866-1871. Both Thomas Belt (1888) and
Seemann (Anon., 1869) describe this as an area of
high rainfall and rich forests. We have discovered
15 additional species collected in the region of the
Javali mines during this period. Most of the species
range from Nicaragua into South America and are
currently known from much wetter forests than
those which now occur in the mines region. All but
two of these species, including the one in question,
are currently found in Nicaragua, but they tend to
be found in the wet forests of Zelaya and Rio San
Juan Departments, or in remnant cloud forests of
western Nicaragua, especially Cerro Oluma, Cerro
Mombachito, Volcan Mombacho, and Volcan Mad-
eras.

Belt's and Seemann’s historical narratives, and
the modern distribution of the species historically
known from the Javali mines region, suggest that
the Cordillera Chontelena, from southern Matagal-
pa through Boaco, Chontales, and Rio San Juan
Departments, has become dramatically drier in the
last 150 years. Deforestation, for mining timbers,
charcoal, and grazing land, has clearly changed the
landscape directly and has probably indirectly
changed the general climate of the region. What
was once apparently wet, evergreen forest is now
mostly scrubby savanna with patches of dry forest
on roc ky slopes, grading into cloud forests at higher
elevations. The vegetation of the southernmost part
of the Cordillera, in Rio San Juan Department, is
less altered but not well known. It is possible that
future collection in this area will result in the re-

Novon 9: 423^124. 1999.

424

Novon

discovery in Nicaragua of species historically
known from the Javalf mines region.

Nkotypification and Synonymy

We assume, due to the failure of our own search¬
es in addition to the failed previous attempts by
Planchon and Lombardi, that it is unlikely that any
original material of Vitis javalensis exists, and the
selection of a neotype is therefore necessary (Greu-
ter et ah, 1994, Art. 9.6). On the basis of Seemann ’s
description, a neotype is here selected, V. javalensis
is transferred to Ampelocissus, and Ampelocissus
costaricensis is placed in synonymy.

AiiipelocisHiis javalensis (Seemann) W D. Ste¬
vens & A. Pool, comb. nov. Basionym: Vitis
javalensis Seemann, J. Bot. 7: 332. 1869. Cis-
sus javalensis (Seemann) Planchon in A. DC.
& C. DC., Monogr. Phan. 5: 624. 1887. TYPE:
Costa Rica. Limon: Cordillera de Talamanca,
Canton de Matina, entre Rfo Barbilla y Que-
brada Canabral. subiendo por la fila, 1()°01'N,
83°25'W, 100-300 m, 20 Oct. 1988 (H), Ger¬
ardo Herrera & Hr man Martinez M. 222H
(neotype, here designated, MO; isoneotype, CR
not seen).

Ampelocissus costaricensis Lundell, Publ. Carnegie Inst.
Wash. 478: 215. 1937. Syn. nov. TYPE: Costa Rica.
I.imon: Hacienda de Zent, without date (11). United
Fruit Company 383 (holotype. US).

Acknowledgments. We thank Dan Nicolson for
his advice and encouragement and Mike Gilbert for
searching for material at BM.

Literature Cited

Anon. 1869. New Publication [a rev iew and excerpts from:
Captain It. Pirn vK If. Seemann. 1869. Doltings on the
Roadside in Nicaragua. Panama and Mosquito]. J. Hot.
7: 271-279.

licit, T. 1888. The Naturalist in Nicaragua. A Narrative of
a residence at the gold mines ol Chontales; journeys in
the savannahs and forests. Edward Rumps, London.
Greuter, A.. F. R. Barrie, 11. M. Runlet. A. G. Chaloner.
V. Deinoulin, I). L. Hawksworth, P. M. Jorgensen. I). 11.
Nicolson. P. C. Silva. P. Trehane & J. McNeill (editors).
1994. International Code of Rotanical Nomenclature
(Tokyo Code). Regnum Veg. 131: i-389.

Lombardi. J. A. In press. Vitaceae-Generos Ampelocissus,
Ampelopsis e Cissus. El. Neotrop. Monogr. 80.

Lundell, C. L. 1937. Appendix. Studies of Mexican and
Central American plants — I. Publ. Carnegie Inst. Wash.
478: 208-221.

Planchon. J. E. 1887. Monographic des Ampelidees
vraies. In A. & C. de Candolle, Monographic Phaner-
ogamarum Prodromi, vol. 5: 305—654.

Seemann, R. 1869. Description of two new species of Vitis
from Central America. J. Rot. 7: 332—333.

A New Species of Psychotria Subgenus Psychotria from Nicaragua

Charlotte M. Taylor

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

ABSTRACT. The new species Psychotria hamilton¬
iana C. M. Taylor is described based on recent col¬
lections Irom the Caribbean coast of Nicaragua; this
species is related to P. neillii C. W. Hamilton &
Dwyer.

RESUMEN. Se describe la especie nueva Psycho¬
tria hamiltoniana C. M. Taylor basada en colec-
ciones recientes de la costa earibena de Nicaragua;
es atm a P. neillii C. W. Hamilton & Dwyer.

During study of recent collections from the Ca¬
ribbean coast of Nicaragua, the following new spe¬
cies was discovered. It belongs to Psychotria L.
subg. Psychotria, which was monographed for Cen¬
tral America by Hamilton (1989). The herbarium
of the Universidad Nacional Autonoma de Nicara-
gua-Leon, where specimens of this new species are
deposited, does not yet have an international ac¬
ronym; it is here referred to as “HULE.”

Psychotria hamiltoniana C. M. Taylor, sp. nov.
TYPE: Nicaragua. Rfo San Juan, municipio de
El Castillo, Reserva Indio-Mafz, Cerro El Dia¬
blo, n°01'N, 84°12'W, 250 m, 9 Jan. 1997,
R. Rueda, /. Coronado, O. Aurauz & F. Flores
5618 (holotype, Universidad Nacional Auton-
oma de Nicaragua-Ledn (HULE); isotype, MO-
4895236). Figure 1.

Haec species a Psychotria neillii C. W. Hamilton &
Dwyer foliorum majorurn petiolis longioribus atque inflo-
reseentiae axe breviore ar bracteis floral i 1 ms longioribus
distinguitur.

Shrubs and small trees to 3 m tall, flowering at
2.5 m tall; stems terete, densely short-pilosulous
often becoming glabrescent with age. Leaves
paired; blades oblanceolate to obovate or occasion¬
ally lanceolate, 14—31 cm long, 6—16 cm wide, at
apex acute to acuminate with tips to 20 mm long,
at base cordulate with sinus to 12 mm deep, pap¬
yraceous, adaxially glabrous, abaxially moderately
to densely pilosulous, often more densely so along
costa and secondary veins; secondary veins 8—18
pairs, ascending to spreading but less than 90°,
usually at least weakly looping to interconnect in
an undulating submarginal vein with 1—2(3) usually
rather weak intersecondary veins present between

pairs of secondary veins, adaxially the costa and
remaining venation plane, abaxially the costa prom-
inulous to prominent and the secondary veins and
reticulated minor venation prominulous; petioles 3—
9 cm long, densely short-pilosulous; stipules dense¬
ly short-pilosulous, caducous or sometimes persist¬
ing on distalmost 1-2 nodes, interpetiolar, trian¬
gular to lanceolate, 18-34 mm long, acute to very
shortly bifid, usually somewhat asymmetrical and
twisted, with a thickened basal triangular portion.
Inflorescences and flowers not seen. Infructesc-
ences terminal, erect, with peduncles 2.5—3 cm
long; panicles subglobose, 3—7 X 2—6 cm, with 2—
3 sets ol secondary axes, with the axes 4—6 per rank
and unequal, with the remains of old flowers and
developing fruits sessile in glomerules of 2-3 or
usually separated along cymose branches; bracts all
deltoid to ovate, 1.5— 2.5 mm long, acuminate; pe¬
duncle, axes, and bracts apparently green, densely
short-pilosulous to hirtellous or the bracts often gla¬
brescent; fruits ellipsoid, ca. 5 X 3.5 mm, not flat¬
tened laterally, pilosulous to hirtellous, orange to
red, with the persistent calyx limb ca. 1 mm long,
puberulous to glabrescent, deeply lobed, lobes ob¬
tuse to rounded; pyrenes with 3—5 low, rounded to
angled, longitudinal ridges.

Distribution , phenology, and habitat. In wet
forest at 120—250 m, southeastern Nicaragua. Col¬
lected in fruit in January, February, and December.

This species is distinguished by its pilosulous
pubescence on vegetative structures, red-black-
drying color, obovate to oblanceolate, petiolate
leaves that are cordulate at the base, shortly pe¬
dunculate inflorescences with relatively short, un¬
equal axes that are produced 4—6 per rank, bracts
1.5— 2.5 mm long, fruits orange to red at maturity,
and pyrenes dorsally with low longitudinal ridges
or angles. Although flowers have not been seen,
these characters clearly distinguish this species.
Psychotria hamiltoniana belongs to Psychotria
subg. Psychotria, and within this, to “Group 1” of
Hamilton (1989). Within “Group 1” it is similar to
P neillii C. W. Hamilton & Dwyer, which can be
distinguished by its leaves 8—18 X 3—7.5 cm, pet¬
ioles 2—8 mm long, stipules 6—14 mm long, pedun¬
cles 3—10 cm long, bracts subtending inflorescence

Novon 9: 425-427. 1999.

426

Novon

Figure I. Psychotria hamiltoniana C. M. Taylor. — A. Fruiting branch with infruetescenee, based on Ruedn el al.
5618. — B. Node witli stipule, based on Ruedn el al. 5822.

axes 3-4 mm long and floral bracts ca. 0.5 mm
long, flowers subsessile to pedicellate with pedicels
to 1 mm long, and calyx limb ca. 0.5 mm long.
Psychotria hamiltoniana is also similar to P. insta¬
ll is Standley, which can be distinguished by its stip¬
ules ca. 12 mm long, leaves with secondary veins
12—20 pairs and diverging 100—105° from the cos¬
ta, sessile inflorescences with the secondary axes

subequal at each node, and flowers all produced in
glomerules. The specific epithet honors Clement W.
Hamilton, who studied the Central American spe¬
cies of this group.

Paratypes. NICARAGUA. Rfo San Juan, municipio
de FI Castillo, Reserva Indio-Mafz, a lo largo del Carlo El
Pav<5n, a 3 km de su desembocadura en el Rfo Bartola,
ll°()l'N. 85°t6'W, Ruedn el al. 5140 (MO. HUEE); a 8

Volume 9, Number 3
1999

Taylor

Psychotria hamiltoniana from Nicaragua

427

km de la cabecera del Rfo Bartola, 11°01'N, 84°I4'W,
Rueda el al. 5266 (\1(), HULE); a 8 km de la cabecera
del Rfo Bartola, en direccion liacia el Cerro El Diablo,
I 1°01'N, 84°14'W, Rueda el al. 5322 (MO. HULE); a lo
largo del Cano Chontaleno, 1]°07'N, 84°12'W. Rueda el
al. 5810 (MO, HULE).

Acknowledgments. I thank the Flora de Nica¬
ragua Project for help with study of the Nicaraguan
materials and preparation of this manuscript; R. E.

Gereau for preparation of the Latin text; and Ri¬
cardo Rueda for access to specimens and providing
additional information.

Literature Cited

Hamilton, C. 1989. A revision of Mesoamerican Psycho-
tria subg. Psychotria (Rubiaeeae). Ann. Missouri Rot.
Card. 76: 67-111, 386-429, 886-916.

Una Especie Nueva de Posoqueria (Rubiaceae) de la
Guayana Colombiana

Charlotte M. Taylor

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Roeio Cortes- Ballon

Herbario Forestal UDBC, Universidad Distrital, A. A. 20305, Santafe de Bogota, Colombia.
Direceion actual: New York Botanical Garden, Bronx, New York 10458-5126, U.S.A.

Resum KN. Se describe e ilustra la especie nueva
Posoqueria tarairensis C. M. Taylor & R. Cortes-
Balhrn (Rubiaceae), que abunda en los bosques de
la Serranfa de Taraira en el sureste colombiano. La
especie posee hojas esclerofilas, una adaptacion tf-
pica a los habitats eon suelos arenosos pobres en
nutrientes, caracterfsticos de la region Guayana.

ABSTRACT. The new species Posoqueria tarairen¬
sis C. M. Taylor & R. Cort6s-Ballen (Rubiaceae),
common in the Serranfa de Taraira, southeastern
Colombia, is here described and illustrated. This
species shows selerophyllous adaptations to sandy
soils that are poor in nutrients, a habitat that is
typical in the Guayana region.

Durante la primera exploration de la Serranfa de
Taraira en el sureste de Colombia, lo eual es flo-
risticamente aim con la region biogeogrdfica de la
Guayana, se deseubrid una especie nueva de Po¬
soqueria que se describe a continuacidn. Steyer-
inark presentd las caracterfsticas de Posoqueria
(Steyermark, 1974) y una sinopsis de los miembros
sudamericanos de este genero neotropical (Steyer-
mark, 1967).

Posoqueria tarairensis C. M. Taylor & R. Cortes-
Ballen, sp. nov. TIPO: Colombia. Vaupes: Se¬
rranfa de Taraira, 10 km al NO del Raudal de
la Libertad, alrededores de la parcela #1,
bosque de catinga, 00°58'S, 69°45'W, 250 m,
1 Sep. 1993 (fl), J. Rodriguez IRR (holotipo,
COL; isotipos, MO-5060886, UDBC).

Haec species a similibus foliorum coriaceorum sat par-
vorum venatione secundaria et minore non manifesta, tubo
corollino 7.5—10.3 cm longo etiam fructu ca. 3 cm diant.
exocarpio 1.0— 2.5 mm crasso distinguitur.

Arbustos o arbolitos hasta de 2 m de altura, algo
suculentos, con drganos vegetativos glabros. Hojas
opuestas; lamina elfptico-oblonga, de 6-13.5 X 3-
5.5 cm, con apice y base obtusos hasta redondea-

Novon 9: 428-430. 1999.

dos, coriacea, haz lustroso, margenes ligeramente
convolutos, costa prominente en ambas caras, 3—4
pares de venas secundarias poco notorias o no evi-
dentes, las de orden menor muy poco o no eviden-
tes; peciolos de 8-18 mm de largo; estfpulas inter-
peciolares, deltoides, 3—5 mm de largo, enteras,
deciduas con las hojas. InHorescencias cimosas,
cortamente puberulas, pedunculos de 7-10 mm de
largo, porcion ramificada de la cima de 0.5— 1.5 cm
de largo (excluyendo f lores y frutos), bracteas del¬
toides hasta angostamente triangulares, 1-2 mm de
largo, agudas, enteras o cilioladas, pedicelos 3-4
mm de largo; flores ca. de 8 por infloresceneia; hi-
panto (ovario fnfero) turbinado, 2.5-3 mm de largo,
glabro o cortamente puberulo; limbo calicino 1.2—
1.5 mm de largo, glabro, lobulos hasta ca. 14 del
largo, deltoides a anchamente triangulares, obtusos,
margenes escariosas y cilioladas; corola hipocrate-
rimorfa, blanca, un poco carnosa, extemamente gla¬
bra o diminutamente puberula sobre los lobulos,
internamente glabra excepto en la garganta y en la
parte basal de los lobulos donde es moderadamente
vilosa; tubo de 7.5—10.3 cm de largo, 1.5—2 mm de
diametro, lobulos ligulados, ca. 12 mm de largo,
ca. 4 mm de audio, redondeados, con margen ci-
liolado, con prefloracion quincuncial; estambres in-
sertos justamente debajo de la garganta, filamentos
de 2-2.5 mm de largo, aplanados, ca. 0.8 mm de
ancho, anteras lanceoloides, ca. 4.5 mm de largo,
loculos dorsalmente puberulos a pilosulos, con un
apendice apical, glandular, deltoide, de ca. de 0.8
mm de largo y dos apendices basales (uno por 16-
culo) ligulados, glandulares, de ca. de 0.5 mm de
largo; estilo filiforme, aplanado, de ca. de 3.8 cm
de largo y 0.3 mm de ancho; estigma bifurcado,
viloso, de ca. de 1.5 mm de largo, lobulos de ca.
ile 0.2 mm de largo. Frutos globosos, de 3 cm de
diametro, lisos, verdes, glabros, pared de 1-2.5 mm
de gruesa; semillas anguladas, de 8-9 mm de largo,
lisas.

Volume 9, Number 3
1999

Taylor & Cortes-Ballen
Posoqueria tarairensis de Colombia

429

Rodriguez 818.

430

Novon

Distribucidn y fenologia. Conocida solo de la
localidad tfpica; colectada con Hores en septiembre
y con frutos en agosto.

Esta especie se distingue de otras especies de
Posoqueria por las hojas coriaceas y relativamente
pequenas con apices obtusos o redondeados y la
venacion muy poco notoria o no evidente, las co-
rolas con el tubo corto con relacion a las demAs
especies de Posoqueria y los trutos con pared re¬
lativamente delgada. Ademas esta especie habita
en bosques de dosel abierto caracterizados por es¬
pecies de porte bajo y diametros reducidos que cre-
cen sobre suelos arenosos y pobres en nutrientes,
denominados bosques de catinga. Esta es la uniea
especie de Posoqueria conocida en este habitat, y
presenta adaptaciones tfpicas a suelos pobres en
nutrientes, tales como las hojas de tamano reduci-
do, coridceas, cerdceas y con apices cortos (“escle-
rbfilas”) y el hdbito relativamente bajo.

El nombre se refiere a la localidad tfpica.

Pardtipo. COLOMBIA. Vaupes: Serranfa de Taraira,
10 km al NO del Baudal de la Libertad, alrededores de
la parcela # I. bosque de catinga, 00°58'S, 69°45'W, 250
m, 24 July 1993, R. Cortes A J. Rodriguez 543 (COL

esteril), 26 July 1993, R. Cortes A J. Rodriguez 601 (COL
esteril), 8 Aug. 1993, R. Cortes & ,/. Rodriguez RIR (COL.
MO).

Agradecimientos. A la Universidad Nacional de
Colombia y a la Royal Holloway College, University
of London, por organizar y financiar la expedieion
“Amazon 1993”; a J. Rodriguez por su valiosa par-
ticipacion en el trabajo de campo; al curador del
Herbaria COL por lacilitar el envfo de las muestras;
a R. E. Gereau por preparar la diagnosis en latfn.
La segunda autora expresa sus agradecimientos a
J. Aguirre, T. Kingston y K. Barlow por su invita-
cion y a la Universidad Distrital por permitir su
participacibn en el proyecto; y la primera a Carmen
Ulloa y Kay Yatskievych por su ayuda en la pre-
paracion del manuserito.

I .iteratura Cilada

Steyermark. J. A. 1%7. Posoqueria. En: B. M. Maguire &
Collal jorators. Flora of the Guayana Highlands — Bart
VII. Mem. New York But. Card. 17: 322-333.

- . 1974. Rubiaceae. En: T. Lasser (editor). Flora de

Venezuela 9: 1—2070. Instituto Botanico, Direccibn de
Recursos Naturales Renovables, Ministerio de AgricuL
tura y Crfa, Caracas, Venezuela.

Nuevos Taxones de Rubiaceae de la Costa Pacifica de
Colombia y Ecuador

Charlotte M. Taylor

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A.

Wilson Devia

Jardm Botanico “Juan Marfa Cespedes,” A. A. 314, Tulua, Valle, Colombia

Alvaro Cogollo

Jardm Botanico “Joaquin Antonio Uribe,” A. A. 51-407, Medellin, Antioquia, Colombia

Claes Persson

Botanical Institute, Systematic Botany, Goteborg University, Box 461, SE 40530, Goteborg,
Sweden. Present address: The New York Botanical Garden, Bronx, New York 10458, U.S.A.

RESUMEN. Las nuevas especies Coussarea sanc-
ti-ciprianii C. M. Taylor, Duroia laevis Devia,
Persson & C. M. Taylor, Duroia sancti-ciprianii
Devia, Persson & C. M. Taylor, Foramen ampin
C. M. Taylor, Foramen monsalveae C. M. Taylor,
Faramea calimana C. M. Taylor y Psychotria
monsalveae C. M. Taylor y la subespecie nueva
Coussarea loftonii subsp. calimana C. M. Taylor
se describen, basadas en material colectado du¬
rante los proyectos de inventarios florfsticos en
Bajo Calinia y el rfo Escalerete en la region de
la costa pacffica del departamento del Valle del
Cauca, Colombia; las distribuciones de algunas
de ell as se extienden hasta el departamento del
Choco en el norte y al noroeste de Ecuador en el
sur.

ABSTRACT. The new species Coussarea sancti-ci¬
prianii C. M. Taylor, Duroia laevis Devia, Persson
& C. M. Taylor, Duroia sancti-ciprianii Devia. Pers¬
son & C. M. Taylor, Faramea ampla C. M. Taylor,
Faramea monsalveae C. M. Taylor, Faramea cali¬
mana C. M. Taylor, and Psychotria monsalveae C.
M. Taylor, and the new subspecies Coussarea lof¬
tonii subsp. calimana C. M. Taylor are here de¬
scribed based on material collected for the floristic
inventory of the Bajo Calima and Rfo Escalerete
project regions in coastal Valle del Cauca Depart¬
ment, Colombia; several ol these species range
north to Choco Department and south into north¬
western Ecuador.

Durante inventario y estudio florfstico de dos sf-
tios en bosque pluvial cerca de Buenaventura, en

la costa pacffica de Colombia, los siguientes taxo¬
nes nuevos se descubrieron. Ambos sftios se en-
cuentran en la region fitogeografiea del Choco
(Gentry, 1986, Forero, 1982). La region de Bajo
Calima ha sido descrita por Faber- Langendoen y
Gentry (1991) y Ladrach y Wright (1995); el otro
sftio cercano es la Reserva Natural Rfo Escalerete,
que es generalmente similar. El rango de distribu-
cion de algunas de estas especies se extiende hasta
el noroeste de Ecuador, en la misma region fito-
geografica.

Coussarea loftonii (Dwyer & M. V. Hayden)
Dwyer subsp. calimana C. M. Taylor, subsp.
nov. TIPO: Colombia. Valle: municipio de
Buenaventura, Bajo Calima, Concesion
Pulpapel/Buenaventura, 3°55'N, 77°W, ea.
100 in, 6 mar. 1985, M. Monsalve B. 759
(holotipo, CUVC; isotipo. MO-3332908). Fi-
gura 1.

Haec subspecies a subspecie typica limbo calycino 4—
5 mm longo, fructu ca. 2.0 X 1.5 cm atque foliis abaxia-
liter foveolis glabris in axillis nervorum secundariorum
munitis distinguitur.

Arbustos o arbolitos hasta 12 m de alto, glabros.
Hojas con laminas elfpticas, 6—18 X 2—7 cm, con
apice acuminado con acumen 5—15 mm de largo,
con base aguda a obtusa, cartaceas, glabras en am-
bas superficies; venas secundarias 4—7 pares, con
domaeios glabros del tipo cripto; pecfolos 6-12 mm
de largo, glabros; estfpulas interpeciolares o lige-
ramente tubulares, caducas, glabras, porcion inter-
peciolar redondeada, 3—4 mm de largo, yemas ovoi-

Novon 9: 431^140. 1999.

432

Novon

9mm

I* igura 1. Coussarea lojlonii subsp. calitnana (.. M. laylor; de Motisulve Sfi4. — A. Kama con inflorescencia y fruto.
— B. Estfpula. — C. Baton floral. — I). Boton floral parcialmente disectado. — K. Antera. — F. Estigma. — (i. Fruto. B.
I- a la misma escala; F. F a la rnisma escala.

des a subglobosas. Infloresceneias terminales,
paniculadas, glabras, pedunculos ca. 1 cm de largo,
panfculas triangulares a redondeadas, ca. 3X3
cm, brcieteas ausentes, pedicelos 1—2 mm de largo
en cfmulas de 2—3; flores en boton con limbo ca-
licino t-5 mm de largo, glabro, denticulado; corola
en boton glabra, blanco-verdosa. Frutos lisos, ca. 2
X 1 .5 cm.

Distribucidn , habitat y fenologta. En bosque
pluvial a 100-130 m en la costa pacffica del Valle
del Cauea en Colombia; localmente infrecuente.
Colectada con flores en boton en septiembre, con
frutos en inarzo.

La especie Coussarea loftonii se encuentra desde
el oeste de Colombia hasta Panama central; las
plantas de Panama tienen el limbo calieino 2—3.5

Volume 9, Number 3
1999

Taylor et al.
Nuevos Rubiaceae

433

mm de largo con dientes lineares, los frutos ca. 13
X 10 mm y las hojas en el enves con domacios
abiertos y pilosulos. Esta nueva subespecie se dis¬
tingue por el limbo calicino mas largo con dientes
muy cortos, los frutos mas grandes y las hojas en
el enves con domacios del tipo cripto que son gla-
bros y casi cerrados. Las corolas maduras de las
plantas colombianas no se han visto, pero sus ca-
racteres vegetativos son casi iguales a las plantas
panamenas, y la similaridad en general entre las
plantas de Panama y las del oeste de Colombia
sugiere que las poblaciones constituyen dos grupos
no completamente aislados. Las plantas colombia¬
nas se describen al nivel de subespecie porque pa-
recen distintas en la distribucibn geografiea, con
poca variacion intermedia evidente en las muestras
disponibles.

El epiteto se refiere a la region de donde viene
el tipo.

Pardtipos. COLOMBIA. Valle: municipio de Buena¬
ventura, Bajo Calima, Concesion Pulpapel/Buenaventura,
3°55'N, 77°W. Monsalve 584 (CUVC. MO), 1696 (CUVC,
MO); Beserva Natural Rfo Escalerete, Devin et al. 4324
(MO. TUI.V, esteril), 5347 (MO. TUI.V. esteril).

Coussarea sancti-ciprianii C. M. Taylor, sp. nov.
TIPO: Colombia. Valle: municipio de Buena¬
ventura, corregimiento Cordoba, vereda San
Cipriano, Reserva Natural Rfo Escalerete, 100
m, 13 abr. 1993, W. Devin. D. Cardenas & A.
Cruz 3966 (holotipo, TULV; isotipos, MO-
4625143, MO-4625148). Figura 2A.

Haec species a congeneris stipulis deltatis 5—13 nun
longis, foliis efoveolatis, inllorescentiis subcylindrieis,
limbo calycino truncato ea. 2 mm longo atque corollae
tubo lubulis lungiore distinguitur.

Arbustos o arbolitos hasta 7 m de alto, glabros.
Hojas con laminas elfpticas a ovadas, 12—30.5 X
4.8—17 cm, con apice agudo a acuminado con acu¬
men hasta 8 mm de largo, con base cuneada a ob-
tusa, papiraceas a eartaceas, glabras; venas secun-
darias 6—10 pares, sin domacios, en la haz la costa
promfnula, las venas secundarias planas a promf-
nulas y la venacion menor plana, en el enves la
costa proininente y las venas secundarias y la ve¬
nacion menor promfnulas; pecfolos 12-40 mm de
largo, glabros; estfpulas interpeciolares o ligera-
mente tubulares, caducas, glabras, porcion inter-
peciolar deltoide, 5—13 mm de largo, yernas lan-
ceoloides, usualmente angostas. Inflorescencias
terminales, paniculadas, glabras o puberulentas,
peduneulos 3—7.5 cm de largo, articulados en la
mitad basal, en la artieulaeion con una estfpula sin
hojas, panfculas subcilfndricas, 5—7 X 2—6 cm,
ejes secundarios no o poco desarrollados, bracteas

hasta 0.5 mm de largo, pedicelos 2-3 mm de largo;
flores producidas en cfmulas de 2—3; limbo calicino
ca. 2 mm de largo, glabro, truncado, entero o di-
minutamente ciliolado; corola en boton blanca, ex-
temamente puberulenta, tubo ca. 14 mm de largo,
lobos ca. 7 mm de largo. Frutos lisos, ca. 18 X 12
mm.

Distribucion, habitat y fenologia. En bosque
pluvial a 30—140 m, en la costa Pacifica del Valle
del Cauca en Colombia. Colectada con flores en
abril, con frutos en abril, julio y agosto.

Esta especie se distingue por las estfpulas del-
toides 5—13 mm de largo, las hojas sin domacios,
las inflorescencias subcilfndricas, el limbo calicino
truncado ca. 2 mm de largo y las corolas con el
tubo mas largo que los lobos. El epiteto se refiere
a la localidad de donde viene el tipo.

Pardtipos. COLOMBIA. Valle; municipio de Buena¬
ventura. along road from Buenaventura to Malaga, Km.
52.4 (from beginning of road at Gallinero), 4°03'N,
77°05'W, Croat & Hay 75712 (CUVC, MO); corregimiento
Veneral, Rfo Yurumanguf, Quebrada Querre. Devin el al.
2974 (MO, TULV); corregimiento Cordoba, vereda San Ci¬
priano. Reserva Natural Rfo Escalerete, 3°50'N, 76°35'W,
Devin et al. 4041 (MO, TULV), 4194 (MO. TULV), 4438
(MO. TULV), 5265 (MO, TULV); Bajo Calima, Concesion
Pulpapel/Buenaventura, 3°55'N, 77°W. Monsalve 172
(CUVC. NY), 248 (CUVC, MO), 1516 (CUVC. MO).

Duroia laevis Devia, Persson & C. M. Taylor, sp.
nov. TIPO: Colombia. Valle: municipio de Res-
trepo, Campamento Rfo Azul, El Chanco, Pro-
yecto Calima III de la CVC, 04°00'N,
76°00'W, 550 m, 18 mar. 1986, W. Devia A.
1266 (holotipo, TULV; isotipos, GB, MO-
4671198, MO-4671199). Figura 3A, B.

Haec species a Duroia soejartoi I). R. Simpson pubes-
centia strigulosa atque foliis papyraceis basi cordulatis
distinguitur.

Arboles hasta 25 m de alto; tallos cuadrangula-
res, estrigulosos a glabrescentes. Hojas opuestas;
laminas elfpticas a ampliamente elfpticas, 16-40(-
62) X 12-28(— 37) cm, papiraceas, con apice ob-
tuso a redondeado y brevemente acuminado con
acumen 3—8 mm de largo, con base cordulada, en
la haz glabras excepto puberulas a estrigulosas en
la costa y las venas secundarias, en el enves pu¬
berulas en toda la superficie; venas secundarias
15—20 pares, en la haz la costa y las venas secun¬
darias planas a subsulcadas y la venacion menor
reticulada y promfnula, en el enves la costa y las
venas secundarias prominentes y la venacion me¬
nor promfnula; pecfolos serfceos, 1.4-5 cm de lar¬
go; estfpulas densamente velutino-serfceas, cadu¬
cas en una caliptra entera, hasta 28 cm de largo.

434

Novon

Figure 2. — A. Coussarea saruti-ciprianii C. M. Taylor, ramita con inflorescencia; tie Devin 3966. - — B. Faramea

calimana C. M. Taylor, ramita con inflorescencia; tie Devin 3061. C-K. Psychotria monsalveae C. M. Taylor; tie Croat
57527. — C. Ramita con inflorescencia. — I). Hoja con apice de la ramita con estfpula. — E. Ctmula con flor. A— D a
la eseala de I cm; E a la escala <le 5 mm.

aguda, algo aplanada desde los lados intrapeciola-
res. Inflorescencias y flores enteras no observadas;
corola estaminada (encontrada on el suelo), blanca,
hipocraterilorme con el tubo gradualmente ensan-
chado desde la base, externamente serfeea, inter-
namente glabra, 18-25 mm de largo, 3-4 mm de
diam. cerca de la base, 6—8 mm de diam. cerca de
la garganta, lobos 6, angostamente triangulares.

25—28 mm tie largo, 5—7 mm de ancho en la base,
agudos a acuminados; anteras sesiles, ca. 13 X 1
mm, incluidas; eslilo ca. 22 mm de largo, estigma
bilobulado, ca. 9 mm de largo; ovario aparente-
mente unilocular, placentacion aparentemente pa¬
rietal. Frutos terminales, solitarios (—3), elipsoides,
basta 8.5 X 5.5 cm, lisos, estrigulosos a glabres-
centes, verde-parduzcos.

Volume 9, Number 3
1999

Taylor et al.
Nuevos Rubiaceae

435

Figure 3. A. B. Duroia laevis Devia, Persson & C. M. Taylor. — A. Ramita con apice y estfpula; de Aulestia et al.
375. — B. Flor estaminada; de Forero et al. 3273. C, D, Duroia sancti-ciprianii Devia, Persson & C. M. Taylor. — C.
Ramita esteril; tie Devia 5050. — D. Fruto; de Devia 5268. A, C y D a la misma escala.

436

Novon

Distribution , habitat y fenologia. En bosque
pluvial a 130-600 m, en la costa pacffica del Choco
en Colombia al noroeste de Ecuador, en la provin-
cia de Esmeraldas; localmente ocasional. Colectada
con Hores es febrero, con Irutos en marzo y sep-
tiembre.

Esta especie se distingue por sus hojas opuestas
y corduladas en la base, la pubescencia adpresa y
tfpicamente estrigulosa, y los frutos I isos y estri-
gulosos a glabrescentes. El epiteto se refiere a los
frutos lisos y la ausencia general de pubescencia
notable de esta especie. La especie mas similar es
Duroia soejartoi D. R. Simpson, descrita de esta
misma region general (Simpson, 1982); D. soejartoi
se distingue por su pubescencia generalmente hir-
suta y sus hojas coriaceas a subcoriaceas con la
base aguda y decurrente.

Pardtipos. COLOMBIA. Choco: carretera Medellin—
Quilxld, adelante de la Ciudad Bolivar. Km 171, rfo La
1‘laya, Forero et al. 1079 (COL, MO); municipio de Novita,
ladera nolle del Cerro Torra. filo al 0 del rfo Surama,
camino al Alto del Oso, 4°55'N, 76°25’W, Forero et al.
3273 (COL); Quibd6-Medellfn road, 33 km L ofTutunen-
do, 5°43'N, 76°22'W, Gentry A- Renteria 24172 (COL).
Vail e: municipio Buenaventura, corregimiento Cordoba,
vereda San Cipriano, Reserva Natural Rfo Escalerete, en
la parcela permanente, 3°4<)'N. 76°52'W. 130 m, Devia et
al. 4748 (JAUM, MO, TULV), 4754 (J ADM. MO, TULV),
4834 (JAUM. MO. TULV), 4848 (JAUM, MO. TULV),
4872 (JAUM, MO, TULV), 4978 (JAUM. MO, TULV),
5145 (JAUM. MO. TULV), 5940 (JAUM. MO, TULV) [to-
dos es!eriles|. ECUADOR. Esmeraldas: canton San Lo¬
renzo. parroquia Mataje, Reserva Etniea Awti, 1°08'N.
78°33'W, 200 m, Aulestia et al. 375 (GB, MO, QCNE);
canton Eloy Alfaro, parroquia Luis Vargas Torres, Reserva
Eeolbgiea Cotacachi-Cayapas, rfo Santiago, Estero Angos¬
tura. (X)°49'S, 78°45' W, Tirado et al. 717 (GB, MO,
QCNE).

Duroia saneti-ciprianii Devia, Fersson & C. M.
Taylor, sp. nov. TIPO; Colombia. Valle: muni¬
cipio de Buenaventura, Bajo Calima, conce-
sidn Pulpapel, S.A. file Carton de Colombia],
frente Hans, km 22 via Malaga, 20 m, 2 mar.
1990, W. Devia A. 3063 (holotipo, TULV; iso-
tipo, MO-4671202). Figura 3C, D.

Haec species a Duroia soejartoi I). R. Simpson pubes-
centia patente pilosa ad hirsuta. petiolis brevioribus, fo-
liorum venatione siqira plerumque sulcata, stipulis usque
ad 14.3 cm longis atque inflorescentiis congestis distin-
guilur.

Arboles hasta 15 m de alto; tallos cuadrangula-
res, patente-pilosos a glabrescentes. Hojas opues¬
tas; laminas elfpticas a ovadas u obovadas, 19-40
X 10-29.5 cm, eartAceas, con Apice obtuso a su-
bagudo y brevemente acuminado con acumen 3—10
mm de largo, con base cuneada a redondeada, en
la haz pilosas o hirsutas esparcidamente en la lam¬

ina y densamente en la costa y las venas secun-
darias, en el enves moderadamente pilosas o hir¬
sutas en toda la superficie; venas secundarias 9-
19 pares, en la haz la costa, las venas secundarias
y la venaeidn menor planas a frecuentemente sul-
cadas, en el enves la costa y las venas secundarias
prominentes y la venacion menor promfnula; pecfo-
los pilosos a hirsutos, 6-20 mm de largo; estfpulas
densamente pilosas a hirsutas, caducas, alargan-
dose en un cilfndro que se rompe por un lado y
persistiendo un rato como una estructura espatacea,
hasta 14.5 cm de largo, el conjunto agudo. Inflo-
rescencias estaminadas congestionadas, con ca. 20
Hores; Hores estaminadas con caliz externainente
densamente serfceo a hirsuto, intemamente espar¬
cidamente serfceo, tubo 10-14 mm de largo, trun-
cado, con lobos 6, Hliformes, ca. 2 mm de largo;
corola en botdn blanca, hipocrateriforme a ligera-
inente infundibuliforme, externamente densamente
retro rso-sertcea, tubo hasta 15 mm de largo, lobos
6, ca. 12 mm de largo, externamente densamente
antrorso-serfceos; anteras sesiles, incluidas, ca. 9
X 0.5 mm; estilo ca. 16 mm de largo, estigina bi-
lobulado, ca. 8 mm de largo; ovario aparentemente
unilocular, placentacibn aparentemente parietal.
Frutos terminales, solitarios (—3), elipsoides a ovoi-
des, 7.5— 8.5 X 4—5.5 cm, lisos, pilosos a hirsutos,
verde-parduzcos.

Distribution , habitat y fenologta. En bosque
pluvial a 20—130 m, en la costa pacffica del Choc6
al Valle del Cauca en Colombia; localmente fre-
cuente con semilleros y plantas jovenes comunes
en sftios con luminosidad relativamente alta del so-
tobosque. Colectada con Hores en botdn en abril,
con frutos en marzo, septiembre y octubre.

Esta especie se distingue por las hojas opuestas,
obtusas a redondeadas en la base y con la venacion
frecuentemente sulcada en la haz, la pubescencia
patente y pilosa a hirsuta, la inflorescencia conges-
tionada y los frutos lisos y pilosos a hirsutos. El
epiteto se refiere a la comunidad de San Cipriano,
protectora de la reserva natural “Rfo Escalerete” y
colaboradora en su estudio botanico. La especie
mas similar es Duroia soejartoi, descrita de esta
misma region general y a la cual varias inuestras
de D. sancti-ciprianii lueron asignadas por Simpson
(1982). No obstante, D. soejartoi se distingue por
su pubescencia adpresa, sus hojas agudas en la
base con la venacion plana en la haz, sus pecfolos
1.5—2 cm de largo y sus inflorescencias estamina¬
das pedunculadas.

Pardtipos. COLOMBIA. Choco: rfo Serrano, afluente
del rfo Atrato, 4—6 km arriba de Guyabal, Forero et al.
1305 (COL). Valle: Bajo Calima. Cabrera 575 (COL, E);

Volume 9, Number 3
1999

Taylor et al.
Nuevos Rubiaceae

437

rfo Cajambre, Barco, 3°28’N, 76°57'W, Cuatrecasas 17144
(F); rfo Calima, I ,a Trojita, Cuatrecasas 16588 (F; esteril);
municipio Buenaventura, correginiiento C6rrloba, vereda
San Cipriano, Beserva Natural Bfo Escalerete, Devia et
al. 8948 (GB, MO); Beserva Natural Rfo Escalerete, en
la parcela permanente, 03°49'N, 76°52'W, 130 m, Devia
et al. 5088 (JAUM. MO, TULV; esteril), Devia et al. 5050
(JAUM, MO, TULV; esteril), Devia et al. 5222 (JAUM,
MO. TULV; esteril), Devia et al. 5287 (J AUM. MO. TULV;
esteril), Devia et al. 5268 (JAUM. MO. TULV). Devia et
al. 5884 (JAUM, MO, TULV; esteril), Devia et al. 5951
(JAUM. MO, TULV).

Fararnea ampla C. M. Taylor, sp. nov. TIPO: Co¬
lombia. Valle: municipio Buenaventura, corre-
gimiento Cordoba, vereda San Cipriano, Re-
serva Natural Rfo Escalerete, en la parcela
permanente, cerca de la Parcela “E,” 3°49'N,
76°52'W, 130 m, 22 sep. 1997, W. Devia A.,
C. M. Taylor & J. Adarve 5601 (holotipo,
TULV; isotipos, COL, JAUM, MO-4942001).
Figura 4A, B.

Haec species a Fararnea calophylla Standley fructu mi-
nore lateraliter eompresso atque tubo coroll i no gracili dis-
tinguitur.

Arbustos o arbolitos hasta 6 m de alto, glabros.
Hojas con laminas elfpticas, 17.5-32 X 6.5-17 cm,
con apice acuminado a obtuso o redondeado y
abruptamente cuspidado con acumen 10-15 mm de
largo, con base cuneada a obtusa, cartaceas, con
enves ligeramente palido; venas secundarias 11—20
pares, con la vena submarginal sinuosa, desarrolla-
da, en arnbas superficies la costa prominente y la
venacion menor reticulada y promfnula, en la haz
las venas secundarias planas a ligeramente promf-
nulas, en el enves promfnulas; pecfolos 11—28 mm
de largo; estfpulas interpeciolares, caducas. porcion
interpeciolar deltoide, 8—15 mm de largo, breve-
mente aristada. Inflorescencias terminales, pani-
culadas, glabras, pedunculos 5.5—11 cm de largo,
panfculas redondeadas, 10—12 X 12-18 cm, brac-
teas reducidas, pedicelos 2—5 mm de largo en cf-
inulas umbeliforme de 3—7; flores con limbo cali-
cino 0.3-0.5 mm de largo, denticulado; corola en
boton hipocrateriforme, azul, glabra, tubo ca. 6 mm
tie largo, lobos ca. 4 mm de largo. Frutos oblatos,
algo aplanados lateralmente, ca. 13 X 16 mm, ne-
gro-azules, jugosos.

Distribution, habitat y fenologi'a. En bosque
pluvial a 50—180 m, en la costa pacffica del Choco
y el Valle del Cauca en Colombia. Colectada con
llores en septiembre y diciembre, con frutos en fe-
brero, septiembre y diciembre.

Esta nueva especie es similar a Fararnea ca¬
lophylla Standley, que se distingue por los frutos
generalmente mas largos, no o casi no aplanados

lateralmente y aparentemente coriaeeos en lugar de
jugosos y los tubos corolinos mas gruesos. El epi-
teto se refiere a las hojas por lo general relativa-
mente grandes. Los frutos tienen un sabor insfpido
excepto un poco dulce (Taylor, obs. pers.).

Paratipos. COLOMBIA. Choco: hills near highest
point of Bagado— Certegui trail, 5°25'N, 76°28'W. Juncosa
1545 (JAUM. MO); confluence of rfo Valle and rfo Mutata.
Warner 169 (MO). Valle: municipio Buenaventura, along
the road between Buenaventura— Malaga at Km 51.3,
4°09'N, 77°11'W. Croat & Watt 70847 (MO); municipio
Buenaventura, correginiiento Cordoba, vereda San Cipria¬
no, Beserva Natural Bfo Escalerete. en la parcela per¬
manente, 3°49'N, 76°52'W, Devia et al. 5598 (JAUM, MO.
TULV); municipio Buenaventura, Bajo Calima region, N
of Buenaventura. Carton de Colombia concession, re¬
served area on lower Rfo San Juan below mouth of Bfo
Calima, 4°10'N, 77°25'W, Gentry 85473 (MO).

Farainea calimana C. M. Taylor, sp. nov. TIPO:
Colombia. Valle: municipio Buenaventura,
Bajo Calima, ca. 15 km N of Buenaventura,
Carton de Colombia Concession, Dindo area,
03°59'N, 77°02'W, 50 m, 28 mar. 1986, A.
Gentry, E. Zardini, M. Monsalve B. & J. D.
Caicedo 53727 (holotipo, CUVC; isotipo, MO-
3393526). Figura 2B.

Haec species a Fararnea calophylla Standley corollae
tubo lobulis breviore atque foliorum venatione tertiaria
quaternariaque vix manifesta distinguitur.

Arbustos o arbolitos hasta 8 m de alto, glabros.
Hojas con laminas elfpticas a elfptico-oblongas,
10—23.5 X 3.5— 9.5 cm, con apice obtuso a redon¬
deado y abruptamente cuspidado con acumen 7—
15 mm de largo, con base aguda a cuneada o algo
obtusa, cartaceas, con enves ligeramente palido;
venas secundarias 9—14 pares, con la vena sub¬
marginal sinuosa, desarrollada, en ambas superfi¬
cies la costa promfnula, las venas secundarias pla¬
nas o ligeramente promfnulas y la venacion menor
no evidente o las venas intersecundarias evidentes,
delgadas y con unas venulas transversales; pecfolos
3—20 mm de largo; estfpulas tubulares, caducas,
porcion interpeciolar 10-13 mm de largo, con una
parte fibre triangular, aristada con arista 4—6 mm
de largo. Inflorescencias terminales, paniculadas,
glabras, pedunculos 3-7 cm de largo, panfculas tri¬
angulares a redondeadas, 5—10 X 6—14 cm, brac-
teas reducidas, pedicelos 5—17 mm de largo en cf-
mulas de 2—3; flores con limbo calicino 0.8—1 mm
de largo, truncado a denticulado; corola hipocra¬
teriforme, blanca a azul, glabra, tubo 8—9 mm de
largo, lobos 15—20 mm de largo. Frutos no obser-
vados.

Distribution, habitat y fenologi'a. En bosque

438

Novon

Figure 4. A. B, Faramea ampla C. M. Taylor; de Warner 169. — A. Ramita con inflorescencia. — B. Flor en bot6n.
C, 1), Faramea monsalveae C. M. Taylor; de Gentry et al. 59594. — C. Ramita con inflorescencia. — 0. Flor en hot on.
A, C a la misma escala.

pluvial a 5-150 m, en la costa pacffica del Choco
y el Valle del Cauca en Colombia. Colectada con
flores en febrero, marzo y abril.

Esta especie es similar a Faramea calophylla,
especie simpatrica que tiene las corolas con el tubo

igual o mas largo que los lobos y las hojas con las
venas secundarias, intersecundarias y submargi-
nales prominentes y la venacion menor evidente,
reticulada y prommula en ambas superficies. El
epiteto se refiere a la region de donde viene el lipo.

Volume 9, Number 3
1999

Taylor et al.
Nuevos Rubiaceae

439

Paratipos. COLOMBIA. Choco: along road between
Medellin and Quibdd, 5.5 kin E of Tutunendo, 23.5 km
E of Quibdd, 5°44'N. 76°29'W. Croat 56217 (J A U M , MO);
lioya del rfo San Juan, quebrada Taparal. afluente del no
San Juan, alrededores de la comunidad indfgena Waunana
de Taparalito, 4°12'N. 77°10'W, Forero et al. 4276 (COL.
MO); area of Baudd, on right side of Rfo Baudo, ca. 13
km upstream the estuary, near and on right back of Que-
brada Carpio, Fuchs & Zanella 21939 (MO). Valle: mu-
nieipio de Buenaventura, Bajo Calima region, along road
between Buenaventura and Malaga, on Carretera Hanz
(lumber extraction road N of main Buenaventura— Mdlaga
road), 6.5 km N of main road. 4°03'N, 77°03'W, Croat
71126 (MO); municipio de Buenaventura, corregimiento
San Isidro, reserva INDERENA-CONIF, Devin & Prado
2606 (MO, TULV); municipio de Buenaventura, Carton de
Colombia Concession, km 22 via Malaga, frente Hans, De¬
vin 3061 (MO. TULV); municipio de Buenaventura, Bajo
Calima, ca. 10 km due N of Buenaventura. Carton de
Colombia concession, 3°56'N, 77°08'W, Gentry 35429
(MO; esteril), 35449 (MO; esteril). Gentry et al. 40319
(MO; esteril); municipio de Buenaventura, Bajo Calima,
ca. 15 km N of Buenaventura, Carton de Colombia con¬
cession, Dindo area, 3°59'N, 77°02'W, Gentry et al. 53635
(MO); municipio de Buenaventura, Bajo Calima, Juancha-
co Palmeras area ca. 10 km NW of Buenaventura, 3°56'N,
77°08'W. Gentry et al. 56877 (MO), 56932 (MO); muni¬
cipio de Buenaventura, Concesion Pulpapel, bosque plu¬
vial tropical, 3°55'N, 77°W, Monsalve 667 (CUVC. MO);
municipio de Buenaventura, Bajo Calima, Concesion Pul¬
papel, van der Werj] & Monsalve 9691 (MO).

Faramea monsalveae C. M. Taylor, sp. nov.
TIPO: Colombia. Valle: municipio Buenaven¬
tura, Concesion Pulpapel, bosque pluvial trop¬
ical, 3°55'N, 77°W, 100 m, 30 ago. 1984, M.
Monsalve B. 274 (holotipo, CUVC; isdtipos,
JAUM, MO-3395866, NY). Figura 4C, D.

llaec species a Faramea torquata Mueller Argoviensis
pedicellis longioribus. floribus paucioribus In cymulas
umbelliformes quinquefloras dispositis atque corollae tubo
lobulis longiore distinguitur.

Arbustos hasta 10 m de alto, glabros. Hojas con
laminas elfpticas o elfptico-oblongas, 10.5—20 X
4. 5-8. 5 cm, con apice acuminado u obtuso a re-
dondeado y cuspidado con acumen 1—2 cm de lar¬
go, con base obtusa a cuneada, cartaceas; venas
secundarias 8—10 pares, no o ligeramente broqui-
dodromas, en ambas superficies la costa y las venas
secundarias promfnulas y la venacion menor reti-
culada y plana o con frecuencia ligeramente pro-
mfnula; pecfolos 8—12 mm de largo; estfpulas per-
sistentes, triangulares a redondeadas, 2—5 mm de
largo, costadas y aristadas, arista 3—5 mm de largo.
Inflorescencias axilares y a veces terminales, 6—20-
floras, pedunculos l por axila o 3( — 5) en el apice
del tallo, 3-5 cm de largo, articulados cerca de la
base con bracteas estipulares reducidas, pedicelos
4—9 mm de largo en ctmulas umbeliformes 3—6-
floras; flores en boton con limbo calicino 2-3 mm

de largo, truneado; corola infundibuliforme, apa-
rentemente blanca, glabra, con tubo ca. 6 mm de
largo, lobos ca. 12 mm de largo, lanceolados, aeu-
minados. Frutos subglobosos a ligeramente oblatos,
ca. 12 X 15 mm, morados, con pedicelos hasta 19
mm de largo.

Distribution, habitat y fenologia. En bosque
pluvial a 50-100 m, en la costa pacffica del Choco
y el Valle del Cauca en Colombia. Colectada con
flores en boton en diciembre, con frutos en enero,
febrero, abril y agosto.

Esta especie es similar a Faramea torquata Mue¬
ller Argoviensis, especie amazonica que tiene los
pedicelos mas cortos, las flores mas numerosas y la
corola con el tubo mas largo que los lobos; y a
Faramea occidentals A. Richard, que tiene las ho¬
jas de textura mas fina, las inflorescencias usual-
mente terminales con las flores mas numerosas en
cfmulas 3-floras y los pedunculos no articulados.
El epiteto es en honor a Miryam Monsalve, botaniea
colombiana quien ha colectado y estudiado exten-
sivamente la flora de la region de Bajo Calima.

Paratipos. COLOMBIA. Choco: trail to Tubadd,
Quibdd— Tutuendo road 14 km NE of Quilxlo, Gentry &
Renteria 24282 (HUA, MO; esteril); Quibdd— Tutunendo
Road ca. 3 krn W of Tutunendo, 5°46'N, 76°35'W, Gentry
et al. 30123 (JAUM, MO; esteril). Valle: municipio de
Buenaventura, Bajo Calima, Cabrera 504 (COL); munici¬
pio de Buenaventura, Bajo Calima Concession, ca. 20 km
N <>( Buenaventura, ca. 300 m NW of “Dindo III,” “Dindo
IV" int ersection on “Dindo IV." 3°59'N. 77°04'W, Faber-
Langendoen A Hurtado 1485 (MO; esteril); municipio de
Buenaventura, Bajo Calima Concession, ca. 25 km NW of
Buenaventura, ca. 9 km NW of San Isidro at intersection
on “Canalete,” near gate, 3°58'N, 77°08'W, Fa her- hi n-
gendoen &: Hurtado 1658 (MO; esteril), 1758 (MO; este¬
ril); municipio de Buenaventura, Bajo Calima, ca. 15 km
N of Buenaventura, Carton de Colombia Concession,
3°56'N, 77°08'W Gentry et al. 40273 (MO); municipio de
Buenaventura, Bajo Calima, road to Juanchaco Palmeras,
3°55'N, 77°02'W, Gentry et al. 47924 (MO); municipio de
Buenaventura, Bajo Calima, Canalete area. Km 40 N of
Buenaventura. W of San Isidro, 3°58'N, 77°00'W, Gentry
et al. 59594 (MO); municipio de Buenaventura, Concesion
Pulpapel, 3°55'N, 77°W, Monsalve 2064 (CUVC, MO).

Psyehotria inonsalveae C. M. Taylor, sp. nov.
TIPO: Colombia. Valle: municipio de Buena¬
ventura, along road from Buenaventura to Bajo
Calima, ca. 4 km from Rfo Calima, near Km
14 marker (14 km from Cali— Buenaventura
Hwy.), 3°56'N, 76°59'W, < 50 m, 21 Mar.
1984, T. B. Croat 57527 (holotipo, MO-
3194188; isotipo, CUVC). Figura 2C— E.

Haec species a congeneris stipulis lanceolatis interpe-
tiolaribus caducis, foliorum nervo submarginali arcuato
absentive, inflorescentiae axibus secundariis geminatis,
floribus sessilibus vel brevipedicellatis in cymulas 3(— 5)-

440

Novon

floras dispositis, limbo calycino 0.5—1 mm longo atque
corollae fauce barbata el tubo ea. 2.5 mm longo distin-
gnilur.

Arbustos hasta 1.3 m de alto, cuando secos con
color pardo-rojizo o -castano; ramitas glabras. Ho-
jas con laminas elfpticas a elfptico-oblongas, 8-21
X 2.5—8 cm, con apice acuminado con acumen
1.2-2 cm de largo, con base aguda a cuneada, pa-
pir&ceas, glabras o a veces puberulas en el enves;
venas secundarias 8-14 pares, con vena submar¬
ginal ausente o debil y conectandose en lazos, en
la haz la costa promfnula, las venas secundarias
planas o a veces algo suleadas y la venacion menor
plana, en el env£s la costa y las venas secundarias
prominentes y la venacion menor promfnula; pecfo-
los 8-20 mm de largo, glabros; estfpulas caducas,
lanceoladas, 6—15 mm de largo, agudas, enteras o
cilioladas, con la costa engrosada y usualmente for-
mando un tridngulo engrosado en la base. Inflores-
cencias terminales, paniculadas, puberulas, pedun-
culos ca. 2.5 cm de largo, panfeulas corimbiformes
a anchamente piramidales, ca. 4.5 X 6 cm, ejes
secundarios 2 por nudo, bracteas lineares, hasta 1
mm de largo, pedicelos hasta 0.5 mm de largo; flo-
res s^siles a brevipediceladas en cfmulas de 3( — 5);
hipanto ca. 1 mm de largo, puberulento; limbo ca-
licino 0.5-1 mm de largo, dentado; corola
infundibulilorme, blanca, externamente glabra, bar-
bada en la garganta, tubo ca. 2.5 mm de largo, 1.2—
1.5 mm de didm. cerca del medio, lobos angosta-
mente triangulares, ca. 2 mm de largo, agudos;
anteras ca. 1 mm de largo, exsertas sobre filamentos
puberulos; estigmas ca. 1.2 mm de largo, exsertos.
Fruto no observado.

Distribucidn, habitat y fe.nologla. En bosque
pluvial a ca. 50 m, conocida solo de la region del
tipo en la costa pacffica del Valle del Cauca en
Colombia. Colectada con flores en marzo.

Esta espeeie pertenece al subgenero Psychotria ;
se distingue de la otras especies de este grupo por
las estfpulas distintivas, lanceoladas, interpeeiola-
res y caducas; las hojas con la vena submarginal
ausente o debil y conectandose en lazos; las inflo-
rescencias con los ejes secundarios en pares; las

flores s6siles o brevipediceladas en cfmulas de 3(-
5); el limbo calicino 0.5-1 mm de largo; y la corola
barbada en la garganta y con el tubo ca. 2.5 mm
de largo. Las flores tienen las anteras y los estigmas
ambos exsertos y producidos al mismo nivel, en-
tonces parecen ser homostilas, estado raro pero ya
conocido de Psychotria subg. Psychotria en Panama
(Hamilton, 1989). Ademcis las flores del tipo tienen
las anteras ya abiertas pero los estigmas apretados
y aparentemente todavfa jovenes, un arreglo que
sugiere que las flores son protandras. El epiteto
tambien es en honor a Miryam Monsalve.

Pardtipos. COLOMBIA. Valle: municipio de Buena¬
ventura, Bajo Calirna, ea. 15 km IN of Buenaventura, Car¬
ton de Colombia Concession, 3°56'N, 77°08'W, Gentry et
nt. 40327 (CIJVC, JAUM. MO); municipio de Buenaven¬
tura, Bajo Calirna, ca. 15 km NW of Buenaventura, 18-
year-old regeneration (Standi 2, Plot 3), behind Pulpapel
camp, 3°59’N, 77°05'W, Gentry et <il. 56830 (CL VC. MO;
esteril), Gentry et al. 56834 (CUVC, MO; esteril).

Agradecimientos. Recibimos apoyo para el tra-
bajo en el campo de la National Geographic Society
(#5621-96); Smurflt Carton de Colombia, Cali, Co¬
lombia; el INCIVA (Instituto Vallecaucano de Inves-
tigaciones Cientfflcas), Cali, Colombia; y el Proyecto
Biopacffico. Nuestros sinceros agradecimientos a R.
E. Gereau por preparar el texto latino y a Miryam
Monsalve por preparar la Figura 1.

Literature Citada

Faber- Langendoen, I). & A. II. Gentry. 1991. The struc-
lure and diversity of rain forests at Bajo Calirna, Choc6
region, western Colombia. Biotropica 23: 2—11.

Forero, E. 1982. La flora y vegetacioii del Chord y sus
relaciones fitogeogrdficas. Revista Inst. Geogr. “Augus¬
tin Codozzi” 10: 77-90.

Gentry, A. II. 1986. Species richness and floristic com¬
position of Chocd region plant communities. Caldasia
15: 71-91.

Hamilton, C. W. 1989. A revision of Mesoameriean
Psychotria subg. Psychotria (Rubiaceae). Part I; Intro¬
duction and species 1 — 16. Ann. Missouri Bot. Gard. 76:
67-1 11.

Cidrach, VL E. & J. A. Wright. 1995. Natural regeneration
in a secondary Colombian rain (orest: Its implications
(or natural forest management in the tropics. J. Sustain¬
able Forest. 3(1): 15—38.

Simpson, I). R. 1982. New species from South America.
HI. Phytologia 51: 310-312.

Nomenclator of Hawaiian Sicyos (Cucurbitaceae)

Warren L. Wagner and Robynn K. Shannon

Department of Botany, MRC 166, National Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560, U.S.A. wagner.warren@nmnh.si.edu

ABSTRACT. In 1987 Harold St. John published 69
names in Hawaiian Cucurbitaceae in four short pa¬
pers. The nomenclature is analyzed, and the com¬
plete nomenclature of the Hawaiian species of Si¬
cyos is presented, including 105 names. Names
described by St. John in Sicyos and the segregate
genus Cladocarpa were incorporated by Telford in
his 1990 treatment, but those names described in
the segregate genera Sarx (1 name) and Sicyocarya
(14 names) were not treated and are here placed in
Telford’s classification. Five of the taxa described
by St. John in 1987 were recognized in Telford’s
taxonomy (Cladocarpa herbstii [= Sicyos herbstii],
Sicyos hillebrandii var. anunu [= S. anunu ], S. eros-
tratus , S. waimanaloensis , and Sicyocarya lanceo-
loidea [as sp. A = Sicyos lanceoloideus ]). Two of
these were transferred to Sicyos by Telford, and one
of them is here transferred to Sicyos [Sicyocarya
lanceoloidea\. All of the other taxa described by St.
John are placed in the synonymy of other species.
Five of the names were not validly published ( Cla¬
docarpa kawelaensis, Sicyocarya cruralis, Sicyocar¬
ya hawaiiensis, Sicyos erostratus var. kealiaensis ,
and Sicyos microcarpus var. kealiaensis ).

After the Manual of the Flowering Plants of Ha¬
waii (Wagner et al., 1990) was sent to the publish¬
er, St. John published a hurried series of papers,
cited in Wagner et al. (1990), in which he pub¬
lished about 880 manuscript names that he had
accumulated over his entire career of studying the
Hawaiian flora. Most of the names were validly
published, but some were invalid or illegitimate.
Over a period of about 30 years, he had developed
lengthy manuscripts with accompanying detailed il¬
lustrations for a number of genera in a style similar
to his treatment of O’ahu Cyrtandra (St. John,
1966). Few of them were ever completed and even
fewer published. When the Wagner et al. (1990)
flora manuscript was in the final stages of being
readied for the publisher, St. John published just
the new taxa and combinations from all of these
manuscripts in Phytologia and privately in a series
of papers that provided little more than a brief Lat¬
in diagnosis and an abbreviated mention of the
type. Not only were the treatments of these numer¬

ous new taxa extremely brief, they were done with¬
out rechecking specimens (unpublished obs. by
WLW). This led to many errors in citation of spec¬
imens, reusing names in the same genus (often even
in the same paper), and reusing the same specimen
as the type of more than one name. More impor¬
tantly. he did not incorporate the taxa in a carefully
evaluated taxonomic system. Now, after nearly 10
years, most of the approximately 880 names (in¬
cluding combinations) have been analyzed. St. John
names in Campanulaceae published in 1987 and
1988 have been studied by T. Lammers (see Lam-
mers, 1998, and papers cited therein).

This is the second of a series of papers analyzing
the names published in 1987 and 1988 by St. John,
in which the nomenclature is evaluated and the
taxa incorporated in a current classification. Ad¬
ditional papers in this series have already dealt or
will in the future deal with names in Phyllostegia
(Wagner, 1999), Stenogyne (both Lamiaceae), and
Cyrtandra (Gesneriaceae) and the special problems
of the invalid and illegitimate names.

Although all names for Hawaiian angiosperms
published through 1986 and a few from 1987 were
included by Wagner et al. (1990) in the Manual of
the Flowering Plants of Hawai'i, a rigorous assess¬
ment of them has never been done. In 1983, as
part of the Hawaiian Flora Project at the Bishop
Museum, a Hawaiian nomenclatural database was
initiated. It included bibliographic and type infor¬
mation, and current classification. Many of the spe¬
cialists contributing treatments to the Manual also
completed database record forms for names based
on Hawaiian specimens. Aspects of the project cur¬
rently being developed include a database intended
for W WW publication, a book, and several papers
on specific problematic groups. In analyzing the no¬
menclature for the genus Sicyos (Cucurbitaceae), a
number of problems were discovered. These in¬
cluded names not validly published or illegitimate,
errors in original publications, typification prob¬
lems, and the incorrect placement of one name (Si-
cyos microcarpus var. kealiaensis ) in synonymy in
the Manual treatment of Telford (1990). We here
present a complete nomenclator ol Hawaiian .Sicyos,

Novon 9: 441^447. 1999.

442

Novon

including a number of not validly published names.
One new combination by Wagner and Herbst is pro¬
vided for an uncommon species from northern
Kaua‘i and the Wai‘anae Mountains, 0‘ahu, which
was treated as “species A” in Telford’s (1990) treat¬
ment. Telford provided nomenclatural forms for
most of the names published through 1986 as part
of his contribution to the Hawaiian Flora Project.
We have confirmed and added to this information,
analyzed the 69 names published in 1987 by St.
John (1987a, 1987b, 1987c, 1987d), and added the
3 combinations made by Telford (1989). Names de¬
scribed by St. John in Sicyos and the segregate ge¬
nus Cladocarpa were incorporated by Telford in his
1990 treatment, but those described in the segre¬
gate genera Sarx (one name) and Sicyocarya (14
names) were not and are here placed in the clas¬
sification. Five of the taxa described by St. John in
1987 were recognized in Telfords taxonomy (Cla¬
docarpa herbstii [= Sicyos herbstii], Sicyos hille-
brandii var. anunu \ = S. anunu ], S. erostratus, S.
wainuinaloensis, and Sicyocarya lanceoloidea [as
sp. A = Sicyos lanceoloidea ]). Two of these were
transferred to Sicyos by Telford, and one of them is
here transferred to Sicyos [ Sicyocarya lanceoloidea ].
All of the other taxa described by St. John are
placed in the synonymy of other species. In cases
where we have not seen any material of a taxon,
the disposition is based on Telford’s (1990) place¬
ment of the name. There are a number of names
lor which the holotype is missing or destroyed, a
couple of situations where no holotype was desig¬
nated, and others where no original material can be
located. We have refrained from designation of lec-
totypes or neotypes in all of these cases following
recommendation 9A of the ICBN (Greuter et al.,
1994). As stability of accepted names does not
seem in question, we prefer to leave these selec¬
tions to those engaged in detailed revision of the
genus. Five of the names were not validly pub¬
lished. Since the emphasis of this paper is nomen¬
clature rather than taxonomy, the accepted names
are arranged alphabetically. All of the indigenous
Hawaiian species of the tour segregate genera de¬
scribed by St. John ( Cladocarpa (H. St. John) H.
St. John, Sarx H. St. John, Sicyocarya H. St. John,
Skottsbergiliana H. St. John) are placed in the ge¬
nus Sicyos as recommended by Jeffrey (1978) and
followed by Telford (1990).

Nomenclature

I . Sicyos alba (H. St. John) Telford, Phytologia

67: 210. 1989. Sarx alba H. St. John, Bot.

Jahrb. Syst. 99: 493. 1978. TYPE: Hawaiian

Islands [U.S.A.]. Hawai'i: Kulani, Sep. 1974,
W. C. Gagn d 698 (holotype, BISH-502382; iso¬
types, BISH [2]).

2, Sicyos anunu (H. St. John) Telford, Phytologia

67: 211. 1989. Sicyos hillebrandii H. St. John
var. anunu H. St. John, Phytologia 63: 187.
1987. TYPE: Hawaiian Islands [U.S.A.]. Ha¬
waii: Kapapala, Ka'u, 4—7 Aug. 1911, C. N.
Forbes 400. H (holotype, BISH-92745).

Sicyos hillebrandii var. douglasii H. St. John, Phytologia
63: 187. 1687. TYPE: Hawaiian Islands (U.S.A. ].
Hawai'i: N slope of Mauna Kea, near where David
Douglas was murdered, 23 Aug. 1949, 0. Degener
A A. Greenwell 21364 (holotype, BISH-522868; iso¬
type, NY).

Sicyos hillebrandii var. lanaiensis H. St. John, Phytologia
63: 188. 1987. TYPE: Hawaiian Islands [U.S.A. |.
Lana'i: Kamoa, 1740 ft., 16 May 1929, G. G. Munro
417 (holotype, BISH-633922; isotypes, PFBG [3 J).

3. Sicyos cucumerinus A. Gray, Proc. Amer.

Acad. Arts 3: 54. 1853; U.S. Expl. Exped.,
Phan. 652. 1854. Sicyocarya cucumerina (A.
Gray) H. St. John, Phytologia 38: 407. 1978.
TYPE: Hawaiian Islands [U.S.A.]. Hawai‘i:
[Mauna Kea, upper part of the forests], 1840,
U.S. Expl. Exped. s.n. (holotype, US-48017).
The material used by A. Gray in descriptions
of S. cucumerinus and the unnamed (3 and y
varieties was mounted on two sheets. Gray’s
species description is of the sheet taken as the
holotype, with short descriptions of the two ad¬
ditional taxa given afterward. They are both
mounted on US-48018. The specimen of va¬
riety fi is S. alba (St. John) Telford and that of
variety y is a lobed-leaf form of S. cucumeri¬
nus.

Sicyos cucumerinus var. triangulata Cogniaux, VI or mgr.
Phan. 3: 898. 1881. Syn. nov. TYPE: Hawaiian Is¬
lands [U.S.A.]. Maui: [East Maui, Haleakala), 27
Jan.— 12 Feb. 1870, H. Wawra 1916 (syntype, W
[presumably destroyed]); H. Wawra 1619 (syntype,
W [presumably destroyed]). Locality taken from Wa¬
wra (1873). Cogniaux cited //. Wawra 1619 anti
191 6; however, Wawra (1873) cited only the latter
collection, and thus we have no information about
the former. Present disposition based on locality and
description.

Sarx kauaiensis H. St. John, Bot. Jahrb. Syst. 109: 257.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Kaua‘i: along Wafalae Stream between 1st and 2nd
bend above Wai'alae Cabin, 3600 ft., 10 Sep. 1980,
K. M. Nagata 2141 (holotype, BISH-520191).
Sicyocarya mauiensis 11. St. John, Phytologia 64: 53.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: [East Maui], Haleakala Crater, Kaupo Cap,
second cove S of Waikeke'ehia, in grove on lower
part of N-facing slope. 6000 ft.. 27 Aug. 1945, H.

Volume 9, Number 3
1999

Wagner & Shannon
Nomenclator of Hawaiian Sicyos

443

,S7. John & A. /,. Mitchell 21198 (holotype, KISH
521561; isotypes, IlISll |2|).

Sicyocarya molokaiensis H. St. John, Phytologia 64: 53.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.).
Moloka'i: Kahua'awi Gulch, partly shaded thicket
near stream, 4 June 1928, 0. Degener 18398 (ho¬
lotype. NY-58603).

Sicyocarya saltuaria H. St. John. Phytologia 64: 54. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawaii:
South Kona, Honaunau Forest Reserve, 3000 ft.,
Nov. 1957, L IT Bryan s.n. (holotype, BISH-92762;
isotype, BISH).

Sicyocarya umbellata H. St. John, Phytologia 64: 54.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A. |.
Maui: [East Maui], liana Hist., along ridge in central
Kipahulu Valley, 3580 ft.. 22 July 1980, P. K. Hi -
gashino & R. A. Holt 9412 (holotype, BISH-463291).

4. Sicyos erostratus H. St. John, Phytologia 63:
187. 1987. TYPE: Hawaiian Islands [U.S.A.].
Molokai: Onini Gulch, 2200 ft., 21 Jan. 1973,
L. E. Bishop 1714 (holotype, BISH-519993).

Sicyos erostratus var. herbstii 11. St. John, Phytologia 63:
187. 1987. TYPE: Hawaiian Islands [U.S.A.]. Mo¬
loka'i: Kalamalua, past Homalani Cemetery, on jeep
road to Molokai Forest Reserve, on S side of road.
600 It.. 3 Apr. 1973, I). Herbs t 2944 (holotype,
BISH -5 19978; isotypes, BISH [3], US).

Sicyos erostratus var. paucipilosus H. St. John, Phytologia
63: 187. 1987. TYPE: Hawaiian Islands [U.S.A.].
Moloka'i: Onini Gulch, 2300 ft., 21 Jan. 1973, L. E.
Bishop 1713 (holotype, BISH-519979).

3. Sicyos herbstii (H. St. John) Telford, Phyto¬
logia 67: 210. 1989. Cladocarpa herbstii H. St.
John. Phytologia 63: 185. 1987. TYPE: Ha¬
waiian Islands [U.S.A.]. Kauai: Waimea Dist.,
Queen’s Pond area of Barking Sands, 10 ft., 26
Feb. 1976, D. Herbst & S. Ishikawa 5691 (ho¬
lotype, BISH-520171 and -520396; isotypes,
BISH [4], US).

Cladocarpa julesii H. St. John, Phytologia 63: 185. 1987.
TYPE: Hawaiian Islands [U.S.A.]. kaua'i: Re' my

542 (holotype, P not seen).

Cladocarpa pauciramosa H. St. John, Phytologia 63: 186.
1987. TV PE: Hawaiian Islands [U.S.A.f Kaua'i: Po-
lihale Cliffs, s.d., C. Christensen 326 (holotype,
BISH-520196; isotypes, BISH [3]).

Sicyocarya lebishopii H. St. John, Phytologia 64: 53. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Kaua'i:
Waimea Canyon, Koaie Branch, on floor of gulch,
2050 It., 3 June 1972, L. E. Bishop 1316 (holotype,
BISH-521557; isotypes, BISH |2|).

6. Sicyos hillebramlii H. St. John, Occas. Pap.
Bernice P. Bishop Mus. 10(12): 7. 1934. Sicyos
laciniatus Hillebrand, FI. Hawaiian Isl. 138.
1888. Nom. illeg., non L. (1753). TYPE: Ha¬
waiian Islands [U.S.A.]. Maui: East Maui,
Kula, 1851—1871, W. Hillebrand s.n. (holo¬
type, B [presumably destroyed]; isotypes.

BISH, K). The BISH sheet has the label print¬
ed by C. N. Forbes for Lydgate’s herbarium
stating the collectors as Hillebrand and Lyd¬
gate, but Hillebrand did not give Lydgate as
one of the collectors.

7. Sicyos hispidus Hillebrand, FI. Hawaiian Isl.
136. 1888. Cladocarpa hispida (Hillebrand) H.
St. John, Bot. Jahrb. Syst. 99: 491. 1978.
TYPE: Hawaiian Islands [U.S.A.]. Maui: East
Maui, Kula, 1851—1871, W. Hillebrand s.n.
(holotype, B [presumably destroyed, photo
BISH]; isotypes, BISH, K not seen). The BISH
sheet has the label printed by C. N. Forbes for
Lydgate’s herbarium stating the collectors as
Hillebrand and Lydgate, but Hillebrand did
not give Lydgate as one of the collectors.

Cladocarpa capitata H. St. John, Phytologia 63: 185.
1987. TYPE: Hawaiian Islands [U.S.A.]. Hawaii:
North Kona Hist., land of Pu'u Anahulu, Pu'u
Wa'awa'a Ranch at base of Pu'u Anahulu Ridge near
Mamalahoa Hwy., 2 KM) ft.. 26 June 1975, I). Herbst
A 5. Ishikawa 5393 (holotype, B1SH-520662; iso¬
types. BISH [2]).

Cladocarpa discoidea 11. St. John, Phytologia 63: 185.
1987. TYPE: Hawaiian Islands [U.S.A.]. Maui: West
Maui, Pohakea Gulch, arid rocky valley, 1 1 July
1927, O. Degener 9893 (holotype, NY-58605). De¬
gener used this collection number more than once;
US sheet of 0. Degener 9893 is Wahlenbergia grac¬
ilis. collected 5 May 1928.

Cladocarpa harrisonae H. Si. John, Phytologia 63: 185.
1987. TV PE: Hawaiian Islands [U.S.A.]. Molokai:
Kulaupapa, 1974, B. Harrison s.n. (holotype, B1SH-
520193; isotypes, BISH |2|).

Cladocarpa lanaiensis H. St. John, Phytologia 63: 186.
1987. TYPE: Hawaiian Islands [U.S.A.f Lanai:
Maunalei, J. Jacobi 1979 (holotype, BISH not seen).
Holotype missing at least since October 1987. The
brief diagnosis by St. John suggests Sicyos hispidus ;
thus this disposition is tentative.

Cladocarpa munroi H. St. John, Phytologia 63: 186. 1987.
TYPE: Hawaiian Islands [U.S.A.). Lana'i: Ka'ena, 25
Apr. 1914. G. C. Munro 87 (holotype, BISH-92762).
Cladocarpa umbellata H. St. John. Phytologia 63: 186.
1987. TYPE: Hawaiian Islands [U.S.A.]. Hawaii:
belt road between Huehue and Pu'u Wa'awa'a, North
Kona, 1900 ft., 9 Nov. 1950, IT //. Hatheway 509
(holotype, B1SH-92743; isotype, BISH).

Cladocarpa waimeaensis H. St. John, Phytologia 63: 186.
1987. TYPE: Hawaiian Islands [U.S.A. ]. Hawaii: 20
mi. along main road from Waimea toward Kona, 18
Aug. 1926. 0. Degener 17416 (holotype, NY-58613).
St. John did not annotate this sheet as the holotype,
but it is the only collection with this number from
this locality at NY. Degener distributed another col¬
lection (type of Sicyos waimanaloensis H. St. John)
from a different locality and date under the same
number.

444

Novon

8. Sieyos lanceoloideus (H. St. John) W. L. Wag¬

ner & D. R. Herbst, comb. nov. Basionym: Si-
cyocarya lanceoloidea H. St. John, Phytologia
64: 52. 1987. TYPE: Hawaiian Islands
[U.S.A.]. Kaua'i: Koke'e on left side of Waimea
Canyon, on side of gully, 7 July 1917, C. N.
Forbes 776.K (holotype, BISH-520175; iso-
types, BISH [4]).

Sieyos eucumerinus A. Gray var. tripartita Cogniaux, Mon-
ogr. Phan. 3: 898. 1881. Svn. nov. TYPE: Hawaiian
Islands | U.S.A.]. Kaua'i: llalemanu, [23 Eeb.— 29
Mar. 1870]. II. Wawra 2091 (syntype, W | presum¬
ably destroyed]); Waimea, 2(HK) — 3(KX) ft.. 1864—
1865, II. Mann & Hi T. Brigham 604 (syntype. 8151 1
1 2 1). Present disposition based on H. Mann & W. T.
Brigham syntype.

Sicyoearya kaalaensis H. St. John. Phytologia 64: 51.
1987. Svn. nov. TYPE: Hawaiian Islands [U.S.A. |.
()‘ahu: Wai'anae Mts., Mt. Ka'ala. Wai'anae Uka,
2600 ft., 18 Apr. 1948. II. St. John 23365 (holotype,
B1SH-519980; isotypes, BISH [8]).

Sicyoearya kalalanensis H. St. John, Phytologia 64: 52.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Kaua'i: Kalalan Valley, 18 July 1909, C. N. Forties
& ('.. II. Dole 59. K (holotype, B1SH-92766; isotype,
IJS).

Sicyoearya kauaiensis H. St. John, Phytologia 64: 52.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Kaua'i: Koke'e Camp, 25 June 1926, 0. Degener
13041 (holotype, NY-19546). St. John's paper indi¬
cated that all types are at BISH unless otherwise
noted, but there is no Degener 13041 collection at
BISH and St. John did not indicate any other insti¬
tution for the type. The NY sheet is the probable
holotype since it is the only sheet of this collection
we have seen, and the first set of Degener’s collec¬
tions are depositeil at NY or BISH. Perhaps St. John
intended to add "NY” to indicate that the type was
there.

Sicyoearya obatae II. St. John, Phytologia 64: 53. 1987.
Syn. nov. TYPE: II awaiian Islands [U.S.A. |. O'ahu:
Wai'anae Mts., Wai'anae Kai-Makaha ridge, dry na¬
tive forest. 2250 ft., 4 Nov. 1973. W. C. Gagin1. J.
Obata & I). Palmer 6 32 (holotype, BISH-521564;
isotypes. BISH [2], ITBG, US).

9. Sieyos Insioceplialus Skottsberg, Acta llorti

Gothob. 10: 184. 1936. Skottsbergiliana lasi-
ocephala (Skottsberg) H. St. John, Pacific Sci.
28: 458. 1974. TYPE: Hawaiian Islands
[U.S.A.]. Hawai'i: N Kona, along road near
Huehue, on 'a'a lava, 26 Sep. 1926, C. Skotts¬
berg 1947 (holotype, GB [on 2 sheets] not
seen; isotypes, BISH). St. John’s lectotypifiea-
tion (1974 [1975]) was superfluous. Skottsberg
mentioned three collections in the protologue,
but segregated his number 1947 in a separate
paragraph and compared it to other collec¬
tions. He referred to “the type specimen” in
the next paragraph, and wrote “type” on the
collection label.

Skottsbergiliana partita H. St. John, Pacific Sci. 28: 460.
1974. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i: 20
mi. from Waimea toward Kona, 'a'a desert along
road, 18 Aug. 1926. 0. Degener 9396 (holotype, GB
not seen; isotype, NY). St. John apparently consid¬
ered 0. Degener 9396 to be a mixed collection, as
he added an "A” to the collection number of a sheet
at NY and annotated it as Skottsbergiliana lasioce-
phala at the same time he annotated the other NY
sheet as an isotype of S. partita.

10. Sieyos inaerophyllus A. Gray, Proc. Amer.
Acad. Arts 3: 54. 1853; U.S. Expl. Exped.,
Phan. 651. 1854. Sicyoearya macrophylla (A.
Gray) H. St. John, Phytologia 38: 407. 1978.
TYPE: Hawaiian Islands [U.S.A.]. Hawai'i: [in
the forests of Mauna Kea, toward their upper
margin], 1840, U.S. Expl. Exped. s.n. (holo¬
type, US-48022).

Sicyoearya beccifer H. St. John, Phytologia 64: 51. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i:
North Hilo Hist., Hilo Forest Heserve, E slope of
Mauna Kea. makai of Keanakohi Bd., 5880 ft., 17
Nov. 1977. F. R. Warshauer 1535 (holotype, B1SH-
463288).

Sicyoearya kilaueaensis II. St. John, Phytologia 64: 52.
1987. Syn. nov. TYPE: Hawaiian Islands | U.S. A . ].
Hawai'i: Kilauea Volcano; mile 23, wet forest, Aug.
1918, ./. F. Rock 16006 (holotype, BISH-S202(X);
isotypes, BISH [4]).

Sicyoearya kipahulnensis H. St. John. Phytologia 64: 52.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: [East Maui], Kipahulu Stream, 22 Nov. 1919,
C. I\. Forbes I709.M (holotype, BISH-92763; iso¬
type, BISH).

Sicyoearya konaensis H. St. John, Phytologia 64: 52. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i:
Kona. McCandless forests, 2500 ft., Eeb. 1958, J. F.
Rock 20013 (holotype. BISH-520402; isotypes,
BISH 1 2 1).

Sicyoearya puberula H. St. John, Phytologia 64: 54. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i:
eastern slopes of Mauna Ke'a. near David Douglas
Historical Monument, ahupua'a of Uiupahoehoe,
5800 ft., 29 Nov. 1969, C. II. Uimonreux 4332 (ho¬
lotype, BISI 1-520676; isotypes, BISH, CBG, US).
Specimen label indicates liquid-preserved flowers
and fruits at Univ. of Hawaii Botany Dept.
Sicyoearya rockii II. St. John, Phytologia 64: 54. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Hawai'i:
North Kona I list., Pu'u Wa'awa'a Ranch (Waihou/
Waimea Paddock). 3000 ft.. 18 July 1975, I). Herbst.
S. Montgomery & G. Spence 5433 (holotype, BISH-
519989; isotypes. BISH [3]). Holotype does not have
Montgomery named as a collector, but isotypes do,
and Herbst (pers. comm. 1997) verified that Mont¬
gomery was one of the collectors.

11. Sieyos maximow ic/.ii Cogniaux, in DC.,
Monogr. (than. 3: 895. 1881. Cladocarpa nuix-
imowiezii (Cogniaux) H. St. John, Bot. Jahrb.
Syst. 99: 491. 1978. TYPE: Hawaiian Islands

Volume 9, Number 3
1999

Wagner & Shannon
Nomenclator of Hawaiian Sicyos

445

[U.S.A.]. O’ahu: 1826—1829, G. Kastalsky s.n.
(holotype, LE not seen).

Sicyos niihauensis H. St. John, Pacific Sci. 13: 175. 1959.
Cladocarpa niihauensis (H. St. John) H. St. John,
Bot. Jahrh. Syst. 99: 491. 1978. TYPE: Hawaiian
Islands [U.S.A.|. Ni’ihau: 1 mi. W of ki’i. 50 ft., 29
Mar. 1949, II. St. John 23567 (holotype, BISH-
501834, -501835, -501836, and -501837; isotypes,
BIS1I [2], PTBG [41).

Sicyos atollensis II. St. John, Pacific Sci. 24: 446. 1970.
Cladocarpa atollensis (H. St. John) H. St. John, Bot.
Jahrh. Syst. 99: 491. 1978. TYPE: Northwestern Ha¬
waiian Islands [U.S.A.]. Ocean Island [Kure Atoll |:
Scaevola “islands” in central plain and nearby, 4 m,
18 Apr. 1923, E. L. Caum 13 (holotype, BISH-
501816; isotypes, B1SH. NY, % not seen).

Sicyos caumii 11. St. John. Pacific Sci. 24; 448. 1970.
Cladocarpa caumii (II. St. John) H. St. John. Bot.
Jahrh. Syst. 99: 491. 1978. TYPE: Northwestern Ha¬
waiian Islands [U.S.A.]. Pearl and Hermes Beef.
Seal Island, eastern half of higher part ot island. 3
m. 27 Apr. 1923, E. L. Caum 53 (holotype, BISII-
501818; isotype, BISH).

Sicyos lamoureuxii H. St. John, Pacific Sci. 24: 451 . 1970.
Cladocarpa lamoureuxii (II. St. John) II. St. John,
Bot. Jahrh. Syst. 99: 491. 1978. TYPE: Eure, near
end ol runway on W side, 12 Sep. 1961, C. II. hi-
moureux 1901 (holotype, BISH-501822; isotype.
BISH). St. John cited “nos. 1899, 1900, 1902-3 are
evidently isotypes (BISH)”; Lamoureux (pers. comm.
1998) indicated he assigned different numbers when
specimens came from different plants, thus we do
not consider these to be isotypes, but rather para-
types.

Cladocarpa lehuaensis H. St. John, Phytologia 63: 186.
1987. TYPE: Hawaiian Islands [U.S.A.]. Uehua: 4
May 1965, C. R. Long 3017 (coll. W. King & l. Hoe-
mon) (holotype, HAW not seen; isotype, BISH). Ac¬
tually. the sheet at BISH here listed as an isotype
may in fact be the holotype. G. Carr (pers. comm.
1998) indicated that there is no sheet of this collec¬
tion extant at HAW', and that many of the Long col¬
lections were given to BISH.

12. Sicyos pachycarpus Hooker & Arnott, Bot.
Beeehey Voy. 83. 1832. TYPE: Hawaiian Is¬
lands [U.S.A.]. O’ahu: Diamond Hill [Head],
1826-1827, A. Collie s.n. (lectotype, designat¬
ed by Telford (1989), K not seen). The K sheet
was a mixed collection of two elements: one of
Sicyos waimanaloensis, annotated by St. John
as lectotype, but which was never published
(ined.); and the other the same taxon as the
type of Sicyos microcarpus H. Mann selected
by Telford as the lectotype. The two elements
have been remounted on separate sheets.

Sicyos microcarpus H. Mann, Proc. Amer. Acad. Arts 7:
167. 1867. TYPE: Hawaiian Islands [U.S.A. |. O'ahu:
1851—1855, J. lid my 541 (holotype, A not seen; iso¬
types, BB not seen, P).

Sicyos remyanus Cogniaux, in DC., Monogr. phan. 3: 898.
1881. TYPE: Hawaiian Islands [U.S.A.]. O’ahu:

1851-1855, J. He my 540 (holotype, P not seen; iso¬
types, BB not seen, P [3] not seen).

Sicyos laysanensis H. St. John, Pacific Sci. 24: 440. 1970.
Sicyos microcarpus subsp. et var. laysanensis (H. St.
John) 11. St. John, Phytologia 63: 190. 1987. TY PE:
Hawaiian Islands [U.S.A.]. Laysan: May 191 1. IT A.
Bryan A (holotype, BISH-501827; isotype. BISH).
Two other specimens, collected in April 1903, were
annotated as “isotype” by St. John, but are not iso-
types.

Sicyos nihoaensis H. St. John, Pacific Sci. 24: 443. 1970.
TYPE: Hawaiian Islands [U.S.A.]. Nihoa: W slope of
Tanager Peak. 2(H) m, 12 July 1924, E. Christopher-
sen A- T. Dranga 9 (holotype, BISH-501829; iso-
types, BISH. US).

Sicyos microcarpus var. ewaensis II. St. John. Phytologia
63: 190. 1987. TYPE: Hawaiian Islands |U.S.A.[.
O’ahu: W side of coral plains at Ewa, near Barber’s
Point, ocean side of Malakole Bd. near Camp Ma-
lakole Military' Bes.. 10 ft.. 21 June 1978, k. M.
\agata & H. Kimura 1656 (holotype, BISH-522860;
isotype, BISH).

Sicyos microcarpus var. forhesii H. St. John, Phytologia 63:
188. 1987. TYPE: Hawaiian Islands [U.S.A.]. Maui:
West Maui. Maunaho'oma. May 1910, C. A. Eorhes
I I.M (holotype, BISH-92753).

Sicyos microcarpus var. haleakalaensis H. St. John, Phy¬
tologia 63: 191. 1987. TYPE: Hawaiian Islands
[U.S.A. |. Maui: [East Maui), Wai’opai, C. A. Eorhes
1351. M (holotype, B1SII not seen). Holotype missing
at least since October 1987.

Sicyos microcarpus var. kalaupapaensis H. St. John. Phy¬
tologia 63: 188. 1987. TYPE: Hawaiian Islands
[U.S.A.], Moloka’i: Kalaupapa beach, 24 May 1918,
./. E Hock 14061 (holotype, BISH-519999; isotypes,
BISH [ 2 1).

Sicyos microcarpus var. kamoaensis H. St. John. Phytologia
63: 189. 1987. TYPE: Hawaiian Islands [U.S.A. |.
Lana’i: Kamoa. 1740 ft., 16 May 1929, G. C. Munro
4IH (holotype, BISH-91960; isotypes, BISH |3|).

Sicyos microcarpus subsp. kanepuuensis H. St. John, Phy¬
tologia 63: 189. 1987. TYPE: Hawaiian Islands
| U.S.A.]. Lana’i: kanepu’u. 28 May 1927. G. ('..
Munro 10 (holotype, BISH- 100149; isotype, BISH).

Sicyos microcarpus var. kauensis H. St. John, Phytologia
63: 191. 1987. TYPE: Hawaiian Islands [U.S.A.].
llawai'i: ka*u. on lava, Apr. 1911. ,/. E Hock s.n.
(holotype, BISH- 190169).

Sicyos microcarpus var. kilaueaensis H. St. John, Phytolo¬
gia 63: 188. 1987. TYPE: Hawaiian Islands | U.S.A.].
llawai’i: Apua, llawai’i [Volcanoes] National Park,
on slope of pali, 2000 ft., 10 May 1943. G. 0. Ea-
gcrlund A A. M. Mitchell 544 (holotype, BISH-
519906).

Sicyos microcarpus var. koelcensis H. St. John, Phytologia
63: 190. 1987. TY PE: Hawaiian Islands [U.S.A. |.
Lana’i: mts. near ko’ele, June 1913, C. A. Eorhes
95. L (holotype, BISH-519997; isotypes, BISH [4]).

Sicyos microcarpus var. koloaensis H. St. John, Phytologia
63: 191. 1987. TYPE: Hawaiian Islands [U.S.A. |.
kaua’i: koloa Hist., ahupua’a of koloa, 1 mi. W of
Poipu Beach Park. 10 ft., 13 Apr. 1963, C. II. hi-
moureux 234 6 (holotype, BISI 1-520014; isotypes,
BISH [31).

Sicyos microcarpus subsp. et var. lanaiensis II. St. John,
Phytologia 63: 189. 1987. TYPE: Hawaiian Islands
[U.S.A.]. Lana’i: kolulu. 24 Mar. 1916, G. G. Munro

446

Novon

535 (holotype, BISH-92754; isotypes, li I S 1 1 |3]). Two
of the isotypes have additional collection data, pos¬
sibly written by Forbes, including a date of "3/29/
16.” This appears to be a misinterpretation of the
original handwritten label on the holotype.

Sicyos microcarpus var. meeboldii 11. St. John. Phytologia
63: 191. 1987. TYPE: Hawaiian Islands [U.S.A.].
Ilawai'i: North Kona, Kaupulehu, 3 mi. E of Hue-
hue. 20(H) ft., 23 June 1948. G. L Webster A R. L.
Wilbur /H37 (holotype. BISH-632025; isotype. TEX
not seen). St. John erroneously published the collec¬
tor number as Webster 1537. which is a specimen of
PbyUostegia grandijlora collected on O'ahu.

Sicyos microcarpus var. obovoideis H. St. John. Phytologia
63: 191. 1987. TYPE: Hawaiian Islands [U.S.A.].
Maui: s.l., s.d., II. h. Clay s.n. (holotype, BISH-
522864; isotype(s), BISH, PTBG, US). Although
published by St. John as collected by Clay, the label
clearly states: "Sent from Maui to H. E. Clay; given
by him to Bishop Museum May 16. 1954." Published
as “obovoideis," but St. John’s handwriting on the
specimen label is “ obovoideus .”

Sicyos microcarpus var. oniniensis II. St. John, Phytologia
63: 190. 1987. TYPE: Hawaiian Islands [U.S.A.].
Moloka’i: Onini Gulch bottom, 2200 ft., 21 Jan.
1973, /,. hi. Bishop 1690 (holotype, BtSH-513075).

Sicyos microcarpus var. paneensis H. St. John, Phytologia
63: 188. 1987. TYPE: Hawaiian Islands [U.S.A.].
O’ahu: Kamokuiki Valley, between Pu’u Iki and Pu’u
Kamaohanui, 12 Apr. 1933. 0. Degener 1 8035 (ho¬
lotype, NY-5861 7: isotypes, NY |2|).

Sicyos microcarpus var. pohalceaensis II. St. John, Phyto¬
logia 63: 188. 1987. TYPE: Hawaiian Islands

[U.S.A.]. Maui: West Maui, Pohakea Gulch, arid val¬
ley floor. I I July 1927. 0. Degener 18400 (holotype,
NY-58614).

Sicyos microcarpus subsp. puberulus II. St. John, Phyto¬
logia 63: 191. 1987. TYPE: Hawaiian Islands

[U.S.A.]. 0‘ahu: Wai'anae, arid roadside near town.
10 Apr. 1932, 0. Degener 18391 (holotype, US-
1862308; isotypes, NY |2|).

Sicyos microcarpus var. trilobata H. St. John, Phytologia
63: 190. 1987. TYPE: Hawaiian Islands [U.S.A.].
Moloka’i: Makolelau Gulch, 760 m. 15 May 1973.
W. C. Gagne' 676 (holotype, BISH-S22869).

Sicyos microcarpus var. woolfordii H. St. John. Phytologia
63: 189. 1987. TYPE: Hawaiian Islands [U.S.A.].
O’ahu: ‘Ewa coral flats. 1 1 Mar. 1957. II. A. Wool ford
s.n. (holotype, BISH-522862; isotype, BISH).

Sicyos niger H. St. John. Phytologia 63: 192. 1987. TYPE:
Hawaiian Islands [U.S.A.f Maui: East Maui, liana
Hist., Nu’u, 10 ft.. 28 Mar. 1971. II. St. John 26771
(holotype. BISH-522866; isotypes, BISH [7|, US).

Sicyocarya atrimaculata II. St. John. Phytologia 64: 51.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
Maui: East Maui, below Haleakala Ranch pipeline,
39(H) ft.. 20 Eel>. 1928, G. C. Munro 801 (holotype,
BISH-52 1555; isotype, BISH).

Sicyocarya rostrata II. St. John, Phytologia 64: 54. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Kaua’i:
Waimea Hist., in Queen’s Pond area of Barking
Sands. 10 ft.. 26 Eeb. 1976. I). Hcrbst A' S. Isliikawa
5691 A (holotype, BISH-520406; isolvpe. BISH).

13. Sicyos semitonsus H. St. John, Pacific Sci.
24; 453. 1970. Cladocarpa semitonsa (H. St.

John) H. St. John, Bot. Jahrb. Syst. 99: 491.
1978. TYPE: Hawaiian Islands [U.S.A.]. Lay-
san: NW end of lagoon in a flat area, 19 Sep.
1964, C. R. Long 2370 (holotype, BISH-
502486 and -502487). Original plate for illus¬
tration also mounted and in the collection
(BISH). St. John cited “other collections with
the same data. Long 2365, 2368, and 2369
(Univ. Hawaii) are evidently duplicates and
therefore isotypes.” But they must be consid¬
ered paratypes, since they have different col¬
lection numbers.

14. Sicyos waiiuanaloensis H. St. John, Phyto¬
logia 63: 192. 1987. TYPE: Hawaiian Islands
[U.S.A.]. O’ahu: talus SW of Waimanalo Land¬
ing, 10 Apr. 1936, (). Degener 17416 (holo¬
type, US-1864571; isotypes, BISH [2], NY
[3]). Degener distributed another collection
(type of Cladocarpa waimeaensis H. St. John,
a synonym of Sicyos hispidus ) from a different
locality and date under the same number.

Sicyocarya kaenaensis H. St. John, Phytologia 64: 51.
1987. Syn. nov. TYPE: Hawaiian Islands [U.S.A.].
O’ahu: Wai’anae Mts., Ka’ena Point, bluffs halfway
between Pu’u Pueo and Pu’u Manini, 3(H) ft., 6 Apr.
1976. G. I). Carr 898 (holotype. BISH-520399; iso¬
types, BISH [2]).

Sicyocarya kokocrateris H. St. John. Phytologia 64: 52.
1987. Syn. nov. TYPE: Hawaiian Islands |U.S.A.].
O'ahu: Koko Crater, Maunalua, on crest of S rim. in
thicket. 250 ft.. I Apr. 1979. //. St. John 26936
(holotype. BISH-52 1548; isotypes, BISH |2], US).

Sicyocarya paucifrugalis H. St. John, Phytologia 64: 53.
1987. Syn. nov. TYPE: Hawaiian Islands (U.S.A.].
Moloka’i: Waiakuilani Gulch bottom, 2700 ft., 25
Jan. 1973, I,, hi. Ilislwp A" T. Pratt 1766 (holotype,
BISH-52 1554).

Sicyocarya protrusa 11. St. John, Phytologia 64: 54. 1987.
Syn. nov. TYPE: Hawaiian Islands [U.S.A.]. Eana’i:
Maunalei Gulch. Maunalei, 500 ft.. 12 Apr. 1938,
II. St. John. A. J. Haines & hi. Y. Hosaka 18818
(holotype, BISH-5215S1; isotypes, BISH |2|).

Invaliim.y Published Names

Cladocarpa kawelaensis II. St. John, Phytologia 63: 185.
1987. No type cited for this name (ICBN Article
37.3; Greater et al.. 1994). - Sicyos hispidus

Sicyocarya cruralis H. St. John, Phytologia 64: 51. 1987.
Sicyocarya hawaiiensis II. St. John, Phytologia 64:
51. 1987. The same collection, Remy 539 (P). was
cited as type of both names. Neither name is validly
published under ICBN Art. 34.2 (Greuter et al.,
1994). = Sicyos macrophyllus

Sicyos erostratus H. St. John var. kealiaensis H. St. John,
Phytologia 63: 187. 1987. Sicyos microcarpus var.
kealiaensis II. St. John. Phytologia 63: 189. 1987.
The same collection, IE C. Gagnd 671 (BISH |2|),
was cited as type of both names. Neither name is
validly published under ICBN Art. 34.2 (Greuter et

Volume 9, Number 3
1999

Wagner & Shannon
Nomenclator of Hawaiian Sicyos

447

al., 1994). Telford included Sicyos erostratus var.
kealiaensis in S. erostratus ; he placed Sicyos micro-
carpus var. kealiaensis in the synonomy of .S'. pachy-
carpus. = Sicyos erostratus

Excluded Name

Sicyos Jauriei H. Leveille, Repert. Spec. Nov. Regni Veg.
10: 150. 1911. TYPE: Hawaiian Islands [U.S.A.].
0‘ahu: Kali k i [Kalihi], Oct. 1909, U. Faurie 877 (ho-
lotype, P not seen; isotype, BM not seen). The type
of this name is a specimen of the common weed,
Momordica charantia L. (see Lauener, 1980).

Acknowledgments. We thank Denis Kearns. Joe
Kirkbride, and Alex Monro for useful reviews of the
manuscript that improved the clarity and accuracy
of it. We appreciate the assistance and support of
the staff of the Bishop Museum Herbarium on our
visits there. We thank the curators of the following
herbaria lor the loan of specimens for this study:
BISH, NY, P. We appreciate the help of Werner
Greuter, John McNeill, and Dan Nicolson in apply¬
ing the ICBN to the problem of simultaneous pub¬
lication of two species using the same specimen.
We thank George Staples and Clyde Imada for an¬
swering questions about specific BISH collections.
We also thank Grady Webster and Tom Wendt for
helping resolve the problem with Sicyos microcar¬
pus var. meeboldii, and Pete Lowry for assistance
with the type of Sicyocarya cruralis.

Literature Cited

Greuter. W.. F. R. Barrie. H. M. Burdet, W. G. Chaloner,
V. Demoulin, D. L. Hawksworth, P. M. Jprgensen. I). H.

Nicolson, P. G. Silva, P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Jeffrey, C. 1978. Further notes on Cucurbitaceae: IV. Some
New-World taxa. Kew Bull. 33: 347—380.

Lammers, T. G. 1998. New names and new combinations
in Campanulaceae. Novon 8: 31—35.

Lauener, L. A. 1980. Faurie’s Hawaiian types at the Brit¬
ish Museum. Notes Roy. Bot. Card. Edinburgh 38: 495—
497.

St. John, II. 1966. Monograph of Cyrtandra (Gesneri-
aceae) on Oahu, Hawaiian Islands. Bernice P. Bishop
Mus. Bull. 229: 1—465.

- . 1974 1 1975]. Skotlsbergiliana new genus (Cucur¬
bitaceae) of Hawaii Island. Hawaiian plant studies 41.
Pacific Sci. 28: 457-462.

- . 1987a. Diagnoses of new species of Cladocarpa

(Cucurbitaceae): Hawaiian plant studies 142. Phytolo-
gia 63: 185-186.

- . 1987b. Diagnoses of new species of Sicyos (Cu¬
curbitaceae): Hawaiian plant studies 143. Phytologia
63: 187-192.

- . 1987c. Diagnoses of new1 species of Sicyocarya

(Cucurbitaceae): Hawaiian plant studies 160. Phytolo¬
gia 64: 51-54.

- . 1987d. Sarx kauaiensis sp. nova (Cucurbitaceae):

Hawaiian plant studies 132. Bot. Jahrb Syst. 109: 257-
259.

Telford, I. 1989. Notes on Sicyos (Cucurbitaceae) in the
Hawaiian Islands. Phytologia 67: 209—213.

- . 1990. Cucurbitaceae. Pp. 568—581 in W. L. Wag¬
ner. D. R. Herbst & S. H. Sohmer, Manual of the Flow¬
ering Plants of Hawai‘i. Univ. Hawaii Press and Bishop
Museum Press, Honolulu.

Wagner, W. L. 1999. Nomenclator and Review of Phyllo-
stegia (Lamiaceae). Novon 9: 265-279.

- , I). R. Herbst & S. H. Sohmer. 1990. Manual of

the Flowering Plants of HawaPi. Univ. Hawaii Press and
Bishop Museum Press, Honolulu.

Wawra, H. 1873. Beitrage zur Flora der hawaPschen ln-
seln. Flora 56: 168-176.

Rediscovery and Assessment of Stenogyne sherffii
Degener (Lamiaceae)

Warren L. Wagner

Department of Botany, MRC-166, National Museum of Natural History, Smithsonian Institution,
Washington, l).C. 20560, U.S.A. wagner.warren@nmnh.si.edu

Stephen G. Weller

Department of Ecology and Evolutionary Biology, University of California, Irvine,
California 92717, U.S.A. sgweller@uei.edu

ABSTRACT. Stenogyne sherffii 0. Degener from the
Ko‘olau Mountains, 0‘ahu, Hawaiian Islands, was
treated as synonymous with the more common spe¬
cies from the WaPanae Mountains, S. kaalae Wa-
wra, in the most recent revision of the genus by
Weller and Sakai in 1990. A single population of
about five plants of this entity was rediscovered in
1994. A plant was propagated from a stem cutting
in the greenhouse at the University of California,
Irvine. Comparison to plants of S. kaalae grown in
the same greenhouse and herbarium material shows
that die rare Ko‘olau plants are distinct in several
characters, especially subentire vs. serrate leaves.
We here recognize this taxon as a subspecies of S.
kaalae because of the clear species synapomorphy
of very dark maroon corolla color and the geograph¬
ical separation of the two taxa.

Stenogyne Bentham is an endemic Hawaiian ge¬
nus of vines in the Lamiaceae consisting of 21 spe¬
cies (Weller & Sakai, 1990; Wagner & Weller. 1991),
with greater diversity on younger than on older is¬
lands. Hawai‘i (8 species) and Maui (9 species) have
the greatest diversity, while lower diversity occurs on
Moloka‘i (3 species), Lana‘i (2 species), 0‘ahu (2
species), and Kaua‘i (3 species). The reduced lower
corolla lip, exserted stamens, abundant nectar pro¬
duction, lack of floral odor, and long falcate corollas
of some Stenogyne species provide circumstantial
evidence for pollination of Stenogyne by honeycreep-
ers (Weller & Sakai, 1990). In the Hawaiian genus
Phyllostegia Bentham, with which Stenogyne and
Haplostachys (A. Gray) Hillebrand share a common
ancestor, the presence of a sweet floral fragrance
among some species and the expanded lower corolla
lip and predominant white coloration indicate pol¬
lination by insects (Weller & Sakai, 1990).

After the publication of the treatment of Stenogyne
by Weller and Sakai (1990), J. Obata rediscovered a
small population of the plant described by Degener

(1943) as S. sherffii from the northern Ko‘olau Moun¬
tains, 0‘ahu. Subsequently, we have propagated it
from a stem cutting in the greenhouse at the Uni¬
versity of California, Irvine. For comparative pur¬
poses we have also grown S. kaalae Wawra, which
occurs at a number of localities in the Wai‘anae
Mountains, 0‘ahu (with a single collection made in
1852 from Nu‘uanu Pali in the southern Ko'olau
Mountains). The plants described as Stenogyne sherf¬
fii are closely allied to S. kaalae, and this taxon was
included with S. kaalae by Weller and Sakai (1990:
838) based on study of the type. Our greenhouse
study shows that the Ko‘olau Mountain plants differ
in several characters, although the relationship to S.
kaalae is very close. The morphological distinctions
coupled with the geographical separation of about
30 km support recognition of the Ko‘olau plants as
a subspecies of S. kaalae.

Stenogyne species on younger islands of Maui
and especially Hawai‘i are more polymorphic than
the species on the older islands of 0‘ahu and
Kaua‘i (Weller & Sakai, 1990). Similar patterns of
greater polymorphism within species on the youn¬
ger islands occur in a number of other genera in¬
cluding Cyrtandra .1. B. Forster & G. Forster (Ges-
neriaceae), Labordia Gaudichaud (Loganiaceae),
Lysimachia L. (Primulaceae), Melicope J. R. Forster
& G. Forster (Rutaeeae), Myrsine L. (Myrsinaceae),
and Phyllostegia (Lamiaceae). Extinction of inter¬
mediate forms and evolution of novel forms on older
islands, as well as increasing isolation resulting
from erosion and dissection of habitats, are likely
explanations for the greater ease in delimiting spe¬
cies on the older Hawaiian Islands. These factors
could explain the patterns of variability seen in
Stenogyne and other genera, where more narrowly
distributed, less polymorphic species are found on
the older islands. Stenogyne sherffii is an example
of a species where geographic isolation has led to

Novon 9: 448-449. 1999.

Volume 9, Number 3
1999

Wagner & Weller
Stenogyne sherffii

449

well-defined, through modest, differentiation from
S. kaalae. In contrast, differences among popula¬
tions found on the island of Hawaii may he of
greater magnitude, but lack consistency on a geo¬
graphic basis. Lack of consistency in the distribu¬
tion of characters prevents recognition of infraspe-
cifie taxa within these species and provides the
rationale for use of a broadly defined species for
many of the Stenogyne taxa found on the island of
Hawai‘i.

Comparison of several closely related taxa of
Stenogyne illustrates patterns of variation related to
island age that require different taxonomic ap¬
proaches and demonstrate why species-level rec¬
ognition is given in most of these cases, while sub¬
species level is used for S. sherffii. Stenogyne
macrantha Bentham and S. scrophularioides Ben-
tham, two polymorphic species from the island of
Hawai‘i, are very distinct in some areas in both
vegetative and floral characters, but show consid¬
erable morphological overlap in other areas. Pop¬
ulations that are intermediate may result from areas
of secondary contact or represent the extremes of
variation in a recently diverged ancestral species.
Without extensive study, these hypotheses cannot
be differentiated, and at present recognition of two
broadly defined species captures the essence of
variation in these entities. In contrast, differentia¬
tion is geographically and ecologically well marked
for a morphologically distinct pair of species from
East Maui; the extinct S. haliakalae Wawra, which
occurred at lower and drier elevations on the south
slopes of Haleakala, and S. rotundifolia A. Gray, a
species still common at higher, wetter elevations on
the north slopes of Haleakala. A third pattern of
variation is exemplified by S. purpurea H. Mann
and S. kealiae Wawra, both endemic to Kaua‘i.
These species differ only in vegetative characters,
but rare and narrowly distributed S. kealiae occurs
sympatrically with the widespread S. purpurea , and
therefore they are considered distinct species. Sten¬
ogyne sherffii and S. kaalae, which like S. kealiae
and S. purpurea on Kaua‘i differ only in vegetative
characters, are completely allopatric in the same
habitat type. The minor, strictly vegetative differ¬
ences and the allopatric distribution of S. sherffii
and 5. kaalae provide the justification for subspe¬
cific recognition.

Ki n to thk Sfbspkciks ok Sm\o(;v\E kaaiak

la. Leaf margins inconspicuously serrulate, the
blades 8—12 cm long, 2—3.5 cm wide, coria¬
ceous; stems glabrous; pedicels 2—5 mm long

. subsp. sherffii

lb. Leaf margins serrate, the blades 4. 2-7. 7 cm
long, 1.7-2. 9 cm wide, thin and membranous;

stems usually pubescent; pedicels 4—7 mm long
. subsp. kaalae

Stenogyne kaalae Wawra subsp. sherffii (0. De-
gener) W. L. Wagner & Weller, comb, et stat.
nov. Stenogyne sherffii 0. Degener, Brittonia 5:
58. 1943. TYPE: Hawaiian Islands [U.S.A.].
0‘ahu: [Ko‘olau Mountains], Kawailoa,
Pe‘ahinai‘a trail, 28 Apr. 1940, 0. Degener &
Ordonez 12999 (holotype, NY; isotype, NY).

Seandent vines, climbing or sprawling and root¬
ing at the nodes; stems 2.5—6 m (in cultivation),
quadrangular, glabrous. Leaves glossy, coriaceous,
lanceolate, the blades 8—12 cm long, 2—3.5 wide,
antrorsely strigulose along the adaxial midvein, oth¬
erwise glabrous, margins inconspicuously serrulate,
apex acute to weakly acuminate, base cuneate, pet¬
ioles 1-2 cm long. Flowers (3)5 per verticillaster,
sometimes some of them on a peduncle up to 2—5
mm long, pedicel 2-5 mm long, retrorsely strigu¬
lose; calyx nearly radially symmetrical, narrowly
campanulate, 9-15 mm long at anthesis, green or
tinged purple, glabrous, the teeth linear-lanceolate,
5—8 mm long; corolla very dark maroon, hispidu-
lous externally, the hairs weakly antrorse, the mar¬
gins and inner surfaces glandular puberulent, the
tube mostly straight, curved slightly at the throat,
11—12 mm long, upper lip 9—10 mm long, lower lip
2-3 mm long. Nutlets 4 mm long, fleshy, dark pur¬
ple.

Distribution. Known only from the Pe‘ahinai‘a
Trail in the northern part of the Ko‘olau Mountains,
0‘ahu, from only two collections (1940 and 1994).
Apparently there is only a single extant population
with perhaps five individuals (J. Obata, pers.
comm.).

Specimen examined. HAWAIIAN ISLANDS [U.S.A.].
Oahu: Ko'olau Mountains, Pe‘ahinai'a Trail, 1994, Obata
s.n. (US), cultivated in 1997 from Obata s.n. (BISH, F, K,
US).

Acknowledgments. We appreciate the helpful
comments by R. Shannon and two anonymous re¬
viewers on the manuscript, and John Obata for pro¬
viding us with cuttings to propagate in the green¬
house.

Literature Cited

Degener, 0. 1943. Stenogyne sherffii, a new mint from
Hawaii. Brittonia 5: 58—59.

Wagner, W. L. & S. G. Weller. 1991. Resurrection of a
Kaua‘i Stenogyne: S. kealiae. Pacific Sci. 45: 50—54.
Weller, S. G. & A. K. Sakai. 1990. Stenogyne. Pp. 83 1 —
843 in W. L. Wagner, I). R. Herbst & S. H. Sohmer,
Manual of the Flowering Plants of Hawai'i. Univ. Hawaii
Press and Bishop Museum Press, Honolulu.

New Species of Justicia (Acanthaceae) from Gerro La Chapa,

El Amparo, Venezuela

Dieter C. Wasshausen

Department of Botany, National Museum of Natural History, Smithsonian Institution,
Washington, I).C. 20560-0166, U.S.A.

Abstract. Current research in the cloud forest of
Cerro La Chapa, El Amparo, hy Winfried Meier has
resulted in the recognition of a new species of Jus¬
ticia. This new species, Justicia chapana, previ¬
ously misidentified as J. galapagana Lindau, is de¬
scribed, illustrated, and compared to the latter, its
closest relative, from the Galapagos Islands.

Justicia is the largest and taxonomically most
complex genus of Acanthaceae, with ca. 600 spe¬
cies worldwide (Graham, 1988; Mabberley, 1997).
It is also the largest genus of the family in Vene¬
zuela, with about 45 species found there. The genus
is characterized by the upper corolla lip rugulate,
androecium of two 2-thecous stamens and no stam-
inodes, 2- or 3-aperturate pollen with a trema area
usually traversed by 2—8 rows of up to 10 insulae
(Fig. 2), and stipitate capsules with four (or fewer)
seeds.

Justicia chapana Wasshausen, sp. nov. TYPE;
Venezuela. Yaraeuy: El Amparo, 7 km N of
Salom, 1 10—1200 m, 1 Dec. 1974, J. Steyer-
mark, Carreno Espinoza, C. Steyermark &
Diederichs 111211 (holotype, US; isotype,
VEN). Figures 1, 2.

Herba, caulibus quadrangularibus, puberulis et glan-
duloso-pilosis. Foliorum lamina lanceolata vel paree ova-
ta, rnembranacea, utrinque paree puberula, costa et venis
lateralibus puberulis, acuminata, basi attenuata. Intlores-
centiae pauciflorae, axillares, dichotomae, folia aequantes,
apice spiciformes, flores in axillis bractearum alternanti-
bus, glanduloso-pilosi; bracteae bracteolaeque subulatae;
calycis segmenta 4, lance-subulata; corolla lavendulacea
vel profunde atrorosea-purpurea, 6—7 mm longa, pilosa,
tubo 3—4 mm longo, labio superiore anguste ovato, apice
euculato et rotundato; stamina exserta; lobis antherarum
leviter superpositis.

Branching herb 0.5 m tall. Stems ascending,
quadrangular-ridged, conspicuously puberulous
and glandular-pilose, trichomes erect, gland-
tipped, others ± spreading. Leaves petiolate, peti¬
oles 10-15 mm long, pubescence similar to that of
stem, blades lanceolate to narrowly ovate, 5—7.5 X
2—3.3 cm, membranous, dull green, sparingly and

inconspicuously puberulent both above and below,
especially on costa and lateral veins, trichomes
erect or appressed, eglandular or glandular, cysto-
liths usually obscure, apex long-acuminate, base
attenuate, margin entire to obscurely crenulate;
leaves subtending inflorescence much reduced,
narrower than cauline leaves. Inflorescence few-
flowered, axillary, dichotomously branched spikes,
these barely exceeding terminal leaves, densely
glandular-pilose, trichomes interspersed both glan¬
dular and eglandular. Flowers alternate, pedicels
almost none; bracts and bracteoles subulate, 3—3.5
X 0.5 mm, pubescence intermixed pilose and glan¬
dular-pilose; calyx 4.5— 5.5 mm long, deeply 4-part-
ed, lobes lance-subulate, 3^1.5 X 0.75—1 mm, pu¬
bescence similar to that of the bracts and
bracteoles; corolla rich lavender to deep rose-pur¬
ple, 6-7 mm long, externally moderately pilose,
tube 3^1 mm long, basally 1.2— 1.5 mm wide, throat
2.5-3 mm wide, upper lip narrowly ovate, 3^1 mm
long, basally 4 mm wide, apieally cucullate and
rounded, folded longitudinally inside, lower lip 3-
lobed, lobes narrowly ovate to elliptic, 1.5—2 mm
long, 1.5 mm wide, undulate at tip; stamens ex-
serted; filaments slender, flattened, 2 mm long, gla¬
brous, anther thecae obliquely attached to a flat¬
tened connective, upper thecae obtuse, 1 mm long,
lower thecae 1 mm long, apiculate at base; pollen
grains 2-porate, bilateral, 50 X 30 pan, trema area
microreticulate, traversed by 2 rows peninsulae
(1 ig. 2); ovary pubescent, brownish, style filiform,
5 mm long, pubescent, stigma oblique. Capsule 4-
seeded, clavate, 13-14 X 3-3.5 X 3 mm, seed¬
bearing-portion brownish, densely pilose, trichomes
both eglandular and glandular; retinacula 2 mm
long, slightly curved; seeds suborbicular, 1.5 mm
diam., densely pubescent with glochidiate tri-
chomes.

Distribution. Endemic to Venezuela; plants oc¬
cur in cloud forests and along upper slopes of riv¬
ers, in the departments of Carabobo, Yaraeuy, Co-
jedes, Lara, Portuguesa, and Trujillo at elevations
from 700 to 2300 m.

Justicia chapana superficially resembles J. gal-

Novon 9: 450-452. 1999.

Volume 9, Number 3
1999

Wasshausen
Justicia from Venezuela

451

Figure 1. A— F. Justicia chapana Wasshausen ( Steyermark el al. 111211). — A. Habit. — B. Bract and bracteoles. — C.
Calyx and pistil. — D. Pistil, ovary, cupula, and glands. — E. Corolla, stamens, and pistil. — F. Capsule, retinacula,
and seeds.

452

No von

Figure 2. Scanning electron (SF.M) photomicrographs of
Justicia chapana pollen. A. It. ./• chapana (, Sleyermark A'
Hunting 105258). — A. Equatorial view. X 1 145. — B.
Equatorial view, X1215.

apagana Lindau from the Galapagos Islands; in
fact, J. Steyermark identified his Venezuelan spec¬
imens as J. galapagana. In J. galapagana the co¬
rollas are considerably larger, 10—14 min long, pur¬
ple, the tube is 4— 6 mm long, the upper lip is
apically somewhat dentate, the middle lobe of the

lower lip is distinctly masked with maroon nectar
guides, and the leaf blades are basally abruptly
narrowed. By contrast, in J. chapana the corollas
are 6—7 mm long, rich lavender to deep rose-pur¬
ple, the tube is 3^k mm long, the upper lip is api¬
cally cucullate and rounded, the middle lobe of the
lower lip is not reported to have maroon markings,
and the leaf blades are basally attenuate. The new
species is placed in Graham’s (1988) section Sar-
otheca , which is characterized by its compound spi-
cate inflorescence with a glandular axis, and by its
type 6 pollen.

Paratypes. VENEZUELA. Carubobo: road to Tina-
quillo, Pittier 1 1986 (US). Yaracuy: Cerro I,a Chapa, sel¬
va nublada al norte de Nirgua, 1200-1400 m, Steyermark,
Hunting cV Wessels-Hoer 100289 (US. V EN), Steyermark &
Hunting 105258 (US. VEN); selva nublada en la li la v las
laderas superiores, entre Salom y Temerfca, 100 m, Stey¬
ermark. Hunting Wessels-Hoer 100859 (US, Y EN); 7.5
km N of Salom. 10°15'N, 68°29'W, 1200-1300 in, Liesner
& Steyermark 12406 (MO, US); Ouebrada Honda, 17.3
kin del pueblo de Arva, 1 100 m, Steyermark 10541 1 (US,
VEN). Cojedes: Dtto. San Carlos, Mun. Romulo Gallegos,
Llamozas 147 (US). Lara: Dtto. Moran, between llumu-
caro Alto and La Mesa, 1 100 in, Sleyermark & CarrerTo
Espinoza 1 1 1089 (US. VEN). Portuguesa: Dtto. Sucre,
la carretera Gavildn— Mijagual— Mesa de Bucaral, 15 km S-
SE of Biseucuy, 700 m. Aymard et al. 8695 (PORT, US);
Dtto. Guanare, along Hfo Guanare, above Mijagual, Ster-
gios et al. 8758 (PORT, US). Trujillo: Dtto. Boeono, La
Morita, above Rfo Saguaras, S ol Campo Elias and ca. 3
km YY of Trujillo-Portuguesa state line. 23(H) m. Dorr et
al. 5881 (NY. US).

Acknowledgments. My special thanks to Cathy
Pasquale lor skillfully preparing the line drawing,
and to the staff of the National Museum of Natural
History SEM Laboratory for their high-quality pol¬
len photomicrographs.

Literature Cited

Graham, V. A. W. 1988. Delimitation and infra-generic

classification of Justicia (Aeanthaeeae). Kew Bull. 43:

551-624.

Mabberley, I). 1997. The Plant-Book; A Portable Dictio¬
nary of the Vascular Plants. Cambridge Univ. Press,

Cambridge.

New Combinations in Chinese Kobresia (Cyperaeeae)

Zhang Shu- re n

Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China

Abstract. Three new subspecific combinations
are proposed in the genus Kobresia: K. myosuroides
(Villars) Fiori subsp. bistaminata (W. Z. Di & M. J.
Zhong) S. R. Zhang, K. royleana (Nees) Boeekeler
subsp. minshanica (Wang & Tang ex Y. C. Yang) S.
R. Zhang, and K. setschwanensis Handel-Mazzetti
subsp. squamaeformis (Y. C. Yang) S. R. Zhang.
Additionally, K. hashgarica Dickore is treated as a
synonym of K. myosuroides subsp. bistaminata and
K. menyuanica Y. C. Yang as a synonym of K. roy¬
leana subsp. minshanica.

In preparing the account of Kobresia for the
forthcoming volume of Flora of China , the f ollowing
new combinations have been found necessary.

Kobresia myosuroides (Villars) Fiori subsp. bi¬
staminata (W. Z. Di & M. J. Zhong) S. R.
Zhang, comb. nov. Basionym: Kobresia bistam¬
inata W. Z. Di & M. J. Zhong, Acta Bot. Bor.-
Oecid. Sin. 6: 275. 1986. TYPE: China. Nei
Mongol: Helanshan, Halawugou, 25 July 1984,
EHNWU [ Northwest University Expedition to
Helanshan ] 6051 (holotype, WNU).

Kobresia hashgarica Dickore. Stapfia, 39: 79. 1995. Syn.
nov. TYPE: China. Xinjiang: Kunlun Shan. Tiznap,
tributary valley W of Kudi, 36°51'N, 76°57'E, 4040
m, 9 July 1992, G. & S. Miehe 8471 (holotype. k:
isotypes. GOET, PE).

Kobresia myosuroides is widespread in the North¬
ern Hemisphere, with a range extending from
northern China to the Arctic region. Plants of north¬
western China (from western Nei Mongol to western
Xinjiang), whic h were described as K. bistaminata
and K. hashgarica and treated here as the subspe¬
cies bistaminata , differ from those of the rest of the
species range in having slightly smaller (2.5— 3.5
mm long) glumes with a narrowly hyaline or not
hyaline margin and smaller nuts 1.8— 2.3 mm long.
In contrast, subspecies myosuroides has glumes (2—)
3—4 mm long with a broadly hyaline margin and
nuts 2—2.8 mm long.

Kobresia royleana (Nees) Boeekeler subsp. min-
shaniea (Y. C. Yang) S. R. Zhang, comb. nov.
Basionym: Kobresia minshanica Wang & Tang
ex Y. C. Yang, Acta Biol. Plateau Sin. 2: 1.
1984. TYPE: China. Gansu: Minxian, 10 Aug.
1978, J. Q. Wang 228 (holotype, NWTC).

Kobresia menyuanica Y. C. Yang, Acta Biol. Plateau Sin.
2: 3. 1984. Syn. nov. TY PE: China. Qinghai: Men-
yuan. 6 Sep. 1980, A. M. Zhou 273 (holotype.
HNWP).

Kobresia royleana is distributed in the central
Asian highlands and Himalayas. Plants of subspe¬
cies minshanica , which are restricted to eastern
Qinghai and Gansu, are rather slender and often
have unisexual lateral spikelets with one female
flower, whereas those of subspecies royleana are
stout and have bisexual lateral spikelets with a bas¬
al female flower and several distal male flowers.

Kobresia setschwanensis Handel-Mazzetti subsp.
squamaeformis (Y. C. Yang) S. R. Zhang,
comb. nov. Basionym: Kobresia squamaeformis
Y. C. Yang, Acta Biol. Plateau Sin. 2: 9. 1984,
sphalm. “squmaeformis.” TYPE: China. Gan¬
su: Xiahe, 3300-3600 m, 14 Aug. 1980, 11. T.
Zhao s.n. (holotype, LZU).

In the protologue of Kobresia squamaeformis, the
species was considered similar to K. pygmaea C.
B. Clarke and K. bellardii (Allioni) Degland (= K.
myosuroides). I have studied the type specimen of
K. squamaeformis and have seen it in the field in
Zeku, Qinghai. It is most closely related to K.
setschwanensis. Although the taxon is distinguish¬
able from K. setschwanensis on the basis of basal
sheaths, glumes, prophylls, and nuts, it can be dis¬
tinguished by being shorter (7—16 cm) and having
unbranched inflorescences 8—12 mm long and uni¬
sexual or sometimes bisexual lateral spikelets with
a basal female flower and a distal male flower. In
contrast, K. setschwanensis is taller (10^10 cm high)
and often has basally branched inflorescences 20—
25 mm long and bisexual spikelets with a basal
female flower and (1)2-6 distal male flowers. Sub¬
species squamaeformis is distributed in eastern
Qinghai and adjacent Gansu, while subspecies
setschwanensis occurs in southern Qinghai, Sich¬
uan, Yunnan, and Xizang (Tibet).

Acknowledgment. I thank Ihsan Al-Shehbaz for
his help with the manuscript.

Novon 9: 453. 1999.

Cypripedium taibaiense (Orchidaceae), a New Species from

Shaanxi, China

Guanghua Zhu

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Shingchi Chen (Xinqi Chen)

Laboratory of Systematic and Evolutionary Botany, Institute of Botany,

Chinese Academy of Sciences, Nanxinchun 20, Xiangshan, Beijing 100093, P. R. China

Abstract. Cypripedium taibaiense G. Zhu & S.
C. Chen, a new species from Taibai Mountain in
Shaanxi, northwestern China, is described and il¬
lustrated. The species belongs to section Cypripe¬
dium subsection Macrantha (Kranzlin) Cribb. It re¬
sembles C. yunnanense Franchet in having small
flowers hut differs from it in having ovaries pubes¬
cent at least along the ribs, oblong staminodes al¬
most twice as long as wide, untwisted petals, and
in having dorsal sepals elliptic-lanceolate more
than twice as long as wide.

During a recent trip of the senior author to China
in 1998, several plants of Cypripedium, which were
said to have been collected from the provinces of
Gansu and Shaanxi in northwestern China, were
bought from a floral market in Beijing. Among the
already withered and poorly preserved flowers, one
was still distinguishable and did not seem to be
any known species of the genus. This plant was
brought to the Institute of Botany, Chinese Acade¬
my of Sciences, for study. It does not match any
species already described but is identical with sev¬
eral unidentified specimens in the PE herbarium,
which were collected from Taibai Mountain in
Shaanxi, China. The features of single purple flow¬
er, villose ovaries, alternate ovate leaves, and stam-
inode nearly sessile indicate that these specimens
belong to the section Cypripedium subsection Ma¬
crantha (Kranzlin) Cribb. The specimens were com¬
pared with species of wide distribution or those
found in neighboring provinces in the subsection
Macrantha, namely, C. franchetii Rolfe (Gansu,
Shanxi, Shaanxi, Henan, Hubei, and Sichuan), C.
micranthum Franchet (Heilongjiang, Jilin, Liaon¬
ing, Inner Mongolia, Hebei, Shandong, and Tai¬
wan), C. smithii Schlechter (Sichuan and Yunnan),
C. tibeticum King ex Hemsl. (Gansu, Sichuan,
Guizhou, Yunnan, and Tibet), and C. Yunnanense
(Sichuan, Yunnan, and Tibet). The comparison in¬
dicated that the specimens belong to a new member

of the subsection. This undescribed species is su¬
perficially similar to C. franchetii, but differs in
having much smaller flowers and less pubescent
ovaries. It resembles C. yunnanense Franchet in its
small flowers, but differs in having ovaries with pu¬
bescence at least along the ribs, oblong staminodes
almost twice as long as broad, untwisted petals, and
in having dorsal sepals elliptic-lanceolate more
than twice as long as wide.

Cypripedium taibaiense G. Zhu & S. C. Chen,
sp. nov. TYPE: China. Shaanxi: Taibai County,
Taibai Mountain, 3300 m, 4 July 1959, J. X.
Yang 629 (holotype, PE). Figure 1.

Inter species subsectionis Maeranthorum flore parvo ad
Cypripedium yunnanense accedens sed ovario pubescenti,
petalis non tortis, staminodio oblongo 1-1.1 cm longo, 5-
6 mm lato subduplo longiore quam latiore, sepalo dorsali
longiore quam latiore.

Terrestrial herb, 13-15(-24) cm tall, with stout
rhizome 4-5 cm long and 4-5 mm thick. Stem gla¬
brous, bearing 2—3 sheaths at base and 3^1 leaves
well spaced above pointing upward. Leaves elliptic
or elliptic-lanceolate, 4.5-11 cm long, 2.8-3. 5 cm
wide, acuminate or subacute, sheathing and am-
plexicaul at base, glabrous above, minutely pubes¬
cent or glabrescent beneath, ciliate. Inflorescence
terminal, 1-flowered; pedunc le 3— 7(— 10) cm long,
minutely pubescent; bract narrowly elliptic or nar¬
rowly elliptic-lanceolate, 6-6.5 cm long, 1.8-2. 2
cm broad, acuminate, minutely pubescent on both
sides or glabrescent above, ciliate on margins. Ped¬
icel and ovary 1.7—2 cm long, pubescent at least
along the ribs; flower 4—4.5 cm across, purple-red;
dorsal sepal elliptic-ovate, 2.2-3 cm long, 1.3-1. 5
cm wide, acuminate, glabrous; synsepal ovate-ellip¬
tic to narrowly elliptic, 2.2-2.8 cm long, 1-1.2 cm
wide, bifid at apex, glabrous; petals lanceolate,
2.6-3 cm long, 7-9 mm wide, villose in basal half
ol adaxial surface; lip deeply pouched, obovate-
subglobose, 2.5-3 cm long, 1.5-2 cm broad, pu-

Novon 9: 454-456. 1999.

Volume 9, Number 3
1999

Zhu & Chen

Cypripedium taibaiense from China

455

456

Novon

bescent inside, glabrous outside, with mouth ea.
1.2— 1.4 cin across; staminode oblong, 1—1.1 cm
long, 5-6 cm wide, longitudinally grooved along the
midrib, carinate beneath, mucronate at apex; fila¬
ments linear, 3 mm long. Flowering from June to
July.

Cypripedium is mairdy a temperate genus, con¬
sisting of about 45 species (Cribb, 1997) distrib¬
uted largely in eastern Asia and North America.
Over two-thirds of the species, about 32, are found
in China (Chen & Xi, 1987; Chen & Tsi, 1998).
Many species in the genus are of horticultural im¬
portance, attracting worldwide attention due to their
large and showy flowers. Some are also of phyto¬
geographic significance, several pairs of vicarious
species being disjunctly distributed in eastern Asia
and eastern North America (Chen, 1983a). Al¬
though this interesting genus was recently revised
by Cribb (1997), field studies are still needed, par¬
ticularly on the poorly known Chinese species.
Among the currently recognized species from Chi¬
na, many, such as C. wardii Rolfe, C. farreri W. W.
Smith, C. micranthum , C. fasciolatum Franehet, C.
palangshanense T. Tang & F. T. Wang, C. yunna-
nense, and C. smithii, are known from only a few
or single collections. A number of new species have
been described from China in recent years, namely
C. shanxiense S. C. Chen (Chen. 1983b), C. wum-
engense S. C. Chen (Chen, 1985), C. subtropicum
S. C. Chen & K. Y. Lang (Chen & Lang, 1986), C.
forrestii Cribb (Cribb, 1992), C. lichiangense S. C.
Chen & Cribb (Cribb & Chen, 1994), ami C. lud-
lowii Cribb (Cribb, 1997). The new species is the
only member of the genus currently known from
Shaanxi Province in China. As the remote regions

of Shaanxi and neighboring Gansu Province contain
many suitable habitats for Cypripedium species,
such as mixed deciduous forest and mossy and
grassy slopes and bogs, it is likely that more un¬
described species will be discovered if more her¬
barium material from those regions becomes avail¬
able.

Paratypes. CHINA. Shaanxi: Taibai County, Taibai
Mountain, grassy slopes, 2600— 3300 m, S. F. Li 9100 9
(MO. WUK), T. P. Wang 1698 (PE, WUK), J. X. Yang 581
(PE. WUK), J. X. Yang 885 (PE, WUK).

Acknowledgment. We are grateful to the cura¬
tors of MO, PE, and WUK for making the speci¬
mens available for this study.

I .iterature Cited

Chen, S. C. 1983a. Two pairs of vicarious species of Cyp-
ripedium (Orchidaceae) from eastern Asia and eastern
North America. Brittonia 35: 159—163.

- . 19831). New laxa of Orchidaceae from China.

Acta Phytotax. Sin. 21(3): 343-347.

- . 1985. Cypri/>edium wumengense and its allies.

Acta Phytotax. Sin. 23(5): 369—375.

- & K. Y. Ling. 1986. Cypripedium subtropicum, a

new species related to Selenipedium. Acta Phytotax.
Sin. 24(4): 317-322.

- & Y. /. Xi. 1987. Chinese cypripediums, with a

discussion on the classification of the genus. Proe. 12th
World Orchid Conf. 141-141).

- & Z. H. Tsi. 1998. Pp. 103—116 in The Orchids

of China. China forestry Press, Beijing.

Cribb, P. J. 1992. The Chinese spotted-leaved cypripedi¬
ums. Bull. Alpine Card. Soe. (it. Brit. 60(2): 165-177.

- . 1997. The Genus Cypripedium. Timber Press,

Portland, Oregon.

- & S. C. Chen. 1994. Further thoughts on the Chi¬
nese spotted-leaved cypripediums. Orchid Bev. 102:
320-323.

Volume 9, Number 3, pp. 289^156 of NOVON was published on 15 September 1999.

Volume 9
Number 4
1999

NOVON

New Combinations in Packera (Asteraceae) and Lectotypification

of Packera ovina

John F. Bain

Department of Biologieal Sciences, University of Lethbridge, Lethbridge,

Alberta, Canada T1K 3M4

ABSTRACT. Four new combinations are made for
Packera-. P. eurycephala var. lewisrosei (J. T. Howell)
Bain, P. contermina (Greenman) Bain, P. mores-
biensis (Calder & Taylor) Bain, and P. ovina
(Greene) Bain. Packera ovina is lectotypfied. The
latter is necessary because the holotype is a mixed
collection whose disposition affects the naming of
Packera contermina.

The segregate genus Packera , including those
species of Senecio previously referred to as mem¬
bers of the aureoid complex of Senecio, will be rec¬
ognized in the upcoming treatment of the Astera¬
ceae for Flora of North America (North of Mexico).
Most of the needed combinations have been made
by previous authors (Freeman & Barkley, 1995;
Love & Love, 1976; Jeffrey, 1992; Weber & Love,
1981), but recent work involving some arctic and
Rocky Mountain species of this complex has
brought to my attention the necessity of making a
few new combinations immediately. They are pre¬
sented below. In addition, I discuss the type ma¬
terial of Senecio ovinus because it has possible im¬
plications for naming Senecio conterminus.

Packera eurycephala var. lewisrosei (J. T. How¬
ell) Bain, comb. nov. Basionym: Senecio lew¬
isrosei J. T. Howell, Leafl. West. Bot. 3: 141.
1942. Senecio eurycephalus var. lewisrosei (J.
T. Howell) T. M. Barkley, Brittonia 20: 283.
1968. TYPE: U.S.A. California: Butte Co.,
near Pulga, L. Rose 39247 (holotype, CAS).

Packera contermina (Greenman) Bain, comb,
nov. Basionym: Senecio conterminus Green¬
man, Ann. Missouri Bot. Gard. 3: 101. 1916.
Based on Senecio Lyallii Klatt, Annal. Naturh-
ist. Hofm. Wien 9: 365. 1894, not Hooker f.
TYPE: [Canada.] Summit of the Rocky moun¬
tains, D. Lyall s.n. (GH).

Packera moresbiensis (Calder & Taylor) Bain,
comb. nov. Basionym: Senecio cymbal arioides
Buek subsp. moresbiensis Calder & Taylor,
Canad. J. Bot. 43: 1399. 1965. Senecio mores¬
biensis (Calder & Taylor) Douglas & Douglas,
Canad. J. Bot. 56: 1710. 1978. Senecio strep-
tanthifolius var. moresbiensis (Calder & Taylor)
Boivin, Naturaliste Canadien 93: 1031. 1966.
TYPE: [Canada.] British Columbia: Queen
Charlotte Islands, Moresby Island, J. Calder ;
D. Savile & R. Taylor 23006 (holotype, DAO).

Packera ovina (Greene) Bain, comb. nov. Basio¬
nym: Senecio ovinus Greene, Pittonia 4: 110.
1900. TYPE: [Canada.] Alberta: Sheep [=
Sofa] Mountain, J. Macoun 11619 (lectotype,
selected here, NDG in part, glabrous individ¬
uals). Figure 1.

Type Material of Senecio ovinus

Senecio ovinus Greene was described in 1900
from collections made from the Sheep Mountain re¬
gion of southern Alberta ( Macoun 11619). The lo¬
cality appears on modern maps as Sofa Mountain,
situated on the eastern boundary of Waterton Lakes

Novon 9: 457-459. 1999.

458

Novon

National Park. Senecio ovinus was recognized as
distinct from S. cymbal arioides Buek (= S. subnu-
dus DC.) by Greenman (1916), primarily based on
its “thickish” rather than thin leaves. The name
was considered a synonym of S. cymbalaria Pursh
(= S. resedifolius Lessing) in most recent treatments

including those by Barkley (1962, 1978, 1981),
Cronquist (1955, 1960), Hulten (1950), and Kowal
(1975), all of whom recognized S. cymbalaria as
distinct from S. cymbalarioides Buek and described
both as occurring in the southern Rocky Mountains.
Packer (1972) also recognized two taxa in the

Volume 9, Number 4
1999

Bain

Packera (Asteraceae)

459

southern Rocky Mountains, hut considered them to
he S. cymbalnrioides Buck and S. conterminus
Greenman. He did not, however, clearly indicate
the relationship ol S. ovinus to the latter two taxa.
Recent treatments following Packer (1972) include
those by Moss (1983) and Douglas (1982). The lat¬
ter author treated S. ovinus as a synonym of S. cym-
balarioides Buek. Thus, all modern treatments rec¬
ognize two taxa in southern Alberta, and treat 5.
ovinus as a synonym of either S. cymbalarioides or
S. conterminus/ S. cymbalaria. The latter taxon is
treated as S. conterminus in the present discussion.

An examination ol the holotype and isotypes of
S. ovinus has revealed that the type collection (Ma-
coun 1161 9) contains individuals of both S. conter¬
minus and S. cymbalarioides. The isotypes at NY
and US are referable to S. conterminus, while those
at CAN and GH are S. cymbalarioides. The holo¬
type sheet at NDG contains individuals of both S.
cymbalarioides and S. conterminus.

According to Article 9.10 of the Code (Greuter
et al., 1994), when a type herbarium sheet contains
parts belonging to more than one taxon, the name
must remain attached to that part that corresponds
most nearly with the original description. In addi¬
tion, Recommendation 9A.5 indicates that the lec-
totype should be chosen to preserve current usage.
A comparison of Greene’s original description with
the holotype sheet reveals that it fits all of the in¬
dividuals relatively well. The presence of floccose
tomentum, at least in the leal axils, is important
inasmuch as it is characteristic of S. conterminus ,
but even with regard to this character, Greene de¬
scribed the petioles of S. ovinus as “floceulent, at
least when young,” implying that not all of the in¬
dividuals he observed had floceulent petioles. Since
the petioles of individuals of S. cymbalarioides are
glabrous, this is not surprising. It should be noted
that Greenman did not indicate having seen the
holotype of S. ovinus when he recognized it as dis¬
tinct in his 1916 monograph. He relied on the type
material from GH and CAN, which, as previously
noted, is comprised entirely of glabrous individuals,
here treated as S. cymbalarioides. His presumably
later annotation of the holotype indicates clearly
that he, too, could see that the specimen is com¬
posed of two elements, which he referred to as S.
conterminus and S. ovinus.

I lectotypify S. ovinus Greene by designating the
glabrous individuals on the sheet as the type ma¬
terial, thus treating it in the same manner as Green¬
man, whose annotation appears on the sheet. The
elements selected are consistent with the original
description and preserve current usage by allowing

the name to remain in synonymy, albeit with S.
cymbalarioides not S. cymbalaria. Were the floccose
individuals (S. conterminus ) to have been chosen as
lectotype, then the name S. conterminus Greenman
would become a synonym of S. ovinus Greene, ne¬
cessitating a name change for those who, like Pack¬
er (1972), recognize two taxa in southern Alberta
and surrounding areas, neither of which is 5. cym¬
balaria.

Acknowledgments. I thank the herbarium cu¬
rators for providing loans of Senecio and Natural
Sciences and Engineering Research Council of
Canada for funding support. The work has benefited
from numerous discussions with T. M. Barkley, J.
L. Golden, and J. G. Packer.

Literature Cited

Barkley, T. M. 1962. A revision of Senecio aureus I. inn.
and allied species. Trans. Kansas Acad. Sci. 65: 318—
408.

- . 1978. Senecio. Pp. 50—139 in C. T. Rogerson,

North American Flora II pt. 10. New York Botanical
Garden, Bronx.

- . 1981. Senecio and Erechtites in North America:

Supplementary notes. Britton ia 33: 523—527.

Cronquist, A. 1955. Vascular Plants of the Pacific North¬
west, Part 5. Univ. Washington Press, Seattle.

- . 1960. Senecioneae. Pp. 416—465 in B. T. Ferris,

Illustrated Flora of the Pacific States, Vol. 4, Bignoni-
aceae to Compositae. Stanford Univ. Press, Stanford.
Douglas, G. W. 1982. The Sunflower Family (Asteraceae)
of British Columbia. Volume 1 — Senecioneae. Occas.
Pap. British Columbia Provincial Museum #23. British
Columbia Provincial Museum, Victoria.

Freeman, C. C. & T. M. Barkley. 1995. A synopsis of the
genus Packera (Asteraceae: Senecioneae) in Mexico.
Sida 16: 699-709.

Greenman, J. M. 1916. Monograph ol the North American
and Central American species of the genus Senecio, Part
II (Continued). Ann. Missouri Bot. Card. 3: 85—164.
Greuter, W., F. B. Barrie, H. M. Burdet, W. G. Chaloner,
V. Demoulin, I). L. Hawksworth. P. M. Jprgensen, 1). 11.
Nicolson, P C. Silva, P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Hulten, E. 1950. Flora of Alaska and Yukon. X. Acta
llniv. Lund 2. 41: 1485-1902.

Jeffrey, C. 1992. The tribe Senecioneae (Compositae) in
the Mascarene Islands with an annotated world check¬
list of the genera of the tribe. Notes on Compositae: VI.
Kew Bull. 47: 49-109.

Kowal, R. L. 1975: Systematics of Senecio aureus and al¬
lied species on the Gaspe Peninsula, Quebec. Mem.
Torrey Bot. Club 23: 1—113.

Love, A. & D. Ixive. 1976. Nomenelatural notes on arctic
plants. Bot. Not. 128: 497-523.

Moss, E. H. 1983. Flora of Alberta, 2nd Ed. Univ. Toronto
Press, Toronto.

Packer, J. G. 1972. A taxonomic and phytogeographical
review of some arctic and alpine Senecio species. Can-
ad. .1. Bot. 50: 507-518.

Weber, W. A. & A. Love. 1981. New combinations in the
genus Packera (Asteraceae). Phytologia 49: 44—50.

New Combinations in Californian Arnica and
Monolopia (Compositae)

Bruce G. Baldwin

Jepson Herbarium and Department of Integrative Biology, University of California, Berkeley,

California 94720-2465, U.S.A.

ABSTRACT. New combinations in Arnica L. and
Monolopia DC. are made in light of evidence from
a phylogenetic study of helenioid Heliantheae that
each of the two genera is paraphyletic. Transfers
herein of Whitneya dealbata A. Gray to Arnica and
Lembertia congdonii (A. Gray) Greene to Monolopia
allow for circumscriptions of Arnica and Monolopia
that better reflect phylogenetic relationships.

Results of a phylogenetic study of helenioid He¬
liantheae or Helenieae sensu lato (Baldwin & Wes¬
sa, 1997, & unpublished) lead me to conclude that
Arnica L. and Monolopia DC. are each paraphylet¬
ic. The following combinations are needed to make
Arnica and Monolopia monophyletic.

Arnica dealbata (A. Gray) B. G. Baldwin, comb,
nov. Basionym: Whitneya dealbata A. Gray,
Proc. Amer. Acad. Arts 6: 550. 1865. TYPE:
U.S.A. California: “dry woods on the Yosemite
trail, alt. 7,000 feet,” 1863, W. H. Brewer 1627
(holotype, GH).

In his description of Whitneya , Asa Gray (1865)
noted morphological similarities to Arnica , but ab¬
sence of pappus in Whitneya made its relationships
difficult to assess. Hoffmann (1890) placed Whit¬
neya with Baileya Harvey & A. Gray ex Torrey and
Psilostrophe DC. in subtribe Riddeliinae on the ba¬
sis of shared persistence of the ray corollas, a treat¬
ment followed by later workers such as Rydberg
(1914) and Keck (1959). Omduff et al. (1967) con¬
cluded from cytologieal and morphological consid¬
erations that Whitneya is closely related to Arnica,
a view subsequently endorsed by Turner and Powell
(1977) and Nordenstam (1977) and implicit in Rob¬
insons (1981) and Karis and Ryding’s (1994)
placement of Arnica and Whitneya together in
Chaenactidinae. Results from phylogenetic analys¬
es of nuclear ribosomal DNA sequences provide
strong support for a nested position of Whitneya
within a group corresponding to Arnica sensu Ma¬
guire (1943) plus A. mallotopus (Franchet & Sa-
vatier) Makino [= Mallotopus japonicus Franchet &
Savatier].

Like other species of Arnica, A. dealbata is a
rhizomatous perennial with opposite leaves and
each ray floret associated with a phyllary. The chro¬
mosome number reported for A. dealbata is 2n =
19 II, corresponding to the diploid condition in Ar¬
nica (see Wolf, 1980, 1987).

Monolopia congdonii (A. Gray) B. G. Baldwin,
comb. nov. Basionym: Eatonella congdonii A.
Gray, Proc. Amer. Acad. Arts 19: 20. 1883.
Lembertia congdonii (A. Gray) Greene, FI.
Fran. 441. 1897. SYNTYPES: U.S.A. Califor¬
nia: Tulare County, “Deer Creek,” Mar. 1881,
J. W. Congdon 439 (GH); “southern part of the
San Joaquin Valley,” Mar. 1881, C. C. Parry
s.n. (GH). [Rydberg (1914) cited the Congdon
collection as type.]

Greene’s removal from Eatonella A. Gray of E.
congdonii (transferred to a new genus, Lembertia
Greene) is upheld by phylogenetic evidence that (1)
E. nivea (D. C. Eaton) A. Gray, the type species of
Eatonella, is the sister group of Hulsea Torrey &
A. Gray; and (2) Lembertia and Monolopia consti¬
tute a elade (Baldwin & Wessa, 1997, & unpub¬
lished). Johnson (1978) concluded from morpholog¬
ical, chromosomal, and distributional studies that
Lembertia and Monolopia constitute a “monolopioid
line” distinct from an “eriophylloid line” compris¬
ing Eriophyllum, Pseudobahia, and Syntrichopap-
pus. Results from phylogenetic analyses of nuclear
ribosomal DNA sequences (Baldwin & Wessa, un¬
published) confirm Johnson’s (1978) hypothesis and
provide robust support for nested placement of
Lembertia within Monolopia sensu Johnson (1978).

Like other species of Monolopia and unlike the
closely related genera Eriophyllum, Pseudobahia,
and Syntrichopappus, members of M. congdonii
possess a small lobe opposite the ray-corolla lam¬
ina, black phyllary hairs, unequal hair cells on the
distal surfaces of the eypselae, and a chromosome
number that falls within the range of 2 n — 10 to
13 II (i.e., 2 n = 11 II) (Johnson, 1978). Monolopia
congdonii and two of the four remaining species of
Monolopia, i.e., M. lanceolata Nuttall and M. stricta

Novon 9: 460-461. 1999.

Volume 9, Number 4
1999

Baldwin

Combinations in Arnica & Monolopia

461

Crum, have broadly overlapping distributions in the
southern San Joaquin Valley and adjacent valleys
of the Inner South Coast Ranges of California
(Johnson, 1978).

Acknowledgments. I thank Ted Barkley, Dave
Keil, and John Strother for their helpful reviews of
the manuscript. I also thank the curator of GH for
specimen loans. This paper is based on research
supported by the National Science Foundation
(DEB-9458237), the Lawrence R. Heckard Endow¬
ment Fund, and Roderick Park and other generous
Friends of the Jepson Herbarium.

Literature Cited

Baldwin. B. G. & B. L. Wessa. 1997. Origin of Madiinae
and Dimeresici from within Chaenactidinae (Asteraceae).
Amer. J. Bot. 84 (6, supplement): 176—177.

(fray, A. 1865. Whitneya , nov. gen. Proe. Amer. Acad.
Arts 6: 549-550.

Hoffmann, (). 1890. Riddellinae. In Engler & Prantl. Die
natUrlichen Pflanzenfamilien I V (5): 253.

Johnson, I). E. 1978. Systematics of Eriophyllinae (Com-
positae). Ph.D. Dissertation, University of California,
Berkeley.

karis, P. (). & (). Ryding. 1994. Tribe Helenieae. Pp.

521—558 in K. Bremer, Asteraceae: Cladistics and Clas¬
sification. Timber Press, Portland, Oregon.

keck. I). I). 1959. Helenieae. Pp. 1129-1161 in P. A.
Munz, A California Flora. Univ. California Press, Berke¬
ley.

Maguire, B. 1943. A monograph of the genus Arnica. Brit-
tonia 4: 386—510.

Nordenstam, B. 1977. Senecioneae and Liabeae — System¬
atic review. Pp. 799—830 in V. H. Heywood, J. B. liar-
borne & B. L. Turner (editors), The Biology and Chem¬
istry of the Compositae, Vol. 11. Academic Press,
London.

Omduff, R. 0., T. Mosquin. I). W. kvhos & P. H. Raven.
1967. Chromosome numbers in Compositae. VI. Sene¬
cioneae. II. Amer. J. Bot. 54: 205—213.

Robinson, H. 1981. A revision of the tribal and subtribal
limits of the Heliantheae (Asteraceae). Smithsonian
Contr. Bot. 51: 1—102.

Rydberg, P. A. 1914. Riddellianae. North American Flora
34: 6-11.

Turner, B. E. & A. M. Powell. 1977. Helenieae — System¬
atic review. Pp. 699—737 in V. H. Heywood, J. B. Har-
borne & B. L. Turner (editors), The Biology and Chem¬
istry of the Compositae, Vol. II. Academic Press,
London.

Wolf, S. .). 1980. Cytogeographical studies in the genus
Arnica (Compositae: Senecioneae). I. Amer. J. Bot. 67:
300-308.

- . 1987. Cytotaxonomic studies in the genus Arnica

(Compositae: Senecioneae). Rhodora 89: 391—400.

New Combinations and New Genera in the North American
Tarweeds (Compositae-Madiinae)

Bruce G. Baldwin

Jepson Herbarium and Department of Integrative Biology, University of California, Berkeley,

California 94720-2465, U.S.A.

Abstract. New combinations anil new genera tor
species in Hemizonia DC., Madia Molina, anil Rail¬
lardiopsis Rydberg allow tor a revised, practical
taxonomy of Madiinae that better reflects phyloge¬
netic relationships. Reinstatement of Anisocarpus
Nuttall, Centromadia Greene, Deinandra Greene,
and Hemizonella A. Gray, abandonment of Raillar-
diopsis Rydberg, and recognition of tour new genera
{Carlquistia, Harmonia, Jensia, Kyhosia ) for tar-
weed lineages without available generic names re¬
sult from changes in circumscription based on
monophyletic groups.

Results of phylogenetic, biosystematie, and cy¬
togenetic studies in Madiinae (Kyhos et al., 1990;
Baldwin et ah, 1991; Baldwin, 1992, 1993, 1996,
unpublished; Carr et ah, 1996) indicate that Hem¬
izonia DC. sensu Keck (1959), Madia Molina sensu
Keck (1959), and Raillardiopsis Rydberg are not
monophyletic. Hemizonia sensu Keck (1959) com¬
prises a minimum of three monophyletic groups
that correspond to section Centromadia (Greene) D.
D. Keck, section Hemizonia , and section Madiom-
eris Nuttall sensu Tanowitz (1982) plus the informal
“Fruticosae” or “Zonamra” (cf. Clausen, 1951;
Keck, 1959). Madia sensu Keck (1959), Raillar¬
diopsis , and the monophyletic Hawaiian silversword
alliance (Argyroxiphium DC., Dubautia Gaudi-
chaud-Beaupr6, Wilkesia A. Gray; see Carr, 1985)
constitute a clade lacking a formal name (the “Ma¬
dia” lineage in Baldwin, 1996) that is extremely
heterogeneous and difficult to diagnose morpholog¬
ically. The Hawaiian group comprises trees, shrubs,
subshrubs, large rosette plants, mat plants, cushion
plants, and vines that are strikingly divergent in
morphological and ecological characteristics from
the primarily Californian herbs in Madia sensu
Keck (1959) and Raillardiopsis (see Carr, 1985;
Baldwin & Robichaux, 1995; Baldwin, 1997). In
addition to the distinctive Hawaiian lineage, a min¬
imum of seven monophyletic groups can be reliably
diagnosed from morphological and molecular data
within die clade comprising Madia sensu Keck
(1959), Raillardiopsis , and the silversword alliance.

Recognition at the generic level of each of the
three lineages in Hemizonia sensu Keck (1959) and
each of the seven lineages in the paraphyletic
group comprising Madia sensu Keck (1959) plus
Raillardiopsis in the traditional (only) sense results
in a practical taxonomy that conforms to the crite¬
rion of monophyly (= holophyly) of taxa. The re¬
vision of Madia sensu Keck (1959) plus Raillar¬
diopsis (hereafter, MADRAIL) also allows for
continued generic distinction of the unusual Ha¬
waiian species in the silversword alliance without
maintaining non-monophyletic California tarweed
genera. I realize that additional nomenelatural
changes in the silversword alliance are necessary
to achieve a taxonomy of the Hawaiian species that
includes only monophyletic genera. The new com¬
binations herein focus only on genera relevant to
treatments in preparation for the Flora of North
America North of Mexico.

Revision ok MADRAIL

In my revised taxonomy of MADRAIL, Madia is
restricted to an x = 8 lineage of radiate, epappose
annuals with compressed or 3-angled ray eypselae,
i.e., the informal section “Madia" sensu Keck
(1959), minus M. minima (A. Gray) D. D. Keck.
Madia minima , with obcompressed ray eypselae,
falls outside Madia, as circumscribed here, and is
returned to the monospecific genus Hemizonella A.
Gray. All of the widespread, ruderal taxa of Madia
sensu Keck (1959) fall within Madia in the new
sense. Madia remains the most widely distributed
genus of Madiinae, with species found across much
of western North America, in Chile and Argentina,
and (mostly naturalized) in other parts of the world.

Anisocarpus Nuttall is resurrected for a taxon
comprising both perennial (sister) species in MAD¬
RAIL with 2n = 7 II; M. madioides (Nuttall)
Greene (= A. madioides Nuttall, the type species
of Anisocarpus) and R. scabrida (Eastwood) Rydberg
(= A. scabridus (Eastwood) B. G. Baldwin), the type
species ol Raillardiopsis. Members of Anisocarpus
in the new sense are the only perennial herbs in
Madiinae that combine well-distributed eauline

Novon 9: 462^-71. 1999.

Volume 9, Number 4
1999

Baldwin

Combinations in Madiinae

463

leaves, radiate heads, ellipsoid or spheric involu¬
cres, and yellow anthers.

Anisocarpus is a western North American genus
found from southern California to southern British
Columbia. The two species constituting the genus
are allopatric and ecologically divergent. Anisocar¬
pus madioides is found in understories of evergreen
forests at low to mid elevations, primarily in the
coast ranges and on the outer coast of California,
Oregon, Washington, and British Columbia. In con¬
trast, A. scabridus is restricted to metamorphic
scree slopes at high elevations in the inner North
Coast Ranges and southern Cascade Range in Cal¬
ifornia.

Anisocarpus scabridus (Eastwood) B. G. Bald¬
win, comb. nov. Basionym: Raillardella sca-
brida Eastwood, Bull. Torrey Bot. Club 32:
216. 1905. Raillardiopsis scabrida (Eastwood)
Rydberg, N. Amer. FI. 34: 320. 1927. TYPE:
U.S.A. California: Snow Mountain, 25 Aug.
1892, K. Brandegee s.n. (holotype, CAS).

Four new genera are erected for the remaining
M ADR AIL lineages. Carlquistia, lor Raillardiopsis
muirii (A. Gray) Rydberg (2 n = 8 II), recognizes
Sherwin Carlquist for his contributions on system¬
atic anatomy in Madiinae, work that led him to de¬
termine that Raillardella A. Gray in the traditional
sense (including Raillardiopsis ) and the Hawaiian
silversword alliance are tarweeds. Harmonia , Jen-
sia , and Kyhosia recognize Harvey Monroe Hall,
Jens Clausen, and Donald W. Kyhos for their con¬
tributions to Madiinae biosystematies across three
successive generations of botanists. Harmonia cor¬
responds to the lineage of four species ol yellow-
anthered, pappose annuals with 2 n = 9 II. Jensia
encompasses two (sister) species of dark-anthered,
pappose annuals with 2 n = 8 II. Kyhosia is a
monospecific genus to which the montane perennial
Madia bolanderi (A. Gray) A. Gray (2n = 6 II) is
transferred.

Carlquistia B. G. Baldwin, gen. nov. TYPE: Rail¬
lardella muirii A. Gray, Bot. Calif. 1: 618.
1876. = Carlquistia muirii (A. Gray) B. G.
Baldwin.

A Madiinae ceteris differt eharacteribus conjuncte: ha-
bitu perenni. herbaceo; rhizomatibus lignosis; foliis cau-
linis. foliis proximalibus oppositis; capitulis discoidis;
squamis papporum subulatis, ciliato-plumosis; 2n = 8 II.

Perennials, clumped to mat-forming, white-hir¬
sute to villous and glandular; rhizomes woody,
loosely branched; aerial stems greenish yellow to
dark purplish, sometimes woody at base, ascending

to erect, leafy throughout, simple or branched, 7-
54 cm high; leaves sessile, opposite from base to
beyond middle of stem, alternate distally; blades
green to gray-green, linear to lanceolate, mostly 9—
42 mm long, entire, flat to strongly crisped-coiled,
the apices acute; heads discoid, borne singly or in
leafy, loosely corymbiforin or rounded eapitules-
cences; peduncles 0—8 cm long; “involucres” ellip-
tic-campanulate, campanulate, or turbinate, 8—13
mm high; “phyllaries” (5) 7 to 16, uniseriate, green
to purple, subequal; receptacles flat, setulose; “pa-
leae” none; florets 7 to 29, bisexual, the corollas
yellow, tubular, 6.5—10 mm long, constricted ba-
sally; anthers yellow; style branches hispidulous;
cypselae black, terete, linear, straight or slightly
curved at acute base, 4—7.5 mm long, covered with
white, antrorse hairs; pappus of 9 to 17, purplish
white to tawny, subulate scales 5—11 mm long, cil-
iate-plumose to tip; 2 n = 8 II.

Members of Carlquistia are the only rhizomatous
perennial herbs in Madiinae with both well-distrib¬
uted cauline leaves and discoid heads. As in other
members of Madiinae with discoid heads, true
phyllaries are absent — the bracts surrounding the
florets and functioning as an involucre are homol¬
ogous with receptacular bracts (Carlquist, 1959).

Distribution, ecology, and relationships. Carl¬
quistia is restricted to granitic exposures and gra¬
nitic alluvium of mid- to high-elevational montane
habitats of the southern Sierra Nevada and Ventana
Double Cone in the Santa Lucia Range of Califor¬
nia. Carlquistia appears to be most closely related
to Madia, which shares the common base chro¬
mosome number of x — 8 (Baldwin, 1996, & un¬
published). Members of the two genera are so dis¬
parate in habit, capitular morphology, and ecology
that treating them as congeners is impractical, i.e.,
would result in a group difficult to diagnose with
macroscopic characteristics.

Carlquistia muirii (A. Gray) B. G. Baldwin, comb,
nov. Basionym: Raillardella muirii A. Gray,
Bot. Calif. 1: 618. 1876. Raillardiopsis muirii
(A. Gray) Rydberg, N. Amer. FI. 34: 320.
1927. TYPE: U.S.A. California: Sierra Nevada,
“the station unknown,” J. Muir s.n. (holotype,
GH).

Harmonia B. G. Baldwin, gen. nov. TYPE: Madia
hallii D. D. Keck, Madrono 3: 5. 1935. = Har¬
monia hallii (D. D. Keck) B. G. Baldwin.

A Madiinae ceteris differt eharacteribus conjuncte: ha-
bitu annuo; capitulis radiatis; phyllariis omnibus ovario
radii perfecte cingentibus, phyllariis omnibus hirsutis ir-
regulariter, apicibus pilorum rectis; bracteis receptaculo-

464

Novon

rum in annulo i ntt*r flosculos radiorum et flosculos dis-
corum, discretis vel imitis infirme; antlieris luteis; cypselis
radiorum teretibus vel compressis; squamis papporum ±
subulatis, ciliato-plumosis; 2 n = 9 II.

Annuals; stems erect, often reddish purple, slen¬
der, to 2.7(-4) dm high, hirsute; leaves sessile,
mostly cauline, evenly distributed or congested at
bases of subumbelliform capitulescences, opposite
in proximal half or at base, alternate distally;
blades narrowly linear, entire or shallowly toothed,
often revolute, mostly hirsute, eglandular or mi¬
nutely stipitate-glandular, the glands black (or yel¬
lowish); capitulescences loosely corymbiform or
rounded to subumbelliform, or heads borne singly;
peduncles erect or reflexed, 1-60 mm long, mi¬
nutely stipitate-glandular, the glands black (or yel¬
lowish); involucres obconic to obovoid, 3—6 mm
high; phyllaries 3 to 8, uniseriate (1 per ray floret),
herbaceous, each completely enveloping a ray ova¬
ry, not bulging downward at base, the apices acute
to acuminate, <half the length of enfolded basal
portion of phyllaries, flat or revolute; abaxial faces
mostly sparsely to densely hirsute, often with
broadly arching or somewhat appressed hairs, often
sericeous to villous near margins, ciliate with soft
hairs interlocking the enfolded margins, eglandular
or minutely stipitate-glandular, the glands black (or
yellowish); receptacles Hat to slightly convex, gla¬
brous or sparsely setulose; paleae not persistent,
restricted to ring at periphery of receptacle (encir¬
cling the disc Horets), the margins free or weakly
fused; ray florets 3 to 8, pist i Hate, the corollas
bright yellow, the laminae flabelliform to obovate,
3-lobed; disc Horets 7 to 30, bisexual or function¬
ally staminate, the corollas bright yellow, 5-lobed,
glabrous or the lube sparsely villous; anthers yel¬
low; style branches elongate, hispidulose; ray cyp-
selae black, terete to compressed, faintly or prom¬
inently adaxially angled, clavate, kyphotic (i.e.,
strongly bowed out abaxially anil slightly concave
adaxially) or weakly arcuate, glabrous, beakless to
distinctly beaked, attachment scar basal; ray pap¬
pus none or of rudimentary broad or subulate
scales, ^1 mm long, fimbriate to plumose; disc
cypselae black, terete, clavate, straight or weakly
arcuate, glabrous or with antrorse hairs; disc pap¬
pus of 7 to 10, stramineous or purplish, lance-at¬
tenuate to subulate, ± Hat (not crisped), fimbriate
or ciliate-plumose scales 0.2—4 mm long; 2 n — 9
11.

Members of Harmonia are the only annuals in
Madiinae that combine yellow anthers, phyllaries
each completely surrounding a ray ovary, terete to
compressed ray cypselae, and pappose disc ovaries.

Unlike Jensia, Harmonia has phyllaries irregularly
hirsute and lacking uncinate hairs, receptacular
bracts free or weakly fused, anthers yellow, pappus
elements plane, and 2 n — 9 II.

Distribution, ecology, and relationships. Har¬
monia is restricted to soils derived from volcanic
ash or serpentine in the North Coast Ranges and
Klamath Ranges, California. The volcanic-ash en¬
demic H. nutans is apparently sister to the remain¬
ing three, serpentine-endemic species (Baldwin,
1993, & unpublished).

Harmonia doris-nilesiae (T. W. Nelson & J. P.
Nelson) B. G. Baldwin, comb. nov. Basionym:
Madia doris-nilesiae T. W. Nelson & J. P. Nel¬
son, Brittonia 37: 394. 198.3. TYPE: U.S.A.
California: Trinity County, along Shasta-Trinity
Forest road 1N40, 0.23 mi. E of junction with
road 1N35 (TIN, R7E, S2), 3600 ft., 14 June
1983, T. W. Nelson & B. Williams 7422 (ho-
lotype, HSC; isotypes, CAS, JEPS; also GH,
MO, NY, US, WTU, not seen).

Harmonia hallii (D. D. Keck) B. G. Baldwin,
comb. nov. Basionym: Madia hallii D. D.
Keck, Madrono 3: 5. 1935. TYPE: U.S.A. Cal¬
ifornia: Napa County, “Near Knoxville,” 27
May 1931, //. M. Hall 12094 (holotype, DS;
isotypes, UC; also B, Cl, GH, K, MO, NY, US,
not seen).

Harmonia nutans (Greene) B. G. Baldwin, comb,
nov. Basionym: Callichroa nutans Greene, Pit-
tonia 2: 227. 1892. Blepharipappus nutans
Greene, Pittonia 2: 247. 1892. Layla nutans
(Greene) Jepson, FI. W. Mid. Calif., ed. 2. 449.
1911. Madia nutans (Greene) D. D. Keck, Ma¬
drono 3: 5. 1935. TYPE: U.S.A. California:
Sonoma County, “Hood’s Peak” (on label), 15
May 1892, F. T. Bioletti s.n. (holotype, NDG).

Harmonia stebbinsii (T. W. Nelson & J. P. Nelson)
B. G. Baldwin, comb. nov. Basionym: Madia
stebbinsii T. W. Nelson & J. P. Nelson, Brittonia
32: 323. 1980. TYPE: U.S.A. California: Trin¬
ity County, along Shasta-Trinity National For¬
est road to Hell-to-Find Lake (T29N, R11W,
S27, NE1/4), 23 June 1978, T. W. Nelson & J.
P. Nelson 4227 (holotype, HSC; isotypes, CAS,
DAV, JEPS, UC; also GH, MO, NY, RSA, US,
WTU, not seen).

Jensia B. G. Baldwin, gen. nov. TYPE: Madia yo-
semitana Parry ex A. Gray, Proc. Amer. Acad.
Arts 17: 219. 1882. = Jensia yosemitana (Par¬
ry ex A. Gray) B. G. Baldwin.

Volume 9, Number 4
1999

Baldwin

Combinations in Madiinae

465

A Madiinae ceteris differt characteribus eonjunete: ha-
bitu annuo; capitulis radiatis; pliyllariis omnibus ovario
radii perfecte cingentibus, pliyllariis omnibus hirsutis ±
uniformiter, apicibus pilorum uncinatis; bracteis recepta-
culorum in annulo inter flosculos radiorum et flosculos
discorum, unitis valde; antheris purpureis; cypselis ra-
diorum compressis; squamis papporum subulatis, crispis
irregulariter, ciliatis; 2 n = 8 II.

Annuals; stems erect, stramineous to reddish
purple, usually slender, to 6 dm tall, hirsute;
leaves sessile, mostly cauline, evenly distributed
or congested at bases of subumbelliform capitu-
lescenses, opposite in proximal half or at base,
alternate distally; blades linear to narrowly spat-
ulate, entire or inconspicuously toothed, plane or
slightly revolute, strigose or hirsute, eglandular or
the distalmost minutely stipitate-glandular, the
glands yellowish or reddish (or black); capitulesc-
ences loosely corymbiform or rounded to subum¬
belliform, or heads borne singly; peduncles erect,
to at least 1 dm long, hirsute or not, eglandular or
minutely stipitate-glandular, the glands yellowish
or reddish or rarely black; involucres obconic to
spheric, 3-5 mm high; phyllaries 2 to 10 in 1
series (1 per ray floret), herbaceous, each com¬
pletely enveloping a ray ovary, usually bulging
slightly downward at base, the apices acuminate,
revolute, abaxial faces hirsute (the stout hairs usu¬
ally uncinate at tips, often from pustules), usually
eglandular, minutely farinose near margins, ciliate
with soft hairs interlocking the enfolded margins;
receptacles flat to slightly convex, glabrous or
sparsely setulose; paleae not persistent, restricted
to ring at periphery of receptacle (encircling the
disc florets), ± strongly fused most of length; ray
florets 2 to 10, pistillate, the corollas bright yel¬
low, often with purplish lines on abaxial surface
of laminae, the laminae obovate, 3-lobed; disc flo¬
rets 1 to 65, functionally staminate, corollas bright
yellow, 5-lobed, glabrous or nearly so; anthers
dark purple; style branches elongate, hispidulose;
cypselae black, compressed, adaxially angled, cla-
vate, kyphotic, glabrous, distinctly beaked, attach¬
ment scar slightly oblique; ray pappus a minute
crown of subulate scales, to 1 mm long; disc ova¬
ries glabrous; disc pappus of 5 to 7, white or pur¬
plish tipped, subulate, crisped, ciliate scales 2.5-
4 mm long; 2n — 8 II.

Members of Jensia are the only annuals in Ma¬
diinae that combine phyllaries each completely
surrounding a ray ovary, dark purple anthers, terete
to compressed ray cypselae, and pappose disc ova¬
ries. Unlike Harmonia , Jensia has phyllaries more
or less uniformly hirsute with hairs usually unci¬
nate, receptacular bracts strongly united, anthers

dark purple, pappus elements crisped, and 2n = 8

II.

Distribution and ecology. Jensia is found from
foothill woodlands to mid-montane meadows in the
western Sierra Nevada, California. The showy,
self -incompatible J. rammii and the diminutive,
self- compatible J. yosemitana are allopatric sister-
species. Jensia rammii occurs in the Sierra foot¬
hills; J. yosemitana occurs in montane Sierran
habitats.

Jensia rammii (Greene) B. G. Baldwin, comb. nov.
Basionym: Madia rammii Greene. Bull. Calif.
Acad. Sci. 1: 90. 1885. Anisocarpus rammii
(Greene) Greene, FI. Francisc. 415. 1897.
TYPE; U.S.A. California: Nevada City, July
1884, C. A. Ramm s.n. (holotype, CAS).

Jensia yosemitana (Parry ex A. Gray) B. G. Bald¬
win, comb. nov. Basionym: Madia yosemitana
Parry ex A. Gray, Proc. Amer. Acad. Arts 17:
219. 1882. Anisocarpus yosemitana (Parry ex
A. Gray) Greene, FI. Francisc. 416. 1897.
TYPE: U.S.A. California: “foot of the Upper
Yosemite Fall,” June 1881, C. C. Parry s.n .
(holotype, GH; isotype, DS).

Kyhosia B. G. Baldwin, gen. nov. TYPE: Aniso¬
carpus bolanderi A. Gray, Proc. Amer. Acad.
Arts 7: 360. 1868. = Kyhosia bolanderi (A.
Gray) B. G. Baldwin.

A Madiinae ceteris differt characteribus eonjunete: ha-
bitu perenni, rhizomatoso, herbaeeo; capitulis radiatis;
pliyllariis omnibus ovario radii ± cingentibus; bracteis re-
ceptaculorum in annulo inter flosculos radiorum et flos¬
culos discorum; antheris atropurpureis; squamis pappo¬
rum ± subulatis, ciliatis; 2 n = 6 II.

Herbs, perennial, with acrid odor; stems erect,
stramineous to reddish purple, usually robust, to 12
dm tall, sparsely to densely hirsute proximally, stip¬
itate-glandular distally, the glands yellowish brown
to black; leaves sessile, basal and cauline, con¬
gested near base, opposite from base to beyond
middle of stem, alternate distally; blades linear or
lance-linear, to 35 cm long, to 1.5 cm wide, the
distal shorter and more widely spaced than on prox¬
imal stem, entire, sparsely to densely hirsute, the
distalmost stipitate-glandular; capitulescences
loosely corymbiform or rounded, usually of 8 or
fewer heads; peduncles erect, to 15( — 25) cm long,
stipitate-glandular, the glands yellowish brown to
black; involucres campanulate to hemispheric, 7—
14 mm high; phyllaries 8 to 12, uniseriate (1 per
ray floret), herbaceous, each ± completely envel¬
oping a ray ovary, ciliate, often ± hirsute, coarsely

466

No von

stipitate-glandular; phyllary apices acuminate, up
to half the length of enfolded basal portion, revo¬
lute; receptacles convex, glabrous or sparsely se-
tulose; paleae not persistent, mostly free or weakly
united, restricted to ring encircling the disc florets
at periphery of receptacle; ray florets 8 to 12, pis¬
tillate, the corollas bright yellow, the laminae con¬
spicuous, oblanceolate to obovate, deeply 3-lobed,
the tubes hirsute and minutely glandular; disc flo¬
rets 30 to 65, bisexual, the corollas bright yellow,
5-lobed, sparsely hirsute and minutely glandular;
anthers dark purple; style branches elongate, his-
pidulose; ray cypselae brown to black, ± terete,
weakly 5-ridged (with adaxial angle evident), cla-
vate, arcuate, about 5-7 mm long, glabrous or
sparsely hispidulose, constricted apically, attach¬
ment scar basal; ray pappus absent or a minute
crown of subulate scales; disc cypselae brown to
black, terete, weakly 5-ridged, clavate, straight or
arcuate proximally, ca. 5—9 mm long, brownish his¬
pid (the hairs ascending); disc pappus of 5 to 10,
stramineous or purplish, broadly lanceolate to sub¬
ulate, ciliate to ciliate-plumose scales ca. 2-5 mm
long; 2n = 6 II.

Members of Kyhosia are the only rhizomatous pe¬
rennial herbs in Madiinae that combine radiate
heads, paleae restricted to a ring at the periphery
of the receptacle, bisexual disc florets, and dark
purple anthers.

Distribution and ecology. Kyhosia is endemic to
the California Floristic Province, where it occurs in
mid-montane wet meadows and on streambanks of
the western Sierra Nevada and in the southern Cas¬
cade Range and Klamath Ranges of California and
southern Oregon.

Kyhosia bolanderi (A. Gray) B. G. Baldwin,
comb. nov. Basionym: Anisocarpus bolanderi
A. Gray, Proc. Amer. Acad. Arts 7: 360.
1868. Madia bolanderi (A. Gray) A. Gray,
Proc. Amer. Acad. Arts 8: 391. 1872. TYPE:
U.S.A. California: “Mariposa Big-tree Grove,”
//. N. Bolander s.n. (holotype, GH; isotype,
DS).

Rkvision ok Hemizonia sensu Keck (1959)

Under my revised taxonomy for species in Hem-
izonia sensu Keck (1959), Hemizonia includes only
the hayfield tarweeds, i.e., II. sect. Hemizonia sensu
Keck (1959) — the II. congesta DC. (the type spe¬
cies of Hemizonia) complex. Hemizonia in the new
sense comprises the annual members of Madiinae
with receptacles chaffy throughout, disc florets

functionally staminate, and cypselae without well-
developed beaks. The hayfield tarweeds are endem¬
ic to seasonally dry, low- to mid-elevational habitats
of the northern and central California Floristic
Province, and are most diverse in the North Coast
Ranges of California. Based on results of phyloge¬
netic analyses, Hemizonia in the new sense is ap¬
parently most closely related to Blepharizonia
Greene, also with 2 n = 14 II (Baldwin, unpub¬
lished).

Centromadia Greene is resurrected to accom¬
modate the spikeweeds [Hemizonia sect. Centro¬
madia sensu Keck (1935, 1959)], the only mem¬
bers of Madiinae with spine-tipped or apiculate
leaves. Members of Centromadia are annual or rhi-
zomatous-perennial herbs found in grassland or
woodland habitats at low elevation, mostly in poorly
drained, somewhat alkaline sites. The spikeweeds
are endemic to the California Floristic Province; C.
pungens subsp. septentrional is has become natural¬
ized in the interior Pacific Northwest of Oregon and
Washington. New infraspecific combinations in
Centromadia are provided below for practical con¬
venience, prior to critical systematic re-evaluation
of the infraspecific taxa.

Centromadia parryi (Greene) Greene subsp.
australis (D. D. Keck) B. G. Baldwin, comb,
nov. Basionym: Hemizonia parryi Greene
subsp. australis D. D. Keck, Madrono 3: 15.
1935. Hemizonia australis (D. D. Keck) D. D.
Keck, Aliso 4: 1 10. 1958. TYPE: U.S.A. Cal¬
ifornia: Orange County, Seal Beach. 29 Sep.
1933, D. D. Keck 2537 (holotype, DS; isotype,
CAS).

Centromadia parryi (Greene) Greene subsp.
congdonii (B. L. Robinson & Greenman) B.
G. Baldwin, comb. nov. Basionym: Hemizonia
congdonii B. L. Robinson & Greenman. Bot.
Gaz. 22: 169. 1896. Centromadia congdonii
(B. E. Robinson & Greenman) C. P. Smith,
Muhlenbergia 4: 73. 1908. Centromadia pun-
gens (Hooker & Arnott) Greene var. congdonii
(B. L. Robinson & Greenman) Jepson, Man.
FI. PI. Calif. 1087. 1925. Hemizonia parryi
Greene subsp. congdonii (B. L. Robinson &
Greenman) D. D. Keck, Madrono 3: 15. 1935.
Hemizonia parryi Greene var. congdonii (B. L.
Robinson & Greenman) Hoover, FI. Vase. PI.
San Luis Obispo Co., Calif. 288. 1970. T\ PE:
U.S.A. California: Monterey County, Salinas,
26 May 1886, J. W. Congdon 151 (holotype,
GH).

Volume 9, Number 4
1999

Baldwin

Combinations in Madiinae

467

Cenlroinadia parryi (Greene) Greene subsp.
rudis (Greene) B. G. Baldwin, comb. nov. Bas-
ionym: Centromadia rudis Greene, Man. Bot.
San Francisco 197. 1894. Hemizonia parryi
Greene subsp. rudis (Greene) D. D. Keck. Ma¬
drono 3: 15. 1935. TYPE: U.S.A. California:
Solano County, Vacaville, 20 Aug. 1892, W. L.
Jepson s.n. (lectotype, selected here, NDG
61004; isolectotype, JEPS 38135). Three po¬
tential syntypes at NDG conform to Greene’s
protologue (“Sacramento Valley, near Vaca¬
ville, Jepson'’). The date and locality of NDG
61004, annotated in D. D. Keck’s handwriting
as the type, are the same as on JEPS 38135,
annotated in Jepson’s handwriting as “part of
original collection.”

Centromadia pungens (Hooker & Arnott) Greene
subsp. laevis (D. D. Keck) B. G. Baldwin,
comb. nov. Basionym: Hemizonia pungens
(Hooker & Arnott) Torrey & A. Gray subsp.
laevis D. D. Keck, Madrono 3: 14. 1935. Hem¬
izonia laevis (D. D. Keck) D. D. Keck, Aliso
4: 110. 1958. TYPE: U.S.A. California: San
Bernardino Valley, 300 m, 8 July 1916, S. B.
Parish 10972 (holotype, DS; isotype, CAS).

Centromadia pungens (Hooker & Arnott) Greene
subsp. maritima (Greene) B. G. Baldwin,
comb. nov. Basionym: Centromadia maritima
Greene, Man. Bot. San Francisco 196. 1894.
Hemizonia pungens (Hooker & Arnott) Torrey
& A. Gray subsp. maritima (Greene) D. D.
Keck, Aliso 4: 110. 1958. TYPE: U.S.A. Cal¬
ifornia: “Alameda,” Oct. 1881, E. L. Greene
s.n. (lectotype, selected here, NDG 61021).
Two potential syntypes at NDG, both from “Al¬
ameda,” conform to Greene’s protologue (“Bor¬
ders of salt marshes about San Francisco Bay,”
no date or collector indicated). The collection
selected here as lectotype was annotated as the
apparent type by D. D. Keck.

Centromadia pungens (Hooker & Arnott) Greene
subsp. septentrionalis (D. D. Keck) B. G.
Baldwin, comb. nov. Basionym: Hemizonia
pungens (Hooker & Arnott) Torrey & A. Gray
subsp. septentrionalis D. D. Keck, Aliso 4:
110. 1958. Hemizonia pungens (Hooker & Ar¬
nott) Torrey & A. Gray var. septentrionalis (D.
D. Keck) Cronquist, FI. Pacific Northwest 529.
1973. TYPE: U.S.A. California: Siskiyou
County, Shasta Valley, H. M. Hall 12907 (ho¬
lotype, DS; isotype, DS).

Deinandra Greene is reinstated to accommodate

the species in Hemizonia sect. Madiomeris sensu
Tanowitz (1982) plus the informal “Fruticosae” or
“Zonamra” (see Clausen, 1951; Keck, 1959). Mem¬
bers of Deinandra in the new sense are the only
annuals, and non-rhizomatous perennials in Madi¬
inae that combine plivllaries each about half-en¬
closing a ray ovary, receptacular bracts usually re¬
stricted to a peripheral ring, yellow ray corollas,
beaked ray cypselae, and leaves lacking apical
spines or large tack or pit glands. The names Dein¬
andra Greene [type: D. fasciculata (DC.) Greene]
and Zonanthemis Greene [type: Z. corymbosa (DC.)
Greene, designated by Tanowitz (1982)] were pub¬
lished simultaneously. Their types fall within H.
sect. Madiomeris sensu Tanowitz (1982); both
names are available at the generic rank for the tax¬
on comprising //. sect. Madiomeris sensu Tanowitz
(1982) plus “Fruticosae” or “Zonamra” [the older
generic name Hartmannia DC. is a later homonym
of Hartmannia Spach], I have chosen to use Dein¬
andra to minimize the number of new combinations
necessary to institute the revised taxonomy of Hem¬
izonia sensu Keck (1959). New infraspecific com¬
binations in Deinandra are provided for practical
convenience, prior to critical systematic re-evalu¬
ation of the infraspecific taxa.

Deinandra in the new sense is nearly endemic
to the California Floristic Province, with the excep¬
tions of one species in the San Benito Islands of
central Baja California (D. streetsii ) and two species
in the Mojave Desert of California {D. arida and D.
mohavensis). Members of Deinandra usually occur
in exposed, low- to mid-elevational habitats, mostly
in seasonally dry sites. Based on results of phylo¬
genetic analyses, Deinandra is apparently more
closely related to the distinctive and reproductively
isolated genus Holocarpha Greene than to the
spikeweeds ( Centromadia ) (Baldwin, unpublished),
despite crossability and partial interfertility be¬
tween some members of Centromadia and Dein¬
andra (Clausen, 1951; Tanowitz, 1977).

Deinandra arida (D. D. Keck) B. G. Baldwin,
comb. nov. Basionym: Hemizonia arida D. D.
Keck, Aliso 4: 109. 1958. TYPE: U.S.A. Cal¬
ifornia: Kern County, Mojave Desert, mouth of
Red Rock Canyon, 2300 ft., 11 May 1935, D.
D. Keck & P. Stockwell 3279 (holotype, DS;
isotypes, CAS, UC; also ARIZ, B, BKL, BM,
BR, BRY, C, CAN, Cl, COLO, CU, E, F, G,
GH, HEL, IA, ILL, IND, K, L, LA, LD, MICH,
MIN, MO, MT, NY, OC, OKL, ORE, OS, OSC,
P, PENN, PH, POM, RM, S, SBBG, TEX, U,
UPS, US, UTC, WIS, WS, WTU, not seen).

468

Novon

Deinandra Clementina (Brandegee) B. G. Bald¬
win, comb. nov. Basionym: Hemizonia clemen-
tirui Brandegee, Erythea 7: 70. 1899. Zonan-
themis clementina (Brandegee) Davidson &
Moxley, FI. S. Calif. 401. 1923. TYPE: U.S.A.
California: San Clemente Island, 25 Aug.
1894, T. S. Brandegee s.n. (lectotype, selected
here, UC 174175). Type not designated by
Brandegee.

Deinandra conjugens (D. D. Keck) B. G. Bald¬
win, comb. nov. Basionym: Hemizonia conju-
gens D. D. Keck, Aliso 4: 109. 1958. TYPE:
U.S.A. California: San Diego County, “river
bottom land near Otay,” 16 May 1903, L.
Abrams 3521 (holotype, UC 407335; isotypes,
CAS; also NY, POM, not seen).

Deinandra corymbosa (DC.) B. G. Baldwin,
comb. nov. Basionym: Hartmannia corymbosa
DC., Prodr. 5: 694. 1836. Hemizonia corym¬
bosa (DC.) Torrey & A. Gray, FI. N. Amer. 2:
398. 1843. Zonanthemis corymbosa (DC.)
Greene, FI. Fran. 425. 1897. TYPE: U.S.A.
California: “Nova California,” 1833, Douglas
2 (holotype, G-DC not seen; photograph and
fragments, UC; isotypes [fide Tanowitz (1982)],
BM, K, not seen).

Deinandra corymbosa (DC.) B. G. Baldwin
subsp. macrocephala (Nuttall) B. G. Bald¬
win, comb. nov. Basionym: Hemizonia macro¬
cephala Nuttall, J. Acad. Nat. Sci. Philadel¬
phia II 1: 175. 1848. Hemizonia angustifolia
DC. subsp. macrocephala (Nuttall) D. D. Keck,
Madrono 3: 12. 1935. Hemizonia corymbosa
(DC.) Torrey & A. Gray subsp. macrocephala
(Nuttall) D. D. Keck, Aliso 4: 109. 1958.
TYPE: U.S.A. California: “At St. Simeon, Up¬
per California,” W. Gambel (holotype, PH not
seen). Tanowitz (1982). who reported seeing
the type, indicated the type locality as “in the
vicinity of Monterey” and the collector as Nut¬
tall.

Deinandra f’rutescens (A. Gray) B. G. Baldwin,
comb. nov. Basionym: Hemizonia frutescens A.
Gray, Proc. Amer. Acad. Arts 11: 79. 1876.
TYPE: Mexico. Baja California: “the interior
of Guadalupe Island,” 1875, E. P<ilmer s.n.
(holotype, GH).

Deinandra greeneana (Rose) B. G. Baldwin,
comb. nov. Basionym: Hemizonia greeneana
Rose, Contr. U.S. Natl. Herb. 1: 24. 1890.
TYPE: Mexico. Baja California: Guadalupe Is¬
land, “south end” (on label), 29 Mar. 1889, E.
Palmer 865 (holotype, US; isotype, CAS).

Deinandra greeneana (Rose) B. G. Baldwin
subsp. peninsularis (Moran) B. G. Baldwin,
comb. nov. Basionym: Hemizonia greeneana
Rose subsp. peninsular is Moran, Trans. San
Diego Soc. Nat. Hist. 15: 286. 1969. Hemizon¬
ia greeneana Rose var. peninsularis (Moran) B.
L. Turner, I’hytologia 69: 15. 1990. TYPE:
Mexico. Baja California: Punta Banda, Banda
Peak, 27 Aug. 1966, Moran 13437 (holotype,
SD not seen; isotypes, CAS, UC; also K,
MEXU, US, “etc.," not seen).

Deinandra halliana (D. D. Keck) B. G. Baldwin,
comb. nov. Basionym: Hemizonia halliana D.
D. Keck, Madrono 3: 12. 1935. TYPE: U.S.A.
California: San Luis Obispo County, 1.5 mi. E
of Cholame, dry flood plain of Cholame Creek,
1200 ft., 4 May 1933, I). D. Keck & W. M.
Heusi 2170 (holotype, DS; isotypes, DS, UC;
also [fide Tanowitz (1982)] MO, NY, US, not
seen).

Deinandra inereseens (H. M. Hall ex D. D.
Keck) B. G. Baldwin, comb. nov. Basionym:
Hemizonia paniculata A. Gray subsp. incres-
cens H. M. Hall ex D. D. Keck, Madrono 3:
11. 1935. Hemizonia inereseens (H. M. Hall ex
D. D. Keck) Tanowitz, Syst. Bot. 7: 328. 1982.
TYPE: U.S.A. California: San Luis Obispo
County, 7.5 mi. SW of Arroyo Grande, 7 June
1931, //. M. Hall 13136 (holotype, DS; iso¬
type, DS).

Deinandra inereseens (H. M. Hall ex D. D.
Keck) B. G. Baldwin subsp. foliosa (Hoover)
B. G. Baldwin, comb. nov. Basionym: Hemi¬
zonia paniculata A. Gray subsp. foliosa Hoo¬
ver, FI. Vase. PL San Luis Obispo Co., Calif.
288. 1970. Hemizonia inereseens (H. M. Hall
ex D. D. Keck) Tanowitz subsp. foliosa (Hoo¬
ver) Tanowitz, Syst. Bot. 7: 330. 1982. TYPE:
U.S.A. California: San Luis Obispo County,
sand hills N of Pozo, 6 Oct. 1965, Hoover 9715
(lectotype, designated by Tanowitz (1982), OBI
not seen; isolectotypes, CAS, UC; also OBI,
RSA, not seen).

Volume 9, Number 4
1999

Baldwin

Combinations in Madiinae

469

Deinandra inerescens (H. M. Hall ex D. D.
Keck) B. G. Baldwin subsp. villosa (Tanowitz)
B. G. Baldwin, comb. nov. Basionym: Hemi-
zonia inerescens (H. M. Hall ex D. D. Keck)
Tanowitz subsp. villosa Tanowitz, Syst. Bot. 7:
331. 1982. TYPE: U.S.A. California: Santa
Barbara County, Gaviota, Tanowitz 1650 (ho-
lotype, UCSB; isotypes (2), JEPS).

Deinandra inartirensis (D. D. Keck) B. G. Bald¬
win, comb. nov. Basionym: Hemizonia marti-
rensis D. D. Keck, Madrono 3: 12. 1935.
TYPE: Mexico. Baja California: 25 mi. E of
San Telmo, in immediate vicinity of San Jose,
foothill region at base of San Pedro Martir
Mountains, 23 Feb. 1931, A. Meling 2 (holo-
type, DS; isotypes, DS, UC; also [fide Tanowitz
(1982)] C, GH, US, not seen).

Deinandra mint bornii (Jepson) B. G. Baldwin,
comb. nov. Basionym: Hemizonia minthornii
Jepson, Man. FI. PI. Calif. 1092. 1925. TYPE:
U.S.A. California: Ventura County, Santa Su-
sana Mountains, 15 Nov. 1923, T. W. Minthorn
s.n. (holotype, JEPS 2628).

Deinandra inohavensis (D. D. Keck) B. G. Bald¬
win. comb. nov. Basionym: Hemizonia mohav-
ensis D. D. Keck, Madrono 3: 9. 1935. TYPE:
U.S.A. California: San Bernardino County,
“Mohave River, at confluence with Deep
Creek, Mohave Desert,” 900 m, 17 Sep. 1933,
D. D. Keck 2531 (holotype, DS; isotype, DS).

Deinandra pallida (D. D. Keck) B. G. Baldwin,
comb. nov. Basionym: Hemizonia pallida D. D.
Keck, Madrono 3: 8. 1935. TYPE: U.S.A. Cal¬
ifornia: Kern County, 5.3 mi. N of Grapevine,
“Head of the San Joaquin Valley,” 6 May
1933, D. D. Keck & W. M. Heusi 2255 (holo¬
type, DS; isotypes, CAS, UC; also B, C, FM,
GH, K, MO, NY, PH, POM, US, not seen).

Deinandra palmeri (Rose) B. G. Baldwin, comb,
nov. Basionym: Hemizonia palmeri Rose,
Contr. U.S. Natl. Herb. 1: 24. 1890. TYPE:
Mexico. Baja California: Guadalupe Island,
Mar. 1889, E. Palmer 874 (holotype, US).

Deinandra pentaetis (D. D. Kec k) B. G. Baldwin,
comb. nov. Basionym: Hemizonia lobbii Greene
subsp. pentaetis D. D. Keck, Madrono 3: 8.
1935. Hemizonia pentaetis (D. D. Keck) D. D.
Keck, Aliso 4: 109. 1958. TYPE: U.S.A. Cal¬
ifornia: San Luis Obispo County, “East side of
the Salinas River, opposite San Miguel,” 27
Apr. 1934, D. D. Keck & J. Clausen 2836 (ho¬
lotype, DS; isotypes, DS; also GH, K, not
seen).

Deinandra streetsii (A. Gray) B. G. Baldwin,
comb. nov. Basionym: Hemizonia streetsii A.
Gray, Proc. Amer. Acad. Arts 12: 162. 1877.
TYPE: Mexico. Baja California: “San Benito
Island,” T. H. Streets (holotype, GH).

The following key institutes a new taxonomy of
Madiinae. The key was constructed for convenience
of identification rather than to reflect phylogenetic
relationships and does not list all distinguishing
characteristics of the taxa or groups of taxa.

Key to Genera of Madiinae

la. Ray oypselae strongly obcompressed; if rays lacking, plants pappose annuals.

2a. Disc florets 1( — 2); minute annuals . Hemizonella A. Gray

2b. Disc florets 3 or more; annuals or perennials.

3a. Plants rhizomatous perennials . Holozonia Greene

3b. Plants annuals.

4a. Pappi of 10 broad, apically obtuse scales . Achyrachaena Schauer

4b. Pappi absent or of elements with acute apex.

5a. Heads calyculate; ray florets 5; disc florets 6, functionally staminate; disc pappus

absent . Lagophylla Nultall

5b. Heads not calyculate; ray florets 0 or 3—27; disc florets 5— 100+, bisexual; disc pappi

usually present . Layia Hooker & Arnott

lb. Ray cypselae compressed or ± terete or somewhat triangular in cross section; if rays lacking, plants peren¬
nials or epappose annuals.

6a. Styles of disc florets hairy proximal to minute branches; annuals . Rleplmripappus Hooker

6b. Styles of disc florets glabrous proximal to branches; annuals or perennials.

7a. Plants perennial herbs, ± scapose; disc pappi of ciliate-plumose, subulate scales .

. Raillardella A. Gray

7b. Plants annuals, leafy-stemmed perennial herbs, perennial rosette plants, or of woody growth form;
disc pappi absent or various.

470

Novon

8a. Heads radiate or discoid; each ray cypsela usually compressed (to 3-angled with broad,
rounded abaxial surface) and completely or mostly enclosed within a phyllary (il ray cypsela
terete with phyllary only enclosing abaxial surface of cypsela, plants high-montane annuals
or perennials).

9a. Plants woody or semi-woody, evergreen; Hawaiian.

10a. Hay florets present; rosette plants . Argyroxiphium DC.

I Ob. Hay florets absent; habits various.

1 la. Leaves 1—4 per node, free or weakly united; trees, (sub)shrubs, mat-plants,

cushion-plants, vines . Dubautia Gaudichaud-Beaupre

1 lb. Leaves 5 or more per node, strongly coalescent at base; rosette plants ....

. Wilkesia A. Gray

9b. Plants herbaceous (aerial stems rarely semi-woody), seasonal; North American or South
American.

12a. Disc pappi absent . Madia Molina

12b. Disc pappi present.

13a. Heads all discoid . Carlquistia B. G. Baldwin

13b. Heads radiate on at least part of plant.

14a. Plants perennials.

15a. Involucres campanulate to hemispheric; anthers dark purple

. Kyhosia B. G. Baldwin

15b. Involucres elliptic or rounded in silhouette; anthers yellow' ....

. Anisocarpus Nuttall

14b. Plants annuals.

16a. Anthers yellow' . Harmonia B. G. Baldwin

16b. Anthers dark purple . Jensia B. G. Baldwin

8b. Heads radiate; each ray cypsela terete to slightly obcompressed with adaxial surface flat¬
tened, slightly bulging, or low-keeled and only abaxial surface enclosed within a phyllary;
plants low- to mid-elevation annuals, perennial herbs, and (sub)shrubs.

17a. Annuals; leaves filiform to narrowly linear and often strongly revolute; mostly with
tack glands; ray corolla lobes often more than half the total length of the limb, the

lateral lobes often spreading.

18a. Hay cypselae beaked; tack glands absent . Osmadenia Nuttall

18b. Hay cypselae not beaked; tack glands present . Calycadenia DC.

17b. Annuals, perennial herbs, or (snb)shrubs; leaves linear or broader and usually not
strongly revolute; mostly without tack glands; ray corolla lobes usually less than half
the total length of the limb, the lobes ± parallel.

19a. Hay cypselae not beaked; corollas white or, less commonly, yellow.

20a. Disc florets bisexual; ray and disc cypselae hairy, often with pappus .

. Blepharizonia Greene

20b. Disc florets functionally staminate; ray cypselae glabrous, without pappus

. Hemizania DC.

19b. Hay cypselae beaked; corollas yellow.

21a. Leaves each with an apical spine . Cenlromadia Greene

21b. Leaves each without an apical spine.

22a. Bracts each with large apical pit gland; receptacles with paleae

throughout . Holocarpha (DC.) Greene

22b. Bracts without large pit glands; receptacles with paleae usually re¬
stricted to bases of outermost disc florets . Deinandra Greene

Acknowledgments. I thank Gerald D. Carr, David
J. Keil, and John L. Strother for helpful reviews of
the manuscript; JLS lor his advice and encourage¬
ment during the study; JLS and Alan R. Smith lor
assistance with the Latin diagnoses; and Victoria C.
Hollowell and Roy Gereau for careful editing ol the
paper. I also thank the curators of GH, HSC, NDG,
and US for loans. This paper is based on research
supported by the National Science Foundation
(DEB-9458237), and by Roderick Park and other
generous Friends ol the Jepson Herbarium.

I .iterature Cited

Baldwin, B. G. 1992. Phylogenetic utility of the internal
transcribed spacers of nuclear ribosomal DNA in

plants: An example from the Compositae. Molec. Phy¬
logenetics Evol. I: 3—16.

- . 1993. Molecular phylogenetics of Madia (Com¬
positae— Madiinae) based on ITS sequences of Iff— 26S
nuclear ribosomal DNA. Amer. J. Bot. 80 (6, suppl.):
130.

- . 1996. Phylogenetics of the California tarweeds

and the Hawaiian silversword alliance (Madiinae;
Heliantheae sensu lata). Pp. 377-391 in I). J. N.
Hind & H. J. Beentje (editors), Compositae: Systein-
atics. Proceedings of the International Compositae
Conference, Kew, 1994, vol. 1. Royal Botanic Gar¬
dens, Kew.

- . 1997. Adaptive radiation of the Hawaiian sil¬
versword alliance: Congruence and conflict of phyloge¬
netic evidence from molecular and non-moleeular in¬
vestigations. Pp. 103—128 in T. J. Givnish <X K. J.

Volume 9, Number 4
1999

Baldwin

Combinations in Madiinae

471

Sytsma (editors). Molecular Evolution and Adaptive Ra¬
diation. Cambridge Univ. Press, Cambridge, U.K.
- & R. II. Robichaux. 1995. Historical biogeogra¬
phy and ecology of the Hawaiian silversword alliance
(Asteraceae): New molecular phylogenetic perspectives.
Pp. 259-287 in W. L. Wagner & V. A. Funk (editors).
Hawaiian Biogeography: Evolution on a Hot Spot Ar¬
chipelago. Smithsonian Institution Press, Washington.
D.C.

- , I). W. Kyhos, J. Dvorak & G. I). Carr. 1991.

Chloroplast l)NA evidence for a North American origin
of the Hawaiian silversword alliance (Asteraceae). Proc.
Natl. Acad. U.S.A. 88: 1840-1843.

Carlquist. S. 1959. Studies on Madinae: Anatomy, cytol¬
ogy, and evolutionary relationships. Aliso 4: 171-236.
Carr. G. 1). 1985. Monograph of the Hawaiian Madiinae
(Asteraceae): Argyroxiphium, Dubautia, and Wilkesia.
Allertonia 4: 1—123.

- . B. G. Baldwin & I). W. Kyhos. 1996. Cytogenetic

implications of artificial hybrids between the Hawaiian
silversword alliance and North American tarweeds (As¬
teraceae: Heliantheae— Madiinae). Amer. J. Bot. 83:
653-660.

Clausen. J. 1951. Stages in the Evolution of Plant Species.
Ilafner, New York.

Keck, D. D. 1935. Studies upon the taxonomy of the Ma¬
dinae. Madrono 3: 4—18.

- . 1959. Subtribe Madiinae. Pp. 1 106—1 129 in P.

A. Munz, A California Flora. Univ. California Press,
Berkeley.

Kyhos, D. W., G. I). Carr & B. G. Baldwin. 1990. Biodi¬
versity and cytogenetics of the tarweeds (Asteraceae:
Heliantheae— Madiinae). Ann. Missouri Bot. Gard. 77:
84-95.

Tanowitz, B. 1). 1977. An intersectional hybrid in Hemi-
zonia (Compositae: Madiinae). Madrono 24: 55—61.

- . 1982. Taxonomy of Hemizonia sect. Madiomeris

(Asteraceae: Madiinae). Syst. Bot. 7: 314—339.

A New Species of Tapirira (Anaeardiaceae) from Ecuador

Anders S. Barfod

Department of Systematic Botany, Institute of Biological Sciences, Aarhus University,
Nordlandsvej 68, DK-8240 Risskov, Denmark

ABSTRACT. Tapirira rubrinervis Barfod, sp. nov., is
described from the coastal plain of Ecuador. It is
easily distinguished from its congeners by the few-
foliolate leaves, the ohovate, unequally sized leaf¬
lets, and the striking red venation of the lower sur-
faces of the leaflets. The variation in leaf
morphology and phenology is discussed.

RKSUMKN. Se describe una especie nueva, Tapi¬
rira rubrinervis Barfod, de la region costera de Ec¬
uador. Se distingue facilmente de sus congeneres
por tener las hojas con pocos folfolos, los folfolos
desiguales y obovados, con llamativa nervadura
roja del enves. Asimismo, se comenta la variacion
en la morfologfa de las hojas y su fenologfa.

A recent revision of Ecuadorian Anaeardiaceae
for the Catalogue of the Vascular Plants of Ecuador
(J0rgensen & Ledn-Yanez, 1999) revealed a new
species of Tapirira. Barfod (1987) described only
two species of Tapirira in Flora oj Ecuador. Tapi¬
rira guianensis Aublet, with two subspecies (Bar¬
fod, 1986), and T. peckoltiana Engler, whic h should
be correctly named T. obtusa (Bentham) J. D.
Mitchell (Mitchell, 1993). Since this flora treatment
was published, T. guianensis has been recorded
several times west of the Andes in the provinces of
Esmeraldas and Carchi, Ecuador. It has been found
on the northern coastal plain along the upper trib¬
utaries of the Cayapas and Mira rivers. In the Cay-
apas river system, it is locally known as “sajo de
arriba,’' or “sajo chi.” “Sajo” is the name for Camp-
nosperma panamense Standley, a rare simple-leaved
species of Anaeardiaceae growing near Borbon (Lit¬
tle & Dixon, 1969).

In 1993, 1 received for the first time material of
an undescribed species of Tapirira growing along
the Mira River and its tributaries in the province
of Carchi, where it is locally known as “cuilde.”
The congener T. guianensis also occurs in that area.
It is called “cuilde bianco.” There was little doubt
that the new material represented a separate spe¬
cies, and it is described here based on nine col¬
lections. Although it has several very distinctive
features, it is probably most closely related to Tap¬
irira guianensis. Like this species it has charta-

ceous leaflets that are inconspicuously hairy be¬
neath or glabrous. The flowers are typical for the
genus.

Kl- 'i TO Eel ADOREAN SPECIES OF TAI'IRIRA

la. [.eaves (f— )3— 5 foliolate, obovate, rounded api-
eally, glabrous, venation bright red beneath . . .

. T. rubrinervis Barfod

lb. [.eaves 5—15 foliolate, ovate to elliptic, rarely ob¬
long, acuminate, with hairs beneath, venation
brown or green.

2a. Leaflets with adpressed hairs beneath or gla¬
brous.

3a. leaflets chartaceous, adpressed-pubes-
cent beneath; sepals less than 0.4 mm
long, petals 1.5-2 mm, venation usually
inconspicuous; disc ca. 1 mm diam . .

. . T. guianensis Aublet subsp. guianensis
3b. Leaflets subcoriaeeous, sparsely adpres-
sed-pubeseent to glabrous beneath; se¬
pals more than 0.5 mm long, petals 2-
2.5 mm, venation conspicuous; disc

1.5—2 nun diam .

. . T. guianensis Aublet subsp. subandina
Barfod & Holm-Nielsen
2b. Leaflets densely tomentose pilose to veluti-

nous beneath .

. . . I obtusa (Bentham) J. 1). Mitchell (syn. T.

peckoltiana Engler)

Tapirira rubrinervis Barfod, sp. nov. TYPE: Ec¬
uador. Carchi; border area with Prov. of Es¬
meraldas, km 20 on Lita-Alto Tambo road,
550 m, 25 June 1991 (fl), H. van der Werfj , B.
Gray & G. Tipaz 11971 (holotype, AAU; iso¬
types, MO, QCA). Figure 1.

A congeneris ceteris differt foliis paucifoliolatis, foliolis
statura inaequale, obovatis, glabris. Nervatio foliorum in¬
fra conspicue rubra.

Evergreen trees to 30 m tall and 65 cm DBH.
Bark unknown. Branehlets with deciduous ferrugin-
eous feltlike pubescence, lentieels irregularly dis¬
tributed, abundant on young portions, unevenly
sized, rounded; leaf-bearing twigs hollow in dried
condition with minute longitudinal ridges. Leaves
imparipinnate, (1 — )3 — 5 foliolate, the number of leaf¬
lets per leaf apparently decreasing toward the tip
of the branches, the distal leaves being trifoliolate
or simple (unifoliolate) ( Tipaz et al. 1031); petiole

No VON 9: 472-175. 1999.

Volume 9, Number 4
1999

Barfod

Tapirira rubrinervis from Ecuador

473

Figure 1. Tapirira rubrinervis Barfod. — A. Flowering branch. — B. Distal floral branch. — C. Flower at staminate
anthesis. — D. Longitudinal section of C. — E. Infructescence. (A-l) based on van der Werff et al. 11971 ; E based on
Tipaz et al. 1339.)

474

Novon

(5-)8-12(-16) cm long, raehis to 15 cm long in the
7-foliolate leaves, both raehis and petiole glabrous,
brown to ferrugineous, terete, with lens-shaped len-
tic-els and fine longitudinal ridges; leaflets opposite,
lateral petiolules 5—10 mm, terminal petiolule up
to 8 cm long, glabrous, with dorsal furrow from at¬
tenuated leaf base; leaflets markedly different in
size, the proximal ones (if present) smallest, 6—8 X
4—5 cm, the terminal one largest up to 25 X 14
cm, elliptic to obovate, chartaceous in dried con¬
dition, lustrous green, glabrous to minutely puber-
ulent above, smooth matte green below, margin en¬
tire, subrevolute, apex rounded to obtuse (rarelv
with short acumen), base obtuse to rounded,
oblique, shortly attenuate; venation brochidodrome
pinnate, midvein prominulent often in groove
above, very prominent and red beneath, secondary
veins oppposite to alternate, prominulent above,
prominent and red below, veinlets only clearly vis¬
ible and somewhat impressed below. Inflorescence
paniculate, arising in the axils of new leaves and
grouped at a subterminal position, at anthesis usu¬
ally superseded by juvenile leaves at the distal
nodes, decreasing in size and degree of develop¬
ment toward the tip of the branch, the proximal
lateral inflorescence being up to 40 cm long at an¬
thesis, slightly curved, with peduncle to 14 cm
long, all axes with ferrugineous pubescence varying
in density, flowers typically borne in congested
groups of three on short monopodial flower-bearing
branches, each subtended by deltoid bracts ca. 0.5
mm long, floral bracts conspicuous, 0. 3-0.5 mm
long, often rounded apically, bracteoles inconspic¬
uous to conspicuous on single lateral flowers toward
the distal end of the flower-bearing branches, in¬
serted immediately below the abscission zone.
Flowers at staminate anthesis subsessile, rarely
with pedicel up to 1 mm, sepals 5, 0. 7-0.8 X 0.7-
0.8 mm, rounded, with sparse pubescence on ab-
axial faces, marginally ciliate; petals 5, 1.5-1. 8 X
0. 7-0.9 mm, ovate to elliptic, with visible venation
in dried condition, reflexed; stamens 10, of unequal
length up to 2 mm long, filaments slender, subulate,
anthers 0.2— 0.4 mm long, rounded; disc about 1.0
mm wide, 10-crenulate; rudimentary pistil 0.5-0. 6
mm long, tomentose to pilose; ovary partly sunken
in disc; styles 5, straight to slightly recurving; stig¬
mas discoid. Flowers at late pistillate anthesis with
developing pistil not available. Green fruits up to
10 mm long, some with wilted, 0.8-1.0-min-long
stamens at the base, apparently with aborted the¬
cae. Mature fruits not seen.

Leaf morphology. The leaves are few-foliolate,
typically trifoliolate. From the material, it appears
that trees with 5-foliolate leaves are common. One

specimen (Tipaz el al. 1031) has simple (unifolio-
late) leaves, at least toward the tips of the inflores¬
cence-bearing branches.

Phenology. The sexual expression within Tap-
irira needs to be studied in detail before any con¬
clusions can be drawn. As interpreted here, T. rub-
rinervis is functionally dioecious like other species
of Tapirira. However, most of the material examined
is in bud and difficult to determine to sex. Fully
open flowers with reflexed petals and unequal sta¬
mens, to 2 mm long, were only found in Quelal 647
and van der Werff et al. 11971. The anthers are
past dehiscence, and a large proportion of the pol¬
len grains has germinated.

In other specimens, e.g., Aulestia 1257, Tipaz et
al. 1339, and Mendez et al. 400, flowers have just
opened. These are generally smaller than the fully
opened flowers described above, with petals 1.2—
1.3 mm long and stamens of equal length. The fil¬
aments are only 0. 3-0.4 mm long and the anthers
rounded. These are interpreted as functionally sta¬
minate flowers at early anthesis. Aberrant pollen
grains were not observed, and the gynoecium was
very similar to that of the long-staminate flowers,
except for being a little smaller. It should be noted
that in the material examined, I did not see pistil¬
late flowers with developing, fertilized gynoecia.

Fruiting material of Tapirira rubrinervis was long
past anthesis and gave few indications of the mor¬
phology ol the assumed pistillate flower. Only the
0.8— 1.0-mm-long, wilted stamens suggest that the
flowers are different in proportions from their sta¬
minate homologues.

Based on this limited evidence, it cannot be ex¬
cluded that Taprira rubrinervis has hermaphroditic
flowers with sexual expression separated temporally
but not spatially. This seems uidikely, since all oth¬
er species of Tapirira are known as functionally di¬
oecious or polygamodioecious.

Common name. “Sajo de arriba” (Esmeraldas),
“Cuilde” (Carchi).

Use. Fibers are extracted from this species ac¬
cording to Quelal et al. 267. What they are used
for is not specified, however. The fruits are eaten
by several birds in the area.

Distribution. Tapirira rubrinervis is only known
from the province of Carchi in Ecuador.

Paratypes. ECUADOR. Carchi: border area between
I’rov. Carchi and Esmeraldas, km 20 Lita— Alto Tambo
road, 7(H) m, 2d June 1991 (Ir). //. ran der Werff, II. Gray
& G. Tipaz 11894 (AAU, MO, QCA, QCNE); Parroquia
Tobar Donoso, Reserva indfgena Awa. bosque primario
noreste de la casa comunal, cerca de R fo Botella,
78°24'W, l°0'N. 650-1000 m, 19-28 June 1002 (Ir). G.
Tipaz, ./. Zuleta & IS. Guanga 1329 (AAU. MO, OCA.
QCNE); Parroquia Chical, Gualpi Medio, Reserva I ndf-

Volume 9, Number 4
1999

Barfod

Tapirira rubrinervis from Ecuador

475

gena Awa, Sendero a San Marcos al norte de la casa com¬
munal. 23—27 May 1992 (11. laid and early anthesis), G.
Tipaz, C. Quelal & G. Cantincuz 1031 (AAU, MO. QCA,
QCNE); Chical. Reserva etnica Awa— Camumbf, 78°16'W,
0°53'N. 1700-1900 m. 20-29 Jidy 1991 (fr). C. Quelal,
C. Aulestia A: F. Nastacuaz 267 (AAU. MO. QCA. QCNE);
Comunidad dc Gualpi Medio, 78°16'W. 1 0 1 'N. 900 m. 21
May 1992 (fl. staminate anthesis), C. Quelal, G. Tipaz &
./. Taicuz 6 47 (AAU, MO, QCA. QCNE); San Marcos de
los Coaiqueres and surrounding perhumid forest on trail
Ghical-Tobar Donoso, 78°16'W. 1°6'N. 800 m, 8 Feb.
1985 (fl. bud). IT 0llgaard, J. Korning, A. Thomsen & 7.
Ilium 57632 (AAU. MO. QCA); Parroquia Alto Tambo, la
Union. Reserva etnica Awa, 78°26'W, 00°52'N, 250 m 22
Mar. 1993 (11. bud and early anthesis), C. Aulestia A: M.
Aulestia 125 9 (AAU. MO, NY, QCA, QCNE); Parroquia
el Chical, San Marcos, 78°14'W, 01°06'N. 900-1100 m.
20-30 Apr. 1993 (fl. bud and early anthesis), P. Mendez,
./. Aulestia & ./. Pai 400 (AAU. MO. NY. QCA. QCNE).

Acknowledgments. The Latin diagnosis was
kindly checked by Benjamin 011gaard. Kirsten
Tind rendered the line drawing. Manuel Macia
helped write the Spanish abstract.

Literature Cited

Rarfod, A. 1986. Two new Anacardiaceae from Ecuador.
Nordic J. Hot. 6: 423-426.

- . 1987. Anacardiaceae. In: G. Hailing & L. An-

dersson (editors), Flora of Ecuador 30: 9—19.
Jprgensen. P. M. & S. Leon-Yanez (Editors). 1999. Cata¬
logue of the Vascular Plants of Ecuador. Monogr. Syst.
Hot. Missouri Hot. Card. 75.

Little, E. L. & R. G. Dixon. 1969. Arboles comiines de la
Provincia de Esmeraldas. FAO, Rome.

Mitchell. J. I). 1993. Tapirira obtusa comb. nov. (Anacar¬
diaceae). Novon 3: 66.

A New Species of Gibsoniothamnus (Sehlegeliaceae) from

Costa Rica and Panama

Kerry Barringer

Brooklyn Botanic Garden, 1000 Washington Avenue, Brooklyn, New York 11225, U.S.A.

Absthact. A new spec ies, Gibsoniothamnus par-
vifolius, is described based on disjunct populations
from Costa Rica and western Panama. The species
is distinguished by its relatively small, acuminate
leaves, few-flowered inflorescences, and its rela¬
tively small, purplish flowers borne on short pedi¬
cels.

Gihsoniothaiiinus parvifolius Barringer, sp. nov.
TYPE: Panama. Chiriquf: NE del campamento
Fortuna, 1000-1200 m, 15 Aug. 1976, M.
Correa, R. Dressier & Garibaldi 2500 (holo-
type, MO; isotypes, NY, PMA). Figure 1.

A Gibsoniothamno epiphytico loliis minoribus anguste
lanoeolatis vel ovato-laneeolatis acuminatis coriaceis, flo-
ribus solitariis vel geminatis differt.

Epiphytic shrub or small epiphytic tree to 4 m
tall; young branches glabrous to sparsely pubes¬
cent, terete to slightly angled; older stems with
gray-brown bark and persistent raised petiole ba¬
ses. Leaves opposite, anisophyllous; larger leaves
with petioles 3-7 mm long, the lamina narrowly
lanceolate to obovate lanceolate, 3—5.5 cm long,
1.0— 1.8 cm wide, coriaceous, glabrous or rarely pu¬
bescent, gland-dotted, the base cuneate, the apex
acuminate to long-acuminate, the veins immersed
above, prominent and often reddish below, 2 or 3
secondary veins per side, the upper surface dark
green, shiny, the lower surface light green, dull,
domatia present in the axils of the secondary veins;
smaller leaves with petioles 2-5 mm long, the lam¬
ina ovate-lanceolate, 1.6— 2.2 cm long, 0.6— 1.4 cm
wide, coriaceous, glabrous or rarely pubescent,
gland-dotted, the base rounded to acute, the apex
rounded to obtuse, rarely acuminate, the veins im¬
pressed above, prominent and sometimes reddish
below, 1-2 secondary veins per side. Flowers ax¬
illary, solitary or in pairs; pedicels 1. 5-2.0 cm long,
sparsely pubescent with glandular trichomes,
slightly thickened above, braeteolate at base, the
bracteoles linear, to 1 mm long; calyx eampanulate,
purple to red-purple, the tube 3-4 mm long, with
5 lobe-like enations that develop from the side of
the calyx tube, the enations 8—9 mm long, 1 mm
wide, narrowly triangular, slightly winged, glabrous.

spreading; corolla cylindric, reddish violet to pur¬
ple or lilac, 1.5— 2.0 cm long, 3 mm diam., the up¬
per lobes broadly ovate, 1.5— 2.0 mm long, 2.0— 2.5
mm wide, the lateral lobes ovate, 1.5— 2.0 mm long,
1.5— 2.0 mm wide, the median lobe ovate, 1.5— 2.0
mm long, 2.0— 2.5 mm wide, the lobes generally
more darkly colored than the tube; stamens 4, at¬
tached about 5 mm above the base of the corolla
tube, with a ring of trichomes at the level of at¬
tachment, the filaments 12-16 mm long, filiform,
curved above, slightly thickened at the base; stam-
inode 1.7— 5.0 mm long; pistil glabrous, ovary glo¬
bose, 1 mm tall, glandular at the base, tin* style
elongate, 15—18 mm long, the stigma clavate to
capitate. Berry globose, to 1 cm diam., green turn¬
ing white, with 5— 10(-20) seeds, the seeds fusi¬
form, 2.0— 2.5 mm long 1.0-1. 4 mm wide, with a
loose, hyaline, reticulate seedcoat.

This species is distinguished by its small leaves,
slightly winged calyx with narrowly triangular ena¬
tions, and few-flowered inflorescence. It is found
growing on the volcanic mountains of Costa Rica
and western Panama in two disjunct populations.
The northern Costa Rican plants grow between 300
and 700 in, while the southern Costa Rican and
Panamanian plants grow between 1000 and 1300
m elevation. The two populations differ slightly and
may eventually prove to be distinct species. The
northern plants tend to have terete, not angled,
branches and smaller staminodes, less than 3 mm
long. The distribution of the northern populations
overlaps slightly with that of Gibsoniothamnus epi¬
phyticus (Standley) L. 0. Williams, native to central
Costa Rica, and the southern populations.

This species is sometimes confused with Gibson¬
iothamnus epiphyticus, but the new species differs
by its smaller, narrowly lanceolate, acuminate, co¬
riaceous leaves, and 1- or 2 -flowered inflorescence.
In contrast, G. epiphyticus has elliptic, subeoria-
ceous leaves, 5.5— 7.5 cm long, and flowers in con¬
densed, paniculate inflorescences. The new species
is also confused with G. alatus A. Gentry, but that
species has large calyx wings, usually extending
more than 4 mm, leaves 4—9 cm long, and white
flowers. Many of the specimens of G. parvifolius are

Novon 9: 476-478. 1999.

Volume 9, Number 4
1999

Barringer

Gibsoniothamnus parvifolius

477

Figure 1. Gibsoniothamnus parvifolius Barringer. — A. Plant habit. — B. Unequal leaves at a node. — C. Flower habit.
— 1). Flower dissected; note the staminode and the enations on the calyx. — E. Fruit with detail of trichome.

tentatively identified as G. pterocalyx A. Gentry.
That species has leaves that are acute and not as
thick as those in G. parvifolius , the larger leaves at
a node have petioles 1.0-1. 3 cm long, the calyces
are green, and the Howers are usually borne in con¬
densed panicles.

D'Arcy (1979) included three specimens of this
species under G. alatus, noting that they probably
represented a distinct species.

Paratypes. Costa Rica. Alajuela: along road between
Canas and Upala. 8 km N of Bijagua, 300 in. 26 June
1976, Groat 36505 (MO); Reserva Biologica Monteverde,
Valle del Rio Penas Blancas 10°20'N, 84°43'W, 900 m.

28 Apr. 1987, Haber & Cru: 7006 (CR. F, MO); Upala,
Dos Rfos, 5 km al sur de Brasilia, margen derecha de Rfo
Pizote. 10°35'N. 85°20'W. 500 m. 30 Oct. 1987. Herrera
1015 (CR. F. MO); Upala, Colonia Libertad, 2 km NK,
cuenca del Rio Cucaracho, 450 m. 6 May 1988. Herrera
1980 (CR. F. MO); Upala. Bijagua. FI Pilon, 10°43'30"N.
85°00'30"W, 600 m, 12 July 1988. Herrera 2071 (MO);
Upala, Bijagua, El Pilon, camino a Chimurria 10°43'45"N.
84°59'58"W 500-600 m, 20 July 1988, Herrera 2131
((31, F, MO); approx. 8 km N of Rid Naranjo on the
Canas— Upala road. 2 km N of Bijagua, 400— 500 m, 21
Sep. 1975, Utley & Utley 3161 (DUKE, F. MO). Guana-
caste: 6 km N of village of Rio Chiquito, slope above
Fake Arenal, 600-700 m, I0°25'N, 84°46'W. 8 May
1986, Haber ex Hello et al. 4831 (MO); 6 km \E ol Rfo
Chiquito village, next to Lake Arenal, 10°27'N, 84°4R'W.

478

Novon

MM) m. 8 May 1988, Haber ex Hello el al. 47119 (MO);
Parque Kinc6n de la Vieja, Hacienda Santa Marfa,
I0°48'N, 85°19'W, 8(K)— 850 m. Herrera 672 (MO); beside
creek in pasture 8 km N of Rfo Naranjo, W of Tenorio
Volcano, 550 m, 8 June 1972, Lent 2559 (F); Fos Avotes,
near Tileran. 600—700 m. 21 Jan. 1926, Standley & Val¬
erio 45610 (LIS). PANAMA. Boeas del Toro: between
(Juebrada Higueron and Gutierrez, 18 Apr. 1968, Kirk-
bride & Duke 754 (MO. NY); Fortuna Dam region, 8°45'N,
82°15'W, l(MM) m, 8 Dec. 1985, McPherson 7H69 (MO.
NY). Ghiriquf: vicinity of Gualaca, 8.6 mi. from Planes
de Hornito. La Fortuna. 4000 ft., 9 July 1980, Antonio
5015 (MO); NF del Campamento Fortuna, Finca Pittf,
1(MM) — I2(M) m. If) Aug. 1976, Correa el al. 2525 (F. NY);
NO del campamento Fortuna, 8°45'N. 82°15'W, l(HM)—
12(H) in, 28 Sep. 1976, Correa el al. 2571 (MO. NY.
PM A); along road between Gualaca and Fortuna Dam site,
10.1 mi. NW of Los Planes de Hornito, 1800 m, 82°17'W,
8°45'N, 8 Apr. 1980, Croat 49815 (MO); 9 mi. from Rfo
Chiriquf Viejo bridge near Nueva California on road to
Rfo Sereno. 7 Apr. 1979, Hammel et al. 6822 (MO); road

over Fortuna Dam, N of lake, 8°45'N, 82°15'W. 1110-
1150 m, 27 Oct. 1985, McPherson 7400 (MO); Fortuna.
Potero de Pittier, 22 Feb. 1976, Mendoza 794 (F); Fortuna
Dam area, slope NW of confluence of Rfo Hornito and Rfo
Chiriquf, 8°44'N. 82°18'W. 1050-1100 m, II Nov. 1980,
Sytsma & Stevens 2267 (MO).

Acknowledgments. The illustration is by Bobbi
Angell. I thank the curators of F, MO, NY, and US
for allowing me to study their collections. I am
grateful to William D’Arcy (MO) and William Bur¬
ger (F) for their help. I also thank the librarians at
BKL and NY, especially Alan Cramer, for their
help.

Literature Cited

D’Arcy, W. G. 1979. Scrophulariaeeae. In R. F. Woodson
& Shery, Flora of Panama. Ann. Missouri Rot. Card.
66: 178-274. [Gibsoniothamnus pp. 220-227. |

New Taxa of Fuchsia (Onagraceae) from Northern and

Central Peru

Paul E. Berry and Elizabeth Hermsen

Department of Botany, University of Wisconsin, 132 Birge Hall, 430 Lincoln Drive, Madison,

Wisconsin 53706, U.S.A.

ABSTRACT. Two new taxa of Fuchsia from north¬
ern and central Peru are described and illustrated:
Fuchsia rivularis subsp. puhescens P. E. Berry &
Hermsen, a member of the large section Fuchsia
from Cajamarca, and Fuchsia mezae P. E. Berry &
Hermsen, a member of the apetalous section Hem-
sleyella from Huanuco.

Section Fuchsia is the largest of the 11 sections
of Fuchsia, with 65 species. In the latest revision
(Berry, 1982), several species were treated rather
inclusively, acknowledging aberrant populations,
but with too tew specimens or too little information
to place them in a different taxon. One of these was
F. rivularis .). F. Macbride, known primarily from
Amazonas in northern Peru, but with disjunct pop¬
ulations occurring across the dry Maranon River
valley to the west in Cajamarca. Now that addition¬
al collections have been received from botanists
from Cajamarca, we feel justified in describing the
western populations as a new subspecies of F. ri¬
vularis. We also describe a distinctive new species
of the apetalous section Hemsleyella from central
Peru.

Fuchsia rivularis subsp. puhescens P. E. Berry
& Hermsen, subsp. nov. TYPE: Peru. Caja¬
marca: trail from Chorro Blanco to San Andres,
Cutervo National Park, cloud forest, 2100—
2200 m, 6°10'S, 78°45'W, 15 Sep. 1991, ,4.
Gentry et al. 74844 (holotype, MO). Figure 1.

ffaec subspecies Fuchsiae rivulari subsp. rivulari sim-
ilis sed ab ea foliis pubescentioribus minoribus apice
emarginato acutove atque tubo floral i angustiore distin-
guitur.

Scandent shrub or liana to 10 m above ground,
with long, flexuous-areuate, mostly unbranched
shoots to several meters long. Young growth densely
puberulent, the branchlets with erect trichomes ca.
0.25 mm long, older branches with exfoliating
brown to red-brown bark. Leaves in whorls of 3 or
4, blades elliptic to obeli iptic, 3-6 X 1-3 cm, api¬
ces variable from rounded and mucronate to obtuse
or acute, bases acute to obtuse; upper leaf surface

green and sparsely pubescent, the lower surface
paler and densely pubescent in young leaves, with
erect trichomes ca. 0.25 mm long, older leaves with
trichomes persistent along the midvein and second-
arv veins of both surfaces, margins subentire to
glandular-denticulate, secondary veins 13 to 15 on
either side of the midvein, the midrib elevated on
lower surface; petioles stout, pubescent, 2-4 mm
long; stipules narrowly lanceolate, 1.5—2 mm long.
Flowers to 4 per node, pendent in the uppermost
leal axils; pedicels drooping, pubescent, 12-18 mm
long, 1—2 mm thick; floral tube narrowly funnel-
form, 47-58 mm long, 3-5 mm diam. at the slightly
bulbous base, narrowing to 2-4 mm wide above,
then widening again to ca. 8 mm wide at the rim,
pubescent outside, more sparsely pubescent inside;
sepal lobes 4, spreading at anthesis, narrowly lan¬
ceolate, 16—17 mm long, ca. 4 mm wide at the base,
the distalmost 2-3 mm free in bud, densely pubes¬
cent outside, glabrous inside; petals 4, glabrous,
narrowly elliptic, 16—18 mm long, 4—5 mm wide at
the midpoint. Stamens 8, the antesepalous ones 12-
13 mm long, the antepetalous ones 8-9 mm long;
anthers ca. 4 X 1.5 mm. Ovary cylindrical, ca. 7
X 2 mm; style sparsely pubescent and protruding
from the floral tube; stigma capitate, slightly 4-
lobed at the apex, 2-2.5 mm long, ca. 1.5 mm wide.
Unripe berries oblong, 8-12 mm long, 3—6 mm
thick, densely pubescent.

Ecology and distribution. Only known to occur
in cloud forest between 2000 and 2500 m elevation
in Cajamarca in northern Peru.

This new subspecies differs from subspecies ri¬
vularis by having short, dense indumentum on the
leaves and young stems, which is much less evident
in subspecies rivularis. Fuchsia rivularis subsp. pu-
bescens also has more slender flowers than subspe¬
cies rivularis, a more pilose style, and smaller,
emarginate to acute leaves. Fuchsia rivularis subsp.
puhescens is geographically separated from popu¬
lations of subspecies rivularis by the dry, low Mar¬
anon River valley, a well-known biogeographic bar¬
rier to mesic montane taxa in northern Peru.

Novon 9: 479-482. 1999.

480

Novon

Figure 1. Fuchsia rivularis subsp. pubescens P. E. Berry & Hermsen. — A. Habit, with details of leaf and stem
pubescence. — B. Flower at antliesis (inverted). — C. Detail of sepal, petal, and stamens from each of the two stamina!
whorls. — D. Upper part of style showing slightly 4-lobed stigma. — E. Longitudinal section through the ovary and
lower section of the floral tube, with enlarged detail showing the nectariferous region at the base of the tube. — F. A
typical whorl of leaves, with detail of pubescence and stipules. A— E drawn from Gentry el al. 74844, F drawn from
Sanchez 804.

Paratypes. PERU, Cajainarca: Jaen Province, Cola-
say District, Agua Fria 2 hrs. walk N and up from Colasay,
2(KX) m, (>°S, 79°10'W, 7 July 1993, Milanowski & Shonle
165 (MO); Prov. Cutervo, Distr. de San Andres, has Gru-
tas. 2350 m, 13 Oct. 1987. Sanchez 804 (MO); Prov. Cu¬
tervo, I,a Pucarilla, between Socata and San Andres, San¬
chez el al. 5926 (MO); Socata— San Andres, Ldpez &
Sagdstegui 5872 (MO); Achira, near Socota, Prov. Cuter¬

vo, Velarde 7065 (Z); El Suro, Prov. Cutervo, Velarde 7022
(Z).

Fuchsia mezae P. E. Berry & Hermsen, sp. nov.
TYPE: Peru. Hudnuco: Quebrada antes de
Utao, 2430 m, 4 Oct. 1965, I. Meza 360 (ho-
lotype, MO). Figure 2.

Volume 9, Number 4
1999

Berry & Hermsen

New Fuchsia Taxa from Peru

481

Figure 2. Fuchsia mezae P. E. Berry & Hermsen. — A. Habit. — B. Young leaf blade and detail of margin. — C. Intact
flower (left) and longitudinal section (right) in advanced bud stage before anthesis. — D. Detail of sepal, and stamens
from each of the two staminal whorls (note absence of petals). — E. Detail of nectariferous band at the base of the
inside of the floral tube. Drawn from Meza 360.

Frutex ea. 0.5 m altus, ramulis erectis. Folia opposita
vel ternala, juniora ovata pubescentia, 2-4 X 1 .5-2.0 cm.
petiolis junioribus 7-14 mm longis. Inflorescentia ex flo-
ribus axillaribus solitariis constans, pedicellis 15-20 mm
longis. Flores nondum maturi, apetali, tubo floral i usque
ad 40 mm longo, sepalis ante anthesin 10-13 mm longis.

Shrub ea. 0.5 m tall, stems erect and mostly bare
of leaves, to 5 mm diam., lower internodes 5—6 cm
long, leal scars opposite or ternate; new Hush of
leaves and flowers subterminal on short side
branches 1.5—4 cm long. Leaves mostly opposite,
young expanding blades densely pubescent with ±
erect whitish trichomes 0.5— 0.7 mm long, blades
ovate, 2^4 X 1.5—2 cm, with 7 to 9 secondary veins
per side, densely pubescent on lower side, sparsely
pubescent on upper surface, margins glandular-
denticulate; petioles 7—14 mm long; stipules brown,
triangular, 1 X 1 mm. Flowers axillary, preanthesal,
pedicels 15-20 mm long, the largest flower with
floral tube 40 mm long, 2-2.5 mm wide at the base,
gradually tapered to ca. 5 mm wide at the base of
the sepals, pink and pubescent outside, inside with
a nectariferous band lining the basal 3 mm of the

tube; sepals 10-13 mm long, basally connate in the
lower 1/2 to 1/3, green in the distal hall and broad¬
er than the floral tube, sepals broadly acute at the
apex, ca. 4 mm wide at their widest; petals 0. Sta¬
mens 8, the antesepalous filaments ca. 10 mm long,
the alternating series ca. 6 mm long; anthers ob¬
long, 3 mm long, 1.5 mm wide. Ovary narrowly cy¬
lindrical, 8—9 mm long, style pubescent in lower 1/3,
stigma clavate, 2 mm long, 1.5 mm wide. Fruit not
seen.

Ecology and distribution. Only known from the
type specimen, found growing along a streamside
at 2430 m in Huanuco, Peru, beginning to leaf and
flower in early October.

Although Fuchsia mezae is known from a single
specimen without fully expanded leaves or flowers,
it is clearly a new species belonging to section
Hemsleyella, which has 14 other recognized species
(Berry, 1985). This section is an unusual group of
species restricted to the tropical Andes and is char¬
acterized by its apetalous flowers, strongly seasonal
flowering (usually during the dry season), common

482

Novon

presence of tubers, and smooth floral nectaries lin¬
ing the base of the Horal tubes. Most of the species
in this section have alternate leaves, but three other
Peruvian species in the section have opposite
leaves, namely F. cestroides Schulze-Menz, F. huan-
ucoensis P. E. Berry, and F. tunariensis Kuntze.
Fuchsia cestroides has much smaller flowers and
leaves, and F. huanucoensis is glabrous, with fun-
nelform floral tubes, lanceolate sepals, and narrow¬
ly elliptic-ovate leaves. Fuchsia mezae is most sim¬
ilar to F. tunariensis but differs from it in its longer,
stiffer indument, narrower floral tubes anti sepals,
and more erect habit. Section Hemsleyella is char¬
acterized by the great rarity of many of its species,
some of them known from just one or a few collec¬
tions. For this reason we feel justified in describing

this species, albeit without a lull range of flowering
and fruiting specimens, because it is clearly dis¬
tinct from other members of the genus, and it calls
attention to the high degree of local endemism in
the mountains of central Peru.

Acknowledgments. This work was supported by
grant 3827-88 of the National Geographic Society.
The illustrations were prepared by Bobbi Angell.

Literature Cited

Berry, I’. E. 1982. The systematics and evolution of Fuch¬
sia sect. Fuchsia (Onagraceae). Ann. Missouri Bot.
Gard. 69: 1-198.

- . 1985. The systematics of the apetalous fuchsias

of South America, Fuchsia sect. Hemsleyella (Onagra¬
ceae). Ann. Missouri Bot. Gard. 72: 213—251.

Contributions Toward a Revision of Begonia Section
Weilbachia (Begoniaceae)

Kathleen Burt -Utley and John F. Utley

Department of Biological Sciences, University of New Orleans, New Orleans, Louisiana, U.S.A.

ABSTRACT. Characteristics of Begonia sect. Weil¬
bachia (Klotzsch & Oersted) A. DC. (Begoniaceae)
are discussed, and four new species in the section,
B. almedana , B. hondurensis, B. faustinoi, and B.
turrialbae, are described, illustrated, and compared
with other species in the section.

The genus Weilbachia was described by Klotzsch
and Oersted (1855) for those species in Begonia
with four staminate and three pistillate perianth
segments and bilocular ovaries with two parted
styles. Although Klotzsch and Oersted did not in¬
clude capsular orientation in their generic descrip¬
tion, both species placed in the genus, W. reptans
Klotzsch & Oersted and W. pustulata (Liebmann)
Klotzsch & Oersted, were characterized as having
nutant capsules. Alphonse De Candolle (1859) re¬
duced Weilbachia to a section within Begonia. Its
description was emended subsequently to include
species with apetalous staminate perianths, pistil¬
late flowers with two “lobes,” and nutant capsules
(De Candolle, 1864). Individuals of many species
within the section have pistillate flowers with two
or three styles (Burt-Utley, 1985), including B. pus¬
tulata , which De Candolle (1864) characterized as
having two styles. When three styles are present, a
stigma borne on one of the styles sometimes ap¬
pears smaller and less well developed than the oth¬
er two. Section Liebmannia Ziesenhenne was de¬
scribed for a species with apetalous pistillate
flowers, three styles, and bilocular ovaries (Ziesen¬
henne, 1976). Although Ziesenhenne (1976) dis¬
cussed the cane-like habit of the type species, he
chose not to include this character in his Latin di¬
agnosis and relied on floral characters exclusively.
Because style number is variable within a species
and section Liebmannia cannot be distinguished
from section Weilbachia using other floral or vege¬
tative characters (Burt-Utley, 1985), section Lieb¬
mannia is herein considered synonymous with sec¬
tion Weilbachia. A similar conclusion was recently
published by Doorenbos et al. (1998) relying solely
on plant habit.

Begonia sect. Weilbachia is the only neotropical
section having species with characteristically biloc¬

ular ovaries. With the exception of sections Gobenia
A. DC. and Tetrachia Brade, which have quadri-
locular ovaries, species in all other neotropical sec¬
tions have trilocular ovaries. Capsules of species in
section Weilbachia are also unusual among neo¬
tropical Begoniaceae in having a succulent or
fleshy outer layer that must erode before the cap¬
sules can dehisce (pers. obs.). In other neotropical
sections, as capsules mature they begin drying
without first shedding an outer layer of tissue and
then dehisce near the base of the loeules or rarely
through the wings. Section Weilbachia includes
about 25 species that are distributed from the states
of Veracruz, Puebla, and Jalisco in Mexico to north¬
western Panama. Of these, almost half are endemic
to Mexico. As a result of our continued research
with neotropical Begonia, a planned revision of
section Weilbachia, and treatment of the family for
Flora Mesoamericana, four species are herein pro¬
posed as new within section Weilbachia.

Begonia sect. Weilbachia (Klotzsch & Oersted)
A. DC., Ann. Sci. Nat. IV. 11: 134. 1859. Weil¬
bachia Klotzsch & Oersted, Abh. Konigl.
Akad. Wiss. Berlin. 1854: 239. 1855.

Begonia sect. Liebmannia Ziesenhenne, The Begonian
43: 64. 1976.

Begonia hondurensis Burt-Utley & Utley, sp. nov.
TYPE: Honduras. Atlantida: along Danto Riv¬
er, slopes of Mt. Cangrejal, 1200 ft., 2 Aug.
1938, T. G. Yuncker, J. M. Koepper & K. A.
Wagner 8773 (holotype, F; isotypes, BM, GH.
MICH, MO, NY, US). Figure 1.

Rhizomata repentia: internodia squamosa. Petioli in-
dumentis (1— )2— 4 mm longis dense squamosi. Florum 6
petala 4.5—8 X 1.4— 3.5(— 5) mm; stamina 22-49. Ovaria
bilocularia.

Herbaceous perennials; rhizomes repent, inter¬
nodes usually elongate and slender, rarely short.
0. 5-6.6 cm long X 1— 3( — 4) mm diam., glandular-
squamose with laciniate scales to 3 mm long; stip¬
ules persistent, narrowly ovate-triangular, 0.8-1. 7
cm X 4.5-9 mm, glandular-pilose, strongly keeled,
with the keel laciniate; petioles 6.5—23 cm long.

Novon 9: 483^189. 1999.

484

Novon

Figure 1. Begonia hondurensis Burt-Utley & Utley. — A. Habit with insert showing petiole scale. — B. Staminate
(lower. — C. Pistillate (lower. — I). Stigmas. — E. Capsule. (A, B, E from Yuncker el al. 8773: C, I) from Evans 1562.)

densely glandular-squamose with broad laciniate
scales ( 1— )2 — 4 mm long, often appearing glandular-
tipped; leal blades oblique to transverse, asymmet¬
rically ovate to obovate or reniform-elliptic, 6.8—
13. 8(— 14.6) X 4.5— 7 (—9) cm, basally cordate,
apically acute to attenuate-acuminate, marginally
sparingly ciliate and irregularly dentate to crenate
at the end ol major nerves, green throughout to red-
maculate at vein junctions and glandular to gla¬
brous above except for villi at the petiole-blade
junction, glandular-hirtellous beneath with tri-
chomes most dense on principal nerves, 8— 10-pal-
matinerved; inflorescences shorter than to exceed¬
ing the foliage, symmetric to weakly asymmetric,
laxly cymose, few-flowered; peduncles 10.5—25 cm
long, sparingly glandular-villous becoming glabrate
with age; bracts deciduous, the lowermost ovate to
narrowly elliptic, 5—8 X 2-5 mm, marginally en¬
tire, externally sparingly glandular, weakly keeled;
staminate flowers: pedicels 4.5—10 mm long; sepals
2, membranaceous, suborbicular to broadly trans¬
versely elliptic, 6—10 X 6.5-10 mm, sparingly
glandular and white to pink externally; petals 2,
oblanceolate, 4.5—8 X 1.4 — 3.5( — 5) mm; stamens
22^49; filaments 0.2— 1.3 mm long, borne on a
raised torus; anthers oblong to narrowly obovate, 1 —

1 .5 X 0/4-0. 9 mm; pistillate flowers: pedicels 5.5—

7.5 mm long; bracteoles wanting; sepals 2, subor¬
bicular to oblong-elliptic, 8.5—11 X 10-11 mm,
sparingly glandular and white to light pink; petal
only 1, obovate, 10 X 4-7 mm; ovary bilocular with
bipartite placentae, 5—8 mm long, sparingly glan¬
dular to glandular-villous; styles 3 or rarely 4, 1.5—

2.5 mm long, connate only briefly basally; stigmas
broadly lunate with ends beginning to spiral to bi-
cornute; capsules with pedicels 8-11 mm long,
strongly nutant, frequently somewhat beaked, thick
and fleshy, drying coriaceous, (9 — ) 10.5 — 12 mm
long; wings 3, unequal, the largest falcate to sub-
falcate, 14—17 X 5—8 mm, the others subequal, tri¬
angular, 3—7 X 8—10 mm.

Distribution. Known only from near La Ceiba
in the Department of Atlantida, Honduras, where it
has been collected between 140 and 365 m on
slopes along rivers and streams.

Begonia hondurensis is distinguished from all
other species in section Weilbachia by a combina¬
tion of characters including squamose internodes
and petioles, coupled with staminate and pistillate
llowers with petals. A squamose indumentum has
been observed on only one other species in the sec¬
tion, B. vestita C. DC. from southern Central Amer¬
ica. Although B. vestita clearly has bilocular ova¬
ries and capsules consistent with its inclusion in
section Weilbachia, it was listed in section Gireou-
dia (Klotzsch) A. DC. (Barkley & Golding, 1974;
Doorenbos et al., 1998) possibly because Casmir
De Candolle (1908) mistakenly characterized the
capsules as trilocular. Begonia hondurensis differs
from B. vestita in its petalous staminate and pistil¬
late flowers, as well as its capsules with narrow
falcate to subfalcate primary capsule wings (14—17
mm high, 5^8 mm wide). In contrast, primary wings
of B. vestita are asymmetrically oblong and (5-)8-
12 mm high and (4— )8 — 13 mm wide.

Volume 9, Number 4
1999

Burt-Utley & Utley
Begonia Sect. Weilbachia

485

Figure 2. Begonia almedana Burt-Utley & Utley. — A. Habit. — B. Staminate flower. — C. Pistillate flower. — I).
Stigmas. — E. Capsule prior to dehiscence. (From Utley & Utley 8425.)

Paratypes. HONDURAS. Allantida: Cordillera Nom-
bre de Dios, Ouebrada Grande on lower N slope of Pico
Bonito, Blackmon ■ & Charley 4178 (BM): base of N slope
of Pico Bonito, E of new CURLA camp building on Que-
brada Grande, ca. Vi km above its confluence with Rio
Bonito, 10 km SW of La Ceiba, Parque Naeional Pico
Blanco, 140 m, Keans 1562 (MO. NOES); Campamento
Quebrada Grande ca. 10 km SW of La Ceiba, at base of
N slope of Pico Bonito, 140 m, Liesner & Mejia 26016
(EAP. MO. NOLS, TEHF), Liesner & Mejia 26085 (MO);
Montana La Manga, 30 km SE de La Ceiba. Nelson, Var¬
gas & Alduvin 8299 (TEHF); Ceiba, Dyer A2I I (US).

Begonia almedana Burt-Utley & Utley, sp. nov.
TYPE: Mexico. Chiapas: cloud forest along
banks of Rfo Jol Uk’um, about 43.5 mi. S of
Palenque near 100 km marker and 7.8 mi. S
of turnoff to Cascadas Agua Azul on MEX 199
to Ocosingo, 1900 ft., 21 July 1989, J. Utley
& K. Utley 8425 (holotype, MEXU; isotypes,
B, BM, C, CAS, DUKE, F, MICH, MO, US).
Figure 2.

Rhizomata repentia; intemodia hirtella. Petioli hirtelli;
laminae 3— 4-lobatae. Florum 5 sepala 7.5— 11. 5(— 13) X
5.5— 9.5 mm; petala 8—10.5 X 4.5— 7(— 8) mm; stamina 30—
76. Ovaria bilocularia glandulosa hirtella sty 1 i duo.

Herbaceous perennials; rhizomes repent, elon¬
gate, generally branching, intemodes short to elon¬
gate, 0.3 — 1 ,6( — 4. 1 ) cm long X 3 — S( — 8) mm diam.,
sparingly glandular-hirtellous with trichomes 0.2—

0.4 mm long; stipules caducous to deciduous,
fleshy but becoming chartaceous upon drying, nar¬
rowly ovate-triangular, (0. 9-) 1. 3-3.1 cm X 4.5— 8.5
mm, marginally entire, glandular to very sparingly
glandular-hirtellous, strongly keeled with the keel
fimbriate and occasionally laciniate; petioles 10.6—

24.8 cm long, glandular-hirtellous with .short
spreading to ascending villi 0.3-0.8 mm long often
appearing glandular-tipped; leaf blades oblique or
rarely straight, asymmetric, suborbicular to broadly
obovate, transversely elliptic or rarely ovate, 9.3—

19.8 X 7.7-17 cm, basally very shallowly cordate
to occasionally deeply cordate, apically attenuate-
acuminate, rarely appearing bilobed with no dis¬
tinct apex, marginally serrulate-denticulate and
asymmetrically palmately lobed, usually with 3—4
major attenuate-acuminate lobes to Vi the blade
length, dark green and sparingly glandular to glan¬
dular-hirtellous above, pale green and sparingly
glandular-hirtellous beneath with trichomes most
dense on principal nerves, (6 — )8 — 1 9(— ll)-palmati-
nerved; inflorescences much shorter than to ex¬
ceeding the foliage, weakly to strongly asymmetric,
laxly cymose, few- to many-flowered; peduncles of¬
ten unusually short, (3 — )9. 1 — 23( — 3 1 ) cm long, glan¬
dular and very sparingly hirtellous; bracts caducous
to deciduous, the lowermost olten smaller than sec¬
ondary bracts, ovate to narrowly ovate, ovate-tri-

486

Novon

Table 1. Comparison of B. almedana

with B. ludicra.

B. almedana

B. ludicra

Stipule length (cm)

(0.9-)l. 3-3.1

0.6-1. 2

Petiolar pubescence (mm)

0.3-0.8

0.3-2

6 sepal size (mm)

7.5 — 1 1 .5(— 13) x 5.5— 9.5

(8-) 12-22 X (7-) 12.5- 18

6 petal size (mm)

8-10.5 x 4.5-7(-8)

1 1-16 X 3-4.5(-6)

angular or rarely transversely broadly ovate, to or-
bieular, 3-15.5 X 3.5— 7.5 mm, apieally atten-
uate-acuminate or infrequently rounded, marginally
entire, sparingly glandular, occasionally keeled;
staminate flowers: pedicels 8—14 mm long, glabrous
to sparingly glandular, sepals 2, fleshy, elliptic to
broadly elliptic, ovate-elliptic, obovate, or rarely
suborbicular, 7.5-1 1.5(-13) X 5. 5-9.5 mm, gla¬
brous to sparingly glandular, pink externally; petals
2, broadly oblanceolate, 8-10.5 X 4.5-7(-8) mm,
white, apieally frequently obcordate, rarely round¬
ed; stamens 30—76; filaments 0.3— 1.3 mm; anthers
narrowly oblong-elliptic to obovate, 1.6— 3.3 X 0.4—
0.5(— 0.7) mm; pistillate llowers: pedicels 6—16 mm
long, ebracteolate or often with 1 or 2 bracteoles to
3.5 mm long; sepals suborbicular to transversely
broadly elliptic, 7-13 X 7.5—15 mm, dark pink;
petal only 1, oblanceolate, 9.5—12.5 X 5—8 mm,
white; ovary 5.5-10 mm long, glandular to glan-
dular-hirtellous; styles 2, connate only briefly ba-
sally or rarely to over half their length; stigmas di-
lated-lunate to weakly bicornute; capsules with
pedicels (0.9— )1. 2— 2.5 cm long, strongly nutant, bi¬
locular with bipartite placentae, thick and fleshy
but drying cartilaginous, (0.9 — ) 1 . 1 — 1 .7 cm long;
locules 6.5— 10(— 1 1.5) mm long; wings 3, unequal,
the largest often falcate, narrowly triangular, oblong
or rarely broadly ovate, (0. 8-) 1.1-2. 2 X 0.9— 1.5
cm, the second and third subequal, triangular,
(0.5-)0. 7-1.2 X 0.8-1. 1 cm.

Distribution. Known from limited collections
from Chiapas, Mexico, in primary forests between
450 and 580 m elevation. It is common at the type
locality on forested banks of the Rfo Jol Uk’um.

Begonia almedana is most similar in its con¬
spicuously lobed leaf blades to a group of species
within section Weilbachia that includes B. purpusii,
B. ludicra, and B. confusa. It differs from the gen¬
erally erect stems of B. purpusii in its repent habit
and flowers with both sepals and petals. Although
both B. almedana and B. purpusii occur in Chiapas,
they are found in different physiographic regions
recognized by Breedlove (1973); B. almedana is
known only from the Gulf Coastal Plain at 450 to
580 m elevation, while B. purpusii occurs in the

Sierra Madre de Chiapas between 1100 and 2400
in, and extends into Guatemala and northwestern
El Salvador. Begonia confusa is only known from
immature flowers on the type collection from Gua¬
temala, but its densely villous sepals distinguish it
from B. almedana. Distribution and a suite of char¬
acters (Table 1) effectively separate B. almedana
from B. ludicra, which ranges from Puebla and Ve¬
racruz to the wet, northern slopes of the Sierra Ma¬
dre in central Oaxaca.

Begonia almedana is named in honor of Frank
Alrneda who first collected this unusual species
along with Dennis Breedlove.

Paratypes. MEXICO. Chiapas: Mpio. Ocosingo, 70
km SW ol Palenque on rd. to Ocosingo along the Rfo Jol
Uk’um, 550 m, Breedlove 47468 (CAS), Breedlove A' Al¬
rneda 48274 (CAS), Breedlove A Alrneda 48291 (CAS),
Breedlove A Keller 49588 (CAS), Breedlove 52589 (CAS);
Mpio. Chontal, rd. between San Cristobal de las Casas and
Palenque, 136 km NE of San Cristobal, Villa Parafso,
1 7°29'N, 92°05'W, 450 m, Harnmel. Martinez A Merello
15645 (MO); banks of Rfo Jol Uk’um 43.5 mi. S of Pal¬
enque on MEX 199 to Ocosingo, 1900 ft.. Utley A Utley
8356 (MEXU).

Begonia faustinoi Burt-Utley & Utley, sp. nov.
TYPE: Mexico. Chiapas: Mpio. Berriozahal, 13
km N of Berriozahal in vicinity of Pozo Tur-
pache, 900 m, J. Utley & K. Utley 8343 (ho-
lotype, MEXU; isotypes, B, C, CAS, DUKE, F,
GH, MICH, MO, NY, US). Figure 3.

Rhizomata repentia; internodia villosa. Uaminae ad
marginem duplicato-dentatae vel duplicato-serratae; petio-
li trichomatibus 1 .5— 3( — 4) mm longis villosi. Bracteae ad
marginem ciliatae laciniatae. Stamina 31—63. Ovaria bil-
ocularia villosa.

Herbaceous perennials; rhizomes repent, often
branching; intemodes slender, very short to elon¬
gate, 0.6—10.5 cm X 1.5— 3(— 4) mm, villous with
trichomes 1—3 mm long; stipules persistent, very
narrowly triangular to lanceolate, 8 — 1 7( — 23) X 3-
6(-8) mm, often strongly keeled, marginally entire,
glabrous to villous throughout; petioles (2.8— )7. 8-
13(— 21.5) cm long, villous with spreading tri¬
chomes 1.5— 3(— 4) mm long; leaf blades charta-
ceous, usually oblique, asymmetrically ovate, (5—)

Volume 9, Number 4
1999

Burt-Utley & Utley
Begonia Sect. Weilbachia

487

Figure 3. Begonia faustinoi Burt-Utley & Utley. — A. Habit with insert showing leaf margin. — B. Staminate flower.
— C. Pistil late flower. — I). Stigmas. — E. Capsule. (From l/tley & Utley 8343.)

7.5— 12.5(-16) X (3.5— )5-8(-10.6) cm. basally cor¬
date, apically attenuate-acuminate, marginally cil-
iate-denticulate to ciliate-serrulate and doubly den¬
tate to doubly serrulate, villous above and below
with trichomes to 3 mm long, (8— )9— 10(-ll)-pal-
matinerved; inflorescences generally shorter than
the foliage, cymose, few-flowered; peduncles (4 — )
6.2-13.5(-17.3) cm long, villous with trichomes
0.5— 3.5(— 5) mm long; bracts deciduous, the low¬
ermost occasionally navicular, elliptic to ovate, 5—
7( — 9.5) X 3.5—6 nun, marginally ciliate-laciniate,
villous; staminate flowers: pedicels (3.5— )7— 22(— 31)
mm long; sepals 2, fleshy but ehartaceous upon
drying, suborbicular to elliptic, ovate or obovate,
5 — 14(— 1 7) X (3— )5.5 — 14 mm, externally villous;
petals 2, narrowly obovate to oblanceolate, 5—13
(-17) X ( 1 .5 — )2.5 — 5( — 6) mm; stamens (20-)31-63;
filaments (0.4— ) 1—2 mm, inserted on a low torus;
anthers narrowly oblong to obovate, 0.8— 1.5(— 1.8)
X (0.3-)0.5—0.7 mm; pistillate flowers: pedicels 6-
1 1 .5( — 14) mm long; bracteoles wanting or if present
1 or 2 and borne on the pedicel, 1-2.5 mm long;
sepals 2, ehartaceous, suborbicular to broadly ob¬
ovate or transversely elliptic, 6— 12. 5(— 15) X 4.5—
13.5(-15) mm, externally villous; petal only 1 or
wanting, oblanceolate, 7-10 X 2-4. 5(— 7) mm; ova¬
ry bilocular with bipartite placentae, 5.5— 8.5
(-10.5) mm long, villous; styles 3, 2-3 mm long.

fused briefly basally; stigmas dilated-lunate or
weakly bicomute, but 1 often smaller than the other
2; capsules with pedicels 10—26 mm long, 9-14
mm long; locules narrowly oblong-elliptic, 6.5—10
X 1.5— 2.5 mm; wings 3, unequal, the largest asym¬
metrically ovate-triangular to oblong, 12-18 X 7-
10.5 mm, marginally eiliate, the second and third
subequal to each other, 3.5—5 X 8-12 mm.

Distribution and habitat. Known only from
Chiapas, Mexico, between 600 and 900 m eleva¬
tion, but should also occur in adjacent parts of
Guatemala. Begonia faustinoi is locally abundant
on karstic formations near Pozo Turpache.

Begonia faustinoi is most similar to B. pustulata
Liebmann, which is not known to occur in Chiapas.
These species share a similar villous indument, sta¬
minate and pistillate flowers with sepals and petals,
as well as nutant, bilocular capsules. Begonia faus¬
tinoi can be distinguished from B. pustulata by its
smooth upper leaf surfaces; those of B. pustulata
are consistently pustulate or bullate. These two spe¬
cies are also distinguished from each other by ad¬
ditional leaf characters and peduncle length sum¬
marized in Table 2.

This species is named in honor of Faustino Mi¬
randa who contributed significantly to knowledge of
the vegetation of Chiapas.

488

Novon

Table 2. Comparison of B. faustinoi

with B. pustulata.

B. faustinoi

B. pustulata

Blade length (cm)

(5-)7.5-l2.5(-l6)

(7.7-)1 2-23

Blade width at widest point (cm)

(3.5-)5-8(-10.6)

(4.6-)8-16.7

Blade margin

ciliate-denticulate to serrulate; doubly
dentate to serrate

ciliate-denticulate

Number of nerves

(8 — )9— I <>(— 1 1)

(9-) 11-14

Petiole length (cm)

(2.8-)7.8- 1 3(-2 1 .5)

(4.5-) 12.5-33

Peduncle length (cm)

(4 — )6.2— 1 3.5(— 1 7.3)

(8.5-) 15-39

Paratypes. MEXICO. Chiapas: Mpio. Berriozabal. 13
km N of Berriozabal. 1000 m, Breedlove & Smith 2 1 OS1)
(OS); 13 km N of Berriozabal near Pozo Turpache and
Finca El Suspiro, (XH) m. Breedlove & Dressier 297 1 1 (DS,
MO), Breedlove & Thorne 30856 (OS, MEXU, MICH,
MO), lltley & Utley 8747 (CAS, MEXU, MICH, MO, US);
Mpio. Palenque, near Agua A/.nl, 600 m, Breedlove & Al-
meda 57270 (CAS).

Begonia turrialbae Burt-Utley & Utley, sp. nov.
TYPE: Costa Rica. Cartago: about 7.2 km E
of Tuis and 5.3 km E of Platanillo along rd. to
Bajo Pacuare and Moravia, 2900 ft., J. Utley
& K. Utley 7261 (holotype, MO; isotypes, CAS,
CR, DUKE, F, US). Figure 4.

Rhizomata repenlia; intemodia villosa. Petioli villosi;
laminae pustulatae. Infloreseentiae paucifloribus. Florum

8: sepala 6.5—17 X 4.5—12 mm; petala 5.5—12.5 X 1 —
3.8 mm; stamina (19— )25-46. Florum 9: sepala 4.5— 8 X
4.5—9 mm; ovaria bilocularia; styli 3.

Herbaceous perennials; rhizomes repent, inter¬
nodes short to elongate, 0.3-5. 4 cm X 1-3.5 mm,
villous with trichomes to 2.5 mm long; stipules per¬
sistent, lance-ovate to lanceolate, 3.5— 6(-8.5) X
2.5^f(-5) mm, marginally entire, sparingly villous
throughout to glabrous; petioles 3— 7(— 14.5) cm
long, glandular and villous with spreading tri¬
chomes to 3 mm long; leal blades simple, oblique,
asymmetrically ovate to oblong, 3.5— 9(— 11) X 2.5—
5(— 7.8) cm, basally cordate with lobes often over¬
lapping, apically acute to acuminate, marginally
ciliate-denticulate, papillose and finely pustulate
above with trichomes 0.5— 2.5 mm long terminating

Figure 4. Begonia turrialbae Burt-Utley & Utley. — A. Habit. — B. Staminate flower. — C. Pistillate flower. — I).
Stigmas. — E. Capsule. (From Utley & Utley 7261.)

Volume 9, Number 4
1999

Burt-Utley & Utley
Begonia Sect. Weilbachia

489

pustules, glandular and villous throughout below
with triehomes to 1.3 mm long, (8— )10— 11-palma-
tinerved; inflorescences cymose, very lew-flowered;
peduncles 2.3-9. 7 cm long, glandular and sparing¬
ly to densely villous with triehomes to 1.5 mm long;
bracts caducous to deciduous, the lowermost ovate
to elliptic, 3— 3.5(-6) X 0.5— 1.5(— 3) mm, glabrous,
marginally entire; staminate flowers: pedicels 6.5-
16 mm long; sepals 2, chartaceous, elliptic to ovate
or obovate, 6.5—17 X 4.5—12 mm, externally vil¬
lous; petals 2, oblanceolate, 5.5—12.5 X 1—3.8 mm;
stamens (19-)25^46; filaments 0.7-1. 3 mm long in¬
serted on a conspicuously raised torus; anthers
narrowly oblong to obovate, 0.7— 1.6 X 0.3— 0.6 mm
diam.; pistillate flowers: pedicels 3.5—9 mm long;
bracteoles wanting; sepals 2, chartaceous, suhor-
bicular to broadly transversely elliptic, 4.5-8 X
4.5—9 mm, externally villous; petal only 1, oblan¬
ceolate, 5—5.5 X 1.5—2 mm; ovary bilocular with
bipartite placentae, 4—7.5 mm long, glandular and
villous; styles 3, 1.5—2 mm long, fused only briefly
basally; stigmas dilated-lunate with the ends just
beginning to spiral to weakly bicomute, often with
1 of the 3 reduced; capsules with pedicels 6—12
(—14) mm long, 7—11 mm long; locules 4.2— 6.5 mm
long; wings 3, unequal, the largest asymmetrically
triangular, 7-15 X 5.5—9 mm, marginally ciliate,
the remaining 2 subequal, asymmetrically triangu¬
lar, 2.5—5 X 5—9 mm.

Distribution and habitat. Known only from
moist, forested stream- and riverbanks between 600
and 900 m elevation in Cartago Province, Costa
Rica.

This is an attractive species with finely pustulate,
often silver-variegated and discolored leaf blades.
It is most similar to B. carletonii Standley, which
occurs along the Caribbean coast of southern Costa
Rica and western Panama. It may be distinguished
readily from this latter species by its finely pustu¬

late leaves and fine, spreading petiolar villi. Leaves
of B. carletonii examined all lacked pustules and
had a petiolar indument of stiff, strongly reflexed
villi.

Paratypes. COSTA PICA. Cartago: Turrialba. across
Rfo Reventazon from Interamerican Institute, 2000 ft..
DeWolf 194 (NY); between 9 and 12.5 km E of Inis or
1 .3 — 4.8 km W of Rajo Paeuare along rd. to Moravia de
Chirripo. 800-900 m, Utley 6 (114 (DUKE); about 2.3 km
E ofTuis, Calle Rival along Rio Tuis. Utley & Utley 7259
(NOUS). Utley & Utley 903b (NOUS).

Acknowledgments. Fieldwork was made possi¬
ble in part by grants from the American Philosoph¬
ical Society, Standley Smith Horticultural Trust,
and the Southwestern Region of the American Be¬
gonia Society. We thank the curators of the follow¬
ing herbaria for loans of specimens and use of fa¬
cilities essential to this research: B, BM, BR, C,
CAS, DUKE. F, G, GH. K, M, MEXU, MICH, MO,
NY, US, XAL.

literature Cited

Rarkley, E. A. & J. Golding. 1974. The Species of the
Begoniaceae, 2nd ed. Northeastern Univ., Boston.
Breedlove, I). 1973. The phytogeography and vegetation
of Chiapas, Mexico. Pp. 149-165 in A. Graham (editor).
Vegetation and Vegetational History of Northern Uatin
America. Elsevier Scientific, Amsterdam.

Burt-Utley, K. 1985. A revision of Central American spe¬
cies of Begonia section Gireoudia (Begoniaceae). Tulane
Stud. Zool. Rot. 25: 1—131.

Candolle, A. de. 1859. Memoire sur la famille des Begon-
iacees. Ann. Sei. Nat. IV. 11: 93—149.

- . 1864. Begoniaceae. Prodromus systematis natur-

alis regni vegetabilis 15(1): 266-408.

Candolle, C. de. 1908. Begoniaceae Novae. Bull. Herb.
Boiss. II. 8: 309-328.

Doorenbos, J., M. S. M. Sosef & J. ,1. E. E. de W ilde. 1998.
The sections of Begonia. Wageningen Agric. Univ. Pap.
98-2: 1-266.

Klotzsch, J. E. & A. Oersted. 1855. Weilbachia. Pp. 239—
241 in J. E. Klotzsch. Begoniaceen-Gattungen und Ar-
ten. Abh. Konigl. Akad. Wiss. Berlin 1854: 121-255.
Ziesenhenne, R. 1976. Begonia alice-clarkae Ziesenh.
The Begonian 43: 63—67.

Typification of Batis maritima L. (Bataceae)

Steve Cafferty and Alex K. Monro

Department of Botany, The Natural History Museum, London SW7 5BD,

United Kingdom

ABSTRACT. No valid type designation for Batis
maritima L. currently exists. A neotype for this
name is therefore selected here to fix the applica¬
tion of the name.

The family Bataceae comprises a single genus,
Batis P. Browne, which itself includes only two spe¬
cies, the generitype B. maritima, typified here, and
B. argillicola P. Royen, a recently described spe¬
cies from Papua, New Guinea. Batis maritima was
originally described by Patrick Browne (1756: 356)
with the polynomial, "Batis maritima erecta, ra-
mosa; foliolis succulentis, subcylindricis .” Linnaeus
(1759: 1289) also provided a generic-species de¬
scription adequate to validate his binomial, togeth¬
er with a reference to Browne. He appears to have
taken this specific epithet from Browne’s polyno¬
mial, and it seems likely that his knowledge of the
plant was based on Browne’s description. During
preparation of an account of Bataceae for Flora Me-
soamericana no valid type designation for the name
was found.

Kellogg (in Howard, 1988: 42) cited Browne’s
description as type, but this does not conform to
Article 8.1 of the International Code oj Botanical
Nomenclature (Greuter et ah, 1994), which states
that the type of a name must be a single specimen
or illustration. Although valid at the time it was
published under the then-current Code (Greuter et
ah, 1988), Kellogg’s designation is now invalid.
There is no accompanying plate in Browne’s work,
and no associated material has been traced in any
of the Linnaean herbaria. Material from the Sloane
herbarium (BM), which might be considered origi¬
nal material since Browne in his work refers to Slo-
ane’s Catalogue, cannot be considered eligible for
lectotypifieation since it was never studied by Lin¬
naeus. Neither would it make a suitable neotype

due to the restricted access of the collection. No
synonyms apart from that of Browne were cited by
Linnaeus, and it appears, therefore, that there are
no original elements for the name and that a neo¬
type is required. The name Batis maritima is in
current use and is taxonomically unambiguous. A
neotype is therefore selected here to fix the appli¬
cation of the name in a way that maintains current
usage.

Batis maritima L., Syst. Nat. (ed. 10) 2: 1289.
1759. TYPE: Jamaica. Middlesex: Clarendon
Parish, near Tarentum farm, 10 ft., 3 Jan.
1960, C. D. Adams 6063 (neotype, selected
here by A. K. Monro, UCWI).

Acknowledgments. The work of the first author
has been supported by the award of a grant by the
Linnean Society of London to the Linnaean Plant
Name Typification Project at The Natural History
Museum. We thank Patrick Lewis of the University
of the West Indies Herbarium for kindly lending
material.

I .iterature Cited

Browne, P. 1756. Civil & Natural History of Jamaica. T.

Osborne & J. Shipton, London.

Oreuter, W., H. M. Burdet, W. C. Chaloner, V. Demoulin,
B. Grolle, I). I,. Hawksworth, I). II. Nicolson, P. C.
Silva, F. A. Stafleu, E. G. Voss & J. McNeill. 1688. The
International Code of Botanical Nomenclature (Berlin
Code). Regnum Veg. 1 18.

- . F. R. Barrie, II. M. Burdet, W. G. Chaloner, V.

Demoulin, I). L. II awksworth, P. M. j0rgensen, I). H.
Nicolson, P. C. Silva, P. Trehane N J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Howard, R. A. (Editor). 1988. Flora of the Lesser Antilles,
4. Arnold Arboretum, Harvard University, Jamaica
Plain.

Linnaeus, C. 1759. Systema Naturae, ed. 10, 2. Salvius.
Stockholm.

Novon 9: 490. 1999.

New Species of Central American Araceae

Thomas B. Croat

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. Five Central American species of
Dieffenbachia, D. beachiana Croat & Grayum, D.
concinna Croat & Grayum, D. grayumiana Croat,
D. hammelii Croat & Grayum, and D. tonduzii
Croat & Grayum, as well as one species of Rho-
dospatha , R. pellucida Croat & Grayum, are de¬
scribed as new.

Because of the impending completion of the Flo¬
ra de Nicaragua, five names in Dieffenbachia and
one in Rhodospatha must be published here in or¬
der to be included in that work. Revisions of both
genera with additional new species (for Central
America only for Dieffenbachia ) are forthcoming.

Dieffenbachia beachiana Croat & Grayum, sp.
nov. TYPE: Panama. Bocas del Toro: along
road between Chiriquf Grande and Fortuna,
13.2 mi. W of Chiriquf Grande, 8°45'N,
82°10'W, 310 m, Croat & Grayum 60130 (ho-
lotype, MO-3230653; isotypes, K, PMA, US).

Planta terrestris, 40—100 cm; internodia 2—6 cm longa,
1.5— 3.5 cm diam.; petiolus 17—46 cm longus, vaginatus
1/4—8/10 longitudinis; vagina decurrens ad apicem; lam¬
ina anguste elliptica vel lanceolata, 16-41 cm longa, 6.5-
15 cm lata. Inflorescentia 1—3 in quoque axilla: pedun-
culus 9—13 cm longus; spatha 10—19 cm longa; spadix
usque 13—15 cm longus; pistila 48—66.

Slender herb 40-100 cm tall; stems briefly
creeping at base; internodes 2—6 X 1.5— 3.5 cm,
medium green to olive-green, sometimes streaked
with cream, semiglossy and obscurely roughened.
Leaves clustered toward apex of stem, erect-arch¬
ing; petioles 17-46 cm long (averaging 26.5 cm),
broadly C-shaped in cross section, flat with acute,
erect margins or sharply to bluntly sulcate adaxi-
ally, dark green to brownish, pale-green mottled,
matte, usually minutely roughened and with scat¬
tered whitish raphide cells visible, drying with scat¬
tered clusters of pustular raised areas with granu-
lar-puberulent progressions; unsheathed portion
(2.5— )10-30 cm long; sheaths decurrent distally,
10-22 cm long (averaging 14.2 cm), comprising
0.25-0.83% of the total length of the petiole (av¬
eraging 0.54% of petiole length); blades narrowly
elliptic to lanceolate, rarely narrowly ovate, 16-44
X 6.5—15 cm, 1.8— 5.3 times longer than wide, 0.7—

1.9 times longer than petioles, inequilateral, one
side 1-2 cm wider than the other, ± equilaterally
acuminate at apex (acumen to 5 mm long), slightly
inequilateral and acute, rounded or truncate at base
(rarely subcordate in Panama), thinly coriaceous to
subcoriaceous, drying papyraceous, usually quilted,
weakly bicolorous, crisp-undulate along the mar¬
gins; upper surface semiglossy (rarely matte), plain
dark green (in Costa Rica, but sometimes mottled
white or cream in Panama); lower surface slightly
paler, semiglossy to weakly glossy or matte below;
midrib flat-raised, often striate and usually conco-
lorous above, convex or thicker than broad beneath;
primary lateral veins 23-36 pairs, sunken above,
convex beneath, drying paler than surface, depart¬
ing midrib at an acute angle, then straight to weakly
curved to the margin at 70°-110° before sweeping
prominently toward apex (to ca. 45° toward apex
and sometimes at the very base, sometimes with
adjacent veins alternately reflexed and ascending);
interprimary veins usually present, not markedly
weaker than the primary lateral veins; midrib and
primary lateral veins puberulent on lower surface
with thick, whitish trichomes; minor veins moder¬
ately distinct and weakly raised beneath. Inflores¬
cences 1-3 per axil; peduncles 9—13 cm long;
spathe 10—19 cm long, uniformly green; gradually
long-tapered to apex, gradually contracted at mid¬
dle, which corresponds to the sterile section of the
spadix; tube 1-2 cm diam. when furled; blade to 3
cm wide when flattened; spadix to 13-15 cm long;
staminate portion of spadix 5—6.5 cm long, drying
4 mm diam.; male flowers 1.8— 2.6 mm diam., ir¬
regularly subrounded, drying smooth and light
brown at apex; pistillate portion of spadix 4.5-6 cm
long, drying 7 mm diam., usually separated from
the staminate portion by a slender, almost com¬
pletely sterile section 2.0-3. 5 cm long, drying 2
mm diam. (the staminate and pistillate portions
sometimes almost contiguous); female flowers 48—
66, 2-4 situated across the width of the spadix,
pistils oblong-ellipsoid, 1. 5-2.0 mm long, 1.0-1. 6
mm diam.; stigma subglobular, about as broad as
the ovary; staminodia narrowly elavate, usually not
at all fused at base, ca. twice as long as pistil.
Infructescence somewhat flattened, yellow-green

Novon 9: 491-502. 1999.

492

Novon

with darker decking; spathe (10— )14— 16 cm long,
mottled green and white, eventually turning oratige
in fruit; fruiting spadix 6—10 cm long; berries or¬
ange, subglobose, 6—8 mm diam.

Dieffenbachic i beachiana ranges from northeast¬
ern Costa Rica to western Panama (Bocas del Toro,
Chiriquf, and Veraguas) at elevations of 40-800 m
in tropical wet forest, premontane wet forest, and
premontane rainforest: In Costa Rica it occurs on
the Atlantic slope of the Central Cordillera and the
Cordillera Talamanca, ranging from the Sarapiquf
region to Tortuguero and Siquirres.

The species is characterized by its moderately
thin, quilted blades with many broadly spreading
primary lateral veins and crisped-undulate margins
with the lower midrib and primary lateral veins pu-
berulent.

Dieffenbachia beachiana is similar to D. galda-
mesiae Croat, sp. nov. ined., a species from central
Panama that also has somewhat puberulent midribs
on the lower blade surfaces. Diffenbachia galda-
mesiae differs in having the pubescence much
shorter, merely granular-puberulent, and has fewer
than 22 veins per side (vs. 23—36 for D. beachiana ),
which arise from the midrib at a 40°— 60° angle.
Dieffenbachia beachiana is also somewhat similar
to D. grayumiana , but that species differs in having
the veins merely minutely granular in slender rows
on drying and in having somewhat more ovate
blades.

Dieffenbachia beachiana initiates inflorescences
in the late rainy season with plants reaching an-
thesis in the dry season (January through April).
Immature fruits have been seen from April to Sep¬
tember, and mature fruits have been seen from July
to December.

The species is named in honor of James Beach,
who first brought it to my attention while he was a
Duke University student at the OTS Field Station
at La Selva.

Paratypes. COSTA HICA. Heredia: area between Rfo
Peje and Rfo Sardinalito, Atlantic slope of Volcan Barva,
10“ 1 8.5' IN, 84°04'W, 480-520 m, Grayum 6899 (MO); be¬
tween Rfo Peje and Rfo Sardinalito. Atlantic slope of Vol¬
can Barva, I0°17.5'N, 84°04.5'W, 700—950 m, Grayum
6986 (MO); between Rfo Peje & Rfo Gudcimo, trail from
main road across Quebrada Cantarana to Rfo Guacimo.
250 m, Grayum & Sclialz 8220 (DUKE): between San Jose
and Pto. Viejo, vicinity of Chilamate. 11.6 mi. N of Car-
iblanco, I0°27'N, 84°05'W, l(M) m. Croat 68858 (B, CR.
MO); Finea I .a Selva, near Rfo Puerto Viejo, ca. 2 km
upstream from confluence with Rfo Sarapiquf. 100 m.
10°06'N, 84°00'W, Beach 1448 (DUKE), Beach 1485
(DUKE, MO). Burger Stolze 5754 (CR, E), Hummel
8122 (DUKE, MO), Hummel 884 6 (DUKE); Zona Protec-
tora. N slopes of Volcan Barba, between Rfo Peje and Rfo
( hide i mo, 250 m. Grayum & Schulz 8220 (DUKE); Sara¬

piquf, Einca La Selva. 100 m. Croat 44288 (MO). Liiudn:
La Colombiana f arm. United Fruit Company, 70 m, Stan-
dley 86840 (US); between Rfo Pacuare and Quebrada
Diablo, ca. 2.5 km E of Siquirres, 10°06'N, 83°29’W,
1(H)— 200 m, Grayum et ul. 7698 (MO); Parque Tortuguero
Estacion Agua Erfa, 10°05'N, 83°34'W, 40 m. Babies
1284 (CR. MO). San Jose: Vazquez de Coronado, Braulio
Carrillo National Park, along San Jose to Siquirres Uwy.,
along trail to Rfo Sueio, site of the Old Carrillo Station,
6(M)-7(M) m, 1()°09'5()"N, 83°57'10"W, Croat 78771 (INB.
MO. W U). PANAMA. Bocas del Toro: 10 km SW of
Chiriquf Grande, 8°52'N, 82°I()'W. 3(M) m. Thompson
4987 (CM): Eortuna Dam— Chiriquf Grande, 7.3 mi. N of
bridge over Eortuna Dam. 3.2 mi. N of Continental Divide,
8°45'N, 82 I5'W. 7(M) m. Croat & Grayum 60255 (MO);
Eortuna— Chiriquf Grande Road, vie. of continental divide,
10 road-miles from continental divide and about 2 road-
miles along road E of hwy., 08°55'4"N, 82°10'4"W. 300
m. McPherson 1 1816 (MO), McPherson Aranda 10095
(MO); Eortuna— Chiriquf Grande, 7.7 mi. W of Chiriquf
Grande. 1.5 mi. Vi of Punta IVna, 8°50'N, 8°10'W, 80 m,
Croat <C Grayum 6 0094 (MO). CULTIVATED: Origin Cos-
la Rica, Henny 5 (MO).

Dieffenbachia concinna Croat & Grayum, sp.
nov. TYPE: Costa Rica. Puntarenas: along
road between Palmar Norte and Panamanian
border 3 km N of jet. to Rincon, 8°48'39"N,
83°16'18"W, 110 m., 10 Sep. 1996, Croat &
Hannon 79191 (holotype, MO-05096177; iso¬
types, AAU, B, CAS, COL, CR, DUKE, F, GH,
HUA, IMB, K. M, MEXU, NY, PMA, US).

Planta lerrestris, 0.6— 1.3 m; intemodia 1.5— 4. ()(— 7.5)
cm longa. 1-4 cm diam.; petiolus 7.5— 25(— 35) cm longus,
vaginatus 2/5— fere omnino; vagina 6.7— 13(— 20) cm longa;
lamina elliptica, ovata-elliptica, raro ovata aut late lan-
eeolata, 16— 36 cm longa, (7.8— ) 10-1 5. 5(— 20.5) cm lata;
nervis primariis lateralibus 6— 12(— 14) utroque; inflores-
centia 2— ft in quoque axilla; pedunculus (2.5— )8— 17.5 cm
longus; spatha I 1-25.7 cm longa; spadix 12-14 cm lon¬
gus; pistila 42—65.

Small erect fieri), 0.6— 1.3 m tall; sap milky, typ¬
ically lacking foul odor; internodes 1.5— 4.0(— 7.5)
cm long, 1-4 cm diam., medium green to dark
green, obscurely marbled with light to medium
green, glossy; petioles 7.5— 25(— 35) cm long (aver¬
aging 17 cm long), moderately erect, semiglossy or
almost matte, dark green to olive-green, blotched-
striate or streaked with whitish or silver-green, dry¬
ing greenish to sometimes light yellow-brown,
sheathed from lower 2/5 to nearly throughout (av¬
eraging 0.63 the length of the petiole); sheath 6.7-
13(-20) cm long (averaging 9.5 tan long), to 1.5 cm
high, the margins erect to involute, asymmetrically
auriculate to rounded and free-ending at apex, one
side obtuse to rounded, the other acute to obtuse,
the unsheathed portion 1—13 cm long, often solid
green or whitish, terete and obtusely sulcate near
the apex or subterete to C- or D-shaped, flat to

Volume 9, Number 4
1999

Croat

Central American Araceae

493

shallowly sulcate, the lateral margins rounded or
sometimes acute; leaves erect-arching; blades usu¬
ally — elliptic, ovate-elliptic, rarely ovate or broad¬
ly lanceolate, 16-36 cm long, (7.8 — ) 10 — 1S.5( — 20.5)
cm wide (averaging 23 X 12 cm), usually broadest
at or near the middle, 1.46—2.9 times longer than
wide (averaging 2 times longer than wide), 0.8— 2.4
times longer than the petiole, equilateral to slightly
inequilateral (one side sometimes to 1 .0 cm wider
than the other), equilaterally acuminate at apex (the
tip sometimes apiculate), broadly acute or usually
rounded at base, slightly inequilateral, moderately
coriaceous, weakly to moderately bicolorous, upper
surface solid dark green or rarely whitish to cream
or yellow-green-splotched or -flecked, semiglossy to
glossy, drying dark olive-green, dark gray-green to
yellowish brown above, lower surface moderately
paler and matte to weakly glossy, drying yellowish
brown to greenish, rarely dark gray-green; midrib
flat to broadly flattened-convex and concolorous to
paler above, sometimes pale in distal 1/3, low-con¬
vex to convex or rounded-triangular and paler to
slightly paler below, drying light brown to dark
brown but with the medial portion usually drying
even darker; primary lateral veins 6 — 12( — 14) per
side, spreading at 45°— 70° angle, sometimes arising
at an acute angle, then spreading, only rarely at a
different angle on opposite sides of the blade,
weakly sunken above, convex to obscurely raised
below, drying moderately inconspicuous and
scarcely raised, yellow-brown, paler than or darker
than surface, mostly concentrated in the lower 1/2;
interprimary veins sometimes present, never more
than 1 per pair of primary lateral veins, usually
much less evident than the primaries; minor veins
inconspicuous to visible and darker than surface.
Inflorescences 2-8 per axil; peduncles (2.5— )8— 17.5
cm long (averaging 10.5 cm long), somewhat dor-
siventrally compressed, drying 3—1 mm diam.;
spathe 11—25.7 cm long, 1.5— 2.0 cm diam. when
furled, constricted ca. 4.5 cm above the base, 4.5—
8.5 cm wide on tube when flattened, 1.8—3 cm wide
at constricted area when flattened, to 3.3 cm wide
on blade when flattened, as long as or up to twice
as long as the peduncle (averaging 1.6 times lon¬
ger), medium to pale green outside, somewhat dark¬
er on tube except whitish on open face, uniformly
paler within, drying dark brown to blackened; spa¬
dix 12—14 cm long, to 2 cm shorter than spathe,
the free portion 6.5— 9.5 cm long, protruding some¬
what forward at anthesis; stipe and spadix axis pale
green; pistillate spadix 4.5—9 cm long, 5—12 mm
diam., narrowed slightly toward the apex; staminate
portion of spadix white, 8—9 cm long, 7—10 mm
diam., tapered slightly toward both ends; sterile sta¬

minate portion 2.5^1 cm long, 5-7 mm diam.,
mostly bare but with a scattering of staminodia at
both ends (especially at base); fertile staminate por¬
tion of spadix 3.5— 6.5 cm long, 7 mm diam., broad¬
est at middle, tapered toward both ends; the flowers
irregularly rounded, 0.8— 1.4 mm diam.; female
flowers 42—65, pistils globose, pale green, 1.0-2. 5
mm diam., relatively closely spaced; stigmas pale
yellow; staminodia 1-5 mm long, up to 3 times lon¬
ger than the pistils, slender, somewhal flattened and
free from one another at base, tapering and some¬
what globular at apex, the tips to 1—2 mm diam.
Infructescence with spathe yellow-orange to bright
orange; spadix to 6 cm long; berries oblong-elliptic,
bright red to orange-red, ca. 1 cm long, 8 mm diam.

Diejfenbachia concinna ranges from southeastern
Nicaragua along the entire Atlantic slope and the
Osa Peninsula of Costa Rica from sea level to 200
m, mostly at less that 100 m, sometimes locally
common in rocky sites along streams. The species
occurs in tropical wet forest and premontane wet
forest.

The species is recognized by having all parts of
the plant appearing glossy (on live plants), as well
as by its relatively small stature, moderately cori¬
aceous, more or less elliptic, weakly inequilateral
blades, and auriculate petiolar sheath. It is most
easily confused with D. oerstedii but differs from
that species in its generally somewhat larger size,
and more coriaceous blades typically broadest at
the middle (vs. broadest below the middle in D.
oerstedii) with more numerous primary lateral veins
(9-14 vs. 4—11 for D. oerstedii). In addition, the
blades of D. concinna are glossy above and at most
rounded at the base, whereas they are typically
weakly glossy to matte and frequently subcordate
in D. oerstedii.

Flowering occurs mostly during the middle of the
rainy season, from July through October, especially
August and September, but may occur as early as
March. A cultivated collection at the Missouri Bo¬
tanical Garden flowered repeatedly over a two-week
period during late May and early June. Collections
with immature fruits have been seen from Decem¬
ber to March. Mature fruits occur primarily from
May through September, especially May to July, but
have been seen as early as March.

Grayum et al. 10588, from the headwaters of the
Rfo Piedras Blancas in Puntarenas Province, is un¬
usual not only in occurring much higher (to 900 m)
than most collections of this species but also in
having a broader leaf (to 19 cm wide) with a longer
petiole (to 30 cm). In other respects it agrees with
other collections of D. concinna.

494

Novon

Paratypes. COSTA RICA. Heredia: near Puerto Vie¬
jo along road near Rfo Sucio, 20 m, Croat 35702 (MO);
Finca La Selva, O l'S Field Station on the Rfo Puerto Viejo
just K of its junction with the Rfo Sarapiquf, 50-200 m,
Croat 78732 (INK. MO), Grayum 2840 (DUKE), Hummel
8784 (l)UKK, MO), llammel 0088 (I)UKK), Hummel 9772
(DUKE), Hummel 10081 (DUKE); near Puerto Viejo along
road near Rfo Sucio, 20 m, Croat 35088 (MO). Limdn:
La Colombiana Farm of the United Fruit Co., 70 m, Stun-
dley 30739 (US); headwaters of Quebrada Mata de Limdn,
westernmost fork. Finca Anai (Sixaola region), 09°34'N,
82°39'W, 20-25 m, Grayum et al. 4447 (MO); Parque
Nacional Tortuguero, Cerro Tortuguero, 1-119 m,
I0°35'N, 83°31'W, Barringer et al. 1973 (F); Cerro Tor-
tuguero junto a la Parra de Tortuguero, 10°35'N, 83°32'W,
60 m. Babies 2090 (CR, 1NB, MO); Cerro Tortuguero, 1 19
m, Gomez-ljuurito 7850 (CR); Ref. Nac. Barra del Colo¬
rado. Rfo Chirripocito— Rfo Sardina, Kef. Nac. Barra del
Colorado ("Sardinal" on Chirripo Atlantico quadrangle),
12 m. 10°38'N, 83°45'W. Grayum 9777 (CR. MO); Rfo
Reventazdn, Finca Montecristo below Cairo, 25 m, Stan -
dley & Valerio 48900 (US); Tortuguero National Park, near
Boca <le las lagunas de Tortuguero, l()°34'N, 83°32'W,
0-30 m. Burger & Antonio 11249 (F, MO); Canton de
Pococf, Refugio de Vida Silvestre Barra del Colorado, SE
base of Cerro del Tortuguero, 10°35'N, 83°31'30"W, ca.
0-00 m. Grayum et al. 1 1 139 (CM, CR, INB, MO, USJ).
Puntarenas: near Esquinas, along Panamerican 11 wy.,
Me A l pin 85—33 (SEL); Orotina— Jacd, valley ol Rfo Grande
de Tdracoles, I km S of Quebrada Ganado, 5 km S of
bridge over Rfo Agujas on road to Jacd, 09°42'50"N,
84°36'30"W, 50 m, Croat 79070 (INB, MO); Rfo Sandalo.
I — 10 in. Dodge & Goerger 10198 (F. MO), Dodge & (mer¬
ger 10020 (F. MO); Canton de Osa, Fila Costena, Fila
Cruces, headwaters ol Rfo Piedras Blancas, Cerro Angu-
ciana, 0B’49'02"N. 83° 1 1'25"W, 900-1000 in. Grayum et
al. 10588 (MO); Canton de Osa, Fila Huacas, ca. 4 km
NE of Panamerican Hwy. at I .as Huacas (Venecia) along
road to Sinaf, 08°52'00"N, 83°17'00"W, 400-500 m, Gra¬
yum & Brans 10156 (M0); Rfo Riyito (valley of Laguna
Chocuaco) and Quebrada Banegas, S of Cerro Rancho
Quemado (ca. 7 km W of Rincon de Osa), 8°41'N,
83°32.5'W, 200-300 m. Grayum et al. 7507 (MO); Rin¬
con— Puerto Jimenez, 40 km W ol Panamerican Hwy., 100
m. Gomez 19531 (MO); near airport, 4 mi. W of Rincdn
de Osa, Raven 21532 (F); along road from Panamerican
Highway to Rincdn, 10 km W ol Interamerican Hwy.,
08°45'08"N, 83°18'W, 2(K) m, Croat & Hannon 79107
(MO); Rincdn-Rancho Quemado, just S of Rincdn near
Rfo Rincdn. 08°41'21"N, 83°29'00"W, Croat & Hannon
79170 (IIUA, IMB, MEXU. MO, QCA, RSA, WU); Rin¬
cdn-Rancho Quemado, ca. 10 km W ol main Rincdn— Pto.
Jimenez Road, 8°32’30"N, 83°32'W, 150—260, Croat
Grayum 59 785 (CR, MO); Palmar Norte vicinity, ill) m,
Allen 5009 (E AP, US): hills above Palmar Norte, I ( H ) — 2(K)
m. Croat 35108 (MO); Quebrada Renjamfn near crossing
ol trail from Palmar Norte to Jalisco, R°58'00”N.
83°28'00"W, 160 m, Grayum et al. 9902 (MO); Osa Pen¬
insula, Nicolson 3393 (US); 6 km W of Rincdn, Grant &
Rundell 92-01928 (CR, MO, US); 4 mi. W of Rincdn de
Osa. 8°42'N, 83°3I 'W, Burger & Stolze 5401 (CR. F); Rin¬
cdn de Osa, ridge between Quebrada Aparicio and Q.
Aguabuena, 200— KM) m, 8°42'N, 83°31'W, Grayum et al.
3982 (CR. MO). 3983 (CR. MO): ca. 5 km W of Rincdn
de Osa, 8°42'N, 83°31'W, 50—200 m. Burger & Liesner
7190 (CR. F. PMA); vie. of Boscosa. at Quebrada Agua¬
buena. 50 m, 08°42'01"N, 83°30'48'W, Croat & Hannon

79238 (COL, 1MB, MO. PMA. EFH, TEX); Parque Na¬
cional, Sector Esquinas, vicinity ol Fila Gamba, hills be¬
hind Esquinas Rain Forest Lodge, along Quebrada Negra,
at end of side road off of Villa Bricefio to Golfito Road,
08°42'(K)"N, 83°12'30"W, 200-300 m, Croat A- Hannon
79291 (CM, DUKE, INB. MO. P, WU); Colfito-Villa Bri-
ceno Road. Rfo Cafiaza, 08°39'30"N. 83°12'W, 80 m, Gra¬
yum 9250 (MO); along shortcut road from Villa Bricefio
on Panama Hwy., W side of Villa Bricefio, ca. 6 km from
Golfito airport, less than 100 m, 8°41,30”N, 83°12'W,
Croat & Grayum 59930 (CM, K. MO); Golfito, Refugio de
Vida Silvestre. 08°39'50"N, 83°10'40"W, 30-KX) m, Mar¬
tin 789 (F); along highway to Golfito from Panamerican
Highway at Rfo Claro, 2.5 mi. SE ol Golfito, 8°36'N,
83°04'W, 60 m elev., Croat & Hannon 07594 (AAU, ( i VI .
DUKE, F, JAliM, MO. MY. NY, P. QCNE, SA, SCZ, SEL,
TEFH, VEN, US, W); Parque Nacional Coreovado, Lower
I xiokout Trail. 08°27-30'N, 83°33-38'W, 10-50 m, Ker¬
ri an 748 (CR); Monkey Woods, Kernan & Phillips 831
(CR); Estacidn Sirena, S of Rfo Sirena along Rfo Cama-
ronal, sea level. 08°28'N, 83°55'W, Knapp 2105 (MO);
Estacidn Sirena, 8°28'N, 83°35'W. 1-20 m. Pro. Quesada
51 (MO); up to I km W of Sirena, 8°29'N, 83°36'W, l.ies-
ner 2871 (CR. MO); Equinas Forest, 25 km Shi of Palmar
Sur, along Pan-Am Hwy., Croat 78318 (MO); Interameri¬
can Hwy., km 287, 2 km NW of Chacarita, 30 km S of
Palmar Sur, 8°48’N, 83°17'W, 20—40 m. Grayum & Flem¬
ing 8119 (MO); Canton Coto Brus, Guaymf Indian Res¬
ervation, Rfo Limoncito near junction of Villa Palacios
school trail. 680 m, Koshear 59 (CR). NICARAGUA. Rio
San Juan: Reserva Indio-Mafz, 10°45’N, 083°59'W. Rue-
da et al. 4070 (MO?).

DieffenbaeWia grayuiniana Croat, sp. nov. TYPE:
Costa Rica. Limdn; Refugio Nacional Barra
del Colorado, forests and pastures between Rfo
Chirripocito and Rfo Sardina (“Sardinal” on
Chirripo Atldntico Quadrangle), 10°38'N,
83°45'W, 12 m elev., Grayum 9773 (holotype,
MO-4370212; isotype, INB).

Herba, 1-1.5 m; intemodia 2.5— 3.0 cm longa, 2— 3.5(—
10) cm diam.; petiolus subteres, (24— )30— 59 cm longus,
vaginatus 0.3—0.55 longitudinis; vagina decurrens ad ap-
icem; lamina anguste ovata (22— )30— 54 cm longa, 10—32
cm lata; nervis primariis lateralibus 13— 1 8(— 22); inllores-
centia 1—3 in quoque axilla; pedunculus — 1 2 cm longus;
spatha 16.5-23.5 cm longa; spadix 15-27 cm longus; pis-
tilla 1 25— 1 Y>( ).

Stout herb, 1-1.5 m tall; stems erect at apical
part, the older portion reclining for up to 1.5 m;
internodes 2.5— 3.0 X 2— 3.5(— 10) cm, dark green,
glossy, variegated with cream-yellow or pale green
(sometimes medium green with dark green lines as
in Croat & Grayum 60149); petioles (24— )30— 59 cm
long, sheathed from 0.3—0.55 their length, matte,
dark olive-green to deep brown, streaked or with
cream-white in lower half, especially near the base,
subterete, weakly flattened on upper surface toward
the apex, sometimes slightly whitish adaxially and
continuing whitish onto lower midrib; sheath 19—
29 cm long, usually decurrent at apex, sometimes

Volume 9, Number 4
1999

Croat

Central American Araceae

495

weakly free-ending (as in Croat & Grayum 60149);
blades narrowly ovate to oblong-ovate, (22— )30— 54
X 10—32 cm, 1.5— 2.6 times longer than wide, rang¬
ing from 0.6 as long as petioles to 1.4 times longer
than petioles but averaging about as long as peti¬
oles, slightly inequilateral (one side 1—1.5 cm wider
than the other), acuminate to abruptly acuminate at
apex, inequilaterally subcordate (sinus to 2.5 cm
deep) or rarely inequilaterally rounded at base
(sometimes one side weakly subcordate, the other
side rounded), subcoriaceous, drying thin and dark
olive-green above, slightly coneolorous to slightly
bicolorous; upper surface usually glossy to semi¬
glossy, dark green, drying dark olive-green, plain
(unvariegated) to usually conspicuously mottled
yellow-cream; lower surface semiglossy to nearly
matte, slightly paler, plain to more weakly mottled
cream drying yellowish green; midrib flat to weakly
flat-raised or broadly convex (sometimes broadly
sunken at base, broadly flat-raised and striate to¬
ward apex) slightly paler and dark-green-spotted to
coneolorous above, convex to narrowly rounded and
whitish or narrowly acute and paler below; primary
lateral veins 13— 18(-22) per side, gradually arising
at a steep angle from the midrib, then spreading in
a broad curve at a 55°— 80° angle (those near the
apex to 25° angle, those near the base sometimes
90°— 110° angle and sometimes forming a sigmoid
curve), deeply sunken above, convex below, form¬
ing a series of weakly developed collective veins
that eventually merge with the margin; at least the
midrib sometimes minutely farinose-granular; mi¬
nor veins moderately obscure below. Inflorescence
1—3 per axil, sometimes subtended by a reduced
leaf with a fully sheathed petiole (the sheath emar-
ginate at apex) and a reduced leal blade 12—15 cm
long, 3.5— 6.5 cm wide; peduncle 8-12 cm long,
drying 2-3 mm diam.; spathe 16.5—23.5 cm long,
3—4 cm longer than the spadix, 4.0— 5.0 cm wide at
base, to 2.5— 3.5 cm wide at constriction, 2.5—3 cm
wide on blade, up to 7 cm wide when flattened,
uniformly light green to medium green on both sur¬
faces, weakly glossy throughout outside, somewhat
glossier within; spadix 15—27 cm long; pistillate
portion 8.0— 9.3 cm long, drying 7-10 mm wide;
staminate portion 5—6.5 cm long, white, tapered to¬
ward apex and slightly toward the base; the sta¬
minate and pistillate portions separated by an al¬
most sterile segment 2-2.5 cm long with a few
scattered pistils in the lower half and a few scat¬
tered staminodia in the upper half; male flowers
1.2— 1.6 mm diam., irregularly angled with rounded
margins; female flowers ea. 125—160, pistils dense¬
ly aggregated, separated from one another by 1/2
to 2.5 times their width, irregularly scattered with

roughly 4—5 covering the width of the spadix, pale
lemon-yellow, depressed-globose, 1.8—2 mm wide;
staminodes white, 3—5 per pistil, free or briefly
united at base, 1.5— 2.2 mm long. Infructescence to
27 cm long, with fruiting portion 10—14 cm long;
berries red-orange, subglobose, ovoid to ellipsoid,
5—7 mm diam.

Dieffenbachia grayumiana ranges along the At¬
lantic slope from northwestern Costa Rica to west¬
ern Panama (then disjunct to northern Colombia),
mostly near sea level but ranging up to 1300 m in
Costa Rica in Alajuela Province (vicinity of Mon-
teverde Reserve) and to 480 m in Heredia Prov¬
ince. It occurs in areas of tropical wet forest, pre-
montane wet forest, and premontane rainforest.
Grayum (pers. comm.) reported the species to occur
mostly in light gaps and disturbed areas in primary
forest at the La Selva Reserve in Costa Rica.

The species is characterized by its narrowly
ovate, typically subcordate, mottled blades, weakly
sheathed, decurrent petioles, and variegated stems
and petioles. Also characteristic are blades that are
frequently glossy on the upper surface and matte
or nearly so on the lower surface. The major veins
are sometimes minutely farinose-granular on the
lower surface. In this regard it is perhaps related
to D. beachiana, a species with puberulent major
veins on the lower blade surface. Aside from pu¬
bescence type, Dieffenbachia beachiana also differs
in having much narrower blades (1.8— 5.3 times lon¬
ger than wide) with 23-36 pairs of primary lateral
veins.

A collection from Bocas del Toro Province in
Panama (Croat 74945) differs in having larger con-
colorous blades (54 X 30 cm), which are only
weakly glossy on the upper surface, with weakly
sunken primary lateral veins. Another colleetion
from Bocas del Toro, Croat & Grayum 60149, hail
stems to approximately 10 cm in diameter, sub¬
stantially larger than those of any other collection
reported.

Paratypes. COLOMBIA. Choco: ea. 10—15 km S of
Qnibdo on road to Istrnina (Panamerican Hwy.), and fi — I < )
km K on road to petroleum exploration camp, 5°35'N,
76°37'W, 00 m, Grayum et al. 7657 (MO); Hfo Nercua,
Quebrada Barrial— Quebrada Ainbrosio, near Camp Curi-
ohe, Duke 11471 (NY, OS); Rio Truando, Schott 6 (NY).
COSTA RICA. Alajuela: Upala Road, 5 km S of Canalete.
near Rio Zapote, I0°48'N, 85°02'W, 100-200 m, Burger
& Baker 9968 (CR, F); Monteverde Reserve, 1300—1400
m, Dryer 1681 (F); Monteverde, upper Rio Penas Blancas,
9° 1 7'N, 84°86'W, 1250— 1350 m. Burger et al. 10734 (CK.
F). Heredia: Volcan Barva, area between Rio Peje & Rio
Sardinalito, I0°18'30"N, 84°04'W. 480-520 m, Grayum
6887 (MO); Rio Sarapiqui, jenseits von La Virgen, 200-
250 m, 1974. Cl. Horich s.n. (M); San Jose-Puerto Viejo,

496

Novon

vie. Chilamate, I 1.0 mi. N of Carihlanco, 1()°27'N,
84°05'W, 100 m, Croat 68.359 (B, MO); La Selva Biolog¬
ical Station, 10°26'00"N. B4o0L00''W. 50-150 m. Beach
1440 (DUKE), Beach 1441 (DUKE), Burger & Malta 4181
(CD. F), Croat 44226 (MO), Croat 44.321 (MO), Croat
787.3.3 (IMS. MO), Folsom 9724 (MO), Grayum 2288
(DUKE), Hammel 8167 (DUKE), Hummel 8617 (DUKE),
llummel 8885 (DUKE), Hummel 8872 (DUKE), I). Smith
597 (DUKE); N slopes of Volcan Barva, between Rfo Beje
& Rfo Guucimo. Gruyum & Schutz .3206 (DUKE). / Cha-
cdn 507 (DUKE, MO); Secondinos Woods, off Occidental.
Kress 84—1622 (SEE), I.imon: Refugio Gandoca-Manza-
nillo, along Rfo Gandoca, 9°35"N, 82°36'W, 0-5 m. Gra¬
yum et al. 80.38 (MO); Guapile.s-Guacimo, vie. Gudcimo,
10°12'N. 83°43'W, 4(H) m. Barringer 2422 (E); Cerro Cor-
onel. E of Laguna Danto. 10°41'N, 83°38'W, 20-170 m.
Stereos 22642 (CR. MEXU. MO); Refugio Nacional Barra
del Colorado, between Rfo Chirripocito & Rfo Sardina,
10°38'N, 83°45'W, 12 m, Gruyum 9820 (MO); Pococf
Cantbn, Barque Nacional Tortuguero, along Rfo Suerte be-
tween Palacios N Cano Penitencia. 10°33'N, 83°33— 35'W,
10 m, Gruyum et al. 1 1 1 16 (INB. MO), 10°29'N, 83°31'W.
tf-25 m. Gruyum el al. 11163 (IMB. MO). Puntarenas:
Rfo Biedras Blancas, Cerro Angueiana. 08°49'02"N,
83° I I '25''W, <KK)-1(KH) m. Gruyum 10588 (CR. INB,
MO). BANAM A. Boras d«4 Toro: W of Chiriquf Grande,
< 50 m. M. Akers 78A (MO); Gualaoa-Chiriquf Grande,
1.4 mi. S ol Bunta Berta, 0°55'N, 82°1 l'W, 2(H) m. Grout
74945 (MO); Chiriquf Grande— Fortuna, 3 mi. W of Chi¬
riquf Grande, 8°47'N. 82°09'W, ca. 10 m, Croat dc Gru¬
yum 60149 (MO); along railroad between Almirante and
Chanquinola, Station Milla 5. 50 m, Croat & Barter 16499
(MO).

DiefTenhachia liaiiinielii Croat & Grayum, sp.
nov. TYPE: Costa Rica. Heredia: Finca La Sel¬
va, the O.T.S. Field Station on the Rfo Sara-
piquf, 50-80 m elev., Grayum 7670 (holotype,
MO-3491533; isotypes, B, CR, K).

Herba plerumque 25-40 cm raro usque 70 cm alta;
intemodia 3.5— 7.5 cm longa, 0.5— 2.0 cm diam.; petiolus
7— 25( — 35) cm longus, anguste C-formatus, non profunde
sulcatus adaxialiter; lamina elliptica aut oblanceolata;
costa plana et elevata adaxialiter; nervis primariis later-
alibus 9—17 utroque; inflorescentia 1-3 in quoque axilla;
pedunculus 4.5—11 cm longus; spatha (8— )10-12.5 cm
longa. Baccae aurantiacae.

Small herb, mostly 25—40 cm (rarely to 70 cm
tall); sap not foul-smelling; stem becoming decum¬
bent and subrhizomatous at base; internodes 1-1.5
cm long on lower portions, 3.5— 7.5 cm long toward
apex, 0.5— 2.0 cm diam., glossy, drying yellowish
brown to gray-green, weakly striate; petioles 7-25
(—35) cm long (averaging 15 cm long), sheathed
(0.3— )0. 4-0. 8 their length, plain green or mottled
with dull yellow-green, drying matte, greenish or
sometimes to yellowish brown, erect; sheath 5.5—
13 cm long, narrowly decurrent at apex and some¬
times difficult to discern where it ends; unsheathed
portion 1—9 cm long (rarely to 15 cm), narrowly C-
shaped to D-shaped, shallowly suloate adaxially.

the margins finely acute; blades erect -arching, nar¬
rowly to broadly elliptic or oblaneeolate, slightly
inequilateral (one side 0.5— 1.0 wider) (the narrower
side usually plane, the wider side usually minutely
undulate), 9-28 X 3.2-13(-15) cm, 1.7-3. 4 times
longer than wide (averaging 2.7 times longer than
wide), acuminate at apex, narrowly to broadly acute
or attenuate and equilateral or slightly inequilateral
at base, rarely obtuse, thinly coriaceous, slightly to
moderately bieolorous; upper surface plain medium
green to dark green and weakly glossy to semi¬
glossy, drying light yellow green to dark brown; low¬
er surface slightly to moderately paler, weakly
glossy to essentially matte, drying usually yellow-
green, sometimes yellow-brown; midrib Hat-raised
or rarely convex (sometimes obscurely sulcate me¬
dially) and slightly paler above, convex to narrowly
convex below, 1-3 mm diam., drying pale brown,
matte, with short, white raphide cells; primary lat¬
eral veins 9—17 per side, moderately obscure, aris¬
ing at 40°-50° angle, weakly curved to the margins,
weakly sunken or slightly raised in weak valleys,
weakly quilted and concolorous above, convex and
weakly pleated-raised below, moderately inconspic¬
uous on drying, slightly darker or slightly paler
than surface; interprimaries about as conspicuous
as the primary lateral veins; minor veins and cross¬
veins clearly visible, and darker than lower surface,
moderately faint on drying. Inflorescences 1-3 per
axil; peduncles 4.5—11 cm long, subterete; spathe
(8 — )10 — 12.5 X 1.2-1. 5 cm, to 3 cm wide when
flattened. 1.4— 2.7 times longer than peduncle, long-
attenuate above tube, cuspidate to acuminate at
apex, uniformly medium-green outside and inside;
spadix 6.5—12.5 cm long; staminate portion mod¬
erately acute at apex, tapered to both ends, (1.7—)

3- 4.5 cm X 6-7 mm (dried); male flowers bluntly

4— 5-sided, irregularly shaped toward the base,
more nearly rounded toward the apex, 2—2.5 mm
diam.; pistillate spadix 5-8.5 cm long (fused with
spathe for up to 4 cm), 6-7 mm wide; female flow¬
ers 26—31, sparse, separated from the staminate
portion by a sterile section (0.8-)2.0— 3 cm long,
this with only a few scattered sterile male flowers;
pistils ovoid, loosely scattered in clusters of 2-4
with up to three across the width of spadix but often
with spaces between the groups of pistils up to
twice the width of the spadix; staminodia up to
twice as long as pistils, free from one another at
base and drying flattened and very thin. Infrue-
tescence with spathe pale orange, to 22 cm long;
berries orange to bright red, oblong-ellipsoid, ca. 1
cm long, 8 mm diam.

Southeastern Nicaragua (Depto. Rfo San Juan)

Volume 9, Number 4
1999

Croat

Central American Araceae

497

and northern Costa Rica from sea level to 100 m
in the tropical wet forest life zone, in wet forests
and swampy areas on the Atlantic slope.

The species is characterized by its small stature,
glossy, decumbent, subrhizomatous stems, weakly
sheathed, matte-drying petiole (decurrent at apex),
and moderately small, more or less oblong-elliptic,
weakly inequilateral plain green leaf blades with
moderately numerous primary lateral veins, and
with one blade margin plane and one minutely un¬
dulate.

In Costa Rica Dieffenbachia hammelii is most
easily confused with D. oerstedii or smaller plants
of D. grayumiana, both of which can be distin¬
guished in having the petiolar sheath aurieulate at
the apex rather than decurrent as in D. hammelii.
At La Selva, where the species was first studied,
D. hammelii is uncommon. One collection (Hammel
8784) reported that the sap was not foul-smelling,
although many species in the genus are.

Flowering plants have been seen from March
through May while mature fruits have been seen in
September. Cultivated plants at the Missouri Bo¬
tanical Garden flowered in mid July and mid Oc¬
tober.

Paratypes. COSTA RICA. Heredia: La Selva biolog¬
ical Station, 10°26'N. 84°01'W, 50-200 m. /. Charon 558
(F). Croat 78731 (INR. MO), Folsom 9329 (DUKE), ( an¬
vil ni 2772 (DUKE, MO), Hammel 8212 (DUKE), Hammel
8273 (DUKE), Hammel 8748 (DUKE. E. MO). Hammel
12320 (MO). Linion: Cerro Coronel, E of Rfo Xapote.
within 1 km of Rfo Colorado, 10°40'N, 83°40'W, 10-40
m, Sterens et al. 24699 (CR. MO), Stevens 24257 (CR.
MO); 2—3.5 airline km SSE of Islas Buena Vista in the
Rfo Colorado, 14-10 airline km SW of Barra del Colorado,
10°40'l\, 83°40'W. 10—120 m. Davidse A Herrera 31077
(CR. MO). Davidse & Herrera 31213 (MO): Refugio Barra
del Colorado, Rfo Chirripocito— Rfo Sardina (“Sardinal" on
Chirripo Atlantieo quadrangle), 10°38'N, 83°45'W, 12 m.
Grayum et al. 9744 ((41, MO); Tortuguero National Park,
near Boca de las Lagunas de Tortuguero, I0°04'N.
83°32'W, 0—30 m. Burger & Antonio 11224 (E); Canton
de Pococf. Parque Nacional Tortuguero, northern terminus
of Lomas de Sierpe, S from Rio Tortuguero, 10.29N,
83.31.30W, ca. 8-25 m, Grayum et al. 1 1 169 (CR, INB.
MO). San Jose: Vkzquez de Coronado, 600—700 m, Brau-
lio Carrillo National Park, San Jose— Siquirres Hwy.. trail
to Rio Sueio. site of the Old Carrillo Station, 10°09'50"N.
83°57'10"W, Croat 78758 (INB, MO. WU). NICARAGUA.
Rfo San Juan: near Cano Chontaleno, 20 km NE of El
Castillo, 200 m. Neill A V incell i 3484 (MO).

Dieffenbachia tonduzii Croat & Grayum, sp. nov.
TYPE: Panama. Bocas del Toro; vie. Valle de
Silencio, along Rfo Changuinola, ca. 1 km
above mouth of Rfo Teribe, vie. Teribe Indian
population, disturbed forest among cacao plan¬
tations, 9°21'45"N, 82°31'45"W, <100 m, 25
June 1994, Croat & Zhu 76452 (holotype,
MO-04611212; isotypes, AAU, B, CAS, COL,
CR, DUKE, F, GH, HUA, IMB, K, M, MEXU,
NY, PMA, SCZ, US, YEN, WU).

Herba 0.5— 1.5 m alta; internodia 1.5— 4.5(— 6) cm longa,

1.5— 3(— 4.5) cm diam.; petiolus 10—24 cm longus, vagin-
atus ubique, vagina rotundata inaequaliter apice, saepe
imbricata laminae basi; lamina inaequilatera, ovata vel
ovato-elliptica aut oblanceolato-elliptica, (15— )25— 48(-63)
cm longa, (8— )15— 32 cm lata, basi plerumque inaequaliter
subcordata. Inflorescentia 1-4, pedunculus (1 — )6 — 1 7 cm
longus, usqua ad 6 cm diam. (in sicco); spatha viridis,
cum tubo 2—2.5 cm diam. (in sicco); spadix pistillatus

8.5— 9.5 cm longus; cum parte pistillata aggregata praeter
summum 1—1.5 cm.

Terrestrial herb 0.5— 1.5 m tall, usually to less
than 1 m tall; internodes usually .solid dark to me¬
dium green, sometimes faintly marbled with gray-
green or yellowish gray throughout (on plants that
also have streaked petioles), initially weakly glossy,
becoming semiglossy to glossy, often with a sub-
velvety sheen, 1.5— 4.5(— 6) cm long, 1.5— 3(-4.5) cm
diam., the petiole scars manila to whitish, curved
downward on the opposite side of the stem and end¬
ing unevenly; petioles 10—24 cm long (averaging
17.6 cm long), held ± erect, medium green (except
sometimes white to pale green at base), almost
matte to weakly glossy, weakly striate (especially
near the base), narrowly rounded to obtusely an¬
gular on abaxial surface and often whitish medially,
sometimes streaked in a variegated pattern
throughout (this whitish coloration continuing onto
the midrib), sheathed virtually throughout; sheath
erect to involute (rolled inward throughout in age),
free-ending and unequally rounded at apex, pro¬
longed to 2 cm beyond the base of blade; un¬
sheathed part obsolete or rarely to 1 cm long (when
evident obtusely flattened); blades inequilateral
(one side 1-3.0 cm wider), ovate to ovate-elliptic
or oblanceolate-elliptic, (15— )25— 48(— 63) cm long,
(8— )15— 32 cm wide, (averaging 34 X 16 cm),
broadest near the middle, sometimes below, fre¬
quently above the middle, 1 ,4-2.5(-3.5) times lon¬
ger than wide (averaging 2.1 times longer than
wide), 1.9—3 times longer than petioles, spreading
to erect-spreading, gradually to abruptly acuminate,
sometimes acute at apex, inequilaterally cordulate
at base, one side sometimes broadly rounded to ob¬
tuse, the other side cordulate, sometimes inequi¬
laterally acute, subcoriaceous, often conspicuously
quilted, moderately bicolorous; upper surface usu¬
ally solid dark to medium green, sometimes con¬
spicuously to sparsely variegated with pale green
or pale yellow throughout much of the surface, the
mottling large or small, but somewhat restricted to
the area midway between the midrib and margin,
matte to weakly glossy, sometimes appearing weak¬
ly velvety, drying gray-green to olive-green or dark
brown; lower surface much paler and matte to weak¬
ly glossy, silvery-green, drying yellowish green to

498

Novon

yellow-brown below; midrib Hat to broadly rounded
and moderately to strongly paler, pale green or
sometimes creamy white above (sometimes only to¬
ward the apex), bluntly acute to obtusely angular
and paler, sometimes whitish or creamy white be¬
low, (0.6-)l— 1.7 cm wide; primary lateral veins
(14— )18— 25(— 30) per side, arising at an acute angle
with the midrib and spreading at 45°-90°, some¬
times reflexed toward the base, prominently to
weakly and obtusely sunken above, convex to
weakly raised and darker than surface or coneolo-
rous below, some of the lowermost with a weak fold
near the base (Croat & Grayum 60112), sometimes
convex-pleated below; interprimary veins almost as
conspicuous as the primaries; minor veins moder¬
ately to distinctly visible, darker than surface be¬
low. Inflorescences 1 — 2( — 4) per axil, often with two
oriented in opposite directions; peduncle (1 — )6 — 17
cm long (averaging 10.3 cm long), 7—8 mm diam.,
weakly glossy, dark to medium green, sometimes
with pale yellow-green streaks; spathe (12— )15-28
cm long (averaging 20 cm long), 2^1 times as long
as peduncle, acuminate at apex, convolute to about
the middle in lower part, matte to weakly glossy
outside, glossy within, solid medium green on both
surfaces, gradually and weakly constricted some¬
what above the middle; spathe tube 1.5— 3.0 cm
diam. when closed, 6.8-9 cm wide when flattened;
constricted area 4.3^L5 cm wide, flattened spadix
(9— )16.5— 25.5 cm long, about as long as or up to
3.0 cm shorter than the spathe; free portion 7.8-

8.5 cm long; pistillate portion (4.8— )6— 11.5 cm
long; mostly sterile portion sparsely flowered to na¬
ked (rarely lacking, as iti Croat 70768), 1.8— 1.0
(—8.5) cm long with 0.6— 1.5 cm totally bare, the
uppermost portion with a few staminodia, some¬
times with a few scattered staminodia throughout,
tin' lower half sometimes with an occasional pistil
and much reduced staminodia, rarely with the fe¬
male flowers ± equidistant and nearly contiguous
with staminate part; fertile staminate portion (4—
)5.5— 10 cm long, 7—10 mm diam., slightly broader
midway, weakly tapering to apex and base, bluntly
pointed at apex; male flowers 5-6 per spiral, ±
rounded in outline, erenulate along margins, trun¬
cate at apex; sterile male flowers irregularly
shaped, 1.8— 2.5 mm diam.; pistillate portion of spa¬
dix to 11 cm long, 9—10 mm diam.; female flowers
(15— )48-62, closely aggregated except in the upper

1.5 cm of spadix, 4—5 in the width of the spadix
(uppermost pistil borne on an almost bare segment
of the spadix); pistils pale cream-yellowish to pale
yellow-green, smooth, 2.3— 3.5 mm diam.; style (id¬
ler stigma has fallen) sharply cupuliform, 1.5— 1.7
mm diam., with a single central pore, stigmas yel¬

low; staminodia elavate, white, 2-3 mm long, most¬
ly contiguous and sometimes fused at base; fruiting
spathe pale yellow; berries red to red-orange, 5—8
mm diam.

Dieffenbachia tonduzii ranges from southeastern
Nicaragua to the Pacific slope of Colombia (Antio-
quia, Choeo, Cauca, Valle) and Ecuador (Esmeral-
das, Loja, and Los Rios), from sea level to 1400 m,
in tropical wet forest, premontane wet forest, and
premontane rainforest in Central America and in
tropical wet forest and pluvial forest and pluvial
premontane forest and tropical wet forest transition
to premontane in Colombia.

The species is characterized by its fully sheathed
petioles, usually matte to weakly glossy, sometimes
weakly velvety blades with cordulate bases and nu¬
merous, broadly spreading primary lateral veins.
Because of the lully sheathed petioles it can be
confused only with D. horichii Croat, sp. nov. ined.,
and D. longispatha var. creberipistila Croat, var.
nov. ined. Both of the latter typically have much
larger leaves that are not at all cordulate at the- base
(though they are rarely broadly and weakly subcor-
date). Dieffenbachia tonduzii is polymorphic with
regard to leaf markings in the same way as D. oer-
stedii and is quite variable in all respects. It ap¬
pears to hybridize with D. oerstedii.

Dieffenbachia tonduzii is similar to D. daguensis
Engler, a Colombian species described from less
than 200 m elevation on the Rio Dagua in Valle
Department. Dieffenbachia daguensis also has
many rather close primary lateral veins and a fully
sheathed petiole, but differs in having the staminate
and pistillate sections of the spadix contiguous or
nearly so. In addition, it differs in having much
shorter petioles (described as being up to 5 cm
long).

A South American species, D. parlatorei Linden
& Andre, also sometimes has petioles fully
sheathed but differs from D. tonduzii in having leaf
blades usually broadest above the middle, semi¬
glossy on the lower surface, and the midrib often
broadly rounded and spongy. It also has the primary
lateral veins arising at a 40°— 60° angle from the
midrib (vs. often at more than 60° and sometimes
uj) to 90° in D. tonduzii).

Croat 70900, from 250 in in Choco Department,
Colombia, appears to be D. tonduzii, but differs in
several ways. It has leaves with the midrib flat-
raised above with the margins undercut. It also has
stems that appear scurfy (though weakly glossy if
rubbed clean). Another difference is that the petiole
sheath is more prominently free-ending and sub-

Volume 9, Number 4
1999

Croat

Central American Araceae

499

acute at the apex. In addition, the free portion of
the petiole is broadly and shaqily sulcate.

A collection from Bajo Calima (Valle Depart¬
ment, Colombia), Croat 79768, differs in having no
sterile segment, with the staminate and pistillate
flowers being closely contiguous. Il otherwise re¬
sembles this species.

Dieffenbachia tonduzii is named in honor of
Adolfo Tonduz (1862-1921), who made the first
collection of the species in December 1898.

Paratypes. COSTA RIGA. Alajuela: Canton de Ala-
juela, Rfo Sarapiquf, junction of road to Virgen del So¬
corro, I0°15'30"N, 84°10'30"W, 700-830 m, Cray urn A
Murakami 9989 (CR, MO), 830 m. Croat 68386 (MO),
Stevens 13564 (MO), San Ramon— Bajo Rodriguez, vicinity
of Km 19.5 NVi of San Ramon. 1()°11'N, 84°30'W, 925
m, Croat 78838 (MO); San Ramon— Bajo Rodriguez, near
Rio Cataratitas, 10°12'30”N, 84°32'W, 850 m. Croat
681)97 (MO); San Ramon— Bajo Rodriguez, 36—37 km N \\
of San Ramon. 10°15'N, 84°34'W. 500—515 m. Croat
68196 (MO); San Ramon— Fortuna, km 26, N of Fortuna,
6<K) m. I). Smith et al. 1059 (DUKE); San Ramon— San
Lorenzo, 4—7 km N of Balsa, KELVIN, 84°32'W, Liesner
A Judziewicz 14797 (MO); San Ramon— Fortuna. km 25 N
of San Ramon. 600 m, I). Smith et al. 1059 (MO); Nar¬
anjo— Aguas Zarcas, 8 km NE of Quesada, 600 m, Hwy.
15, Croat 46945 (MO); Finca Los Ensayos, 1 1 km NW of
Zarcero, 850 m. Croat 43629 (MO); Vo I can Arenal, trail
above Rio Agua Caliente, Funk et al. 10626 (CR). Funk
et al. 10718 (CR); Canas— Upala, 4 km NNE of Bijagua,
vie. of Rio Zapote, 400 m, Croat 36260 (MO); 10 km N
of Bijagua, 200 m. Croat 36472 (MO); Bosque Eterno de
Los Ninos, Rfo Penas Blancas valley, loguna Poco Sol.
Haber A Zuchowski 1 1 1 75 (MO). Cartago: Rfo Las Vuel-
tas, Tucurrique, 635 m, Tonduz 12874 (US); Camino Raiz
de Hide, SE of Platanillo (Tsipirf), 1200-1400 m, Croat
36727 ( MO), Croat 36747 (MO), Croat 36821 (MO); Finca
El Cedral, Orosi, Rfo Naranjo, 9°48'N, 83°52'W, 1400 m,
Irnl 4042 (F. MO. NY. SEE); Monumento Nacional Guay-
abo, Turrialba, Santa Teresita, above Rfos Guayabo, Lajas
& Torito, 9°57'N, 83°41'30"W, 700— 1800 m. Rivera 1718
(CR. K); Canon de Rfo Orosi I. Chacon et al. 1488 (CR.
MO); Monumento Nacional Guayabo, Turrialba,
9°58'20"N, 83°41'45"W, A. Perez I (CR); 1.5 mi. E of
Caclii, 10.2 mi. NE of junction of main road at Parafso,
Croat 47088 (MO); Cartago Canton. Rfo Reventazon at
CAT1E, Turrialba, 9°53'30"N. 83°39'W, 510-630 m, Gra-
yum et al. 9469 (CR. MO. NY). 630 m. Ijent 6 39 (E), 540
in. Lent 694 (F); 3 km E of Turrialba, “Los Espaveles”
nature trail. Rfo Reventazon, 9°54'N, 83°39'W, 525—600
m, Liesner el al. 15330 (MO); Turrialba-Limon. 24 km NE
ol Turrialba, 1.5 km E ol Tres Equis, 9°58'N, 83°34'W,
450-525 m. Liesner et al. 15365 (MO); 12 air km S of
Turrialba. 4 km SE of Pejibaye along Rfo Gato, 9°48'N,
83°42'W, 700 m. Liesner 14394 (MO); Parque Nacional
Tapantf. Sendero Oropendola, 9°45'40"N, 83°47'27"W,

1 1(H) m, Nilsson et al. 632 (CR), 09°54'06"N, 83°47'27"W,
1200—1300 m. Nilsson et id. 377 (CR). Guanacaste: Cor¬
dillera de Guanacaste, Rincon de la Vieja, NW of Que-
brada Grande, 500 m. Barringer et al. 4039 (F); El Ar¬
enal, 485—600 m, Standley A Valerio 45206 (US),
Standley A Valerio 45262 (US). Heredia: Occidental
Sendero, Kress 84-1630 (SEE); 8 km E of San Ramon,
10°20'N. 84°04'W, Loiselle 106 (MO); I a Selva Reserve,

Croat 44317 (MO). Folsom 10116 (DUKE), Grayurn 2780
(DUKE, E, MO), Harnmel 8874 (DUKE, F. MO), Schatz A
Grayurn 706 (DU KE); near Puerto Viejo, along road to Rfo
Sucio, 20 m, Croat 35682 (MO). Croat 35753 (F. MO); N
slopes of Volccin Barba, between Rfo Peje & Rfo Guacimo,
along Quebrada Cantarana, 300—400 m, Grayurn A Schatz
3174 (MO). Limon: F inca Castilla. 30 m. Dodge A Goer-
ger 9489 (MO); forests of Suerre and Dos Bocas, drainage
of Rfo Parismina and Reventazon, sea level. Shank A Mo¬
lina 4288 (EAP); Finca Montecristo on Rfo Reventazon
below Cairo, ca. 25 m, Standley A Valerio 48999 (US);
Rfo Segundo, Asuncion 500—600 m, Gomez et al. 23477
(MO); Reserva Indfgena Talamanca Sukut. mouth of Rfo
Sukul at Rfo Uren, 9°24'30”N, 82°58,I5"W, llammel et
al. 17548 (MO); Finca Anai, Quebrada Mata de Limon
(Sixaola region), 9°35'N, 82°39'W, Grayurn A Schatz 5279
(CR, MO); Refugio Barra del Colorado, between Rfo Chir-
ripoeito and Rfo Sardina, 10°38'N, 83°45'W, 12 m: Gra-
yum 9804 (CR, MO); Parque Nacional Tortuguero, Lomas
de Sierpe, 4 km NE of the P.N. over the Rfo Sierpe.
10°24'N, 83°33 ' W, 10°25'N, 83°34'W. 40-100 m. R. Ro¬
bles 2050 (CR. G, MO); Parque Tortuguero, Estaeion Agua
I ria, Robles 1 142 (CR), Robles 1158 (CR, MO); Tortuguero
Canton. Bribrf— Suretka, 100—200 m, Barringer et al.
3525: between Bribrf and Sixaola, NW of Paraiso, 20—70
m, 82°40'W, 9°39'N, Barringer et al. 3479 (CR. F); Li¬
mon— Shiroles, along Rfo Sixaola, 0.9 mi. SW of Bambu,
6.5 mi. SW of Bribrf, 50 m, Croat 43298 (MO); Rfo Ca-
tarata between Bribrf & Rfo Sixaola. 9°37'N, 82°49’W,
50— 100 m. Baker A Burger 90 (F. MO); ca. 10 km S of
Punta Cahuita, ca. 3 mi. S of turnoff to Bribrf. 70 m, Croat
43201 (MO); 7 km SW of Bribrf, 1(H)— 250 m, Gomez et
al. 20405 (B, MO); near Rfo Sixaola, ca. 1/2 mi. SW of
Bambu, ca. 3 mi. NE of Bratsi. 15 m. Croat 43266 (CR.
MO); Compiles, bosques de Toro Amarillo, 300 m, Leon
720 (CR. F); Canton Limon. headwaters of Quebrada Kak-
ebeta below divide between Rfo Xikiari and Rfo Boyei,
9°47'00"N, 83°20'30"W. 900-1000 m, Grayurn 10858
(MO); lurrialba— Limon, 1 1 mi. S of Siquirres, 650 m,
Croat 43333 (MO). San Jose: Vazquez de Coronado,
Braulio Carrillo National Park, San Jose to Siquerres
Hwy., frail to Rfo Sucio, site of the Old Carrillo Station,
10°09'50"N, 83°57' I0"W, 600-700 m, Croat 78787 (MO);
Puerto Viejo, near Rfo Sucio, 20 m, Croat 35753 (MO);
La Hondura. 1300—1700 m, Standley 36314 (US). NIC¬
ARAGUA. Rivas: Isla Ometepe, NW slopes of Volean
Maderas. 1 1°26-27'N, 85°30-31'W, 1000-1350 m. Ste¬
vens 6511 (MO). Zelaya: road to Colonia Manantiales (Go-
lonia Somoza), 6.3 km S of bridge at Colonia Yolaina,
1 1°36-37'N, 84°22'W, 200-300 m, Stevens 6420 (MO);
Rfo Punta Gorda, Atlanta, mouth of Cano el Guineo,
1 1°33'N, 84°02'W, 10 m, Moreno A Sandino 12855 (MO),
Sandino et al. 4875 (MO); Rfo Punta Gorda, Atlanta, 2
km al S de Carolina del Sur, 1 1°32'N, 84°01'W, 60 m.
Moreno A Sandino 12891 (MO); Rfo Punta Gorda, Atlan¬
ta, Cano Negro en la desembocadura del Rfo Chiquito,

1 1°35'N, 84°00'W, 30 m. Moreno A Sandino 12917 (MO);
Rfo Punta Gorda, Atlanta, desembocadura del Cano del
Oro en el Rfo Chiquito, 1 1°36'N. 84°00'W, 10 m, Moreno
A Sandino 12955 (MO); Rfo Punta Gorda, Cano El Gui¬
neo, Tdlez et al. 4875 (CR. MEXU, MO). PANAMA. Bo¬
cas del Toro: along Continental Divide trail, 1.1 mi. from
main highway, 8°44’N. 82°17'W, Croat 60355 (MO); Con¬
tinental Divide trail, 8°45'N, 82°15'W, McPherson 9865
(MO); Almirante— Ojo de Agua, 3—6 mi. W of Almirante,
30-200 m, Croat 38229 (F, MO); Valle del Sileneio, along
Rfo Changuinola, ca. 1 km above mouth of Rfo Teribe,

500

Novon

vicinity Teribe Indian population, 09°21'40"N,
82°3I'40"W, < 100 in. Croat X /hit 76450 (MO); Chiri¬
quf Grande— Fortuna, I mi. N of Continental Divide, 950
m. Churchill & Churchill 6 252 (MO); 1.0 mi. IN of Con¬
tinental Divide, 08°48'N, 82°13'W, 850 m. Croat 74930
(MO); 1.3.2 mi. W of Chiriquf Grande, 8°45'N, 82°10'W,
310 m. Croat i t Grayum 60139 (8. MO. PM A); \ of For¬
tuna Dam, 08°45'N, 82°15'W; 450-700 m, McPherson
1 1 129 (MO), McPherson 7371 (MO); 6.6 mi. N of bridge
over Fortuna Dam. 8°45'1N. 82°18'W, 780 m. Croat 66732
(MO); Fortuna— Chiriquf Grande, 8.1 km S of Punta Pena,
<)8°49'N. 82° 1 3 ' W, Croat 74952 (MO); 1.5 km W of Punta
Pena. 8°50'N. 82°1()’W. 80 m. Croat & Grayum (MO); Rio
Cricamola, between Finca Hi. Louis & konkintoe, 10—15
m. 460 m. Woodson el al. 1909 (F. MO); Almirante— Chan-
quinola railroad, mile 7.5 S of Changuinola, < KM) m,
Croat X Porter 16249 (MO); Changuinola. forest near Lu¬
zon. Kennedy 3253 (MO). Chiriquf: Gualaca— Fortuna, 7.9
mi. NF of Los Planes de Hornito, 8°44'N, 82°16'W, 1300
m. Croat 49932 (MO): 7.7 mi. NW of Los Planes, I KK)
m. Croat 48778 (MO); 7.9 mi. NW of Los Planes, 8°44'N.
82°I6'W. 1300 in. Croat 49932 (MO); 10 mi. NW of Los
Planes de Hornito. 8°45'N. 82°17'W. 1260 m. Croat
50049 (MO); vie. Fortuna Dam. valley of Rfo Chiriquf.
along aquaduct for IRHK facilities, 8°45'N. 82°18'\\.
1200-1300 in. Croat 66533 ( 8. DUKE, MO. NY. US);
Fortuna Dam area, 5.3 mi. N of center of dam, 1.4 mi. W
of main road near Continental Divide, 08°44'1\, 82°17'\Y,
Croat & /dm 76328 (MO); 4.5—5 km N of dam over For¬
tuna Lake, 8°43'N, 82°I7'W, km 1 1 I, more than 500 m,
Gordon 339 (PM A). Veraguas: NW of Santa Fe, I I km
N of Kscuela Agrfcola Alto de Piedra, 450—500 m, Croat
27490 (MO), Cochrane et al. 3817 (F); 0.6 mi. N of Fs-
cuela Agrfcola Alto Piedra, 730-735 in. Croat A- Folsom
33989 (MO). Croat X Folsom 34042 (MO); 1 hour walk
NW of Fscuela Agrfcola. 3000'. Antonio 2994 (MO): trail
to summit of Cerro Lute, ca. 3 km above Fscuela Agrfcola
Alto Piedra, 1050—1150 m. Croat 48903 (MO). 2< >( K >-
28(H)'. Sylsma & Antonio 3006 (MO); trail to summit of
Cerro Tute, 8°33'N, 81°08'W, 800-950 m, Croat 66993
(MO); mountains 3—5 mi. N of Santa Fe, 5( K ) — 1 ( X K ) m.
Gentry 3035 (MO); along 1st stream N of Fscuela Agrfcola
Alto Piedra, 7(K) m. Croat 49070 (MO). COLOMBIA. An-
tioquiu: Murrf— La Blanquita, Rfo Murn', 06°35'N,
76°50'W, Gentry et al. 75903 (MO). Cauca: Coteje, Rfo
Timbiquf, lehmann 887 6 (K). Choeo: Quebrada Anton,
15 km W of Santa Cecilia, 6 km W of Choeo— Risaralda
border, 20 km F of Plava del Oro. 05°20'30"N,
76° 1 3'45"W. 240 m. Croat 70900 (MO); Serranfa de Bau-
d6. Las Animas— Pato, Rfo Pato, ca. 4 km SW of Pato on
property of Sr. Guttierez, 5°30'N, 76°46'W, 150 m, Croat
56112 (CIIOCO. MO); Rfo Picliima. Communidad ituli-
gena Waunana, 4°25'N, 77°I7'W. 1(H) m. (COL. MO);
Quibd6-Medellfn, al Km 185, 14 km F of Tutunendo,
5°44'N, 76°43'W, 3(H) m. Croat 56282 (CIIOCO. MO): km
208.5, 9 km W of Tutunendo, 5°39'N. 76°40'W, < 100
in. Croat 56205 ((3)1.. JAUM, MO); km 85 W of Bolfvar.
350 in. Croat 49310 (MO): San Jose del Palmar— Novita.
vicinity of Santa Rosa at end of existing road. 4°57'N,
76°30'W. 350 m. Croat 56625 (MO); Ouibdo— Medellin.
25 mi. F of Quilwl6, 2(M) m, Croat 52300 (MO); ca. 2 km
F of Plava de Oro, 5°20'N. 76°23'W. 2(H) m. Croat 57427
(CIIOCO. MO); Pueblo Rico (Risaralda) to Istmina (Cho-
c6), I km W ol Guarato and Rfo Guarato at Risaralda and
Choeo border. 285 in. Croat 70868 (CM. MO); F side of
Serranfa del Darien, approach from Acandf, 8°30'N,
77°2(>'W. I(>0 m. Juncosa 619 (MO). Valle: Cordillera Oc¬

cidental, western slopes, Rfo Digua, Piedra de Moler,
900-1180 m. Cualrecasas 15053 (US). ECUADOR. Es¬
meralda*: Canton Floy Alfaro, Comuna de Corriente
Grande (Rfo Chimbagal, tributary of Rfo Cayapas),
(M)°41'N. 78°50'W. 150—200 m, Ydriez et al. 1387 (MO);
Reserva Ecologies Cotacachi-Cayapas, Charco Vicente,
Rfo San Miguel, (K)°43'N. 78°53'W. 150 m, Palacios X
Tirado 11287 (MO. OCNF): Lita-Carondolet Road, km
If). 800 m. Schwerdtfeger 021422 (MO); Canton Ouinimle,
NF ol Las Golondrinas, Sitio La Bella Jungla, Cooperativa
Unidos Venceremos, (K)°20'N. 79° 1 2 ' W', 3(X) m. Palacios
11452 (MO); (Juininde, Bilsa Biological Reserve. Montan¬
as <le Maehe, 35 km W of Santa Quinind6, 5 km W of
Santa Isabela, 0°2I'N. 79°44'W, 400-600 in. Pitman &
Hass 995 (MO); Fila de Bilsa. 7 km F of San Jos£ de
Bilsa, 80 km due SW of Fsmeraldas, 12 km SF of FI Salto
on Atacames-Muisne Road, 00°37'N, 79°51'W, 280 m.
Gentry et id. 72955 (MO). Los Rios: Caserfa de Bimbo,
10.5 km N of Patricia Pilar, 0°35'N, 79°I2'30”W. 550—
575 in. Croat 57000 (CM. MO); Centinela. Montanas de
I la. 13 km F of Patricia Pilar. (M)°26'S. 79°30'W. (MM) m.
Hammel A- Trainer 15836 (MO); Centinela, 12 km F ol
Patricia Pilar. 650 m. Gentry et al. 26705 (MO); FI Cen¬
tinela, 550—600 m. Gentry el al. 24700 (MO).

Rhodospatha pelliieida Croat & Grayum, sp. nov.
TYPE: Panama. Bocas del Toro: along road be¬
tween Chiriquf Grande and Fortuna, 7.7 mi.
W of Chiriquf Grande, 1.5 mi. W ol Punta
Pena, 80 m, 8°50'N, 82°10'W. 19 Mar. 1985,
Croat & Grayum 60106 (holotype, M0-
3123550; isotypes, B, K, PM A, US).

Planta epiphytiea; internodia brevia, 1-2 cm diarti.; pe-
tiolus 8.5—23 cm longus. vaginatus ubique; genieulum V-
sulcatum; lamina oblongo-elliptica, inaequilatera, 21-41
cm longa, 7.2—14.5 cm lata; nervis primariis lateralibus
21—33 utroque.

Appressed-climbing epiphyte; stems to 2-2.5 m
above ground, the flowering branches somewhat di¬
vergent; internodes shorter than broad, 1—2 cm
diam. near apex (drying frequently less than 1 cm
diarn.) and often flattened, glossy, medium to dark
green, tapered below, drying brown, matte, densely
and finely granular. Leaves distichous, somewhat
clustered toward the stem apex, spreading or as¬
cending; petioles medium green, 8.5—23 cm long,
ca. 5 mm diam. below the genieulum, sheathed
throughout their length, drying matte, yellowish
brown; sheath erect, persistent intact with the mar¬
gin sometimes drying darker brown; genieulum C-
shaped, V-sulcate adaxially, 1—1.5 cm long, some¬
what shrunken on drying; blades reflexed from
petiole on upper leaves, directed in line with the
petioles on the lower leaves, moderately coriaceous,
conspicuously bicolorous, oblong-elliptic, inequi¬
lateral, tapering ± equally toward both ends, slight¬
ly inequilateral and gradually acuminate to obtuse
and abruptly acuminate at apex, acute to attenuate
at base, broadest at the middle or just above the

Volume 9, Number 4
1999

Croat

Central American Araceae

501

middle, 21—41 X 7.2—14.5 cm, 2.5— 4.5 times lon¬
ger than wide, 1.6— 3.4 times longer than petioles,
dark green and glossy above, drying gray-green to
olive-green or yellowish brown; matte to weakly
glossy, yellow-green to silvery-green and paler be¬
low, drying yellow-brown to yellow-green; midrib
paler than surface, obtusely V-suleate above, thick¬
er than broad and ± concolorous with surface be¬
low; primary lateral veins 21—33 pairs, somewhat
to deeply impressed above, darker than surface be¬
low and convex, but drying only weakly raised to
Hat, 5—17 mm apart, departing from midrib at 30—
47°, moderately arcuate to the margin, sometimes
prominently downturned just before merging with
the midrib; interprimary veins 1 alternating be¬
tween adjacent primary lateral veins, much less
conspicuous than the primary lateral veins; both
the primary lateral and interprimary and sometimes
the minor veins margined with thick black lines on
drying (those of the interprimary and minor veins
usually intermittent and appearing as rows of pel¬
lucid dots and streaks on living plants and appear¬
ing above as weak dark green bumps) and also with
a sparse arrangement of pale raphide cells aligned
with the dark black lines; minor veins 1—3 between
alternate primary and interprimary veins, usually
very weak, branched toward the margin; cross veins
usually inconspicuous, often branched, mostly ex¬
tending throughout the surface; surface minutely
dark-granular, often sparsely covered with pale
raphide cells on lower surface. Inflorescences 1 or
2, from the upper leaf axils; peduncles arching,
(5—) 12-28 cm X 6—8 mm diam., usually longer
than petioles (up to 13 cm longer), sometimes
equaling petioles or rarely somewhat shorter than
petioles, drying light brown; spathe coriaceous,
acuminate at apex and up to 1.4 cm diam. when
furled, hemispherical when open, 8—15 cm long, to
ca. 5 cm wide, 5—13 cm diam. when opened flat,
the veins impressed, white to greenish white, matte
outside, glossy and white inside, dark brown on
drying, often promptly deciduous, sometimes per¬
sistent; spadix cylindroid, weakly tapered toward
both ends, white to cream at anthesis, soon becom¬
ing gray or green or grayish green after anthesis,
6.3—18.5 cm long, drying 6—15 mm diam., some¬
times shorter but usually 1—3 cm longer than the
spathe, stipitate 4—9 mm on the back side, held at
ca. 35° angle to the peduncle; flowers 4-sided, 2-
2.8 mm in both directions, 12-13 visible in the
principal spiral, 8—9 visible in the alternate spiral;
pistils mostly 4-sided, sometimes irregularly 5—6-
sided, 1.4— 2 mm wide, the style truncate, gray,
matte and minutely granular, sometimes smooth,
brown; stigma broadly elliptic, slit-like, 0.9— 1.1 X

0. 5-0.8 mm, depressed medially, drying black with
a pale ring-like base; stamens included; filaments
densely embedded with short, pale raphide cells;
anthers 1.2 mm long, 0.7 mm wide, the thecae ob¬
long, connivent; pollen white. Infructescence to 15
cm long, to 2 cm diam.; seeds reniform, light
brown, glossy, 0.8—1 mm long, the outer margin
deeply and sharply suleate.

Rhodospatha pellucida ranges from Nicaragua to
Panama on the Atlantic slope and into Colombia
and Ecuador on the Pacific slope, from sea level to
360 m in premontane moist forest and wetter parts
of tropical moist forest life zones.

This species is characterized by its relatively
small size, with stems less than 2 cm in diameter
and leaf blades usually less than 12 cm wide. The
leal blades dry yellowish brown to yellowish green
and have the primary lateral veins as well as many
interprimary veins and minor veins bordered with
continuous or intermittent dark black lines (ap¬
pearing as irregular rows of pellucid-punctate
markings on the lower surface and as minute
bumps on the upper surface on fresh leaves). Also
characteristic is the dark reddish granular or
brownish granular lower blade surface on magnifi¬
cation.

Rhodospatha pellucida is perhaps most closely
related to R. killipii Croat ined., an undescribed
species restricted to the Amazon basin. These spe¬
cies have leaves of similar color and shape with a
scattering of pale raphide cells visible on the lower
surface. The latter differs in having blades that dry
thicker, with minute, close ridges on the lower sur¬
face, and lacking the intermittent pellucid pune-
tations that appear as darkened dots and dashes
along the interprimary veins on the lower surface
of dried leaves as found in R. pellucida.

Rhodospatha pellucida is also similar to R. for¬
get ii N. E. Brown, a species apparently also from
Costa Rica. Rhodospatha forgetii differs in having
thicker blades that are not at all dark granular-
punctate on the lower surface, and have more prom¬
inently raised primary lateral veins without the
dark margins or the intermittent dark dots and
streaks along the interprimary veins and the minor
veins.

Paratypes. COLOMBIA. Clioco: km IBS. 14 km E of
Tutunendo, 5°44'N, 76°43'W, Croat 56285 (CHOCO.
JAUM, MO); near Rio Ir6, ca. 10 km S of Istmina,
5°14'30"N, 76°4I'00"W, 50 m, Croat 57389 (MO). Nari-
iio: Tumaco, Corr. Llorente, Vereda el Carmen. Diaz et al.
1104 (COL). COSTA RICA. Alajuela: 300 m. Ciudad
Quesada to Los Chiles, 9 km M of Ciudad Quesada, Rio
Peje, 10°22'N. 84°28'W, Hatnmel et al. 15306 (CR, MO);
San Carlos, 3 km S of Boca de Arenal, Rio San Carlos,

502

Novon

Hacienda Boca Arenal, 100 m, 10°30'N, 84°30'W, Hum¬
mel & (ie Nevers 15325 (MO); 22 kin NK of Quesada l»y
air, 4 km W of Mnelle San Carlos, 10°28'N, 84°30'W,
Liesner 14139 (MO), Liesner 14157 (MO); 2 km N of Santa
Rosa, 15 km N of Boca Arenal on Quesada— Muelle San
Carlos— Cos Chiles road, 1 < K ) m, 10°38'N, 84°31'W, Lies¬
ner el al. 15033 (CR, MO); 15 km WNW of Quesada by
air. I km W of Jabillos, 150-200 m. 10°23'N, 84°33'W,
Liesner el at. 15149 (MO); along road from Canas to Upala,
13.8 km N of Bijagua, 100-150 rn, Croat 36426 (MO);
new Volcdn Arenal, Taylor & Taylor 1 1609 (MO). Guan-
acaste: Cake Arenal, from Rfo Chiquito to Rio Cano Ne¬
gro, 10°27'N, 84°50'W, 550-850 m, Hammel et al. 15155
(MO). Heredia: 1 .a Selva Reserve, at confluence of Rio
Sarapiquf and Rfo Puerto Viejo, 10°26'N, 84°01'W, 50—
1(K) m, Grayum 7668 (MO), Hammel 9517 (MO), Croat
78734 (INB, MO), Grayum & Sleeper 2896 (DUKE, MO);
Puerto Viejo, 50 m. /larger & Stolze 5836 (CR. F, US),
Croat 35713 (MO). Union: in Tropical wet forest, and
Premontane wet forest. Cordillera de Talamanca along Rfo
Barbilla, from jet. with Rfo Dantas to ca. 1 km down¬
stream, 10°01'N, 83°25'W, 100 m, Grayum et al. 8966
(MO); between Quebrada Mata de Cimon & Quebrada Ti-
gre, Finca Anai, (Sixaola region), 09°34'N, 82°40'W, 25—
30 m, Grayum et al. 4450 (CR, MO); trail from Mata de
Cimon to Finca Anai, 09°33'N, 82°38'W, 15 m, Grayum
& Schulz 5260 (CR. MO); Rfo Frfo to Cimon, vie. Rfo
Blanco, W of Guapiles, Quebrada Danta, 3 mi. S of main
highway along old railroad bed, I0°I2'N, 83°49'W, 360
in. Croat 68424 (MO): Ref. Nac. Barra del Colorado, 10—

15 m, Grayum 9845 (CR, MO); hills 2 airline km SSE of
Islas Buena Vista in the Rfo Colorado, 14 airline km SW
of Barra del Colorado, I0°40'N 83°40'W, 10-20 m. Dav-
idse & Herrera 31091 (MO); Cerro Coronel, K of Caguna
Danto, 1.0°41'N, 83°38'W, elev. 20—170 m, Stevens 24558
(MO); Hacienda Tapezco, Davidson & Donahue 8499
(CAM), Davidson <& Donahue 8759 (CAM), Davidson &
Donahue 8979 (CAM. MO). ECUADOR. Esnieraldas:
8.6 km beyond bridge over Rfo Esnieraldas (near San Ma¬
teo, on road to Esnieraldas airport), ca. 6.6 km beyond
Universidad Teenologiea Euis Vargas Torres- Eslac ion Ex¬
perimental Mutile, along Rfo Mutile (tributary ol Rfo Es-
meraldas), 0°52'N, 79°33'W. 80 m, Croat 55628 (QUA,
MO); Esmeraldas— Muisne, 50 km SW of Atacames, Bal-
slev Steere 3104 (QCA). NICARAGUA. Kio San Juan:
Ca Cupe, ca. 25 km ENE of Boca de Sabalo, Salick &
Stijfhoorn 8269- b (MO). PANAMA. Bocas del Toro: For-
tuna— Chiriquf Grande, on side road 10 mi. from continen¬
tal divide, on trail off pipeline trace, 8°55'N, 82°10'W.
300 m, McPherson 12566 (MO, I’M A); Fowler hill above
RR station at Milla 7.5. Cro<it 16402 (MO). Colon: Sa-
banitas— Portobello, Rfo Piedras Cumber Road, departing
main highway 6.7 mi. E of Sabanitas, 3.9 mi. up logging
road in direction of Santa Rita Ridge, 09°22'30"N,
79°41'30"W, 250 m, Croat 75156 (CM, K. MO). San
Bias: El Llano— Cartf road. Km 26.5, 9°19'N, 78°55'W,
200 m, de Nevers et al. 7846 (MO).

Acknowledgment. Research lor this article was
supported in part by NSF grant DEB9306891.

h ive New Taxa and Two New Combinations of Aeanthaceae from

Central America

L. H. Durkee

Department of Biology, Grinnell College, Grinnell, Iowa 50112-0806, U.S.A. Present address:
115 Woolf Lane, Ithaca, New York 14850, U.S.A.

ABSTRACT. Five new taxa from Central America,
Anisacanthus nicaraguensis (compound dichasial
cymes, 4-merous red corollas, 3-colporate pollen),
Aphelandra scabra var. angustifolia (narrowly ellip¬
tic leaves), Justicia cauliflora (some inflorescences
on old wood, lanceolate bracts, greenish white co¬
rolla), J. chiriquiensis (cymose panicles, oblaneeo-
late bracts, 4-porate pollen), and J. nicaraguensis
(clusters ol 2—3 flowers to short spikes, corollas or¬
ange, dicolporate pollen), are described. The new
combination Justicia rothschuhii (Lindau) Durkee,
based on Chaetothyiax rothschuhii Lindau, is made,
and the new name Justicia danielii Durkee, which
replaces Chaetothyiax cuspidatus D. N. Gibson, is
provided.

In preparation of the family treatment of Acan-
thaceae for the Flora de Nicaragua and Flora Me-
soamericana, examination ol recent collections has
revealed four new species and one new variety. In
each case these taxa were collected from areas that
have not been well collected in tin* past. Also, as
the result ol combining Chaetothyiax with Justicia
(Graham, 1988), two Central American species of
the lormer genus must now he translerred to the
latter.

Anisacanthus nicaraguensis Durkee, sp. nov.
TYPE: Nicaragua. Departamento Esteli: “Ku-
kamonga,” 18.5 in N ol Esteli, 13°14'N,
86°21'W, elev. ca. 850 m, 19 Jan. 1985, Pedro
P. Moreno 25325 (holotype, MO; isotype,
CAS). Figures 1, 6A.

Frutex ad 3 m. Folia ovata, 27—75 mm tonga. 9-25 mm
lata, apiee acuminato, basi acuta ad obtusa; petiolis ad 27
mm. I nflorescentiae spiciformes secundae vel eymae com-
positae sine floribus basibus; bracteae lineares-subulatae,
3.5— 4.5 mm longae. Corollae rubrae, 4-merae. 33—35 mm
longae. Fructus clavati, 15—17 mm longi, glabri, semini-
l)us 4.

Shrub to 3 m tall. Young stem quadrate-fluted,
glabrous except lor puberulence at the nodes; older
stems subquadrate or terete, brown epidermis
shredding in long strips. Leaves ovate, blades 27—

75 mm long, 9—25 nun broad, apicallv acuminate,
basally acute to obtuse, margins entire to erenulate,
glabrous, cystoliths to ca. 0.2 mm long, abundant
and plainly visible on both surfaces ol dried leaves;
petioles to 27 mm long, glabrate. Inflorescences ba¬
sically compound dichasial cymes lacking the cen¬
tral or basal flower, to 2.5 cm long, inflorescence
branches sometimes appearing as secund spikes,
rachises minutely puberulent with hairs ca. 0.1 mm
long plus occasional, slightly longer glandular
hairs; bracts and bracteoles linear-subulate, 3.5—
4.5 mm long, 0.3— 0.4 mm broad at base, sparingly
puberulent. Flowers with short pedicels 0.5— 1.5
mm long; calyx 5-merous, 11-15 mm long, seg¬
ments equal, deeply divided, long subulate, 9—13
mm long, glandular-puberulent; corollas red, 4-
merous, 33—35 mm long, puberulent outside, upper
lip single-lobed, 29 mm long, 0.8— 0.9 mm broad,
lower lip 3-lobed, lobes 15—17 mm long, 2.5—3 mm
broad, apically acute; stamens 2, exserted, fila¬
ments to 33 mm long, attached near base ol corolla,
glabrous, anthers bithecous, thecae parallel, equal,
3.6— 3.8 mm long, pollen prolate, 50—55.3 X 27.2—
29 /cm, 3-colporate, pseudocolpi on either side of
each pore, insulae lacking; style to 42 mm long,
extending beyond stamens, stigma bilobed. Fruit a
clavate capsule, 15—17 mm long, 5.8— 6.9 mm
broad, glabrous, tuberculate, stipe 6—8 mm long;
seeds 4, 5.8— 6.5 mm long, 4.4— 5.1 mm broad, gla¬
brous, margins conspicuously thickened.

Habitat, distribution, and phenology. Anisacan¬
thus nicaraguensis is known only from the type col¬
lected in flower and fruit in January. The collection
area, Cuesta de Kukamonga, is a mesa that has
been collected repeatedly in recent years, leading
W. D. Stevens of MO to conclude that the plant is
rare in this generally arid area that, he told me, has
resulted from an old basaltic lava flow now eroded
into a series of mesas, flat-topped hills, or moun¬
tains with steep slopes. The soil that overlies the
basalt is thin and black. He also related that the
area has a distinctive vegetation dominated by Aga¬
ve, columnar cacti, thorny shrubs, and small trees,
with occasional oaks where the soil is a little deep-

Novon 9: 503-510. 1999.

504

Novon

1 cm

Figure 1. A-K, Anisacanthus nicaraguensis Durkee {Moreno 25325). — A. H. Habit. — C. Fruit. — D. Corolla. — E.
Enlarged view of inflorescence.

er, and numerous epiphytes. The “Cuesta” is where
the Pan American Highway ascends the first mesa
going north from Estelf.

Etymology. The species epithet refers to the
country in which the type collection was made.

According to Daniel (1982), Anisacanthus is a
genus of ca. 16 species of perennial shrubs with a
center of distribution in mostly dry habitats of Mex¬
ico. Its closest relative is presumably the genus

Carlowrightia, which grows in similar habitats, but
its species are typically shorter plants with 5-mer-
ous corollas (vs. 4-merous) that lack the bright red
or orange color characterizing species of Anisacan¬
thus. This is the first known collection of the genus
Anisacanthus in Nicaragua. Anisacanthus nicara¬
guensis most closely resembles A. tetracaulis Leon¬
ard of Honduras, but differs from this species in its
longer corollas (33—35 mm vs. 20-23 mm), longer

Volume 9, Number 4
1999

Durkee

Anisacanthus, Aphelandra & Justicia

505

sepals (9-13 mm vs. 5-7 mm), longer bracts and
bracteoles (3.5— 4.5 mm vs. 1-3 mm), and longer
thecae (3.6-3. 8 mm vs. 2. 9-3.0 mm). The other
species of the genus are from Mexico north ol Chia¬
pas, the adjacent United States, and South Ameri¬
ca. No other species of the genus have been de¬
scribed from Central America.

Aphelandra scabra (Vahl) Smith var. angustifol-

ia Durkee, var. nov. TYPE: Panama. San Bias:
on solid ground at Playon Chico, Rio Grande
(Ukupseni), between Neba Dummat and camp
Kanadi, 09°15'N, 78°15'W, elev. 50-100 m, 2
Nov. 1991, H. Herrera , J. Morris & J. Mojica
1061 (holotype, MO; isotypes, CAS, PMA).
Figure 2.

Folia anguste elliptica, 6.7-15.5 cm longa, 0.6— 1.3 cm
lata; oorollae 4.3— 4.6 cm longae.

This variety differs from the type variety (which
ranges from Mexico to Brazil) by its leaves, which
are narrowly elliptic, 6.7-15.5 cm long, 0.6-1. 3 cm
broad, glabrous above, with appressed hairs be¬
neath that are more dense on the costa; by its pu-
berulous petioles, 2-10 mm long; and by its corolla
that is 4.3— 4.6 cm long (slightly longer than that of
the type variety) and glabrous outside.

Habitat, distribution, and phenology. This va¬
riety is described as “common to the edge of the
river* in the vicinity of the type locality and “along
the creek” in the other area of collection. No other

' information is given. These collections were made
at lower altitudes, 50-150 m, of the Caribbean
coastal region and Pac ific watershed area of Pan¬
ama. These are apparently similar habitats although
separated by the Cordillera de San Bias. Additional
collection would probably reveal a more extensive
range for the variety.

Etymology. The varietal epithet refers to the
conspicuously narrow leaves.

Pnratype. PANAMA. Panama: along El Elano-Carti
road, along creek E of road, 9°15’N, 79°00,W, 150 m. 8
Sep. 1989, G. McPherson 1403 6 (MO).

Justicia cauliflora Durkee, sp. nov. TA PE: Pana¬
ma. Panama: 5—10 km NE of Altos de Pacora,
ea. 750 m, 7 Mar. 1975, S. Mori & J. Kallunki
4061 (holotype, MO). Figures 3, 6B.

Suffrutex ad 2 m. Folia elliptica, 4.5—20 cm longa, 1.4—
6.3 cm lata, apice acuminato, basi attenuata; petiolis 4—
20 mm longis. Inflorescentiae spiciformes, terminales et
axil lares, ad 5 cm longae; bracteae lanceolatae, 14—15 mm
longae, 3-4 mm latae. Corollae labiae superae viridi-al-
bae, labiae inferae lavandulaceae. Fructus clavati, 20-22
mm longi, seminibus 4.

Suffrutescent herb to 2 m tall. Young stem qua¬
drangular, strigose. Leaves elliptic, 4.5—20 cm long,
1. 4—6.3 cm broad, apically acuminate, basally at¬
tenuate, margins crenate, glabrous above, puberu-
lent along costa and main veins beneath, cystoliths
visible on both surfaces; petioles 4-20 mm long,
strigose. Inflorescences in what appear as terminal

506

Novon

Figure 3. Justicia caulifiora Durkee (Mori & Kallunki 4991). — A. Habit. — B. Inflorescence on older, woody stem.
— C. Anther with detail of appendage on lower theca. — D. Capsule.

Volume 9, Number 4
1999

Durkee

Anisacanthus, Aphelandra & Justicia

507

and axillary spikes to 5 cm long, 1.5 cm broad,
some occurring on old wood as well, bracts lance¬
olate, 14—15 mm long, 3—4 mm broad, puberulent,
mostly opposite but occasionally subopposite, each
bract with one flower most olten, but occasionally
with two flowers per bract, suggesting that the in¬
florescence is thyrsoid with dichasia usually re¬
duced to one flower; bracteoles narrowly lanceolate,
13—13.5 mm long, 1.7-2 mm broad, puberulent;
peduncle to 18 mm long, strigose. Flowers 5-mer-
ous, calyx segments equal, narrowly lanceolate, 12—
15 mm long, 1.7—2 mm broad, puberulent; corolla
bilabiate, 2.7— 2.8 cm long, glandular-puberulent,
tube ca. 1.6 cm long, upper lip emarginate, green¬
ish white, ca. 13 mm long, 6 mm broad, lower lip
3-lobed, lavender, ca. 12 mm long, 12 mm broad,
lobes rounded, ca. 3 mm long, 4 mm broad; sta¬
mens two, exserted to the tip of the upper lip, fil¬
aments attached just below the apex of the corolla
tube, ca. 11 mm long, glabrous, thecae superposed,
upper theca 1.5— 1.8 mm long, lower theca with
basal appendage, 1.7-1. 9 mm long excluding ap¬
pendage, appendage 0.6— 1.0 mm long, pollen pro¬
late, 49—57.2 X 31-40.3 gm, 2-colporate, one row
of insulae on each side of the pore, ca. 8 insulae
per row, fitting Graham’s (1988) Type 5 pollen most
closely. Capsule clavate, 20—22 mm long, 6 mm
broad, 5 mm thick, puberulent, seeds 4, suborbie-
ular, ca. 4 mm diam.

Habitat, distribution, and phenology. Found in
premontane rainforest.

Justicia cauliflora is recognized by its spicate in¬
florescences (some occurring on old wood) with lan¬
ceolate bracts, 14—15 mm long, greenish white co¬
rolla ca. 2.8 cm long that has a lavender lower lip
and a glandular-puberulent indumentum, stamens
with superposed thecae but only the lower cell with
a basal appendage, and filaments that are attached
at the apex of the corolla tube. It appears to have
some affinity to J. fortunensis T. F. Daniel & Wass-
hausen of Panama but differs from it in bract shape
(lanceolate vs. broadly elliptic to subcirculate to
obovate), length of calyx lobes (12—15 mm vs. 4.5—
6.5 mm) and length of corolla (27-28 mm vs. 8—
10 mm). It also resembles J. veraguensis T. F. Dan¬
iel & Wasshausen of Panama, which has obovate to
broadly elliptic bracts, shorter calyx lobes (5-6
mm), and a shorter corolla 10—13 mm long.

Justicia cauliflora is found in open areas of pre¬
montane rainforest near Altos de Paeora of the
Province of Panama and along streams in lower al¬
titude forest on the Atlantic slope of Veraguas Prov¬
ince.

Paratype. PANAMA. Veraguas: Atlantic slope, Rio

Concepcion to Rfo Rarrera, ca. 100 m, 16 Oct. 1978,
Hummel 5180 (MO).

Justicia ehiriquiensis Durkee, sp. nov. TYPE:
Panama. Chiriquf: Fortuna Dam, along Que-
brada Arena, S of continental divide, 08°45'N,
82°15'W, 1050 m alt., 8 Sep. 1987, G. Mc¬
Pherson 11699 (holotype, MO). Figures 4, 6C.

Suffrutex ad 1.5 m. Folia elliptica ad elliptico-ovata,
10—15 cm longa, 3—8 cm lata, apice acuminato, basi acuta
ad obtusa; petiolis 0.5—3 cm longis. Inflorescentiae thyr-
siformes, terminates, ad 14 cm longitudine; bracteae ob-
lanceolatae, 4.5— 6.5 mm longae, 0.7— 1.3 mm latae. Cor-
ollac albae, 16—18 mm longae, staminibus exsertis.
Fructus non observatus.

Suffrutescent herb to 1.5 m tall. Stem quadran¬
gular, puberulent in younger portions. Leaves ellip¬
tic to elliptic-ovate, 10—15 cm long, 3—8 cm broad,
apically acuminate, basally acute to obtuse, mar¬
gins crenulate to entire, glabrous above, puberulent
on costa below, cystoliths abundant and prominent
on both surfaces of dried specimens; petioles 0.5—
3 cm long, puberulent. Inflorescences in loose ter¬
minal cymose panicles (thyrses) ca. 14 cm long, 8
cm broad, rachises puberulent; bracts oblanceolate,
4.5— 6.5 mm long, 0.7— 1.3 mm broad, with short,
glandular trichomes plus longer eglandular tri-
chomes; bracteoles similar but slightly smaller.
Flowers 5-merous, calyx segments equal, lanceo¬
late, 5—6 mm long, 1—1.3 mm broad, pubescence
that of the bracts and bracteoles; corolla white, pu¬
berulent outside, bilabiate, 16-18 mm long, the
tube 7—9 mm long, basally ca. 4.5 mm broad, the
upper lip erect, keeled, apically emarginate, lower
lip with pale purple markings, 3-lobed, the lobes
apically rounded, each ca. 2 mm long, 3 mm broad;
stamens two, exserted to about the tip of the upper
lip, thecae superposed, ca. 3 mm long, basally
spurred, pollen subprolate, 56.7-66.7 X 43.8-^48.3
/u,m, 4-porate, trema not differentiated, whole sur¬
face microreticulate, studded with rows of smooth,
obtuse verrucae, rows separated at each trema area,
somewhat resembling Type 9 pollen as described
by Graham (1988) except that the verrucae are ob¬
tuse instead of conical. Fruit not observed.

Habitat, distribution, and phenology. Known
only from the type collection, this species was in
flower in September and was collected along a
stream in rainforest in the vicinity of Fortuna Dam.

Etymology. The specific epithet refers to the
province in which the type was collected.

Among Mesoamerican Justicia, J. ehiriquiensis is
recognized by its crenulate leaves, inflorescences
in cymose panicles, oblanceolate bracts 4. 5-6.5
mm long, 4-porate pollen, and white corolla 16—18

508

Novon

B

1 cm

inm long. It bears some resemblance to J. crenata
(Leonard) Durkee but can be distinguished by its
crenulate leaves (vs. erenate), cymose panicles (vs.
spicate), oblanceolate bracts (vs. elliptic), and
white corolla 16—18 mm long (vs. red with yellow-
tipped corolla to 50 mm long). It can be distin¬
guished from J. pectoralis Jacquin by its elliptic to
elliptic-ovate leaves (vs. lanceolate), cymose pani¬
cles (vs. spicate), oblanceolate braets 4.5— 6.5 mm
long (vs. subulate bracts ca. 2 mm long), calyx 5—
6 mm long (vs. 2 mm long), and corolla 16-18 mm
long (vs. 8—12 mm).

Justicia danielii Durkee, nom. nov. Replaced
name: Chaetothylax cuspidatus D. N. Gibson,
Fieldiana Rot. 34: 58. 1972. Not Justicia cus-
pidata Vahl, 1790. TYPE: Guatemala. Alta
Verapaz: valley of Santa Lucfa, on road from
Gualan to Cobdn, 650 m, Pittier 1792 (holo-
type, F).

This species has also been collected in Hondu¬
ras.

Justicia nicaraguensis Durkee, sp. nov. TYPE:
Nicaragua. Zelaya: along trail from Cerro Sas-
laya to San Jose del Hormiguero, between
Cano Majagua and Cano Sucio, tall evergreen
forest, ca. 13°45'N, 84°59'-85°00'W, 600-800
m, 10 Mar. 1978, W. D. Stevens 6848 (holo-
type, MO; isotype, CAS). Figures 5, 6D.

Caulis ligneus, volubilis vel frutex ad 4 m. Folia ovata
ad laneeolata, 3.8—10.2 cm longa, 1-3.7 cm lata, apice
acuminate, basi obtusa ad rotundata; petiolis ad 5 mm.
Inflorescentiae in fasciculae parvae vel spicas ad 15 mm
longas; bracteac longi-subulatae, 9—11 mm longae, glan-
dulari-pubescentes; corollae aurantiacae. Fructus clavati,
curvi, 33 mm longi, seminibus 4.

Liana or shrub to 4 m tall. Young stem subterete
with unilineately retrorse yellow pubescence. Leaves
ovate to lanceolate, blades 3.8-10.2 cm long, 1-3.7
cm broad, apically long acuminate, basally obtuse to
rounded, margins entire to undulate, sparingly pilose
mostly on costa on both surfaces, cystoliths abundant
and easily seen in dried leaves above, less visible
beneath; petioles to 5 mm long, puberulent. Inflo¬
rescences ranging from terminal and axillary clusters
of 2—3 flowers to short, compact spikes to 15 mm
long, peduncles 9—14 mm long with appressed-pi-
lose triehomes; bracts and bracteoles long-subulate,

Volume 9, Number 4
1999

Durkee

Anisacanthus, Aphelandra & Justicia

509

Figure 5. Justicia nicaraguensis Durkee. — A. Habit. — B. Bract and bracteoles. — C. Enlargement of bract surface
trichomes. — D. Calyx. — E. Corolla. — F. Capsule. — G. Anther thecae. (A from Pipoly 6289 ; B-G from Stevens 6848.)

bracts 9—11 mm long, bracteoles 8—8.5 mm long,
both densely glandular-puberulent externally. Flow¬
ers opposite or alternate, one per bract, sessile to
subsessile; calyx 7.5—12 mm long, 4-merous, seg¬
ments equal, long-subulate, 6.5-10 mm long, with

external eglandular puberulence plus smaller glan¬
dular hairs; corollas orange, bilabiate, 35-41 mm
long, externally pubescent with trichomes 0.3— 0.6
mm long, upper lip apically acute, not divided. Ki¬
ll mm long, lower lip 3-lobed, 10-13 mm long, 4-

Eigure 6. Scanning electron micrographs of pollen. — A. Anisacanthus nicaraguensis ( Moreno 25825). — B. Justicia
cauliflora (Harnmel 5180). — C. Justicia chiriquiensis ( McPherson 1 1699). — I). Justicia nicaraguensis ( Pipoly 6289).
Scale lines represent 10 /tm.

510

Novon

4.5 mm broad, lobes apically rounded, 3-4 mm long,
middle lobe 2.2— 2.5 mm broad, lateral lobes 1.7—
2.2 mm broad; stamens attached at the base ol the
upper lip, filaments 8.5— 9.5 mm long, thecae un¬
equally attached at a subperpendicular angle, upper
theca 2.2 mm long, lower theca 1.5 mm long, both
lacking basal appendages, pollen prolate, 52—54.4
X 32.5—35.5 pun, dicolporate, 4 rows of insulae in
each trema region, most closely resembling Graham’s
(1988) Type 7 pollen; style 32-35 mm long, gla¬
brous. Capsule clavate, curved, 33 mm long, 4 mm
broad, seed-bearing area elliptic, not constricted,
gradually tapering into a stipe 18 mm long, minutely
glandular-puberulent; seeds 4, mature seeds not ob¬
served.

Habitat, distribution, and phenology. The three
collections of Justicia nicaraguensis observed for
this description were found along a trail in tall ev¬
ergreen forest and at a finca in disturbed secondary
forest, the species thus appearing to occur in both
primary and secondary forest areas. All collections
were from the same general area of Zelaya Depart¬
ment, Nicaragua, and were collected in March,
each displaying flowers and fruits. Elevation 300-
800 m.

Etymology. The species epithet refers to the
country in which the original collection was made.

Among Nicaraguan Justicia, J. nicaraguensis is
recognized by its shrubby and/or scandent viny
habit and ovate leaves with long acuminate apices
and obtuse to rounded bases; short petioles to 5
mm long; inflorescences of axillary headlike clus¬
ters of 2-3 flowers or short, compact spikes to 15
mm long; 4-merous calyx; orange corollas 35—41
mm long; and curved, gradually tapered capsule
that is 33 mm long. It is not easily confused with
other Justicia species from Nicaragua, but it shares
some characters with J. sulfurea (Donnell Smith) D.
N. Gibson of Guatemala in its leal size and shape,
its inflorescences of headlike clusters, and its calyx
that is 4-merous and of similar length, but differs
in its viny habit, its orange corollas that are longer
(35-41 mm vs. 20-28 mm), and corolla lips that
are about one-half the length of die corolla tube

(vs. about equal). Justicia nicaraguensis also resem¬
bles J. madrensis T. F. Daniel, an endemic from
Chiapas, Mexico, in its 4-merous calyx, color and
length of corolla, leaf shape and size, and inflores¬
cence type, but differs in its viny habit, longer
bracts (9-11 mm vs. 1.8-2. 5 mm), inflorescence
length (15 mm vs. 55 mm), longer calyx (7.5—12
mm vs. 5.5—9 mm), and length and shape of cap¬
sule (33 mm vs. 20 mm, elliptic head with no con¬
strictions vs. elliptic head with medial constriction).

None of the species of Justicia from Costa Rica
that do not also occur in Nicaragua are likely to be
confused with Justicia nicaraguensis.

Paratypes. NICARAGUA. Zelaya: along same trail as
holotype, in tall secondary growth evergreen forest, ca.
I3°45'N, 84°58— 59'W, elev. 300-450 m, Steren.s 7015
(MO); Finca S of La I’imienta, ca. 13°45'N, 84°59'W, Pi-
poly 6280 (CAL. MO).

Justicia rothschuhii (Lindau) Durkee, comb. nov.
Basionym: Chaetothylax rothschuhii Lindau,
Bull. Herb. Boissier 3: 492. 1895. TYPE: Nic¬
aragua. Matagalpa: on a river bank near Maz-
maz, 200 m, 22 Mar. 1894, Rothschuh 410
(holotype, B destroyed; photo at F); Nicaragua.
Chontales: ca. 2.8 km N of Cuapa, disturbed
evergreen forest on hillside, 12°17'N,
85°23'W, 400-500 m, W. D. Stevens 6148
(neotype, MO; isoneotype, CAS).

This species has also been collected in Guate¬
mala, Honduras, and Costa Rica.

Acknowledgments. I thank the curator of MO
lor providing the specimens for examination, and
Laura Goddard, Gretel Kiefer, Allison Shaw, and
Kristin Stuchis for the illustrations. I also thank
Allison Shaw and Lenore Durkee who made the
scanning electron micrographs.

Literature Cited

Daniel. T. F. 1982. Anisacanthus andersonii (Acantha-
ceae), a new species from northwestern Mexico. Bull.
Torrey But. Club 109: 148-151.

Graham, V. A. W. 1988. Delimitation and infra-generic
classification of Justicia (Acanthaceae). Kew Bull. 43:
551—624.

A New Species of Corallorhiza (Orchidaceae) from
West Virginia, U.S.A.

John V. Freudenstein

Department of Biological Sciences, Kent State University, Kent, Ohio 44242, U.S.A.
Current address: Museum of Biological Diversity, The Ohio State University,
1315 Kinnear Rd., Columbus, Ohio 43212, U.S.A.

Abstract. Corallorhiza bentleyi Freudenstein is
described from West Virginia, U.S.A. The species
is eleistogamous and is currently known from only
one population. It is most similar to C. striata Lind-
ley, specifically variety involuta (Greenman) Freu¬
denstein from southern Mexico, particularly in the
structure of the labellum and size of the flower.

Corallorhiza Gagnebin comprises 11 species (in¬
cluding the one herein) of temperate-montane leaf¬
less orchids (Freudenstein, 1997). The genus is
limited to the New World, except for C. trifida Chat-
elain, which is circumboreal. The centers of diver¬
sity are southern Mexico and the northeastern Unit¬
ed States and Canada. The present taxon is
currently known from only a single population but
should be sought elsewhere in the region; given its
eleistogamous nature, plants could be dismissed
easily as individuals of the much more common C.
maculata (Rafinesque) Rafinesque in bud.

Corallorhiza bentleyi Freudenstein, sp. nov.
TYPE: U.S.A. West Virginia: Monroe Co., ca.
43 km NW of Roanoke, Virginia, 18 July
1998, Freudenstein 2550 (holotype, AMES).
Figure 1.

Species Corallorhizae striatae var. involulae (Greenman)
Freudenstein maxime similis, sed callo ampliore labelli et
floribus cleistogamis differt.

Plants to ca. 20 cm tall, reddish to yellow, with
up to 20 flowers. Ueaves reduced to sheaths. Rhi¬
zomes coralloid, roots absent. Flowers closed, peri¬
anth connivent, perianth segments 3-veined, those
other than the labellum with very faint darker red¬
dish striping, sometimes with yellowish perianth and
reddish ovary and rachis. Dorsal sepal lanceolate,
5.9-6. 1 mm long, 2. 1-2.5 mm wide. Lateral sepals
lanceolate, somewhat falcate, strongly cymbiform,
5.5— 6.1 mm long, 2. 1-2.5 mm wide. Lateral petals
narrowly ovate, weakly cymbiform, acute, 5.0-5.8
mm long, 2.2-3.0 mm wide. Labellum unlobed,
ovate, acute, with prominent, thickened, involute
margin, 4.9-5. 7 mm long, 3. 1-3.5 mm wide and with

slightly upturned apex; callus formed from two fused
lamellae at the base of labellum, ca. 1/2 as long as
the labellum, 2. 5-2.8 mm long, 1.5— 1.8 mm wide.
Column curved forward, 3. 0-3. 8 mm long, 0.8— 1.0
mm wide, with a small depression adaxially at its
base. Rostellum triangular. Pedicellate ovary narrow¬
ly obovoid, 6.0— 9.0 mm long, 2.5— 3.0 mm diam.,
without adaxial mentum at summit.

Etymology. The species is named for Stanley
Bentley, of southwestern Virginia, who discovered
and first recognized the distinctness of this popu¬
lation.

This unusual Corallorhiza is known from a single
population in an area ca. 4 m X 4 m on an aban¬
doned revegetated railroad grade, with neither tracks
nor ties present, in Appalachian deciduous forest of
Acer pensylvanicum L., Liriodendron tulipifera L.,
Quercus, and Rhododendron. The population has in¬
creased in number since it was discovered in 1996.
Four plants in fruit were noted that year, with eight
stems in 1997, 20 in 1998, and 27 in 1999 (S. Bent¬
ley, pers. comm.). Although individuals of Corallor¬
hiza are often clonal, the variation in color seen in
this population suggests that more than one clone is
present. Searches in the immediate area have not
yet uncovered additional individuals. Neither were
specimens detected in a monographic study of the
genus (Freudenstein, 1997) or in a recent search of
material at WVA. However, it seems likely that other
individuals may exist, since the site was an active
railroad earlier in the 20th century, and the plant
could almost certainly not have been growing at its
present location at that time.

The species is clearly most similar to Corallo¬
rhiza striata Lindley, which is the only other spe¬
cies to have an unlobed labellum with involute
margins, fused lamellae at the base of the labellum,
and the absence of a small spur (often called a
“mentum”) at the summit of the ovary. Corallorhiza
striata is a morphologically variable species that
has a wide distribution across North America. The
population of C. striata nearest to the West Virginia

Novon 9: 51 1-513. 1999.

512

Novon

Figure I. Corallorhiza bentleyi Freudenstein. — A. Habit. — B. Perianth, dissected. — C. Column, ventral view. Drawn
by the author from a photograph and spirit material collected at the type locality.

locality is in central New York, where the species
is very rare. However, the flowers of C. bentleyi
differ strikingly in size from northern North Amer¬
ican C. striata , being only one-half to one-third
their size. Remarkably, Corallorhiza bentleyi is
most similar to the small Mexican forms of C. stri¬
ata that have been called C. involuta Greenmail,
but were most recently recognized as C. striata var.
involuta (Greenmail) Freudenstein (Freudenstein,

1997) because of the essentially continuous cline
of decreasing flower size from northeastern North
America southwestward to southern Mexico. The
taxa differ in the proportions of the fused lamellae
at the labellum base, being ca. half the length of
the labellum in C. bentleyi and ca. one-quarter the
length in C. striata var. involuta.

Corallorhiza bentleyi and northern C. striata also
differ in flowering phenology. Corallorhiza striata is

Volume 9, Number 4
1999

Freudenstein

Corallorhiza bentleyi from W. Virginia

513

an early-flowering species; in northern Michigan,
where the plant can be locally frequent, it is typi¬
cally in flower during the first week in June. Cor¬
allorhiza bentleyi flowers much later, during mid to
late July. In montane regions of Mexico, C. striata
var. involuta is also in flower during mid to late July.

The other clear difference between C. bentleyi
and C. striata var. involuta is that the flowers in
the former are cleistogamous. Corallorhiza striata
varies in the degree to which the perianth expands,
with the northeastern plants often having widely ex¬
panded flowers that are at least sometimes insect-
pollinated (Freudenstein, 1997), while flowers of
southwestern and Mexican populations tend to be
less widely opened, but no truly cleistogamous
flowers have been reported previously in this spe¬
cies. Many instances are known of species in which
both chasmogamous and cleistogamous flowers are
produced on the same plant, including Viola spp.
(Russell, 1965), Ajuga iva (L.) Schreber (Ruiz De
Clavijo, 1997), Lamium amplexicaule L., and Im¬
patient pallida Nuttall (Stewart, 1994; see Uphof,
1938, arid Lord, 1981, lor reviews). Perhaps less
common are species with two types of individuals,
each producing only cleistogamous or chasmoga¬
mous flowers. These have sometimes been given
taxonomic standing (Uphof, 1938). Among orchids,
a peloric cleistogamous form is known from Mexico
in Corallorhiza odontorhiza (Willdenow) Poiret as
variety pringlei (Greenman) Freudenstein forma ra¬
dio Freudenstein, and a more widespread, primar¬
ily cleistogamous variety is also known in that spe¬
cies (var. odontorhiza ; Catling, 1983; Freudenstein,
1997). Some populations of Nervilia gassneri B.
Pettersson are cleistogamous (Pettersson, 1989,
1990). A number of instances of wholly cleistoga¬
mous species are also known (“complete cleistog-
amy” of Uphof, 1938). Recently described species
from Orchidaeeae include Cheirostylis notialis D.
L. Jones, a cleistogamous or nearly cleistogamous
species (Jones, 1997), and Epipactis futakii Mered’a
& Potucek (Mered’a & Potucek, 1998).

Although cleistogamy is rare, autogamy is known
in many temperate orchid species and is accom¬
plished in different ways (Catling, 1983). The flow¬
ers of most Corallorhiza have a distinct pollinium
stalk derived from the rostellum (a hamulus; Freu¬
denstein, 1994). In C. striata the hamulus is well
developed and functions in insect pollination.
Cleistogamous forms of C. odontorhiza have essen¬
tially no hamulus (Catling, 1983; Freudenstein,
1994). Corallorhiza bentleyi does have a hamulus,
but since the flowers do not open it is not utilized;
the pollinia germinate in situ and grow down to the
stigma on either side of the rostellum.

The cleistogamous nature of C. bentleyi could be
a clue as to its origin. Two possibilities exist: either
it was dispersed by chance from Mexico to West
Virginia, or it is a relictual population of a taxon
with a formerly more extensive distribution. The
chances of an orchid seed being transported this
distance and finding the appropriate fungus with
which to interact seem remote. Since there are no
currently known cleistogamous populations of C.
striata in Mexico, the evolution of cleistogamy
would probably have occurred after dispersal, ar¬
guing against a recent dispersal. The relict hypoth¬
esis seems more likely, since it is easier to conceive
of a formerly more extensive distribution that has
contracted with changing environmental and per¬
haps pollinator conditions. The absence of a polli¬
nator would have favored any tendencies toward au¬
togamy and cleistogamy, and it is perhaps simply
fortuitous that the current population has survived.

Acknowledgments. I am grateful to Stan Bentley
for bringing this population to my attention, to John
B. Nelson and Lawrence Magrath for comments on
the manuscript, and to Donna Ford-Werntz for as¬
sistance at WVA. This research was supported by
NSF grant DEB-9615437.

Literature Cited

Catling, P. M. 1983. Autogamy in eastern Canadian Or-
chidaceae: A review of current knowledge and some
new observations. Naturaliste Canad. 110: 37-53.
Freudenstein, .1. V. 1994. Character transformation and
relationships in Corallorhiza (Orchidaeeae: Epidendro-
ideae). II. Morphological variation and phylogenetic
analysis. Amer. J. Bot. 81: 1458—1467.

- . 1997. A monograph of Corallorhiza. Harvard

Pap. Bot. 10: 5-52.

Jones, D. L. 1997. A taxonomic revision of Cheirostylis
(Orchidaeeae) in Australia. Muelleria 10: 75—83.

Lord, E. M. 1981. Cleistogamy: A tool for the study of
floral morphogenesis, function and evolution. Bot. Bev.
(Lancaster) 47: 421-449.

Mered’a. P. & 0. Potucek. 1998. Epipactis futakii, sp. nov.
(Orchidaeeae), a new cleistogamous flowering plant
from Slovakia. Preslia 70: 247—258.

Pettersson, B. 1989. Pollination in the African species of
l\ervilia (Orchidaeeae). Lindleyana 4: 33-41.

- . 1990. Studies in the genus Nervilia (Orehida-

ceae) in Africa. Nordic J. Bot. 9: 487—497.

Buiz De Clavijo, E. 1997. A comparative study of the
reproductive biology of two Ajuga species (Lamiaceae)
in the southwest of the Iberian Peninsula, lnt. J. PI. Sci.
158: 282-291.

Russell, N. H. 1965. Violets (Viola) of central and eastern
United States: An introductory survey. Sida 2: 1—113.
Stewart, S. C. 1994. Genetic constraints on mating system
evolution in the cleistogamous annual Impatiens palli¬
da: Inbreeding in chasmogamous flowers. Heredity 73:
265-274.

Uphof, J. C. T. 1938. Cleistogamic flowers. Bot. Bev. (Lan¬
caster) 4: 21—49.

A New Species of Miconia Ruiz & Pavon (Melastomataceae) from

Espfrito Santo, Brazil

Renata Golden berg

Universidade Federal do Parand, Setor de Cieneias Bidlogicas, Departamento de Botanica,
Caixa Postal 19031, Centro Polit^cnico, 81531-970, Curitiba, PR, Brazil

ABSTRACT. Miconia capixaba belongs to Miconia
sect. Miconia, series Paniculares Naudin. It is dis¬
tinguished from related species by the dense, stel¬
late trichomes on young branches, inflorescences,
and hypanthia, as well as abaxial leal surfaces, and
by cuneate leal bases, small panicles, triangular
inner calyx lobes, and a papillose (otherwise gla¬
brous) ovary apex. This species occurs in montane
forests in the state ol Espfrito Santo, Brazil, a flora
noted for other endemic taxa within Melastomata¬
ceae.

Miconia capixaba R. Goldenberg, sp. nov. TYPE:
Brazil. Espfrito Santo: “Municfpio de Santa
Teresa, Estagao Biologica de Santa Lucia, ar-
vore nao amostrada, proximo & parcela 63,” 21
July 1993 (H), L. D. Thomaz s.n. (holotype,
MBML 9300; isotypes, CEPEC, K, MO, UEC).
Figure 1A— F.

Haec species congeneris sect. Miconiae ser. Panicular-
ium Naudin ramis, lamina foliari subtus, inflorescentia hy-
panthioque dense stellato-puberulis, lamina foliari basi
cuneata modice decurrenti, paniculis parvis laxis, calycis
laciniis internis triangularibus, ovario ad apieem papilloso
distinguitur.

Tree 4—16 m tall, young branches slightly com¬
pressed, older branches terete, internodes 1.4—3 cm
long, along with inflorescences, hypanthium, and
petioles densely covered with stellate, yellowish
brown to lerrugineous trichomes ca. 0.1 mm diam.
Blades 6.5—12 X 1.4— 3.1 cm, lanceolate, apex acu¬
minate, base cuneate and shortly (1—2 mm) decur¬
rent, entire to repand on the upper half, charta-
ceous, lamina adaxially stellate-furfuraceous, soon
glabrous, lustrous and dark olive-green, abaxially
also densely stellate pubescent, white, but nerves
yellowish brown, shortly 3-plinerved, with a faint
pair of inframarginal nerves, secondary veins free,
diverging 1—2 mm from the base; primary, second¬
ary, and transverse veins impressed above, but pri¬
mary and secondary veins prominent below and
with transverse ones faint, reticulation barely visi¬
ble above and below, the areoles 0.7-1 mm diam.
Petioles 5—10 mm long, canaliculate, slightly sili¬

cate. Panicles 1-2.5 X 0.9-1 cm, terminal, pyr-
amidate or elongate, depauperate, with 2 branchlets
per node; bracteoles ca. 0.6 mm long, narrowly del¬
toid to linear, persistent; flowers sessile (laterals in
each dichasium on short, ca. 1 mm, peduncles), 5-
merous. Hypanthium ca. 1.6 mm long, campanu-
late, without longitudinal ridges; calyx tube 0.2-0.3
mm long, caducous, inner lobes ca. 0.6 mm long,
triangular, acute; outer lobes inconspicuous. Petals
1.7 X 1 mm, papillose, margins papillose-ciliate,
obovate to oblong, apex deeply emarginate, with 2
irregular lobes 0.3 mm long. Stamens 10, in two
series, white, only slightly heteromorphic, the larg¬
er ones with filaments 1.9— 2.1 mm long, anthers
1.7— 1.8 mm long, linear-oblong, arcuate, thecae
1.3-1. 4 mm long; the smaller ones with filaments

1.5— 1.9 mm, anthers 1.4— 1.5 mm, linear-oblong,
straight, thecae ca. 1.1 mm; anthers of both cycles
with a narrow apical pore, ca. 0.1 mm diam.; con¬
nective prolonged 0. 3-0.4 mm below the thecae,
dorsally thickened and not appendaged, ventrally
with two minute lobes, sometimes directed back¬
ward and appearing like dorsal appendages. Ovary
0.7— 0.9 mm high, fused to the hypanthium for Vi
of its length, 3-celled, with 2-3 ovules in each cell,
apex rounded, obscurely 5-lobed, papillose; style

4.5- 5 mm long, filiform, apex truncate, not thick¬
ened. Mature fruits not seen.

Miconia capixaba is morphologically similar to
species in the section Miconia ser. Paniculares
Naudin and section Glossocentrum Bentham &
Hooker. The morphology of its anthers indicates re¬
lationship to the former (connective ventrally bi-
lobed vs. dorsally calcarate or unappendaged).
Within Miconia ser. Paniculares, Miconia weddellii
Naudin has glomerulate inflorescences (Wurdack,
1974) and leaves densely villose below. Miconia
burchellii Triana and M. pohliana Cogniaux differ
from M. capibaxa in many vegetative characters
(leaves broader, 5— 7-nerved, with rounded to cor¬
date bases). Miconia irwinii has less dense pubes¬
cence and rounded, wider calyx lobes. Miconia ci-
poensis R. Goldenberg ined. (in press) lias similar
inflorescences and flowers, but shorter and wider

INovon 9: 514-516. 1999.

Volume 9, Number 4
1999

Goldenberg

Miconia capixaba from Brazil

515

Figure 1. Miconia capixaba R. Goldenberg (all drawings from L. D. Thomaz 819). — A. Branchlet with an inflores¬
cence. — B. Inflorescence unit showing the external appearance of the hypanthium and calyx. — C. Longitudinal section
of hypanthium and ovary. — D. Petal. — F,. Smaller stamen, frontal-lateral view. — F. Larger stamen, dorsal and frontal

view.

1 mm

516

Novon

leaves with dendritic triehomes and a stellate ovary
apex.

Within Miconia sect. Glossocentrum, M. capixaba
is vegetatively similar to M. cubatanensis, which
has stamens with a single dorsal appendage and a
completely inferior ovary with its apex covered hy
stellate triehomes. Miconia brasiliensis (Sprengel)
Triana and M. pauciflora Cogniaux also have quite
similar flowers, hut with unappendaged or dorsally
calcarate stamens and leaves with more sparsely
distributed triehomes. Miconia setosociliata Cog¬
niaux has ciliate leaf margins (often not evident in
some collections from Espfrito Santo, e.g., W. Boone
446, US), clubbed dendritic hairs on the branch-
lets, and blades with the nerves basally fused by a
membrane. Finally, M. pepericarpa DC. has 4-mer-
ous flowers and dendritic triehomes on the lower
leaf surface.

Miconia capixaba occurs in the mountains of the
state of Espfrito Santo, along the Brazilian coast
between Bahia and Rio de Janeiro. These moun¬
tains have a particularly rich flora (Thomaz & Mon-
teiro, 1997), with several endemic melastomes,
such as Dolichoura espiritusanctensis Brade, Meri-
ania tetramera Wurdack, Merianthera burlemarxii
Wurdaek, and Leandro fallacissima Markgraf. The
melastome flora of Espfrito Santo also includes
some poorly known species, otherwise collected
only in adjacent Rio de Janeiro ( Miconia longicau-
data Cogniaux, M. octopetala Cogniaux, M. setoso¬
ciliata Cogniaux) or Bahia ( Meriania callophylla
(Naudin) Triana).

The epithet “capixaba” is derived from the Bra¬
zilian word for people or things native to the state
of Espfrito Santo.

Paratypes. BRA/11. . Espfrito Santo: "Municfpio de
Santa Teresa, Estayao Biologica de Santa Lucia,” 20 Aug.
1996 (fl), H. Goldenberg et al. 375 (MBML, UKC, US);
“Estayao Biologica de Santa Lucia, mata atlantica de en-
eosta. beira do rio, 650-800m,” 30 Aug. 1994 (fl). L. D.
Thomaz 819 (HRCB, UEC); "Estayao Biologica de Santa
Lucia, planta u" 1543,” 21 July 1993 (fl). /,. I). Thomaz
848 (MBML. VIES).

Acknowledgments. I thank W. Rodrigues for
help with the Latin diagnosis, E. Kickhoffel for the
drawings, A. B. Martins, V. Bittrich, and R. Romero
for comments, and L. D. Thomaz and H. Q. B. Fer¬
nandes from the Museu de Biologia Mello Leitao at
Santa Teresa, Espfrito Santo, for sending speci¬
mens. 1 also acknowledge the Departamento de Bo-
tanica, Universidade Estadual de Campinas and the
UEC Herbarium for support when I was working on
this species.

Literature Cited

Cogniaux, C. A. 1891. Molastomataeeae. In A. & C.
DeCandolle (editors). Monographiae Phanerogamarum
7: 1—1256. G. Masson, Paris.

Thomaz, L. D. & R. Monteiro. 1997. Composiyao florfstica
da Mata Atlantica de encosta da Estayao de Santa Lu¬
cia, Municfpio de Santa Teresa, ES. Bol. Mus. Biol.
Mello Leitao (N. S£r.) 7: 1—29.

Wurdack. J. .). 1962. Melastomataceae of Santa Catarina.
Sellowia 14: 109-217.

- . 1974. Certamen Melastoniataceis XXIII. Phyto-

logia 29: 135-151.

A New Calyptranthes (Myrtaceae) from Nicaragua

Bruce K. Holst

Marie Selby Botanical Gardens, 811 South Palm Avenue, Sarasota. Florida 34236, U.S.A.

bholst@virtu.sar.usf.edu

Abstract. In preparation for the treatment of Ca¬
lyptranthes (Myrtaceae) for the Flora tie Nicaragua ,
a summary and identification key to the five known
species from the country are presented. One of
these, Calyptranthes amarulenta B. Holst, a new
species from central Nicaragua, is described and
illustrated. The new species is characterized by
having bicolorous leaf blades when dry that are
rounded to obtuse at the apex, a biconvex midvein
on the upper leaf surface, and a glabrous inflores¬
cence.

The known Nicaraguan Calyptranthes flora in¬
cludes five species. Three of these are widespread:
C. chytraculia (L.) Swartz from mostly lowland,
coastal regions in southern Mexico and the Greater
Antilles to northern Colombia, C. hylobates Stan-
dley ex Amshoff from lowland, mostly evergreen
forests in Mexico, Costa Rica, and Panama, and C.
pallens Grisebach, a polymorphic species found
from low to moderately high elevations in semide-
ciduous forests of southern Florida, Mexico, Central
America, and the West Indies. The fourth species,
C. amarulenta B. Holst, is endemic to semideci-
duous forests at low elevations in central Nicara¬
gua, and the fifth, a species of uncertain status, is
known from mid-elevation cloud forests in Hon¬
duras and Nicaragua. The following key distin¬
guishes the Nicaraguan species.

hi A TO THK Sl’KClh.S Oh CALYPTRANTHES IN NICARAGUA

la. Branches of the inflorescence and/or flowers
densely pubescent.

2a. Inflorescence tomentose . C. chytraculia

2b. Inflorescence sericeous .

. C. pallens var. williamsii

lb. Branches of the inflorescence and flowers gla¬
brous. or at most with a few sparse hairs.

3a. Young stems narrowly winged or 4-angled;

peduncles and pedicels filiform . . C. hylobates
3b. Young stems terete or compressed, never an¬
gled; peduncles and pedicels robust.

4a. Leaf apices rounded to obtuse; midvein

biconvex on upper leaf surface .

. C. amarulenta

4b. Leaf apices abruptly long-acuminate;
midvein sulcate on upper leaf surface
. C. sp.

Calyptranthes amarulenta B. Holst, sp. nov.
TYPE: Nicaragua. Chontales: Hacienda Vera¬
cruz, including Cerro La Batea and Cerro Los
Charcos, 120-475 rn, 12°11'N, 85°21'W, 17
July 1983, W. D. Stevens 22252 (holotype, MO;
isotypes. BM, CR. EAP, F, HNMN, JBSD. K.
MEXU, NY, SEL, SP. U, US). Figure 1.

Frutex vel arbor usque ad 8 m altus, ramulis compres-
sis. Folia petiolo 4—6 mm longo sicco nigro, lamina in
sicco bicolori, elliptica vel obovata, obtusa vel rotundata,
costa supra biconvexa. venis lateralibus obscuris vel sub-
manifestis. Panicula 8 ad 13 flora. Flos glaber; calyptra
apiculata, hypanthio 1—2 mm longo. Fructus glaber. glo-
bosus.

Shrub 2-3 m tall or tree to 8 m tall; twigs slightly
compressed, glabrescent, the young vegetative parts
with sparse, scurfy trichomes. Leaves short-petio-
late, the petioles 4—6 mm long, black when dry,
shallowly sulcate; blades bicolorous when dry, the
upper surface dull to slightly lustrous, olive- to
gray-green, lower surface brownish green, broadly
elliptic to obovate, 5. 2-7. 2 X 2. 5-4.2 cm, charta-
ceous, glabrescent; midvein broad, biconvex; sec¬
ondary veins scarcely evident, 10 to 13 on each
side; marginal vein 1-2 mm from blade margin and
parallel to it, as prominent as the secondaries; apex
obtuse to rounded; margin plane to slightly revo¬
lute; base cuneate; glands barely impressed-punc-
tate on upper surface, convex or plane on lower
surface. Inflorescence glabrous, twice branched,
solitary or paired on abortive axes; bracts caducous,
not known; peduncle 3.2— 4.5 cm long; secondary
branches 3—4.5 cm long. Flowers 8 to 13 per pan¬
icle, glabrous, sessile or seemingly long-pedicellate
on 1-flowered branches that are 8—10 mm long;
bracteoles caducous, not known; petals, stamens,
style, and stigma unknown; hypanthium prolonged
1—2 mm beyond the ovary; calyptra apiculate, 3
mm wide. Fruits glabrous, globose, ca. 1 cm diam.;
seed 1.

Calyptranthes amarulenta is known from the de¬
partments of Chontales and Matagalpa in Nicaragua
where it is locally common in primary or disturbed,
semideciduous forests at 120—500 m elevation. It
is known by the common name “coralillo.” The spe-

Novon 9; 517-519. 1999.

518

Novon

Figure I. Calyptranthes amarulenta B. Holst. — A. Habit (leaf venation slightly exaggerated to show detail). — B.
Flower, past anthesis. — C. Transverse section of flower. — D. Fruit. — E. Cross section of leaf blade. Drawn from
Stevens 22428 (fmit); 22282 (habit, flowers, leaf detail).

cific epithet denotes the fruits, which are described
as very bitter and inedible (Stevens 22428).

Calyptranthes amarulenta is readily distin¬
guished from all other Calyptranthes in Nicaragua
by having blunt leaves that dry bicolorous and a
broad, biconvex midvein. It appears to be most
closely related to C. millspaughii Urban of southern
Mexico and Belize, from which it differs by having

glabrescent (vs. densely appressed-pubescent with
coppery hairs) inflorescences and flowers and ob¬
tuse to rounded (vs. abruptly acuminate) leaf api¬
ces.

Paratypes. NICARAGUA. Chontales: ca. 2.8 km
above (N of) Cuapa, ca. 12°17'N, 85°23'W, 400-500 m,
4 Sep. 1077, W. I). Stevens 2700 (MO); Hacienda Vera¬
cruz, including Cerro Fa Batea and Cerro Los [Charcos],

Volume 9, Number 4
1999

Holst 519

Calyptranthes amarulenta from Nicaragua

120-475 m, 12°H'N, 85°21'W, 4-6 Aug. 1983, W. I).
Stevens 2242H (MO. SEE). Vlatagalpa: Rancherfa, 1 1 km
al NE fie Muy muy, aprox. 280 m, 12°46'N, 85°31'W. 20-
22 Aug. 1984. P. P. Moreno 24440 (MO, SEE), 24441
(MO).

Acknowledgments. I thank the stall at MO for

providing the collections for this study, Valerie
Renard at SEL lor providing the illustration, and
Leslie Landrum, Victoria Hollowell, Roy Gereau,
and an anonymous reviewer for suggesting improve¬
ments to an earlier draft of this paper.

A New Species of Roucheria and a New Species of Hebepetalum
(Hugoniaceae) from the Venezuelan Guayana

Antony Jardim

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Paul E. Berry

Botany Department, University of Wisconsin, 132 Birge Hall, 430 Lincoln Drive, Madison,

Wisconsin 53706, U.S.A.

ABSTRACT. A new species of Roucheria and a new
species of Hebepetalum (Hugoniaceae) from south¬
ern Venezuela are described and illustrated. Rouch¬
eria sipapoensis is a montane, small-leaved shrub
known only from Cerro Sipapo in Amazonas state,
while Hebepetalum neblinae is a lowland, wet forest
tree from white sand areas in southernmost Vene¬
zuela and northern Brazil. Keys to the species of
both genera are provided.

Roucheria and Hebepetalum are closely related
woody taxa in the Linales. Cronquist (1988) placed
both genera in the Hugoniaceae, although they had
earlier been treated in a more broadly circum¬
scribed Linaceae (Planchon, 1847; Bentham,
1862). While preparing the Hugoniaceae treatment
for the Flora of the Venezuelan Guayana (Ramirez
et ah, 1999) and undertaking a revisionary study
of Hebepetalum and Roucheria, we discovered one
new species in each genus. Venezuela is the center
of diversity of these genera, with five species of
Roucheria and all three species of Hebepetalum oc¬
curring there. The species commonly known as
Roucheria punctata (Ducke) Ducke is now consid¬
ered a synonym of an earlier name, R. Columbiana
Hallier (Ramirez et ah, 1999).

Cronquist’s circumscription of Hugoniaceae is
supported here based on results of revisionary work
of these genera, which point to the presence of dru¬
paceous fruit, woody habit, and crenulate/subcren-
ulate leal margins with glands as supporting shared
characters with the family. Hebepetalum is distin¬
guished from Roucheria by its petals with a clawed
base which are inwardly pubescent, the presence
of trichomes of variable density at the apex of the
ovary, and broquiodromous leaf venation with sec¬
ondaries joined in a series of conspicuous and well-
formed arches. In Roucheria the venation is camp-
todromous, with the secondary veins tightly
parallel, and lacking the series of conspicuous
arches. Both genera share a submarginal collecting

(fimbrial) vein, which is formed by the fusion of the
secondaries. The leaf venation in Roucheria is rem¬
iniscent of that of many members of the Clusiaceae
and Ochnaceae. Including the species described
below, we now recognize eight species in Roucheria
and three in Hebepetalum.

Roucheria sipapoensis Jardim & P. E. Berry, sp.
nov. TYPE: Venezuela. Amazonas: Cerro Si¬
papo, Paraque, higher plateau, among rocks,
1798 m, 12 Dec. 1948, B. Maguire 27625 (ho-
lotype, MO; isotypes, NY, WIS). Figure 1.

Haec species inter congeneros vegetative glabros
Roucheriae momalveae A. II. Gentry foliis subsessilibus
decurrentibus etiam R. elatae Ducke inflorescentia ex
panicula axillari \^X cm longa constante maxime accedit,
sed a hac foliis valde coriaceis atque habitu humiliore
altitudine 10 m non excedente, ab ilia foliis plerumque
minoribus in sicco griseo-olivaceis atque inflorescentia
paniculata, ab ambabus lamina foliari marginibus ad bas-
im valde revoluta distinguitur.

Shrub or small tree 3-10 m tall; stems glabrous,
nodes round. Leaves alternate, simple, ovate-ellip¬
tical, coriaceous, subsessile, olive-gray upon dry¬
ing, with a thin layer of wax, 1.8-8. 5 cm long, 0.5—
2.5 cm wide, broadly acute at the apex, decurrent,
with inrolled margins at the base, glabrous, dark
green on the adaxial surface, light green abaxially;
midvein raised abaxially, impressed adaxially, nu¬
merous strongly parallel secondary veins terminat¬
ing at the margin in a submarginal, continuous col¬
lecting vein, raised on both surfaces; margins
crenate-subserrate with numerous dentations, 40—
55 glands per leaf, inconspicuous, oval, present in
each sinus between dentations and flush, not raised
in sinus; petioles O-^ mm long; stipules triangular,
caducous, 0.7— 0.9 mm long. Inflorescence axillary
panicles grouped toward the branch apices, axis 1—
4 cm long, canaliculate. Flowers regularly arranged
on branches of inflorescence, each branch 1- to
several-flowered; bracteoles obovate, 1—1.5 mm

Novon 9: 520-523. 1999.

Volume 9, Number 4
1999

Jardim & Berry

Hugoniaceae from Venezuelan Guayana

521

Figure 1. Roucheria sipapoensis Jardim & F. K. Berry. — A. Habit and detail of leaf margin. — B. Flower slightly past
anthesis. — C. Ovary and styles, frontal view. — 1). Cross section of ovary. — E. Remnants of filaments and staminal
tube. Based on Maguire 27625.

522

No von

long, subtending inflorescence axes, along axes,
and subtending flowers; pedicels 1-2 mm long; se¬
pals 5, ovate, acute, 2-3 mm long in young fruit,
persistent; petals unknown, already fallen off; sta¬
mens 10; filaments 2.8 mm long, enlarged at base
where connate with base of staminal ring, staminal
ring 1.0— 1.1 mm wide; ovary 3-4 mm long, the 3
styles free, ascending and each 1—1.2 mm long,
glabrous. Fruit a three-seeded, ovoid, fleshy drupe
5-8 mm long, 5 mm broad at base; seeds angular.

Ecology and distribution. This species occurs
as a shrub or small tree near the summit of Cerro
Sipapo on rocky substrate. It is known only from
this mountain in western Amazonas state, Venezue¬
la, at altitudes between 1500 and 2000 m.

This species is different from other Roucheria
species in both habitat and morphology. It is the
only species in the genus that occurs in upper mon¬
tane areas, and it is one of two species in the genus
with subsessile, decurrent leaves, the other being
R. monsalveae from the hyperwet, lowland forests
of Bajo Caliina, Valle Department, Colombia. The
parenchyma cells of Roucheria sipapoensis in the
petiole and lamina are the thickest seen for any
species in the genus (~ 14—14.5 gm thick), and the
epidermal cells are strongly lignified, giving the
leaves an especially coriaceous texture.

Roucheria sipapoensis can be distinguished veg-
etatively from R. monsalveae by the auricular leaf
blade bases (the leaf margins are strongly inrolled
at the base), and the generally smaller blades (1.8—
8.5 X 0.5— 2.5 cm in R. sipapoensis vs. 3.5—15 X
1. 5—4.0 cm in R. monsalveae ), which also dry some¬
what olive-gray compared to the dark brown-red
drying color in R. monsalveae. Roucheria sipapoen¬
sis further differs in having paniculate, subterminal
inflorescences, as opposed to the axillary, sessile
fascicles of R. monsalveae. In the Ramirez et al.
(1999) treatment for the Venezuelan Guayana, this
species was called “ Roucheria sp. A .”

Paratypes. VENEZUELA. Amazonas: Cerro Sipapo
(Paraque) tepui, near summit, peak I, 1798 m, 12 Dee.
1948. Maguire & Politi 27625 (NY. VEN): Cerro Sipapo
in Sabana Grande, vicinity of Campo Grande, 1500 m, 21
Jan. 1949, Maguire & l>oliti 28667 (NY, VEN); Cerro Si¬
papo (Paraque) tepui among rocks. South Rim and East
Rim, 1829-1981 m, 26-28 Jan. 1949, Maguire & Politi
88626 (NY. VEN).

Hebepelalum neblinae Jardim & P. E. Berry, sp.
nov. TYPE: Venezuela. Amazonas: Rfo Ma-
warinuma, near Cerro de la Neblina in terra
firrne forest, 00°50'N, 66°10'W, 140 m, 11
Feb. 1985, B. Boom et al. 5729 (holotype, NY;
isotype, F).

Haec species a congeneris foliis apice emarginatis dis-
tinguitur.

Tree ea. 10 m tall. Mature stems frequently with
hollow pith. Leaves alternate, entire, repand, ob-
ovate, coriaceous, emarginate at apex, decurrent at
base, 7.0—15.5 cm long, 2.8— 7.8 cm wide; second¬
ary venation broquiodromous, 2-5 mm between
veins, terminating in a submarginal collecting vein,
intersecondary veins present, midvein raised abax-
ially, submarginal collecting vein irregular, 0.05—
3.0 mm from margin; repand margin with small,
inconspicuous glands; petioles slightly bulbous at
base, adaxially acanalate, 3—9 mm long; stipules
caducous. Inflorescence a terminal, many-flowered
panicle with 1 or 2 linear leaf-like bracts, to ap¬
proximately 3 cm long and 0.8 cm wide; bracteoles
triangular, 1.5— 2.0 mm long; pedicel 0.5 mm long
in flower, elongating to 2 mm long in fruit; sepals
5, ovate, acute, ca. 2 mm long, 1 mm wide; petals
undeveloped; stamens 10, of two heights united in
a basal ring 0.3 mm wide, filaments 10, five 0.9
mm long, five 0.5 mm long; anthers saggitate, bi-
valved with longitudinal dehiscence; ovary 1 mm
long; locules 5, commonly two of these aborted; pla-
centation axile; styles 5, stigma villous. Fruit a 5-
seeded drupe 5—6 mm long; endocarp bony with 5
costae and intercostal ridged processes.

Ecology and distribution. This species is
known from three localities in southern Venezuela
and northern Amazonian Brazil, all in wet, lowland,
nonflooded forest on white sand.

Hebepetalum neblinae is distinct from other spe¬
cies of Hebepetalum in the strongly coriaceous
leaves with emarginate apices. Unique features in
the leaf anatomy of this species include scattered
druse crystals in tin' leal blade, contrasting with the
typical condition in the genus of absence of crys¬
tals, and subepidermal, abaxial mucilage cells, also
absent in other species in the genus. In the Ramirez
et al. (1999) treatment for the Venezuelan Guayana,
this species was called “ Hebepetalum sp. 4.”

Paratypes. BRAZIL. Amazonas: Munieipio Sta. Isa¬
bel do Rio Negro, 0.7 km alrds da cidade de Sta. Isabel,
proximo ao antigo aeroporto, eaalinga alia, solo arenoso

humoso, 00°22'S, 64°59'W, 100 m. 10 Oct. 1987, C. A.
Ferreira 6226 (INPA).

Kk> to tuk Gknkha ok Nkothopicai. Huconiackak

la. Leaf venation broquiodromous (secondary veins

joined together in a series of conspicuous, well-
formed arches near the margin), petals clawed at
base, ovary pubescent . Hebepetalum

lb. Leaf venation camptodromous with closely par¬

allel secondary veins, not joined in a series of
conspicuous arches, petals not clawed at base,
ovary glabrous . Roucheria

Volume 9, Number 4
1999

Jardim & Berry

Hugoniaceae from Venezuelan Guayana

523

Key to the Species oe Roucheria

la. Leaves and young twigs pubescent .

. Roucheria schomburgkii Planchon

lb. Leaves and young twigs glabrous.

2a. Inflorescence an axillary fascicle.

3a. Leaf base cuneate .

. Roucheria calophylla Planchon

3b. Leaf base decurrent .

. Roucheria monsalveae A. H. Gentry

2b. Inflorescence axillary or terminal panicles,
never fasciculate.

4a. Flowers in axillary short panicles 1—4
cm long.

5a. Leaves strongly coriaceous, lamina
margin abaxially inrolled, shrubs to
small trees <10 m tall . . Roucheria si-
papoensis Jardim & P. E. Berry'

5b. Leaves not coriaceous, lamina
margin not inrolled, trees >15 m

tall . Roucheria elata Planchon

4b. Flowers in terminal panicles >5 cm
long.

6a. Flowers in glomerules .

. Roucheria laxiflora H. Winkler

6b. Flowers never in glomerulous

clusters .

.... Roucheria columbiana Hallier f.

Key to Species of Hebepetalum

la. Leaf apex emarginate .

. Hebepetalum neblinae Jardim & P. E. Berry

lb. Leaf apex acuminate.

2a. Fruit with 4 locules, 4-stylous . . . Hebepetalum
roraimense Secco & S. M. B. Silva
2b. Fruit with 5 locules, 5-stylous . . . Hebepetalum
humiriifolium (Planchon) Bentham

Acknowledgments. The first author thanks the
Missouri Botanical Garden for a Bovard fellowship,
permitting the study of these plants, and the second
author acknowledges support from the National Sci¬
ence Foundation (grants BSR-8717303 and BSR-
9201044).

Literature Cited

Bentham, G. 1862 (reprinted 1965). Genera Plantarum.

Vol. 1: 241-246. Verlag J. Cramer, Weinheim.
Cronquist, A. 1988. The Evolution and Classification of
Flowering Plants, 2nd Ed. Allen Press, Lawrence, Kan¬
sas.

Planchon, J. E. 1847. Observations sur VAmoreuxia , DC.
( Euryanthe , Cham, et Schlecht.) et description des nou-
veaux genres Roucheria et Lobbia: comme introduction
a des memoires distincts sur les Cochlos pe rrnees , Liases
et Aristolochiees, families auxquelles ces genres seront
respectivement rattaches. London J. Bot. 6: 139—144.
Ramirez, N., P. E. Berry & A. Jardim. 1999. Hugoniaceae.
Pp. 618—623 in P. E. Berry, K. Yatskievych & B. K.
Holst (editors). Flora of the Venezuelan Guayana, Vol.
5, Eriocaulaceae— Lentibulariaceae. Missouri Botanical
Garden Press, St. Louis.

A New Name and a New Lectotypification in Neotropical Plants

(Ebenaceae, Solanaceae)

Sandra Knapp

Department of Botany, The Natural History Museum, Cromwell Road, London SW7 5B D,

United Kingdom, sk@nhm.ac.uk

Abstract. A new name in Diospyros, D. pana-
mense S. Knapp, is coined to replace the homonym-
ic D. whitei S. Knapp, and the name Brunfelsia
cuneifolia J. A. Schmidt in Martius is lectotypified
correctly with an extant specimen.

The following errors have come to light and are
corrected here.

During the preparation of the account of Eben¬
aceae lor Flora Mesoamericana, two new taxa of
Diospyros were described (Knapp, 1997). One of
these, D. whitei S. Knapp, was named in honor of
the late Frank White, life-long student of the genus.
Unfortunately, the epithet had been published just
prior to that date for an African species, D. whitei
Dowsett-Lemaire & Pannell (Dowsett-Lemaire &
Panned, 1996), also in honor ol the late Frank
White. This necessitates a new name for the Me-
soameriean taxon.

Diospyros panamense S. Knapp, nom. nov. Re¬
placed name: Diospyros whitei S. Knapp, No-
von 7: 258. 1997. Not Diospyros whitei Dow¬
sett-Lemaire & Panned, Bud. Jard. Bot. Nat.
Belg. 65: 399. 1996. TYPE: Panama. San
Bias: Rio diablo y vecinidad de Duque Sui, a
unos 10 km de la costa frente a la Isla de
Nargana, ruta hacia Cerro Ibedon, 80—110 m,
9°22'N, 78°35'W, //. Herrera et al. 1175 (ho-
lotype, BM; isotypes, MO, PMA).

The new epithet is derived from the country of
origin of the only known specimens, Panama.

After the publication of the monograph of Brun¬
felsia (Solanaceae) by the late Tim Plowman (Plow¬
man, 1998), edited by S. Knapp and J. R. Press,
Fred Barrie (MO) kindly alerted to us to our (SK &
JRP) incorrect lectotypification of Brunfelsia cunei¬
folia J. A. Schmidt using a photo of a destroyed B
sheet, as specified by Plowman in his unpublished
thesis. We also contravened Article 37.5 of the
Code (Greuter et al., 1994) by not citing the her¬
barium in which the photograph was housed, ren¬
dering the lectotypification in Plowman (1998) in¬
valid.

In describing Brunfelsia cuneifolia, Schmidt
(1864) cited two specimens: “In Brasilia austral-
iore: Sellow nr. 4016 in hb. Berol. In sylvis palu-
dosis inundatisque prope S. Joz£ prov. S. Pauli:
Riedel nr. 1467 in hb. Petropol.” Reidel 1467 in
LE is clearly a specimen of B. ohovata Bentham
(Plowman, 1998). The Sellow collection in B cited
by Schmidt ( Sellow 4016) was destroyed in the Ber¬
lin herbarium, but two unnumbered Sellow collec¬
tions housed at F and M appear to be duplicates
as they “conform in every way” (Plowman, 1998)
to the photograph of the destroyed Berlin sheet.
Many Sellow duplicates were distributed without
numbers. Plowman’s treatment of the specimen in
F as an isosyntype is accepted here, and thus it is
possible to lectotypify B. cuneifolia. Plowman
(1974, 1998) indicated that he wished to lectotypify
B. cuneifolia with the destroyed B sheet, as good
photographs of it exist. This, however, is contrary
to Article 9.9 of the Code (Greuter et al., 1994),
which states that original material must be used for
lectotypification; the photographs were taken later
and did not form part of the protologue. It is also
contrary to Article 9.1 1: “If a holotype or previously
designated lectotype has been lost or destroyed and
it can be shown that all other original material dif¬
fers taxonomically from the destroyed type, a neo¬
type may be selected to preserve the usage estab¬
lished by the previous lectotypification.” Since the
previous lectotypification (Plowman, 1998) was in¬
valid, duplicates of the syntypes exist, and one of
the two syntypes of B. cuneifolia is attributable to
another taxon ( B . ohovata Bentham), it is clear the
species needs correct lectotypification to prevent
future confusion. I correctly and unambiguously
lectotypify B. cuneifolia J. A. Schmidt here, using
the unnumbered Sellow specimen at F annotated
by Plowman as “isotype” as the new lectotype.

Brunfelsia cuneifolia J. A. Schmidt, in Martius,
FI. Bras. 8(1): 259. 1864. TYPE: Brazil. “In
Brasilia australiore,” Sellow [4016] (lectotype,
designated here, F; isolectotype, M; photo-

Novon 9: 524-525. 1999.

Volume 9, Number 4
1999

Knapp

Neotropical Plants

525

graphs of destroyed sheet at B (F neg. 621824,
as Sellow 4016), F, NY, US).

Acknowledgments. I thank Alison Strugnell
(FHO) and Fred Barrie (F, MO) for alerting me to
these errors, and Mike Gilbert (MO, BM) and two
anonymous referees for careful review of the man¬
uscript.

Literature Cited

Dowsett-Lemaire, F. & C. Panned. 1996. A new Diospyros
(Ebenaceae) from the Congo Republic. Bull. Jard. Bot.
Belg. 65: 399-403.

Greuter, W., F. R. Barrie, H. M. Burdet, W. G. Chaloner,
V. Demoulin, I). L. Hawksworth, P. M. Jprgensen, 1). H.
Nicholson, P C. Silva. P Trehane & .1. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Knapp, S. 1997. Two new species of Diospyros (Ebena¬
ceae) from Mesoamerica. Novon 7: 256—260.

Plowman, T. 1974. A Revision of the South American Spe¬
cies of Brunfelsia (Solanaceae). Ph.D. Thesis, Harvard
University, Cambridge.

- . (Edited S. Knapp & .1. R. Press.) 1998. A revi¬
sion of the South American species of Brunfelsia (So¬
lanaceae). Fieldiana, Bot. n.s. 39: 1—135.

Schmidt. J. A. 1864. Scrophularinae. In: C. F. P. Martins,
Flora Brasiliensis 8(1): 230—339.

Bauhinia wuzhengyii (Leguminosae, Caesalpinioideae), a New

Chinese Species

Supee Saksuwan Larsen

Department of Systematic Botany, Institute of Biology, University of Aarhus,
DK-8240 Risskov, Denmark

ABSTRACT. Bauhinia wuzhengyii, a new species
from Yunnan, China, is described and illustrated.
Its relationship within the section Loxocalyx is dis¬
cussed. The section Loxocalyx Bentham, containing
the new species, is reinstated based on re-assess-
ment of the obliquely campanulate hypanthium and
free gynophore.

During a visit to the Chinese Kunming Institute
of Botany in May 1997, the author, who has been
invited to act as co-author of Bauhinia for Flora oj
China, went through the herbarium of the genus.
Here a specimen of an undescribed species was
discovered. This sheet had been annotated (in
1959) by the eminent Chinese botanist Wu Zheng-
yi as Bauhinia cinnamomifolia sp. nov. This name,
however, has been ascribed to section Cinnanwm-
ifolia under Phanera, a group to which this species
clearly does not belong. From discussions with
Prof. Wu it was evident that he had no plans to
describe this species. It is therefore a great plea¬
sure to dedicate this new species to him.

Bauhinia wuzhengyii S. S. Larsen, sp. nov. TYPE:
China. Yunnan: Malipo County, Tian Bas, in
thin forest on mountain slope, 480 m, 27 July
1958, Hu Yue Ying & Wen Shao Kong 580812
(holotype, KUN). Figure 1.

Bauhinia wuzhengyii species nova sectionis Ijoxocaly-
cis, Bauhiniae wallichii MacBride affinis a qua differt fo¬
lds integris, stigmate subsessile.

Liana, tendrils not seen. Flowering branches te¬
rete, glabrous. Leaf lamina entire, elliptic, 9.0— 10.3
X 4.8— 5.3 cm, glabrous both sides, base attenuate,
apex abruptly acuminate; nerves 3; stipules not
seen; petiole 2.3—3 cm long, thickened at both
ends. Inflorescence terminal, an elongated raceme,
with 2 racemes evident on type, one with a reduced
leaf at base. Inflorescence axis up to 19 cm long,
finely grayish pubescent; bracts 4—5 mm long, se¬
taceous, finely pubescent outside; bracteoles simi¬
lar hut minute, 0.5 mm long, subopposite, inserted
below the middle of the pedicel; pedicel 2.5—3 cm
long, finely grayish pubescent. Flower buds ellip¬

soid, 9—10 X 5—6 mm, finely grayish pubescent;
hypanthium obliquely campanulate, enlarged adax-
ially, ca. 4 mm, twice the length of abaxial portion
(Fig. 2A, B), apically pubescent except middle part
of adaxial margin (Fig. 2F). Calyx 3-4 mm long,
campanulate (occasionally slightly bilabiate. Fig.
2B), open in hud, split in upper half to 5 lobes,
each with acute apex, finely grayish pubescent out¬
side, inside glabrous to sparsely pubescent. Corolla
with 5 subequal petals, 4 lateral ones spathulate,
7—10 mm long including a 2— 3 -mm claw, densely
brownish pubescent on both sides, becoming gla¬
brous toward claw; posterior petal ± obovate-ob-
long, carinate, basally fleshy, ca. 2 mm, with claw
indistinct, upper portion, ca. 5 mm, finally reflexed,
inside glabrous, outside pubescent in upper por¬
tion, marginally and basally glabrous. Fertile sta¬
mens 3; filaments ca. 10 mm long, glabrous; an¬
thers 1.5-2 mm long, glabrous, dehiscing
longitudinally; staminodes 5—7, 3 posterior forming
a 3-dentate connate structure, 1.5 mm long (Fig.
2F); middle connate staminode slightly larger, in¬
flated at base to a low rim, the 2 lateral teeth in¬
serted behind the rim; 1-2 minute staminodes at
both sides of the teeth and 1-2 minute staminodes
between the fertile stamens. Pistil with thic k stipe,
free, 2—3 mm long, brownish woolly pubescent, in¬
serted at lower abaxial side of hypanthium (Fig.
2F); ovary ca. 9 mm long, brownish woolly pubes¬
cent throughout, stigma subsessile, ± punctate,
style glabrous, very short, ca. 1 mm. Ovules ca. 8.
Pods not seen.

Distribution and habitat. Oidy known from the
type locality.

Discussion

Bauhinia wuzhengyii is here described based on
a single collection in KUN, as the author was un¬
able to locate any iso- or paratypes (see Note 1).
After my experience of revising the genus Bauhinia
for Flora Malesiana and the Indochinese floras, this
new species is, however, so striking in both flower

Novon 9: 526-529. 1999.

Volume 9, Number 4
1999

Larsen

Bauhinia wuzhengyii from China

527

Hofotype

Set u4u+uet wujAen^y// S’- S'- Larstn

Dcp! of Sysl. Bol.. Univ.

Dei Supee Saksuwan Lirsen, 1W8

>upw Saksu

of Aarhus

!' « sj ,v *

1 T Y PUS

C - (j - , >f

g w « »» w as wr?

*»* -jt.

«* H JpWSfAj^-

»«

** »r*#H

»*

*

*

- IJ|yr (W sm wn UN|V Of- AARIIUs" HMUfARItiM AAII

f/o. rwt z

C*lWfy; tfu ytiC >.Vj 4- W+» >!»+> K*«J

, /Wlf

UjiaLGi . ~n<*« Mul.'j’O Y«—~>
|W4*C t La XUa Jv,

. oi-t. <> -»m

Wf£. »****»* — „

-Wift'r'V

4m. «AU

'■ ' ■

Doternihmvit: ,S.'?57 ■

Figure 1. Bauhinia wuzhengyii S. S. Larsen; holotype specimen, Hu Yue Ying & Wen Shut) Kong 580812 (KUN).

528

Novon

Figure 2. Bauhinia wuzhengyii. — A. Floral bud, lateral view, showing typical campanulate calyx. — B. Floral bud,
lateral view, showing occasional bilabiate condition. — C. Open flower (anthers missing). — D. Posterior petal, upper
surface. — E. Lateral petal, upper surface. — F. Calyx and hypanthium seen from inside showing ovary and the dentate,
staminodial body. Drawn from Hu Yue Ying & Wen Shoo Kong 580812.

and leal morphology that I am not in doubt that it
deserves specific status.

Bauhinia wuzhengyii belongs to the section Lox-
ocalyx Bentham (Svn.: series Loxocalyx (Bentham)
Wunderlin, Larsen & Larsen, 1987) together with
B. wallichii MacBride and B. japonica Maximowicz.
These three species have a tree gynophore (stipe)
inserted at the lower or upper abaxial side of the
hypanthium. The new species is easily distin¬
guished from these two other species by having en¬
tire leaves with 3 nerves, a subsessile stigma, and
a 3-dentate, staminodial structure. Bauhinia japon¬
ica has bilobed leaves (VS— % of the length), while
B. wallichii has leaves emarginate to bilobed (VS of
the length); both species have 9— 13-nerved leaves,
distinct style, and a 5-digitate staminodial struc¬
ture. Bauhinia japonica is a coastal species distrib¬
uted from Japan to Hainan, extending to a single
locality along the coast of Kwangtung (Larsen &
Larsen, 1982). Bauhinia wallichii is distributed
from India (Silhet, Assam) to Myanmar and north¬
ern Vietnam (Larsen et al., 1980). Bauhinia wal¬
lichii was recently reported from southeastern Yun¬

nan by Zhang and Chen (1996) and from northern
Thailand (S. S. Larsen, in press).

The tri-nerved leaves with entire margins of Bau¬
hinia wuzhengyii are similar to those in section
Tubicalyx Wunderlin, Larsen & Larsen (1987). De¬
spite superficial similarities between floral struc¬
tures, the Tubicalyx group is distinguished by hy-
panthia not being oblique, gynophores adnate to the
hypanthium wall, and all the staminodes being free
and not connate as a compound structure.

The pollen grains of Bauhinia wuzhengyii are 3-
eolporate (Fig. 3A) with rugulose tectum (Fig. 3B,
C). Pollen of this new species belongs to the curtisii
pollen type, grouping together with B. japonica (S.
S. Larsen, 1975) and B. wallichii (S. S. Larsen, in
press).

In conclusion, the morphology and palynology
point to this species being related to Bauhinia wal¬
lichii and B. japonica. It must, however, be empha¬
sized that very little pollen has heen available for
SEM studies. Further studies, including sectioning
of pollen, are highly needed for a comparison with
oilier members of the curtisii pollen type.

Volume 9, Number 4
1999

Larsen

Bauhinia wuzhengyii from China

529

Figure 3. Bauhinia wuzhengyii. SEM micrographs of non-acetolyzed pollen. —A. Whole pollen grain, oblique view.
— B, C. Detail of tectum at mesocolpia and apocolpia. Scale bar for A = 10/am; scale bar for B and C = 5 pm.

Note 1. Zhang (1993) reported a Bauhinia
strychnoidea Craib from Yunnan, probably referring
to B. strychnifolia Craib, as B. strychnoidea Prain
is a Malaysian species. This record is based on the
same collecting number as the type of B. wuzhen¬
gyii. Dr. Zhang (pers. comm.) has seen material ei¬
ther in Kunming (KUN) or in Xishuangbanna
(HITCB), but his determination does not appear on
the holotype specimen.

Note 2. Since the work by Wunderlin et al.
(1987), knowledge on the variation and structure of
floral characters in Bauhinia has considerably in¬
creased. Consequently, the section Loxocalyx Ben-
tham (1865) is reinstated under Bauhinia subg.
Phanera. This was earlier referred to as series Lox¬
ocalyx Wunderlin, Larsen & Larsen (1987); Zhang
(1996) raised series Loxocalyx to a subsection,
transferring it to section Lasiobema, but in this we
cannot follow him for the diagnostic reasons listed
below.

Bauhinia [subg. Phanera ] section Loxocalyx
Bentham (1865), emended description:

Differs from other sections of subgenus Phanera
by the campanulate, 5-dentate calyx open in bud;
the hypanthium being obliquely campanulate, with
the adaxial side twice as long as the abaxial and
with the posterior staminode forming a dentate or
digitate, fleshy structure; the gynophore (stipe) free,
inserted on the lower or upper abaxial side of the
hypanthium, and the stigma ± punctiform.

Acknowledgments. The author is indebted to
Hu Chi Ming for translating the Chinese texts on
the label, to Benjamin 011gaard for latinizing the
diagnosis, to Anni Sloth for technical work on pol¬
len and photography, and to Oratai Kirdkaew for
preparing the line drawing from sketches made by
the author. The stay in Yunnan was sponsored by
the Academia Sinica; a grant from the Carlsberg
Foundation covered the travel expenses to China.
Warm thanks are also extended to the staff of the
botanical institutes in Kunming and Xishuangban¬
na.

Literature Cited

Bentham, G. 1865. Bauhi nieas. In: Bentham & Hooker (..

Genera plantarum I: 575—576. London.

Larsen, K. & S. S. Larsen. 1982. Notes on some Asian
Bauhinia. Nordic .). Bot. 2: 329—332.

- , - & J. E. Vidal. 1980. Legumineuses-Cae-

salpinioidees. Flore du Cambodge. du Laos et du Vi¬
etnam 18. Museum National d’Histoire Naturelle, Paris.
Larsen, S. S. 1975. Pollen morphology of the Thai species
of Bauhinia (Caesalpiniaeeae). Grana 14: 114—131.

- . 1999. Bauhinia wallichii MacBride (Legumino-

sae-Gaesalpinioideae), a species new to Thailand. Thai
Forest. Bull. 27: in press.

Wunderlin, B. P., k. Larsen & S. S. Larsen. 1987. Reor¬
ganization of the Cercideae (Fabaceae: Caesalpinioi-
deae). Biol. Skr. 28.

Zhang, D. 1993. Some additional taxa of Bauhinia (Le-
guminosae) from China. Nordic J. Bot. 13: 399—402.

- & T. Chen. 1996. Three species of Bauhinia L.

(Leguminosae) new to China. J. Trop. Subtrop. Bot. 4(4):
16—17.

Landoltia (Lemnaceae), a New Genus of Duckweeds

Department of Ecology and

Donald H. Les

Evolutionary Biology, The University of Connecticut, Storrs,
Connecticut 06269-3043, U.S.A.

Daniel J. Crawford

Department of Plant Biology, The Ohio State University, Columbus, Ohio 43210, U.S.A.

Abstract. Morphological, allozyme, and cpDNA
(r6cL) sequence data provide evidence for the dis¬
tinctness of Spirodela punctata from species in both
Lenina and Spirodela (Lemnaceae). We propose the
recognition of a new genus, Landoltia , to better re¬
flect current phylogenetic concepts in the Lemna¬
ceae. Landoltia is distinguished by its reduced
frond prophyllum, frond nerves (3 to 7), roots (up
to 7), root tracheids, external anther locules, and
also by well-supported molecular evidence provid¬
ed by allozymes and cpDNA sequences. The new
combination Landoltia punctata is made to accom¬
modate this taxonomic modification.

Duckweed classification remains equivocal be¬
cause phylogenetic relationships are difficult to rec¬
oncile in this diminutive, reduced family (Les et
al., 1997b). The most comprehensive phylogenetic
hypotheses for duckweed taxa were presented in a
monograph by Landolt (1986) and form the basis
of modern classification for the family Lemnaceae.

In the most recent monograph of the duckweed
family (Lemnaceae), Landolt (1986) recognized four
genera: Lenina, Spirodela, Wolffia, and Wolffiella.
Few additional genera have been proposed in past
taxonomic treatments of duckweeds. One example
is Staurogeton Reichenbach, which was elevated
from subgeneric to generic rank by Schur (1866) to
accommodate the morphologically distinctive Leni¬
na trisulca L. (Landolt, 1986). Most contemporary
classifications continue to assign this taxon to the
genus Lenina. Den Hartog and van der Plas (1970)
subdivided Wolffiella to create the two genera Pseu-
dowolffia and Wolffiopsis. Few taxonomic treatments
recognize either of these segregate genera as dis¬
tinct from Wolffiella. To our knowledge, division of
either Spirodela or Wolffia into subsidiary genera
has not yet been suggested, although some authors
have transferred certain species from these genera
into either Lenina or Wolffiella. Landolt s recent ge¬
neric concept of Spirodela recognizes a paraphylet-
ie taxon, with .S’, polyrhiza and S. intermedia as sis¬

ter species, but with S. punctata associating with
Lenina (Fig. 1A).

Although Landolt’s evolutionary trees reflect a
keen understanding of duckweeds and a compre¬
hensive evaluation of published taxonomic litera¬
ture on the group, they were constructed using non¬
explicit phylogenetic methods, i.e., not by cladistic
analyses. We have re-analyzed results of these ear¬
lier studies using cladistic methodologies to test hy¬
pothetical relationships proposed by Landolt. In ad¬
dition, we have generated several molecular data
sets to supplement the clearly limited number of
phylogenetically informative characters available
for these morphologically simple plants. Although
this work remains in progress, the results of our
preliminary analyses of morphological, biochemi¬
cal, allozyme, and DNA (rftcL) sequence data
(Crawford & Landolt, 1993, 1995; Crawford et al.,
1995, 1997; Crawford et al., 1996; Les et al., 1994,
1997a, 1997b) are reasonably concordant with Lan¬
dolt’s classification, but differ in a number of details
from his phylogenetic diagrams. In particular, ge¬
neric subdivisions used in his classification of Lem¬
naceae are inconsistent with results of these phy¬
logenetic analyses and merit reconsideration.
Herein we summarize evidence that warrants the
taxonomic segregation of a new duckweed genus.
We consider this taxonomic action essential for a
classification that reasonably depicts our best, cur¬
rent estimate of phylogenetic relationships in the
Lemnaceae.

A specific diagram of intergeneric duckweed re¬
lationships that summarizes the phylogenetic trees
originally appearing in Landolt (1986) was provid¬
ed to us by E. Landolt. We compared these hypo¬
thetical relationships to published allozyme studies
(Crawford & Landolt, 1993) and to previous cla¬
distic analyses of morphological, anatomical, and
biochemical data (Les et al., 1997b). Intergeneric
duckweed relationships were also examined using
preliminary results of a phylogenetic analysis of

Novon 9: 530-533. 1999.

Volume 9, Number 4
1999

Les & Crawford
New Duckweed Genus

531

A.

Wolffia
Wolffiella
Lemna

Spirodela punctata
Spirodela polyrhiza
Spirodela intermedia

B.

Wolffia
Wolffiella
Lemna

Spirodela punctata
Spirodela polyrhiza
Spirodela intermedia

Wolffia p,

Wolffiella

Lemna

Spirodela punctata
Spirodela polyrhiza
Spirodela intermedia

Figure I. Phylogenetic position of Spirodela punctata as
indicated in several recent studies. — A. Landolt (1986)
hypothesized the association of 5. punctata with Lemna,
in a paraphyletic concept of Spirodela. — B. Non-molecular
data (Les et ah, 1997b) place S. punctata in a separate
clade from lemna , but also apart from other Spirodela
species (bootstrap % shown). — C. Molecular (rftcL) data
provide strong support (bootstrap % shown) for the dis¬
tinctness of S. punctata from either lemna or Spirodela
(Les et al., 1997b). All evidence points to an isolated po¬
sition of S. punctata in the Lemnaceae, and its recognition
as a distinct genus is compatible systematically with any
ol these results.

rbcL sequence data for the Lemnaceae (Les et al.,
1997b).

Non-molecular data (Fig. IB) resolve the entire
genus Spirodela as paraphyletic. However, S. punc¬
tata lies distinct from the other Spirodela species
as a separate branch, and with good internal sup¬
port (75% bootstrap value). Strong bootstrap values
(99% and 84%, respectively) from molecular (rbcL)
data (Fig. 1C) support S. polyrhiza and S. inter¬
media as sister species, and also their distinction
from S. punctata. Allozyme data (Crawford & Lan¬
dolt, 1993) show a moderate genetic identity be¬
tween S. polyrhiza and .S’, intermedia (/ = 0.404),
yet they share no electrophoretically detectable al¬
leles with S. punctata (/ = 0.000). Various mor¬
phological features (Table 1) are consistent with a
phylogenetic position for S. punctata intermediate
between Lemna and other Spirodela species. By in¬
spection of these features, the species S. punctata
is not only morphologically distinct from both Lem¬
na and Spirodela, but intermediate and transitional
between these genera.

Presently, Spirodela Schleiden comprises three
distinct species: Spirodela intermedia W. Koch, S.
polyrhiza (L.) Schleiden, and S. punctata (G. Meyer)
C. H. Thompson. Landolt (1986) placed the former
two species into Spirodela sect. Spirodela, and seg¬
regated the latter in section Oligorrhizae W. Koch.
Spirodela punctata has been taxonomically prob¬
lematic because it possesses features similar to
both Spirodela and Lemna (Table 1; Landolt, 1986).

Meyer (1818) originally named Spirodela punc¬
tata as Lemna punctata , but it was not until 50
years later that Hegelmaier (1868) transferred the
taxon (as S. oligorrhiza) to Spirodela. In Meyer’s
time, all Lemnaceae were included in the genus
Lemna. The new genera Spirodela and Wolffia were
created in 1839 and 1844, respectively, and Wolf¬
fiella was established in 1895 (Landolt, 1986). It is

Table 1. Morphological features compared among species of Spirodela and Lemna. Spirodela punctata is interme¬
diate between Lemna and other Spirodela species for the character states indicated (from Landolt, 1986, 1998; Shill,
1979).

Feature

5. intermedia; S. polyrhiza

S. punctata

lemna

Prophyllum at hase of frond

present

present, but reduced

absent

No. of veins in frond

7 to 16

3 to 7

1 to 5

No. of roots

7 to 2 1

1 to 7 (12)

1

Boot tracheids

extend to tip

basal only

absent

Dorsal ineristem of new
fronds

on one side1

on both sides

on both sides

Lxternal anther locules

do not extend above inter¬
nal locules

extend slightly above inter¬
nal locules

extend above internal loc¬
ales

1 Lateral on other side.

532

Novon

tht- merit of Hegelmaier (1868, 1895) to have clear¬
ly separated the genera Lenina and Spirodela. He
also demonstrated accurately and comprehensively
the special position of S. punctata (called S. oli-
gorrhiza ) within the genus Spirodela.

Spirodela punctata is very polymorphic in rela¬
tion to size, pigmentation, number of roots, and
veins. The high level of variability led Hegelmaier
(1895) to distinguish two species (S. oligorrhiza and
5. pusilla ) and to describe three other species of
questionable status. Growth experiments (Landolt,
1986; Landolt & kandeler, 1987) and allozyme
studies (Crawford & Landolt, 1993) have demon¬
strated that genetic variation in S. punctata is rath¬
er limited, and many of the differences used to dis¬
tinguish former taxa are induced environmentally.

Thompson transferred L. punctata to Spirodela in
1898. Landolt (1986) observed that some authors
have merged the genera Lemna and Spirodela be¬
cause of their similar appearance and because the
features of S. punctata are transitional between the
genera. Yet, he clearly differentiated Spirodela (in¬
cluding 5’. punctata) from Lemna by the reduced
prophyllum at the base of its fronds, druse crystals,
pigment cells, multiple roots, better developed tra-
cheids, and other anatomical/morphological fea¬
tures (Landolt, 1986). To our knowledge, the ge¬
neric distinctness of S. punctata from both
Spirodela and Lemna has not been suggested pre¬
viously.

We have now examined relationships of duck¬
weed genera using morphological, anatomical, fla-
vonoid, allozyme, and rbc L sequence data. As sum¬
marized in Figure 1, analyses of these data sets
indicate that Spirodela punctata represents an iso¬
lated clade distinct from both Spirodela and Lemna.
Cladograms constructed from either morphological
or flavonoid data (or their combination) show high
internal support (75—97% bootstrap values) for the
distinctness of S. punctata from section Spirodela
(Les et ah, 1997b) and support Landolt’s original
phylogenetic concept that recognized Spirodela as
paraphyletic with respect to the position of S. punc¬
tata (Landolt, 1986).

Allozyme data (Crawford & Landolt, 1993, & un¬
published) indicate a complete lack of genetic
identity between S. punctata and any species in
either Lemna or Spirodela , yet the two species of
Spirodela sect. Spirodela do retain a moderate ge¬
netic; identity. Chloroplast DNA (r6cL) sequence
data (Les et ah, 1997b) resolve S. punctata in a
clade between Spirodela and Lemna but not within
either genus. In summary, these results echo the
transitional nature and unsettled taxonomic status
of .S', punctata manifest in prior systematic inves¬

tigations. Phylogenetically, our broad-based/wide-
ranging studies indicate that S. punctata is indeed
transitional between, but not a member of either
Lemna or Spirodela. It is for this reason that we
establish a new genus to better reflect this revised
hypothesis of duckweed relationships.

Landoltia D. H. Les & D. J. Crawford, gen. nov.
TYPE: Lemna punctata G. Meyer: Prim. FI.
Esseq. 262. 1818. = Landoltia punctata (G.
Meyer) D. H. Les & D. J. Crawford.

Herbae ex radieibus 2 ad 7 (raro 1 vel 8 ad 12) usque
ad 7 cm longis, omnibus prophyllum perforantibus; tur-
ionibus absentibus. Frondes in summa aqua natantes, ova-
tae ad lanceolatae, 1 .5— 2.0-plo longiores quam latiores,
supra nitidae viridesque serie mediana papillarum orna-
tae, subtus laeves rubraeque; nervis 3 ad 7. Flores infre-
quentes; antherae loculis externis super internos positis.
Fructus ala laterali in parte supera praedita; seminibus I
vel 2 manifeste 10 ad 15-eostatis.

Roots 2 to 7 (rarely 1 or 8 to 12), up to 7 cm
long, all perforating the prophyllum. Turions ab¬
sent. Fronds floating on the surface of the water,
ovate to lanceolate, 1.5—2 times longer than wide,
above shining and green with a medial series of
papillae, below smooth and red; veins 3 to 7. Flow¬
ers infrequent; external locules of the anther above
the internal locules. Upper part of fruit with a lat¬
eral wing; seeds 1 or 2 with 10 to 15 distinct ribs.

Landoltia punctata (G. Meyer) D. H. Les & I). J.

Crawford, comb. nov. Basionym: Lemna punc¬
tata G. Meyer, Prim. FI. Esseq. 262. 1818.
Spirodela punctata (G. Meyer) C. H. Thomp¬
son, Rep. (Annual) Missouri Bot. Gard. 9: 28.
1898. TYPE: Chile. Tierra del Fuego Island,
Orange Harbor, leg. Wilkes expedition 1838
(neotype, US not seen; isoneotypes, DS, GH,
KANU, MO not seen).

Lemna oligorrhiza kurz, J. Finn. Soc.. Bot. 9: 267. 1866.
Spirodela oligorrhiza (Kurz) Hegelmaier, Die l.em-
naceen 147. 1868. TYPK: India. Calcutta, Kurz IH65
(holotype, CAF? not seen; isotypes, k, MEF, FI not
seen).

The generic name Landoltia commemorates Eli¬
as Landolt for his outstanding contributions to the
systematies and biology of Lemnaceae in his more
than 45 years of research on duckweeds.

Acknowledgments. We thank D. Keil for his as¬
sistance in preparing the Latin diagnosis and R.
Rutishauser for his helpful comments. This work
was supported in part by NSF grant DEB-9806537
to the authors.

Volume 9, Number 4
1999

Les & Crawford
New Duckweed Genus

533

Literature Cited

Crawford, I). J. & E. Landolt. 1993. Allozyme studies in
Spirodela (Lemnaceae): Variation among nonspecific
clones and divergence among the species. Syst. Bot. 18:
389-394.

- & - . 1995. Allozyme divergence among

species of Wolffia (Lemnaceae). PI. Syst. Evol. 197: 59—
70.

- . - . I). H. Les & E. Tepe. 1995. Allozyme

divergence among species of Wolffiella (Lemnaceae).
Amer. J. Bot. 82 (6, suppl.): 122.

- . - & - . 1996. An allozyme study of

two sibling species of Lemna (Lemnaceae) with com¬
ments on their morphology, ecology, and distribution.
Bull. Torrey Bot. Club 123: 1-6.

- . - , - & E. Tepe. 1997. Allozyme var¬
iation and the taxonomy of Wolffiella (Lemnaceae).
Aquatic Bot. 58: 43—54.

Den Hartog, C. & F. van der Plas. 1970. A synopsis of
the Lemnaceae. Blumea 18: 355—368.

Hegelmaier, F. 1868. Die Lemnaceen. Eine monographis-
che Untersuchung. Engelmann, Leipzig.

- . 1895. Systematisch Uebersicht der Lemnaceen.

Bot. Jahrb. 2 1 : 268—305.

Landolt, E. 1986. The family of Lemnaceae — A mono¬
graphic study, vol. I. Veroff. Geobot. Inst. ETH Stiftung
Rifbel Zurich 71: 1—566.

- . 1998. Anatomy of the Lemnaceae (duckweeds).

Pp. 1—122 in E. Landolt, I. Jager-Zurn & If. A. A.
Schnell (editors), Handbuch der Pflanzenanatornie,
Band 13, Teil 4: Extreme Adaptions in Angiospermous
Hydrophytes. G. Borntraeger, Berlin, Germany.

- & R. Kandeler. 1987. The family of Lemnaceae —

A monographic study, vol. 2. Veroff. Geobot. Inst. ETH
Stiftung Rifbel Zurich 95: 1-638.

Les, D. H., E. Landolt & D. J. Crawford. 1994. Molecular
systematics of the Lemnaceae. Amer. J. Bot. 81 (6,
suppl.): 168-169.

- , - & - . 1997a. Systematics of Lem¬
naceae: Inferences from micromolecular and morpho¬
logical data. PI. Syst. Evol. 204: 161—177.

- , D. J. Crawford. E. Landolt, R. Aakjar & E. Tepe.

1997b. Systematics of Lemnaceae revisited. Amer. J.
Bot. 84 (6, suppl.): 211.

Meyer, G. E W. 1818. Primitiae Florae Essequeboensis
adjectis descriptionibus centum circiter stirpium nova-
rum, observationibusque criticis. H. Dieterieh, Gottin¬
gen.

Schur, P. J. F. 1866. Enumaratio plantarum Transylvaniae.
Vindobonae.

Shih, C. Y. 1979. SEM studies of the flowering of duck¬
weed, Lemna perpusilla , 6746. Scan. Electron Micros¬
cop. 1979: 479^186.

Thompson, C. H. 1898. A revision of the American Lem¬
naceae occurring north of Mexico. Rep. (Annual) Mis¬
souri Bot. Gard. 9: 21—42.

A New Combination in Gabonese Synsepalum (Sapotaceae)

Gordon McPherson

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Lee J. T. White

Wildlife Conservation Society, 185th St. & Southern Blvd., Bronx,
New York 10460-1099, U.S.A.

Abstract. Gambeya nyangensis is shown to have
been based on a mixed collection. The name is
lectotypified and the species, with an amended de¬
scription, is transferred to Synsepalum.

Gambeya nyangensis Pellegrin was recognized in
1924 and illustrated four years later (Pellegrin,
1928), based on Le Testu 1 786 , a collection from
the Tchibanga region of southwestern Gabon. In¬
cluded in this gathering were flower-bearing
branchlets as well as detached seeds. Pellegrin’s
assumption that these two elements represented
one species appears to have misled him in his ge¬
neric placement of the species. He commented that
in its general habit and reticulate loliar nervation,
G. nyangensis approached G. africana (G. Don)
Pierre, but offered no other justification for his rec¬
ognition of the new species in Gambeya. However,
the fact that the seeds of both G. africana and G.
subnuda (Baker) Pierre (type species of Gambeya)
resemble those assumed to belong to his new taxon
in having a linear hilum may also have influenced
his decision. That some doubt about it existed in
his mind is suggested by his statement that the
flowers of G. nyangensis are very similar to those
of Chrysophyllum sect. Zeyherella, in particular
those of C. longipedicellatum De Wildeman.

Aubreville (1961) stated flatly that G. nyangensis
is not really a Gambeya. Nevertheless, lacking the
intact fruit, fie felt it best to leave the species there,
while echoing Pellegrin’s thoughts on Zeyherella
and adding Pachystela and Ecclinusa to the list of
generic candidates. Aubreville correctly pointed
out that the seed of Le Testu 1786 precludes the
use of either Pachystela or Zeyherella.

Recent collecting in central Gabon has yielded
specimens that resolve this puzzle. In January 1983
researchers at the Station d’Etudes des Gorilles et
Chimpanz£s in the Lope Reserve found chimpanzee
dung containing a Sapotaceae seed that did not cor¬
respond to any species illustrated by Aubreville
(1961). Seeds of the same species were found sev¬

eral times over the years, during November— Janu¬
ary, in both chimpanzee and gorilla dung (White &
Abernethy, 1997: 186), but it was not until Novem¬
ber 1990 that a fruiting specimen was located and
collected. It proved impossible to identify this spec¬
imen, and a number of trees were monitored for
flowering activity. In June 1996 flowering material
was collected that matches very well the flower¬
bearing branchlets of the type of G. nyangensis.
Since the recently collected seeds have a much
wider hilum than do those of the type (>15 mm vs.
ca. 3 mm), it is now evident that the type collection
is a mixture, and the name is lectotypified below
in accordance with Articles 9.9 and 9.10 of the
Code (Greuter et al., f994).

In Aubrevilles key to genera in the Flore du
Gabon , the recent material would lead to Afroser-
salisia, but that genus is now recognized as a syn¬
onym of an inclusive Synsepalum, based on the
convincing work of Pennington (1991). Other Ga¬
bonese genera similarly now included in Synsepal¬
um are Pachystela, Tulestea, and Vincentella. The
necessary combination in Synsepalum is made be¬
low, and an amended description of the species is
presented.

Synsepalum nyangense (Pellegrin) McPherson &
L. White, comb. nov. Basionym: Gambeya
nyangensis Pellegrin, Bull. Mus. Hist. Nat.
(Paris) 1924: 327. TYPE: Gabon. Mayombe
bayaka, Tchibanga, 16 Sep. 1914, Le Testu
1786 (lectotype, designated here, P flowering
branchlets; isolectotype, MO).

Amended description: Tree 20—35 m, 40-80 cm
DBH, sometimes with narrow buttresses 2—4 m
high, trunk somewhat fluted, twisted, or irregular.
Bark orange-brown or orange-gray, flaky. Exudate
white. Leaves mostly clustered at ends of branch-
lets, these appressed-pubescent but quickly gla-
brate, 5-8 mm diam.; blades obovate or oblong,
(1 1 — ) 1 5—2 5(— 3 1 ) cm long, (3.7-)5.5-10(-13) cm
wide, base attenuate and smoothly merging with the

Novon 9; 534-535. 1999.

Volume 9, Number 4
1999

McPherson & White
Gabonese Synsepalum

535

petiole, apex rounded but with a short (to 5 mm),
obtuse acumen in undamaged leaves, midrib prom¬
inent abaxially, much less so adaxially, lateral veins
(9—) 10— 12 on each side ol the midrib, the finer ve¬
nation subparallel and visible on both surfaces,
these at first appressed-pubescent but quickly gla-
brate; petiole (1.5— )2— 4.5 cm long, glabrate; stip¬
ules absent. Flowers in fascicles in the axils of fall¬
en leaves; pedicels 4—5 mm, accrescent in fruit to
10-15 mm long, pubescent; calyx of 5 (rarelv 6)
sepals, 3 mm long, 2 mm wide, fused one-third to
one-half of length, fawn, pubescent abaxially anti
adaxially; corolla of 5 (rarely 6) petals, glabrous,
corolla tube 1 mm long, corolla lobes 1 .5—3 mm
long, imbricate in bud, green; stamens 5, inserted
on the throat of the corolla tube, filaments ca. 1
mm long, usually kinked centrally, anthers apicu-
late, ca. 1 mm long; staminodes absent; ovary 5-
locular, pubescent with pale brown hairs, style gla¬
brous. Fruits 4—5.7 cm long (on drying), 2.5—3 cm
diam., yellow-orange or yellow, flesb translucent
white, sweet; seed 3.4— 4.7 cm long including 2—3
mm apiculum, ca. 2 cm wide, ca. 1.5 cm thick,
somewhat laterally compressed, hilum occupying
ca. two-thirds of the seed surface, its surface shal¬
lowly reticulate, testa smoother but not shiny, em¬
bryo with plano-convex cotyledons, endosperm ab¬
sent.

Flowering is in June-July, and fruits ripen from
November to January. Immature seeds are eaten by
black colobus, Colobus satanas. Ripe fruits are con¬
sumed in large quantities by gorillas ( Gorilla g. go¬
rilla) and chimpanzees ( Pan t. troglodytes ), both of
which swallow seeds intact and disperse them in
their dung, and by smaller primates that may trans¬
port seeds short distances from the parent tree.

The trees occur in patches in mature forest, often
along ridge lines at ca. 500-700 in throughout the
Lope Reserve, reaching densities of 3-5 adult trees
per hectare on Mount Yindo, at 00°13'S, 11°34'E,
where they are particularly common.

Specimens studied. GABON. Nyanga: Mayombe bay-
aka, Tchibanga, 16 Sep. 1614, Ac Testa 1786 (MO. 1’).
Ogooue-Ivindo: Lope-Okanda Reserve, bill SW of
SEGC. ca. 200 in. 15 Nov. 1993, McPherson 16189 (LBV.
MO): I, ope Reserve, West Woods Transect, open forest, 25
Nov. 1990. L. White 0228 (MO); Lope Reserve, Camel
Ridge, Nov. 1993, L White 1 172 (MO); Lope Reserve,
Camel Ridge, mature forest, 600 m. 17 Nov. 1995, A.
White 1527 , 1528. 1529 (all MO); Lope Reserve. Camel
Ridge, mature forest, 700 m, 25 June 1996. A. White 1597
(MO).

Acknowledgment. This work was supported in
part by National Science Foundation Grant No.
BSR -9024745.

I .iterature Cited

Anbreville, A. 1961. Sapotacees. In Flore du Gabon I: 1 —
162. Museum national d’Histoire naturelle, Paris.

G renter. W., F. R. Barrie. II. M. Burdet, Vi. G. Chaloner.
V. Demoulin, D. L. Hawksworth, P. M. Jorgensen. I). H.
Nicholson, P. C. Silva. P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Pellegrin, F. 1924. Plantae Letestuanae novae ou Plantes
nouvelles reeollees par M. Le Testu de 1907 a 1919
dans le Mayombe congolais. Bull. Mus. Hist. Nat. (Par¬
is) 1924: 327.

- . 1928. La Flore du Mayombe d'apres les reeoltes

tie M. Georges Le Testu 2: 14. Le Tendre, Caen.
Pennington, T. I). 1991. The Genera of Sapotaceae. Royal
Botanic Gardens, Kew, Richmond, and New York Bo¬
tanical Garden, New York.

White, I.. & K. Abernethy. 1997. A Guide to the Vege¬
tation ol the Lope Reserve. Wildlife Conservation So¬
ciety, Multipress-Gabon, Libreville.

Two New Brazilian Velloziaceae, Vellozia auriculata and Vellozia
gigantea, and a Key to the Related Dracenoid Species of Vellozia

Renato de Mello-Silva and Nanuza Luiza de Menezes
Departamento de Botanica, Instituto de Biociencias, Universidade de Sao Paulo,
Cx. Postal 11461, 05422-970 Sao Paulo, SP, Brazil

Abstract. Two new Species of Velloziaceae, Vel¬
lozia auriculata and V. gigantea, from Espinhago
Range, Minas Gerais, Brazil, are described. Both
are characterized by a large dracenoid habit. Vel¬
lozia auriculata has large tepal appendages and is
endemic to the Ambrosio Range, northeast of the
Diamantina plateau. It reproduces in an unusual
way, with conspicuous vegetative propagation by
lodging branches. Vellozia gigantea may attain a
height of up to 6 m, being the largest species of
Vellozia so far described. Its one known population
is located within the limits of the National Park of
Serra do Cipo. Morphological and anatomical de¬
scriptions as well as illustrations of both species
are presented. Comments are made on their geo¬
graphic distribution and putative taxonomic rela¬
tionships. A key to the accepted species with dra¬
cenoid habit related to Vellozia auriculata and V.
gigantea, and comments on the delimitation of this
group are also presented.

Since the revision by Smith and Ayensu (1976),
many additional Brazilian Velloziaceae have been
described (Smith & Ayensu, 1979, 1980; Smith,
1985, 1986; Mello-Silva & Menezes, 1988; Mello-
Silva, 1991, 1994, 1996, 1997; Menezes & Semir,
1991). This suggests that there are still a large
number of undescribed species. Factors that sup¬
port this conclusion are the restricted distribution
of most of the species as well as the difficult access
to wide areas of the Espinha^o Range in central
Brazil, where the family reaches maximum diver¬
sity. The two species herein described are spectac¬
ular, both in terms of their morphology and their
geographical distribution, constituting examples of
the richness and uniqueness of the flora of the Bra¬
zilian campos rupestres (for maps, see Mello-Silva,
1994: 273, and Mello-Silva & Pirani, 1994: 149).

Vellozia auriculata and V. gigantea can be in¬
cluded in the group of Vellozia species with dra¬
cenoid habits, reaching more than 2 m high. Other
dracenoid members of this group are V. compacta
Martius ex Schultes & Schultes f., V. glabra J. C.
Mikan, V. piresiana L. B. Smith, and V. spiralis L.

B. Smith (Mello-Silva, 1995). Their stems are
thickened at the base with leaves apically concen¬
trated and soon deciduous, leaving persistent leaf
sheaths. The flowers are long-pedicellate with violet
tepals, numerous stamens (at least 18), and con¬
spicuous staminal appendages, and the capsules
have apical dehiscence. In addition, the anatomical
structure of the leaf lamina is quite characteristic
of the group, presenting an aquiferous hypodermis
extending adaxially to bundle sheaths and furrows,
as well as fibro-vascular bundles with two phloem
strands. For phylogenetic interpretation of these ex¬
ternal and anatomical characters, see Menezes et
al. (1994) and Mello-Silva (in press).

KkY TO THU AcCKITKI) Sl’KClKS Oh THE DkACENOII)

Group ok Vei.io/.ia

la. Leaf lamina arcuate; peduncle and hypanthium
always smooth.

2a. Plants to 6 m tall; leaf arrangement densely
imbricate; leaf sheaths brown but apically
cinereous, obliquely truncate and sharp
pointed at apex (Serra do Cipo) ... V. gigantea
2b. Plants to 3 m tall: leaf arrangement laxly
imbricate; leaf sheaths lustrous brown
throughout, evenly truncate at apex (Serra
do Cipd to Diamantina) . V. glabra

lb. Leaf lamina plane; peduncle and hypanthium
with emergences, rarely tlx* hypanthium solely
smooth.

3a. Margins of leaf lamina smooth, serrate only

at the base . b piresiana

3b. Margins of leaf lamina serrate throughout.

4a. Outer tepals auriculate; hypanthium
and capsule spheroid to obloid, ± as
long as broad (Serra do Ambrosio) . . .

. V. auriculata

4b. Outer tepals not auriculate: hypanthium
and capsule oblong to obtriangular, lon¬
ger than broad (Ouro Branco to Grao-
Mogol).

5a. Apex of leaf sheath and base of
leaf lamina with short-eiliate, en¬
tire margins; hypanthium evenly
covered with emergences (Grao-

Mogol) . V. spiralis

5b. Apex of leaf sheath and base of leaf
lamina with serrate margins; hy¬
panthium generally with emergenc¬
es only at angles and base, if pres-

Novon 9: 536-54 1 . 1999.

Volume 9, Number 4
1999

Mello-Silva & Menezes
Brazilian Velloziaceae

537

ent (Ouro Branco to Diamantina)
. V. compacta

Vellozia aurieulata Mello-Silva & N. L. Menezes,
sp. nov. TYPE: Brazil. Minas Gerais: Mun. Rio
Vermelho, Pedra Menina, Morro do Ambrosio,
31 Mar. 1985 (fl, fr), N. L. Menezes, A. M.
Giulietti, M. G. L. Wanderley, M. G. Sajo &
M. Meguro in CFCR 7678 (holotype, SPF; iso¬
types, BHCB, K, MBM, MO, SR US). Figures
1A-I, 3C-E.

Ab omnibus speciebus familiae tepalis extemis auri-
culatis optime distincta. Inter affines hypanthio capsu-
laque spbaeroideis vel obloideis diagnoscitur.

Solitary, dracenoid plants, 1.5— 3.5 m tall. Stems
much-branched, to 20 cm diam. at base and 1.5-
2.8 cm at apex. Leaves spirotristichous, straight;
leal sheaths brown; leaf lamina plane, 14—33 cm X
8—13 mm, linear-triangular, glabrous, soon decid¬
uous, the margins serrate. Flowers 1 to 3, biseriate
with petaloid perianth; peduncles 5—13 cm long,
with glandular emergences toward apex; hypanthi-
um 1.2— 1.5 cm long, 1.0— 1.3 cm diam., spheroid
or obloid, trigonous, glabrous except for few small
glandular emergences at base and apex. Tepals ca.
5 X 2.5 cm, elliptic-lanceolate, violet, glabrous, the
outer ones auriculate, the auricles 1.0— 1.8 X 0.7-
1.5 cm, obscuring the hypanthium. Stamens 36,
with purple lacerate appendages, filaments 0. 7—1.0
cm long, violet, anthers ca. 1.5 cm long, yellow;
style 3.5— 4.0 cm long, violet, stigma 3—5 mm diam.,
trilobate, yellow. Capsule 1. 8-2.0 cm long, 1.8— 2.0
cm diam., dehiscence by apical slits on the loculi.
Leaf anatomy (Fig. 3C— E). Lamina dorsiventral.
Trichomes absent but some emergences present
(Fig. 3C, arrow). Cuticle thickened on both surfac¬
es. Abaxial furrows about one-half thickness of
lamina, papillate. Stomata confined to deep blade
furrows, tetracytic. Adaxial epidermis 2— 3-seriate,
with fiber strands separated by non-thickened cells;
abaxial epidermis 1-2-seriate, with fibers. Aquif¬
erous 1-seriate hypodermis present on both surfac¬
es, but less developed abaxially, extending to
furrows as a bridge of bulliform cells. Palisade me-
sophyll 3—4 cell layers thick, abaxially merging
with spongy parenchyma. Fibro-vascular bundle
surrounded by endodermis as leaf-bundle sheaths.
Three passage cells of endodermis flanking the xy-
lem with thickened walls facing the transfusion tra-
cheids (Fig. 3E). Endodermis extending adaxially
and abaxially to hypodermis and involving pericy-
cle fibers on both sides of bundles. Phloem strands
2, protophloem separated by parenchymatous cells
(Fig. 3E, arrow). Metaxylem with two lateral exten¬
sions, the transfusion tracheids.

Vellozia aurieulata is found in the Ambrosio and
Bocaina Ranges, dividing the Rio Jequitinhonha
and Rio Doce basins, within the municipalities of
Rio Vermelho and Itamarandiba, Minas Gerais. The
new species is locally abundant, and in the Am¬
brosio Range is a significant floristic feature (Me¬
guro et al., 1994; Pirani et al., 1994). It appears to
be endemic to that region. Its vegetative reproduc¬
tion is an obvious feature in this Minas Gerais land¬
scape. Very tall plants (to 2 m) eventually spread
their branches on the ground; these branch apices
then become erect with adventitious roots devel¬
oping. These self-sufficient branches eventually de¬
tach from the mother plant, and it is common to
find young individuals, generated by this process,
surrounding an old or already dead plant (Fig. 1A).
This same phenomenon was also observed by Mor-
awetz (1983) for Vellozia cf. glochidea Pohl (prob¬
ably V. dasypus Seubert) in the Itapoa restinga, Ba¬
hia, and by Mello-Silva (1991) for V. abietina
Martius, V. maxillarioides L. B. Smith, V. minima
Pohl, and V. scoparia Goethart & Henrard, in Minas
Gerais, and in a much more specialized way in V.
prolifera Mello-Silva, from Grao-Mogol, Minas Ger¬
ais. Vellozia aurieulata differs from all similar spe¬
cies by its auriculate tepals and by the spheroid to
obloid hypanthium and fruit. These auricles of the
outer tepals occur nowhere else among the Vello¬
ziaceae. However, their adaptative significance, it
one exists, is obscure.

Paratypes. BRAZIL. Minas Gerais: Mun. Itamaran¬
diba, Penha de Franga, elev. 1040 m. I 1 Mar. 1995 (st),
S. Splet 865 (SPF, LJB not seen): Mun. Rio Vermelho, Ped¬
ra Menina. Serra do Ambrosio, Fazenda Varzea do Anjo,
Morro do Ambrosio, 15 July 1984 (fr), M. G. L. Wanderley,
E. Varanda, A. Furlan. A. M. Giulietti & R. M. Harley in
CFCR 4507 (K. not seen, SPF, U); Espigao do Meio, 13
Oct. 1984 (fr), R. Mello-Silva. J. R. Pirani, E. M. Isejima
& M. Meguro in CFCR 5464 (F, SPF, UB); 830 m alt., 8
Sep. 1986 (fr), N. L. Menezes, R. Mello-Silva, T. B. Cav¬
alcanti, I. Cordeiro & J. C. C. Gongalves in CFCR 10227
(CTES, G. SP, SPF), 10 June 1991 (fl), R. Mello-Silva , M.
L. F. Salatino , A. Salatino & P. Affonso 413 (SPF); Morro
da Virada do Mato Virgem, 31 July 1985 (fr), R. Mello-
Silva, J. R. Pirani & M. Meguro in CFCR 7837 (NY. SPF,
LEG).

Vellozia gigantea N. L. Menezes & Mello-Silva,
sp. nov. TYPE: Brazil. Minas Gerais: Santana
do Riacho, Parque Nacional da Serra do Cipo,
proximo a sede do IBAMA do Alto do Palacio,
24 Mar. 1989 (fl, fr), R. Mello-Silva, J. R. Pir¬
ani & M. Meguro in CFSC 11319 (holotype,
SPF; isotypes, K, MBM, MO). Figures 2A— E,
3 A, B.

Ab omnibus speciebus familiae habitu giganteo optime
distincta. Affinis est V glabrae a qua foliis imbricatioribus

538

Novon

figure 1. Vellozia auriculata Mello-Silva & N. L. Menezes. — A. Habit. — 15. Flowering shoot. — C. Leaf margin. —
I). Glandular emergences on pedicel. — E. Glandular emergences on hypanthium. — F. Outer tepal. ahaxial surface.
— G. Inner petal, adaxial surface, with stamens and lacerate appendages. — II. Style apex and trilobate stigma. — I.
Fruiting shoot. A. from Meguro et al. (1994). 15 II. Menezes in CFCR 767H. I. Mello-Silva in CFCR 5464.

Volume 9, Number 4
1999

Mello-Silva & Menezes
Brazilian Velloziaceae

539

Figure 2. Vellozia gigantea IN. I,. Menezes & Mello-Silva. — A. Habit. — II. Fruiting shoot. — C. Leaf margin. D.
Flower. — E. Longitudinal fertile seetion showing hypanthium, axile placentation, and two of the three locules typical
of the family; an entire stamen and lacerate appendages, style and stigma; tepals removed. A. from a photograph by
Mello-Silva. B, E, Mello-Silva in CFSC 11319. C. Menezes 1327. D, Pereira 993.

540

Novon

figure 3. Cross section of median region of lamina. — A, B. Vellozia giganlea, from Menezes 1327. — C— E. Vellozia
auriculala, from Menezes in Cb'CR 7673. — En. Endodermis. — Mp. Metaxylem. — Pc. Passage cell. — Pp. Protophlo¬
em. — Px. Protoxylem. — St. Stomata. — Tt. Transfusion tracheids. — C. arrow, indicating emergence. — E. arrow1,
indicating parenchyma cells. Seale bar = 200 yam (A. C). 100 yam (B. I)). .">() yum (E).

vaginae foliornm apieibus cinereis ol>li<|iie trnnealisqne
differt.

Solitary, dracenoid plant, 1.5—6 m tall. Stem much
hranclted, to 60 cm diam. at base and 3—5 cm at the
apex. Leaves spirotristichous, arcuate; leal sheaths

brown, cinereous toward the apex; leal lamina 1 3—15
X 0.8-1. 5 cm, linear-triangular, glabrous, soon decid¬
uous, the margins serrate. Flowers 1 to 4; peduncles
smooth, 9—25 cm long, glabrous; hypanthium smooth,
2—3 cm long, 0.8-1 cm diam., eylindrie-trigonous.

Volume 9, Number 4
1999

Mello-Silva & Menezes
Brazilian Velloziaceae

541

glabrous. TepaLs 6-9 X 1—1.5 cm, narrowly elliptic,
violet, glabrous. Stamens 36, with purple lacerate ap¬
pendages basally, filaments 5—8 mm long, violet, an¬
thers 2.5— 3.0 cm long, yellow; style 4—5 cm long,
whitish, stigma ca. 6 mm diam., trilobate, yellow. Cap¬
sules 2.5— 3.5 cm long, 1.2— 1.8 cm diam., dehiscence
by large apical slits on the loculi. Leaf anatomy (Fig.
3 A, B). Similar to Vellozia auriculata except as noted.
Trichomes and emergences absent. Abaxial furrows
only about one-third thickness of lamina, papillate.
Adaxial epidermis 3—t-seriate, with fiber strands sep¬
arated by non-thickened cells; abaxial epidermis 1-
2-seriate, 4-seriate at entrances of furrows, with fibers.
Aquiferous 2-seriate adaxial hypodermis extending to
furrows as bulliform cells; abaxial hypodermis only 1-
seriate. Palisade mesophyll 4—5 cell layers thick.
Phloem strands 2, protophloem separated by fibers.

Vellozia gigantea appears to be extremely rare,
much more so than V auriculata. There is only one
known population, within the Cipo Range National
Reserve, in Minas Gerais. Individuals up to 6 m
tall inhabit a sandstone outcrop of ca. 20,000 m2
on the northwestern edge of the park, thus forming
a unique landscape. These plants do not regularly
flower, anti the production of new individuals ap¬
pears to be limited. Although the plants are pro¬
tected by the Park limits, the frequent visiting of
tourists threatens this population. Vellozia glabra ,
found in the Serra do Cipo and in the Diamantina
plateau, is the species most similar to V. gigantea.
Both possess smooth, oblong hypanthia and long,
arching leaves. Other related species differ with
more or less glandular hypanthia and more or less
straight leaves. Vellozia gigantea can be distin¬
guished from V. glabra by its leaf arrangement,
which is more imbricate; its leaf sheaths are matte
but apically cinereous. Vellozia glabra has less im¬
bricate leaves, and its leaf sheaths arc lustrous
brown throughout. Given its large stature (to 6 m),
mature individuals of Vellozia gigantea are unlikely
to be confused with any of its smaller relatives.

Paratypes. BRAZIL. Minas Gerais: Mini. Santana do
Riacho, Serra do Cipo, entroncamento da estrada Coneei-
§ao do Mato Dentro— Morro do Pilar, 27 Apr. 1978 (fl, fr),
G. Martinelli 4370 (RR, SIT); Parque Ksladual da Serra
do Cipo, IB Dez. 1979 (II). N. /,. Menezes. ./. Semir. M. G.
Sajo. M. C. //. Mamede, It. Ml ikes. M. C. hi. 1 moral & \L
T. LI. Rodrigues 870 (SPF, US). 29 Jim. I9BB (fr), I. Cor-
deiro. S. Romaniuc Neto, M. G. L. Wanderley & K. L. M.
Catharine in CFSG 1 1 187 (SP. SPK), 5 Jan. 1993 (fl, ir),
M. Lucca el al. 74 (BMCB); elev. 1400 m, 5 Jan. 1993
(fl), M. Lucca et al. 70 (BIICR). 8 Mar. 1993 ((I). M. Pe¬
reira & M. Lucca 998 (BHCB), I May 1993 (fl, Ir). J. R.
Pirani, A. M. Giulietti & K Barros in Ch'SC 13048 (SPF),
13 Apr. 1994 (fl. fr), A. L. Menezes 1327 (SPF), II Jan.
1999 (fl, fr). It. (i. horzza , S'. Buzato & A. Christianini

1070 (SPF), 25 Jan. 1999 (fr). It. C. horzza. S'. Buzato &
A. Christianini 1082 (SPF).

Acknowledgments. Thanks are due to Antonio
Salatino and Ramon Arthur Clark for the English
version, to Emiko Naruto for the illustration of the
plant habits and for inking in the other drawings of
Figures 1 and 2, and to Victoria C. Hollowell and
an anonymous reviewer for helpful comments.
Fieldwork was partly financed by CAPES, CNPq,
WWF, and Universidade de Sao Paulo. Both au¬
thors are research fellows of CNPq.

I .iterature Cited

Meguro, M., J. R. Pirani, A. M. Giulietti & R. de Mello-
Silva. 1994. Phytophysiognomy and composition of the
vegetation of Serra do Ambrosio, Minas Gerais, Brazil.
Rev. Bras. Bot. 17: 149-166.

Mello-Silva, R. de. 1991. A new species of Vellozia from
the Fspinhayo Range, Brazil, with some consideration
on the section Xerophyloides. Kew Bull. 46: 321—326.

- . 1994. A new species, new synonyms, and a new

combination in Brazilian Velloziaceae. Novon 4: 271-275.

- . 1995. Aspectos taxonbmieos, biogeografieos,

morfologicos e biologicos das Velloziaceae de Grao-Mo-
gol. Minas Gerais, Brasil. Bob Bot. Univ. Sao Paulo 14:
49-79.

- . 1996. Two new species of Vellozia (Velloziaceae)

from Minas Gerais, Brazil. Bot. J. Linn. Soe. 120: 257—
263.

- . 1997. Vellozia sessilis I.. B. Sm. ex Mello-Silva

(Velloziaceae), especie nova de Goias, Brasil. Bob Bot.
Univ. Sao Paulo 16: 65—69.

- . In press. Partial cladistic analysis of Vellozia and

characters for the phylogeny of Velloziaceae. In k. L.
Wilson & I). Morrison (editors), Proceedings of the
Monocots II and Grasses III Conference. Royal Botanic
Gardens, Sydney.

- & N. L. tie Menezes. 1988. Duas especies novas

de Velloziaceae de Minas Gerais. Acta Bot. Bras. 1
(suph): 195—207.

- & J. R. Pirani. 1994. Gualteria rupestris (Annon-

aeeae), a new species from Minas Gerais, Brazil. Novon
4: 146-150.

Menezes, N. L., R. de Mello-Silva & S. J. Mayo. 1994. A
cladistic analysis of Velloziaceae. Kew Bulb 49: 71—92.

- & J. Semir. 1991. Burlemarxia. a new genus of

Velloziaceae. Taxon 40: 413-426.

Morawetz, W. 1983. Dispersal and succession in an ex¬
treme tropical habitat: Coastal sands and xeric wood¬
land in Bahia (Brazil). Sonderbd. Naturwiss. Ver. Ham¬
burg 7: 359-380.

Pirani, J. R., A. M. Giulietti, R. de Mello-Silva & M.
Meguro. 1994. Checklist and patterns of geographic dis¬
tribution of the vegetation of Serra do Ambrosio, Minas
Gerais, Brazil. Rev. Bras. Bot. 17(2): 133-147.

Smith, L. B. 1985. Nolulae brasilianae I. Bradea 4: 1 33 —

1 34. 157-160.

- . 1986. Notulae brasilianae lb Bradea 4: 21 1—214.

- & E. S. Ayensu. 1976. A revision of American Vel¬
loziaceae. Smithsonian Contr. Bot. 30: i— viii + 1 — 172.

- & - . 1979. Velloziaceae Brasiliae I. Bradea

2: 326-328.

— - & - . 1980. Velloziaceae Brasiliae II. Bra¬

dea 3: 105—1 14.

Una Nueva Combinacion en Festuca (Poaceae)

Maria A. Negritto y Ana M. Anton

Institute Multidisciplinario de Biologfa Vegetal (Universidad Nacional de C6rdoba— CONICET),
Casilla de Correo 495, 5000 Cdrdoba, Argentina

Rksl MEN. Se da a conocer una nueva combina-
cidn para una especie sudamericana originalmente
descripta bajo Poa: Festuca nigriflora (Hitc hcock)
Negritto & Anton (Poaceae). El material tfpico es
descripto e ilustrado.

Abstract. A new combination is made for a
South American species originally described under
Poa: Festuca nigriflora (Hitchcock) Negritto & An¬
ton (Poaceae). The typical material is described and
illustrated.

Poa L. y Festuca L. son generos muy afines euya
independencia se sostiene desde antiguo sobre la
base de unos pocos caracteres: lemmas aquilladas
y cariopsis con hilo oval en Poa, lemmas redon-
deadas en el dorso e hilo linear en Festuca. A ellos
se agrega la versatilidad reproductiva que se man-
ifiesta en Poa, ante la casi uniformidad de Festuca
en estos aspectos.

Por tratarse de taxones numerosos, ambos cuen-
tan con unas 450-500 especies, las excepciones
son frecuentes; en tales casos y ante lfmites inter-
gen^ricos imprecisos, las especies han sido asig-
nadas sin un andlisis exhaustivo a uno u otro ge-
nero, lo que ha resultado en transferencias y/o
sinonimias; de hecho, solo en el Index to Grass Spe¬
cies (Chase & Niles, 1962) aparecen 35 taxones
descriptos originalmente en Festuca y que fueron,
en un tiempo u otro, transferidos a Poa; un ejemplo
reciente lo constituye la propuesta de Laegaard
(1998) cuando propone Festuca vaginalis (Ben-
tham) Laegaard, bas&ndose en Poa vaginalis Ben-
tham. Por otra parte, en el Catdlogo de la Familia
Poaceae en la Republica Argentina (Zuloaga et al.,
1994), puede constatarse la existeneia de unas diez
sinonimias y cinco de las transferencias citadas.

El estudio de los ejemplares tipos y del material
adicional que se cita m&s abajo, permitio elaborar
la siguiente description:

Festuca nigriflora (Hitchcock) Negritto & Anton,
comb. nov. Basonimo: Poa nigriflora Hitch¬
cock, Contr. U. S. Natl. Herb. 24: 330. 1927.
TIPO: Peru. On rocky lake shore, forming
tough clumps, at an altitude of 4200 m, at
Huaron, 12 June 1922, Macbride & Feather-
stone 11S5 (holotipo, US no visto; fragmento
del holotipo, BAA; isotipo, F). Figura 1A— I.

Plantas perennes, monoelinas, cespitosas, de
40—50 cm de alto; canas erguidas, 2-nodes; entre-
nudos largos, el superior ca. 15 cm tie long. Lem¬
mas de 10-15 cm de long, por 2.5 mm de lat.,
involutas, ascendentes, coriaceas, algo rfgidas, gla-
bras en su cara abaxial y pilosas en la adaxial,
apice agudo a subagudo. Lfgulas truncadas, de 0.5-
1 mm de long., borde superior ciliado. Vainas de
7-15 cm de long, por 5.5 mm de lat., glabras. Pa-
nojas exertas, angostas, elfpticas, densas, de 9-14
cm de long, por ca. 1 cm de lat., con ramas adpre-
sas; ramas basales 2, de hasta 4.5—5 cm de long.,
generalmente desprovistas de espiguillas en la 1/2
inferior. Eje, ramas y pedicelos de la panoja an-
gulosos y ligeramente escabriusculos. Espiguillas
subcoriaceas, generalmente 2-floras, de 5.5 mm
long, por ca. 2 mm lat., raro 3-floras y alcanzando
hasta 8 mm long. Glumas obtusas, algo aquilladas,
desiguales, coridceas, dpice finamente ciliado, la
inferior de 2.5 mm long, por ca. 0.7 mm de semi-
lat., 1-3-nervia, la superior obtusa, de 3—3.5 mm
long, por ca. 1 mm lat., 3-nervia. Lemmas negruz-
cas, subcoriaceas, algo redondeadas, obtusiusculas,
de 4—5.5 mm long, por 1.2 mm de semilat., nervios
inconspicuos, glabras o algo escabriuseulas en la
mitad superior. Pdleas subcoriaceas, mayores que
las lemmas o de la misma longitud, de (4.5) 5 (5.5)
mm long, por 0.8 mm de lat., escabriuseulas en la
mitad superior sobre las quillas y entre ellas. Flores
perfectas, anteras 3, de 2.5 mm de long. Cariopsis
de 2.3— 2.7 mm de long., comprimidas dorsiven-
tralmente, con surco dorsal e hilo linear tan largo
como el fruto.

Material examinado. PERU. Dpto. Jiiiu'n: Uuar6n,
shady and somewhat moist crevice in rocky lakcshore, As-
plund 1 1800 ( K). Prov. Yauli: Antajascha, entre Casapalca
y Corpacancha, Tovar M84 (USM).

Los materiales adicionales analizados han per-
mitido ampliar al area de distribucion de la espe¬
cie, la que, al tiempo de su descripcion, estaba
restringida a la localidad del tipo. Creee exclusi-
vamente en los Altos Andes del Peru, entre los
4200—4600 m, en la estepa graminosa.

Tovar publico en 1972 una revision de las es¬
pecies peruanas del genero Festuca; all! reconoce

Novon 9: 542-544. 1999.

Volume 9, Number 4
1999

Negritto & Anton

Nueva Combinacion en Festuca

543

Figura 1. Festuca nigriflora (Hitchcock) Negrillo & Anion ( Mucbride & Featherstone 1135, F). — A. Habito. — B, F,
G. Cariopsis, vista hilar, escutelar y transcorte respectivamente. — 1). Palea. — C. Kspiguilla. — E. Antecio basal, nhtese
la lemma mds corta que la palea. — H. Flor. — I. Lfgula. Dibujo Nidia Flury.

544

Novon

37 entidades, en su mayorfa nativas de la Cordille¬
ra de los Andes. Entre ellas, la que resulta mas
alfn a F. nigriflora es F. densiflora Tovar, especie
que solo se conoce de la loealidad tipo (pajonal de
puna de la region del Cuzeo, a 4500 m). La mayor
diferencia entre ellas la constituye la pigmentaeion
de las lemmas y en menor medida de las glumas,
la que les oonfiere a las panojas de F. nigriflora su
earaeterfstica eoloraeion negruzca. Otras diferen-
eias se resuinen en la siguiente clave:

la. Panojas de 9—14 cm long.; ramas basales de 4.5—

5 cm long, desprovistas de espiguillas en su me¬
dio inferior; espiguillas generalmente 2-floras, de
5.5 turn (raro 3-floras y de liasta 8 mm long.);
glumas obtusas; lemmas negruzcas . . F. nigriflora

lb. Panojas de 15—18 cm long.; ramas basales de
hasta 9 cm long., floridas desde la base; espi¬
guillas 3-4-floras, de 7.5 mm long.; glumas lan-
ceoladas, agudas; lemmas pajizas ... F. densiflora

Literature Citada

Chase, A. & C. 1). Niles. 1962. Index to Grass Species,
Vol. 3. G. K. Hall, Boston.

Laegaard, S. 1998. New species and names in Ecuadorian
grasses (Poaceae). Novon 8: 23—30.

Tovar, 0. 1972. Revisibn de las especies peruanas del
genero Festuca (Gramineae). Mem. Mus. Hist. Nat. “Ja¬
vier Prado” 16: 1-93.

Zuloaga, F. 0., E. G. Nicora, Z. E. Riigolo de Agrasar, O.
Morrone, J. Pensiero & A. M. Gialdella. 1994. Catdlogo
de la Familia Poaceae en la Repiiblica Argentina. Mcn-
ogr. Syst. Rot. Missouri Bot. Gard. 47.

A New Combination in Polygala (Polygalaceae) for Southeast Asia

C. A. Pendry

School of Botany, Trinity College, Dublin 2, Ireland. Correspondence address:
Royal Botanic Garden Edinburgh, 20a Inverleith Row, Edinburgh, EH3 5LR,

United Kingdom

Abstract. The new varietal combination Polyg¬
ala glaucoides L. var. laotica (Gagnepain) Pendry
is published herein. The taxon is known from Cam¬
bodia, Laos, Vietnam, and Thailand.

During preparation ol accounts of the Polygala¬
ceae for the Flora of Thailand and La Flore du
Cambodge, du Laos et du Vietnam , it has been
found that a new varietal combination needs to be
published.

Wight (1834), in his Prodromus Florae Penin-
sulae Indiae Orientalis, wrongly used the name Po¬
lygala ciliata L. for the collection Wallich 4177
(collected by Wight in southern India). The use of
the name P. ciliata was mistaken because that
name correctly refers to a quite different taxon
which had already been transferred to Salomonia
by De Candolle (1824) as S. ciliata (L.) DC.

Gagnepain (1909) treated Poly gala ciliata sensu
Wight as consisting of three varieties. Variety typica
has the syntypes Wallich 4177 and Thomson s.n.
from Maisor (= Mysore), Carnatic in southern In¬
dia, and this taxon was later named Polygala bol-
bothrix by Dunn (1915). The other two varieties,
var. laotica Gagnepain and var. brachystachya
(A.W. Bennett) Gagnepain, can be ascribed to P
glaucoides L., but because of the misapplication of
the name P. ciliata , it is necessary to publish a new
combination. Variety brachystachya is referable to
the element of P. glaucoides L. that contains the
type of the name P. glaucoides (van der Meijden,
1988) and must therefore be called variety glau¬
coides. Variety laotica is published here as a new
combination within P glaucoides. It is distin¬
guished from variety glaucoides by its indumentum:
in addition to the usual short, curved hairs seen on
all specimens of P. glaucoides, variety laotica pos¬
sesses distinctive 1-mm-long, spreading, stiff, white
hairs, which are found on the stems, leaves, inflo¬
rescences, and capsules.

There has been considerable confusion about the
correct name for P. glaucoides and several other
related species (notably P. chinensis L.); a full dis¬
cussion and complete synonymy are found in van
der Meijden’s (1988) Flora Malesiana account.

Polygala glaucoides L. var. glaucoides. TYPE:
Herb. Hermann 270, vol. 3: 24 (BM).

Polygala ciliata I.. var. brachystachya (Blume) Gagnepain,
Hull. Bot. Soc. Fr. 56: 20. 1909. TYPE: "In grami-
nosis circa Buitenzorg,” Blume (not seen).

Polygala glaucoides L. var. laotica (Gagnepain)
Pendry, comb. nov. Basionym: Polygala ciliata
L. var. laotica Gagnepain, Bull. Bot. Soc. Fr.
56: 19. 1909. TYPE: Cambodia. Oudon (=
Oudong, Kandal), Thorel 2066 (lectotype, des¬
ignated here, P).

Additional material seen. CAMBODIA. Kampot:
kamchay, Thorel s.n. (P). Kompong Speu: Poilane
I73B4 (P). Siem Reap(?): Compong-luong (= kampong
Hluong), Thorel s.n. (P). THAILAND. Katcliahuri:
Nawng ke, 5 rn. Collins 1629 (ABD); Praehaup khiri
khan: llua Hin, Kerr 13454 (UK, BM, L). VIETNAM.
Dong Nai: Pierre s.n. (P).
kr\ to Tin: Variktiks ok Polycala glaucoidks

la. Stems, leaves, petioles, and inflorescence with
indumentum of short curved hairs only; capsule

with only short, curled hairs on margin .

. P. glaucoides var. glaucoides

II). Stems, leaves, petioles, and inflorescence with
indumentum of short curved hairs and 1-mm-
long, stiff, straight hairs; capsule with short,
curled and long, straight hairs on margin ....
. P. glaucoides var. laotica

Acknowledgments. I thank the Royal Society for
financial support for my work, and the School of
Botany, Trinity College, for hosting my project. I
thank curators of ABD, L, and P for loaning their
specimens, and BK, BM, and P for their assistance
during my visits. I also thank David Middleton,
Matthew Jebb, Mark Newman, and Robert Mill lor
their advice in the preparation of this manuscript.

Literature Cited

Candolle, A. De. 1824. Prodromus Systematis Naturalis
Regni Vegetabilis, Vol. I. Treuttel & Wurtz, Paris.
Dunn, S. T. 1915. Polygalaceae. In: J. S. Gamble, FI. Ma¬
dras I: 58.

Gagnepain, F. 1909. Nouveautes Asiatiques de l’herbier
du Museum. Bull. Bot. Soc. Fr. 56: 15—22.

Meijden, R. van der. 1988. Polygalaceae. FI. Mai. Ser. 1,
10(3): 455-539.

Wight, R. 1834. Prodromus Florae Peninsulae Indiae Or¬
ientalis. Parbury, Allen, London.

Novon 9: 545. 1999.

A New Species of Poupartia (Anacardiaceae) from Madagascar

Armand Randrianasolo and James S. Miller
Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. Poupartia orientalis Capuron ex A.
Randrianasolo & J. S. Miller is described from
Madagascar, and its distribution, habitat, and phe¬
nology are discussed. Poupartia orientalis is similar
to P. silvatica H. Perrier hut is distinct in its ev¬
ergreen habit and subcoriaceous leaflets with fewer
and less prominent tertiary veins.

Poupartia Cominerson is a genus presently con¬
sidered to comprise seven species restricted to
Madagascar, the Mascarenes, and the Seychelles.
Perrier de la Bathie (1946) recognized five species
in Madagascar, but Poupartia cajfra (Sonder) H.
Perrier was transferred to Sclerocarya Hochstetter
by Kokwaro (1986), and P. gurnmifera Sprague, a
species that occurs in Madagascar and Seychelle
Islands, was transferred to Operculicarya H. Perrier
by Capuron (1975). More recently, Friedmann
(1994) accepted P. gurnmifera in his treatment of
Anacardiaceae of the Seychelles over Operculicarya
gurnmifera (H. Perrier) Capuron, a decision agree¬
ing with Eggli’s (1995) revision that excluded O.
gurnmifera from his concept of Operculicarya.
Three other species of Poupartia are endemic to
the Mascarene Islands (Friedmann, 1997), bringing
to seven the total number of species in this genus
prior to this publication. Latin American and Asian
taxa previously considered to belong to Poupartia
have all been transferred to other genera (Hill,
1937; Metcalf, 1931; Cheng & Ming, 1980). A
forthcoming revision of Poupartia will provide ad¬
ditional information about P. gurnmifera as well as
a key for identifying all known species (Randriana¬
solo & J. S. Miller, in prep.).

During preparation of a revision of Anacardi¬
aceae for Madagascar, it became apparent that a
group of specimens collected in the northeastern
and mid-eastern regions of the island represents an
undescribed species. Although no specimens of
male individuals were observed, study of the female
flowers and fruits indicates that this undescribed
species belongs to Poupartia.

Poupartia orientalis Capuron ex A. Randriana¬
solo & J. S. Miller, sp. nov. TYPE: Madagas¬
car. Toamasina (Est): foret sublittorale, sur sa¬
bles, au Sud de SoanieranTvongo (P.K. 154),
27 Nov. 1962 (H and yf r), SF 22124 Capuron
(holotype, P; isotype, MO). Figure 1.

Arbor usque ad 15—20 m alta. Folia persistent ia, alter-
na; foliolis 7 ad 13, oppositis vel subopposilis, terminali
obovato, lateralibus inaequaliter ovatis, 3.5— 9.0 cm longis,
apice aeuminatis, base inaequaliter obtusis, supra et infra
glabratis; petiolulis 3—4 mm longis. Inflorescentia axillar¬
is, racemosa, 5—7 cm longa; pedicel lis '1—5 mm longis.
Flores leminei 5-rneri; lobulis corollinis reflexis, ovatis,
2. 5-3. 5 mm longis; stylis 5. Fructus drupaceus, late ovo-
ideus ad subglobosum, 1.5—2 cm longus, 5-loeularis, 5
operculatus.

Evergreen trees, 15-20 in tall; young twigs gen¬
erally glabrous and with lenticels, rarely with
sparse hairs at the tips of branches or branchlets.
Leaves alternate, often clustered at the end of
branches or branchlets, imparipinnately compound,
15—30 cm long; leaflets 7 to 13, opposite or subop¬
posite, subcoriaceous, lateral leaflets asymmetri¬
cally ovate and the terminal ones obovate, 3.5—9
cm long, 1.5—4 cm wide, the apex long acuminate,
the base unequally obtuse, sometimes very shortly
decurrent, the margin entire and slightly revolute,
the adaxial surface pubescent when young, almost
glabrous when older except for a few hairs left on
the midvein and on the margin at the leaflet base,
abaxial surface pubescent when young, glabrous
when mature, venation pinnate, generally brochi-
dodromous but sometimes basal lateral veins reti-
culodromous, the midvein prominent on the lower
surface, the lateral veins arcuate, very slightly
prominent on the lower surface, the tertiary veins
very few. not prominent, ramified and admedial,
sometimes connecting two adjacent lateral veins;
petiolule of terminal leaflet 5—30 mm long; petio-
lules of lateral leaflets 3—1 mm long, pubescent or
glabrescent. Male inflorescence and flowers un¬
known. Female inflorescence axillary, racemose, 5—
7 cm long, pubescent, sometimes glabrous; bracts
very caducous, not observed. Flowers generally 5-
merous, unisexual, small, ca. 3 mm long; pedicels
4—5 mm long, scabrous or pubescent; calyx lobes
debate, ca. 1 mm long and ca. 1 mm wide, pubes¬
cent, imbricate; corolla lobes reflexed at anthesis,
ovate, 2. 5-3. 5 mm long, ca. 2 mm wide, glabrous,
yellow or pale yellow, imbricate. Staminodes in 2
whorls of 5; filaments 1—1.5 mm long (episepalous
ones longer than epipetalous), straight and broad¬
ened at the base, inserted basally on the outer sur¬
face of the disc, glabrous, white; anthers ovate, ca.

Novon 9: 546-548. 1999.

Volume 9, Number 4
1999

Randrianasolo & Miller

Poupartia orientalis from Madagascar

547

Figure 1. Poupartia orientalis Capuron ex A. Randrianasolo & J. S. Miller. — a. Branch with inflorescences. — b.
Female flower with one petal removed. — c. Leaflet tertiary veins. — d. Infructescence. — e. Fruit. — f. Cross section
of fruit. — g. Polar view of endocarp opercula. — h. Inflorescence branches. Drawn from the type.

548

Novon

0.3 mm long, yellow, dorsifixed, introrse, dehiscent
by longitudinal slits, sterile and glabrous; disc
fleshy and crater-shaped, ea. 1.5 mm diam. and ea.
0.5 mm thick; ovary very widely ovate to subglo-
bose, 1—1.2 mm long, ea. 1 mm broad, glabrous, 5-
loeular of which 4 are fertile (only two ovules de¬
velop. the rest abort), ovules anatropous, with
apical plaeentation; styles 5, distinct and short,

O. 2-0. 5 mm long, latero-apical, around the top,
with five capitate stigmas. Fruits drupaceous, very
widtdy ovoid to subglobose, 1.5—2 cm long, 1.3—2
cm broad, glabrous; exocarp thin, mesocarp thin
and resinous, endoearp bony and thick, with 5 op-
ercules.

Habitat and distribution. Poupartia orientalis is
known from northeastern and mid-eastern littoral
and sublittoral wet forests of Madagascar, where it
grows on sandy or lateritic soils.

Phenology. This species flowers in October—
November and fruits from December through June.

Vernacular name. Sakoanala, Sakoala.

Discussion. In previous publications, species of
Poupartia have been described either as polyga-
modioecious (Perrier de la B&thie, 1946) or dioe¬
cious (Friedmann, 1994, 1997). Our observations
of specimens of the Malagasy species included in
this genus confirm that male and female flowers are
consistently found on separate individuals. In ad¬
dition. the male flowers always display a reduced
gynoecium with the fertile stamens, and female
flowers have reduced and sterile stamens or stam-
inodes with the functional gynoecium. Sometimes
female flowers seem to have regular-sized stamens,
but the anthers of these stamens lack pollen grains
and are sterile and non-functional. These observa¬
tions suggest that Poupartia species are dioecious,
and although Poupartia orientalis is known only
from specimens with female flowers, it appears not
to be an exception.

Poupartia orientalis and P. chapelieri (Guillau-
min) H. Perrier, both from eastern Madagascar, are
the only two evergreen species of Poupartia in
Madagascar. However, despite their overlapping
distribution and ecological similarity, numerous
characters distinguish these two species from one
another. Poupartia orientalis has 5-locular fruits
with five opercules (Fig. 1) compared with the uni¬
locular, but bi-operculate fruits of P. chapelieri. In
addition, P. orientalis has pedicels that are 4—5 mm
long and petals that are 2.5— 3.5 mm long, as com¬
pared to the nearly sessile and smaller flowers of

P. chapelieri , which have pedicels 0.3— 0.5 mm long
and petals ca. 2 mm long. Poupartia silvatica H.

Perrier is perhaps the species most easily confused
with P. orientalis , as it has a similar inflorescence
and flower size. However, P. silvatica differs in be¬
ing deciduous, in having more membranaceous leaf¬
lets with more ramified, closely spaced tertiary
veins, and in occurring only in the northwest and
west of Madagascar.

Rene Capuron, a French botanist working in
Madagascar, was the first person to recognize that
specimens collected from northeast Madagascar
probably represented a new species, and he anno¬
tated the specimens as such. However, he did not
publish the name Poupartia orientalis; we honor
him and his contributions to the Malagasy flora by
accepting the name that he proposed.

Paratypes. M ADAGASCAR. Antsiranana: Antalaha
Kst (Nord), table basaltique d'Ambanitazana, pres
d'Andrapengy (aw N. d" Antalaha). I4°40'S. 5()°12'K, II
June 1967, SF 27731 Capuron (P). Toainasina: Bricka-
ville. Ambodikijy Coharindava, 18°47'30"S, 48°40'K, 15
June 1954, SF 100.19 (P); Maroanlsetra, Farankaraina,
1 5°25'S. 49°52'E, 12 Feb. 1955, SF 12915 (MO. P. 2
sheets); Maroantsetra, Jardin Botanique n° 17. 15°26'S,
49°45'F. 12 Dec. 1952, SF 7238 (P).

Acknowledgments. Funds to support the study
ol Anacardiaceae of Madagascar were provided by
the Beneficia Foundation and National Geographic
Society. We are grateful to Antony Jardim for review
comments and Barbara Alongi for the marvelous
illustration. We also express our gratitude to the
staff of the Paris Herbarium for making their spec¬
imens available on loan.

I .iterature Cited

Capuron. B. 1975. Contribution A 1‘etude de la llore fo¬
rest ifere de Madagascar. Adansonia, ser. 2. 14(4): 57 1 —
572.

Cheng, M. A T. Ming. 1980. Anacardiaceae. In Flora Rei-
publicae Popularis Sinicae 45(1): 66-135.

Fggli. U. 1995. A synoptical revision of Operculicarya
(Anacardiaceae). Bull. Mus. Nall. Hist. Nat.. 4eme ser.,

1 7(3-4): 149-158.

Friedmann, F. 1994. Anacardiaceae. Pp. 417-422 in:
Flore des Seychelles. Dieotyledones. Editions de
rOrstom, Paris.

- . 1997. Anacardiacees. 77 erne famille. Flore des

Mascareignes. Published by The Sugar Industry Re¬
search Institute, Mauritius; Orstom, Paris; and Boyal
Botanic Gardens, Kew.

Hill, A. W. 1937. The method of germination of seeds
enclosed in a stony endoearp. Ann. Bot. I: 239—256.
Kokwaro, J. (). 1986. Anacardiaceae. Flora of Tropical
East Africa. A. A. Balkema, Rotterdam/Boston.

Metcalf. 1. P. 1931. Notes on Chinese plants. J. Arnold
Arbor. 12: 270-275.

Perrier de la Bathie, II. 1946. Anacardiacees. 114 feme
famille. Flore de Madagascar el des Comores. Museum
National d’Histoire Naturclle, Paris.

Validation of Some Suprageneric Names in Magnoliophyta

James L Reveal

Norton-Brown Herbarium, University of Maryland, College Park, Maryland 20742-5815, U.S.A.

Alexander R. Doweld

National Institute of Carpology (Gaertnerian Institution), P.O. Box 72, BUS-11915,

Moscow, Russian Federation

Abstrac t. Eight superordinal and twelve ordinal
names not validly published by A. L. Takhtajan in
1997 are formally validated. One name proposed
by Cronquist in 1980 is validated. In addition,
three names not properly validated by J. L. Reveal
are included. A new family name, Exbueklandi-
aceae, is proposed for the illegitimate family name
Bucklandiaceae. Engelhardtiaceae is validated to
accommodate Engelhardtia, Alfaroa, Alfaropsis,
and Oreomunnea, all distinct from the Juglanda-
ceae s. str.

In 1997, A. E. Takhtajan proposed numerous
new suprageneric names in his review of tin' clas¬
sification of the flowering plants (Magnoliophyta).
Unfortunately, a few proved not to be validly pub¬
lished as required by the International Code oj Bo¬
tanical Nomenclature (Greuter et al., 1994). These
names are now validated with the blessing of Pro¬
fessor Takhtajan. Hydatellales, proposed by Cron¬
quist in a Takhtajan (1980) paper was not validly
published (Art. 32.4, 33.2); it is here validated. Two
names proposed previously by Reveal (1992,
1993), Rhizophorales and Torricelliales, are also
validated. Engelhardtiaceae is established as a dis¬
tinct family for four genera: Engelhardtia Lesehen-
ault ex Blume (5 spp., northern India to China and
the Philippines), Alfaroa Standley (7 spp., Mexico
to Colombia), Alfaropsis Iljinskaja (1 sp.. Southeast
Asia), and Oreomunnea Oersted (2 spp., Mexico to
Panama), which are distinct in pollen morphology,
seed and fruit anatomy, and moq>hology from Jug-
landaceae (s. str.; Iljinskaja, 1990). Exbucklandi-
aceae, recently introduced by Doweld (1998), is
proposed as a valid, substituted family name for the
illegitimate Bucklandiaceae J. Agardh. The family
is restricted to a single genus, Exbuchlandia R. W.
Brown (2 spp., eastern Asia and Malesia).

Barbeyanae Takhtajan ex Reveal & Doweld, su-

perord. nov. TYPE: Barbeya Schweinfurth;

Barbeyaceae Rendle, in Dyer, FI. Trop. Afr.

6(2): 14. 1916, nom. cons.

Arbusculae; folia opposita, exstipulata; flores dioici, re-
gulares, ebracteati v. ebraeteolati: petala nulla; sepala 3—
4. basi eonnata; stamina 6— 9(— 12); antherae oblongae, lon-
gitudinaliter dehiscentes; ovarium I -3-partitum, carpella
basi eonnata, uniovulata; ovulum anatropum, pachychal-
azale; fructus (capsetum) indehiseens, ealyeis limbo per-
sistente; semina exalbuminosa, embryo rectus, cotyledon-
ibus duobus planis instruetus.

Trees; leaves opposite, simple, entire, pinnately
veined, estipulate; inflorescences reduced cymes of
short, 3-flowered dichasia, ebracteate; flowers
small, wind-pollinated, dioecious, the calyx of 3—4
valvate sepals, accrescent basally, the petals none;
stamens 6-9(-12), the anthers basifixed, tetraspor-
angiate, dehiscing longitudinally; ovary superior,
1— 3-merous, the carpels typically connate basally,
each with a linear style terminated by a decurrent,
plumose stigma; ovules solitary, anatropous, pachy-
chalazal; fruits an indehiscent capsetum; seeds ex-
albuminous; embryo straight with two flattened cot¬
yledons.

Buxanae (Muller Argoviensis) Takhtajan ex Re¬
veal & Doweld, superord. et stat. nov. Basio-
nym: subtrib. Buxinae Muller Argoviensis, in
DC., Prodr. 16(1): 9, 11. 1869 [Latin diagnosis
in key and description in Latin]. TYPE: Buxus
L.; Buxaceae Dumortier, Anal. Fam. PL: 45.
1829, nom. cons.

Casuarinanae (Lindley) Takhtajan ex Reveal &
Doweld, superord. et stat. nov. Basionym: ord.
Casuarinales Lindley, Nix. 14.: 167. 1833 [Lat¬
in diagnosis in key]. TYPE: Casuarina Adan-
son; Casuarinaceae R. Brown, in Flinders, Voy.
Terra Austr. 2: 571. 1814, nom. cons.

Ceratophyllanae Takhtajan ex Reveal & Doweld,
superord. nov. TYPE: Ceratophyilum L.; Cer-
atophyllaeeae Cray, Nat. Arr. Brit. PI. 2: 395,
554. 1821, nom. cons.

Novon 9: 549-553. 1999.

550

Novon

Herbae submersae, caule ramisque teretibus, nodoso-
articulatis; folia verticillata, sessilia, exstipulata, dicho- v.
trichotome dissecta; flores monoici, apetali; in floribus
masculis sepala libera, 9-l()(-12); stamina (5— )8— 18(-27),
anlherae bilooulares, longitudinaliter dehiscentes; in flo¬
ribus femineis calyx ut in masculis, ovarium unicum (fol-
liculare), sessile; stylus terminalis, apice hinc stigmatosus;
ovulum unicum, pendulum, orthotropum, unitegminatum;
nuculae monospermae, coriaceae, involucro persistente
stipatac, stylo persistente apiculatae; semen exalbuminos-
um, cotyledonibus duobus, plumula polyphylla.

Submerged aquatic herbs, rootless, monoecious;
leaves simple, verticillate, essentially sessile, di-
chotomously divided into linear serrulate segments,
estipulate, the stomata lacking; flowers minute, sol¬
itary, axillary, apetalous, actinomorphic, the calyx
of 9-10(— 12) bractlike sepals, hypogynous; stamens
(5— )8— 18(-27), spirally arranged, developing cen-
tripetally, the anthers oblong, tetrasporangiate, de¬
hiscing longitudinally; pollen inaperturate; ovary
superior, solitary (follicular), the style fairly long,
the stigma decurrent ventrally; ovule solitary, pen¬
dulous, orthotropous, unitegmic; fruits a nutlet with
a persistent stylodium and basal to lateral horns;
seeds minute, essentially exalbuminous; embryo
large, with two oblong, fleshy cotyledons, the radi¬
cle shorter than the cotyledons, the plumule well
developed.

Dioscoreanae (R. Brown) Takhtajan ex Reveal &
Doweld, superord. et stat. nov. Basionym: fam.
Dioscoreaceae R. Brown, Prodr.: 294. 1810
[Latin description], TYPE: Dioscorea L.; Dios¬
coreaceae R. Brown, Prodr.: 294. 1810, nom.
cons.

Lactoridanac Takhtajan ex Reveal & Doweld, su¬
perord. nov. TYPE: Lactoris R. Philippi; Lac-
toridaceae Engler, in Engler & Prantl, Nat.
Pflanzenfam. 3(2): 19. 1888, nom. cons.

Frutices; folia alterna. petiolata, simplicia, stipulae in¬
terpet iolares; flores in monochasia axillaria conferti, po-
lygami-monoici, cyclici, apetali, 3-part it i; sepala 3, libera;
stamina 6, dicyclica; gynoeceum apocarpum, 3-carpellare;
ovula 4—8, anatropa, bitegminata; fmctus (capsetum)
ventrihiscens; semina albuminosa; embryo minutissimus.

Small shrubs with swollen nodes; leaves alter¬
nate, simple, distichous, with large ochrelike inter-
petiolar stipules united to the petioles; inflorescences
an axillary monochasia; flowers polygamo-monoe-
eious, cyclic, apetalous, trimerous, the sepals free;
stamens 6 in two whorls, the anthers extrorse, de¬
hiscing longitudinally; ovary of 3 free carpels nar¬

rowing to a stylodium bearing a short, decurrent
stigma; fruits a trimerous ventrihiscent capsetum;
ovules 4—8 per carpel, anatropous, bitegmic; seeds
small, albuminous; embryo minute and essentially
undifferentiated.

Poanae Takhtajan ex Reveal & Doweld, superord.
et nom. nov. Basionym: fam. Gramineae Jus¬
sieu, Gen. PL: 28. 1789, nom. cons, et nom.
alt. [Latin description]. TYPE: Poa L.; Po-
aceae (R. Brown) Barnhart, Bull. Torrey Bot.
Club 22: 7. 1895, nom. cons.

Rhizophoranae (Persoon) Takhtajan ex Reveal &
Doweld, superord. et stat. nov. Basionym: fam.
Rhizophoraceae Persoon, Syn. PI. 2: 3. 1807
[Latin description]. TYPE: Rhizophora L.; Rhi¬
zophoraceae R. Brown, in Flinders, Voy. Terra
Austr. 2: 549. 1814, nom. cons.

Anisopliylleales (Bentham & Hooker f.) Takhtajan
ex Reveal & Doweld, ord. et stat. nov. Basio¬
nym: trib. Anisophylleeae Bentham & Hooker
f., Gen. PI. 1: 678, 683. 1865 [Latin diagno¬
sis], TYPE: Anisophyllea R. Brown ex Sabine;
Anisophyllaeeae Ridley, FI. Malay. Penins. 1:
700. 1922.

Chrysobalanales (DC.) Takhtajan ex Reveal &
Doweld, ord. et stat. nov., Basionym: trib.
Chrysobalaneae DC., Prodr. 2: 525. 1825 [Lat¬
in description]. TYPE: Chrysobalanus L.;
Chrysobalanaceae R. Brown, in Tuckey, Narr.
Exped. Congo: 433. 1818, nom. cons.

Flagellariales (Meisner) Takhtajan ex Reveal &
Doweld, ord. et stat. nov. Basionym: subfam.
Flagellarioideae Meisner, PL Vase. Gen.: Tab.
Diagn. 406. 1842 [Latin diagnosis in key].
TYPE: Flagellaria L.; Flagellariaceae Dumor-
tier. Anal. Fam. PL: 59—60. 1829, nom. cons.

Griseliniales (J. R. Forster & G. Forster ex Cun¬
ningham) Takhtajan ex Reveal & Doweld, ord.
et stat. nov. Basionym: fam. Griseliniaceae J.
R. Forster & G. Forster ex Cunningham, Ann.
Nat. Hist. 3: 261. 1839 [Latin diagnosis],
TYPE: Griselinia Forst.; Griseliniaceae J. R.
Forster & G. Forster ex Cunningham, Ann.
Nat. Hist. 3: 261. 1839.

Volume 9, Number 4
1999

Reveal & Doweld

Suprageneric Names in Magnoliophyta

551

Hydatellales Cronquist ex Reveal & Doweld, ord.
et stat. nov., based on fam. Hydatellaceae U.
Hamann, New Zealand J. Bot. 14: 195. 1976
[Latin description]. TYPE: Hydatella Diels;
Hydatellaceae U. Hamann, New Zealand J.
Bot. 14: 195. 1976.

Hypoxidales (Bemhardi) Takhtajan ex Reveal &
Doweld, ord. et stat. nov. Basionym: trib. Hy-
poxideae Bemhardi, Flora 23: 426. 1840 [Lat¬
in diagnosis], TYPE: Hypoxis L.; Hypoxida-
ceae R. Brown, in Flinders, Voy. Terra Austr.
2: 576. 1814, nom. cons.

Lowiales Takhtajan ex Reveal & Doweld, ord. nov.
TYPE: Lowia Scortechini; Lowiaceae Ridley,
FI. Malay. Penins. 4: 291. 1924, nom. cons.

Herbae perennes, acaules; folia basalia. disticha, pe-
tiolata; infloreseentiae e rhizomate erumpentes, cymosae,
bracteosae; flos zygomorphicus; sepala 3, in tubum lon-
gum coalita; petaia 3, inaequalia, petalum merlianum, aliis
dissimile, ex floribus exterius; stamina 5, epigyna, unila-
teralia; antherae basifixae, biloculares; ovarium inferum,
3-loeulare, in quoque loculo ovula numerosa, 2-seriata;
stylus apice trifurcatus; ovula anatropa, bitegminata; fruc-
tus capsinaris, dorsihiseens; semina albuminosa, arillo tri¬
lobate cincta.

Small to mid-sized perennial, rhizomatous herbs;
leaves basal, distichous, sheathing at the base with
a long well-developed petiole and lanceolate to el¬
liptic blade, the lateral veins arching and converg¬
ing apically; inflorescences a short, terminal few-
flowered monochasial cyme, originated directly
from the rhizome, bracteose; flowers zygomorphic
and somewhat orchid-like, bisexual, the sepals 3,
connate basally into a slender tube, the petals 3
with the middle one large and the others small;
stamens 5, epigynous, the anthers bilocular, basi-
fixed; pollen inaperturate; ovary inferior, 3-loeular,
the stigmas 3, laciniate; ovules numerous, biseriate,
anatropous, bitegmic; fruits a dorsihiscent eapsine
(inferior capsule); seeds globose, aril late, albumin¬
ous.

Medusagvnales Takhtajan ex Reveal & Doweld,
ord. nov. TYPE: Medusagyne Baker; Medusa-
gynaceae Engler & Gilg, Syllabus, ed. 9—10:
280. 1924, nom. cons.

Frutex ramosus; folia opposita, exstipulata, brevipetiol-
ata; paniculae terminales, trichotomae; pedicelli graciles;
bracteae bracteolaeque nullae; sepala 5; petaia 5; stamina
numerosa, hypogyna, antherae basifixae, biloculares, lon-

gitudinaliter dehiscentes; ovarium multiloeulare (20—25),
loculis angustis, sty 1 i validi, tarde decidui, stigmata cap-
itala; ovula in quoque loculo duo, medio affixa, superpos-
ita, pendula vel adscendentia; fructus sciridialis; carpella
a basi lateri-lobihiscentia, sursum divergentia, apice per-
sistentia; semina oblonga, alata; endospermium minimum;
embryo rectus.

Shrubs or small trees; leaves opposite, simple,
leathery, estipulate; inflorescences a terminal pan¬
icle; flowers andromonoecious (bisexual and male
only in the same inflorescence), actinomorphic, hy-
pogynous, foetid, the sepals 5, imbricate, reflexed,
basally connate, persistent; petals 5, free, imbricate
and contorted in bud; stamens numerous, free, spi¬
rally arranged in 4-5 levels, the anthers basifixed,
dehiscing longitudinally; ovary multilocular (20—
25) with narrow locules, the robust styles free, the
stigmas capitate; ovules 2 per loeule, superposed,
one pendulous and one erect; fruits a specialized
lateri-lobihiscent (columellate) sciridium with car¬
pels remaining united at the apex; seeds small,
winged, exotegmic, the endosperm limited to a nar¬
row band of thin-walled cells; embryo straight.

Rapateales (Meisner) Colella ex Reveal & Doweld,
ord. et stat. nov. Basionym: subfam. Rapateo-
ideae Meisner, PL Vase. Gen.: Tab. Diagn.
405. 1842 [Latin diagnosis in key]. TYPE: Ra-
patea Aublet; Rapateaceae Duinortier, Anal.
Fam. PL: 60, 62. 1829, nom. cons.

Rhizophorales (Persoon) Tieghem ex Reveal &
Doweld, ord. et stat. nov. Basionym: fam. Rhi-
zophoraceae Persoon, Syn. PI. 2: 3. 1807 [Lat¬
in diagnosis]. TYPE: Rhizophora L.; Rhizo-
phoraceae R. Brown, in Flinders, Voy. Terra
Austr. 2: 549. 1814, nom. cons.

Torricelliales Takhtajan ex Reveal & Doweld, ord.
et stat. nov. TYPE: Torricellia DC.; Torricelli-
aceae (Wang) H. H. Hu, Bull. Fan Mem. Inst.
Biol., Bot., ser. 5, 5: 311. 1934.

Suffrutices, ramis teretibus; folia exstipulata, simplicia,
petiolo basi subdilatato; panicula terminalia multiflora;
flores dioici, actinomorphici; calycis tubus cum ovario
connatus, limbus superus; petaia 5, marginem disci epi-
gyni inserta; stamina 5, cum petalis inserta, iisdem alter-
na; ovarium inferum, 3-4-carpel lare, syncarpum; ovulum
anatropum, unitegminatum; bacca infera (uvina) 3— 4-lo-
cularis, loculis uniovulatis; semen lineare; embryo in ap¬
ice endospermii copiosi brevis.

Shrubs to small trees; leaves alternate, simple,
palmately veined, estipulate; inflorescences a ter-

552

Novon

initial panicle, dioecious; flowers small, actinomor-
phic, bracteolate, the male flowers with a 5-lobed
calyx forming a tube connate with the ovary and 5
valvate petals, the female flowers with a minute 3—
5-lobed calyx and no corolla; stamens 5, the an¬
thers basifixed, tetrasporangiate, dehiscing longi¬
tudinally; ovary inferior, the carpels 3—4, syncar-
pous; ovules 1 per locule, anatropous, unitegmic;
fruits an inferior berry (uvina), 3— 4-locular; seeds
linear; embryo small, endosperm copious.

Xanthorrhoeales Takhtajan ex Reveal & Doweld,
ord. nov. TYPE: Xanthorrhoea Smith; Xan-
thorroeaceae Dumortier, Anal. Fam. PI.: 60,
62. 1829, nom. cons.

Mediae perennes vel suffrutices; folia graminea, angus-
ta, linearia, alternata, basi dilatata, semivaginantia; spica
terminalis, typhoides; perianthium 6-partitum, dicyclioum;
stamina (>, dicyclica. imo perigonio inserta; antherae bas-
ifixae v. dorsifixae, biloculares; ovarium triloculare; ovula
plurima; stylus trisulcus, stigma simplex; capsula trilo-
cularis, dorsihiscente-trivalvis; semina in loculis 1-2.
compressa; embryo linearis.

Xeromorphic, rhizomatous shrubs or perennial
herbs; leaves usually linear, liguliform, usually tuft¬
ed, alternate, not sheathing or with a short open
sheath; inflorescences a panicle, dense Typha-hke
spike, globular head, or the flowers solitary and ax¬
illary or terminal; flowers small, mostly dry and glu-
maceous, the perianth of 6 segments in two whorls,
free or shortly connate; stamens 6, bicyclic, the in¬
ner whorl attached to the base of the inner perianth
segments, the outer free or hypogynous, the anthers
basifixed or dorsifixed, bilocular, dehiscing longi¬
tudinally; ovary superior, trilocular; ovules 1—2, an-
atropus or rarely campylotropous, bitegmic; fruits a
dorsi-laterihiscent capsule; seeds globose to flat¬
tened; embryo straight, sometimes slightly curved,
the endosperm copious.

Zygophyflales (Bartling) Takhtajan ex Reveal &
Doweld, ord. et stat. nov. Basionym: Irib. Zyg-
ophylleae Bartling, Ord. Nat. PL: 391. 1830
[Latin diagnosis]. TYPE: Zygophyllum L.; Zyg-
ophyllaceae R. Brown, in Flinders, Voy. Terra
Austr. 2: 545. 1814, nom. cons.

Engelhardtiaceae Reveal & Doweld, fam. nov.
TYPE: Engelhardtia Leschenault ex Bluine.
The family also includes Alfaroa, Alfaropsis,
and Oreomunnea.

Arbores; folia alterna, abrupte pinnata, foliolis inaequi-
lateralibus; spicae monoicae, paniculato-ramosae; in flo-
ribus tnasculis calycis bracteae trifidae, bracteolae hi- vel
tetrafidae, stamina 5—13, antherae biloculares, loculis
connectivo supra eosdem producto adnatis, longitudinali-
ter dehiscentibus, ovarii mdimenlum nullum; in florihus
femineis calyx ut florihus masculis, starninurn rudimenta
nulla, ovarium inferum, basi hi- vel incomplete 4-loculare,
dissepimentis duplicatis; ovulum unicum, sessile, erec¬
tion, orthotropum, pachychalazale; stylus terminalis; stig¬
mata 2— A, rarius plura, commissuralia; fructus ascadinaris
cum involucre cupuliformi hasi connatus, apice trialatus,
indehiscens, monospermus; semen erectum, exalbuminos-
um; embryo hi- vel quadrilobatus.

Trees, monoecious; leaves alternate, pinnately
compound, estipitate; inflorescence a spike; flowers
small, the bracts 3-lobed, the bracteolae (2 — )4-
lobed; stamens 5—13, the anthers bilocular, dehisc¬
ing longitudinally; ovary inferior, 2- (at the base) or
incompletely 4-locular; ovule solitary, sessile, erect,
orthotropous, pachychalazal, the style terminal with
2-4, rarely more, stigmas, commissural; fruits an
ascadina with an adherent cupuliform involucrum,
trialate, indehiscent; seed solitary, exalbuminous;
embryo massive, (2— )4-lobed.

Exbueklandiaeeae Reveal & Doweld, nom. et
fam. nov. TYPE: Exbucklandia R. W. Brown
(= Bucklandia Griffith, nom. illeg., Syming-
tonia van Steenis, nom. illeg.); digeneric fam¬
ily (inch Chunia H. T. Chang, sine Mytilaria
Lecomte). A replacement name for Bueklan-
diaceae J. Agardh, Theor. Syst. PL: 155. 1858,
nom. illeg. (Art. 18.3) [Latin description].

Arbores, ramulis abrupte terminales; folia alterna, exst-
ipulata, petiolata; flores capitati, calyx inferne cum ovarii
hasi connatus; petala perigyna. numero varia; stamina 10—
14, perigyna; antherae hasi affixae, biloculares, longitu-
dinaliter dehiscentes; ovarium semiinferum, biloeulare;
ovula in quoque loeulo numerosa (5—9); styli duo; fructus
eapsularis, dorsihiscens, incomplete bivalvis; semina nu¬
merosa, alata, albuminosa; embryo linearis.

Trees; leaves alternate, ovate, often tricuspidate,
palmately veined, estipulate, petiolate, the stipules
generally large; inflorescences an aggregate of flow¬
ers in dense spikes; flowers bisexual, the calyx ad¬
herent at the base of the ovary, the petals perigy-
nous; stamens 10—14, perigynous, the anthers
basifixed, bilocular, dehiscing by valves; ovary
semi-inferior, bilocular; ovules 5—9 per locule,
anatropous, the styles 2; fruits a dorsihiscent cap¬
sule, incomplete bivalvate; seeds numerous, alate;
embryo straight in copious endosperm.

Volume 9, Number 4
1999

Reveal & Doweld

Suprageneric Names in Magnoliophyta

553

Acknowledgments. We thank Margo Semenova
lor assistance with the Latin descriptions and a
helpful reviewer lor useful comments. We also
thank Armen L. Takhtajan for his comments. This
study is part of the Indices Nominum Supragener-
icorum Plantarum Project from the Norton-Brown
Herbarium (MARY) and presented with the coop¬
eration of the International Association lor Plant
Taxonomy and the National Agricultural Library.
The research is a contribution to the Systema Sper-
matophytorum Project of the National Institute ol
Carpology (Gaertnerian Institution), Moscow.

I .iterature Cited

Doweld, A. It. 1998. Carpology, seed anatomy, and taxo¬
nomic relationships of Tetraeentron (Tetracentraceae)

and Trochodendron (Trochodendraceae). Ann. Hot. (Con¬
don) 82: 413-443.

Greater, W., F. It. Barrie, H. M. Burdet, W. G. Chaloner,
V. Demoulin, D. L. Hawksworth, P. M. Jprgensen. I). H.
Nicolson, P. C. Silva. P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Iljinskaja, I. A. 1990. On the taxonomy and phylogeny of
the family Jnglandaceae. But. Zhum. (Leningrad <X
Moscow) 75: 792-803. | In Russian.]

Reveal. J. L. 1992. Validation of ordinal names of extant
vascular plants. Novon 2: 238—240.

- . 1993. New ordinal names for extant vascular

plants. Phytologia 74: 173—177.

Takhtajan, A. L. 1980. Outline of the classification of flow¬
ering plants (Magnoliophyta). Bot. Rev. 46: 225—359.

- . 1997. Diversity and Classification of Flowering

Plants. Columbia Cniv. Press, New York.

New Combinations in Mimosaceae

M. de Lourdes Rico Arce

Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3AB, United Kingdom.

1 . rico@rbgkew.org. u k

ABSTRACT. Three new combinations, one in Abar-
ema and two in Albizia (tribe Ingeae of the Mimo¬
saceae), are proposed based on the study of fertile
material from Central and South America. Two of
these are necessary for the completion of the Mi¬
mosaceae treatment for Flora de Nicaragua.

In the course of preparing treatments for genera
of the tribe Ingeae for the Flora de Nicaragua, it
has been necessary to propose new combinations
in the Mimosaceae. In the past 20 years, large parts
of neotropical Pithecellobium and Albizia have been
divided into several small genera; this fragmenta¬
tion has not yet been generally accepted by forest¬
ers, agronomists, and parataxonomists. Nielsen
(1981) reviewed the genera in the tribe Ingeae and
stated that, historically, classification of the group
lias been based largely on fruit characters, leading
in many cases to “pod-genera” (Nielsen, 1981:
173). Most genera were difficult to separate when
only in flower, so he proposed a classification for
the whole tribe (covering 20 genera), based on veg¬
etative. floral, and carpological characters. Barneby
and Grimes (1996) established a framework of sev¬
en generic alliances (containing a total of 20 gen¬
era) for mainly neotropical taxa of tribe Ingeae. A
few cultivated Old World species of Albizia were
included. Barneby and Grimes’s (1996) classifica¬
tion was based largely on growth and branching
patterns, as well as on developmental characters of
vegetative and floral buds.

The first species under consideration in the pres¬
ent paper falls within the Abarema alliance of Bar¬
neby and Grimes (1996). Abarema contains ap¬
proximately 45 species, with inflorescence, fruit,
and seed characters being consistent within the ge¬
nus. The species was first described as Pithecellob¬
ium (“ Pithecolobium ”) acreanum Macbride (1943),
but was tentatively transferred to the genus Hydro¬
chorea by Barneby and Grimes (1996: 33) as ‘ ' Hy¬
drochorea (?) acreana (Macbride) Barneby &
Grimes.” The genus Hydrochorea Barneby &
Grimes is mainly characterized by a lomentaceous
lruit. Hydrochorea acreana was known to Barneby
ami Grimes by only two flowering specimens from
the Brazilian state of Acre, and they commented

(1996: 34) that “pending discovery of the fruit, its
affinity cannot be assessed with finality.” They went
on to state, “The foliage, except for smaller petiolar
nectaries, is more reminiscent of Abarema adeno-
phora."

Robleto 678 (from Nicaragua) and Herrera 7008,
Rivera 2850, and Zumbado 98 (all from Costa Rica)
are fruiting specimens that all possess leaflets and
relatively small petiolar nectaries that closely
match those of Pithecellobium acreanum. Together
these four collections represent two disjunct distri¬
butions of the species under discussion here.

Zumbado 98 was cited by Barneby and Grimes
in their exsiccatae (1996: 283) as Abarema macra-
denia (Pittier) Barneby & Grimes, another disjunct
species (Central and South America; Barneby &
Grimes, 1996: 59). However, A. macradenia has a
thick coriaceous fruit and obliquely oblong-elliptic
or very obtusely rhombic leaflets, and the duplicate
specimen of Zumbado 98 at K lacks these char¬
acteristics.

Robleto 678 (BM) from Nicaragua was not cited
in the Barneby and Grimes (1996) exsiccatae, but
was given by them as the basis of Abarema ricoae
Barneby & Grimes, nom. provis (1996: 110).

In their discussion under this species Barneby
and Grimes stated, “In foliage the described speci¬
men closely resembles Balizia acreana, and was so
identified in 1988 by L. Rico (BM ).” In clarification,
this specimen was not identified as Balizia acreana
but as Abarema acreana (Macbride) L. Rico, ined.
The genus Balizia Barneby & Grimes was not de¬
scribed until 1996. Barneby and Grimes’s statement
that Robleto 678 closely resembles Balizia acreana
is an error. Presumably, the reference is to Hydro-
chorea acreana. A key difference between their pro¬
visional species Abarema ricoae and Hydrochorea
acreana is peduncle length. Peduncle length oi Bal¬
izia acreana (= Hydrochorea acreana) is cited as 2.5
cm (Barneby & Grimes, 1996: 1 10) but is elsewhere
mentioned as 20-33 mm (1996: 34). Peduncle
length for Abarema ricoae is given as ±11 cm in the
description but only 11 cm in the discussion. Careful
measurement of the K isotypic of Pithecellobium
acreanum ( Krukoff 5681, a flowering specimen) re-

Novon 9: 554-556. 1999.

Volume 9, Number 4
1999

Rico Arce

New Combinations in Mimosaceae

555

veals peduncles of up to 3.5 cm long. Zumbado 98
(K), a specimen in fruit, has a peduncle about 7 cm
long, and Rivera 2350 (K), also in fruit, has a pe¬
duncle 9.5 cm long. Evidently the assumption (Bar-
neby & Grimes, 1996: 110) that “peduncles and
pedicels of other abaremas do not elongate appre¬
ciably alter anthesis” is not supported. Abarema n-
coae (as represented by Robleto 678) is at the upper
end of peduncle length range for Pithecellobium
acreanum (= H. acreana).

Also noteworthy is the fact that Hydrochorea
acreana closely resembles Abarema adenophora
(Ducke) Barneby & Grimes, and the two ultimately
may prove to be conspecific alter more fieldwork
has been done. In this case, Abarema adenophora
would have priority.

What is clear is that Hydrochorea (?) acreana is
in fact an Abarema, and Abarema ricoae Barneby
& Grimes nom. provis. is conspecific. The neces¬
sary combination is proposed below.

Abarema acreana (Macbride) L. Rico, comb. nov.
Basionym: Pithecellobium (“ Pithecolobium ”)
acreanum Macbride, Publ. Field Mus. Nat.
Hist. Bot. Ser. 13(3.1): 51. 1943. Hydrochorea
(?) acreana (Macbride) Barneby & Grimes,
Mem. New York Bot. Card. 74(1): 33. 1996.
TYPE: Brazil. Rio Acre: on terra firma, mouth
of Rio Macauhan, 24 Aug. 1933, Kruhoff 5631
(holotype, F not seen; isotype, K).

The second species under discussion in this pa¬
per was first described in 1825 as Inga pedicellaris
DC. and has subsequently had a complicated no-
menclatural history', with the epithet being trans¬
ferred by Bentham in 1844 to Pithecellobium , by
Kuntze in 1891 to Feuillea, by Killip ex Record in
1940 to Samanea, by Kleinhoonte in 1940 to Ma-
crosamanea, and most recently by Barneby and
Grimes in 1996 to Balizia. It has also been de¬
scribed by Vellozo in 1829 as Mimosa terminalis.

Balizia pedicellaris (DC.) Barneby & Grimes tails
within section Balizia of Barneby and Grimes’s
small genus Balizia and is very closely related to
(perhaps even conspecific with) B. elegans (Ducke)
Barneby & Grimes, the only other species in sec¬
tion Balizia. The third species of Balizia, B. leu-
cocalyx (Britton & Rose) Barneby & Grimes, was
accommodated in the monotypic section Leucosa-
manea (Barneby & Grimes, 1996: 36), and they
included the combination Albizia leucocalyx (Bar¬
neby & Grimes) L. Rico (1992) in synonymy.

In their key to the three species of Balizia, Bar¬
neby and Grimes (1996: 35) separated B. pedicel laris
from B. elegans by calyx and corolla lengths (both

longer in B. elegans) and by the distance between
the transverse fibers of the pod mesocarp (greater in
B. elegans). They stated that B. pedicellaris is wide¬
spread in South America, while B. elegans has a
bicentric distribution in lower Amazonian Brazil and
southeastern Central America (Nicaragua, Costa
Rica). Zamora (1991: 132), in his treatment of the
Mimosaceae for Costa Rica, gave calyx and corolla
measurements for Pithecellobium elegans Ducke that
confirm the Costa Rican material is best placed in
this species. However, my measurements ol all ma¬
terial from Nicaragua fall exactly within those of B.
pedicellaris, and I conclude that the Nicaraguan
specimens are better placed in that taxon.

When considering the genus Albizia, Barneby and
Grimes (1996) dealt with Old World species only
when they had been introduced into the Americas.
This leaves an inconsistency in that several neotrop¬
ical Albizia species have been placed in segregate
genera (e.g., Hesperalbizia, Pseudosamanea, Blan¬
ch etiodendron) by Barneby and Grimes (1996), but
most of the paleotropical species have not been dealt
with. Albizia, when considered pantropically, re¬
mains a genus with a great diversity of floral and
fruiting morphological characters. In this context,
Balizia pedicellaris and B. elegans are comfortably
accommodated in Albizia by their whitish seeds
(characteristic of other Albizia species) and inflores¬
cences that are similar to some Malesian species,
e.g., A. dolichadena (Kostennans) Nielsen and A. ro-
sulata (Kostennans) Nielsen. These latter two spe¬
cies also have fruits not dissimilar to those of Bali¬
zia, i.e., indehiscent or tardily dehiscent, cracking
between the seeds but not through the sutures. The
main fruit difference between those of Balizia and
these two Asian taxa is that the Asian species have
coiled fniits. On balance it seems wiser to adopt a
broader concept of Albizia until the genus has been
monographed across its pantropical range.

Balizia leucocalyx has already been recognized
as an Albizia (Rico, 1992). Based on the above ar¬
gument, B. pedicellaris and B. elegans are below
transferred into Albizia, thus effectively placing the
entire genus Balizia as a synonym of Albizia.

Albizia pedicellaris (DG.) L. Rico, comb. nov.
Basionym: Inga pedicellaris DC., Prod. 2: 441.
1825. Pithecellobium (“ Pithecolobium ”) pedi-
cellare (DC.) Bentham, in Hooker, London J.
Bot. 3: 219. 1844. Feuillea pedicellaris (DC.)
O. Kuntze, Rev. Gen. PI. 1: 88. 1891. Sama¬
nea pedicellaris (DC.) Killip ex Record, Trap.
Woods 63: 4. 1940. Macrosamanea pedicellaris
(DC.) Kleinhoonte, in Pulle, FI. Suriname 2
(2): 329. 1940. Balizia pedicellaris (DC.) Bar¬
neby & Grimes, Mem. New York Bot. Gard.
74(1): 37. 1996. TYPE: Cayenne (G-DC.).

556

Novon

Mimosa terminalis Vellozo, FI. Flum. 11: t. 30. 1829.
TYPE: “Habitat silvis maritimis.”

Albizia elegans (Ducke) L. Rico, comb. nov. Bas-
ionym: Pithecolobium elegans Ducke, Arch.
Jard. Bot. Rio Janeiro 3: 64. 1922. Balizia ele¬
gans (Ducke) Barneby & Grimes, Mein. New
York Bot. Card. 74(1): 40. 1996. TYPE: Brazil.
Para: “Alcobaga prope H. Tocantins,” 17 July
1916, Ducke 16271 (lectotype, designated by
Barneby & Grimes (1996: 40), MG not seen;
isolectotypes, BM, G, K, P, US not seen).

Acknowledgments. I thank G. P. Lewis for his
comments on multiple drafts of this paper, one

anonymous reviewer, and the journal editor for con¬
structive criticism.

l.iterature Cited

Barneby, R. C. A J. Vi. Grimes. 1996. Silk Tree, Guana-
caste, Monkey’s Earring. A Generic System for the Syn-
androus Mimosaceae of the Americas. Part I. Abarema,
Albizia and Allies. Mem. New York Bot. Card. 74(1):
1-292.

Nielsen. I. 1981. Ingeae. Pp. 173-179 in B. M. Polhill &
P. H. Raven (editors). Advances in Legume Systematic*.
Part 1. Royal Botanic Gardens, Kew.

Rico, A. M. de I.. 1992. New chromosome counts in neo¬
tropical Albizia. Havardia, and Pithecellobium, and a
new combination for Albizia (Leguminosae-Mimoso-
ideae— Ingeae). Bot. J. Linn. Soc. 108: 269-274.

Zamora, V. N. 1991. Tratamiento de la familia Mimosa¬
ceae (Fabales) de Costa Rica. Brenesia 36: 63-149.

Novedades Taxonomicas en Gochnatia (Asteraceae, Mutisieae)

Gisela Sancho

Departamento Cientffico tie Plantas Vasculares, Museo de La Plata, 1900, La Plata, Argentina

RESUM EN. En la presente nota dos de las tres var-
iedades no tipicas de la espeeie Gochnatia pani-
culata: G. paniculata var. lanuginosa y G. panicu-
lata var. polycephala son consideradas sinonimos
de G. paniculata. Gochnatia paniculata var. densi-
cephala es elevada a la categorfa de espeeie; la
nueva combinacion propuesta es: Gochnatia den-
sicephala (Cabrera) Sancho.

ABSTRACT. Two non-typical varieties, G. panicu¬
lata var. lanuginosa and G. paniculata var. poly¬
cephala, are sunk back into the synonymy of G.
paniculata, which as recognized here includes no
intraspecific taxa. Gochnatia paniculata var. den-
sicephala is raised to the rank of species; the new
combination proposed is: Gochnatia densicephala
(Cabrera) Sancho.

El genero Gochnatia Kunth pertenece a la tribu
Mutisieae e incluye alrededor de 70 especies dis-
tribuidas principalmente en centro y sudamerica;
solo dos especies tienen su area de distribut ion en
Asia y una en el sur de los Estados Unidos. Ac-
tualmente se aceptan seis secciones dentro de
Gochnatia (Cabrera, 1971): (1) sect. Gochnatia, que
incluye alrededor de 43 especies centro y suda-
mericanas; (2) sect. Pentaphorus, con dos especies
sudamerieanas; (3) sect. Moquiniastrum, que inclu¬
ye 20 especies sudamerieanas; (4) sect. Leucomeris,
con dos especies asiaticas; (5) sect. Hedraiophyl-
lum, con 7 especies americanas; (6) sect. Discoseris,
que inc luye tres especies brasilenas.

Gochnatia paniculata, una de las especies de la
seccion Moquiniastrum, incluye hasta el presente
cuatro variedades (Cabrera, 1971): var. paniculata,
var. densicephala, var. lanuginosa y var. polycepha¬
la. A1 tratar la seccion Moquiniastrum, Cabrera
(1971), considera al tipo de pelo y posicion de la
pubescencia en las caras tie las hojas como car-
acteres diterenciales de las especies. De este modo,
se forman dos grandes grupos de especies: las que
presentan hojas pubescentes en ambas caras (G.
barrosii Cabrera, G. mollissima (Malme) Cabrera, G.
paniculata, G. rusbyana Cabrera y G. velutina (Bon-
gard) Cabrera) y aquellas con hojas pubescentes
solo en la cara inferior (como las restantes espe¬
cies). Sin embargo, solo G. paniculata presenta tres

variedades con hojas pubescentes en ambas caras
y solo una con la cara inferior pubescente. Las cua¬
tro variedades fueron difereneiadas entre si por la
posicion y densidad de la pubescencia y la longitud
de los pedunculos de los capftulos (Cabrera, 1971).

Sobre la base de lo anteriormente expuesto y
como resultado de mi trabajo sobre la seccion Mo¬
quiniastrum (Sancho, 1997, ined.) en el que he exa-
minado mas de 90 ejemplares de G. paniculata,
considero que dos de las variedades no tipicas {po¬
lycephala y lanuginosa) deben ser incluidas en la
sinonimia de G. paniculata y que la var. densice¬
phala debe elevarse a la categorfa de espeeie. Las
modificaciones propuestas son las siguientes:

Gochnatia paniculata (Lessing) Cabrera, Notas
Mus. La Plata Bot., 15: 43. 1950. Basonimo:
Spadonia paniculata Lessing, Syn. gen. Com¬
pos.: 100. 1832. Moquinia paniculata (Les¬
sing) DC., Prodr. 7(1): 23. 1838. Gochnatia
paniculata (Lessing) Cabrera var. paniculata.
Revista Mus. La Plata, Secc. Bot. XII (66): 81.
1971. TIPO: Brasil. “Prope St. Paulo in Cam-
po Brasiliensum,” Sellow s.n. (holotipo, prob-
ablemente destruido, B); Brasil. Sao Paulo:
“Casa Branca, 23-9-89,” Loefgren 16994 (neo-
tipo, designado aquf, US!; isoneotipo, NY!). Fi-
gura 1.

Moquinia tornentosa Bongarcl, Mem. Acad. Imp. Sci.
Saint-Petersbourg, S6r. 6, 2: 43, 1838. TIPO: Brasil.
“Brasilia. In carnpis collibusque siccis pr. CaetG
Sept. 1824,” Riedel 569 (holotipo, LE no visto).

Moquinia polycephala Gardner, London J. Bot. 6: 458.
1847. Gochnatia paniculata var. polycephala (Gard¬
ner) Cabrera, Revista Mus. La Plata, Secc. Bot. XII
(66): 85. 1971. TIPO: Brasil. “Minas Gerais, Serro
do Frio, Aug./40, a shrub 6 feet high,” Gardner 4809
(holotipo, BM no visto; isdtipos, US!, K! y fotograffa
UP!, no vistos: G, MO, NY: fotograffa Serie del f ield
Museum 1868 of NY!, P, S, W). El isotipo de US,
tiene en su etiqueta ano 1842.

Moquinia congesta Gardner, London J. Bot. 6: 459. 1847.
TIPO: Brasil. “Minas Gerais, Shrub 4 to 5 feet high,
July/40,” Gardner 4810 (holotipo. BM no visto; isd-
tipos, G!, K!. US!, no vistos: NY; fotograffa Serie del
Field Museum 15861 of NY!, P. W). El isotipo de
G. tiene en su etiqueta fecha: “Juillet, 1842.”

Moquinia desertorum Gardner, London J. Bot. 6: 459.

Novon 9: 557-561. 1999.

558

Novon

K.L.M

400

Figura I. Gochnatia paniculala. — A. Kama fertil. — B. Krdctea externa del involucro. — C. Krdctea interna del
involucre. - — I). Flor hermafrodita. — E. Flor funcionalmente femenina. — F. Estambre de flor hermafrodita. — G. Es-
taminodio de flor funcionalmente femenina. — H-J. Variacidn de la forma de la hoja. — K-M. Pubescencia de cara
superior de la Idmina. (A— C. E, G de h)efgrea 16994, US; I). F de Regncll III 305, S; H y K de tirade 13550. UP; I
y I, de Ixtefgren 334. S; J y M de Dusdn 17107a, G.)

Volume 9, Number 4
1999

Sancho

Novedades en Gochnatia

559

1847. TIPO: Brasil. “Minas Gerais, Between Bio San
Francisco and Formigas, July 1840, a shrub 6 to 8
leet,” Gardner 4808 (holotipo, BM no visto; isotipos,
F!, K! y fotografta LP!, US!, no vistos: G, (ill, NY,
P, S, W). El isotipo de F, tiene en su etiqueta fecha:
“Juillet, 1842.'’

Maquinia lanuginosa Baker, en Martius, FI. bras. 6(3):
345. 1884. Gochnatia lanuginosa (Baker) Cabrera,
Notas Mus. I .a Plata, 15: 43. 1950. Gochnatia pan -
iculata var. lanuginosa (Baker) Cabrera, Revista
Mus. La Plata, Secc. Bot. XII (66): 83. 1971. TIPO:
Brasil. “Goyaz, Serra de Duro, Sept. 1839, 2-3 feet
high,” Gardner 3245 (lectdtipo, designado por Ca¬
brera (1971), K!). Paraiectotipo. Pohl 554 (k no vis¬
to, fotografta UP!).

La diferencia entre las variedades paniculata y
polycephala estaba sustentada principalmente por
las longitudes de los pedunculos de los capftulos:
sesiles o casi sesiles en la variedad paniculata y
en gran parte pedunculados en la variedad poly¬
cephala (Cabrera, 1971). Si bien se han hallado
ejemplares en los extremos del rango de longitud
de los pedunculos, en la mayorfa de los ejemplares
de las dos variedades existe superposicion o son
variables en su longitud de acuerdo al lugar de la
panoja en que se eneuentran. La variedad lanugi¬
nosa lue diferenciada de las restantes variedades
por su densa pubescencia en toda la planta y hojas
anohamente ovadas (Cabrera, 1971). Estos carac-
teres presentan superposicion con las variedades
polycephala y paniculata, de manera que es im-
posible la diferenciacion de muchos ejemplares ya
que existe continuidad en la longitud de los pe¬
dunculos, de la pubescencia y el ancho de las ho¬
jas. La Figura 1 muestra la secuencia de aumento
en el ancho de la hoja (H — J) y su correspondiente
aumento en la longitud de los pelos (K— M), de ma¬
nera que puede observarse la continuidad de ambos
caracteres.

Observacion. No he podido hallar el tipo de
Spadonia paniculata Lessing, leg. Selloiv s.n. Fu-
eron consultados los herbarios B, BM, BR, F, G,
GH, HAL, K, M, MO, NY, P, POM, S, SP, UPS, US,
W, WU pero en ninguno de ellos se encuentra de-
positado ese ejemplar tfpico. Por otro lado tampoco
el Dr. Cabrera, en su exhaustiva revision del genero
Gochnatia (1971), pudo hallar dicho especimen.
Por estas razones he decidido neotipificar la espe-
cie Spadonia paniculata designando como neotipo
al ejemplar Loefgren 16994 depositado en US, que
posee un isoneotipo en NY.

Ejemplares adicionales examinados. BRASIL. Minas
Gerais: Mun. Tres Cora^oes, Hatschhach 29835 (MO, NY,
S, US); Paraopeba, Faz. do Bebedouro, Heringer 4028
(LP); Mun. Belo Horizonte, Mello Barreto 4201 (F), 4207
(F); Passos, Regnell 111.805 (S); Lagoa Santa, Warming s.n.
(S); cachoeira do campo, Casaretto 2840 (G); Serra do Ca-

tiara, Duarte 3035 (NY); Serra do Barranco, Mahelles
1717 (R); Lavras, Pereira-ESAL 2347 (UEC); Mun. Pas¬
sos, Vichnesky 14798 (UEC); Campus UFMG, Fernandes
et Neto 065 (UEC); Lavras, Altinopolis, Dac et al. 2878
(UEC); Lavras, Calegeri Lopes & Vichneswki 15797 (UEC);
Diamantina, Zickel et al. 21688 (UEC); Carrancas, Sernir
et al. 19577 (UEC); Massa, Brade 13550 (LP); “environs
de R. Janeiro et D’Ouro Preto,” Glaziou 15044 (K); La¬
vras, Maia 167 M. (LP); Minas, Glaziou 19557 (k); Mun.
de Gouveia, Corrego do Tigre, Hatschhach 27016 (k. S);
Serra do Lenheiro, Valka Alves 1354 (NY). Sao Paulo:
Serra da Bocaina, Glaziou 78 (LP, R); Jardim Botanico, 5—
8-1940, Handro s.n. (LP); Campos de Butantam, Joly 450
ABJ (LP); Butantam, Hoehne 538 (NY, US); Butantam, 24—
10—1918, Hoehne s.n. (SP); Campos da Bocaina, Glaziou
11098 (G, R); Serra da Bocaina, Brade 21174 (G, F, NY.
S); Cotia, Kulhmann s.n. (SP); Atibaia, Serra do Itapatinga,
Tamandare & Brade 967 (US); Itarare, Dusen 9938 (S,
US); Rod. Itarare-Bom. Sucesso, Mottos 13994 (SP); Sao
Jose dos Campos, Loefgren 334 (S); Sao Jose dos Campos,
Mimura 551 (NY, SP, US); Serra da Bocaina, Shepherd &
Shepherd 12827 (UEC, US); Capital, von Emelen 1435
(US); Mun. Sao Jose dos Campos, Eiten & Mimura 3352
(US); Est. Exp. de Itarare, Leitao Filho 879 (LP, UEC);
Rio Preto, Camargo A Marinis 8 (HAS); Perto de Mogf,
Mottos 8153 (HAS); Butantam, Diogo 538 (R); Campos da
Bocaina, Glaziou 11098 (R); Itapeva, Souza et al. 8705
(UEC); Riberao Preto, Sard 8497 (UEC); Burr he! I 4769-
2 (k, NY). Parana: Itarare, Dusen 9935 (US); Jaguariai-
va, Fda. Cajuru, Hatschhach 20057 (LP); Jaguariaiva. Du¬
sen 9571 (NY, S), 13087 (S, SI). 15105 (G, S), 15105a
(S), 17107 (G, NY, S, US), 17107a (MO, S); Villa Velha.
Dusen 14825 (S); Rio Tibagi, Hatschhach 14688 (LP. NY.
US); Mun. Ortigueira, Pedra Branca, Hatschhach 53307
& Silva (MO); Mun. Sao Jeronimo da Serra, Fda. Nho O'.
Hatschhach 24788 A Guimaraes (MO); Mun. Cerro A/.ul.
Serra da Canha, Hatschhach 32613 (MO, SP); Bocaiuva
do Sul, Santa Ana, Hatschhach 27084 (S). Rio de Janei¬
ro: near Rio de Janeiro, Glaziou 11098 (k) (los ejempla¬
res Glaziou 11098 de K, G y R tienen igual colector y
numero pero diferente fecha o localidad).

Gochnatia densicephala (Cabrera) Sancho, comb,
et stat. nov. Basonimo: Gochnatia paniculata
var. densicephala Cabrera, Revista Mus. La
Plata. Secc. Bot. 12(66); 86. 1971. TIPO:
“Near Rio de Janeiro, Nov. 1879,” Glaziou
11072 (holotipo, K!). Figura 2.

Los caracteres que diferencian a Gochnatia den¬
sicephala de G. paniculata son: (1) Hojas glabras
en la cara superior de la hoja (sin pelos en T, “mal-
pigiaceos,” y solo con pelos glandulares). (2) Capf¬
tulos siempre dispuestos en panojas de glomerulos
mientras que en G. paniculata las panojas pueden
ser glomeriformes, espiciformes o racemiformes.

Si bien las hojas inmaduras de G. densicephala
presentan ocasionalmente pelos en T en ambas ca-
ras, esta condicion es comun a todas las hojas jove-
nes de las restantes especies de la seccion Moqui-
niastrum independientemente del estado adulto.
Las hojas maduras de G. densicephala se caracteri-
zan por ser brillantes y glabras en el haz.

560

Novon

1 OO yum

!' igura 2. Corhnatia densiccphala. — A. Rama fertil. — 15. Braetea externa del involucro. — C. Braetea interna del
involuero. — I). Aquenio de flor hermafrodita. — K. Flor hermafrodita. — F. Flor funeionalmente femenina. — G. Fstam-
bre de flor hermafrodita. — H. Fstaniinodio de flor funeionalmente femenina. — 1. Pubescencia de eara inferior de la
lamina. (A-K. G de Leoncini 88, I.P; I de Yamamoto 1096. I KC; F-H de Hoehne s.n., 1-6-1927, SP).

Volume 9, Number 4
1999

Sancho

Novedades en Gochnatia

561

Distribut ion y habitat. Brasil, en el sur del es-
tado de Minas Gerais, Rio de Janeiro, y norte tie
Sao Paulo. Freeuente en campos del Cerrado, en
morros o sierras, colectado hasta los 1100 m s.m.
Su distribucion coincide con parte del area de G.
paniculata (en los estados de Minas Gerais, Rio tie
Janeiro y Sao Paulo). Esta ultima espeeie se ex-
tiende mas al sur hasta el estado de Parana.

Ejemplares adicionales examinados. BRASIL. Minas
Gerais: Mun. Belo Horizonte, Pampulha, Williams & Assis
7393 (LI’); Poyos de Caldas, Campo do Saco, linker &
Roppa 433 (I.P. SI), 453 (SI), 464 (SI): Poqos de Caldas,
Fax. Teixeira, Leoncini 88 (LP, SI); 1838, Claussen 137
(K); voyage e/1816—1821, A. de Sainl-Hilaire 1490 (K);
Ouro Preto, Damazio 1410 (LP); Manhuayu, Hatschbach
49391 (K. MO): Caldas, Regnell 1-213 (k); Mono das Ped-
ras, near Belo Horizonte, Williams <8: Ass is 8017 (K); Loros
tie Caldas, Campo do Saco, Yamamoto et al. 1010 (ULC),
1096 (UEC); Poyos de Caldas. Pereira et al. 870 (UEC);
Jaboticatubas, Serra de Cipo, Sernir & Sazima 2725
(UEC); Pottos de Caldas. Leitao Filho et al. 1681-A (UEC);
Serra do Tritiba, Silveira <£• Silveira 2166 (R); Pico da
Bandeira, Shepherd et al. 5771 (UEC); Sao Sebastiao da
Praia. Sarti & Santos Filho 14418 (UEC); Campo do Saco.
Ijeitao Filho et al. 1126 (UEC); Ouro Preto, entre 1 tabirito
y Belo Horizonte. Davis & Sheperd 1). 59696 (UEC): La¬
vras, Serra tla Bocaina, Dae et al. 8231 (UEC). Sao Pab¬
lo: 25-7-1945. Hoehne s.n. (LP); Pico da Serra Negra. 1 —

6—1927, Hoehne s.n. (SP); Pico do Morro Pellado, perto
das Thermas de Lindoya, 4—6—1927, Hoehne s.n. (SP);
Mun. Guarulhos, Martins & Tamashiro 11243 (UEC). Rio
de Janeiro: Rio de Janeiro, Glaziou 12053 (G, k); Morro
da Cruz, Nova Friburgo, Siqueira 1030 et al. (UEC). Sin
Loc. Definida: Villa Franca et Pr. Goyaz, Riedel & Land
2365 (NY, US); Claussen s.n. (LP).

Agradecimientos. Agradezco a Susana E. Fre-
ire, Liliana Katinas y Jorge V. Crisci por la lectura
crftica del manuscrito, a Dan H. Nicolson por sus
comentarios acerca de la tipificacion de Spadonia
paniculata, al Consejo Nacional de Investigaciones
Cientificas y Tecnicas (CONICET) por el financia-
miento de mis investigaciones y a los curadores tit*
los herbarios F, G, HAS, K, LP. MO, NY. R. S, SI,
SP, UEC, y US por el prestamo de los materiales
estudiados.

I .iteratura Citada

Cabrera, A. L. 1971. Revision del genero Gochnatia
(Compositae). Revista Mus. La Plata. Secc. Rot. Xll(66):
1-160.

Sancho, G. 1997, inedito. Revision sistematica, analisis
cladfstico y biogeografico de la sect. Moquiniastrum tlel
genero Gochnatia (Asteraeeae, Mutisieae, Gochnati-
irtae). Lesis Doctoral, Facultad de Ciencias Naturales y
Museo. UNI.P. Argentina.

Typification of Two Linnaean Names in I lie Basellaceae

Karen Sidwell

Department of Botany, The Natural History Museum, Cromwell Road,
London SW7 5BD, United Kingdom

ABSTRACT. During the writing of the treatment of
Basellaceae for Flora Mesoamericana it became ap¬
parent that typification of the Linnaean names Bas-
ella alba L. and B. lucida L. required clarification.
A note to formally fix the application of these two
names is provided here.

The genus Basella consists of five species, four
in east and southeast Africa and Madagascar, the
other pantropical and widely cultivated (Sperling,
1987). The three Linnaean names for species of
Basella are now considered conspecific, and typi¬
fication of two of these names is clarified below.

In Species Plantarum, Linnaeus first described
two species of Basella L., B. rubra L. and B. alba
L. These two species were distinguished on a single
leal character: Basella rubra was described as
“ Basella foliis plants while Basella alba in con¬
trast was described as “ Basella foliis undatis ” (Lin¬
naeus, 1753: 272). A third species, Basella lucida
L., was described by Linnaeus in 1759, differing
from the other two in having subcordate, rather than
ovate, leaves and densely terminally branched,
rather than simple, peduncles.

Basella rubra , B. alba, and B. lucida were first
treated as a single species by Roxburgh (1832),
who used the name B. alba. Within B. alba Rox¬
burgh recognized five “varieties,” based on local
taxonomic knowledge, and he listed these five sub¬
groups as two “wild sorts” and three “cultivated
sorts” (Roxburgh, 1832: 275). Graham (1839) fol¬
lowed Roxburgh's account, but was erroneously
credited by Smith (1981) as the first to combine the
names under B. alba. Baillon (1887), on the other
hand, considered B. alba a variety of B. rubra, and
several authors subsequently followed that work
and accepted B. rubra as the name for the single
spet •ies. As Roxburgh (1832) was the first author to
place B. alba and B. rubra in synonymy, and in
doing so adopted B. alba, it (and not B. rubra ) is
the correct name under Article 11.5 of the Code
(see van Steenis, 1957; Smith, 1981, and references
therein; Greuter et al., 1994). Additionally, B. alba
is the name in current use, and the types desig¬
nated here are intended to maintain this usage.

Basella rubra was lectotypified by Verdcourt

(1968), who selected the unpublished drawing of a
fruiting plant in Herb. Hermann 5: t. 207 (lecto-
type, BM) that formed the basis for Linnaeus’s Flo¬
ra Zeylanica account, a choice Smith (1981) also
suggested as a logical lectotype and that has been
followed by Kellogg (1988). Basella rubra L. was
accepted as the type of the genus Basella L. by
Jarvis et al. (1993).

The types ol Basella alba L. and B. lucida L.
have not previously been clarified. A single syno¬
nym from Thran (Hurt. Carolsruh. 11, n. 125.
1747), “ Basella flore albo, foliis & caulibus viridi-
bus ,” was cited in the original description of B.
alba, and this has subsequently been cited as the
type of B. alba (miscited as Thran, Hurt. Carolsruh.
10, n. 100. 1747, by Verdcourt, 1968; Sperling,
1987; and Kellogg, 1988). However, the synonym
in the protologue does not have a visual element
and it cannot therefore serve as the type, which
must be a single specimen or illustration (Greuter
et al., 1994: Art. 8.1). There are no extant original
elements for the name Basella alba. Three speci¬
mens annotated as B. alba exist in the Linnaean
herbarium in Stockholm. However, none are anno¬
tated by Linnaeus and therefore cannot be consid¬
ered original material for the name and eligible for
leetotypification. A neotype is therefore required to
formally fix the application of the name Basella
alba.

In Species Plantarum edition 2 (1762: 390), Lin¬
naeus added a reference to an illustration of Basella
alba in Plukenet (1691: tab. 63, fig. 1). Plukenet
(1696: 252) used the polynomial “ Mirabili Peruvi-
anae affinis tinctoria Betaefolio sea n dens" for the
species and cited the illustration in van Rheede
(Hortus Malabaricus 7, tab. 24. 1688) under that
name. The plate in Plukenet clearly illustrates Bas¬
ella alba L. as it is now circumscribed. Having cit¬
ed the distribution of Basella alba as “Syria?” (Lin¬
naeus, 1753: 272) in the first edition of Species
Plantarum, Linnaeus followed Plukenet in 1762,
stating that B. alba was from China. As possible
neotypes, the Chinese specimens available were
considered of rather poor quality. Therefore, the
neotype selected here is a specimen with plenty of

Nov ON 9: 562-563. 1999.

Volume 9, Number 4
1999

Sidwell

Typification in Basellaceae

563

fertile material, and detailed locality data, from Ne¬
pal.

Although Basella lucida L. was recognized by
Moquin-Tandon (1849) as a distinct species, it has
otherwise been considered a synonym of B. alba L.
by most authors since Linnaeus. In his treatment
of the three species of Basella, Linnaeus (1762:
391) added a note that the species were clearly very
closely related: Affines nimium sunt hae tres spe¬
cies, sed constantes se servant in hortis." The short
description of B. lucida (Linnaeus, 1762) fits within
the current circumscription of Basella alba, and
there is no doubt that the name should be placed
in synonymy.

The nomenclature and synonymy of Linnaean
names within Basella are as follows:

Basella alba L„ Sp. PI. 272. 1753. TYPE: Nepal.
Mahakali Zone: Kanehapur District, 15 mi. W
of Dhangarhi, 7 Dec. 1966, Nicolson 2848,
(neotype, BM; isoneotype, US not seen).

Basella lucida L. Sp. PI. (ed. 2) 391. 1762. TYPE: Tai¬
wan. Taipei). May 1909, Faurie 109 (neotype, BM).

Basella rubra L„ Sp. PI. 272. 1753. TYPE: Herb. Her¬
mann 5 t. 207 (lectotype, designated by Verdcourt
(1968), BM).

Acknowledgments. I thank Steve Cafferty and
Charlie Jarvis of the Linnean Plant Name Typifi¬
cation Project for providing nomenclatural assis¬
tance and advice with this paper.

Eiterature Cited

Baillon, II. 1887. Series des Baselles In H. Baillon, llis-
toire des Plantes 9: 145-148. Librairie Hachette, Paris.

Graham, .1. 1839. Catalogue of the Plants Growing in Bom¬
bay and its Vicinity. Government Press, Bombay.
Greuter, W., F. R. Barrie, H. M. Bordet. W. G. Chaloner.
V. Demoulin, I). L. Hawksworth, P. M. Jorgensen. I). H.
Nicholson, P. C. Silva, P. Trehane & J. McNeill. 1994.
International Code of Botanical Nomenclature (Tokyo
Code). Regnum Veg. 131.

Jarvis, C. E„ F. R. Barrie, I). M. Allan & J. L. Reveal.
1993. A list of Linnaean generic names and their types.
Regnum Veg. 127: 23.

Kellogg, E. A. 1988. Basellaceae. In R. A. Howard, Flora
of the Lesser Antilles: Leeward and Windward Islands
4: 207. Harvard University, Massachusetts.

Linnaeus, C. 1753. Species Plantarum. Stockholm.

- . 1759. Systema Naturae. Stockholm.

- . 1762. Species Plantarum, Ed. 2. Stockholm.

Moquin-Tandon, C. H. B. A. 1849. Basellaceae. In A. de
Candolle, Prodromus Systematis Regni Vegetabilis 13:
220-230. Paris.

Plukenet, E. 1691. Opera Omnia Botanica 1. Phytogra-
phia, tab. 63, fig. 1. Londini.

- . 1696. Opera Omnia Botanica IV, Almagestum

Botanicum pp. 252—253. Londini.

Rheede tot Draekestein, H. A. van. 1688. Hortus Mala-
baricus 7, tab. 24, p. 45. Amstelodami.

Roxburgh, W. 1874. Flora Indica: or Descriptions of In¬
dian Plants. Reprinted literatim from Carey's edition of
1832, Calcutta.

Smith. A. C. 1981. Flora Vitiensis Nova. Lawai, Kauai,
Hawaii.

Sperling, C. 1987. Systematics of the Basellaceae. Un¬
published Ph.l). Thesis, Harvard University, Cam¬
bridge, Massachusetts.

Steenis, C. G. G. J. van. 1957. Basellaceae. In C. G. G.
J. van Steenis, Flora Malesiana, 1st ed., 5: 300—304.
Noordhoff-Kolff N.V., Djakarta.

Thran, C. 1747. Hortus Carolsruhanus. Loeraci.
Verdcourt, B. 1968. Basellaceae. In E. Milne-Rehead &
R. M. Polhill, Flora of Tropical East Africa. Royal Bo¬
tanic Gardens, Kew.

Two New Combinations in Apocynaceae, Asclepiadoideae

from Mesoameriea

W. I). Stevens

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. The combinations Matelea fulvida (F.
Ballard) W. D. Stevens and Dictyanthus asper (Mill¬
er) W. D. Stevens are proposed to make them avail¬
able for the forthcoming Flora de Nicaragua.

Matelea fulvida (F. Ballard) W. D. Stevens, comb,
nov. Basionym: Gonolobus fulvidus F. Ballard,
Bot. Mag. 163: t. 9611. 1940. Based on Vin-
cetoxicum grandiflorum Standley, Publ. Car¬
negie Inst. Wash. 461: 83. 1935. Matelea
grandiflora (Standley) Woodson, Ann. Missouri
Bot. Card. 28: 235. 1941. Not Vincetoxicum
grandiflorum (R. Brown) Kuntze. Revis. Gen.
PI. 2: 424. 1891. TYPE: Belize. Machaca: 12
Aug. 1933, Schipp S575 (holotype, F; isotypes,
MO, UC, US).

Vincetoxicum grandiflorum Standley is a later
homonym of Vincetoxicum grandiflorum (R. Brown)
Kuntze. The replacement name Gonolobus fulvidus

F. Ballard predates by one year Woodsons “comb,
nov.,” which is effectively also a replacement name.
Therefore, Ballard’s epithet must be used for this
Mesoamerican species.

Dictyanthus asper (Miller) W D. Stevens, comb,
nov. Basionym: Cynanchum asperum Miller,
Card. Diet. (ed. 8): n. 6. 1768. Matelea aspera
(Miller) W D. Stevens, Phytologia 32: 396.
1975. TYPE: Mexico. Veracruz: “E. Vera
Cruz,” 1730, Houstoun s.n. (holotype, BM).

Ongoing studies of the large and complex genus
Matelea have demonstrated that some elements, in¬
cluding Dictyanthus, deserve generic status. Full
synonymy for this species, when treated as Matelea,
can he found in Stevens (1975).

Literature Cited

Stevens, W. I). 1975. Notes on the genus Matelea (Apo-
evnaeeae s.l.). Phytologia 32: 387-4-06.

Novon 9: 564. 1999.

A New Marsdenia (Apocynaceae, Asclepiadoideae) from

Baja California

W. D. Stevens

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Veron ica Juarez- Jaimes

Institute) de Biologfa, UN AM, Apartado Postal 70-233, Del. Coyoaean, 04510 Mexico,

I). F., Mexico

ABSTRACT. Marsdenia carterae, a new species of
Apocynaceae (Asclepiadoideae, Marsdenieae) from
Baja California Sur, Mexico, is described and il¬
lustrated. This is the first North American species
of Marsdenia with verticillate leaves and is also
notable in lacking a developed corona and for its
non-twining habit.

Marsdenia carterae W. D. Stevens & Juarez-Jaimes,
sp. nov. TYPE: Mexico. Baja California Sur:
eastern escarpment of Cerro Mechudo between
Portezuelo de San Antonio and La Sabanilla,
Sierra de la Giganta, 675 m, 24°49'N,
110°44'W, occasional on cliffs and talus, 4
Nov. 1971, R. Moran 18990 (holotype, UC;
isotype, MO). Figure 1.

Distribution map and additional images can be
found in W’TROPICOS (http://mobot.mobot.org/
Pick/Search/piek. html).

A speeiebns ceteris Marsdeniae regionis foliis verticil-
latis habitu haud volubili corona obsoleta distinguenda.

Shrub to 1.5 m tall and 2.5 m wide, branches
decumbent or pendent, stems with thick, dark,
corky bark on older wood, bark thinner and whitish
on younger wood, branchlets roughened by raised
leaf scars, densely pubescent with appressed, ±
curly, multicellular hairs to ea. 0.15 mm long, in¬
ternodes 1 — 5( — 25) mm long, ribbed except when
still herbaceous; latex white; roots unknown.
Leaves 3, rarely 1 or 2, per node, without pseu¬
dostipules but with 1—3 colleters in each stipular
area, blades 5—8.5 cm long, 4-8 mm wide, very
narrowly lanceolate to lorate or linear, apex acute
to rounded or minutely mucronate, base acute, mi¬
nutely pubescent when young, glabrescent except
hairs ± persistent along midrib of both sides, mid¬
rib prominent below, sulcate above, especially near
petiole, venation pinnate, lateral veins obscure, ca.
7—13 pairs, middle veins arising ca. 55° to midrib.

margin thickened and ± revolute, colleters 0-3, at
adaxial base of blade; petiole 4—6 mm long, with
indumentum of stem. Inflorescence extra-axillary, 1
per node and mostly not on adjacent nodes, cy-
mose, congested-racemiform, or congested-panicu-
liform, sometimes bearing a fully formed leal op¬
posite the first flower or branch, 5-15-flowered,
with indumentum of stem, peduncle 2-5 mm long,
pedicel 3—4 mm long, bracts to 2.5 mm long, 1.1
mm wide, lanceolate; calyx divided to base, with

1- 2 colleters below each sinus within, lobes 2.2-

2.5 mm long, 0.9-1. 4 mm wide, ovate or elliptic,
with rounded tips, green with scarious margins, pu¬
bescent outside, glabrous inside; corolla rotate with
ascending lobes, aestivation narrowly imbricate and
dextrorse, glabrous outside, glabrous inside except
throat densely barbate with stiff, white, unicellular
hairs to ca. 0.5 mm long, these spreading at sinuses
and retrorse between, cream-white, tube 1.2— 1.4
mm long, lobes 2.3-3. 1 mm long, 1.5-2. 2 mm
wide, elliptic with obliquely rounded tips; gynos-
tegium ca. 2.2 mm tall, stipe ca. 1.7 mm long, co¬
rona obsolete, anthers rectangular, ca. 1.0 mm long,

1.5 mm wide, terminal appendages ovate, ca. 0.8
mm long, 0.6 mm wide, ± erect and appressed to
stvle apex, wings straight, parallel, ca. 0.5 mm long,
0.3 mm wide; corpusculum ca. 0.25 mm long, 0.08
mm wide, linear, dark red-brown, translator ca.
0.08 mm long, translucent and unwinged, pollinia
erect, ca. 0.25 mm long, 0.15 mm wide, obovate,
pale yellow; style apex conical, smooth, shallowly

2- lobed at apex, ca. 0.4 mm wide. Follicles single
or occasionally paired, divergent when paired, fu¬
siform-attenuate to napiform, somewhat asymmet¬
rical, smooth, 7-12 cm long, 1.3-1 .8 cm wide,
densely tomentose, follicle wall ca. 0.5 mm thick,
herbaceous; seeds obovate, flat, ca. 11 mm long, 7
mm wide, pale yellow-brown with a pale brown
margin 0.4— 0.5 mm wide, margin smooth and en¬
tire, surface smooth and glossy, coma about 2.5 cm
long, pale tawny.

Novon 9: 565-567. 1999.

566

Novon

Figure 1. Marsdenia carterae W. I). Stevens & Juarez- Jaimes. — A. Flowering branch. — B. Fruiting branch. — C.
Flower, corolla partially opened. — I). Pollinarium. Drawn from the type collection.

Volume 9, Number 4
1999

Stevens & Juarez-Jaimes
Marsdenia carterae from Baja California

567

This species seems to he restricted to the eastern
slopes of the Sierra de la Giganta, on dry cliffs and
talus slopes, 75—1050 m in elevation. Sierra de la
Giganta is an elongate range on the southeastern
coast of Baja California. The few collections suggest
that it flowers and fruits throughout the year.

Marsdenia is a genus of perhaps 200 species
found throughout the tropics and subtropics. This
is the first North and Central American species with
vertieillate leaves and only the second with a non¬
twining habit; the other non-twining Mexican spe¬
cies, M. neriifolia (Decaisne) Woodson, is an erect
shrub. Marsdenia carterae is superficially similar in
habit and leaf arrangement to M. ericoides Schlech-
ter, an unrelated species from New Caledonia
which (nay occupy similarly dry habitats. While the
corona of Marsdenia is often reduced to a fleshy
ridge on the back of each anther, this species seems
to entirely lack the structure, a character state per¬
haps unique among species of Marsdenia without
long, tubular corollas. Marsdenia is the only New
World genus unambiguously placed within the tribe
Marsdenieae, characterized by erect, uniformly fer¬

tile pollinia, and this species clearly falls within
the genus, although probably in a rather isolated
position. No clearly related species have been iden-
tified. There is no adequate infrageneric classifi¬
cation of the genus, and this species would not fit
well into any of the inadequate ones.

Paratypes. MEXICO. Baja California Sur: nearly
vertical N-facing slope near summit of I ’i Ion de las Parras,
W of Loreto, Sierra de la Giganta, 1050 m. 25°58'N.
I 1 1 °30' W, 14 Mar. 1961, A. Carter & //. Sharsmith 4220
(MO, UC); steep N-facing slope, Canada de Tripin 8W of
Puerto Escondido, Sierra de la Giganta, 400 m,
25°47.25'N, 111°21'W. I Dec. 1961,4. Carter 4353 (MO,
UC); at mouth of Cajon del Salto, Bahia |Puerto| Agua
Verde, Sierra de la Giganta, 75 m, 25°30'N, 1 1 1°05'W.
25 Aug. 1971, A. Carter 56 05 (UC); lower cliffs, N base
of Cerro Mechudo, Sierra de la Giganta, 700 m, 24°48'N,
I 1 0°43'W. 2 Nov. 1971. R. Moran IH902 (UC).

This remarkable new species is dedicated to An-
netta Mary Carter (1907—1991), who first discov¬
ered the species in 1961 and brought it to the se¬
nior author’s attention many years ago. The
illustration was prepared by Alba Arbelaez. Bruce
Hansen and Mark Fishbein provided useful com¬
ments on an earlier version of the manuscript.

Dos Nuevas Especies de Simira (Rubiaceae) de Colombia

Charlotte M. Taylor

Missouri Botanical Garden, P.O. Box 299, St. Louis Missouri 63166-0299, U.S.A.

RESUMEN. Se describen dos especies nuevas bas-
adas en colecciones recientes de Colombia: Simira
cesariana C. M. Taylor, del departamento de Cesar,
y S. hirsuta C. M. Taylor, del Canon del Rio Claro
en el departamento de Antioquia.

ABSTRACT. Two new species are described based
on recent collections from Colombia: Simira ces¬
ariana C. M. Taylor, from Cesar Department, and
S. Iiirsuta C. M. Taylor, from the Rio Claro Canyon
in Antioquia Department.

Durante el estudio de colecciones recientes de
Colombia, se descubrieron las dos especies nuevas
tratadas aqui. El genero Simira Aublet (antes tra-
tado ampliamente bajo el sinonimo Sickingia
Willdenow; Steyermark, 1972) se caracteriza en
parte por los frutos, que son capsulares, lenosos,
globosos, lisos y loculicidos con dos valvas hemis-
fericas, o estas a veces tardiamente partiendose otra
vez en dos para un total de cuatro valvas. Tenemos
hoy dia un entendimiento limitado de este genero,
entonces es diflcil decir con certitud las afinidades
de estas especies nuevas. Sin embargo, estas dos
especies nuevas se distinguen claramente de las
otras especies de Simira ya conocidas por los ear-
acteres mencionados debajo de cada descripcion.

Simira cesariana C. M. Taylor, sp. nov. TIPO: Co¬
lombia. Cesar: ca. 5 km W of Manaure on road
to La Paz, 10°22'N, 73°08'W, 540 m, 26 Apr.
1987, A. Gentry & //. Cuadros 57130 (holoti-
po, MO-3585016). Figura 1A, B.

Haec species a congeneris pubescentia puberula brev-
iter pilosa vel hirsutula, loliis tenuibus basi abrupte trun-
catis vel cordulatis, limbo calycino truncato vel late et non
profunde lobulato 1.2— 1.5 mm longo atque friictu diametro
3.5— 4.0 cm distinguitur.

Arboles, floreciendo a 10 m de altura; tallos pub¬
erulos o corto-pilosulos a glabrescentes. Hojas
opuestas; Idminas oblanceoladas, 8—15 cm de lar¬
go, 4—7 cm de ancho, con apice agudo a ligera-
mente acuminado, con base angosta y luego abrup-
tamente truncada a cortamente cordada,
papiraceas, en la haz glabras excepto puberulas en
la costa y las venas secundarias, en el enves mod-
erada a densamente hirsutulas; venas secundarias

13-17 pares, debilmente extendiendose hasta las
margenes, a veces con domacios de tricomas en las
axilas, en la haz la costa y las venas secundarias
planas a ligeramente sulcadas y la venacion menor
reticulada y plana a promfnula, en el enves la costa
prominula a prominente, las venas secundarias y la
venacion menor promfnulas; pecwlos 4—7 mm de
largo, densamente puberulos a pilosulos; esttpulas
interpeciolares, caducas, triangulares a laneeola-
das, ligeramente resinosas, en la superficie externa
densamente senceas, en la superficie interna gla¬
bras, hasta 15 mm de largo, agudas a acuminadas.
Inflorescencias terminales, subsesiles (o tripartidas),
4-5 cm de largo, con los ejes, bracteas y pedicelos
densamente puberulos a pilosulos, algunos ejes se-
cundarios subtendidos por hojas reducidas (o br&c-
teas foliaceas) de 1.5— 3.5 cm de largo pero los ejes
menores ebraeteados o raramente con bracteas an-
gostamente triangulares a angostamente elfpticas,
1.5—2 mm de largo; pedicelos hasta 2 mm de largo;
flares sesiles y pediceladas en cfmulas, con hipanto
angostamente eilmdrieo a angostamente turbinado,
densamente pil6sulo a puberulo, 1.5—2 mm de lar¬
go; Umbo calicino con exterior densa y completa-
mente pilosulo a puberulo o glabrescente cerca de
la margen, el interior glabro, 1.2— 1.5 mm de largo,
truncado o amplia y brevemente lobulado; corola ,
estambres, estilo y estigina no observados. Frutos
globosos a ligeramente oblados, lenosos, lisos, 3.5—
4 cm de diametro; semillas hemisfericas, 25—28 X
10—13 mm, con alas ca. % de este tamano.

Distribution, habitat y fenologta. En bosque
semihumedo a 440—540 m en el noreste de Colom¬
bia. Colectada despues de la antesis en abril, con
frutos maduros en enero.

Esta especie se distingue por la pubescencia cor-
ta, hojas con la base abruptamente truncada a cor¬
tamente cordada, limbo calicino 1.2— 1.5 mm de
largo y truncado o amplia y brevemente lobulado y
los frutos 3.5-4 cm de diametro; se conoce solo de
la localidad tipiea. Desafortunadamente, las inflo-
rescencias e infructescencias no se preservaron
bien en las muestras estudiadas. El epiteto se re-
fiere al departamento colombiano donde se ha en-
contrado esta especie. Simira cordifolia (Hooker f.)
Steyermark es similar a esta especie nueva. Simira

Nov ON 9: 568-570. 1999.

Volume 9, Number 4
1999

Taylor

Nuevas Especies de Simira

569

Figura I. A. B. Simira cesariana C. VI. Taylor, basada on Gentry A' Cuadros 57130. — A. Rama con Irutos inmaduros.
— B. Cimula de la inflorescencia con Ires (lores despues de la antesis, con hipanto y limbo calieino. C, I). Simira
hirsula C. M. Taylor, basada en Cogollo A Borja 1577. — C. Rama eon inflorescencia despues de la antesis. — 1). Flor
en la antesis, parcialmente disectada. A, C escala de 5 cm; B. I) escala de 5 mm.

cordifolia se distingue de S. cesariana por los or-
ganos vegetativos glabros a puberulos, en cada caso
menos pubescente; las hojas usualmente mayores,
13-23 X 5—15 cm, y de forma diferente, usual¬
mente mas anehas en el medio o la mitad basal;
las inflorescencias de forma diferente, peduncula-
das y no o solo ligeramente tripartidas; y el limbo
calieino mas corto, hasta 1 mm de largo.

Pardtipo. COLOMBIA. Cesar: 5 km W ol Manaure,
I0°22'N, 73°08'W, Gentry et al. 00094 (MO).

Simira liirsuta C. M. Taylor, sp. nov. TIPO: Co¬
lombia. Antioquia: Municipio de San Luis,
Canon del Rio Claro, sector sur-oriental,
05°53'N, 74°39'W, 350-450 m, 29 Apr. 1984,
A. Cogollo & R. Borja 1577 (holotipo, JAUM;
isotipo, MO-5057408). Figura 1C, D.

Haee species a congeneris pubeseentia hirsuta, foliis
hasi rolundatis truncatisve, limbo ealyeino 2.0— 2.2 mm
longo, corollae tubo ea. 4 mm longo ac lobulis ea. 3 mm
longis, filamentis hirtellis, stylo Iriehomalibus adscenden-

tilms serieeis induto atque fructu diametro ea. 3 cm dis-
tinguitur.

Arboles hasta 25 m de alto, floreciendo a 8 m
de altura; tallos hirsutos a glabrescentes. Hojas
opuestas; Idminas elfpticas a ligeramente oblanceo-
ladas, 7—18 cm de largo, 2.5— 6.5 cm de aneho, con
apice agudo a usualmente acuminado con acumen
hasta 12 mm de largo, con base breve a amplia-
mente redondeada a truncada, papiraceas, en la haz
glabras excepto hirsutulas en la costa y a veees
tambien las venas secundarias, en el enves espar-
cidamente hirsutas en la lamina y moderada a den-
samente hirsutas en costa y venas secundarias;
venas secundarias 12—13 pares, debil a claramente
uniendose en una vena submarginal, aparente-
mente sin domacios en las axilas, en la haz la ven-
acion toda plana o la costa y las venas secundarias
a veces ligeramente sulcadas, en el enves la costa
prommula a prominente, las venas secundarias
promfnulas y la venacion menor reticulada y prom-
rnula; peciolos 3—22 mm de largo, hirsutos; estipulas

570

Novon

interpeciolares, caducas, triangulares a lanceola-
das, en la superficie externa hirsutas, en la super-
fieie interna glahras, hasta 20 mm de largo, agudas
a aeuminadas. Inflorescencias terminales, subsesiles
(o tripartidas), 4—5 cm de largo, piramidales a an-
gostamente piramidales, 5—9 X 3-7 cm, con los
ejes, braeteas y pedicelos hirsutos, con braeteas an-
gostamente triangulares, 1.5—10 mm de largo; ped¬
icelos hasta 2 mm de largo; flares aparentemente
protandras, sesilcs y pediceladas en cfmulas o fas-
cfculos de 5—15; hipanto angostamente turbinado,
puberulo a corto-pildsulo, 1.5—2 mm de largo; lim¬
bo calicino externamente hirsutulo, internamente
glabro, cupuliforme o infundibuliforme, 2-2.2 mm
de largo incluyendo los ldbulos 0.3—1 mm de largo,
deltoides; corola infundibuliforme, amarilla palida,
exterior brevemente serfceo con tricomas ascenden-
tes, interior glabro excepto por un anillo piloso un
poco abajo del apiee del tubo, tubo ca. 4 mm de
largo, ca. 1 mm de diametro en la base y ca. 2.5
mm de diametro en el dpiee, ldbulos 5, angosta¬
mente triangulares, ca. 3 mm de largo, agudos; an-
teras ca. 2.5 mm de largo, exertas sobre filamentos
ca. 2.5 mm de largo c hirtelos; estigmas ca. 1.5 mm
de largo, exertos sobre un estilo ca. 4 mm de largo
y serfceo con tricomas ascendentes. Frutos globo-
sos, lenosos, lisos, ca. 3 cm de diametro; semillas
maduras no observadas.

Distribucidn , hdbitat y fenologia. En bosque
humedo con substrato de calice en el noroeste de
Colombia, a 325—520 m. Colectada con flores de
marzo a mayo y en diciembre, con frutos en marzo,
abril y julio a octubre.

Esta especie se distingue de otras especies de
Simira por su pubesceneia hirsuta, las hojas redon-
deadas a truneadas en la base, el limbo calicino 2-
2.2 mm de largo, la corola con el tubo ca. 4 mm
de largo y los 5 ldbulos ca. 3 mm de largo, los
filamentos hirtelos, el estilo serfceo y los frutos ca.
3 cm de diametro; se conoce solo de la region tfp-
iea. El epiteto se refiere a su pubesceneia distin-
tiva. Otras especies similares son Simira cordifolia
(Hooker f.) Steyermark, S. klugei (Standley) Stey-
ermark y S. mollis (Standley) Steyermark. Simira

cordifolia se distingue de S. hirsuta por los organos
vegetativos y ejes de la inflorescencia glabros a
puberulos, en cada caso menos pubescentes; el lim¬
bo calicino mas corto, hasta 1 mm de largo; los
filamentos puberulos, i.e., menos pubescentes; y el
estilo glabro, i.e., menos pubescente. Simira klugei
se distingue de S. hirsuta por la pubesceneia di-
ferente, hirtela y mas eorta; las hojas menores, 3-
8 X 2—6 cm, y menos pubescentes, glabras en el
enves; el limbo calicino mas corto, hasta 1 mm de
largo; los filamentos puberulos, i.e., menos pubes¬
centes; y el estilo glabro, i.e., menos pubescente.
Simira mollis se distingue de 5. hirsuta por las
flores sesiles en lugar de pediceladas; el limbo cal¬
icino mas largo, 6—8 mm de largo, y de forma di-
ferente, prolundamente lobulado; y la corola de pu-
bescencia diferente, densamente tomentulosa
externamente.

Pardtipos. COLOMBIA. Antioquia: Municipio de
San Luis, Parque Ecoldgico Canon del Rio Claro,
05°53'N, 74°39'\\. sector occidental, margen izquierda,
subida a la Dolina, Cogollo A- Borju 463 (JAUM, MO),
sector sur-oriental. margen derecha, Cogollo 659 (JAUM.
MO), sector central, margen izquierda. Cogollo & Borja
736 (JAUM. MO), sector occidental, margen izquierda,
Cogollo 1013 (JAUM, MO), sector sur-oriental. margen
izquierda, Cogollo 1466 (JAUM. MO), sector sur-occiden-
tal. Cogollo A- Borja 1616 (JAUM, MO), Cogollo 1702
(J AUM, MO), sector nor-occidental, margen izquierda, Co¬
gollo 1906 (JAUM, MO). Cogollo A Borja 2026 (JAUM,
MO); municipio de Puerto Triunfo, alrededores de la Cruta
“El Condor." 5°56'N, 74°50'W, Cogollo et at. i276
(JAUM. MO).

Agradecimientos. Agradezeo al personal y los
institutos de los herbarios HUA, JAUM y TULV el
acceso a las colecciones, hospedaje y facilitar via-
jar en Colombia; a R. E. Gereau (MO) la prepara-
cion de la diagnosis en latin; a A. Arbelaez (MO)
la revision del espanol; y en particular a los colegas
W. Devia (TULV) y A. Cogollo (JAUM) la colabor-
acion generosa que me han ofrecido dvrrante mis
estudios.

Literatura Citada

Steyermark. J. A. 1672 .Simira. In: It. M. Maguire & Col¬
laborators. Flora of the Guayana Highlands. Mem. New

York Rot. Card. 23: 299-309.

New Taxa and Combinations in the Ocotea helicterifolia
(Lauraceae) Species Group

Henk van der Werff

Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. In the course of preparing a treatment
of the Ocotea helicterifolia group for Flora
Mesoamericana, the existing treatment was found to
he outdated. A review of the group is here pre¬
sented and includes the description of the following
new species: 0. congregata van der Werff, 0. cor-
rugata van der Werff, 0. gordonii van der Werff,
and 0. patula van der Werff, as well as the follow¬
ing new combinations: 0. betazensis (Mez) van der
Werff, 0. bourgeauviana (Mez) van der Werff, 0.
purpurea (Mez) van der Werff, and 0. tonii (Lun-
dell) van der Werff. A key to the species of the
group is presented, new synonymy is given, and the
specimens studied are listed.

Among the species of Ocotea in Central America
is a group characterized by the presence of an erect
indument on the leaves (especially on the lower
surface) that is discernible to the touch, with dense¬
ly to moderately pubescent twigs, tepals that are
partially papillose (sometimes only along the mar¬
gin or near the tip), glabrous or somewhat papillose
anthers with the four cells arranged in two super¬
posed pairs and, at least in some species, relatively
well-developed staminodia. The term papillose is
useil here for a dense cover of very short, curly
hairs. In a few species the anthers each have a
small, sterile tip and the anther cells do not fill the
anther completely, as is the case in most species of
Ocotea. Although this group is easy to recognize,
its taxonomy is confused. The presence of well-de¬
veloped staminodia has resulted in the inclusion of
several species in Phoebe , and later those were
transferred to Cinnamomum. Other species have
been placed in Nectandra, sharing a common pap-
illosity of the tepals and anthers. Nectandra beli-
zensis (Lundell) Allen resembles in vegetative char¬
acters the O. helicterifolia group, but has typical
Nectandra stamens (short, broad, with the anther
cells arranged in a shallow arc, not in two super¬
posed pairs as in Ocotea ). Nectandra belizensis is
known from Belize, Costa Rica, and Panama. Roh-
wer (1991) gave a brief discussion of the group and
provided a list of taxa included in it. He also re¬
vised ihe species with the general flower structure

of O. helicterifolia , but which did not have the pu¬
bescent leaves and twigs. He noted that the 0. hel¬
icterifolia group is related to the Ocotea sinuata
group, which differs in having tongue-shaped an¬
thers each with a conspicuous sterile tip.

In the course of writing a treatment of Ocotea for
Flora Mesoamericana, I found several undescribed
species in the O. helicterifolia group and several
other species that needed to be transferred to Oco¬
tea. A few species of this group occurred outside
the area covered by Flora Mesoamericana. In ad¬
dition to the novelties, I decided it might be useful
to publish a key to all species I recognize as be¬
longing to the 0. helicterifolia group, as well as new
synonyms. In the most recent treatment of Central
American Lauraceae (Allen, 1945), most species of
the O. helicterifolia group were included in Phoebe,
and these species were mostly separated based on
leal size and leaf shape. I found these characters
less important than Allen did, and rely more on
such characters as inflorescence type (racemose or
paniculate-cymose), flower characters (flowers gla¬
brous or pubescent; inner surface of tepals glabrous
or pubescent; receptacles glabrous or pubescent in¬
side), and leaf position (alternate or clustered). Use
of these characters leads to better-defined species,
although some of the species appear quite variable
and may be further divided at some later point.
Specifically, the Costa Rican specimens of O. hel¬
icterifolia seem slightly different and occur at lower
altitudes than specimens collected north ol Costa
Rica. Likewise, specimens of 0. purpurea from
Panama seem different (fewer lateral veins, for ex¬
ample) than those from Honduras northward. How¬
ever, splitting these species can only be done using
vegetative characters (leaf shape and size), and be¬
cause I regard these characters as weak and not
reliable, I am reluctant to further divide these rath¬
er variable species.

In several species the upper rim of the recepta¬
cle carries a ring of hairs. These hairs are easily
visible and may suggest the receptacle itself is pu¬
bescent or the tepals are pubescent on the inner
surface. It is necessary to split a receptacle open
in order to ascertain whether the receptacle is pu-

Novon 9: 571-583. 1999.

572

Novon

hescent inside: a view from above is not sufficient.
I only describe tepals as pubescent on the inner
surface if the hairs extend at least halfway the

length of the tepals; a tuft of hairs visible at or near
the base of the tepals is not enough to score the
inner surface of the tepals as pubescent.

Kn TO Sl'K.CIKS OK Till OCOUA HMJCTKRIh'OIJ t GUOl I’ IN Cl MIHI AmKUICA

la.

Inflorescences racemose or rarely with one or two lateral cymes.

2a. Leaves obovate; tufts or lines of white hairs common along major veins on lower leal surface .

. 15. 0. sp. nov.

2h. Leaves elliptic; axillary tufts of hairs lacking or only a few on each leaf.

5a. Leaves bu Hate-rugose, the major veins strongly impressed on the upper leaf surface .

. 4. 0. corrugata

.4b, Upper leaf surface smooth or nearly so, the major veins immersed, not strongly impressed.

4a. Outer surface of tepals glabrous; anthers sessile . 5. 0. gordonii

4b. Outer surface of tepals at least sparsely pubescent; stamens with filaments 1/5 or more of
the length of the anthers.

5a. Hairs on lower leaf surface ascending, covering most of the lamina; indument gray . . .

. 8. (). mollicella

5b. Hairs on lower leaf surface erect, most of the lamina visible; indument brown or

ferruginous . II. 0. purpurea

Inflorescences paniculate.

6a. Leaves clustered.

7a. Outer surface of tepals glabrous . 2. 0. bourgeauviana

7b. Outer surface of the tepals pubescent.

8a. Petioles to 6 mm long; leaves 15—25 X 5—6 cm; twigs with yellowish brown indument ....

. 12. 0. lonii

6b.

8b. Petioles at least 10 mm long; leaves 9—17 X 4—7 cm; twigs with brown indument

. 3. 0. congregata

Leaves alternate, evenly distributed along the twigs.

9a. Receptacle pubescent inside.

10a. Leaves elliptic, to 15 cm long . 1.0. betazensis

l()b. Leaves obovate, 18— 40 cm long.

I la. Indument completely covering young twigs and inflorescences . 14. O. valerioides

I lb. Surface of twigs and inflorescences visible between the indument . 1.0. lentil

9b. Receptacle glabrous inside.

12a. Outer surface of tepals pubescent.

15a. Leaves broadly elliptic; tufts of white hairs lacking in axils of lateral veins or along

the major veins . 9. O. patula

15b. Leaves obovate; tufts of white hairs present in axils of lateral veins and along the

major veins . 15. 0. sp. nov.

12b. Outer surface of tepals glabrous.

14a. Inner surface of tepals pubescent; filaments evident, ea. 1/5 the length of the anthers.

15a. Leaves to 10 cm long, the tips obtuse, acute or shortly acuminate .

. 10. O. praelermissa

15b. Leaves 15—20 cm long, acuminate or gradually narrowed into a slender tip .

. 2. O. bourgeauviana

14b. Inner surface of lepals glabrous; anthers sessile.

16a. Surface of young twigs completely covered by the indument; eupule cup-shaped

. 15. O. raleriana

16b. Surface of young twigs partially visible between the indument; eupule only shal¬
lowly bowl-shaped or platelike . 6. 0. hei licteri folia

1. Ocolea belazensis (Mez) van der Werff, comb,
nov. Basionytn: Phoebe betazensis Mez, Jahrb.
Koenigl. Bot. Cart. Berlin 5: 192. 1889. SYN-
TYPES: Mexico. Oaxaca: Liebmann 2(C),
3(C), 22(C), 23(C), Galeotti 2335 not seen,
Juergensen 575 not seen. Figure 1.

Oreodaphne mexicana Meisner var. diminuta Meisner,
Prodr. 15(1); 118. 1861. Ocotea mexicana (Meisner)
Hemsley var. diminuta (Meisner) Hemsley, Biol.
Centr. Amer., Bot. 4; 75. 1882. TV PL: Juergensen
575 not seen.

Ocotea betazensis (as Phoebe betazensis) has long
been included in Ocolea helicterifolia, from which it
differs in the elliptic to broadly elliptic (not obovate)
leaves, the stamens with well-developed filaments, the
dense tomentellous-tomentose indument on the young
twigs that completely covers the surface (not hirsute
with the surface partially visible), and the generally
longer petioles than seen in 0. helicterifolia. Ocotea
betazensis is currently only known from cloud forests
between 1900 and 2600 m altitude in Oaxaca, but
can be expected in neighboring Chiapas.

Volume 9, Number 4
1999

van der Werff

Ocotea helicterifolia Species Group

573

Figures 1-4 (clockwise from top left). — 1. Ocotea betazensis: Cedilla 1197. — 2. Ocotea bourgeauviana : Breedlove
50583. — 3. Ocotea congregata: Mendez Ton 9594. — 4. Ocotea corrugata : Wendt et at. 6765 (isotype).

574

Novon

2. 0< •otea bourgeam iana (Mez) van der Werff,

comb. nov. Basionym: Phoebe bourgeauviana
Mez, Jahrb. Koenigl. Bot. Gart. Berlin 5: 194.
1889. Cinnamomum bourgeauvianum (Mez)
Kostermans, Reinwardtia 6: 20. 1961. TYPE:
Mexico. Veracruz: Bourgeau 2234 (isosyntype,
MO). Figure 2.

Nectandra longieuspis I, umlcll. Wrightia 5: 34. 1074. Syn.
nov. TYPE: Guatemala. Izabal, Contreras 1 1 186 (iso¬
type, MO).

Phoebe chinantecorum R. E. Schultes, Bot. Mus. I ,eafl. 9:
170. 1941. Syn. nov. Cinnamomum chinantecorum
(R. E. Schultes) Kostermans, Reinwardtia 6: 20.
1961. TYPE: Mexico. Oaxaca: Schultes & Reko 827
(holotype, GH).

Ocotea bourgeauviana can be recognized by the
combination of paniculate-cymose inflorescences,
glabrous flowers, pubescent inner surface of the te-
pals, and the mostly clustered leaves. The indu-
inent on the twigs is hirsute, with the surface of the
twigs partially visible between the indument. Leaf
size ranges from 13 to 20 cm long. It is known from
Mexico (Veracruz, Chiapas), Guatemala, and Hon¬
duras at altitudes ranging from 100 to 1200 m, but
is infrequently collected. Provisionally placed here
are some collections from Veracruz (Mexico) with a
glabrous inner surface of the tepals and slightly
smaller flowers (4—5 mm diam. vs. 5—6 mm in 0.
bourgeauviana). Because these specimens differ in
only one solid character (the glabrous inner surface
of the tepals) from O. bourgeauviana , I am reluctant
to recognize them as a distinct species and list
them in the specimens studied as 0. aff. bourgeau¬
viana.

3. Ocotea congregata van der Werff, sp. nov.

TYPE: Mexico. Chiapas: Mpio. Oxchuc, Cas-
cada de Coralito, Shilom Ton 8930 (holotype,
MO). Figure 3.

Ocoteae tonii similis, sed loliis brevioribus, latioribus,
petiolis plus quam 10 mm longis, indumento castaneo dif-
ferl.

Small to medium-sized trees, to 15 m tall. Twigs
terete, solid, densely brown-tomentose or tomentel-
lous, the hairs erect and twisted, covering the sur¬
face of the young twigs completely, terminal buds
densely brown-tomentose. Leaves 9—17 X 4—7 cm,
elliptic to broadly elliptic, firmly chartaceous, clus¬
tered, the base rounded or obtuse, rarely acute, the
apex obtuse to acute, midrib and lateral veins im¬
pressed, tertiary venation weakly impressed on the
upper surface, midrib and major veins prominently
raised, smaller veins raised on the lower surface;
upper surface moderately pubescent when young.

the hairs erect, soon becoming glabrous, the pu¬
bescence denser and persisting along ihe major
veins, lower surface moderately densely pubescent,
the hairs erect and discernible to the touch, the
surface readily visible between the hairs, the in¬
dument denser and tomentose along the midrib and
lateral veins; domatia lacking; petioles 10—25 mm,
round, with a similar indument as the twigs. Inflo¬
rescences 5—12 cm, paniculate-cymose, the flowers
in compact clusters, densely hirsute-tomentose,
mostly in the axils of bracts, infrequently in the
axils of leaves. Flowers 7-9 mm diam., white, per¬
fect, the receptacle densely pubescent outside;
pedicels ca. 3 mm long. Tepals 3 mm long, elliptic,
the outside moderately to sparsely pubescent, the
inside pubescent near the base, otherwise glabrous,
spreading at anthesis, outer 6 stamens 1.6 mm, ses¬
sile or nearly so, with a few hairs near the base,
otherwise glabrous, the cells arranged in 2 pairs,
introrse, at the tip with a narrow, sterile border,
inner 3 stamens 1.7 mm, the filament 0.5 mm, with
a few hairs, the cells in 2 pairs, extrorse-lateral,
glands present at the base of the filaments, stami-
nodia 3, minute, stipitiform, hidden between the
hairs on the top of the receptacle; pistil 1.5 mm,
glabrous, the style 0.4 mm, receptacle cup-shaped,
appressed pubescent or glabrous inside. Fruit el¬
lipsoid, 2 X 1.3 cm, the cupule deeply cup-shaped
when young, bowl-shaped at maturity, 1 cm diam.,
with a single margin, the tepals deciduous. Flowers
March, April, November; fruits August, October.

Ocotea congregata is named after the erect sta¬
mens grouped in a rather tight cluster. The new
taxon is currently only known from the Mexican
state of Chiapas and occurs between 800 and 1370
m altitude. It can be recognized by the slightly bul-
late, loosely clustered leaves, the pubescent flowers
arranged in paniculate-cymose inflorescences, the
brown to dark brown indument on the twigs, and
the rather long (10 mm or more) petioles. Its closest
relative is Ocotea tonii , which differs in its narrow¬
er, longer leaves, shorter petioles (to 6 mm long),
and the yellowish brown indument on the young
twigs.

Paratypes. MEXICO. Chiapas: M pio. Ocosingo,
Breedlove 15672 (CAS), 88017 (CAS. MO), 52580 (CAS.
MO), Martinez S. I 7084 (MO), Quintanilla 49 (MO); Mpio.
San Cristobal de las Casas, Mendez Ton 9594 (CAS, MO):
Mpio. Oxchuc, Shilom Ton 8508 (MO).

4. Ocotea eorrugata van der Werff, sp. nov.
TYPE: Mexico. Oaxaca: Mpio. Sta. Maria Chi-
malapa. Sierra de Tres Picos, alt. 1150—1250
m, T. Wendt, Hernandez G., Tenorio, Torres, Sa¬
lazar, Soto & Rocha 6765 (holotype, MEXU;
isotype, MO). Figure 4.

Volume 9, Number 4
1999

van der Werff

Ocotea helicterifolia Species Group

575

Ad gregem Ocoteae helicterifoliae pertinens, sed foliis
corrugatis diversa est.

Small tree, to 4 m. Twigs terete, densely brown-
tomentellous, the surface of the young twigs not vis¬
ible, solid; terminal buds densely tomentellous.
Leaves 7-14 X 2.5-7 cm, elliptic or broadly ellip¬
tic, firmly chartaceous, alternate, the base variable,
from acute or obtuse to rounded, the apex acute or
obtuse, the upper surface sparsely pubescent with
erect or ascending hairs when young, soon becom¬
ing glabrous, the lower surface sparsely pubescent,
the hairs erect and discernible to the touch, the
indument denser and tomentellous along the major
veins, midrib, lateral veins, and tertiary venation
impressed on the upper surface, strongly raised on
the lower surface, the leaves bullate to rugose, lat¬
eral veins 5 to 7, domatia absent, petioles 7-15
mm, flat above, with a similar indument as t he
twigs. Inflorescences 3—5 cm, racemose, less than
10-flowered, densely hirsute, the flowers grouped
near the tip of the inflorescence. Flowers 7-8 mm
diam., cream-colored, perfect, receptacle densely
pubescent on the outside; pedicels 2—3 mm long.
Tepals 3 mm long, elliptic, on both surfaces pu¬
bescent near the base, otherwise glabrous, spread¬
ing at anthesis; outer 6 stamens 1.5 mm, the fila¬
ments ea. 0.4 mm, with some hairs near the base,
otherwise glabrous, the cells introrse, arranged in
2 pairs, a sterile tip lacking, inner 3 stamens 1.5
mm long, the filament 0.4 mm long, with 2 glands
near the base, the cells in 2 rows, the upper row
lateral, the lower one lateral extrorse, staminodia 3,
clavate, with a few hairs, difficult to see among the
hairs on the rim of the receptacle, pistil glabrous,
ca. 1 mm long, the style very short, receptacle cup¬
shaped, appressed pubescent inside. Fruit and cu-
pule unknown. Flowers April.

Although Ocotea corrugata is only known from
the tvpe collection, it is here described largely be¬
cause of its very distinct, bullate to rugose leaves;
additional characters are the racemose inflores¬
cences and the pubescent inside of the receptacle.
It belongs to the Ocotea helicterifolia group, where
it does not seem to have a close relative. Racemose
inflorescences are uncommon in this complex, but
I do not think that O. corrugata is closely related
to the other species with racemose inflorescences
such as 0. purpurea and 0. gordonii. These species
differ greatly from each other in other characters,
and racemose inflorescences have probably arisen
several times in this complex by reduction of the
more common paniculate-cymose inflorescence.
This species was described as abundant at the type

locality. However, the remoteness of this locality
made it difficult to bring back more material.

5. Ocotea gordonii van der Werff, sp. nov. TYPE:

Panama. Chiriquf: vicinity of Fortuna Dam, G.

McPherson 10421 (holotype, MO). Figure 5.

Inter speciebus Ocoteae helicterifoliae gregis inflores-
centia racemosa, receptaculo intus pubescente et pedicel-
lis longis recedit.

Small trees, to 10 m tall. Twigs terete, solid,
densely tomentose, the surface not or scarcely vis¬
ible between the hairs; terminal buds densely to¬
mentose. Leaves 12-22 X 5—8 cm, narrowly to
broadly elliptic, chartaceous, alternate and evenly
distributed along the twigs, the margin flat or some¬
times folded downward, the base acute or obtuse to
almost rounded, the tip acuminate with an acumen
to 2 cm long or acute, the upper surface with some
erect, curled hairs when young, but this indument
rapidly wearing off, the hairs denser and becoming
tomentose along the major veins, the lower surface
copiously pubescent, the hairs erect and soft to the
touch, denser and tomentose along the midrib and
lateral veins, midrib and lateral veins somewhat im¬
pressed, tertiary venation slightly raised on the up¬
per surface, midrib, lateral veins, and tertiary ve¬
nation raised or prominently raised on the lower
surface; lateral veins 6 to 8; domatia absent; peti¬
oles 8-14 mm, flattened on the upper side, with a
similar indument as the twigs. Inflorescences 8-16
cm, racemose, moderately densely to sparsely pu¬
bescent, the hairs erect or spreading, to 0.6 mm
long; in axils of bracts or, less frequently, of normal
leaves. Flowers 8—10 mm diam., white, glabrous,
fragrant, perfect, pedicels ca. 1 cm long. Tepals 6,
4—4.5 mm long, elliptic, spreading at anthesis, the
outer 3 with a basal triangular papillose patch, oth¬
erwise glabrous, the inner 3 uniformly papillose on
the inner surface; stamens 9, 4-celled, the outer 6
weakly papillose, the cells arranged in 2 rows,
opening introrse-lateral, the anthers sessile or near¬
ly so, sterile tip short, 0. 2-0.3 mm, inner 3 stamens
1.5 mm long, the anther sessile, the cells arranged
in 2 rows, laterally extrorse, the anthers weakly pa¬
pillose and with a few hairs near the base, glands
present at the base of the inner 3 stamens, stami¬
nodia not seen, receptacle cup-shaped, pubescent
inside. Fruits and cupules not known. Flowers Feb-
ruary-April.

Ocotea gordonii is part of the Ocotea helicteri¬
folia complex because of the papillose inner sur¬
face of the tepals, the relatively large flowers, the
long spreading indument of stems and leaves, and
the spreading tepals. Within this group it stands

576

Novon

Figures 5-8 (clockwise from lop left). — 5. Ocolea gordonii: McPherson 10421 (holotype). — (>. Ocolea heliclerifolia:
Linden 1641 (syntype). — 7. Ocotea heliclerifolia-. Mendez Ton 5902. — 8. Ocolea heliclerifolia (Costa Hiea): Herrera
486 6.

Volume 9, Number 4
1999

van der Werff

Ocotea helicterifolia Species Group

577

apart due to its racemose inflorescences, the long
pedicels, and pubescent receptacle. Only one other
species in this group with racemose inflorescences
is known from Panama, and this species, 0. pur¬
purea, has pubescent flowers ca. 5 mm in diameter
and leaves less than 11 cm long. Vegetatively, there
is a stronger resemblance between the new species
and O. valeriana and O. helicterifolia , both of which
occur in Costa Rica. These two species have pa-
niculate-cymose inflorescences and glabrous recep¬
tacles. Two other species from Costa Rica and Pan¬
ama, 0. lentii and 0. valerioides, have pubescent
receptacles like 0. gordonii, but diller in their pa-
niculate-cymose inflorescences and obovate leaves.

This species is dedicated to Gordon McPherson,
whose excellent collections have contributed much
to our knowledge of the Panamanian flora.

Paratypes. PANAMA. Veraguas: mountains above
Sta. Fe, Hummel & Kress 8534 (MO). Chiriqui: vicinity
of Fortuna Dam. McPherson 10573 (MO).

6. Ocotea helicterifolia (Meisner) Hemsley, Biol.
Centr. Amer., Bot. 3: 73. 1882. Oreodaphne
helicterifolia Meisner, Prodr. 15(1): 123. 1864.
Phoebe helicterifolia (Meisner) Mez, Jahrb. Ko-
enigl. Bot. Gart. Berlin 5: 193. 1889. Cinna-
tnomum helicterifolium (Meisner) Kostermans,
Reinwardtia 6: 21. 1961. TYPE: Mexico.
Chiapas: Linden 1641 (syntype, K). Figures 6,
7, 8.

Nec tundra corzoana Lundell, Arightia 4: 102. 1969. Svn.
nov. Phoebe corzoana (Lundell) Lundell. Wrightia 5:
342. 1977. Cinnarnomum corzoanum (Lundell) kos-
tentians. Reinwardtia 10: 422. 1988. TV I’F: Mexico.
Chiapas: Shilom Ton 3560 (holotype, I.L).

Ocotea tenejapensis Lundell. Wrightia 4: 108. 1969. Syn.
nov. TYPE: Shilom Ton 779 (holotype, LL).

Phoebe nectandroides Mez, Jahrb. Koenigl. Bot. Gart. Ber¬
lin 5: 194. 1889. SYNTYPES: Botteri 1018 not seen.
Galeotti 7004 (BR). Juergensen 937 not seen, Lieb-
mann 4 not seen. Oersted 21 not seen.

Phoebe obtusata Lundell. Contr. Univ. Mich. Herb. 6: 21.
1941. Syn. nov. Cinnarnomum obtusatum (Lundell)
Kostermans, Reinwardtia 6: 22. 1961. TYPE: Mex¬
ico. Chiapas: Matuda 1887 (isotype, CAS).

Ocotea helicterifolia is characterized by its alter¬
nate leaves, glabrous flowers and inside of the re¬
ceptacle, as well as the hirsute indument of the
twigs, with at least a small part of the surface vis¬
ible between the hairs. In Costa Rica it can be
confused with O. valeriana, which has a denser in¬
dument on twigs and inflorescences. Ocotea helic¬
terifolia occurs from southern Mexico to Panama at
altitudes of 1000-1900 m; in Costa Rica it occurs

at lower elevations (50—600 m) and has more ob¬
ovate leaves. These collections from Costa Rica
possibly represent an undescribed taxon. Provision¬
ally included in O. helicterifolia is Ocotea teneja¬
pensis, known only from the type collection. This
collection differs from typical 0. helicterifolia in
that the outer 6 stamens have filaments about 1/3
the length of the anthers, the lower leaf surface is
very sparsely pubescent, and small axillary tufts of
hairs are often present on the lower leaf surface. I
have not found other collections with these char¬
acters, and therefore include 0. tenejapensis as a
somewhat aberrant form in O. helicterifolia.

Because O. helicterifolia has a wide range of dis¬
tribution and shows some variation in vegetative
characters, I include three figures, including one of
the type, of this species.

7. Ocotea lentii W. C. Burger, Fieldiana Bot.. n.s.

23: 86. 1990. TYPE: Costa Rica. Cartago: Lent
794 (isotype, MO). Figure 9.

Ocotea lentii is best recognized by its large (18 —
40 cm long), obovate, alternate leaves and the hir¬
sute indument on twigs and inflorescences. Part of
the surface remains visible between the hairs. Oco¬
tea lentii is an infrequently collected species known
only from the provinces Cartago and Guanacaste in
Costa Rica. Most collections have been made be¬
tween 700 and 1400 m altitude.

8. Ocotea molliceUa (Blake) van der Werff, Fiel¬

diana Bot., n.s. 23: 88. 1990. Phoebe molli-
cella Blake, Contr. Gray Herb. 52: 64. 1917.
Cinnarnomum mollicellum (Blake) Koster¬
mans, Reinwardtia 6: 22. 1961. TYPE: Costa
Rica. Tonduz 11676 (isotype, G). Figure 10.

Ocotea mollicella is characterized by its race¬
mose inflorescences, ascending (not erect) hairs on
the lower leaf surface, and the gray color of the
indument. Its leaves are narrowly elliptic to ellip¬
tic-lanceolate and generally less than 8 cm long.
Ocotea mollicella is a rarely collected species only
known from cloud forests in Costa Rica and occur¬
ring between 1400 and 2300 m altitude.

9. Ocotea patula van der Werff, sp. nov. TYPE:

Costa Rica. Puntarenas: Canton de Osa, Fila
Costena, Aguilar et al. 2715 (holotype, MO).
Figure 11.

Ocoteae Valerianae similis, sed floribus pubescentibus
pedicellis tomentellis brevioribusque recedit.

Small trees, to 8 m. Twigs terete or slightly
ridged, solid, densely yellowish brown tomentose

578

Novon

Figures 9—12 (clockwise from top left). — 9. Ocotea lentii : llerrera 8938. — 10. Ocotea mollicella : Tonduz 11678
(isotype). — 1 1. Ocotea patula.'. Aguilar 2715 (holotype). —12. Ocotea praetermissa: Burger et al. 12065 (holotype).

Volume 9, Number 4
1999

van der Werff

Ocotea helicterifolia Species Group

579

when young, the surface completely covered by the
indument, the indument becoming whitish with age,
terminal buds densely yellowish brown tomentose.
Leaves 12-24 X 9—14 cm, broadly elliptic, char-
taceous, alternate, the base obtuse to rounded, the
apex obtuse or shortly acuminate, the upper surface
with some erect hairs when young, soon glabres-
cent, the midrib and lateral veins tomentellous, the
lower surface sparsely to moderately pubescent, the
hairs erect, the indument denser and tomentellous
along the midrib and lateral veins, midrib, lateral
veins, and tertiary venation immersed on the upper
surface, raised to prominently raised on the lower
surface, domatia lacking, lateral veins 7—8; petioles
17-26 mm long, with a similar indument as the
twigs, shallowly canaliculate on the upper surface.
Inflorescences 10-16 cm long, paniculate-cymose,
densely yellowish brown tomentellous, in the axils
of eataphylls, rarely in the axils of leaves. Flowers
ca. 7 mm diam., white, perfect, sparsely to mod¬
erately pubescent; receptacle densely pubescent
outside; pedicels ca. 2 mm long, tomentellous. Te-
pals ca. 2.5 mm long, elliptic, spreading or some¬
what reflexed at anthesis, the inner surface mod¬
erately to sparsely papillose; stamens 9, 4-celled,
the outer 6 ca. 1.2 mm long, weakly papillose, the
cells arranged in 2 pairs, introrse, a short (0.2 mm)
sterile tip present, the anthers sessile or nearly so,
inner 3 stamens as long as the outer 6, also weakly
papillose, sessile or nearly so, the cells in 2 pairs,
extrorse, glands present at the base, staminodia not
seen. Pistil ca. 1 .5 mm long, the style as long as
the ovary, glabrous, receptacle cup-shaped, gla¬
brous inside. Fruits not known. Flowers December.

Ocotea patula is named alter the spreading te-
pals at anthesis. It is known only from two collec¬
tions made on the same day in the same general
area, but probably from different trees. One of the
collectors indicated it was growing on a limestone
substrate. It is part of the O. helicterifolia group
and resembles O. valeriana. However, it differs
from this species in its densely tomentellous inflo¬
rescences (with the surface entirely or almost en¬
tirely covered by the indument), by its shorter (2—
3 mm vs. 6—8 mm) and tomentellous pedicels, as
well as its pubescent flowers.

Paratype. COSTA It 1C A. Puntarenas: Fila Costena,
Hammel et al. 19217 (MO).

10. Ocotea praetermissa van der Werff, Novon
6: 482. 1996. TYPE: Costa Rica. Cartago:

Burger et al. 12065 (holotype, MO). Figure 12.

Ocotea praetermissa can be recognized by its al¬

ternate, rather small leaves (rarely exceeding 10
cm), paniculate inflorescences, glabrous flowers
with a pubescent inner surface of the tepals, and a
glabrous receptacle. It is a species of montane for¬
ests from 2000—3200 m elevation, mostly found in
Costa Rica, but with a few collections from Panama.
This species has been misidentified as Ocotea (or
Phoebe) pittieri and was included in Rurger and van
der Werff (1990) under that name.

1 1 . Ocotea purpurea (Mez) van der Werff, comb,
nov. Basionym: Phoebe purpurea Mez, Jahrb.
Koenigl. Bot. Cart. Berlin 5: 196. 1889.
TYPE: Guatemala. Alta Verapaz, von Tuer-
kheim 371 not seen (B not seen, type photo¬
graph, MO). Figure 13.

Nectandra capituliforma Lundell, Wrightia 5: 33. 1974.
Syn. nov. TYPE: Guatemala. Contreras 11235 (iso¬
type, MO).

Ocotea purpurea is characterized by its racemose
inflorescences (one or two cymes are rarely present
at the base of the inflorescence), pubescent flowers,
and rather small (to 11 cm), (narrowly) elliptic, flat
leaves. The leaf and inflorescence characters are
clearly visible on the type photograph. It is among
the few species in the O. helicterifolia complex with
racemose inflorescences and pubescent flowers; the
other ones are an undescribed species from Mexico
with large, obovate leaves and conspicuous domatia
along the major veins, 0. corrugata with bullate or
rugose leaves, and O. mollicella with ascending,
gray pubescence on the lower leaf surface (brown
and erect in O. purpurea ). Ocotea purpurea occurs
in southern Mexico, Guatemala, Honduras, and
Panama at altitudes ranging from 1400 to 2600 m.
The name Phoebe bourgeauviana has frequently
been applied to this species, but the type of P. bour¬
geauviana has glabrous flowers and a paniculate-
cymose inflorescence.

12. Ocotea tonii (Lundell) van der Werff, comb,
nov. Basionym: Nectandra tonii Lundell,
Wrightia 4: 106. 1969. TYPE: Mexico. Chia¬
pas: Municipio de Tenejapa, A. Shilom Ton
2014 (isotype, NY). Figure 14.

Ocotea tonii is characterized by its narrowly el¬
liptic, clustered leaves with short petioles, yellow¬
ish brown tomentum on the young twigs, and pa¬
niculate-cymose inflorescences with pubescent
flowers. It is only known from Chiapas at altitudes
between 800 and 1500 m.

13. Ocotea valeriana (Standley) W. C. Burger,
Fieldiana Bot., n.s. 23: 96. 1990. Phoebe val-

580

Novon

figures 13—16 (clockwise from top left). — 13. Ocotea purpurea: Breedlove 9467. — 14. Ocotea tonii: Mendez Ton
5264. — 15. Ocotea valeriana : llerrera el ai. 475. — 16. Oeotea valerioides: Grayum .3519.

Volume 9, Number 4
1999

van der Werff

Ocotea helicterifolia Species Group

581

erinnn Standley, Publ. Field Mus. Nal. Hist.,
Bot. Ser. 18(2): 460. 1937. Cinnamomum val-
erianum (Standley) Kostermans, Reinwardtia
6: 24. 1961. TYPE: Costa Rica. El Copey, Ton-
duz 11746 (holotype, F not seen). Figure 15.

Nectandra austinii Allen. J. Arnold Arbor. 26: 374. 1945.
Syn. nov. TV PE: Costa Rica. Austin Smith 1*2226
(holotype, A not seen).

Phoebe smithii Allen, J. Arnold Arbor. 26: 317. 1945.
Cinnamomum smithianum Kostermans, Reinwardtia
6: 23. 1961. TYPE: Costa Rica. Austin Smith P.C.
367 (holotype, F not seen).

Ocotea valeriana resembles 0. helicterifolia in its
paniculate-cymose inflorescences, alternate leaves,
glabrous flowers, and glabrous inside of the recep¬
tacles. They are best separated by the type of in-
dument on the young twigs. In O. valeriana this
indument consists of short, matted hairs and longer,
erect hairs; usually the short, matted hairs cover
the surface of the twigs completely. The indument
of O. helicterifolia consists oidy of long, erect hairs,
which may be densely distributed, but do not cover
the surface of the twigs completely. The cupules of
0. valeriana are also deeper than those of 0. hel¬
icterifolia. Ocotea valeriana is not rare in Costa
Rica and is known from a very few collections in
Panama. It occurs from 800 to 2200 m. I accept
the concept of 0. valeriana used in Burger and van
der Werff (1990), where W. C. Burger, who had ac¬
cess to the holotype, transferred Phoebe valeriana
to Ocotea. Although I have not seen the types of
Nectandra austinii and Phoebe smithii, 1 place both
species in synonymy of O. valeriana based on the
paratypes of these species I have seen (for N. aus¬
tinii: Austin Smith P2114 (GH); for P. smithii : Aus¬
tin Smith HS23 (MO)).

14. Ocotea valerioides W. C. Burger, Fieldiana
Bot., n.s. 23: 97. 1990. TYPE: Costa Rica.
Hartshorn 1530 (isotype, MO). Figure 16.

This species is similar to O. lentii in its large,
obovate leaves, paniculate-cymose inflorescences,
and pubescent inside of the receptacles. Ocotea
valerioides differs in its denser indument, complete¬
ly covering the surface, on the twigs and inflores¬
cences. It is known from a few collections in Costa
Rica and Panama, where it occurs in lowland rain¬
forest between 50 and 500 m altitude.

15. Ocotea sp. nov.

This species can be readily recognized by its ob¬
ovate leaves with conspicuous tufts of whitish hairs
in the axils of the lateral veins and along the major

veins. Its inflorescences are racemose or have sec¬
ondary axes terminating in a cyme. Flowers are pu¬
bescent. A description of this species is being pre¬
pared by F. Lorea-Hernandez. The new species is
only known from the state of Guerrero, Mexico.

Imperfectly Known Species

Ocotea macrophylla Kunth, Nov. Gen. Sp. 2:
131. 1818. Nectandra macrophylla (Kunth)
Nees, Systema Laurinarum 292. 1836. TYPE:
Colombia. Quindiu, Bonpland s.n. (isotvpe, P).

The type of 0. macrophylla is a fruiting speci¬
men; thus, flowering specimens can only be asso¬
ciated with this name based on vegetative similar¬
ities. Rohwer (1991) recognized several recent
collections from Colombia as 0. macrophylla and
accepted 0. macrophylla as the only species in the
0. helicterifolia group occurring south of Panama.
However, recent collections show that probably
more than one species is present in northern South
America; lack of sufficient flowering collections
makes it impossible to accurately describe the taxa
involved. All collections have alternate leaves and
paniculate-cymose inflorescences.

Provisionally I distinguish the following entities:

1. Ocotea macrophylla sensu typi. The type col¬
lection has a rather lax, hirsute indument on the
twigs, which does not completely cover the surface.
The hairs on the lower leaf surface are also rather
long and straight. The inflorescence or infructes-
cence is sparsely hirsute. In addition to the type, I
place here Sanchez 1371 (MO) from Antioquia, Co¬
lombia, and Ruiz Teran 1558 (MO). Breteler 4928
(MO), both from Merida, Venezuela. The Ruiz Teran
specimen is a fruiting one, while the Sanchez and
Breteler collections have flowers. These flowers are
glabrous, even on the inside of the tepals. Anthers
of the outer six stamens are sessile or nearly so.

2. Sanchez 1891 (MO) from Antioquia, Colombia.
Indument on the twigs is densely tomentellous/to-
mentose, covering the surface completely, as is the
indument on the inflorescences. The flowers are pu¬
bescent.

3. A group of several specimens characterized
by a dense, tomentellous indument on the twigs,
covering the surface completely. Inflorescences are
sparsely pubescent, and the flowers are glabrous or
nearly so. This group occurs in Colombia (Boyaca:
Lawrance 21 7 & 281 (MO); Antioquia: Fernandez
7 (MO) and Giraldo 287 (MO); Valle: Croat 70677
(MO) and Albert 2451 (MO)) and Ecuador (Carchi:
Palacios 12794 (MO), and van der Werff 10772
(MO)).

582

Novon

OCOTKA SPECIES, PoSSIBIA UNDESCRIBED

A few collections from Puebla (Mexico) resemble
Ocotea helicterifolia vegetatively, but differ in hav¬
ing rather small flowers (only 5—6 mm diam.) and
in having stamens with distinct filaments (filaments
1/3 as long as the anthers). 1 have not seen fruiting
material of this taxon. These collections (Rzedowski
3186.1, Tenorio 8686, Ventura 783, and Campos
225, all in MO) may well represent an undescribed
species, but 1 prefer to wait until more collections
are at hand before describing it.

Acknowledgments. I thank the curators of CAS,
F. GH, K. and TEX for the loan of specimens, B.
Stannard (K) for his efforts in locating the type of
(). helicterifolia, and T. Wendt (TEX) for duplicates
of his excellent Lauraceae collections.

Literature Cited

Allen, C. k. 1945. Studies in the Lauraceae VI. A prelim¬
inary survey of the Mexican and Central American spe¬
cies. J. Arnold Arbor. 2b: 280—434.

Burger, C. Vf. & H. van der Werff. 1990. Flora Costari-
censis, Lauraceae. Fieldiana Bot.. n.s. 23: 1 — 129.
Rohwer, J. G. 1991. Borderline cases between Ocotea.
Net tundra, and Phoebe (lauraceae): The marginal spe¬
cies of the 0. helicterifolia group, including the 0. hey-
deana group. Bot. Jahrb. Syst. 1 12(3): 365-397.

List ok Sckciks Rkcognizkd

1. Ocotea betazensis (Mez) van der Werff

2. Ocotea bourgeauviana (Mez) van der Werff

3. Ocotea congregata van der Werff

4. Ocotea corrugata van der Werff

5. Ocotea gordonii van der Werff

6. Ocotea helicterifolia (Meisner) Hemsley

7. Ocotea lentii W'. C. Burger

8. Ocotea mollicella (Blake) van der Werff

9. Ocotea pat ala van der Werff

10. Ocotea praelermissa van der Werff
1 I. Ocotea purpurea (Mez) van der Werff

12. Ocotea tonii (Lundell) van der Werff

13. Ocotea taleriana (Standley) W. C. Burger

14. Ocotea valerioides W. C. Burger

Specimens Studied

If collected by a team, only the last name of the
first collector is listed. For example, Breedlove &
Smith is included as Breedlove. Type collections
are in bold face. The number in parentheses fol¬
lowing the collector refers to the species listed
above, for example, Aguilar 438 is species # 6 from
the list (= O. helicterifolia).

Aguilar 438 (6). 2715 (9), 2819 (6); Allen 15209 (6);
Almeda 3762 (10); Angulo 200 (10); Antonio 1602 (10);
Aranda 1247 (II), 1 2b I (II), 1265 (II), 1317 (11); Aus¬
tin Smith 11523 (13). H679 (10). P21 14 (13).

Beaman 5402 (2); Bello 640 (13). 1977 (13). 4316 (13),
4340 (13); Bourgeau 2231 (2): Brant 2871 (2). 2886
(11); Breedlove 6307 (12), 6969 (II). 9467 (11). 15672
(3), 21696 (6), 23329 (6). 24742 (6), 25130 (6), 27588
(6). 27637 (I 1). 28182 (1 1), 31020 (6), 31371 (6). 32107
(6), 32239 (6), 32668 (6), 32699 (6), 33017 (3). 33543
(II), 34800 (6), 35276 (6), 38043 (2), 38781 (6), 42794
(II), 44395 (6), 4441 1 (6), 47853 (6), 48210 (6), 49332

(6), 49348 (6), 49371 (6). 49746 (6), 50583 (2), 51313

(6). 52589 (3). 52683 (6). 52998 (0). 53010 (6). 55689

(6), 55690 (6), 57479 (6). 58528 ((>). 60245 (6). 60289

(6), 66260 (6). 68637 (6). 68935 (II). 70059 (12), 72693
(II); Burger 1 1747 (6). 12065 (10), 12097 (13).

Cascante 644 (13); Chavarria I 153 (13); Campos 853
(6). 954 (6). 1020 (6). 1328 (6). 1877 (6). 2569 (6). 3012
(6), 3175 (6); Cedillo 667 (1), 1197 (1). 1204 (1). 1257

( I ) . 3060 (aff. 2), 3323 (aff. 2). 3406 (aff. 2); Chacon 1557

(13) . 1634 (14); Choc 89 (6); Churchill 4326 (6). 5926
(6); Contreras 5209 (II), 11186 (2), 11205 (2). 11233

(II) , 11235 (11). 11318(11); Croat 469 13(1 3), 476 1 3
(6). 642 1 5 (6).

Davidse 28547 (10), 28551 (10). 28627 (10). 28809

(10) , 28922 (10), 30477 (6). 34826 (1 1); Dryer 1336 (13).

Estrada 515 (14); Evans 1474 (II).

Galeotti 7004 (6); Gentry 13575 (6), 44065 (6); Gereau
2096 (1); Gomez 18814 (7). 23619 (14); Gonzalez 159

(14) ; Grayum 3519 (14). 8335 (6). 8870 (14); Grijalva
3705 (6).

Haber 314 (13). 5064 (13), 5591 (13). 5665 (13), 5701
(13), 5826 (13). 5998 (13). 6643 (13). 7938 (13), 8198

(13). 9155 (13). I 1267 (13); von Hagen 2022 (II). 2031

(11) ; von Hagen 2070 (I I); Hammel 7013 (10). 8534 (5),

15679 (6), 16504 (6 vel alf.). 17836 (14), 18049 (13),
19217 (9), 20062 (6). 20553 (8); Hartshorn 1463 (13),
1530 (14), 2130 (13); Hawkins 452 (11). 588A (II). 647
(11), 712 (11): Hazlett 582 (II). 1341 (6). 2516 (II);

Heath 594 (6). 2033 (6); Herrera 475 (13), 3778 (13),

4866 (6), 4946 (13), 4986 (14), 5144 (14), 8938 (7);
House 975 (II), I 187 (1 1). 1300 (1 1).

lshiki 1389(11), 1396 (11), 1488(11), 1551 (11).

Jimenez 564 ( 13), 827 (13); Johnson 578 (6): Juzepczuk
1513 (12).

kirkbride 486 ( 1 4).

Laughlin 528 (6); Lent 794 (7), 2070 (7); Liehiiiami
770/3 (1), 770/22 (1). 771/2 (I). 771/23 (1); Liesner
1789 (6); Linden 1641 (6); Lopez L. 181 (11); Lorea
5542 (6); Lorence 4711 (I); Lot 222 (6); Lundell 18944
(II), 19425 (6). 20435 (II), 20999 ( I 1 ), 2 1 1 70 ( 1 1 ).

MacDougall H288 (6); Martinez S. 17084 (3). 20768
(6): Matuda 1887 (6). 5020 (II), 5125 (6). 5352 (11),
5375 (1 1), 5400 (6). 5437 (II). 6095 (6). 16224 (6); Maya
1226 (6). 1316 (6), 2094 (6). 2865 (6). 2882 (6). 2989
(6). 3081 (6). 3098 (II), 3951 (6). 3967 (6). 4176 (11),
4232 (II). 4276 (6). 4292 (II); McPherson 9259 (13),
10421 (5), 10573 (5); Meave 1314 (6); Mejia 12 (11),
256 (II), 283 (II), 294 (II), 334 (II), 379 (11); Mendez
C. 7966 (6), 8503 (3), 8888 (6). 8930 (3); Mendez Ton
4214 (2). 4802 (II), 5035 (12), 5068 (11), 5264 (12),
5902 (6), 5931 (11), 9594 (3); Miller 2653 (6). 2943 (6

Volume 9, Number 4
1999

van der Werff

Ocotea helicterifolia Species Group

583

s.l.): Molina 203 (6), 8138 (6), 8180 (6). I 1321 (6), I 1769
(6). 12289 (6). 23320 (11). 24212 (II). 24289 (I I), 31 102
(6); Montenegro 1843 (6); Mora I 10 (14). 416 (7). 308 (7);
Moraga 171 (7); Morales 39 (13). 248 (13). 377 (6), 436
(10), 665 (6). 1053 (6). 4806 (10). 5812 (8). 5989 (8):
Moreno 7477 (6). 8253 ((>). 15427 (6). 1 5647 (6). 15826
(6). 15862 (6). 19376 (6). 193908 (0). 19424 (0). 21103
(6).

Nee 24906 (0). 29832 (afl. 2); Nelson 2419 (0).

Poveda 3920 (14).

Quesada 368 (6). 941 (7): Quintanilla 49 (3).

Ramirez 139 (6). 384 (6). 452 (6): Rarnirez-Mareial 656
(6): Raven 19804 (6). 21630 (6); Rios 179 (7); Rivera 260
(10). 664 (6).

Sandino 2330 (6). 4700 (O); Sanliz. Ruiz 830 ((>); Schul¬
tes 827 (2); Shank 55 (0): Sliilom Ton 779 |0). 2014
( 12), 3045 ( 12). 3560 (0). 7399 (II); Sinara 777 (aff. 2);

Soto 226 (6); Sousa 9337 (1); Stafford 316 (6); Standley
8044 (6). 20611 (6). 69379 (6). 69525 (6), 71446 (6),
89985 (6). 91546 (6); Stevens 17055 (6), 20397 (6).
22153 (6); Steyemiark 44207 (6). 44280 (6). 44648 (6).

Taylor 4460 (10); Tenorio 3457 (12 vel aff.), 5870 (1),
1 1270 (I ); Thorne 40325 (6): Tonduz I 1676 (8); Torres
470 ( 1 ). 494 ( 1 ). 1 4(H) (6), 2066 ( 1 ). 2928 ( 1 ), 4832 (11),
4871 (11). 4953 (I). 4973 (1). 6598 (6). 8866 (6). 9153
(6). 10444 (6), 10830 (6), 11620 (6), 11911 (6); von
Turckheim II 1651 (11). II 2164 (6).

Vazquez Torres 378 (2).

Wendt 2677 (aff. 2). 3000 (aff. 2). 3000A (aff. 2). 3861
(2), 4315 (6). 4653 (11), 4866 (11). 5137 (6), 5622 (2).
6765 (4); Wheelwright 209 (13); Williams 13888 (10),
15760 (6). 16469 (13). 20540 (6). 28623 (13), 28996 (13),
40390 (6); Wilson 40770 (6). 40859 (6).

Yneker 5777 ( 1 1).

Zamora 674 (14), 1287 (6); Zuniga 282 (6).

A New Species of Caryodaphnopsis (Lauraceae) from Vietnam

Henk van der Werff

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166, U.S.A.

Nguyen Kim Duo

Botany Department, Centre of Ecology and Biological Resources, Ngliia Do, Tu Liem, Hanoi,

Socialist Republic of Vietnam

AhstkacT. Caryodaphnopsis bilocellata van der
Werff & Dao, collected in the Cue Phuong National
Park in Vietnam and differing from the other Asian
species of the genus in having 2-celled anthers, is
described.

Caryodaphnopsis is a small genus of Lauraceae,
characterized hv its opposite leaves and strongly
unequal tepals, with the outer three much smaller
than the inner three. A unique wood anatomical
feature of Caryodaphnopsis is the presence of py¬
ramidal, trihydrate calcium oxalate crystals in the
ray cells (Richter, 1981). The genus is known from
Southeast Asia (southern China, Vietnam, Laos,
Cambodia, Indonesia, and the Philippines) and
from the Neotropics (Brazil, Peru, Ecuador, Colom¬
bia, Panama, and Costa Rica). The Asian species
have been revised by Kostermans (1974), who ac¬
cepted seven species. The six neotropical species
were recently described or transferred to
Caryodaphnopsis (van der Werff & Richter, 1985;
van der Werff, 1986, 1988, 1991; Zamora et al.,
1988), but have never been revised. Most species,
including all Asian species, have nine 4-eelled sta¬
mens. However, the neotropical C. inaequalis (A.
C. Smith) van der Werff & H. G. R ichter has nine
2-celled stamens, and the neotropical C. tomentosa
van der Werff has only six 2-celled stamens. Such
variation in stamen configuration within the same
genus is unusual among Lauraceae, but has been
reported in several genera (van der Werff & Rit li¬
ter, 1985). We attach more importance to the op¬
posite leaves, strongly unequal tepals, and the wood
anatomy for generic delimitation and accept the
variation in stamen numbers and locelli numbers
i n Caryodaph nops is.

During recent fieldwork in Vietnam, a species of
Caryodaphnopsis was found that differed from the
other Asian species in having nine 2-celled, and
not 4-eelled, stamens; its description follows here.

Caryodaphnopsis bilocellata van der Werfl &
Dao, sp. nov. TYPE: Vietnam. Ninh Binh Prov¬
ince: Cue Phuong National Park, //. van der
Werff et al. 14250 (holotype, MO; isotypes,
HN, KIN, QRS). Figure 1.

A oongeneris asiaticis antheris biloeellatis recedit.

Trees, 25 m tall. Twigs terete, glabrous; terminal
buds glabrous. Leaves opposite, chartaceous, ellip¬
tic or ovate-elliptic, 7-18 X 3-9 cm, glabrous on
both surfaces, the base obtuse or acute, the apex
acute, smaller leaves trinerved, larger leaves slight¬
ly triplinerved with the basal lateral veins leaving
the midrib ca. 3 mm from above the base ol the
leal, the lower surface glaucous, midrib and lateral
veins slightly impressed on the upper surface,
raised on the lower surface, tertiary venation im¬
mersed on both surfaces, petioles 10—14 mm long,
glabrous. Inflorescences axillary, 5-15 cm long,
glabrous, paniculate-cymosely branched. Flowers
hermaphrodite, glabrous, green. Tepals 6, the outer
3 triangular, scale-like, ca. 0.4 mm long, the inner
3 broadly triangular, ca. 1.7 mm long, erect with
the tip incurved, glabrous on the outer surface,
sparsely pubescent on the inner surface; stamens
9, 2-celled, the outer 6 ca. 0.8 mm long, the fila¬
ments ca. 0.4 mm long, pubescent, the anthers gla¬
brous, the locelli opening introrse; inner 3 stamens
opening extrorse, 1 mm long, the filaments 0.3 mm
long, pubescent, with 2 globose glands attached
near the base of the anther; staminodia 3, ca. 0.5
mm long, dorsally pubescent, pressed against the
ovary, with a triangular tip; ovary glabrous, 1 mm
long, with a slender style; receptacle shallow, gla¬
brous inside. Fruit unknown. Flowers April.

Floral measurements were taken from alcohol-
preserved material, and floral size of dried flowers
will be smaller (for instance, length of inner tepals
1.2 mm).

Vegetatively, Caryodaphnopsis bilocellata is quite
similar to C. metallica Kostermans and C. henryi
Airy Shaw and shares with those species the mostly

Novon 9; 584—586. 1999.

Volume 9, Number 4
1999

van der Werff & Dao 585

Caryodaphnopsis bilocellata from Vietnam

Figure 1. Caryodaphnopsis bilocellata van der Werff & Dao. — 1. Habit. — 2. Detail of inflorescence. — 3. Flower
showing outer tepals. — 4. Inner tepal. — 5. Outer stamen. — 6. Inner stamen with basal glands. — 7. Staminode. —
8. Distil.

glabrous leaves, twigs, inflorescences, and flowers.
However, the flowers of C. henryi have, at maturity,
spreading tepals with the inside of the tepals and
the stamens readily visible. Caryodaphnopsis henryi
also has some scattered hairs on inflorescences and
flowers, which are lacking in C. bilocellata. The
flowers of C. metallica have tepals with a dense
indument on the inner surface, and this indument
is readily visible even in half-open flowers; the te¬

pals in this species are also erect to spreading and
are not incurved as in C. bilocellata. The other
Asian species differ vegetatively from the new spe¬
cies in the presence of indument on flowers, inflo¬
rescences, twigs, or leaves. The most commonly
collected species, C. tonkinensis (Lecomte) Airy
Shaw, has larger flowers (inner tepals ca. 2.5 mm
long in dried specimens) that are obviously pubes¬
cent on the outside.

586

Novon

The only other species of Caryodaphnopsis with
nine 2-celled stamens is the neotropical C. inae-
qualis. We do not consider this species to he a close
relative of C. bilocellata because of differences in
leaf venation (trinerved or triplinerved in C. bilo¬
cellata, pinnately veined in C. inaequalis) and dis¬
tribution (northern Vietnam for C. bilocellata , low¬
land Brazil and Peru for C. inaequalis). We think
that the reduction of 4-celled stamens to 2-celled
stamens has taken place at least twice in
Caryodaphnopsis, once in the Neotropics and once
in Asia.

Paratype. VIETNAM. Ninli llinli Province: Cue
Phuong National Park, van <ler V/erff el <il. 1 4 195 (UN.
MO. QKS).

Acknowledgments. Fieldwork in Vietnam was
supported by National Geographic grant #5635-96.
The Curator of P generously made Caryodaphnopsis

specimens from Asia available for study. The illus¬
tration was made by the second author.

I .iterature Cited

kostermans, A. J. 0. II. 1974. A monograph ol
Caryodaphnopsis A. Shaw. Heinwardtia 9: 123—137.

Richter, II. G. 1981. Anatomie des sekundaren Xylems
und der Rinde der I .auraceae. Sonderb. Naturwiss. Ver-
eins Hamburg 5: 1-148.

Werff, H. van der. 1986. A new species of Caryodaphnop¬
sis (I .auraceae) from Peru. Syst. Rot. I I: 415—418.

- . 1988. Fight new species and one new combina¬
tion of neotropical Lauraceae. Ann. Missouri Rot. Card.
75: 402-419.

- . 1991. New species of Lauraceae from Ecuador

and Peru. Ann. Missouri Rot. Card. 78: 409—423.

- & H. C. Richter. 1985. Caryodaphnopsis Airy

Shaw, a genus new to the Neotropics. Syst. Rot. 10:
166-173.

Zamora. N.. L. J. Poveda A. X E. Canessa A. 1988. Una
uueva especie de Caryodaphnopsis Airy Shaw (I. aura¬
ceae) para la Region Neotropical. Ann. Missouri Rot.
Card. 75: 1 160—1 166.

A New Species and a New Combination in the Genus Parodiolyra
(Poaceae: Bambusoideae: Olyreae)

Fernando Zuloaga

Institute) de Botanica Darwinion, Casilla de Correo 22, San Isidro (1642), Argentina

Gerrit Davidse

Missouri Botanical Garden, P.0. Box 299, St. Louis, Missouri 63166-0299, U.S.A.

Abstract. Parodiolyra colombiensis, endemic to
eastern Colombia, is described and illustrated. It
differs from the three currently recognized species
of Parodiolyra in its pitted female lemmas and in
its 3-nerved lower female glumes. Its filiform fe¬
male pedicels, thickened rachilla internode be¬
tween the female glumes, prolongation of the rach¬
illa between the glume and the anthoecium, and
small, awidess female spikelets are characters it
shares with the previously known species of Paro¬
diolyra. Because Olyra micrantha possesses these
last four characters (except its female spikelets are
awned), and in addition has a relatively short hi-
lum, it is transferred to Parodiolyra. Because P.
colombiensis and P. micrantha share pitted lemmas,
they are believed to be more closely related to each
other than to any other species in the genus Par¬
odiolyra. A key to the five species of Parodiolyra
is provided.

In 1992 Davidse received a grass specimen (Res-
trepo & Matapi 467) as part of a collection for rou¬
tine identification that appeared to be an unde¬
scribed species of Parodiolyra. The defining
characters of the genus Parodiolyra are (Soder-
strom & Zuloaga, 1989): (1) female spikelets fall
entire by disarticulating below the glumes; (2) fe¬
male pedicels are filiform, not clavate; (3) a con¬
spicuous, thickened internode is present between
the lower and upper glume; (4) there is a prolon¬
gation of the rachilla between the upper glume and
the anthoecium; (5) the hilum extends for only Vi—
% the length of the caryopsis, not its entire length.
The Restrepo & Matapi 467 collection had female
spikelets with the first four characters. However,
these spikelets were young and without caryopses,
which makes it impossible to ascertain the hilum
type of the new species. After consultation with Zu¬
loaga, we decided to delay the description of the
new species in the hope that we might be able to
study more mature female spikelets from a new col¬
lection. In the meantime the site of the original

collection was revisited early in 1999 by Marfa Vic¬
toria Arbelaez (HUA) with the intent to re-collect
this species. However, she reported (pers. comm.)
that the original population had been destroyed in
order to expand the airstrip. We were recently able
to study duplicates of the original collection, which
(although also without caryopses) were more com¬
plete and slightly more mature; this allowed us to
strengthen our initial conclusions. Since this spe¬
cies grows in a remote area and it may be a long
time before it is re-collected, it seems prudent now
to report our findings. Even though some uncer¬
tainty remains about its generic classification, since
we do not know its hilum type, we now name it in
order to make this undoubtedly previously unde¬
scribed species available.

Parodiolyra colombiensis Davidse & Zuloaga,
sp. nov. TYPE: Colombia. Caqueta: Region de
Araracuara, alrededores de la pista aerea,
00.25'S, 72.30' W, 200-300 m, 11 Nov. 1991,
hierba, raquillas de la infrutescencia moradas,
frutos verdes, D. Restrepo & A. Matapi 467
(holotype, COAH-017796; isotypes, COAH-
020068, MO-05102566). Figures 1, 2.

Gramen perenne, monoecium; culmis superis nodis
ramificantibus; laminae ovatae-lanceolatae, 8—14.5 cm
longae, 1.7-3. 6 cm latae, glabratae, basi asymmetricae;
pseudopetioli 0.2 mm longi. fnflorescentiae panieulatae,
laxae, 6-10 cm longae. Spiculae femineae ellipsoideae,
4—5 mm longae, 1.3 mm latae, glumis inter se inaequali-
bus; gluma infera 3-nervia, nervis tessellatis; gluma su-
pera 5-nervia, nervis tessellatis; anthoecia ellipsoidea, 3.4
mm longa, 1.4 mm lata, excavata; lemma 5-nervium, apice
pilosum; palea 2-nervia. Spiculae masculae lanceolatae, 4
mm longae, 0.7 mm latae, scabrae; lemma 3-nervium, ac¬
uminatum; palea 2-nervia, stamina 3.

Perennial monoecious herb. Culms branching at
the upper nodes; intemodes cylindrical, 3-7 cm
long, rigid, hollow, glabrous; nodes shortly pubes¬
cent or glabrous, dark, compressed. Sheaths striate,
one margin ciliate, otherwise glabrous; auricles
membranous, up to 4 mm long, brownish, the mar-

Novon 9; 587-591. 1999.

588

No von

Figure I. Parodiolyra colombiensis Davidse & Zuloaga. — A. Habit. — B. Male spikelet, lemma view. — C. Male
spikelet, palea view. — 1). Male spikelet, detail ol lower anti upper glume. — K. Palea of male spikelet. — F Male
flower. — (i. Female spikelet, lower glume view. — H. Female spikelet, upper glume view. — I. bower female glume,
ventral view. — J. Upper female glume, ventral view. — K. Female anthoecium, dorsal view. - — b. Female anthoeeium.
ventral view. (. Restrepo A' Matapi 467.)

Volume 9, Number 4
1999

Zuloaga & Davidse
Parodiolyra

589

Figure 2. Detail of the female anthoecium of Parodiolyra
colombiemis. —A. Apex of pitted lemma and palea, lem¬
ma with bottle-like microhairs. — B. Detail of base. — C.
Middle portion of palea and borders of the lemma. — D.
Detail of pitted area in the apical portion of the palea.
(Restrepo & Matapi 467.)

gins ciliate. Ligules small, membranous-ciliate.
Blades ovate-lanceolate, 8—14.5 cm long, 1.7— 3.6
cm wide, flat, basally truncate and asymmetric, the
apex acuminate, the lower margins ciliate or gla¬
brous, otherwise scabrous, glabrous on the rest of
its surface; pseudopetiole brownish, ca. 0.2 mm
long, the adaxial surface densely pilose, the abaxial
surface glabrous. Inflorescence a panicle 6—10 cm
long, lax and open, terminal and exserted, some¬
times with one additional, axillary, slightly smaller,
partially exserted panicle; branches 1—6 per node,
the uppermost branches bearing only female spike-
lets, the branches in the lower half ol the panicle
bearing both female and male spikelets, with the
male spikelets borne below the female spikelets to¬
ward the base of the branches, the lowest branch
sometimes bearing only male spikelets; main axis
terete, scaberulous or sparsely hispid, purplish, the
axis of the branches smooth, glabrous or sparsely
hispid; pedicels of the male and female spikelets
similar, not thickened, scaberulous, 1.5—7 mm long,
purplish. Female spikelets narrowly ellipsoid, 4-5
mm long, 1.3 mm wide, disarticulating below the
glumes; lower and upper glumes subequal, acumi¬
nate, with a 0.6-mm-long, thickened internode be¬
tween them. Lower glume 4—5 mm long, scabrous
on the outer surface and toward the apex of the
inner surface, 3-nerved, the nerves anastomosing.
Upper glume 4. 2—4.6 mm long, scabrous on the out¬
er surface and toward the apex on the inner surface,
5-nerved, the nerves anastomosing. Rachilla inter-
node between the upper glume and anthoecium

somewhat elongated and slender (a stipe). Anthoe¬
cium ellipsoid, 3.4 mm long, 1.4 mm wide, shorter
than the lower and upper glume, indurated, stra¬
mineous, pitted, with rounded excavations over the
entire surface, the lemma with conspicuous bottle¬
like microhairs at the apex, 5-nerved; lodicules 3,
truncate, 0.4 mm long. Caryopsis unknown. Male
spikelets lanceolate, 4 mm long, 0.7 mm wide, pur¬
plish, with 2 small glumes at the base; lemma 3-
nerved, acuminate, scabrous; palea 2-nerved, sca¬
brous; stamens 3, the anthers 2.3—3 mm long.

The species is only known from the type collec¬
tion. The Araracuara region lies in an area of low
sandstone mesas in southeastern Colombia. This
area has a mosaic of vegetation types and has not
been well collected.

The size ol the plant is not known. It probably
has erect, self-supporting culms, but notes on its
habit are not available.

Although male spikelets of the Olyreae typically
lack glumes, the occasional occurrence of rudimen¬
tary glumes has been also reported in P. lateralis
(J. Presl ex Nees) Soderstroni & Zuloaga by Sod-
erstrom and Zuloaga (1989).

The characteristic bottle-like microhairs present
at the apex of the female lemma in P. colombiensis
are also conspicuous in both P luetzelburgii and P.
ramosissima.

Excluding the unknown liiluin character, P. col¬
ombiensis differs from the three currently recog¬
nized species of Parodiolyra in having a female
lemma surface pitted with round excavations, rather
than entirely smooth; however, all four species have
some pitting on the female palea. In P lateralis , P
luetzelburgii (Pilger) Soderstrom & Zuloaga, and P
ramosissima (Trinius) Soderstrom & Zuloaga, the
pits are restricted to the base and the apex, whereas
the palea is uniformly pitted in P. colombiensis. In
this respect P. colombiensis is similar to Olyra mi-
crantha Kunth, a widespread South American spe¬
cies. However, they clearly differ from each other
by the characters listed in Table 1. Olyra is poly¬
morphic for pitted anthoecia. Excluding O. micran-
tlui, 9 of the 23 currently recognized species of
Olyra possess pitted female anthoecia, while the
remainder have entirely smooth anthoecia.

Soderstrom and Zuloaga (1989) had already not¬
ed that 0. micrantha is atypical in Olyra because
its female spikelets disarticulate below the glumes,
the pedicels of the female spikelets are slender at
the apex, not thickened, and its glumes are sepa¬
rated by a conspicuous internode in female spike¬
lets. We have observed that there are two disartic¬
ulation points in the female spikelet of 0.

Table 1. Comparison of the five recognized species of Parodiolyra.

590

Novon

a:

1 .A

T C £ °

z v _c o;

• C C -n X

lr?r?

3 lA iA 1 'S.

CO -O £ Jc

cd _ « 6 -

JL B § lA 1 I

1 i T I .ts

1-0 03 tO CO Q. Q,

m E e e -a -o

I J C c V l

c e

d> a; _z

lO 03 ^ O o z ,

iA § 4 ,A 1 H

1 I i cs j

i I G I i .

= l 1 E ■-

3 T Q.

a; 5
rS _c" g

c g i £■ E -E

4

+ i-

S 4

-0 X

+ ^ + u A A

i

+ ro

03

ec ,1

> °?

4

z _•

- 3 X

J 2 ,A

micrantha, one below the anthoecium and the other
one below the glumes. Female spikelets without a
developed caryopsis disarticulate below the glumes,
although not as readily as in P. colombiensis (as
judged from herbarium specimens). However, in fe¬
male spikelets with mature caryopses, the anthoe¬
cium disarticulates leaving behind the empty,
somewhat persistent glumes.

Olyra species tend to have relatively large fe¬
male spikelets and Parodiolyra small ones. Com¬
paring their spikelet size by means of female an-
thoecia (since that eliminates having to deal with
awns, which tend to be highly variable in length
within the same species), Olyra species range from

5.5 mm in 0. wurdachii Swallen and 0. latispicula
Soderstrom & Zuloaga to 15.8 mm in 0. holttumi-
ana Soderstrom & Zuloaga. In Parodiolyra female
anthoecia vary from 1.9 mm in P luetzelburgii to
3.3 mm in P. ramosissima (Soderstrom & Zuloaga,
1989). The female anthoecia of 0. micrantha (2.8 —

3.5 mm) and P colombiensis (3.4 mm) fall within
the range of Parodiolyra. It seems likely that the
thickened female pedicels of Olyra species are
functionally related to their larger spikelets, espe¬
cially since their much smaller male spikelets have
slender pedicels, similar to the slender female and
male pedicels of Parodiolyra.

Soderstrom and Zuloaga (1989) did not investi¬
gate the hilum type of 0. micrantha as an aid in
the elucidation of the taxonomic position of this
species. We have done so now, and the hilum is
approximately %— % as long as the caryopsis (Ven¬
ezuela: David.se & Huber 22984 , David.se et al.
18746 ; Paraguay: Zardini & Guard 15512, Zardini
& Velasquez 15757, Zardini & Aguayo 15842, all
MO). In contrast, in all other Olyra species the hi¬
lum is approximately as long as the caryopsis.

On the basis of all these characters, it seems
more consistent to include 0. micrantha in Paro¬
diolyra, and we make this transfer here. Primarily
because they share pitted female lemmas, P. mi¬
crantha and P. colombiensis are probably more
closely related to each other than to any other spe¬
cies in the genus.

Parodiolyra micrantha (Kunth) Davids** & Zu¬
loaga, comb. nov. Basionym: Olyra micrantha
Kunth, in Humboldt. Bonpland & Kunth. Nov.
Gen. Sp. 1: 199. 1816. TYPE: Venezuela.
Amazonas: Maypures rapids of the Orinoco
River, Humboldt & Bonpland s.n. (holotype, P,
microfiche, fragment US-2877940).

Kk'i to thk Si’kciks ok Parodiolyra

la. Female and male spikelets awned: male spike-

Volume 9, Number 4
1999

Zuloaga & Davidse
Parodiolyra

591

lets 7.4— 9.3 mm long; blades 13—32 cm long;

inflorescence 10—32 cm long . P. micrantha

II). Female and male spikelets without awns; male
spikelets 3.5— 5.1 mm long; blades 2.5—14 cm
long; inflorescence 3—10 cm long.

2a. Lemma of female spikelets pitted; lower

glume of female spikelets 3-nerved .

. P. colombiensis

2b. Lemma of female spikelets smooth; lower
glume of female spikelets 5— 9-nerved.

3a. Female spikelets 2—2.6 mm long, black
at maturity; female lemma glabrous, ob¬
tuse; leaf blades 2.5—5 cm long .

. P. lateralis

3b. Female spikelets 3—5.8 mm long, light-
colored at maturity; female lemma with
conspicuous bicellular microhairs at the
apex, acute; leal blades 4—13 cm long.

4a. Female spikelets 3—3.6 mm long,

1.5— 1.7 mm wide; male spikelets
4.1— 5. 1 nun long; panicles with the
lower branches whorled; ovary gla¬
brous; lower margins of the blades
long-eiliate . P. luetzelburgii

4b. Female spikelets 5—5.8 mm long,

1 .9—2.8 mm wide; male spikelets
3.5— 4.3 mm long; panicles with the
lower branches alternate or oppo¬
site; ovary pilose toward the sum¬
mit; lower margins of the blades
glabrous or scabrous, not ciliate
. P. ramosissima

Acknowledgments. We thank the Herbario
Amazonico Colombiano and the Tropenbos Program
for making collections of the new species available
for our study, Maria Victoria Arbelaez for her at¬
tempt to re-collect the species, Ricardo Callejas
Posada for help with communications, and Vladi-
miro Dudas for the excellent illustration.

Literature Cited

Soderstrom, T. It. & F. 0. Zuloaga. 1989. A revision of
the genus Olyra and the new segregate genus Parodi¬
olyra (Poaceae: Bambusoideae: Olyreae). Smithsonian
Contr. Bot. 69: 1-79.

Note

Rogers McVaugh — A Protest

For 90 years my name has been Rogers Mc¬
Vaugh, not one letter more nor less. It was my moth¬
er’s wish that I be given a distinctive name, un¬
cluttered by a middle name that would never be
used. As a small boy I was proud ol it because it
put me on a par with my lather, whose name was
Roy McVaugh. In later life I have enjoyed it, and
have fought off the rather frequent attempts by ac¬
quaintances to change it to “Roger” or “Rodgers.”
Now it seems the time has come to admonish those
who wish to add something to a name that is al¬
ready long enough. I am not averse to seeing my
name in print, but I object to having it changed by
those who should know better, namely editors who
have not done their homework carefully enough.

The ultimate indignity surfaced not long ago, in
the recently published history of the American So¬
ciety of Plant Taxonomists (Schofield, 1998). There
my likeness appeared on page 30 above the name
of “Rogers M. McVaugh,” thus basely excluding me
from the elite group of my compeers, LeRoy
Abrams, Lincoln Constance, and Robert Ornduff,
who “had no middle initials” (ibid, page 32).

The seed of this gratuitous “M.” seems to have
been planted (appropriately enough) in Hortus
Third (New York: Macmillan, 1976: 1199), where
my name appeared in the list ol “Authors cited” as
"Michael Rogers McVaugh (1909).” The date of
birth is mine, of course, but the name? Doubtless
the editor had somehow (how in the world?) con¬
fused my name with that of the real Michael Rogers
McVaugh (bom 1938), who is not a professional
botanist but an eminent historian of medieval med¬

icine, who might even be embarrassed at seeing his
name linked so closely with the nomenclature of
the Spermatophytes.

The seed in the “Third Garden" grew and un¬
derwent some metamorphoses before it was brought
again to public notice in a world catalog of the
published names in the genus Euphorbia and its
close relatives (Oudejans, 1990). The Euphorbia-
eeous names for which I had been responsible were
duly listed, with the author reported as “[M.] R.
McVaugh.” The implication seemed to be that I had
been clandestinely suppressing my proper given
name through all these many years and many pub¬
lications, until at last it could be corrected for all
to see.

It is still not clear to me how the “M.” came to
be moved from the front of the name in 1990 to the
middle of the name in 1998. However that may be,
I disclaim the “M.” without any reservation! I re¬
pudiate it utterly, whether it be naked or bracketed,
initial or medial. I hope and expect to live the rest
of my life as the holotype of an unqualified and
unique binomial.

Literature Cited

Oudejans, Rob. C. II. M. 1990. World Catalogue of Spe¬
cies Names Published in the Tribe Euphorbiae (Eu-
phorbiaceae) with Their Geographical Distribution. Pp.
i— viii. 1 — 444. Utrecht, published by the author.
Schofield, Eileen k. 1998. A History of the American So¬
ciety of Plant Taxonomists. The First Sixty Years — 1936
to 1996. t.p.. i-vi. 1—80. Published by ASPT.

— Rogers McVaugh, Biology , University of North Carolina,
Campus Box .1280. 404 Coker Hall. Chapel Hill, North
Carolina 27599. U.S.A.

Volume 9, Number 1, pp. 1-126 of NOVON was published on 16 March 1999.
Volume 9, Number 2, pp. 127-288 of NOVON was published on 26 May 1999.
Volume 9, Number 3, pp. 289-456 of NOVON was published on 15 September 1999.
Volume 9, Number 4, pp. 457—592 of NOVON was published on 30 December 1999.

Novon 9: 592. 1999.