Contributions

FROM

The New York Botanical Garden

VOLUME X

1921-1923

Published by the aid of the

David Lydig Fund

Bequeathed by Charles P. Daly

/?

CONTRIBUTIONS

FROM

I he New York Botanical Garden

Volume X

(Nos. 226-250)

With 16 Plates and 31 Figures
1921-1923

TABLE OF CONTENTS

Some plants from Tropical Sea Gardens, by Marshall Avery Howe.

Scrophulariaceae of Colombia — I, by Francis W. Pennell.

A Graft-Chimera in the Apple, by A. B. Stout.

A Botanical Excursion to the Big Cypress, by John Kunkel Small.

"Veronica" in North and South America, by Francis W. Pennell.

Types of Flowers and Intersexes in Grapes with reference to Fruit

Development, by A. B. Stout.
Notes on Rosaceae — XIII, by Per Axel Rydberg.
A Rearrangement of the Bolivian Species of Centropogon and Sipho-

campylus, by H. A. Gleason.
Sterility and Fertility in Species of Hemerocallis, by A. B. Stout.
Phytogeographical Notes on the Rocky Mountain Region— X.

Grasslands and other open Formations of the Montane Zone of the

Southern Rockies, by Per Axel Rydberg.
Studies of West Indian Plants — X, by Nathaniel Lord Britton.
Cyclic Manifestation of Sterility in Brassica pekinensis and B. chi-

nensis, by A. B. Stout.
Studies of West Indian Plants — XI, by Nathaniel Lord Britton.
Alternation of Sexes and intermittent Production of Fruit in the

Spider Flower (Cleome spinosa), by A B. Stout.
Sterility in Lilies, by A. B. Stout.
Notes on Rosaceae — XIV, by Per Axel Rydberg.

The vegetational History of the Middle West, by Henry Allan Gleason.
Evolution and Geographical Distribution of the genus Vernonia in

North America, by Henry Allan Gleason.
Notes on Fabaceae — I, by Per Axel Rydberg.
The Taxonomic and Morphologic Status of Ophioglossum Alleni

Lesquereux, by Arthur Hollick.
Studies in tropical Ascomycetes — II. An interesting Xylaria from

Porto Rico, by Fred J. Seaver.
Notes on Fabaceae — II, by Per Axel Rydberg.
Studies of Lythrum Salicaria — I. The Efficiency of Self-Pollination,

by A. B. Stout.
The Physiology of Incompatibilities, by A. B. Stout.
Genera of North American Fabaceae— I, by Per Axel Rydberg.

PUBLICATIONS OF

The New York Botanical Garden

Journal of The New York Botanical Garden, monthly, illustrated, contain-
ing notes, news, and non-technical articles of general interest. Free to all
members of the Garden. To others, 10 cents a copy; $1.00 a year. Now
in its twenty-fourth volume.

Mycologia, bimonthly, illustrated in color and otherwise; devoted to
fungi, including lichens; containing technical articles and news and notes of
general interest, and an index to current American mycological literature.
$4.00 a year; single copies not for sale. [Not offered in exchange.] Now in
its fifteenth volume.

Addisonia, quarterly, devoted exclusively to colored plates accompanied by
popular descriptions of flowering plants; eight plates in each number, thirty-
two in each volume. Subscription price, $10.00 a year. [Not offered in
exchange.] Now in its eighth volume.

Bulletin of The New York Botanical Garden, containing the annual reports
of the Director-in-Chief and other official documents, and technical articles
embodying results of investigations carried out in the Garden. Free to all
members of the Garden; to others, $3.00 per volume. Now in its twelfth
volume.

North American Flora. Descriptions of the wild plants of North America,
including Greenland, the West Indies, and Central America. Planned to be
completed in 34 volumes. Roy. 8vo. Each volume to consist of four or more
parts. Subscription price, $1.50 per part; a limited number of separate
parts will be sold for $2.00 each. [Not offered in exchange.]

Vol. 3, part 1, 1910. Nectriaceae — Fimetariaceae.

Vol. 6, part 1, 1922. Phyllostictaceae (pars).

Vol. 7, parts 1-8, 1906-1922. Ustilaginaceae — Aecidiaceae (pars). (Parts
1 and 2 no longer sold separately.)

Vol. 9 (now complete), parts 1-7, 1907-1916. Polyporaceae — Agaricaceae
(pars). (Parts 1-3 no longer sold separately.)

Vol. 10, parts 1-3, 1914-1917. Agaricaceae (pars).

Vol. 14, part i, 1923. Sphaerocarpaceae — Marchantiaceae.

Vol. 15, parts 1 and 2, 1913. Sphagnaceae— Leucobryaceae.

Vol. 16, part 1, 1909. Ophioglossaceae — Cyatheaceae (pars).

Vol. 17, parts 1-3, 1909-1915. Typhaceae— Poaceae (pars).

Vol. 21, parts 1-3, 1916-1918. Chenopodiaceae— Allioniaceae.

Vol. 22, parts 1-6, 1905-1918. Podostemonaceae — Rosaceae.

Vol. 24, parts 1-3, 1919-1923. Fabaceae (pars).

Vol. 25, parts 1-3, 1907-1911. Geraniaceae — Burseraceae.

Vol. 29, part i, 19 14. Clethraceae — Ericaceae.

Vol. 32, parts 1 and 2, 1918-1921. Rubiaceae (pars).

Vol. 33, part 1, 1922. Ambrosiaceae— Carduaceae (pars).

Vol. 34, parts 1-3, 1914-1916. Carduaceae: Helenieae— Anthemideae.

Memoirs of The New York Botanical Garden. Price to members of the
Garden, $1.50 per volume. To others, $3.00.

Vol. I. An Annotated Catalogue of the Flora of Montana and the Yellow-
stone Park, by Per Axel Rydberg. ix + 492 pp., with detailed map. 1900.

Vol. II. The Influence of Light and Darkness upon Growth and Develop-
ment, by D. T. MacDougal. xvi + 320 pp., with 176 figures. 1903.

Vol. III. Studies of Cretaceous Coniferous Remains from Kreischerville,
New York, by A. Hollick and E. C. Jeffrey, viii -|- 138 pp., with 29 plates.
1909.

Vol. IV. Effects of the Rays of Radium on Plants, by Charles Stuart
Gager. viii -f 478 pp., with 73 figures and 14 plates. 1908.

Vol. V. Flora of the Vicinity of New York: A Contribution to Plant
Geography, by Norman Taylor, vi + 683 pp., with 9 plates. 1915.

Vol. VI. Papers presented at the Celebration of the Twentieth Anniversary
of The New York Botanical Garden, viii -f- 594 pp., with 43 plates and many
text figures. 19 16.

Contributions from The New York Botanical Garden. A series of technical
papers written by students or members of the staff, and reprinted from
journals other than the above. Price, 25 cents each. $5.00 per volume.
In the tenth volume.

THE NEW YORK BOTANICAL GARDEN

Bronx Park, New York City

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 226

SOME PLANTS FROM TROPICAL
SEA GARDENS

By MARSHALL AVERY HOWE

NEW YORK
1921

Reprinted, without change of paging, from Natural History, 20:
560-568, F 1021. [ 'November-December, 1920."]

Some Plants from Tropical Sea Gardens

By

MARSHALL AVERY HOWE

Curator of the Museums and Herbarium of the New York Botanical Garden

[Reprinted from Natural History, Vol. XX, Xo. 5, pp. 560-508, 1920]

THE MERMAID'S WINEGLASS, Acetabulum crenulatum, an exquisite green seaweed, photo-
graphed (natural size) soon after being taken from the waters of Biscayne Bay at Miami, Florida.
This dainty plant is of occasional occurrence in shallow bays of Bermuda, southern Florida, and the
West Indian islands. In Biscayne Bay it sometimes covers areas of considerable size to the exclusion
of nearly all other kinds of marine vegetation. The elegant, radiately chambered cups are bright green
when living, but they have a delicate coating of lime and they usually become chalky white soon after
being gathered

Some Plants from Tropical Sea Gardens

By MARSHALL AVERY HOWE

Curator of the Museums and Herbarium of the New York Botanical Garden

CVJ

en

A VISITOR from the north is often dis-
appointed by his first sight of a
L tropical strand, which commonly
shows little or no conspicuous vegetation
between the tide lines. Particularly is this
true if he is familiar with the rocky coasts
of northern New England, where a large
share of the richly abundant marine plant
life is exposed freely to view with every
ebbing tide. Probably the usual poverty
of the strictly littoral marine flora in the
tropics is due chiefly to the scorching
effects of the tropical sunshine, although
there are, of course, here as elsewhere, nu-
merous more or less wide areas where a
bottom of loose shifting sand, allowing no
stable foothold or anchorage, precludes the
development of any conspicuous vegeta-
tion.

But there are also extensive rocky
shores, submerged reefs, and bottoms
strewn with old corals or calcareous peb-
bles, where light, heat, aeration, and suit-
able anchorage combine to furnish ideal
conditions for the development of ma-
rine gardens. If the observer can wade
into such a place at low tide, especially on
a calm morning before the daily trade
wind arises to ruffle the surface of the
water, the sight is one that is long to be
remembered. Or perhaps he can row over
it on one of the calm days such as occa-
sionally occur in late spring or in summer,
or can view it even in less placid weather
from a glass-bottomed boat at some winter
resort.

In addition to the graceful, often bril-
liantly colored or iridescent algae — the plants
proper — there are commonly also in such a
garden stately corals and sea fans, which
are colonial animals. These animals, because
they are attached and have no more power
of locomotion than a tree, are ordinarily
looked upon as plants by the uninitiated. In
fact, the "sea gardens" that are exhibited
through glass-bottomed boats to patrons of

southern winter resorts are sometimes almost

exclusively "zoological gardens." Besides

£># the colonial animals firmly attached to the

home spot there are often also gaily colored

tropical fish swimming in and out among
the other organisms and giving a touch of
active life to these submarine beauty spots.

Although the adjective "tropical" is used
in our title and elsewhere in this discus-
sion, the wealth of the marine vegetation
in the subtropics of our North American
coasts and adjacent islands is probably
even greater than that of the tropics,
strictly speaking. In respect to number
and variety of species of "seaweeds" or
algae the richest areas in the northern
half of the western hemisphere — at least
the richest visited by the writer — would
appear to be Bermuda, the Florida Keys,
and the coast of California. It is possible,
however, that Guadeloupe,! wholly within
the tropics, deserves to be considered in
this connection. Parts of the Bahamas, of
Cuba, Jamaica, and Porto Bico, and of the
Caribbean coasts of Panama are well sup-
plied with marine algae, but, taken as a
whole, they do not give the collector the
impression of wealth that he obtains from
prowling about in the waters of the Florida
Keys and Bermuda.

The "Sargasso Sea," in any such magni-
tude and character as was described by
some of the early navigators and as was
represented on some of the maps made only
fifty years ago, seems to be more or less of
a myth. Yet, floating mats 2 of Sargassum,
several feet or, rarely, several rods in
width are frequently met with, as one
steams southward from New York or Hali-
fax, or cruises about among the West In-
dian islands. Floating Sargassum is found
particularly in the path of the Gulf Stream,
which sometimes brings it far to the
north, occasionally casting it ashore after

1 The algal flora of this island has been inten-
sively studied by Frencli scientists and the list of
marine species and varieties attributed to it
reaches the imposing total of 811, a number con-
siderably greater than that thus far attributed in
any published paper to any of the three other
regions. The determinations on . which the
Guadeloupe list is based are not altogether criti-
cal, however, and the sum total is swollen by the
inclusion of numerous varietal or form names.

- These mats consist chiefly of two species, Sar-
gassum natans and S. fluitans, which are cer-
tainly known only in a free floating condition.

561

562

NATURAL HISTORY

THE MERMAN'S SHAVING BRUSH,
Penicillus capitatus, from Bermuda. (About
one half natural size)

a storm on Long Island, Martha's Vine-
yard, and Nantucket. As is well known,
it is accompanied by a characteristic fauna
of bryozoa, crustaceans, small mollusks,
and the like, and fishes which go along
with it to feed on the small animals that
it carries. Species of Sargassum are diffi-
cult to limit and define, but it is safe to
say that a dozen or more occur in the West
Indian region. Most of these are surf
plants, growing firmly attached to rocks
and reefs in exposed places. They are
found near the low-water mark, and, with
their cousins of the genus Turbinaria, they
take the place, in the south, of the north-
ern rockweeds of the genera Fucus and
Ascophyllum, both taxonomically and eco-
logically, although never so conspicuous
ami massive as the latter often are.

The genus Sargassum, like all the other
alg», belongs to the large group of plants
known in the books as "thallophytes" — a
group in which the plant body, according
to the bookmakers, is not differentiated
into root, stem, and leaf; yet it is very
difficult, if not impossible, to frame a defi-
nition of a leaf which may not apply to
the leaflike structures shown by the species
of Sargassum. In other words, a Sargassum

is a "leafless" plant that appears to have
leaves.

The larger algae are divided into three
great groups, which are often spoken of in
an untechnical way as the "greens," the
"browns," and the "reds," these names
being abbreviated translations of the tech-
nical group names, which have been based
upon the prevalent colors shown by repre-
sentatives of these groups. The "browns"
include the largest kinds of seaweeds, such
as the kelps and rockweeds of the North
Atlantic and the giant kelps of our Pacific
coast, individuals of which often reach a
length of more than one hundred feet and
the extensive beds of which are now being
used as a source of potash for the Ameri-
can farmer and gardener. In the American
tropics, the more conspicuous members of
the brown group include, besides Sargas-
sum and Turbinaria, already mentioned,
representatives of several other genera.

It is perhaps in the large group of algae
known colloquially as the "greens" that
one finds the most interesting and curious
as well as the most beautiful of the tropi-
cal sea plants. In these, the leaf-green or
chlorophyll, characteristic of plants in
general, is found essentially free from ad-
mixture with other pigments. In many of
them, however, the plant body is more or
less coated or permeated with lime, so that
it is often whitish, at least in the older
parts, or speedily becomes so after being
taken from the water and exposed to the
light.

Of the group of calcified green algae the
four species of Penicillus, indigenous to the
sea bottoms of Bermuda, southern Florida,
and the West Indies, are among the most
curious. In these, as is suggested by the
Latin name, the plant body takes the form
of a brush, its component filaments being
closely interwoven to make a cylindric or
flattened stalk and then, at the top, set
free and forking, each enclosed in a porous
sheath of lime, they constitute the head or
bristles of the brush. One sort, perhaps the
commonest, is sometimes known as the
merman's shaving brush, or is occasionally
referred to by the less particular and
meticulous as the mermaid's shaving brush.
A relative of the plant just described and
one that grows in similar situations is the
mermaid's fan. There are several species
of this generic group (Udotea), in all of

SOME PLANTS FROM TROPICAL SEA GARDENS

563

which the delicate root hairs penetrate the
calcareous sand and attach to themselves
small particles of this sand, thus making
anchorage hawsers in a bottom that is not
altogether stable. The rootlike part, in the
attached algae, is supposed to serve simply
as a holdfast. These sea plants often affix
themselves to smooth hard rock, or some-
times to pieces of iron or glass, from which
they can evidently draw little or nothing
in the way of nourishment. Their food is
apparently derived directly from the sur-
rounding water and air and is taken in
directly by the general surface of the
plant. In the case of these Udoteas and
their relatives, however, with their highly
developed systems of sand-burrowing rhi-

zoids, it may be suspected that these
rhizoids play a part in the gathering of
food as well as in anchoring the plant.
What are commonly called "sea fans" by
frequenters of the sea are organisms of a
very different nature from those here de-
scribed under the name of mermaid's fan.
They are larger and are colonial animals
related to the corals.

Another interesting and attractive mem-
ber of the group of lime-coated green algae
is the sea fir (Rhipocephalus phoenix), the
center of whose distribution appears to be
the Bahama Banks. This plant is usually
from two to six inches high and when grow-
ing on the sea bottom is very suggestive of
the little German Christmas trees that were

THE MERMAID'S FAN, Udotea conglutina, from Bimini Harbor, Bahama Islands. (Natural size)

A CALCIFIED AND SEGMENTED GREEN SEAWEED, Halimeda simulans, from Porto Rico.

(About four fifths natural size)

once sold in this country for holiday deco-
rations. It is dark green when living but
is soon bleached to a chalky white after
being killed and exposed to the light. In
general structure, it is somewhat interme-
diate between the brushes and the fans,
being brushlike in general habit but hav-
ing numerous small overlapping fans for
branches.

The writer once enjoyed the privilege of
being becalmed for two days on the
Bahama Banks in a small sloop. On these
banks are hundreds of square miles where
the water is mostly from one to twenty
feet deep. The bottom here is chiefly of
more or less compacted oolitic sand, con-
sisting of clean white nearly spherical
granules that suggest fish roe both in form
and size. Such a bottom rarely becomes
muddy to any appreciable extent and the
waters above it are wonderfully clear, so
much so that when a breathless calm makes
the surface of mirror-like smoothness one
can see the vegetation and animal life of
the sea bottom in twelve or twenty feet of
water almost as distinctly as if the water
were only two feet deep. On wide areas of

564

the Bahama Banks the merman's shaving
brushes, the mermaid's fans, the sea firs,
and their relatives are the dominant fea-
tures of the marine flora, sometimes forming
a continuous carpet on the floor of the sea.

Prominent among the relatives of the
Udoteas are eight or ten kinds of Hali-
meda, in all of which the calcified plant
body is regularly jointed or segmented. In
most of the species of this genus the lime
is abundant and the system of joints gives
a certain degree of flexibility to a plant
that would otherwise be quite rigid and
stonelike. One of the species, at least,
sometimes occurs in great masses, and the
quantities of lime left by its decay are, in
places, important factors in reef building
and land forming. The fact that it is a
more rapid grower than the corals seems to
give it a certain advantage over them in
the matter of secreting and depositing
lime, even though the proportion of lime
in its make-up is not so great.

Perhaps the most dainty and exquisite of
the green algae is the mermaid's wineglass,
Acetabulum crenulatum, which grows in
shallow bays and protected places, ranging

SOME PLANTS FROM TROPICAL -SEA GARDENS

565

as far north as Bermuda and southern
Florida. It is rarely found in any great
quantity, yet occurs in considerable abun-
dance on old shells and pebbles, mostly in
ten or fifteen feet of water or less, in the
bays lying between the keys and mainland
of southern Florida. The actual height of
the plants is usually from two to four
inches. The elegant cup-shaped disk which
surmounts the graceful stalk is largely
reproductive in function, each of its radial
chambers containing at maturity, in the
present species, from 200 to 500 sub-
globose, firm-walled spores, scarcely visible
to the unaided eye, each of which produces
on germination a number of smaller motile
cells which are sexual in nature. At least
three other species of this genus occur in
the West Indian region.

Among the larger green seaweeds of the
warmer parts of the earth are some that
are not calcified and of these the species of
Caulerpa — a dozen or more of them in Ber-
muda, southern Florida, and the Wesl
Indies — deserve especial mention. These
present themselves in a great variety of
graceful and attractive forms, some of them
suggesting delicate feathers, others looking

like clusters of green grapes, the inflores-
cence of grasses, the twigs of cypress trees,
or sprays of running pine. They are found
in tide pools, on the roots of the red man-
grove in lagoons, and creeping on the sea
bottom down to a depth of a hundred feet
or more. Individual plants of some of the
kinds get to be four or five feet long. A
curious thing about them is that, although
they are plants of considerable size, no one
has yet certainly detected in them any
spores or other special reproductive organs.
They seem to maintain themselves by
simply continuing to grow at one end
while dying off at the other, or to propa-
gate their kind by accidentally detached
fragments. It nevertheless seems probable
that they produce some sort of minute
reproductive cells which have thus far
escaped observation and detailed descrip-
tion. Doubtless much remains to be learned
about the life history of the Caulerpas and
many of their relatives by some one so
situated that he can watch the living
plants continuous]}- throughout the year,
with a compound microscope and needed
accessories at hand.

A species of J'ahinui, taking the form of a

A GREEN SEAWEED, Caulerpa racemosa, that suggests bunches of grapes, photographed (not
in place) at the mouth of the Guanica Harbor, Porto Rico. (About one half natural size)

566

NATURAL HISTORY

balloon or an irregularly oval sac, ranging
in size from the dimensions of a robin's egg
to those of a hen's egg, and filled with a
fluid protoplasm, is often found washed up
on beaches in the West Indian region and
is, in Bermuda at least, often referred to
as "sea bottles." Dark green and iridescent
in life, it becomes clear and translucent
after being killed and exposed to the light
for a time and may be as attractive then
as when living. Children sometimes pick
it up on the beaches and by skilfully
exerted pressure playfully squirt the liquid
contents into each other's faces. These
little "bottles" grow in shallow water
mixed in with seaweeds of the soft mossy
kinds or under shelving rocks near the low-
water line. Another kind of Valonia, con-
sisting of somewhat smaller ovoid or bottle-
shaped segments that branch and cohere
in large masses, is often beautifully iri-
descent and very attractive when seen
growing in the water.

It is in the large group of marine plants
known nontechnically as the red algae that
we find the greatest variety in the tropics,
although as individuals the "reds" are
rarely so numerous or so conspicuous as are
many of the "browns" and "greens." Many
of the most interesting and beautiful of
the "reds" are so small that their wondrous
symmetry and beauty are revealed only to
the user of a hand lens or a compound
microscope. Several are so delicate or
have such a soft gelatinous texture that
when floated out on paper and dried under
pressure they adhere so closely to the paper
and have so little substance that they are
sometimes taken for paintings by those
who do not stop to realize that no human
hand could trace lines of such delicacy and
symmetry as these "flowers of the sea"
often possess. One of the coarser red sea-
weeds of the tropics, Bryothamnium tri-
quetrum, looks a little more like some
land plants. The main axes have three
rows of short toothed or fringed branch-
lets, giving these axes a three-angled or
three-winged effect. The individual plants
form dense clumps one or two feet in
thickness, but they sometimes grow asso-
ciated in large numbers, forming extensive
beds.

The red algae, like the greens and the
browns, contain chlorophyll, the green color
substance common to plants in general, but

they have also another pigment that modi-
fies or obscures the green, so that the plants
appear to be of some shade of red, pink, or
purple, or sometimes almost black. The
red pigment is soluble in fresh water, and
the green is not, so that red seaweeds
washed up on the beach and exposed to
rain often become green or show zones or
spots of green in the more exposed parts.
On the other hand, the green is soluble in
alcohol while the red is not, so that the two
pigments may be easily separated. The
red algae that are really and strikingly red
are, with few exceptions, inhabitants of
deeper water than the greens and the
browns and are usually collected by dredg-
ing—or by being found washed ashore, par-
ticularly after a storm.

Several kinds of red algae as well as
of "browns" are extensively used by the
Chinese and Japanese as articles of food.1
The agar-agar of commerce, derived from
red algae, is a food in the Orient, but is
known in America best as a nutrient me-
dium for laboratory cultures of bacteria
and fungi. Another product of red sea-
weeds, known as "funori," is manufac-
tured by the Japanese to the amount of
two or three million pounds a year and is
used by them for sizing for cloth, for
which purpose it seems to have certain ad-
vantages over sfarch. Most of these ma-
rine algae from which the Japanese derive
products that sell for several millions of
dollars a year have close relatives in
American waters, but apart from the re-
cent development of the kelp potash in-
dustry in California and the use of sea-
weeds as a fertilizer for the land by
farmers living in the vicinity of the sea,
the inhabitants of the United States have
thus far made little practical use of the
plant life of the ocean.

On the coast of Massachusetts the Irish
moss or carrageen (Chondrus crispus i
used for making sea-moss jellies or pud-
dings, is collected to the value of a few
thousand dollars a year, and the "dulse"
(Elwdymenia palmata), which is eaten raw
as a sort of salad or relish, is gathered in
still smaller quantities and offered for sale

1 See paper, by the writer, on "Some Economic
Uses and Possibilities of the Seaweeds." Journal
of the New York Botanical Garden, XVIII, 1917,
pp. 1-15.

SOME PLANTS FROM TROPICAL SEA GARDENS

56'

in the water-front markets of New York
and Boston. But with the increase of the
population of our country, and with the
certain advance of science and its applica-
tions, it is probable that the future will
witness a widely extended utilization of
the plant resources of the sea.

In at least three of the natural families
of the red algae the plant body takes up
lime from the sea water and becomes more

animals, even though, as would now appear,
these coral animals are often of secondary
importance. But the coral-like seaweeds
are always plants, however coral-like they
may look. In microscopic structure and
modes of reproduction they are just as
truly plants as are any of the seaweeds of
the soft mossy kinds. Yet it is not at all
surprising that those who have not studied
such things sometimes confuse these hard

A LARGE RED SEAWEED, Bryotkamnium triquetrum, of the American tropics and subtropics.
photographed (not in place) at the mouth of Guanica Harbor, Porto Rico. (About one third natural
size.) At Key West, Florida, this species forms large mats on the floor of the ocean, mostly at a depth
of from twenty to fifty feet

or loss hard and stonelike. In one of these
families in particular (the Corallinami\
so called on account of a superficial resem-
blance to the corals) the plant, except for
the inconspicuous reproductive cells and
almost equally inconspicuous apical or
superficial layers of new tissue, is almost
as hard as any limestone rock. It has be-
come apparent in recent years that in
many parts of the world these lime-secret-
ing sea plants are and have been an im-
portant factor in the building of reefs, a
line of activity that in the past has been
attributed almost exclusively to the coral

stonelike plants with the corals and that
even the naturalists of a hundred years
ago often did likewise.

The lime-secreting plant often forms a
crust that gradually creeps over, covers,
and smothers coral animals. Plant crusts
of these kinds are usually small and thin
and are probably of not much importance
in building reefs. But there are many dif-
ferent kinds of these hard lime-secreting
plants and some of them occur in large
masses and form extensive reefs. Certain
species are of a delicate pink color, so that
one may easily imagine what an attractive

568

NATURAL HISTORY

ONE OF THE STONY, LIME-SECRETING
CORAL-LIKE SEA PLANTS, Goniolithon stric-
turn, from Key West, Florida. (About one half
natural size)

AN ENCRUSTING SEAWEED, Goniolithon
solubilc, growing on a living coral, from Porto
Rico. (About two thirds natural size.) The
lime-secreting plant is conquering the lime-se-
creting colonial animal, gradually covering and
smothering the coral polyps

display a reef must make when exposed,
or even when visible through calm
water.

There are coralline alga? which grow at-
tached to the shells of members of the
clam or mussel family. The intimate at-
tachment is probably unwelcome to the
clam, although the overshadowing pres-
ence of the shrubby stonelike alga may ren-
der the clam a service by covering and
protecting it from some of its enemies.
The alga? in general are not very particular
as to their points of attachment, although
something substantial and firm in the way

of a substratum seems usually to be pre-
ferred.

Living corals are found only in the
warmer seas; coral-like plants occur not
only in the tropics, but also in temperate
and frigid waters. Explorers in the Arctic
regions have reported great beds of them on
the floor of the ocean, mostly in water that
is from 60 to 120 feet deep.

A few years ago the Royal Society of
London sent a party of naturalists to the
South Pacific to study the mode of origin
of the so-called coral islands. The island
of Funafuti of the Ellice Islands group
was chosen for special study because it
was believed to be "a true coral island."
By means of a drill, borings were made to
a depth of a little more than 1100 feet, the
cores brought up were carefully studied,
and the various groups of animals and
plants that had contributed to the up-
building of this island were ranked in
order of their relative importance. The
first rank was given to red algse of the
coralline family; the second to lime-se-
creting green alga? of the genus Halimeda;
third rank was awarded to the group of
microscopic animals known as Foramini-
fera; and fourth rank to the corals. So
Funafuti seems to be "a true coral island'"
which, strictly speaking, is not a coral
island at all!

American geologists are finding evidence
that certain limestones, now high and dry,
in various parts of the United States and
the West Indies, are made up chiefly of the
remains of lime-fixing plants that flour-
ished when those parts of the earth's crust
were under the surface of the sea. So
these hard limy sea plants, living and dying
century after century and rising "on step-
ping-stones of their dead selves," are not
only making land today, but their an-
cestors and relatives did the same thing
thousands of years ago.i And no small
part of the pleasure and satisfaction of
exploring the sea gardens of the tropics
comes from observing bits of evidence as
to how this great work is still being ac-
complished.

1 For a further discussion of this subject, see a
paper by the writer on "The Building of Coral
Reefs." Science, XXXV, No. 909, May 31, 1912,

PI'. s;;7-42.

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL
GARDEN No. 227

SCROPHULARIACEAE OF
COLOMBIA-I.

BY

FRANCIS W. PENNELL

NEW YORK
1920

Reprinted, without change of paging, from Proceedings of the Academy of Natural
Sciences of Philadelphia. 1020

Issued Oct. 4, 1920

136 PROCEEDINGS OF THE ACADEMY OF MayOet ,

SCROPHULARIACEAE OF COLOMBIA— I.
BY FRANCIS W. PENNELL,.

For a period of eight months, during 1917 and 1918, the writer
was engaged in scientific work in Colombia. No attempt will now
be made to tell the story of his adventures there, nor to give more
than the briefest summary of his impressions of Tropical or Andine
vegetation. A short narrative of my explorations and a compara-
tive sketch of the plant-life seen has already been presented in the
Journal of The New York Botanical Garden for June, 1918. 1 will
simply say that my travels took me from the northern seacoast to
the Andes east of Neiva in 3° north latitude, from the prairies and
lowland forest of the Orinoco drainage, from the Magdalena and
Sinu valleys, upward through every zone of vegetation to the sum-
mits of each of the three ranges of the Andes The greatest diver-
sity of life was seen, and the collections of plants, brought from
nearly all points visited, inc.ude much that is new to science.

From July 6 to August 16, 1917, it was my privilege to work
with Dr. Henry H Rusby, of the College of Pharmacy, New York
City — a companionship which to a botanical no\ice in a strange
land was invaluable. All specimens made on and before August 16,
while numbered consecutively with those made by myself alone
later, are to be cited as Rusby & Pennell. . . .

Detailed maps of Colombia are difficult to obtain, and many of
our collecting-stations were at small towns, or single houses. Con-
sequently, although in the lists of specimens given I am stating the
Department in which each point is located, it seems important
to give a full list of the localities from which our plants have come.
Arranging these in the order of our itinerary, and grouping them
according to broad natural areas of topography, should enable any-
one to place approximately any Station. For each point the De-
partment is stated. The names of houses are placed in quotation
marks.

Northern Seacoast:

July 6, 1917. Cartagena, Bolfvar.

o

1920.] NATURAL SCIENCES OF PHILADELPHfA. 137

Along Rio Magdalena:

July 8, 1917. Barranquilla, Bolivar
" 10, " Calamar,
" 11, " El Banco, Magdalena
" 12, " Gamarra and Carpentiera, Magdalena
" 14, " Puerto Berrio, Antioquia
" 15, " Buenavista, Caldas.
Plain of Tolinia :

■Tnly 16, 1917. Mariquita and San Lorenzo, Tolima.
Plain of Upper Magdalena:

July 19, 1917. Girardot, Cundinamarca
" 21, " Espinal to Cuamo, Tolima
" 22, " Cuamo to Rio Saldafia, Tolima
" 22, " Rio Saldafia to ^atagaima, Huila
" 24, " Quebrada de Angeles, "

25, " Quebrada de Angeles to Rio Cabrera, Huila
" 26, " Rio Cabrera to Villavieja, "

" 27, " Villavieja to Neiva,
" 30, " Neiva, Huila.
Cordillera Oriental:

July 30 to August S, 1917. Excursion from Neiva over the crest of the

Cordillera to "Balsillas", and return.
Along Upper Magdalena :

August 8-9, 1917. Neiva, Huila

12, " Natagaima, Huila

13, " Boca Saldafia, Huila.
Railroad from Girardot to Bogot a :

August 14, 1917. Portillo, Anapoima, San Joaquin, Bospicio, La Esper-
anza. Cachipay, Zipacon and Anolaima, Cundina-
marca.
Upper western slopes of Cordillera Orient al :

August 16, 1917. Bogota, Cundinamarca.
Eastern slopes of Cordillera Oriental:

August 22, 1917. Chipaque, Cundinamarca
" 23, " Caqueza, "

" 24, Caqueza to Rio Sananie, Cundinamarca

" 24, " Quetame to "Susumuco," "

" 25-26, 1917. "Susumuco,"
Plain of Meta:

August 26 to September 2, 1917. Villavicencio, Meta.
Eastern slopes of Cordillera Oriental:

September t. !'M7. Villavicencio to "Buenavista," Meta

4, "Buenavista" to " Pipiral," Cundinamarca

5, " "Pipiral" to "Susumuco,"

5, " "Guayabetal,"

6, "'Guayabetal" to ''Monte Redondo," Cundinamarca
" 7. '■ "Monte Redondo" to Quetame,

" 7-8, " Caqueza,
8, " ll>.
Summit of Cordillera Oriental:

September 8, 1917. Paramo de Cruz Verde, Cundinamarca.
Upper western slopes of Cordillera Oriental:

September 12, 1917. Bogota and Mt. Guadalupe, Cundinamarca
13, " Bogota (Rio San Francisco),
15, " Tequendama, "

" 17, " Bogota (Monserrate). "

" 18, 23, " Bogota* (Chapinero),
Summit of Cordillera Oriental:

September 20, 1917. Paramo de Cruz Verde, Cundinamarca.
Upper western slopes of Cordillera Oriental:

138

PROCEEDINGS OF THE ACADEMY OF [May-Oct ,

September 20, 26, 1917. Bogota (Rio San Cristobal), Cundinamarca
" 22, 24, " " (Rio del Arzobispo),

" 24,25, " " (LasCruces),

26, " " (Cerro de Focha),

Summit of Cordillera Oriental:

September 27, 1917. Paramo de Choachi, Cundinamarca.
Upper western slopes of Cordillera Oriental:

September 30, 1917. Bogota (San Cristobal), Cundinamarca
October 4-8,

6, " " (Chapinero),

6, " Rio Teusaca,
12, " Bogota (Rio San Cristobal)
" 13-15, " Sibate to El Peiion,
" 20-24, " Zipaquira to Mt. Chuscal, "

" 23, " Nemac&n, '•

" 28, " Tequendama,

29, " Sibate to El Peiion,
Summit of Cordillera Oriental:

November 14, 1917. Paramo de Cruz Verde, Cundinamarca.
Lower western slopes of Cordillera Oriental :

November 28-30, 1917. Fusagasuga, Cundinamarca
December 1-4, " Pandi and Icononzo, Cundinamarca.
Plain of Upper Magdalena:

December 4-5, 1917. Melgar and Girardot, Cundinamarca.
Eastern slopes of Cordillera Central:

December 11,1917. Libano, Tolima

15-17, " "Rosalito," between Murillo and Paramo de Ruiz,
Tolima
" 16-17, " Paramo de Ruiz, Tolima

17-18, " Murillo,

18-29, " Libano ("La Trinidad" and "La Virginia"),
Tolima
"El Convenio," Tolima.

" 29-30, "

Plain of Tolima :

December 30, 1917.
January 6, 1918.
" ' 3,7, "
Along Rio Magdalena:
January 9-10, 1918.
11-13, "
14, "

14, "

15, "

15. "

16, "
18-19, "

Plain of Bolivar:

January 24, 1918.

25, "

26, "
" 27 "

28,' "
30, "
Along Rio Sinu:

February 3, 1918.

4, ■"

5, "

6, "
8, "

San Lorenzo, Tolima

Guayabal and San Felipe, Tolima

Honda and Mariquita, "

Brazuela de Perales, Antioquia
Puerto Berrio and Malena, Antioquia
Vuelta de Acufia, "

Opposite Boca Carare, "

Canabetal, Bolivar.
Boca de Rosario, Puerto Nuevo and Boca Sogomosa,

Santander
El Banco, Magdalena
Magangue, Bolivar.

Buenavista, Bolivar

Since and Corozal, Bolivar

Sincelejo, "

Chinu and Sahagun, "

Cienaga de Oro and Cerete, Bolivar

Monteria. "

"Medellin," Bolivar

"Los Hurtados," Bolivar

Morales, "*

Barro Blanco, "

Tierra Alta and Boca Tai, Bolivar

1920. NATURAL SCIENCES OF PHILADELPHIA. 139

" 10-11, " "Angostura" and Frasquillo, "
" 13-14, " Boca Verde.
Upper slopes of Cordillera Occidental:

February 23, 1918. Paramo de Chaquiro, Bolivar
25, " Cascada Chorron. "

Along Rio Antizales :

February 25-26, 1918. Antizales, Bolivar.
Along Rio Esmeralda:

February 26, 1918. Boca Antizales, Bolivar

27, " "Las Dantas" to "Puerto Canoa," Bolivar

28, " "Puerto Canoa" to "Salvajin."
Along Rio Sinu:

March 1, 1918. Boca Esmeralda, Bolivar

4, " Boca Verde, "

5-6, " Frasquillo and "Angustura," Bolivar

7-10, " Boca Tai and Tierra Alta,

11, " Morrocoquiel, "

12, " "Los Hurtados,"
21, " Monteria, "
23, " Vilches,
25, " "El Pueblo," below Lorica.

Northern Seacoast :

March 26 27, 1918. Cartagena and Turbaco, Bolivar
April 1, " Santa Marta, Magdalena.

Of chief interest has been the comparison of the vegetation of
different altitudinal life-zones, and these upon the slopes of the
three divergent Cordilleras of the Andes. The central axis of the
Andes; entering Colombia from the South, soon divides into three
ranges, all of which, rising from the midst of a broad low Tropical
plain, reach high elevations. Also in the northeast is the wholly
isolated Sierra Nevada de Santa Marta. Upon each mountain
system, one ascends from Tropical lowland, forest or prairie ("Sa-
bana"), through Subtropical forest, through Temperate forest or
"Sabana," to the "Paramo," as the treeless chill slopes above
timber-line are called. Temperature and moisture cause the floras
of the different zones to differ, and similarly the isolation of the
different mountain systems accounts for a divergence in the floras
of the same zone on each chain.

Dr. Frank M. Chapman, in his "Distribution of Bird Life in
Colombia," has given us a masterly presentation of this problem,
and I adopt his system of life zones and his terminology of each.
As stated in his summary on page 85 of volume 36 of the Bulletin
of The American Museum of Natura1 History, these are:

Tropical Zone— sea-level to 4,500-6,000 ft. (1350-1800 meters).
Subtropical Zone— 4,500-6,000 ft. to 9,000-9,500 ft, (2,700-2,850

meters) .
Temperate Zone— 9,000-9,500 ft, to 11,000-13,000 ft. (3,300-3,900

meters) .
Paramo Zone— 11,000-13,000 ft, to snow-line, 15,000 ft. (4,500

meters).

140 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

My observations have ied me to occasional slight modifications
of his limits, as in placing the lower limit of the Paramo above
Bogota at only 3100 meters. Here local conditions explain such
a change. But in general outline, and in nearly all details, his
scheme may be adopted for plants as for animals. Plants seem
more subject to geographical differentiation than animals, so that
in certain genera the same zone on different Cordilleras has related,
but never the same, species. A fuller discussion of this subject will
be given in the concluding paper of this series — now I wish but
to make the summary of distribution accompanying each species
of this study intelhgible.

The vegetation of a land so diverse as Colombia is immensely
rich; consequently upon a short expedition it has been impossible
to foLow out in comparative study any wide number of families.
A few groups well selected, and so far as possible all their species
considered, will give data for geographical botany nearly as precise
as would the comparing of many families. I have studied the
Scrophulariaceae, keeping a record of each species, and making a
careful floral description of each. For the Temperate and Paramo
zones, and for the Tropical prairies, this family furnishes an excel-
lent index to floral areas. My collections were mainly in these
regions, and of herbaceous plants, so that from a phytogeographical
viewpoint, the selection of this family has been justified. It is my
hope to follow this study with that of some allied group requiring
forest environments, probably of the Gesneriaceae.

The present paper gives the results of a study of only about one-
half the Scrophulariaceae of Colombia — those which we may cail
the Antirrhinoid genera. These fall into several tribes, each pre-
dominant or restricted to a special life-zone. The Gratioleae are
mainly Tropicai, while wholly so are the Russelieae and Angelonieae.
The Mimuleae, Hemimerideae and Fagelieae, the last with many
conspicuous species, predominate in the Temperate Zone of the
mountain-slopes. The genus Bartsia, of the Rhinanthoid Scrophu-
lariaceae and so yet to be studied, is best developed in the Paramo
Zone.

Necessarily the chief basis of this study has been my own collec-
tions. I have however revised all the collections from Colombia
which I know to be in the United States. These are surprisingly
meagre, and from widely scattered localities. The chief are those
made by H. H. Smith in the Sierra Nevada de Santa Marta in
1899-1901; by I. F. Holton near Bogota and in Valle in 1852-1856;

1920.] NATURAL SCIENCES OF PHILADELPHIA. 141

and — more imperfect— certain series of specimens made in southern
Colombia by F. C. Lehmann and at widespread stations over the
country by Jose Triana. To the custodians of the herbaria which
have loaned me specimens, the United States National Museum, Gray
Herbarium, and Field Museum of Natural History, as well as to my
colleagues at The New York Botanical Garden, I am under obligation.

Also, I would mention my indebtedness to that group of Colom-
bian workers whom I had the privilege of meeting in 1917, the
growth of whose museum at Bogota has been phenomenal. Es-
pecially would I thank Brothers Ariste-Joseph and Ydinael, Her-
manos Cristianos of the Universidad de La Salle. A further word
of appreciation must be given to Sr. Santiago Cortes of Bogota,
who, working long alone, has been able to give to the world only
the first -volume of his "Flora de Colombia."

The following study is primarily systematic, and keys are given
throughout. New species and those seen by the writer are care-
fully described. Synonomy for Colombia is cited fully, otherwise
only those names are included which are of first descriptions of
Colombian species. The original statement of distribution, or of
type-specimen is quoted, and the effort is made to firmly estab-
lish the nomenclature used. But it must be said that, as most of
the types are in Europe and inaccessible to me, and as Colombia's
flora is as yet very partially known, we cannot be certain of the
identity of some of these. But I believe that, with very few excep-
tions, the names now used will be permanent.

For each species a statement of environment and distribution
is given, the latter made as definite as our knowledge permits, and
analysed accordingly to life zone and Cordillera. Lastly is given
a list of specimens seen, these grouped under the Departments which
at present (1920) are in force. State-outlines in Colombia have
been so shifting that these limits have not always been easily ascer-
tainable. The herbaria in which specimens may be consulted are
indicated by the symbols:

A — The Academy of Natural Sciences of Philadelphia.

C — Field Museum of Natural History, Chicago, 111.

H — Gray Herbarium of Harvard University, Cambridge, Mass.

U— United States National Museum, Washington, D. C.

Y — New York Botanical Garden, Bronx Park, New York City.

My own collections may all be consulted at The New York Bo-
tanical Garden. Duplicates are being distributed to many her-
baria.

142 PROCEEDING* OF THE ACADEMY OF [May-Oct.,

At the conclusion of the systematic portion of this study there
is planned a synopsis of the geographic distribution of the Scrophu-
lariaceae of Colombia, and also a series of brief sketches of those
collectors in Colombia to whom reference will have been made in
the text.

Key to Axtirrhixoid Scrophulariaceae of Colombia.1

Corolla with the posterior lobes external in the bud.

I AXTIRRHIXOIDEAE.)

Capsule septicidal, or loculicidal by a~ simple median split, the
septum breaking from the capsule-wall or rupturing. Cor-
olla not spurred. Leaves opposite or whorled in threes
(except in Capraria).
Corolla, even if saccate anteriorly, without a horn-like process
at the base of the anterior lobes. Capsule septicidal, or
loculicidal. Seeds, if reticulate, with lines not raised' or
wing-like.
Stigma two-lipped.

Capsule septicidal, or secondarily also somewhat loculicidal,
splitting to base; placentae simple. Sepals distinct
or nearly so (except in Vandellia and Torenia). Leaves,
or rarely only the capsule, somewhat glandular-punc-
tate. Inflorescence simply racemose (if several pedi-
cels are in one axil, then no common peduncle is
evident). Corolla yellow, blue or white.

I. Gratioleae.
Capsule loculicidal (only tardily septicidal if at all), or
indehiscent ; placentae branched and widely spreading.
Sepals united over one-half length. Leaves and cap-
sule not glandular-punctate. Inflorescence racemose,
or of axillary cymes a single one of which is terminal
to the primary peduncle. Corolla yellow, two-ridged
and pubescent within on the anterior side.

II. Mimuleae.
Stigma capitate.

Corolla conspicuously zygomorphic, the tube scarcely de-
veloped and the anterior lobes much exceeding the
posterior ones. Capsule without placental hairs, and
dehiscing only distally. Seeds ridged, not reticulate.
Leaves opposite, or the upper alternate.
Stamens four; anther-sacs with membranous walls. Cor-
olla orange, flattened, its lobes all evident, the tube
split to base between the posterior lobes. Sepals
five, less than one-half the length of the capsule.

lThe warning must be given that the keys to tribes and genera are prepared
for Colombian species, and contrasts may not hold for extra-limital genera and
species.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 143

Capsule scarcely dehiscing loculicidally. Seeds
blackish. Inflorescence simply racemose. Stem
quadrangular. III. Hemimerideae.

Stamens two; anther-sacs with firm walls. Corolla yel-
low, its lips concave-saccate or the posterior much
reduced, the individual lobes scarcely or not evi-
dent, the tube not split to base between the posterior
lobes. Sepals four, at least one-half the length of
the capsule. Capsule dehiscing Joculicidally as well
as septicidally. Seeds brown. Inflorescence cy-
mose, two pedicels of each cyme being terminal to
the primary peduncle. Stem terete or nearly so.

IV. Fagelieae.
Corolla red, nearly regular, tubular, the short lobes nearly
equal. Capsule filled with slender hairs between which
are the scattered seeds, dehiscing to base septicidally.
Seeds reticulate, not ridged. Leaves whorled in threes,
and the stem with six angles. Inflorescence of axillary
cymes, two pedicels of each being terminal to the
primary peduncle. V. Russelieae.

Corolla violet-blue, saccate anteriorly and with a fine horn-
like process at the base of the anterior lobes. Capsule
loculicidal. the septum only tardily if at all splitting sagit-
tally. Seeds reticulate, the reticular lines raised into
wing-like processes. Inflorescence simply racemose.

VI. Angelonieae.
Capsule loculicidal, the septum and adjacent capsule-wall per-
sisting, the remaining wall splitting irregularly. Corolla
blue, with a spur at the base of the anterior petal. Leaves
alternate. VII. Axtirrhixeae.

Corolla with the antero-lateral or anterior lobes external in the bud.

(RHINANTHOIDEAE.4)
I. GRATIOLEAE.

Leaves alternate, serrate. Stamens five. Corolla essentially

regular, the five lobes equally distinct. 1. Capraria.
Leaves opposite or whorled in threes. Stamen- four, three or
two (the posterior one lost). Corolla more or less zygo-
morphic, the two posterior lobes united over one-half*their
length.
Leaves entire to serrate-dentate. Capsule globose to lanceo-
late in outline. Seeds not regularly cylindric nor spirally
ridged, disposed in more than one row within each valve
of the capsule.
Corolla with the ridges to the antero-lateral sinuses, if
developed, low and not projecting beyond those points
(so anterior filaments simple). Style not with a semi-
persistent callose base. Septum rupturing, so that the

2 To be considered in a second paper.

144 . PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

placental mass eventually stands free. Pedicels fre-
quently bibracteolate. Stem, if quadrangular, with
the angles not conspicuously ridged or winged.
Anther-sacs proximate, no connective arms developed,
Seeds reticulate. Leaves sessile or nearly so.
Pedicels bibracteolate. Sepals five.

Pedicels bibracteolate at base (remote from the
calyx). Corolla yellow, pubescent within at
base of posterior lobes. Sepals unequal, and
leaves serrate. Plant repent-ascending.

2. Mecardonia.
Pedicels bibracteolate at apex (just beneath calyx).

Corolla violet-blue or white.
Filaments four, all with anthers. Bractlets 1
mm. or less long, much shorter than the
sepals.
Sepals unequal, the outer much larger than
the narrow innermost. Corolla pubescent
within at base of posterior lobes, or gla-
brous, violet-blue or white. Capsule glo-
bose-ovoid to oblong. Leaves serrate to
entire, and stems, erect or ascending.

3. Caconapea3
Sepals uniform. Corolla pubescent within at

base of the anterior lobes. Capsule de-
pressed-globose. Leaves serrate and stem
repent-ascending. 4. Conobea.

Filaments two, the anterior rudimentary or want-
ing. Bractlets 5-10 mm. long, equaling or
exceeding the nearly uniform sepals. Cor-
olla pubescent at base of the posterior lobes,
white or pinkish-tinged. Leaves serrate and
stem ascending or erect. 5. Gratiola.
Pedicels not bracteolat'e. Sepals four or five. Cor-
olla blue or white.
Corolla glabrous within. Sepals unequal, the inner-
most narrowest. Leaves palmately veined,
entire or slightly undulate. Pedicels tending
to deflex in fruit. Plants repent.
Outermost sepal cordate, much exceeding the
linear-attenuate innermost ones; five sepals
always present. Capsule oblong or ovoid-
oblong, acute, brown, much shorter than the
sepals. Styles united to apex. Corolla blue
or white. 6. Monocardia.

3Branna monnieri (L.) Pennell, a repent herb, with broadly rounded entire
leaves, corolla with distinct posterior lobes, and outer sepal scarcely longer than
the innermost, is widespread in Tropical America, and must surely o'ccur on moist
semi-brackish sands along the Colombian coast. See, Proc. Acad. Nat. ScL, Phila.
1919: 243, 1920.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 145

Outermost sepal ovate-oblong, scarcely longer than
the lanceolate innermost, one of which may
be lost. Capsule nearly globose, obtuse, pale-
brown, little shorter than the sepals. Styles
distinct near apex. Corolla white.
Corolla 4 mm. long, appearing four-lobed be-
cause the three petals forming the anterior
lip are all evident. Stamens four. Sepals
five or four. 7. Macuillamia.

• Corolla 2 mm. long, appearing three-lobed be-

cause the anterior petal is lost, leaving the
anterior lip two-lobed. Stamens three
(only one of the antero -lateral pair devel-
oped). Sepals four. 8. Hydranthelium.
Corolla white, densely hirsute within over bases of
all lobes. Sepals four, uniform. Leaves pin-
nately veined, serrate-dentate. Pedicels per-
manently ascending -spreading. Plant erect.

9. Scoparia.
Anther-sacs separated on short arms of the connective.
Seeds longitudinally striate, the striae frequently
tuberculate. Corolla blue or white. Plants erect.
Pedicels bibracteolate, 1 mm. long or less. Corolla
blue, pubescent within on the anterior side. Cap-
sule acuminate. Seeds tuberculate-striate. Leaves
cordate-clasping at base. Tall herb.

10. Stemodia.
Pedicels not bracteolate. Corolla pubescent within

on the posterior side, or glabrous. Seeds min-
utely roughened-tuberculate or smooth. Leaves
narrowed at base.
Leaves sessile or nearly so, in whorls of three. Cor-
olla 9-13 mm. long, glabrous within, blue.
Capsule acuminate. Seeds minutely roughened-
tuberculate. Style semi-persistent. Tall herb.

11. 1' i(<ni iiea.
Leaves evidently petiole* I. opposite. Corolla 4-8

mm. long, pubescent within on the posterior
side, blue or white. Capsule acutish. Seeds
smooth, with rounded ridges. Style soon de-
ciduous. Low herbs. 12. Lendneria.
Corolla violet-blue or white, with two raised ridges (each
formed by the adherence of a filament) to the antero-
lateral sinuses, and which frequently project as knob-
like processes beyond those points (the free portion
of the filament appearing as a lateral outgrowth of
the adherent portion). Style with a semi-persistent,
frequent lv enlarged and callose base. Septum per-
sistent, with the attached placentae. Pedicels never

146 PROCEEDINGS OF THE ACADEMY OF [May~Oct.,

bracteolate. Stem quadrangular, the angles ridged or
slightly winged.
Sepals united over one-third length. Filaments all with
anthers. Seeds not with transverse lines. Leaves
petioled, serrate-dentate. Angles of stem slightly
winged.
Pedicels 1-2 mm. long. Sepals united nearly one-
half their length, much shorter than the capsule.
Corolla straight, the tube strongly horizontally
flattened, the posterior lip purple-brown, else-
where corolla white throughout, pubescent within
on ridges to antero-lateral sinuses; the free por-
tions of the anterior filaments appearing as up-
curved from the apices of these ridges. Capsule
accuminate, finely pubescent to glabrous. Seeds
tuberculate. Leaves oval, rounded, narrowed at
base. Plant repent, pubescent.

13. Vandellia.
Pedicels 10-25 mm. long. Sepals united over three-
fourths length, equaling or slightly longer than
the capsule. Corolla decurved, the tube scarcely
or not flattened horizontally, blue or white, gla-
brous within on the ridges to antero-lateral sin-
uses; the free portions of the anterior filaments
appearing as outgrowths proximad to the apices
of these ridges which therefore terminate as short
knobs. Capsule mucronate or acute, glabrous.
Seeds shallowly pitted. Leaves ovate or lanceo-
late-ovate, acute, truncate-cuneate at base. Plants
repent to erect, glabrous. 14. Torenia.

Sepals distinct or nearly so. Antero-lateral filaments
without anthers. Seeds with fine transverse lines.
Leaves sessile, mostly clasping, slightly crenate or
entire. Angles of stem ridged, not winged. Corolla
blue. Plant diffused-ascending, glabrous.

15. Ilysanihes.
Leaves pinnatifid. Capsule linear-attenuate in outline.
Seeds regularly cylindric, spirally ridged, disposed in one
row within each valve of the capsule. Corolla purple-
blue. Plant low, branched, erect. 16. Schistophragma.

II. MIMULEAE.

Capsule oblong, dehiscing loculicidally, its walls dry, mem-
branous, brownish. Calyx-tube longer than and enclos-
ing the capsule, its ribs slightly winged, its lobes decidedly
unequal, the posterior longest. Corolla 10-15 mm. long,
yellow, with many red-brown spots within throat on the
anterior side. Leaves broadly ovate, shortly petioled, 1-4
cm. long. Inflorescence of axillary simple pedicels. Stem
obscurely angled, not winged. Plant low, repent-ascend-
ing. 17. Mimulus.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 147

Capsule globose, indehiscent, its walls fleshy, white. Calyx-
tube shorter than the capsule, its ribs not winged, its lobes
slightly unequal, the posterior longest. Corolla 15-18 mm.
long, yellow throughout. Leaves lanceolate, cordate-
clasping at base, 15-20 cm. long. Inflorescence of axillary
several-branched cymes, borne upon conspicuous peduncles.
Stem with angles narrowly winged. Plant tall, erect,
widely branched from base. 18. Leucocarpus.

III. HEMIMERIDEAE. 19. Alonsoa.

IV. FAGELIEAE. 20. Fagelia.

V. RUSSELIEAE. 21. Russelia.

VI. ANGELONIEAE. 22. Angelonia.

VII. AXTIRRHINEAE. 23. Linaria.

1 CAPRARIA Linne.
Capraria L., Sp. PI. 628. 1753.

Type species, Capraria biflora L.

1. Capraria biflora L.

Capraria biflora L., 1. c. 628. 1753. "Habitat in Curassao." Specimens

from Curacao seen in herbarium New York Botanical Garden.
capraria lanceolata Vahl, Eel. Am. 2: 47. 1798. "Habitat ad St. Mar-

tham. von Rohr." Not C. lanceolata L. f.,Suppl. 284. 1781. Von

Rohr's plant was the pubescent form prevalent in Colombia, but appears

to have represented an abnormal state in which the pedicels are short

and arise from an abbreviated branch.
Capraria semiserrata Willd., Sp. PL 3: 324. 1800. Xew name for C.

lanceolata Vahl.
Capraria semiserrata berterii A. DC; Benth. in DC. Prod. 10: 429. 1846.

•Ad Stam. Martham (Bert,! in h. DC.)." An entire-leaved state.

A widespread species of lowland Tropical America, growing mostly
on and near the seashore, but extending inland on river-banks,
plains and waste land. Very variable, but with most states found
in the same region or even in the same colony. Varies in size of
leaves, in form of leaves from linear-lanceolate to nearly ovate, and
in serration from entire to sharply serrate-dentate; varies in the
length of the pedicels; varies in length of and attenuation of the
sepals, in size of corolla from 8 to 10 mm. long, and somewhat pubes-
cent or glabrous within anteriorly, and in capsules from oblong to
ovoid, and from round and emarginate to acutish. Most con-
spicuously does the plant vary in pubescence, from glabrous through-
out, through states with the stem pubescent and the pedicels glabrous
or the pedicels sparsely pubescent — pubescent either with short or
long hairs — to forms densely hirsute on stems, pedicels, sepals and

148 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

the midribs and margins of the leaves. The last state (forma hirta*
Loes. in Bull. Herb. Boiss. ser. II, 3: 284. 1903. "Habitat in Guate-
mala, in dept. Chiquimula in ruderalibus ad S. Juan Ermita: Sel[er]
n. 3314." Isotype seen in herbarium New York Botanical Garden)
prevails in Colombia.

River-banks, flats and sandy waysides, at altitudes below 200
meters, Tropical zone; the typical form near the Caribbean shore
(doubtless also along the Pacific shore), forma hirta Loes. (indicated
irr lists by asterisk) along the lower river-courses and across the
Sabana of Bolivar.

Antioquia. *Brazaela de Perales, on Rio Magdalena (river-flat,
alt. 150 m.), Pennell 3704; *Vuelta de Acufia, on Rio Magdalena
(sandy loam, alt. 125-130 m.), Pennell 3789.

Bolivar. "Calamar (along Rio Magdalena, alt. 15-25 m.), Rusby
& Pennell 20; Cartagena, I. F. Holton 582 (H, Y), (roadside, alt.
5-10 m.), Rusby & Pennell 3 (somewhat pubescent with short
hairs). *Since (edge of thicket, alt. 120-170 m.), Pennell 4039;
*Turbaco (thin loam over white rock, alt. 150-200 m.), Pennell
4761; *Vilches, on Rio Sinn (orchard, alt. 20-50 m.), Pennell 4711.

Magdalena. *Bonda (open damp place, alt. 45 m.), H. H. Smith
1331 (C, H, U, Y). Don Jaco (near the coast), H. H. Smith 551
(C, H, U, Y). Playa Brava (open plain near the coast), H. H.
Smith 2823 (Y). Santa Marta (railroad bank, alt. 0-10 m.), Pen-
nell 4771. Around Rio Frio (between the Cienaga de Santa Marta
and the foothills, alt. 0-100 m.), H. Pittier 1611 (U).

2 MECARDONIA Ruiz and Pa von.

Mecardonia R. & P. [Fl. Peruv. et Chil. Prod. 95. 1794, generic diagnosis];
Syst. Veg. Fl. Peruv. et Chil. 164. 1798.

Type species, M. ovata Ruiz & Pavon,5 of Peru.

4 This state has been described also as Capraria lanceolate, Vahl, and C. semi-
serrata Willd. (above), as C. hirsute H. B. K., Nov. Gen. et Sp. 2: 355. 1818,
from Mexico, and as C. biflora /3 pilosa Griseb., Fl. Brit. W. I., 427, 1861, from
the Bahamas. The first name should be used if this be accounted a distinct
species, the last if a variety, and hirta if a form.

5 Mecardonia ovata Ruiz & Pavon, Syst. Veg. Fl. Peruv. et Chil. 164. 1798.
"Habitat in Peruviae plateis ad Chinchao et Cuchero vicos." Description
appears to be certainly that of M. procumbens (Mill.) Small, a wide-spread plant
to be expected in Peru. The only discrepancy is in describing the calyx as hepta-
phyllous, although it is stated that the two small outer leaves are deciduous.
Possibly the bractlets at the base of the pedicel were noted approximating the
flower in very young buds, or more probably an error of vision was made, and
because later it could not be checked the imagined bractlets were considered
to be deciduous.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 149

1. Mecardonia procumbens (Mill.) Small.

Erinus procumbens Mill., Gard. Diet. ed. VIII. n. 6. 1768. "Houst.

MSS."
Herpestis caprarioides H. B. K., Nov. Gen. et Sp. 2: 368. 1818. "Crescit

locis arenosis, siccis Regni Novo-Granatensis in ripa fluminis Magdalenae

juxta El Pefion [Humboldt & Bonpland]."
Monniera procumbens (Mill.) Kuntze, Rev. Gen. 463. 1891.
Mecardonia procumbens (Mill.) Small, Fl. S. E. Un. St. 1065. 1338. 1903.
Bacopa procumbens (Mill.) Greenm. in Field Col. Mus., Bot. Ser. 2: 261.

1907.
Herpestis procumbens (Mill.) Urb., Symb. Bot. 4: 558. 1911.

Moist open soil, along streams in loam or sand, at altitudes below
1200 meters. Tropical zone, doubtless throughout lowland Colom-
bia, largely as a weed. Also in the Sabana of Bogota, at an altitude
of 2600 meters, probably introduced. Widespread throughout
Tropical America.

Antioquia. Opposite Boca Carare (forest along R. Magdalena,
alt. 125 m.), Pennell 3829.

Bolivar. Boca Verde on Rio Sinu (field along river, alt. 100-200
m.), Pennell 4568; Canabetal (sand along river, alt. 90-100 m.),
Pennell 3874; Sincelejo (thicket, alt. 150-200 m.), Pennell 4068.

Caldas. Buena Vista (moist open sand, alt. 170-180 m.), Rusby
& Pennell 63.

Cundinamarca. Bogota (desiccated soil in meadow, alt. 2600
in.), Pennell 1909 [small-leaved form which has been described as
Herpestis chamaedryoides H. B. K.]; Icononzo (loam slope, alt. 900-
1000 m.), Pennell 2777.

Huila. Natagaima (soil frequently overflowed by river, alt. 400 -
450 m.), Rusby & Pennell 1182; Quebrada de Angeles above Nata-
gaima (gravel, alt. 450-600 m.), Rusby & Pennell 286.

Magdalena. Masinga (damp clearings, alt. 90-750 m.), H. H.
Smith 1326 (Y), Agua Dulce, H. H. Smith 1326 (Y).

Meta. Villavicencio (moist meadow along Rio Guatiquia, (alt.
500 m.), Pennell 1556.

Tolima. Honda (moist sandy field, alt. 250-300 m.), Pennell
3620; Libano (field on "La Trinidad," alt. 900-1200 m.), Pennell
3368.

Valle. La Paila, I. F. Holton 579 (Y).

3. CACONAPEA Chamisso.
Caconapea Cham, in Linnaea 8: 28. 1833.
Type species C. gratioloides Cham., of Brazil.
Pedicels 5-10 mm. long, one to an axil. Corolla 5-7 mm. long,
pubescent or puberulent within over base of posterior lobes;
lobes violet, throat yellow within, especially on the anterior

150 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

side and lined with violet, the posterior lobes united nearly to
apex.
Corolla 7 mm. long, pubescent within. Anthers all uniform.
Sepals strongly dissimilar, the outer ovate, all puberulent.
Capsule septicidal, with thick protuberant placentae, in de-
hiscence portions of the septum remain adhering t3 the walls.
Seeds conspicuously reticulate. Leaves lanceolate-elliptic,
crenate-serrate. Stem sharply quadrangular, 1-3 dm. tall.

1. C. auriculata.
Corolla 5 mm. long, puberulent within. Anthers of posterior
filaments larger. Sepals slightly dissimilar, the outer lanceo-
late, all somewhat pubescent. Capsule loculicidal, with thin
placentae, placed along median line of the septum which
ultimately breaks free from the lateral walls. Seeds finely
reticulate. Leaves linear, entire. Stem obscurely quad-
rangular, .3-.5 dm. tall. 2. C. debilis.
Pedicels less than 2 mm. long, usually several to an axil. Corolla
2-3 mm. long, glabrous within; lobes white or violet, throat
white and without lines. Placentae thin, attached to median
line of septum. Seeds finely reticulate.
Leaves tapering to the narrow base. Stem pilose with spreading

white hairs. 3. C. axillaris.

Leaves rounded-clasping at base. Stem appressed-pubescent or
glandular-dotted.
Corolla deciduous, white throughout, its posterior lobes united
nearly to apex. Exterior sepals rounded, all glabrous or
obscurely puberulent. Stem whitish with appressed re-
flexed hairs. 4. C. appressa.

Corolla persistent, lobes violet-blue, its posterior lobes united
only three -fifths length. All sepals acute, each with a
tuft of white hairs at apex. Stem yellowish with sessile
glands. 5. C. conferta.

1. Caconapea auriculata (Rob.) Pennell, comb. nov.

Herpestis auriculata Rob. in Proc. Am. Acad. 26: 172. 1891. "Wet places
near Guadalajara [Mexico]; November, 1889 [C. G. Pringle] (n. 2937)."
Later collections of Pringle from near Guadalajara, 4623 and 6148, seen
in herbarium of Columbia University at The New York Botanical Garden.

Bacopa auriculata (Rob.) Greenm. in Field Col. Mus., Bot. Ser. 2: 262.
1897.

Stem 1-3 dm. tall, sharply four-angled, glabrous to slightly puberu-
lent above. Leaves 1-1.7 cm. long, .5-. 8 cm. wide, oblong-lanceo-
late, crenate, often dentate at and near base, obtuse, rounded-clasp-
ing at base; glabrous or glabrate. Pedicels solitary, 5-10 mm. long,
puberulent. Bractlets linear-subulate, less than 1 mm. long.
Sepals obscurely puberulent: outermost 4 mm. long, ovate, obtuse
to acute, obscurely veined; two median slightly shorter and more
narrowly ovate; two innermost 3 mm. long, linear-attenuate,

1920.] NATURAL SCIENCES OF PHILADELPHIA. 151

somewhat hyaline. Corolla 7 mm. long; posterior lobes united
nearly to apex; pubescent within tube, especially on anterior
side, pubescent in horizontal line over bases of posterior lobes;
within tube yellow, extending to base of anterior lobes, distally
and over lobes violet-blue, with longitudinal fine deep-violet
lines. Filaments glabrous, the postero-lateral pair slightly shorter,
its anthers equalling those of the antero lateral pair; posterior
filament represented by a tiny knob. Style glabrous, at apex bifid
and bearing two plate like stigmas. Capsule 2.5-3 mm. long,
globose -ovoid, glabrous, dehiscing septicidally, the lateral portions
of the septum adhering to the capsule wall. Placentae pro-
truding into the cells, not coalescent. Seeds .4 mm. long, cres-
centic-cylindric, truncate at the apex, brown, with evident longi-
tudinal lines and cross reticulations.

Wet open soil, sandy or loam, at altitudes of 500 to 800 meters,
Tropical zone, llanos of upper Magdalena valley. Doubtless in the
Sabina of Bolivar and elsewhere in northern Colombia. Ranges
northward to Mexico.

Cundinamarca. Melgar (moist open clayey loam, alt. 500-600
m.), Pennell2885.

Huila. Neiva (seepy place in plain, alt. 550-600 m.), Rusby &
Pennell 1067.

Tolima. San Lorenzo (open springy loam, w. of, alt. 600-800
m.), Pennell 3531.

2. Caconapea debilis Pennell, sp. nov.

Stems ascending, repent and matted below, becoming erect and
.3-.5 dm. tall, obscurely four-angled, finely pubescent with refiexed
appressed white hairs. Leaves .4-.6 cm. long, .07-. 1 cm. wide,
linear, obtuse, clasping (but not dilated) at base, nerveless beneath;
finely pubescent on the midrib beneath or glabrous throughout.
Pedicels solitary, 5-7 mm. long, finely pubescent with refiexed hairs.
Bractlets linear-subulate, less than .5 mm. long. Sepals: outermost
3.5 mm. long, linear-lanceolate obtuse; two median slightly shorter
but nearly similar, these three green, glabrous or nearly so, except
for a frequent terminal tuft of short hairs; the two innermost 3
mm. long, lanceolate-linear, acuminate, with broad scarious margins
and ciliate with minute white hairs. Corolla 5 mm. long; posterior
lobes united nearly to apex; externally minutely pubescent, within
slightly pubescent over bases of the posterior lobes, elsewhere
glabrous; within tube pale, yellowish on anterior side, lined with

152 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

violet, distally over lobes violet. Filaments glabrous, the pos-
tero-lateral pair shorter, its anthers more than twice larger than
those of the antero-lateral pair. Style glabrous, with two plate-like
stigmas. Capsule 2 mm. long, oblongpyriform, rounded and retuse
at apex, glabrous, dehiscing loculicidally. Placentae thin, linear,
flattened against the persistent septum. Seeds about .2 mm. long,
oval, slightly crescentic, yellowish-brown, obscurely ridged to some-
what reticulate at maturity.

Type, shallow pool in llano, east of Villavicencio, Meta, altitude
about 450 meters, collected in flower and fruit September 2, 1917,
F. W. Pennell 1623 in herbarium New York Botanical Garden.

Nearest to Herpestis reptans Benth. of Brazil, which is described
as having leaves .2-3 cm. wide, linear-lanceolate, prominently
nerved beneath, all sepals obtuse and seeds yellow.

Shallow pools in llano, at an altitude of 450 meters, Tropical
zone in Meta.

3. Caconapea axillaris (Benth.) Pennell, comb, now

Herpestis axillaris Benth. in DC. Prod. 10: 396. 1846. "Ad aquas stag-
nates in campis Deluvia Sanctae Marthae (Purdie!) ... (v. in
herb. Hook.)."

Monniera axillaris (Benth.) Kuntze, Rev. Gen. 463. 1891.

Tropical zone, in Magdalena. No specimens seen.

4. Caconapea appressa Pennell, sp. nov.

Stem .7-1.5 dm. tall, obscurely four-angled, below glabrous, above
pubescent with reflexed white hairs. Leaves 1-2 cm. long, .3-4
cm. wide, lanceolate-linear, entire, tapering from the rounded-
clasping base, glabrous, glandular-dotted. Pedicels 1-5 in an axil,
1-2 mm. long, pubescent with reflexed hairs. Bractlets filiform-
subulate, less than 1 mm. long. Sepals glabrous, densely glandular-
punctate: outermost 3 mm. long, ovate; the next nearly as wide, the
median one-sided, narrower, these three obtuse, somewhat reticulate,
glabrous; the two innermost narrowly lanceolate, attenuate, costate,
with the margins hyaline and ciliolate. Corolla 2-2.5 mm. long;
posterior lobes united nearly to apex; glabrous throughout; white
throughout. Filaments glabrous, the postero-lateral pair slightly
shorter and its anthers slightly smaller. Style glabrous, bearing two
approximate stigmas. Capsule 2 mm. long, narrowly cylindric
oblong, glabrous, dehiscing septicidally and loculicidally, none of the
septum adhering to the capsule wall. Placentae narrow, thin, a little
raised line median to the persistent broad septum. Seeds .5 mm.
long, cylindric, tapering to each end, brown, with longitudinal ridges
and fine cross lines.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 153

Type, moist depression in llano east of Villavicencio, Meta, alti-
tude about 450 meters, collected in flower and fruit, August 28 and
September 2, 1917, F. W. Pennell 1460, in herbarium New York
Botanical Garden.

Moist soil, in llanos, at an altitude of 450 meters, Tropical zone,
in Meta.

5. Caconapea conferta Pennell, sp. nov.

Stem .5-1.8 dm. tall,' obscurely four-angled, below glabrous, above
puberulent with sessile yellowish glands. Leaves 1-2 cm. long,
.2-.4 cm. wide, narrowly lanceolate, tapering from the rounded-
clasping base, entire, densely glandular-dotted. Pedicels 1-5 in an
axil, less than 1 mm. long, puberulent with sessile glandular dots.
Bractlets filiform-subulate, less than 1 mm. long. Sepals densely
glandular-puberulent, each tipped with a tuft of white hairs : outer-
most 3 mm. long; narrowly ovate; two median narrower and one-
sided, these three acuminate; two innermost 2.5 mm. long, lanceolate-
attenuate, somewhat hyaline-margined and more ciliolate. Corolla
3 mm. long; posterior lobes united f length; glabrous throughout; its
tube white, lobes violet-blue, darker distally within. Filaments
glabrous, bluish, the postero-lateral pair slightly shorter; anthers all
of about the same size, light-yellow. Style glabrous bearing two
stigmas. Capsule 2 mm. long, ellipsoid-oblong, glabrous, dehiscing
septicidally and loculicidally, none of the septum adhering to the
capsule-wall. Placentae narrow, thin, median to the persistent broad
septum. Seeds .3 mm. long, cylindric, slightly irregularly curved,
rounded, brown, with fine longitudinal ridges.

Type, moist depression, in llano east of Villavicencio, Meta, alti-
tude about 450 meters, collected in flower and fruit August 28 and
September 2, 1917, F. W. Pennell 1435; in herbarium New York
Botanical Garden.

Moist soil in llanos, at an altitude of 450 meters, Tropical zone,
in Meta.

4. CONOBEA Aublet.
Conobea Aubl., Hist. PI. Guiane Fr. 2: 639. pi. 258. 1775.
Type species, C. aquatica Aubl., of Guiana.

1 . Conobea scoparioides (C. & S.) Benth.

Sphaerotheca scoparioides Cham. & Schlechtd. in Linnaea 2: 606. 1827.

"E Brasilia acquinoctiali misit Sellow."
Conobea scoparioides (C. & S.) Benth. in DC. Prod. 10: 391. 1846.

154 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Tropical zone, collected only in Choco, but doubtless widespread
in northern and eastern Colombia. Widespread through eastern
lowland South America.

Choco. Novisa, J. Triana (H, Y).

5. GRATIOLA Linne.
Gratiola L., Sp. PI. 17. 1753.
Type species, G. officinalis L., of Europe.

1. Gratiola bogotensis Cortes, sp. nov.

Spreading extensively by rootstocks. Aerial stems erect or de-
cumbent at base, succulent, slightly puberulent, 1-4 dm. tall. Leaves
oblong-lanceolate, 1.5-2.5 cm. long, .3-. 8 cm. wide, clasping by a
rounded base, distally dentate and glandular-punctate, finely puberu-
lent to glabrate. Pedicels 1-2 mm. long. Bractlets similar to and
equaling or slightly exceeding the lanceolate calyx-lobes, 5-10 mm.
long. Corolla 12-14 mm. long, its tube four-angled, yellowish,
with fine brown lines, its lobes spreading, white, somewhat pinkish-
tinged or at times the corolla strongly pink. Antero -lateral fila-
ments evident, each capped by two small rudimentary anthers.
Capsule ovoid, 5 mm. long. Seeds .5 mm. long, obovoid, conspicu-
ously alveolate-reticulate.

Type, wet grassy place, base of mountain above Chapinero, near
Bogota, Cundinamarca, altitude 2700-2800 meters, collected in
flower and fruit September 23, 1917, Pennell 2108 in Herb. New
York Botanical Garden. This was collected in company with Sr.
Santiago Cortes, who designated it by the name here given.

A near ally or possibly geographical variety, of Gratiola 'peruvi-
ana L., Sp. PI. 17. 1753, based upon Feuillee's description and
drawing of a plant found in the mountains of Chile, at 26° S. L.
Feuillee's plant is well-matched by a specimen collected by Otto
Kuntze in February, 1892, at Ervilla, Chile, and which has the
stem less fleshy, the leaves broader and the pedicels slightly longer
than does our plant.

Wet grassy places, springheads and ditches, at altitudes of 2600
to 3200 meters, Temperate zone, ascending, in dwarf form, to Par-
amo, Cordillera Oriental, and Cordillera Central, southward at
least into Ecuador.

Cundinamarca. Bogota (ditch in meadow, alt. 2600 m.), Pen-
nell 1908, (wet grassy place near Chapinero, alt. 2700-2800 m.)
Pennell 2108, (southwest of Las Cruces, open spring-head, alt.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 155

2600-2700 m.) Pennell 2158; Sibate (wet loam, alt. 2700-2800 m.),
Pennell 2451; Ubaque (wet ditch, edge of paramo, alt. 3000-3200
m.), Pennell 1902, Zipaquira (Alt. Aquila, edge of pool, just below
paramo, alt. 3100 m.), Pennell 2532.

Narifio. Tuquerres (alt. 3000 m.), Triana (H, Y).

Tolima. Murillo (pool, alt. 2600-2800 m.), Pennell 3155.

6. MONO CARDI A6 Pennell. gen. nov.

Stems much branched, terete, repent, the apices ascending.
Leaves sessile, slightly clasping, oblong to orbicular, entire, pal-
mately-veined from base, obscurely glandular-dotted and not odor-
ous. Pedicels axillary, 6-20 mm. long, pubescent, not bracteolate.
Calyx of 5 very dissimilar sepals; outermost (posterior) heart-
shaped; two anterolaterals smaller, obliquely, or but one-half cor-
date-or truncate-ovate, three outer prominently reticulate; two pos-
tero-laterals (innermost) linear-attenuate, only % length of outer-
most and usually slightly longer than the capsule. Corolla 3-7
Mini, long, the widely-spreading lobes longer than the tube, the two
posterior lobes united nearly throughout; glabrous throughout, bine
or white. Stamens four, glabrous, didynamous (the posterior fila-
ments shorter and usually anthers smaller); anthers uniform, the
narrow sacs closely connivent, each opening its entire length. Style
glabrous, little exceeding the sepals. Stigma- distinct, flat. Cap-
sule 2-4 mm. long, much shorter than calyx, oblong or ovoid-oblong,
acute, brown, septicidal and loculicidal; the septum with adherent
thin placentae, persisting plate-like. Seeds .2 .3 mm. long, oblong,
blackish-brown, ridged and with cross-reticulations.

Type species, M. violacea Pennell.

Hydrotrxhi Small, in general aspect like this genus, differs in pos-
sessing a circle of bristles surrounding the ovary, pubescence within
over the bases of the corolla-lobes, two bract coles below the calyx,
and more conspicuous glands which exhale a strong aromatic odor.

Pedicels and sepals with spreading hairs. Stems and leaves beneath

pubescent. Corolla blue.

Corolla 6-7 mm. long. Calyx 5-7 mm. long. Leaves 1.2-1.8

cm. long, and nearly as wide. Stems densely hirsute, 2-3

dm. long. 1. M . violaa a.

Corolla 4-5 mm. long. Calyx 4-6 mm. long. Leaves .7-1.8 cm.

long, ovate-oblong. Stems pubescent, less than 1.5 dm.

long.

'Name from (l.6vo<;, one, and Kapsia, heart, in allusion to the single large cor-
date sepal.

156 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Leaves 1.2-1.8 cm. long, the younger pubescent along midrib

beneath. Sepals 5-6 mm. long. 2. M. lilacina.

Leaves .7-9 cm. long, the younger pubescent over the entire

surface beneath. Sepals 4-5 mm. long. 3. M. humilis.

Plant glabrous throughout. Corolla white, 3-3.5 mm. long.

4. M. albida.

1. Monocardia violacea Pennell, sp. nov.

Stems fleshy, 2-3 dm. long or longer, densely hirsute with yellow-
ish hairs. Leaves ovate- or oval -orbicular, 1.2-1.8 cm. long, 1-1.4
cm. wide, obtuse, pubescent beneath along midrib proximally,
distally glabrous and obscurely reticulate. Pedicels 10-20 mm.
long, hirsute with spreading hairs. Sepals all ciliate: the three
outer 5-7 mm. long, the innermost pubescent on the midrib. Cor-
olla 6-7 mm. long; tube yellow, distally purplish, the lobes violet.
Filaments violet-bluish; anthers white. Capsule 3-4 mm. long,
narrowly oblong. Seeds .3 mm. long, brown.

Type, wet loam, along trail in forest, near Villavicencio, Meta,
altitude 450 meters, collected in flower and fruit August 28, 1917,
F. W. Pennell 1480; in Herb. New York Botanical Garden. Only
specimen seen.

Wet loam in forest at an altitude of 450 meters. Tropical zone,
in Meta, and in Panama. Doubtless wide-spread in northern South
America.

2. Monocardia lilacina Pennell, sp. nov.

Stems slightly fleshy, .5-1.5 dm. long, pubescent with spreading
hairs. Leaves ovate-oblong, 1.2-1.8 cm. long, .8-1.1 cm. wide;
obtuse, pubescent beneath along midrib proximally. Pedicels 8-12
mm. long, pubescent with spreading hairs. Sepals all ciliate: the
three outer 5-6 mm. long, two innermost pubescent on the midrib,
shorter. Corolla 4-5 mm. long, -violet within throat, the lobes pale-
blue. Filaments of anterior stamens pale-blue, of posterior violet;
anthers pale-blue. Capsule 2.5-3 mm. long, ovoid-oblong. Seeds
.2-25 mm. long, blackish.

Type, wet loam, along trail in forest, near Villavicencio, Meta,
alt. 450 meters, collected in flower and fruit August 28, 1917. F. W.
Pennell 1476; in Herb. New York Botanical Garden.

Wet loam in forest, at altitudes of 450 to 500 meters, Tropical
zone, in Meta.

Meta. Villavicencio (wet trail in forest, alt. 500 m.), Pennell
1378, (alt. 450 m.), Pennell 1476, (wet place near Rio Guatiquia,
alt. 500 m.), Pennell 1547.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 157

3. Monocardia humilis PenneH, sp. now

Stems not fleshy, .2-1 dm. long, pubescent with spreading hairs.
Leaves ovate-oblong, .7-.9 cm. long, .3-.8 cm. wide, obtuse, at least
when young hirsute over entire under surface, not evidently reticu-
late. Pedicels 6-9 mm. long, pubescent with spreading hairs.
Sepals all ciliate: the three outer 4-5 mm. long, pubescent over
entire outer surface. Corolla 4-5 mm. long, violet within throat,
the lobes pale-blue. Filaments of anterior stamens pale-blue, of
posterior violet; anthers pale-blue. Capsule 2-3 mm. long, ovoid-
oblong. Seeds .2-.25 mm. long, blackish.

Type, sandy soil, seepy place in plain east of Neiva, Huila, alt.
550-600 meters collected in flower and fruit August 8, 1917, Rusby
& Pennell 1065; in Herb. New York Botanical Garden.

Wet open soil, sandy or loam, at altitudes below 800 meters,
Tropical zone, llanos of upper Magdalena valley, and in Panama.
Doubtless in the Sabana of Bolivar.

Huila. Neiva (sandy seepy place in plain east of N., alt. 550-
600 m.), Rusby & Pennell 1065.

Tolima. San Lorenzo (swale west of S. L., alt. 600-800 m.),
Pennell 3544.

4. Monocardia albida Pennell, sp. nor.

Stems not fleshy, .3-1.5 dm. long, glabrous. Leaves ovate-
oblong, .7-1.5 cm. long, .3-1 cm. wide, obtuse, glabrous. Pedicels
7-14 mm. long, glabrous. Sepals glabrous: the three outer 4-5 mm.
long. Corolla 3-3.5 mm. long, dull white throughout. Filaments
and anthers white. Capsule 2-3 mm. long, narrowly elliptic-oblong.
Seeds .2-.3 mm. long, blackish.

Type, wet loam, trail in forest, Yillavicencio, Meta, alt. 450
meters, collected in flower and fruit August 28, 1919, F. W. Pennell
1477; in Herb. New York Botanical Garden.

Wet loam in forest at altitudes of 450 to 500 meters, Tropical
zone, in Meta.

Meta. Yillavicencio (wet trail in forest, alt. 450 m.), Pennell
1477, (moist meadow near Rio Guatiquia, alt. 50C m.), Pennell
1555 (plant much smaller than 1477).

7. MACUTLLAMIA Raflnesque.

M acufflamia Raf. [Neogenyton 2.1825, generic description only.] Autik. Bot.
44, 1840.

Type species, Monniera rctundifolia Mich., of Blinois.

158 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

1. Macuillamia limosa Pennell, sp. nov.

Extensively repent. Stem slightly succulent, finely pubescent
rather densely so distally. Leaves 1.2-1.7 cm. long, 8-10 mm. wide
obovate-elliptic, entire, narrowed at base, rounded at apex, with
7 or 9 longitudinal veins. Pedicels slender, 10-15 mm. long, finely
pubescent, in flower ascending, soon reflexing below the leaf-like
bracts. Sepals 2.5-3 mm. long, obtuse, somewhat pubescent; two
outer ovate-oblong; two median lanceolate-oblong; the innermost
one narrower or wanting. Corolla 4 mm. long, the lobes spreading,
slightly longer than the tube, the two posterior united to apex;
glabrous throughout, white. Filaments white. Anthers purplish.
Styles distinct near apex; stigmas semi -capitate. Capsule 2-2.5
mm. long, nearly globose, obtuse. Seeds .6 mm. long, cylindric-
oblong, reticulate, brown.

Type, open pool in clayey loam, Melgar, Cundinamarca, altitude
400-500 meters, collected in flower and fruit December 4-5, 1917,
F. W. Pennell 2927; in Herb. New York Botanical Garden.

Open pools and ditches, in shallow still water, at altitudes below
500 meters, Tropical zone, in the Magdalena and Cauca valleys,
and doubtless through northern Colombia. Ranges northward to
Mexico.

Antioquia. Puerto Berrio (shallow water, alt. 125-135 m.),
Rusby & Pennell 32.

Cundinamarca. Melgar (pool in clayey loam, alt. 400-500 m.),
Pennell 2927.

Tolima. Espinal to Cuamo (ditch, loam, alt. 350-400 m.), Rusby
& Pennell 179.

Valle. La Paila, I. F. Holton 581 (H, Y). ■

8. HYDRANTHELIUM Humboldt, Bonpland and Kunth.
Hydranthelium H. B. K., Nov. Gen. et Sp. 7: 202. pi. 646. 1825.
Type species, H. callitrichoides H. B. K., of Venezuela.

1. Hydranthelium braunii Ernst.

Hydranthelium braunii Ernst, in Vargasia 1: 189. 1870.
"Halle el 19 de Setiembre de 1869 en uno de los pozos de la sabana delante
el camposanto de la Merced [Caracas]," Venezuela.

Open wet soil, at an altitude of 30 meters, Tropical zone, in Mag-
dalena; doubtless eastward near the Caribbean coast through Vene-
zuela.

Magdalena. El Libano plantation, (Santa Marta region, open
land on border of swamp and flooded during heavy rains, alt. 30 m.),

1920.] NATURAL SCIENCES OF PHILADELPHIA. 159

H. H. Smith 2544 (C, H, U, Y). (Plants small, largest .8 dm.
long, and in flower only, youngest leaves slightly undulate-lobed.)

9. SCOP ARIA Linne.
Scoparia L., Sp. PL 116. 1753.

Type species, S. dulcis L.

1. Scoparia dulcis L.

Scoparia dulcis L., 1. c. 116. 1753. "Habitat in Jamaica, Curassao";
ex L., Hort. Cliff. 320. 1737. "Crescit in Curassao & Jamaica." No
specimens from Curasao seen but the plant here considered unquestion-
ably occurs there.
Capraria dulcis (L.) Kuntze, Rev. Gen. 459. 1891.

Open soil, loam or sand, river-banks, fields, along trails and in
towns, at altitudes below 1500 meters, Tropical zone, doubtless
throughout lowland Colombia. Ranges throughout Tropical Amer-
ica, a weed of South American origin.

Antioquia. Vuelta de Acufia on Rio Magdalena (sandy loam
shore, alt. 125-130 m.), Pennell 3790.

Bolivar. Boca Verde on Rio Sinu (cacaotale, alt. 90-120 m.),
Pennell 4233; Buenavista, east of Since (open grassy place), Pennell
3991; Calamar (along Rio Magdalena, alt. 15-25 m.), Rusby &
Pennell 17; Vilches on Rio Sinu (loam, alt. 20-50 m.), Pennell 4713.

Cundinamarca. Girardot (field, alt. 350-400 m.), Rusby & Pen-
nell 113; Pandi (open slope, alt. 900-1100 m.), Pennell 2816.

Huila. Cordillera Oriental, east of Neiva (open foot-hill, alt.
700-1500 m.), Rusby & Pennell 400.

Magdalena. Bonda (alt. 45 m., common weed in open places
below 900 m.), H. H. Smith 1330 (C, H, U, Y).

Meta, Villavicencio (streets, alt, 525 m.), Pennell 1372, (road-
side, alt. 500 m.), Pennell 1572.

Tolima. Libano (field, alt. 700-900 m.), Pennell 3426.

Valle. La Paila, I. F. Holton 587 (Y).

10. STEMODIA Linne.

Type species, S. maritime/, L., of Jamaica.

Stemodia L., Syst, Nat. ed. X, 1118. 1759.
1. Stemodia durantifolia (L.) Sw.

Capraria durantifolia L., Syst. Nat. ed. X. 1116. 1759. ". . . Sloan.

Jam. t. 174." Ex Sloane, Jam. 196. pi. 124. f. 2: -Grows in the sandy

savannas [of Jamaica]."
Stemodia durantifolia (L.) Sw., Obs. Bot, 240. 1791.
Stemodacra durantifolia (L.) Kuntze, Rev. Gen. 466. 1891.

Occurs in two color-forms, blue, and lavender or "pale-pink."

160 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

River-flats and moist, or frequently desiccated, ditches, in open
land, at altitudes below 500 meters, Tropical zone, near the Carib-
bean Coast, along the Rio Magdalena and on the Sabana of Bolivar.
Ranges from Mexico to Brazil and in the West Indies.

Antioquia. Brazue'a de Perales (river flat along Rio Magda-
lena, alt. 150 m.), Pennell 3698 [corolla blue].

Bolivar. Cartagena (moist arroyo, 12 km. s. e. of C, alt. 50-100
m.), Pennell 4729 [corolla blue], (open ditch, 14 km. s. e. of C, alt.
50-100 m.), Pennell 4730 [corolla lavender]; Since (desiccated pool
in prairie, alt. 120-170 m.), Pennell 4047 [corolla lavender].

Huila. Quebrada de Angeles, above Natagaima (alt. 450-500
m.), Rusby & Pennell 263 [corolla blue].

Magdalena. Carpentiera (along Rio Magdalena, alt. 50-60 m.),
Rusby & Pennell 28; Mamatoca (open boggy ground, 5 m. s. of M.,
alt. 30 m.), H. H. Smith 1360 (C, H, U, Y) [corolla pale pink].

11. UNANUEA (Ruiz and Pa von.) gen. nov.
Unanuca R. & P., (Ic. Fl. Per. Ined.),

Erect, much-branhed shrubby herbs or low shrubs, at least 8-10
dm. tall. Stems 4-angled. Leaves mostly whorled in threes,
lanceolate to ovate, acute to acuminate, serrate to dentate, shortly
petioled. Pedicels axillary, slender, as long as or longer than the
calyces, not bracteolate. Sepals five, uniform, linear to lanceolate,
entire. Corolla 9-13 mm. long, tubular-campanulate with spread-
ing lobes, the posterior united three fourths length; externally
puberulent, within glabrous, purple-blue. Stamens four, didy-
namous (the posterior filaments shorter), glabrous, anthers-sacs
elliptic, each stalked on a short arm of the connective. Style gla-
brous. Stigmas distinct, flattened. Capsule brown, ovate in out-
line, acuminate, dehiscent septicidally (even through septum) and
slightly also loculicidally; placentae adherent to septum, roughened
by funicles. Seeds oblong, minutely roughened-tubercular.

Differs from Stemodia L., which has corolla-lobes not widely
spreading, pubescent or glabrous within on the anterior side, two
bracteoles on the pedicel beneath the calyx, and the leaves sessile
and clasping; from Lendneria Minod, which are herbs low and
spreading, with smaller corollas densely pubescent within over the
bases of the posterior lobes and with evidently petioled leaves.
Both Stemodia and Lendneria are genera of the Tropical life-zone,
while Unanuea is of the Subtropical and Temperate zones.

Type species, Stemodia suffruticosa H. B. K., of Ecuador.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 161

1. Unanuea dentata (Minod) Pennell, comb. nov.

Slemodia suffruticosa H. B. K., f. dentata Minod in Bull. Soc. Bot. Geneve
ser. II. 10: 201. 1918. "In Andibus Ecuadorensibus (R. Spruce, n.
5066!)." Isotype seen in Gray Herbarium of Harvard University.

At an attitude of 2600 meters, Temperate zone, southern Cor-
dillera Central southward into Ecuador.

Cauca. Mozoco, Moras Valley, Tierra Adentro (alt. 2600 m.),
H. Pittier 1326 (U). ["A shrub, 1 m. high; fl. deep purple."]

12. LENDNERIA Minod.
Lendneria Minod, in Bull. Soc. Bot. Geneve, ser. II. 10: 240. 1918.
Type species, Capraria humihs Soland.

Pedicels 1-2 mm. long. Corolla 4 mm. long, its lobes blue, the
posterior united two-thirds length. Hairs within corolla over
base of posterior lobes not knobbed. Anther-sacs circular,
slightly separated on very short connective -arms. Capsule
globose, 2 mm. long, much shorter than the sepals. Seeds
pale yellowish, nearly cylindric. 1. L. humilis.

Pedicels longer, mostly 7-20 mm. long. Corolla 8 mm. long, its
lobes white, the posterior united nearly to apex. Hairs within
corolla oyer base of posterior lobes knobbed. Anther-sacs
oblong, distinctly separated on stout connective-arms. Cap-
sule in outline oblong-ovate, 4-5 mm. long, about equaling the
sepals. Seeds brown-black, nearly oblong. 2. L. angulata.

1. Lendneria humilis (Soland.) Minod.

Capraria humilis [Soland. in] Ait., Hort. Kew. 2: 354. 1789. "Nat. of
the East Indies. John Gerard Koenig, M.D. Introd. 1781, by Sir
Joseph Banks, Bart." Identified by Bentham in DC. Prod. 10: 383.
1846, as species here considered, his determination with an "!" Our
plant has been occasionally reported as an introduction into the Old
World Tropics.

Stemodia parviflora Ait., Hort. Kew., ed. II. 4: 52. 1812. "Nat. of South
America. Cult. 1759 by Mr. Ph. Miller." Miller's plant was derived
from Houston who collected at Cartagena in Colombia as well as in
Mexico and the West Indies. The original introduction of Houston,
published as Erinus penicillatus Mill., Gard. Diet. n. 5. 1768, differs
so essentially from the account of Aiton's plant, especially in denoting
a plant with glabrous stems and leaves, as to lead to the supposition that
the latter was described from specimens of a different origin.

Stemodia arenaria H. B. K., Nov. Gen. et Sp. 2 : 357. pi. 175. 1818. " Cres-
cit in ripa inundata fluminis Magdalenae prope Banco et El Pefion inter
Mompox et Morales."

Lendneria humilis (Solander) Minod in Bull. Soc. Bot. Geneve, ser. II. 10:
240. 1918.

Corolla-tube yellowish, lobes blue-violet, tube and lobes, es-
pecially on the posterior side, with deep violet lines. Seen also
(Pennell 4709) with corolla very pale, a distinct color-form.

Moist soil, river-banks and waste-land, frequently a weed near
habitations, at altitudes below 200 rneters, Tropical zone, along the

162 PROCEEDINGS OF THE ACADEMY OF

Rio Sinu in Bolivar, the Rio Cauca in Valle, and the Rio Don Diego
in Magdalena; doubtless wide-spread elsewhere. Ranges from
Mexico to Argentina and in the West Indies.

Bolivar. Boca Verde, on Rio Sinu (gravelly river-bank, alt.
90-120 m.), Pennell 4197, (field along river, alt. 100-200 m.), Pennell
4567; Frasquillo, on Rio Sinu (grove along river in village, alt.
90-120 m.), Pennell 4610; Vilches, on Rio Sinu (shaded yard, alt.
20-50 m.), Pennell 4708 [corolla blue], 4709 [corolla pale-blue],
4710 [corolla intermediate in color between 4708 and 4709].

Magdalena. Open sandy ground by the Rio Don Diego, near the
sea, H. H. Smith 2730 (Y).

Valle. La Paila, I. F. Holton 580 (H, Y).

2. Lendneria angulata (Oersted) Pennell, comb. nov.

Stemodia angulata Oersted in Kjoeb. Vidensk. Meddel. 1853: 22. 1853.
"I Naerheden af Cartago i Costa-Rica."

Certainly distinct from Stemodia jorullensis H. B. K., Nov. Gen.
et Sp. 2: 358. 1818, which is described as 1-1 \ feet tall and with
leaves in threes, incised or doubly serrate.

Moist soil in shade, river-banks and waste land, at altitudes
below 600 meters, Tropical zone, along the Rio Sinu in Bolivar,
and in Magdalena; doubtless wide-spread in northern Colombia.
Ranges northward to Guatemala.

Bolivar. Frasquillo, on Rio Sinu (shady soil along river, alt.
70-100 m.), Pennell 4192.

Magdalena. Minca (clamp door-yard, in crevices of bricks, alt.
600 m.), H. H. Smith 1328 (C, H, U, Y).

13. VANDELLIA Browne.
Vandellia Browne; L., Mant. PI. 1: 12, 89. 1767.
Type species, V. diffusa L.

1. Vandellia diffusa L.

Vandellia diffusa L., Mant. PI. 1 : 89. 1767. "Habitat in Insula S. Thomae.
D. D. Browne." Description inaccurate in describing the calyx as quad-
ripartite (but with upper lobe subbifid), its lobes equal, the lower lip of
corolla as bilobed, and the capsules as one-celled, but is nevertheless
certified by Bentham, DC. Prod. 10: 416. 1846 as being the plant here
considered.

Lindernia diffusa (L.) Wettst, in Nat. Pflanzenfam. 43b : 79. iggi.

Pyxidaria diffusa (L.) Kuntze, Rev. Gen.: 464. 1891.

Moist open soil, along trails and in towns, at altitudes beiow
1500 meters, Tropical zone; widely distributed through tropical
America, in situations to suggest its having been naturalized. Ap-

1920.] NATURAL SCIENCES OF PHILADELPHIA. 163

parently introduced from the Ethiopian Region of the Old World
Tropics.

Cundinamarca. Icononzo (along trail in forest, alt. 1400-1800
m.), Pennell 2871.

Magdalena. River Don Diego (open sandy wet ground near sea,
alt. 0-10 m.), H. H. Smith 2729 (H, Y).

Meta. Villa vicencio (moist depression in llano, e. of, alt. 450
m.), Pennell 1466; (streets of town, alt, 525 m.), Pennell 1575,
(gravel along Rio Guatiquia, alt. 500 m.), Pennell 1590.

Tolima. Mariquita (prairie, depression, alt. 250-300 m.), Pen-
nell 3639.

14. TORENIA Linne.
Torenia L., Sp. PI. 619. 1753.
Type species, T. asiatica L., of India.

Bracts leaf-like, the upper smaller; inflorescence an elongate raceme.
Pedicels 20-25 mm. long. Sepals 4-5 mm. long. Corolla 5-7
mm. long, its posterior lobes united f-f length and equaling
the anterior; anterior lobes widely horizontally spreading;
within glabrous, blue on posterior lobes and distally on anterior
lobes, these yellowish-white proximally with an arch of deep-
violet at the base of the anterior lobes. Capsule 3-4 mm. long,
obtuse and mucronate. Leaves ovate, 1-1.5 cm. long. Plant
repent-ascending. 1. T. Crustacea.

Bracts minute, subulate; inflorescence congested at the nodes.
Pedicels 10-15 mm. long. Sepals 9-11 mm. long. Corolla
8-9 mm. long, its posterior lobes united nearly to apex and
shorter than the anterior lobes; anterolateral lobes placed
sagitally, the anterior lobe horizontal and longest; within
pubescent over bases of anterior lobes, white or on anterior
side somewhat blue-violet. Capsule 8-10 mm. long, acute.
Leaves lanceolate-ovate, 1.2-2.5 cm. long. Plant ascending-
erect.
Corolla on anterior lobe blue-violet, and on anterolaterals with

blue-violet streaks. Leaves dentate. 2. T. thouarsii.
Corolla white throughout. Leaves crenate-serrate.

2a. T. thouarsii nivea.

1, Torenia Crustacea (L.) Cham. & Schlecht.

Capraria Crustacea L., Mant. 87. 1767. "Habitat in Amboina; China."
Aniboina plant,, described by Etumphius, Herb. Amb. 5: 461. pi. 170 f. 3,
verified by Dr. E. D. Merrill, Interp. Rumph. Herb. Amb. 468, as the
plant here considered.

Torenia Crustacea (L.) C. & S. in Linnaea 2: 570. 1S27.

Vandellia Crustacea (L.) Benth., Scroph. Ind. 35. 1835.

Lindernia Crustacea (L.) F. Muell., Census 97. 1882.

Pyxidaria crustacea (L.) Kuntze, Rev. Gen. 2: 464. 1891.

164 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Moist open soil, along trails and in towns, at altitudes of 500 to
600 meters, Tropical zone; introduced from Oriental Region of
Old World Tropics.

Cundinamarca. Melgar (moist loam, open slope, along trail,
alt. 500-600 m.), Pennell 2878, 2879 [form smaller throughout].

Meta. Villavicencio (streets and yards, alt. 525 m.), Pennell
1527.

2. Torenia thouarsii (Cham. & Schlecht.) Kuntze.

Nortenia thouarsii Cham. & Schlecht. in Linnaea 3 : 18. 1828. "In Brasiliae
provincia'Rio Janeiro in uliginosis post Botafoginam . . . legit Bey-
rich, in insulis Madagascaria et Mauritii Depetit Thouars.
Willd. Hb. n. 11,547 (planta madagascariensis a Thouarsii comm.)."

Torenia nortenia Steud., Nom., ed. II. 2: 692. 1841.

Torenia thouarsii (Cham. & Schlecht.) Kuntze, Rev. Gen. 468. 1891.

This has usually been known as Torenia parviflora Ham.

Moist banks, in edge of forest, at an altitude of 500 meters, Trop-
ical zone; wide-spread through lowland tropical South America,
growing as if a native plant. However, as this American plant ap-
pears to be indistinguishable from the plant of the Old World, and
the remaining species of the genus are all Palaetropic it appears
nearly certain that our plant is an introduction. From the Ethi-
opian and Oriental regions of the Old World Tropics.

Meta. Villavicencio (moist shaded bank near Rio Guatiquia,
alt. 500 m.), Pennell 1528, (moist meadow and swamp in forest,
near Rio Guatiquia, alt. 500 m.), Pennell 1560.

2a. Torenia thouarsii nivea Pennell, var. nov.

Leaves smaller, 1.2-1.5 cm. long, crenate-serrate, rather than
dentate. Corolla white throughout. Probably a color-form iden-
tical with plants of the Old World Tropics and introduced there-
from.

Type, wet sand along Rio Guatiquia, altitude 500 meters, col-
lected in flower and fruit August 30, 1917, F. W. Pennell 1531; in
herbarium New York Botanical Garden.

15. ILYSANTHES Rafinesque.
Ilysanthes Raf., Ann. Nat. 13. 1820.
Type species, I. riparia Raf., of the Ohio.

1. Ilysanthes inaequalis (Walt.) Pennell.

Gratiola inaequalis Walt., Fl. Carol. 61. 1788. [South Carolina.]
Ilysanthes inaequalis (Walt.) Pennell in Torreya 19: 149. 1919.

Wet soil near streams, at an altitude little above sea -level, Trop-
ical zone, in Magdalena. Probably wide-spread, and also in the

1920.J NATURAL SCIENCES OF PHILADELPHIA. 165

Subtropical and Temperate zones. Through lowland South Amer-
ica south to Paraguay; ranges northward into Temperate North
America.

Magdalena. Rio Buritaca (observed only in a swampy place,
bank of R. B., close to the sea — 50 m. e. of Santa Marta), H. H.
Smith 1329 (C, H, U, Y).

16. SCHISTOPHRAGMA Bentham.
Schistophragma Benth. in DC. Prod. 10: 392. 1846
Type species, S. pusilla Benth.

1. Schistophragma pusilla Benth.

Schistophragma pusilla Benth. in DC. Prod. 10: 392. 1846. "In Mexico
pr. Tehuantepec (Alaman!) . . . (v. s. comm. a. el. DC.)." De-
scription from a dwarf plant, which explains the variety following.

Schistophragma pusilla major Benth., 1. c. 392. 1846. "In campis aridis
pr. Sta Martha CPurdie!) . . . (v. in herb. Hook.)."

Conobea pusilla (Benth.) B. & H., Gen. 2: 951. 1876.

Open dry, stony soil, at altitudes below 300 meters, Tropical
zone, in Magdalena, and doubtless spread across northern Colombia.
Ranges northward to Mexico.

Magdalena. Bonda (alt. 60 m.), H. H. Smith 1970 (C, H, U, Y).

["Rare on open, stony and dry ground, hillsides or banks below
1000 f."]

17. MIMULUS Linne.
Mimulus L., Sp. PI. 634. 1753.
Type species, M. ringens L., of Virginia.

1. Mimulus glabratus II. B. K.

Mimidus glabratus H. B. K., Nov. Gen. et Sp. 2: 370. 1818. "Crescit
prope Moran Mexicanorum, alt. 1330 hex. [ = 2527 in.]. Varies, even
in same colony, with leaves obviously petioled or nearly sessile, and
with all parts of the plant, including the flower, relatively large or rela-
tively small. The latter state is doubtless the basis of M. andicola H.
B. K. from Ecuador.

Corolla lemon-yellow, within throat on anterior side golden and
spotted with many red-brown spots.

Along streams, springheads, swales and brooks, at altitudes of
2300 to 3200 meters, Temperate zone, ascending as a dwarfed plant
to Paramo, and descending rarely into the Subtropical zone, both
slopes of Cordillera Oriental, in Cundinamarca. Doubtless through-
out this and the other Cordilleras. Ranges through the Andes
southward to Bolivia and with many breaks, through the mountains
of Central America and Mexico, northward to Colorado and the
plains of North Dakota.

166 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Cundinamarca. Bogota (ditch in field near Rio San Cristobal,
alt. 2800 m.), Pennell 2194, 2279, (wet open spring-head, alt. 2700-
2800 m.), Pennell 2296; Chipaque (wet roadside, alt. 2300-2400
m.), Pennell 1326; Sibate (wet roadside, alt. 2620 m.), Pennell 2387;
Zipaquira (springhead in meadow, alt. 2650 m.), Pennell 2533; Mt.
Chuscal, west of Zipaquira (swale on paramo, alt. 3100-3200 m.),
Pennell 2600; Guasca (alt. 2700 m.), Triana.

18. LEUCOCARPUS D. Don.
Leucocarpus D. Don in Sweet. Brit. Flow. Gard. II. pi. 124. 1831.
Type species, Conobea alata Graham, of Mexico.

1. Leucocarpus perfoliatus (H. B. K.) Benth.

Mimulus perfoliatus II. B. K., Nov. Gen. et Sp. 2: 371. 1818. "Crescit
in Regno Novo-Granatensi. ... A Mutisio cum Bonplandio com-
municatus." Described as with leaves connate, but this appearance has
misled subsequent workers, including Bentham in his characterization
of Leucocarpus. The plant actuallv has opposite cordate-clasping leaves.

Leucocarpus perfoliatus (H. B. K.) Benth. in DC. Prod. 10: 335. 1846.

This species has usually been called L. alalus (Graham) D. Don,
based upon Conobea alata Graham (1830) of Mexico, a plant more
carefully described and under an appropriate name. The character
of length of calyx-lobes, used by Bentham, is of no significance,
the lobes varying in the same specimens and usually seeming rela-
tively longer when in the bud.

Corolla yellow throughout, with two pubescent ridges within
throat anteriorly. Plant shrubby below, from a perennial root
sending up a clump of many long densely floriferous stems. Fruit
fleshy, chalky-white.

Along stream banks in forest, at altitudes of 1350 to 2000 meters,
probably from all slopes of the Cordilleras, and also on the Sierra
Nevada de Santa Marta. Subtropical zone. Ranges from Mexico
to Bolivia. The seeds are doubtless carried by birds.

Huila. Cordillera Oriental, east of Neiva (along rocky stream in
forest, alt. 1500-2000 m.), Rusby & Pennell 600.

Magdalena. Las Nubes (damp clearing near stream, alt. 1350 m.),
H. H. Smith 1405 (C, H, U, Y).

Tolima (?). "Forets de Quindio" (2200 m.), J. Triana (U).

Valle. Jicaramata, "circum flumen Toluam", I. F. Holton 578.

19. ALONSOA Ruiz and Pavon.
Alonsoa R. & P., Syst. Veg. Fl. Peruv. et Chil. 150. 1798.
Type species, A. caulialata R. & P., of Peru.
Leaves coarsely serrate or dentate, the largest 5-6 cm. long. Cor-
olla 10 mm. long. Filaments thick. Anthers infundibuliform-

1920.] XATURAL SCIENCES OF PHILADELPHIA. 167

explanate. Capsule 9-10 mm. long, narrowly ovoid, con-
spicuously attenuate. Seeds black, the furrows nearly as wide
as the intervening ridges. Stem above, pedicels and calyces
usually glabrous, rarely somewhat glandular-pubescent.

1. A. meridionalis.
Leaves uniformly serrate, the largest 8-9 cm. long. Corolla 5-6

mm. long. Filaments thin. Anthers widely and flatly ex-
planate. Capsule 6 mm. long, pyramidal, shortly attenuate.
Seeds dark-brown, the furrows much narrower than the inter-
vening ridges. Stem above, pedicels and calyces glandular-
pubescent, densely pubescent at the bases of the petioles.

2. A. serrata.

1. Alonsoa meridionalis (L. f.) Kuntze.

Scrophvlaria meridionalis L. f., Suppl. 280. 1781. "Habitat in Nova

Granada. D. Mutis." Type probably from Bogota.
Hemimeris mulisii H. B. K., Nov. Gen. et Sp. 2: 376. 1817. ''Crescit

prope Santa Fe de Bogota [Humboldt & Bonpland]."
Alonsoa mutisii (H. B. K.) G. Don, Gen. Syst. 4: 513. 1838.
Alon.soa meridionalis (L. f.) Kuntze, Rev. Gen. 2: 4.">7. 1891. The plant

of Venezuela noted by Kuntze, and described as so variable in color, is

A. -parviflora (H. B. K.) G. Don.

Corolla uniformly dull-orange. Filaments dull-yellow. Anthers
yellow.

Waysides and grassy slopes, around the margins of the Sabana
of Bogotc4, on the western slope of the Cordillera Oriental; at alti-
tudes of 2600 to 2900 meters. Temperate zone.

Cundinamarca. Zipaquira, Pennell 2564 (Y); Bogota, Holton
(Y), Pennell 1309 (Y), Pennell 2099 (Y), Pennell 2332 (Y); (Plateau
de Bogota), J. Triana (Y); El Pcnon, s. w. <>f Sibate, Pennell 2410
(Y), [pedicels unusually glandular-pubescent].

2. Alonsoa serrata Pennell, sp. now

Stem 6 dm. tall, four-angled, angles slightly winged, stem glabrous
below, glandular pubescent above. Leaves opposite, 8-9 cm. long,
the blades ovate, uniformly serrate, slightly paler beneath, glabrous,
on petioles less than one-half length of blade. Racemes indefinite.
terminal on the stem and branches; bracts sessile, the lowermost
ovate and somewhat serrate, nearly 2 cm. long, the upper lanceolate.
smaller. Pedicels spreading, 10-11 mm. long, glanduiar-pubescent.
Sepals oblong-lanceolate, acute, glandular-pubescent, 3 4 mm. long.
Corolla 5-6 mm. long. Filaments slender. Anthers widely ex-
planate. Style 2-2.5 nun. long. Stigma capitate. Capsule 6 mm.
long pyramidal, slightly attenuate to an obtuse apex, glabrous.
Seeds .8-.9 mm. long, cylindric, dark-brown, with about 8 rounded
longitudinal ridges separated by deep narrow furrows; the whole
surface finely alveolate-reticulate.

168 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Type, Santa Marta Mountains, collected in flower and fruit about
April, 1899, H. H. Smith 1497; in Herb. New York Botanical Garden.
The note for 1497 is stated by Smith to have been lost, but he tells
us that the plant is "probably from Valparaiso, 4500 ft. [ = 1350 m.]"
altitude. I suspect that this plant came from much higher than
this.

Rusby has compared this with Purdie's plant from Santa Marta
and has written on our sheet "Purdie's plants are more hairy than
this," an excellent confirmation of Purdie's specimens cited in DC.
Prod. 10. 250 : being this same species.

Magdalena. Valparaiso, Santa Marta Mts., H. H. Smith 1497 (C,
H, U, Y).

20. FAGELIA Schwencke.

Fagelia Schwencke, [in Verh. Bataafsh. Genootsch. Rotterdam 1 : 474. pi. 13.

1774, generic diagnosis only]; J. F. Gmel., Syst. Nat. 40. 1791.
Calceolaria L., in Kongl. Vetensk. Acad. Handl. 31: 288, 1770, not Calceolaria

Fabr., Enum. Meth. PI. Host. Med. Helmstad. ed. II. 37. 1763.

Type species, F. flavicans J. F. Gmel., probably from Ecuador.
Anther-sacs proximate on the simple filament, both alike and fertile.
Anterior lip of corolla ( = sac) not over twice width of posterior
lip ( = hood). Leaves entire to coarsely serrate-dentate, the
blades at times triangular. [Cheiloncos Kranzl.]
Capsule ovate or broader, no longer than broad, thick-walled,
shorter than or but slightly exceeding the sepals. Corolla
8-25 mm. long, with the posterior lip broadly truncate to
notched. Filaments stout, not or but little longer than the
, oblong anthers. Inflorescence corymb-like, both secondary
branches developed, and at least some of the lateral flowers
without bracts.
Shrubs. Glutinous above, on stems, pedicels, sepals and leaves.
Leaf-blades lanceolate, entire to slightly serrate, shortly
petiolate. Corolla appearing broader than long because
the sac is pressed tightly against hood ; orifices to lips broad
and rounded, so sac is shallow.
Calyx 4-6 mm. long. Leaves glaucous beneath. Plant less
glutinous. Branches of the inflorescence once branched,
so that flowers occur in fours. 1. F. microbefaria.

Calyx 2-3 mm. long. Leaves slightly paler beneath. Plants
very glutinous. Branches of the inflorescence irregularly
twice branched, so that flowers occur usually in clusters
of more than four. 2. F. fruticosa.

Herbaceous throughout. Not glutinous, but often with stalked
glands. Leaf-blades lanceolate to broadly triangular, ser-
rate to doubly dentate. Corolla evidently elongated sagit-
tally, with sac usually not pressed against hood and so

1920.] NATURAL SCIENCES OF PHILADELPHIA. 169

with its orifice evident; orifice to hood narrow, or trun-
cate; to sac truncate, with sac deep.
Calyx 4-5 mm. long, shorter than or about equaling the
capsule. Leaves sessile. Stem densely glandular-pubes-
cent.
Leaves narrowed at base, elliptic-lanceolate, dentate, on
both surfaces densely ferruginous-pubescent with dark-
jointed hairs. Stem below densely, above sparsely
pubescent. Inflorescence of a few wide-spreading
branches. Corolla 8-10 mm. long. Anthers about
1.5 mm. long. 3. F. lehmanniana.

Leaves rounded-clasping at base, lanceolate, crenate-serrate
(with spinulose serrations), above glabrate, beneath
paler and finely pubescent. Stem below apparently
glabrate, above pubescent with short gland-tipped
hairs. Corolla about 12-15 mm. long. Anthers about
2.5 mm. long. 4. F. crenata.

Calyx 7-12 mm. long, longer than the capsule. Leaves
petioled, petiole at times broadly winged.
Anther-sacs stiffly divaricate (so anthers straight), open-
ing throughout or from distal apices. Capsule with
gland-tipped hairs. Corolla slightly pubescent within
at base. Style 1.5-4 mm. long. Wing of petiole less
than one-third width of blade or wanting.
Leaves 3-5 mm. long, the petioles wingless.

Stem pubescent throughout with glandless hairs.
Calyx-lobes obtuse to acute. Corolla 15-20
nun. long. Style 1.5 mm. long. Leaves 3-4 cm.
long, obtuse to acutish, simply or somewhat
doubly crenate-serrate, beneath pale and densely
pubescent. Secondary branches of the inflores-
cence not or scarcely branched.

5. F. saxalilis.
Stem pubescent distally with gland-tipped hairs.

Calyx-lobes acuminate. Corolla about 10 mm.
long. Style 3 mm. long. Leaves 4-5 cm. long,
acuminate, irregularly somewhat doubly serrate
with acute teeth, beneath lighter green and some-
what pubescent. Secondary branches of the in-
florescence repeatedly branched.

6. F. bogotensis.
Leaves 11-17 cm. long, the petioles winged proximally

or throughout, the wing connate with that of oppo-
site leaf.
Petioles broadly winged proximally, distally very nar-
rowly margined; leaf-blades broader than long,
irregularly sha) lowly crenate-dentate. Corolla
with posterior lobes not united to apex, so leaving
a deep narrow orifice into hood. Anther-sacs
broadly contiguous. 7. F. trilobata.

170 PROCEEDINGS OF THE ACADEMY OF [May-Oct.

Petioles nearly uniformly winged throughout; leaf-
blades longer than broad, sharply doubly serrate-
dentate. Corolla with posterior lobes united
nearly or quite to apex, so leaving slight if any
median orifice into hood. Anther-sacs narrowly
contiguous.
Corolla 8-10 mm. long; hood with slight median
orifice. Anthers 2 mm. long, the sacs opening
throughout. Style 1.5-1.8 mm. long. Capsule
3-4 mm. long. Winged petiole usually 12-18
mm. wide, and somewhat dentate. Plant pu-
bescent above with some gland-tipped hairs.

8. F. data.
Corolla 13-15 mm. long; hood truncate, without
apical median orifice. Anthers 3.5 mm. long,
the sacs opening from distal apex but not
throughout. Style 4 mm. long. Capsule 5-6
mm. long. Winged petiole 8-13 mm. wide,
entire. Plant pubescent throughout with gland-
less hairs. 9. F. nevadensis.
Anther-sacs curved (so anther horseshoe-shaped), opening
from proximal apices partially or throughout. Capsule
pubescentwith glandless hairs. Corolla glabrous within
at base. Stvle 5 mm. long. Wing of petiole more than
one-third width of blade.
Leaves sharply dentate, above glabrate, beneath paler
and finely pubescent. Stem glabrate below, some-
what hirsute above. Calyx 8-10 mm. long.

10. F. tolimensis.
Leaves obtusely dentate, above pubescent, beneath softly

pubescent to tomentose. Stem hispid below, hirsute-
pubescent above. Calyx 9-11 mm. long.

11. F. perfoliata.
Capsule narrowly pyramidal, longer than broad, thin-walled,

nearly twice as long as the sepals. Corolla 5-6 mm. long,
nearly globose, with posterior lip attenuate to a shallowly
notched apex. Filaments slender, several times longer than
the hemispheric anthers. Inflorescence appearing as if with
flowers axillary, normally one secondary branch developing
repeatedly through an indefinite number of nodes.

12. F. ovata.
Anther-sacs separated on two arms of the connective (filament

wanting or very short), dissimilar, the anterior tending to
become sterile. Anterior lip two to four times the width of the
posterior. Leaves pinnately lobed. [Aposecos Benth.]
Anterior anther-sac fertile, brown or yellowish. Style .6-. 8 mm.

long. Calyx 3-5 mm. long. Pedicels and stems above

pubescent with gland-tipped hairs.

1920.] NATURAL SCIENCES OF PHILADELPHIA. 171

Leaves 2-4 cm. long, the blades with one or two partial pairs of
segments, the sinuses of which are narrow and reach only
one-half to two-thirds the distance to the midrib. Calyx
3-4 mm. long. 13. F. micrantha.

Leaves 2-10 cm. long, the blades with usually three pairs of
segments, the proximal sinuses of which are usually broad
and reach nearly to the midrib. Calyx 4-5 mm. long.

14. F. radiculoides.
Anterior anther-sac sterile, yellow or orange-yellow. Style 1-2
mm. long. Calyx 5-9 mm. long.
Anterior anther-sac light-yellow. Corolla 5-7 mm. long. Leaf-
blades with shallow crenately-toothed lobes. Distally finely
pubescent with glandless hairs. 15. F. crenatiloba.

Anterior anther-sac orange-yellow. Corolla 7-21 mm. long.
Leaf-blades with deep sharply serrate lobes.
Stems above and pedicels pubescent with few-celled gland-
tipped hairs. Blades of lowermost leaves with shallow
lobes. Base of petioles very narrowly connate. Calyx-
lobes slightly serrate, obviously ciliate with gland-tipped
hairs. Style 1-1.3 mm. long. Capsule pubescent with
short gland-tipped hairs. 16. F. chelidonioides.

Stems above and pedicels hirsute with many-celled dark-
jointed hairs. Blades of all leaves pinnatisect nearly
to the midrib. Base of petioles obviously connate.
Calyx-lobes decidedly serrate, hirsute on the back and
margin. Style 1.8-2 mm. long. Capsule pubescent with
glandless (or in pinnatisecta with interspersed gland-
tipped) hairs.
Corolla 15-21 mm. long. Calyx-lobes ovate, acute. Cap-
sule 6 mm. long. Leaves 5-8 cm. long, 3-5 cm. wide.

17. F. scalaris.
Corolla 8-10 mm. long. Calyx-lobes lanceolate or nar-
rowly ovate, acuminate. Capsule 4 mm. long. Leaves
3-5.5 cm. long, 2-3 cm. wide. 18. F. pinnatisecta.

1. Fagelia microbefaria (Kranzl.) Pennell, comb. nov.

Calceolaria microbefaria Kranzl, in Ann. k. k. Naturh. Hofm. Wien 22: 193.
1907. "Kolumbien, Ostkordilloron, Provinz Pamplona, zwischen Urban
i in. I Las Vetaa in 3300 m u. d. M. (Linden Nr. 730!)."

Stem 1-2 meters tall, much branched, woody, with grayish-brown
bark, the twigs reddish- or yellowish brown, glutinous and finely
pubescent. Leaves 6-8 cm. long, the blades lanceolate, acuminate,
serrate to entire, at times slightly revolute, 12-17 mm. wide; each
narrowed to a petiole 5-10 mm. long; blades above green, pubescent
on the midrib or pulverulent or quite glabrous, beneath glaucous,
sparsely puberulent to glabrous, reticulate; somewhat glutinous on
upper surface. Corymb bractless, the secondary branches slightly

172 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

if at all branched, so that the inflorescence simulates a stalked four-
flowered umbel. Peduncle and pedicels somewhat glutinous and
pubescent with brown hairs. Calyx 4-6 mm. long, the lobes tri-
angular-ovate, obtusish to acute, entire, puberulent, glutinous.
Corolla: posterior lip 6-7 mm. long, 7-9 mm. wide, with broad
rounded orifice; anterior lip 12-13 mm. long, 8-10 mm. wide, with
rounded orifice opening into two-thirds or more of its length; sac
pressed against hood so that corolla appears broader than long;
externally finely puberulent, within glandular-pubescent proximally,
especially about bases of filaments. Filaments stout, less than 1
mm. long. Anther 2.5 mm. long, brown, the walls thin, the sacs
contiguous, opening throughout and eventually through the thin
connective. Style 3 mm. long. Capsule 5 mm. long, broadly ovate,
acute, puberulent. Seeds .4-.5 mm, long, oblong, obtuse, ridged,
red brown.

Thickets, along streams and at edge of forest, also in thicket-
islands in Paramo, at altitudes of 2800 to 3300 meters; Temperate
zone of western slope of Cordillera Oriental, from Santander to
Cundinamarca.

Cundinamarca. Rio Frio, west of Zipaquira (along stream-
banks), Pennell 2570, 2605; Mt. Chuscal, west of Zipaquira (thicket-
island in paramo), Pennell 2584; Sibate (bushy hillsides southwest
of), Pennell 2389.

2. Fagelia fruticosa Pennell, sp. nov.

Stem 1-2 meters tall, much branched, woody, with grayish bark,
the twigs reddish and glutinous-puberulent or slightly pubescent,
Leaves 5-6 cm. long, the blades lanceolate, acuminate, slightly
serrulate to entire, at times slightly revolute, 12-13 mm. wide;
each narrowed to a petiole 4-7 mm. long; blades above dark-green
and puberulent, beneath paler and reticulate, on both surfaces
strongly glutinous. Corymb bractless, the secondary branches soon
branching so as to simulate an umbel. Peduncle and pedicels
glutinous and somewhat pubescent with brown hairs. Calyx 2-3
mm. long, the lobes broadly ovate, acute, entire, puberulent, glutin-
ous. Corolla: posterior lip 6-7 mm. long, 7-8 mm. wide, with broad
rounded orifice; the anterior lip 12-13 mm. long, 8-9 mm. wide,
with rounded orifice opening into two-thirds or more of its length;
sac pressed against hood so that corolla appears broader than long;
externally finely puberulent, within pubescent proximally, especially
near the filaments. Filaments stout, less than 1 mm. long. An-

1920.] NATURAL SCIENCES OF PHILADELPHIA. 173

thers nearly 2 mm. long, brown, the walls thin, the sacs contiguous,
opening throughout and through connective. Style 3 mm. long.
Capsule 5 mm. long, broadly pyramidal, acute, glandular-puberu-
lent. Seeds.

Plant more glutinous and drying blacker than Fagelia microbe-
faria.

Type, forest at margin of Paramo de Ruiz, Tolima, altitude 3200-
3500 meters, collected in flower December 16, 1917, F. W. Pennell
2998; in Herb. New York Botanical Garden.

Shrub belt about and below paramo, Temperate zone, eastern
slope of the Cordillera Central, in Tolima.

3. Fagelia lehmanniana (Kranzl.) Pennell, comb, nov,

Calceolaria lehmanniana Kranzl. in Fedcle, Rep. Nov. Spec. 1 : 100. 1905.
"Columbien: Dpto. Cauca; an feuchten Orten an den oberen Gehangen
des Vulcan de Sotara, und auf dem Paramo de Barbillas in 3000 bis 3300
m ii. d. M. (F. C. Lehmann, no. 6134!)"; not C. lehmannii (Hieron)
Hieron. in Engl. Bot. Jahrb. 20. Beibl. 49: 57. 1894.

At altitudes of 2800 to 3300 meters, Temperate Zone, southern
Cordillera Central, from Cauca to Pasto.
Pasto. Puruquai, J. Triana in 1851-7 (Y).

4. Fagelia crenata (Lam.) Kuntze.

Calceolaria crenata Lam., Encyc. Meth., Bot. 1: 556. 1785. "Trouvce au
Perou par M. Joseph de Jussieu (v. s. in herb. Juss.)." Jussieu collected
mainly in Ecuador.

Fagelia crenata (Lam.) Kuntze, Rev. Gen. 495. 1891.

At an altitude of 3000 meters, Temperate Zone, southern Cor-
dillera Central, from Pasto to Ecuador.

Cauca. Valle de Quintero above Pitaio, R. Palo basin, Central
Cordillera, H. Pitticr 1425 (U).

Pasto. "Tuquerres et Puruquai," J. Triana in 1851-7. "Juda-
bolsa."

5. Fagelia saxatilis (H. B. K.) Kuntze.

Calceolaria saxatilis H. B. K., Nov. Gen. et Sp. 2: 382. 1817. "Crescit
locis saxosis montis Chimborazo in summa planitie Sisgun, alt. 1750
hex. [ = 3325 meters] . . . [Humboldt & Bonpland]."

Fagelia saxatilis (H. B. K.) Kuntze, Rev. Gen. 460. 1891.

Stem about 1 meter tall, much branched, herbaceous, green or
reddish, pubescent with white hairs, densely so above. Leaves 3-4
cm. long, the blades ovate, cordate or truncate at base, obtuse, each
narrowed to a petiole 5-10 mm. long; blades above green, finely
pubescent, beneath pale and densely pubescent, with some sessile
glands; petiole white-lanate. Corymb bracted at base, its secondary
branches usually quite simple. Pedicels lanose with glandless white

174 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

hairs. Calyx 8-14 mm. long, the lobes ovate, obtuse to acute, ob-
scurely lobed, lanose-pubescent. Corolla: the posterior lip 3-4 mm.
long the two lobes united nearly to apex (so strongly hooded, with
shallow median orifice); anterior lip 20-25 mm. long, 10-12 mm.
wide, with orifice opening about one-third of its length; sac upcurv-
ing toward hood, the orifice opening externally; externally glandular-
pruinose, within slightly pubescent at base on anterior side. Fila-
ments 2 mm. long. Anthers 2 mm. long, brown, the walls thick,
the sacs broadly contiguous, opening throughout and through the
thin connective. Style 1.5 mm. long. Capsule 5-6 mm. long,
broadly globose-pyramidal, obtuse, thick-walled, puberulent with
gland-tipped hairs. Seeds .5 mm. long, oblong, distally acute,
finely ridged, brown.

Thickets, along streams and at edge of forest, at altitudes of
2700 to 3300 meters; Temperate zone of eastern slope of Cordillera
Central, from Tolima to Ecuador. Collected also above Bogota,
where probably an escape.

Tolima. "Rosalito," near Paramo de Ruiz (along stream in
meadow), Pennell 2990. Also recorded by Kranzlin from Paramo
de Ruiz, Purdie.

Cundinamarca. Guadalupe, above Bogota, Bro. Ariste-Joseph
A230 (U).

6. Fagelia bogotensis Pennell, sp. nov.

Stem probably about 1 meter tall, branched, herbaceous, pubes-
cent with white hairs, distally these gland-tipped. Leaves 4-5 cm.
long, the blades triangular ovate, cordate, acuminate, irregularly
dentate with acute lobes 3-4 cm. long, 2.5-3 cm. wide; each on a
wingless petiole 10-20 mm. long; blades above green, pubescent,
beneath lighter green and moderately pubescent, without sessile
glands; petiole pubescent, some hairs gland-tipped. Corymb leafy-
bracted at base, its secondary branches becoming much branched.
Peduncles and pedicels hirsute with gland-tipped hairs. Calyx
10-12 mm. long, the lobes ovate, acuminate, entire, hirsute-pubes-
cent. Corolla: the posterior lip 4 mm. long, the two lobes united
nearly to apex (so strongly hooded, with shallow median orifice);
anterior lip about 10 mm. long, with orifice opening less than ^ of
length; sac upcurving toward hood, the orifice opening externally,
externally glandular-pruinose, within somewhat pubescent at base.
Filaments 1.5 mm. long. Anthers 2.5 mm. long, brown, the walls
thick, the sacs broadly contiguous, opening from proximal end,

1920.] NATURAL SCIENCES OF PHILADELPHIA. 175

eventually to the distal apex. Style 3 mm. long. Capsule 6-8
mm. long, urceolate -pyramidal, acute, thick Availed, pubescent with
short gland-tipped hairs. Seeds .3 mm. long, oblong, mucronately
acute at each end, ridged, brown

Type, Bogota, Cundinamarca, alt. 2600 meters, collected in 1851-
1857 by J. Triana; in Herb. Columbia University at The New York
Botanical Garden.

At an altitude of slightly over 2600 meters, Temperate zone of
western slope of Cordillera Oriental, in Cundinamarca.

7, Fagelia trilobata (Hemsl.) Rusby.

Calceolaria trilobata Hemsl., Biol. Centr. Am., Bot. 2: 439. 1881-2. "Guate-
mala, Volean de Fuego, 7000 to 10.000 feet (Godman & Salvin, 239).
Colombia. Hb. Kew. The description was mainly drawn up from
Hotton's [ = Holton's] Colombian specimen, n. 575." Species an aggre-
gate, and to be typified by I. F. Holton 575. An isotype of this, in
Herb. Columbia University at The Xew York Botanical Garden is labeled
"Rio Arzobispo, in montibus juxta Bogotam, legit ... 23 Oct.
1852." A redescription, from this specimen, is given below.

Fagelia trilobata (Hemsl.) Rusby in Mem. Torr. Bot. Club 6: 93. 1896.
As to synonoiny only.

Stem probably about 1 meter tall, branched, herbaceous, pubes-
cent with white hairs, distally these gland-tipped. Leaves 15-17
cm. long, the blades triangular, slightly three-lobed, cordate, taper-
ing to an acute tip, irregularly crenate-dentate, with callous-tipped
lobules, 9-10 cm. long, 11-12 cm. wide, each on a petiole 6-7 cm.'
long its wing distally very narrow, proxhnally expanding to 3-4
cm. wide and connate with that of opposing leaf; above green, pub-
escent, beneath pale-green and slightly pubescent, more so on the
veins, narrow-winged portion of petiole pubescent, some hairs with
glandular tips. ( )orymb leafy-bracted at base, its secondary branches
much branched. Peduncles and pedicels hirsute with gland-tipped
hairs. Calyx 12 mm. long, the lobes ovate, acuminate, obscurely
lobate, hirsute pubescent. Corolla: the posterior lip 5-6 mm. long,
the two lobes not united to apex, so not hooded, with deep, narrow
median orifice: anterior lip 10 12 mm. long, with orifice opening
much less than one-half of length; sac upcurving toward hood;
externally glandular-pruinose, within somewhat pubescent at base.
Filaments 2 mm. long, widening distally. Anthers 3.5 mm. long,
brown or yellowish, the walls thick; the sacs broadly contiguous,
opening throughout, the septum between very thin, and ultimately
(?) breaking. Style 4 mm. long. Capsule glandular-puberulent,
not seen mature.

176 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Thickets along stream, at an altitude between 2600 and 3000
meters, Temperate zone of western slope of Cordillera Oriental, in
Cundinamarca.

8. Fagelia alata Pennell, sp. nov.

Stem about 1 meter tall, little branched herbaceous, pubescent
with white hairs, distally lanose and with short-stalked glands.
Leaves 11-17 cm. long, the blades triangular-ovate, cordate, acum-
inate, coarsely doubly dentate (dentate with the lobules triangular
and dentate), 7-10 cm. long, 6-8 cm. wide; each on a petiole 4-7
cm. long, this broadly winged throughout (in middle 10-18 mm.
wide), entire to crenate-dentate, proximally slightly expanding and
connate with that of opposing leaf; above green, beneath pale green,
on both surfaces slightly pubescent, more so on younger growth.
Corymb bractless, its secondary branches repeatedly branched.
Peduncles and pedicels pubescent with longer white glandless, and
with shorter gland-tipped hairs. Calyx 8-9 mm. long, the lobes
lanceolate, acuminate, slightly serrate-dentate or some entire,
glandular-puberulent. Corolla: the posterior lip 2-3 mm. long,
5 mm. wide, arched, the two lobes united very nearly to apex (so
hooded with slight median aperture); anterior lip 8-9 mm. long,
7 mm. wide, with orifice opening much less than | length (not strongly
» upcurving toward hood) ; externally minutely glandular-puberulent,
within minutely pubescent at base on anterior side, lemon-yellow
margin of sac very finely purple-spotted. Filaments .5-. 7 mm.
long, narrowing distally. Anthers 2 mm. long, yellowish, the walls
thick; the sacs narrowly contiguous, opening throughout, the sep-
tum between thin and ultimately breaking. Style 1.5-1.8 mm.
long. Capsule 3-4 mm. long, urceolate-pyramidal, emarginate,
pubescent with short gland-tipped hairs. Seeds .2-.3 mm. long,
oblong, distally obtuse, ridged, black -brown.

Type, moist bank in forest, loam soil, western slope of Cordillera
Oriental, east of Neiva, Huila, altitude 1800-2300 meters, collected
in flower and fruit August 1-8, 1917, Rusby & Pennell 579, in Herb.
New York Botanical Garden.

Moist banks in forest at an altitude between 1800 and 2300 meters,
Subtropical zone of the western slope of Cordillera Oriental, in Huila.

9. Fagelia nevadensis Pennell, sp. nov.

Stem erect, nearly 1 meter tall, branched, pubescent, lanose
distally, with long glandless white dark-jointed hairs. Leaves 16
cm. long, the blades ovate, cordate, acuminate, doubly and sharply

1920.] NATURAL SCIENCES OF PHILADELPHIA. 177

dentate, (dentate with lobules irregularly and sharply dentate), 10
cm. long, 8-9 cm. wide, each on a petiole 6 cm. long, uniformly
winged throughout (in middle 8-13 mm. wide), entire, proximally
slightly expanding and connate with that of opposing leaf; above
green, beneath slightly paler, slightly pubescent on both surfaces.
Corymb leafy-bracted at base, its secondary branches long, each
dividing above a long peduncular portion into six to eight pedicels.
Peduncles and pedicels hirsute with glandless white hairs and with
short-stalked glands. Calyx 11-12 mm. long, the lobes ovate,
acuminate, entire, pubescent, Corolla: the posterior lip about 5
mm. long, 6 mm. wide, arched, the two lobes united to apex (so
hooded, truncate without apical aperture); anterior lip 13-14 mm.
long, 12-14 mm. wide, with orifice opening much less than one-half
length of sac (sac strongly upcurving toward hood) ; externally gla-
brous, within pubescent at base on anterior side. Filaments .8 mm.
long. Anther straight, 3.5 mm. long, grayish, its walls firm; sacs
contiguous, permanently separated by a firm septum, each opening
by a slit from the distal apex which does not reach the proximal
end. Style 4 mm. long. Capsule 5-7 mm. long, pyramidal, some-
what obtuse, puberulent, with short-stalked glands. Seeds .2-3 mm.
long, irregularly oblong, ridged and transverse-lined, dark-brown.

Type, damp hillside, clearing at Las Nubes, slopes of Sierra Ne-
vada de Santa Marta, Magdalena, collected in flower and fruit
December 15, 1898-1901, Herbert H. Smith 1404, in herbarium
New York Botanical Garden; isotypes in Gray Herbarium, United
States National Museum, and Field Museum of Natural History.
Said to be from "4500 feet" [= 1350 meters], such a low elevation
for a plant of this genus as to force the suspicion that datum is
erroneous. The specimen is more probably from some slope much
higher, surely over 2000 meters altitude.

). nov.

10. Fagelia tolimensis Pennell. sp.

Stem lax, ascending. 1 to 2 meters long, branched, reddish, her-
baceous, glabrate below, above slightly hirsute with dark-jointed
hairs. Leaves 8-10 cm. long; the blades triangular-lanceolate,
cordate or truncate, acuminate, irregularly and somewhat doubly
dentate (dentate with lobules shallowly triangular and irregularly
acutely toothed), 6-7 cm. long, 2.5-3.5 cm. wide; each on a petiole
2-3 cm. long, broadly winged throughout (in middle 12-15 mm.
wide), irregularly shallowly crenately dentate, proximally expanding
and connate with that of opposing leaf; above dark-green, minutely

178 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

pubescent, becoming glabrate, beneath pale-green, permanently
pubescent, especially on the veins. Corymb leafy-bracted at base,
its secondary branches soon much branched (pedicels long and
slender) Peduncles and pedicels hirsute with dark-jointed hairs,
these of various lengths. Calyx 8-10 mm. long, the lobes ovate,
acuminate, entire, pubescent. Corolla: the posterior lip 3-4 mm.
long, 5 mm. wide, arched, the two lobes united to apex (so hooded,
without median aperture); anterior lip 12-14 mm. long. 10 mm.
wide, with orifice opening about one-half length (sac strongly up-
curving toward hood), externally slightly pubescent to glabrate,
within glabrous throughout. Filaments 1.2 mm. long. Anthers
horseshoe-shaped, brown; each sac 2 mm. long, contiguous, splitting
its entire length, septum between sacs thin, but apparently not rup-
turing. Style 5 mm. long, proximally pubescent. Capsule pubes-
cent with white glandless ha'rs; not seen mature.

Type, moist mossy loam, margin of forest, "Rosalito" (east of
Paramo de Ruiz), Tolima, altitude 2800-3100 meters, collected
in flower December 15-17, 1917, F. W. Pennell 2979; in Herb. New
York Botanical Garden.

Nearest to Fagelia purpurascens (Sodiro) Pennell, comb, now,
of Ecuador, but appears distinct in having leaves more sharply cut,
stem less pubescent, pedicels more slender and sepals shorter.

Moist soil, edge of forest, at an altitude between 2800 and 3100
meters, Temperate zone of eastern slope of Cordillera Central, in
Tolima.

11. Fagelia perfoliata (L. f.) Kuntze.

Calceolaria perfoliata L. f., Suppl. 86. 1781. "Habitat in Nova Granada.

Mutis." Type probably from Bogota.
Fagelia perfoliata (L. f.) Kuntze, Rev. Gen. 460. 1891.

Stems lax, ascending, 1-2 meters long, somewhat branched, red-
dish-brown, pubescent throughout, above densely so, with hairs not
or slightly dark-jointed. Leaves 8-13 cm. long, the blades tri-
angular-lanceolate to ovate, cordate to truncate, acuminate, irregu-
larly and somewhat doubly dentate (dentate with lobules shalbwly
triangular and shallowly dentate), 5-8 cm. long, 3.5-5.5 cm. wide;
each on a petiole 3-5 cm. long, broadly winged throughout (in
middle 15-20 mm. wide), slightly crenate -dentate to entire, prox-
imally expanding and connate with that of opposing leaf; above
green, beneath whitish green, pubescent on both surfaces, densely
canescent beneath. Corymb leafy-bracted at base, its secondary
branches soon much branched (pedicels long and slender). Pe-

1920.] NATURAL SCIENCES OF PHILADELPHIA. 179

duncles and pedicels villous, with spreading dark-jointed hairs,
these mostly long. Calyx 9-11 mm. long, the lobes ovate, acumi-
nate, entire, pubescent. Corolla: the posterior lip 4-5 mm. long,
5-6 mm. wide, arched, the two lobes united to apex (so hooded with-
out median aperture); anterior lip 13-14 mm. long, 7-8 mm. wide,
with orifice opening about one-half length (sac strongly upcurving
toward hood) , externally slightly pubescent to glabrate. within
glabrous throughout. Filaments about 1 nun. long. Anthers horse-
shoe shaped, brown, each sac 1.8 mm. long, contiguous, splitting
from proximal end its entire length, septum between sacs thin and
rupturing. Style 5 mm. long, proximally pubescent. Capsule 5
mm. long, urceolate -pyramidal, acutish, pubescent with white gland-
less hairs. Seeds .5 nun. long, lanceolate, distallj' acuminate, ridged,
and transverse-lined, brown.

Moist bushy slopes, along streams in shrub-zone, at altitudes of
27(H) to 3000 meters, Temperate zone of western slope of < 'ordil'era
Oriental, in Cundinamarca.

Cundinamarca. Rio San Cristobal, near Bogota (bushy moun-
tain-slope, alt. 2800-3000 m.), Pennell 2380: Chipaque (moist road-
bank above, alt. 2800 2900 mj, Pennell 1317: Sibate" (bushy slopes
near, alt. 2700-3000 m.), Pennell 2485; Bogota alt. 2700 m.),
J. Triana (U, Y).

12. Fagelia ovata smith) Kiintz.

Calceolaria <»;it<i Smith, Ic. Ined. 1:3 pi. 3. 1789. "ExPeru semina setulit

Dombey."
Fagelia ovata (Smith) Kuntz, Rev. Gen. 4(30. 1891.

Stem erect or ascending, 1 4 dm. tall, branched, pubescent with
white gland-tipped hairs. Leaves 2.5 3.5 cm. long, the blades
ovate, narrowed to neatly truncate at base, acute, obscurely simply
serrate with shall >w teeth, 2.3 3 cm. long, 1-1.5 cm. wide, each on a
petiole .2 .5 cm. long, not winged, lanose with gland-tipped hairs;
green, 1 eneath slightly paler, pubescent on both surfaces. In-
florescence apparently axillary; actually at each node two pedicels
occur, at right angles to which are two leaves from the axil of one of
which usually develops the branch which indefinitely repeats this
manner of branching. Pedicels pubescent with gland-tipped hairs.
( alyx 3 mm. long, the lol es ovate, acutish, entire, pubescent. Cor-
olla: the posterior lip 3-4 mm. long, 4-5 mm. wide, arched, the two
lobes united thioughout, distally attenuate to a slightly notched
apex (so hooded); anterior lip 5 mm. long, 5 mm. wide, widest at
base, with orifice opening much less than one-half length of sac

180 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

(sac, upcurved toward hood, its anterior surface incurved forming
a broad pouch into which anthers shed pollen and into which the
stigma grows, apparently ensuring self-pollination); externally gla-
brous, within with a few hairs about base. Filaments 1.5 mm.
long. Anthers .6 mm. long elliptic or hemispheric, yellow, the
walls thin; sacs contiguous by a broad contact, opening throughout.
Style 1.2 mm. long. Capsule 5-6 mm. long, narrowly pyramidal,
obtusish, thin-walled sparsely puberulent with short-stalked glands.
Seeds .1-2 mm. long, oval, obtuse, ridged and transverse-lined,
brow n.

Type, moist roadside below Chipaque, Cundinamarca, altitude
1800-2200 m., collected in flower August 23, 1917, F. W. Pennell
1327; in Herb. New York Botanical Garden.

Moist shaded banks, at altitudes of 1300 to 2200 meters, Sub-
tropical zone of eastern slope of Cordillera Oriental, in Cundina-
marca.

Cundinamarca. Chipaque (moist roadside below, alt. 1800-2200
m.)i Pennell 1327; Quetame to Monte Redondo (moist cliff, along
stream in woodland, alt. 1300 m.), Pennell 1352, (moist bank, alt.
1400-1500 m.), Pennell 1854.

13. Fagelia micrantha Pennell, sp. nov.

Stem spreading and laxly ascending, 2-3 dm. long branched,
sparsely pubescent below with short few-celled gland-tipped hairs.
Leaves 2-4 cm. long, the blades acute or obtusish at apex, irregu-
larly pinnately lobed with 1-2 pairs of segments, the incisions usually
reaching \ to f the distance to the midrib, the lobules and main
portion of the blade irregularly dentate, 1 5-3 cm. long, 1.5-2.5 cm.
wide; each on a petiole .5-1 cm. long, slightly winged, proximally
clasping stem and slightly connate with that of opposing leaf; abo\e
deep-green, with sparse scattered pubescence soon becoming gla-
brous, beneath pale-green and nearly glabrous. Corymb leafy -
bracted at base (the two primary flowers developed), the secondary
branches scarcely or not longer than the primary pedicels, usually
once dividing, and bracted with reduced leaves. Peduncles and
ped"cels pubescent with short gland-tipped hairs. Calyx 3-1 mm.
long, the lobes oblong-ovate, obtuse, slightly serrate, pubescent
proximally. Corolla: the posterior lip about 2 mm long and 2 mm.
wide, arched, the two lobes united and slightly hooded, free toward
apex (leaving a narrow arched aperture into hood), anterior lip
5-6 mm. long, 4-5 mm. wide, with narrow base, hooded almost

1920.] NATURAL SCIENCES OF PHILADELPHIA. 181

entire length (sac strongly upcurving against hood) : externally
glabrous or minutely puberulent at base, within glabrous. Fila-
ments none. Anther with two sacs separated "on two arms of the
connective, both sacs fertile; opening throughout, the anterior pro-
jecting into orifice, smaller; posterior arm (with sac) 1.5 mm. long,-
anterior arm (with sac) 1 mm. long. Style 6-8 mm. long. Cap-
sule glabrous, not seen mature.

Type, along streamlet, edge of forest, "Rosalito" (between Murillo
and Paramo de Ruiz) Tolima, altitude 2800-3100 meters, collected
in flower December 17, 1917, F. W. Pennell 3145; in Herb. Xew
York Botanical Garden. Growing with 3119, F. crenati oba.

Along streamlets, edge of forest, at an altitude between 2800
and 3100 meters, Temperate zone of eastern slope of Cordillera
Central, n Tolima.

14. Fagelia radiculoides Pennell, sp. now

Stem spreading and laxly ascending, 1-15 dm. long, branched,
glabrate, above pubescent with spreading gland-tipped hairs. Leaves
2-10 cm. long, the blades acute to acuminate at apex, pinnately
lobed with usually three pairs of lanceolate segments, the incisions
extending nearly to the midrib, the lobules and terminal segment
irregularly serrate-dentate, 1.5-6 cm. long, 1.5-6.5 cm. wide; each
on a petiole .5-4 cm. long, narrowly winged, proximally slightly
expanding and connate with that of opposing leaf; above deep-green,
with scattered hairs or glabrous, beneath glaucous, glabrous or pu-
bescent on the midrib. Corymb leafy-bracted at base (the two
primary flowers developed), the secondary branches long and re-
peatedly dividing, bracted throughout. Peduncles and pedicels
pubescent with gland-tipped hairs. Calyx 4 5 mm. long, the lobes
ovate, acute to obtusish, obscurely slightly serrate, pubescent prox-
imally and on margin with gland-tipped hairs. Corolla: the pos-
terior lip 1.5-2 mm. long, 2 mm. wide, arched, the two lobes united
and slightly hooded, free toward apex (leaving a narrow arched
aperture into hood); anterior lip 6-7 mm. long, 4-5 mm. wide,
narrowed at base, hooded almost entire length (sac strongly upcurv-
ing against hood); externally and internally glabrous. Filaments
none. Anther with two sacs separated on two arms of the con-
nect ive, both sacs fertile, opening throughout, of about equal size,
each arm (with sac) about 1 mm. long. Style .7-8 mm. long.
C.'ipsule 3-4 mm. long, broadly globose-pyramidal, rounded and
slightly notched, somewhat pubescent with gland-tipped hairs.

182 PROCEEDINGS OF THE ACADEMY OF [May-Oct.

Seeds .5-6 mm. long, oblong, obtuse, ridged (with rounded ridges)
dark-brown.

Type, moist rocky canon, Rio San Francisco, above Bogota,
Cundinamarca, altitude 2700-2800 meters, collected in flower and
fruit September 13, 1917, F. W. Pennell 1942; in Herb. New York
Botanical Garden.

Moist or wet soil, springheads, swales and cliffs, partially shaded
or open, at altitudes of 2600 to 3200 meters, occasionally descending
to 1500 meters, Temperate zone, ascending to Paramo, where
dwarfed, and to Subtropical zone, where more rank; on both slopes
of the Cordillera Oriental, in Cundinamarca.

Cundinamarca. Zipaquira (springhead in meadow, alt. 2650 m.),
Pennell 2534; Mt. Chuscal, west of Zipaquira, (swale on paramo,
alt. 3100-3200 m.), [only 1-2 dm. tall; leaves in some plants more
pubescent], Pennell 2602; Bogota (moist rocky canon on Rio San
Francisco above, alt. 2700-2800 m.), Pennell 1942, (moist bank,
base of mount, alt. 2700-2800 m.), Pennell 2293; Sibate (wet road-
bank, alt. 2600-2800 m.), Pennell 2386; Ubague (moist loam in
shrub-zone abo\e, alt, 2700-3000 m.), Pennell 1898; Monte Redondo
to Quetame (wet bank, alt. 1400-1500 m.) [plant especially rank],
Pennell 1855.

15. Fagelia crenatiloba Pennell, sp. nov.

Stem ascending, 3-6 dm. long, little branched, glabrous or nearly
so, above pubescent with white several-celled glandless hairs. Leaves
5-7 cm. long, the blades obtuse or acutish at apex, pinnately lobed
with 2-3 pairs of segments, the incisions rarely extending over
§-f the distance to the midrib, the lobules and main portion of
blade crenately dentate, 3.5-5 cm. long, 3-4 cm. wide; each on a
petiole 1.5 2 cm. long, slightly margined, glabrous or nearly so,
proximally somewhat expanding and connate with that of opposing
leaf; above green and with scattered pubescence, beneath pale-green
and the midrib and principal veins pubescent. Corymb leafy-
bracted at base (the two primary flowers developed) , its secondary
branches longer, once or twice dividing and bracted throughout
with reduced leaves. Peduncles and pedicels finely pubescent with
several-celled glandless hairs. Calyx 5-6 mm. long, the lobes ovate,
obtuse or acutish, slightly serrate, slightly pubescent, especially
proximally. Corolla: the posterior lip 1-2 mm. long, 2-2.5 mm.
wide, arched, the two lobes united and slightly hooded, free toward
apex (leaving a narrow or triangular slit-like aperture into hood);

1920.] NATURAL SCIENCES OF PHILADELPHIA. 183

anterior lip 5-7 mm. long, 4-5 mm. wide, narrowed at base, hooded
almost entire length (sac strongly upcurving against hood); exter-
nally glabrous or finely puberulent proximally, within glabrous.
Filament none or very short. Anther with the two sacs separated
on two arms of the connective, each (including sac) about 1.5 mm.
long; posterior sac .6-7 mm. long, whitish, opening throughout,
fertile, concealed within hood; anterior sac shorter, projecting into
orifice, partially or wholly sterile. Style 1 mm. long. Capsule
nearly globose, finefy pubescent with glandless hairs; not seen
mature.

Type, along streamlet, edge of forest, "Rosalito," (between
Murillo and Paramo de Ruiz), Tolima, altitude 2800-3100 meters,
collected in flower December 17. 1917, F. W. Pennell 3119; in Herb.
New York Botanical Garden.

Along streamlets, edge of forest, at an altitude between 2800 and
3100 meters, Temperate zone of eastern slope of Cordillera Central,
in Tolima.

16. Fagelia chelidonioides (H. B. K.) Kuntze.

Calceolaria chelidonioides H. B. K., Nov. Gen. el Sp„ 2:378. 1818. "Crescil
in radicibus inontis Javirac prope Quito, alt. 1500 hex. [ = ca. 2850 m.]
[Humboldt & Bonpland]."

Fagelia chelidonioides (H. B. K.) Kuntze, Rev. Gen. 2: 459. 1891.

Fagelia diversifolia Pennell, in Addisonia 4: 73, pi. 163. "Type . . collected
on a moist bank at Chipaque. Department of Cundinamana. Colombia, at
an altitude of about 8700 feet, August 23, 1917, my number 1320, and is
preserved in the hebarium of the New York Botanical Garden." Specimens
seen later appear to unite this with the plant from Ecuador.

Stem erect or ascending, 3-9 dm. tall, little branched, sparsely
pubescent, more so about nodes, with few-celled gland-tipped hairs.
Leaves 4-15 cm. long, the blades acute to acuminate at apex, the
lower ovate and shallowly pinnately lobed, irregularly serrate-
dentate, the upper pinnately lobed nearly to the midrib with two or
three pairs of oval or ovate, irregularly doubly serrate-dentate
segments, (the odd terminal segment largest), 3-10 cm. long, 2.5-
7.5 cm. wide, each on a petiole 1-5 cm. long, slightly margined,
glandular pubescent, proximally slightly expanding and clasping the
stem, usually slightly connate with that of opposing leaf; green
above, pale green beneath, with scattered pubescence on both
surfaces. Corymb leafy-bracted at base (the two primary flowers
developed), its secondary branches elongated, several times dividing
and bracted throughout with reduced leaves. Peduncles and pedi-
cels finely pubescent with few-celled gland-tipped hairs. Calyx
7-8 mm. long, the lobes ovate, acuminate, slightly serrate (the

184 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

few serratures not callous-tipped), slightly pubescent on the back,
and conspicuously ciliate with gland tipped hairs. Corolla: the
posterior lip about 3 mm. long, 3-4 mm. wide, arched, the two
lobes united and hooded, but not to apex (leaving a narrow slit like
aperture into hood), anterior lip 10-15 mm. long, 11-13 mm. wide,
narrowed at base, hooded almost entire length (sac strongly up
curving against hood); externally glabrous or puberulent on pos-
terior lip, within pubescent about base and within posterior lip.
Filament none. Anther with the two sacs separated on two arms
of the connective, each about 1.5-1.8 mm. long; posterior sac 1.2
mm. long, yellowish, opening throughout, fertile, concealed within
hood; anterior sac short, orange yellow, sterile, projecting into the
orifice, the club-like dark connective arm serving as a lever against
which entering insect pushes, thus forcing the fertile sac out through
the slit like aperture of the hood and against back of insect. Style
1.1-1.3 mm. long. Capsule 8 mm. long, globose-oyramidal, obtuse,
pubescent with short gland-tipped hairs. Seeds .6-7 mm. long,
oblong, obtuse, ridged (with high rounded ridges), brown.

Moist soil, roadside ditches and banks, frequently cultivated and
possibly introduced from Ecuador, at altitudes of 2000 to 2700
meters, Subtropical zone of eastern slope of Cordillera Oriental,
in Cundinamarca; also obviously from cultivation at Bogota and
at "Balsillas", east of Neiva in Huila; also from the Subtropical zone
of the Cordillera Occidental, in Valle. In Ecuador.

Cundinamarca. Chipaque (moist bank, alt. 2600-2700 m.),
Pennell 1320; Ubague (moist soil, alt. 2000-2500 m.) , Pennell 1877;
[Zipaquira (moist ditch on hill — alt. 2900 m.), evidently escaped
from garden, Pennell 2567].

[Huila. "Balsillas," on Rio Balsillas (cult, in garden, alt. 2000-
2100 m.), Rusby & Pennell 692.]

Cauca. Cuesta de Tocota, road from Buenaventura to Cali,
western Cordillera; alt. 1500-1900 m., H. Pittier 698 (U).

17. Fagelia scalaris Pennell, sp. nov.

Stem erect or ascending, about 1 meter tall, little branched,
pubescent to hirsute above with many-celled dark-jointed not or
scarcely gland-tipped hairs. Leaves 5-8 cm. long; the blades acumi-
nate at apex, pinnatisect (cut nearly to midrib) with 2 or 3 pairs
of lanceolate-ovate, irregularly serrate-dentate segments (the odd
terminal segment largest), 4-6 cm long, 3-5 cm. wide; each on a
petiole 1-2 cm. long, very narrowly winged, hirsute, proximally

1920.] NATURAL SCIENCES OF PHILADELPHIA. 185

slightly expanding and connate with that of opposing leaf; green,
beneath paler, with scattered pubescence on both surfaces Corymb
leafy-bracted at base (the primary flowers sometimes not developed),
its secondary branches (one or both developed) long, once or twice
dividing and bracted throughout with reduced leaves. Peduncles
and pedicels hirsute with yellowish-white dark -jointed hairs. Calyx
6 mm. long, the lobes ovate, acute, serrate (with shallow callous-
tipped serratures), densely hirsute. Corolla: the posterior lip
2-3 mm. long, 4 mm. wide, arched, the two lobes united and hooded,
but not to apex (leaving a slit- like aperture into hood), anterior
lip 15-21 mm. long, 14-18 mm. wide, narrowed at base, with narrow
orifice, hooded almost entire length (sac strongly upcurving against
hood); externa'ly pubescent with short hairs on posterior lip, with
longer hairs on base of anterior lip, within pubescent at base. Fila-
ments very short or not developed. Anthers straight, 4 mm. long,
the two sacs separated on two arms of the connective; posterior
sac 1.5 mm. long, yellowish, opening throughout, fertile, concealed
within hood; anterior sac short, orange-yellow, sterile, projecting
into the orifice, the club-like dark connective-arm serving as a lever
against which entering insect pushes, thus forcing the fertile sac
out through the slit- like aperture of the hood and against back of
insect. Style 2 mm. long. Capsule 6 mm. long, globose pyramidal,
obtuse, pubescent with glandless hairs. Seeds .6-.7 mm. long,
oblong, obtuse, ridge-angled, brown.

Type, swale, "Balsillas," on Rio Balsillas, altitude 2000-2100
meters, collected in flower and fruit August 3, 1917, Rusby & Pen-
nell 710; in Herb. New York Botanical Garden.

Swales, at an altitude of 2000 to 2100 meters, Subtropical zone
of eastern slope of Cordillera Oriental, in Huila.

18. Fagelia pinnatisecta Pcnnt'll, sp. now

Stem ascending, 2-6 dm. tall, little branched, pubescent above
with many-celled dark-jointed not or scarcely gland-tipped hairs.
Leaves 3-5.5 cm. long, the blades acuminate at apex, pinnatisect
(cut nearly to midrib) with three pairs of lanceolate, irregularly
serrate -dentate segments, (the odd terminal segment largest) 2-3.5
cm. long, 2-3 cm. wide; each on a petiole 1-2 cm. long, narrowly
winged, somewhat hirsute, proximally slightly expanding and con-
nate with that of opposing leaf; green and pubescent above, beneath
pale, and hirsute pubescent on the main veins, sparsely so over sur-
face. Corymb leafy-bracted at base (the primary flowers not

186 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

developed), its secondary branches long, once or twice dividing and
leafy-bracted throughout with reduced leaves. Peduncles and pedi-
cels hirsute with yellowish-white dark-jointed hairs. Calyx 7-9
mm. long, the lobes lanceolate or narrowly ovate, acuminate, serrate
(with not or scarcely callous-tipped serratures), hirsute, especially
proximally. Corolla: the posterior lip 2-3 mm. long, 2-3 mm.
wide, arched, the two lobes united and hooded, but not to apex
(leaving a narrow aperture into hood) ; anterior lip 8-10 mm. long,
6-7 mm. wide, narrowed at base, hooded over f length (sac strongly
upcurving against hood); externally and within slightly pubescent
about base. Filaments very short or none. Anther with the two
sacs separated on two arms of the connective, each about 1.6-1.8
mm. long; posterior sac 1.2 mm. long, yellowish, opening through-
out, fertile, concealed within hood; anterior sac short, orange-yellow,
sterile, projecting into the orifice, the club-like dark connective-
arm serving as a lever as in F. scalaris. Style nearly 2 mm. long.
Capsule 4 mm. long, broad-globose, rounded, pubescent with gland-
less and some gland-tipped hairs. Seeds .6-7 mm. long, oblong,
obtusish, ridge-angled, brown.

Type, swale, "Balsillas," on Rio Balsillas, altitude 2000-2100
meters, collected in flower and fruit August 3, 1917, Rusby & Pen-
nell 721, in Herb. New York Botanical Garden. From the same
swale as 710, F. scalaris.

Swales, at an altitude of 2000 to 2100 meters, Subtropical zone
of eastern slope of Cordillera Oriental, in Huila.

21. RUSSELIA Jacquin.
Russelia Jacq., Enum. PL Carib. 25. 1760.
Type species, R. sarmentosa Jacq., of Cuba.

1. Russelia colombiana Pennell, sp. nov.

Herb, or shrubby below, diffuse, reaching 5 feet long. Stem
6-angled below, sharply 4-angled above, glabrous or with sparse
pubescence. Leaves in threes, the upper opposite, ovate, 5 cm.
long, 3 cm. wide, truncate at base, strongly acuminate, sharply
serrate-dentate with ascending teeth (bracts lance-ovate, coarsely
toothed), glabrous nearly from the first, green, with brown wax dots
on upper surface. Inflorescence much elongated, of axillary cymes.
Cymes hirtellous, 5-15-flowered. Calyx 4 mm. long, with brown
wax dots, its lobes narrowly ovate with caudate pubescent tips
nearly equaling the length of the body, slightly pubescent. Corolla
red, 10-11 mm. long, its lobes 1.5 mm. long, the posterior united
|-| their length, externally glabrous, within on anterior side pubes-

1920.] NATURAL SCIENCES OF PHILADELPHIA. 187

cent with yellow hairs. Stamens and pistil glabrous throughout.
Capsule brown, globose-ovoid, 4.5-5 mm. long, with a slender beak
1-1.5 mm. long.

Related to R. sarmentosa Jacq. of Cuba, which differs in the
stem being 4-angled, its leaves smaller, with rounded teeth and
obtuse at apex, its sepals with shorter caudate tips, its corollas
slightly larger, 12-14 mm. long, and its capsules smaller, excluding
the beak, only 4 mm. long.

Type, in mountain forest, on the Agua Dulce road, between
Santa Marta and the Sierra Nevada, altitude 450 meters [ = 1500
feet], collected in flower and fruit November 22, 1898, Herbert H.
Smith 1361; in Herb. New York Botanical Garden; isotypes in
United States National Herbarium, Gray Herbarium and Field
Museum of Natural History.

Forest, at an altitude of 450 meters, Tropical zone on lower slopes
of Sierra Nevada de Santa Marta in Magdalena.

22. ANGELONIA Humboldt and Bonpland.
Angdonia Hurab. & Bonpl., PI. Aequin 2: 92. 1809.
Type species, A. salicariaefolia H. & B.

1, Angelonia salicariaefolia Humb. & Bonpl.

A. salicariaefolia Humb. & Bonpl., PL Aequin. 2: 92. pi. 108. 1809. "Hab-
itat in America meridionali ad Caracas." . . . ''Croit sur les collines
arides de gneiss, qui avoisinenl la ville <le Caracas, a une hauteur de cinq
ou six cents toises [ca. 1000-1200 m.] ou-dessus du niveau de 1 'ocean."
Specimen from Caracas, Otto Kuntze 1407, seen in Herb. New York
Botanical Garden.

Gravelly slopes, rather moist, along the lower western slopes
of the Cordillera Oriental, and the similar eastern slopes of the
Cordillera Central, doubtless continuously encircling the upper
Magdalena Valley; in the Cauca valley; extending eastward along
the northern lower slopes of the Venezuelan Andes; at altitudes
of 450-1400 meters. Tropical zone.4

Cundinamarca. Anapoima, J. Triana (Y); Fusagasuga, I. F.
Holton 577 (Y); Fusagasugd to Pandi, Pcnnell 2714 (Y); Icononzo,
Pennell 2761 (Y).

Tolima. San Lorenzo, (first foothill of Cordillera Central, west
of), Pennell 3517 (Y).

4 Angelonia angustifolia Benth.

Specimens collected from plants cultivated at "Medellin," on the bank of
the Rio Hinu, Bolivar, Pennell 4141 (Y) appear to be this commonly cultivated
species of Mexico. The two plants are readily distinguished:

Herbage densely glandular-pubescent. Leaves lanceolate., clasping at base.

1. .1 . salicariaefolia.
Herbage glabrous or sparsely glandular-pubescent. Leaves linear-lanceolate,
narrowed at base. 2. A. angustifolia.

188 PROCEEDINGS OF THE ACADEMY OF [May-Oct.,

Huila. Neiva (open slope of first foothill of Cordillera Oriental,
east of), Rusby & Pennell 1082 (Y); Quebracla de Angeles, above
Natagaima, Rusby & Pennell 284 (Y).

Valle. Cali, H. Pittier 632 (V).

23. LINARIA Miller.
Linaria Mill., Gard. Diet. ed. IV. 1754.
Type species, Antirrhinum linaria L., Sp. PI. 616. 1753, of Europe.

1. Linaria texana Scheele.

Linaria texana Scheele in Linnaea 21: 761. 1848. "Zwischen Houston
und Austin [Texas] haufig: Romer." Type not seen nor verified, but
description evidently of plant here characterized.

Meadow-land, on the Sabana of Bogota, at an altitude of 2600
to 2650 meters, certainly introduced. Widespread through western
temperate North America, and collected extensively in Andine and
Temperate South America, probably always as a weed.

Cundinamarca. Sibate (meadow on sabana, alt. ),

Pennell 2469; Hacienda de Tequendama, I. F. Holton (Y).

_

%

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 228

A GRAFT-CHIMERA IN
THE APPLE

A. B. STOUT

NEW YORK
1920

Reprinted, without change of paging, from Journal of Heredity, Vol. XI, No. 5,

May-June, 1920

A GRAFT-CHIMERA IN THE APPLE

Evidence That the Two Distinct Types of Fruits on the Same Tree Are Not
Due to Bud Sporting or Top-Grafting

A. B. Stout

New York Botanical Garden, New York City

| HE accompanying plate illus-

; trates two distinct sorts of fruits

M. i

T

borne on an apple tree that has
evidently never been top-grafted. The
fruit shown at the right is typical of
the King variety; the other is nearly
identical with the fruits of the Rock-
bury Russet. The two types of fruit
are quite distinct in respect to size,
color, character of skin, flavor, and
texture, and the leaves of the branches
bearing them are noticeably different
especially as to size.

The tree which bears these two
kinds of foliage and fruits stands in
the vicinity of Geneva, N. Y., in an
orchard owned by Mr. T. D. Whitney.
Mr. Whitney helped plant the tree in
1862, has resided on the place ever
since, and has for many years observed
the dual nature of the tree.1

At the present time the tree is large
and well developed and is about 30
feet in height and in spread. Most
branches bear the Russet fruits. About
20 of the smaller branches bear King
fruits and these branches are well
scattered, being found among the tip
branches of all of the large main divi-
sions of the trunk.

EVIDENCE OF THIS AS A CHIMERA

The occurrence of two more or less
distinct kinds of fruit on the same tree
may be due to any one of three causes,
as follows: (1) vegetative variation or
bud sporting, (2) the usual consequence
of top-grafting, or (3) an unusual and
somewhat indirect result of grafting,
which gives a plant in which the two
kinds of cells belonging to stock and
scion become associated together in
the same branches, giving what is now
known as a chimera.

Dr. U. P. Hedrick, of the Geneva
Experiment Station, is convinced that
bud sporting has not occurred in the
tree in question. He does not consider
it probable that these two types of
fruit which differ so widely in several
characters can be so closely related as
to be parent stock and bud sport.

Mr. Whitney is positive that the
tree was never top-grafted. His con-
tinuous association with the orchard
from the time of its planting to date,
his wide knowledge of apples, and his
definite recollection of this particular
tree make this point seem certain. The
scattering branches which bear King
fruits have not grown from King scions
that were grafted to branches of a
Russet tree.

CHIMERAL FRUITS FOUND

Very definite evidence that the King
branches are not simple top-grafts is
also seen in the chimeral fruits which
are sometimes found. Such fruits have
a segment that is King while the rest
are Russet. These show that the two
cells of the King and Russet varieties
are associated together in single twigs,
leaves and fruits. Such a combination
has thus far not been produced as a
direct result of grafting but chimeras
of this class or rank very frequently
occur in bud sports, they have now
been produced experimentally, and
they have also arisen incidentally as
indirect products of grafting, giving
what has frequently been called "mixed
plants."

Several cases of "mixed plants" have
long been known to horticulturists and
for many years these plants aroused
much discussion as to their origin and
nature. One of these is the Bizzaria or -

1 The writer recently had the pleasure of visiting this interesting tree together with Dr. U. P.
Hedrick, Mr. O. M. Taylor, and Mr. W. O. Gloyer, of the staff of the New York Experiment
Station, all of whom shared in the observations recorded here.

Stout: A Graft-Chimera in the Apple

135

ange which appeared in Florence, Italy
about 1644. On some of its branches
oranges are produced, on others citrons
develop, and on other branches the
fruits are part orange and part citron
as shown on page 522 of Vol. 5 of this
Journal (December, 1914). Another
plant known as Cytisus Adami origin-
ated in 1825 as a branch of a plant
grown from a graft between Cytisus
pur pur ens and Laburnum vulgar e.
Some branches of this plant are pure
Cytisus purpureus, others are pure
Laburnum vulgar e, others are various
mixtures of the two, and others bear
leaves that are intermediate in charac-
ter. Several types of plants are known
which appear to be mixtures of the two
species Crataegus monogyna and Mespi-
lus germanica. All of these have been
propagated vegetatively and have been
objects of more than usual interest.
How these plants originated has been
a matter of no little discussion and
speculation among botanists and horti-
culturists.

GRAFTING TO PRODUCE CHIMERAS

The experimental production of
chimeras in rather recent years has
shown very clearly how such plants can
arise incidentally through the practice
of grafting. By a simple but ingenious
arise incidentally through practice of
grafting. By a simple but ingenious
method of experimentation Dr. Wink-
ler, now Director of the Botanical
Garden at Hamburg, produced such
plants under observation and control.
He used the tomato and nightshade,
two distinct and well-known species
with marked differences in leaves,
flowers and fruit. He made grafts,
and when the scions were well estab-
lished he decapitated the branches by
cutting through the points of contact
between scion and stock, thus exposing
on the cut surface the two kinds of
tissue and the lines of contact between
them. On this surface a callus formed
from which buds arose. If a bud arose
entirely from the part that was night-
shade the branch was nightshade only;
if from tomato tissue the branch was
pure tomato. If, however, a branch

arose over the line of juncture it was
composed partly of tomato and partly
of nightshade tissue.

TWO KINDS OF CELLS IN SAME BRANCH

Such branches were called "chime-
ras." The simpler of these show
vertical lines of differentiation; one
sector bears the leaves, flowers, and
fruits of the tomato while the rest of
the branch bears those of the night-
shade. Such branches and the plants
grown from them by vegetative propa-
gation are called "sectoral chimeras."
In such an association of two kinds of
cells, each sort retains its own character
and the leaves, flowers, and fruits of
the two sectors are readily to be
identified.

Occasionally, however, branches
arose which produced leaves, flowers
and fruits that were intermediate or
mixed in character. One of these
(named Solatium tubingense) bears
simple leaves like the nightshade but
the leaves are more or less lobed and
are hairy as the tomato. Another
(Solatium proieus) resembles the to-
mato more than the nightshade; the
stems and leaves are hairy but the
fruits are smaller than those of the
tomato. Several types of intermediates
were produced and for a time it was
believed that these were true graft-
hybrids resulting from the actual
fusion of certain vegetative cells of the
tomato and the nightshade in the
region of contact in the graft. Later,
however, a study of the internal and
minute structure of the cells in these
plants revealed that the two kinds of
cells characteristic of the tomato and
the nightshade are both present, and
that one kind exists as one or more
continuous layers covering the other
kind. It was found that Solatium
tubingense has one outside or periclinal
layer of tomato cells covering a core of
nightshade tissue and that 5. proteus
has two such periclinal layers. Other
types of the intermediates have one or
more outermost layers of nightshade
cells covering tissue of the tomato.
The plants were thus found to be
periclinal chimeras. A photo of four

236

The Journal of Heredity

of these intermediates or periclinal
chimeras is shown in The Journal of
Heredity, Vol. 5, No. 12, and an excel-
lent discussion of how such chimeras
are produced is there given.

Similar study of Cytisus Adami and
the Crataegus- Mespilus so-called graft-
hybrids showed that they are also
periclinal chimeras with one or more
cell-layers of one species covering a
body of cells of the other species.

INTERACTION OF DIFFERENT KINDS OF

CELLS

The rather intimate association of
the cells of two different species in the
same stem, leaf, flower, and fruit is
especially interesting in view of the
mutual interactions, mechanical and
physiological, that may occur. The
evidence indicates that the two kinds
of cells remain independent in respect
to their own hereditary characteristics,
but in the periclinal chimera they
interact producing organs that are
intermediate in character. Slight dif-
ferences in the arrangement of the
same two kinds of cells also produces
vegetative and floral structures and
fruits that are quite different in
appearance.

While the investigations indicate
that the supposed graft-hybrids are
in nature only periclinal chimeras, the
production of true graft-hybrid through
a fusion of vegetative cells is still to
be considered possible. In fact Dr.
Winkler has presented evidence that
this is the case in one adventitious
branch that arose from a decapitated
graft between the tomato and the
nightshade.

The production of sectoral and peri-
clinal chimeras by experimental means
readily shows how such branches can
arise incidentally on plants grafted by
the ordinary methods. New and
adventitious buds may arise from the
region of union between stock and
scion, especially if the upper part of
the scion dies. If such a bud arises
over the line of contact it may develop
as a chimera. Dr. Winkler's studies
show that sectoral associations develop
more frequently than do the periclinal.

CHIMERAS THROUGH BUD SPORTING

It should be noted that intra-varietal
sectoral and periclinal chimeras fre-
quently arise through bud sporting.
Such partial bud sports have been
described frequently in the pages of
this journal, especially for such
conspicuous cases as the loss of green
color which is carried on into various
new branches.

Some of the albomarginate types of
variegated plants (of Pelargonium for
example) are clearly periclinal associa-
tions of green and white cells all of
which belong to the particular variety.
In the original sport, most probably, a
single cell lost the ability to produce the
green pigment and this cell was so
placed in the growing point that its
multiplication by division gave one or
more layers of white cells. Once
established the relative position of the
white and the green cells was main-
tained rather uniformly in the subse-
quent development of branches.

It is also to be noted that, in many
sorts of variegated plants, though the
pattern strongly simulates a chimeral
arrangement, it is really due to physio-
logical conditions affecting develop-
ment of color in the leaf as a whole.
In these the colored and colorless
areas often cut across cell layers or
histogens. Much remains to be
learned regarding the development of
such local areas of infectious chlorosis,
as well as the development of various
patterns of anthocyanin coloration in
flowers and foliage.

PROBABLE ORIGIN OF THIS CHIMERA

Mr. Whitney's tree bearing Russet
and King fruits is evidently a chimera
that has grown from a bud that arose
on the line of contact between scion
and stock, and is hence to be considered
as a graft-chimera. Presumably a
scion of King was grafted to the root
of a seedling of the Russet in the
method of ordinary nursery practice.
At any rate the tree is clearly an inter-
varietal chimera: some branches are
composed of sectors of tissue of two
varieties; lateral branches arising from

Stout: A Graft-Chimera in the Apple

237

the sector of King are pure for King;
those arising from the part that is
Russet are pure for that variety, and
those that happen to arise from the
line of contact continue to be sectoral
chimeras. It is possible that in some
of the branches the two kinds of tissue

are in periclinal relationship and that
some fruits possess a skin of one
variety and a core or body of the other.
Rather careful examination of a large
number of fruits by one competent to
judge the flavor would be necessary to
determine this point.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 229

A BOTANICAL EXCURSION
TO THE BIG CYPRESS

By JOHN K. SMALL

NEW YORK
1920

Reprinted, without change of paging, from Natpbal History, 20
488-500. September-October, 1920.

A Botanical Excursion to
the Big Cypress

By
JOHN KUNKEL SMALL

Head Curator of the Museums and Herbariums of the New York Botanical Garden

[Reprinted from Natural History, Vol. XX, No. 4, pp. 488-500, 1920]

A Botanical Excursion to the Big Cypress

By JOHN KUNKEL SMALL

Head Curator of the Museums and Herbarium of the New York Botanical Garden

-

THE most extensive physiographic
trinity or the largest prairie-marsh-
swamp region, and at the same time
the least known area in the eastern United
States, is in southern Florida. The "Big
Cypress," or the Big Cypress Swamp, lies
south of the Caloosahatehee River between
the Everglades and the Gulf of Mexico. The
greater part of our population is ignorant
even of this geographic designation. To the
few who have seen it printed on maps the
name signifies nothing, or conveys but a
vague idea. Only a score or two of survey-
ors, hunters, and prospectors, out of the
hundred million inhabitants of the United
States, have any definite knowledge of its
physical geography.

The second week of May, 1917, we were
on the very edge of the Big Cypress when we
navigated Lake Hiepochee during a cruise to
Lake Okeechobee. The day we returned from
that cruise, which was described in former
papers,i an opportunity to explore some of
the mysteries of the Big Cypress unexpectedly
presented itself. Mr. W. Stanley Hanson, a
bird inspector with the United States Bio-
logical Survey, and a naturalist well ac-
quainted with the Big Cypress, had come to
Miami across country from Fort Myers,
whence he was about to retrace his course.
The opportunity to accompany him on a trip
through largely unknown territory was a
temptation too great to be resisted. Conse-
quently, we prepared a Ford for a week's
run, and the next day set out for Fort
Myers. Miami and Fort Myers are about
120 miles distant from each other, in a di-
rect line, but the intervening area could
have been conveniently, or at least expedi-
tiously, traveled only in an aeroplane. The
shortest course possible for us followed a
curve more than 250 miles in length.

In order to bring us to our most distant
objective which lay across the Everglades
only about sixty miles from Miami, we had
to make a detour around the Everglades

1 Journal of the New York Botanical Garden,
Vol. XIX, 1918, pp. 279-90. The American
Museum Journal, Vol. XVIII. 1918, pp. 684-
700

488

and Lake Okeechobee at their head. The
facilities for making an examination of the
country and a collection of specimens of its
vegetation were generously furnished by Mr.
Charles Deering, of Miami.

The first stage of our course lay along the
eastern coast of Florida between Miami and
Fort Pierce. Miami, Fort Pierce, and Fort
Myers are about equidistant one from the
other, or, straight lines connecting the three
places would form an equilateral triangle.
The territory included in the triangle, made
up mostly of everglades, prairies, cypress
swamp, and pineland, together with Lake
Okeechobee situated near one side of the
triangle, was essentially uninhabited, ex-
cept for the scattered settlements in the
Caloosahatehee River region. Between
Miami and Fort Pierce pinelands and sand
dunes (scrub2) predominate; between Fort
Pierce and Fort Myers are pinelands and
prairies; while between Fort Myers and
Miami lie prairies, cypress swamps, and the
Everglades.

Mr. Hanson preceded us to West Palm
Beach, where we overtook him. It was late
in the evening when we reached Stuart,
where we had to spend the night because of
a high wind which made the ferryman hesi-
tate to carry us across the Saint Lucie River.
An early start the next morning brought us
to Fort Pierce in time for breakfast. Thence
we started on the second leg of the tri-
angle, proceeding in a southwesterly direc-
tion.

Between Miami and Fort Pierce our
course took us through not fewer than
forty towns.3 After leaving Fort Pierce
only four settlements were encountered, two
established settlements and two embryonic
colonies.

After Fort Pierce disappeared from view
we sped westward through pinelands and
across the Halpatiokee Swamp, where count-
less turtles and snakes basked in the sun
about the water pools that lined the road.

- These are quiescent inland dunes of snow-
white sand.

3 These lie outside of the triangle of unin-
habited territory referred to in a previous para-
graph.

■ a> a ? 5

i 2 S s '-5

bo

bo

ts ce a •-

• •£ ~, c3 0*
"£&* 5 « 2

s -a"5

P g (8 O.

£3 c3 :3

UJ
DC
<

t runs

ately re
or cypi
all aqu

Q.

■ / - - —
s 8.Sfl o

Z

<

S«4-l - ' —

a

z

M H 'S'S H

LU

is-s fc ° —

5 a ft «» -g

I

C 3 j- M

i-

^ o bcg-d

.2 - O — <£

O H S - «■=

UJ n « a 5 J?

cc ■"-^_, M
I— - ai o e £

co g a* * g

] o a>

■2 S'o

C3 ~ o

Z S i J- M *

^ - c Z~

a 2

- = ^

O Sj o"S g

Z &-.x =.= =

< ■ -s g s ►,

I- £.=! li-a

< -5 a-1 ais

— « £* =

= ■ - 2 B •

: £3:

0 (D -n

o > '" " r-

S?o °
" "3 a? a>

tu

1!

' I S3 fl
•5--0 «

•3*9

h" bOjS j*

o 1a *j "■ ■

5 b »
■3 «

©_ > H

e Ml 3 i-4 ■

•5 . O" bo

c3 s u . C

® «j T3 O

•■= oS«""

3 * 2 S
■SsflJ

4'JU

XATUHAL n/STOh'V

After crossing the swamp another stretch of
sandy pine woods was traversed with diffi-
culty, as the combined power of the engine
and the pushing ability of the occupants of
the car were necessary to get through the
twelve miles of loose sand. Finally the
Onoshohatchee River and the first habita-
tion in about forty miles came into view.
We soon reached Okeechobee City— then a
settlement of several scores of houses. In
the fall of 1913 when we went up the
Onoshohatchee River from Lake Okeechobee
this place had been indicated on the map
and staked out by the surveyors, but had
not yet been colonized.

At this point we again left civilization be-
hind. From Okeechobee City to Fisheating
Creek the country was devoid even of roads,
and we took to an old trail dating back
perhaps to a period before the Seminole
wars. By degrees Okeechobee City disap-
peared as we hurried around the curves, not
to say coils, in the trail, and after passing
some miles of pinelands we suddenly came
into the bottoms or prairies of the Kis-
simmee River. These bottom lands are like
immense lawns, perfectly level, carpeted
with a turf of various grasses, and often
extending as far as the eye can see. There
were thousands of semiwild cattle grazing
on the broad green prairies.

All had gone well thus far, but at the
Kissimmee River a series of apparently pre-
destinated troubles began. The trails on
either side of the river were connected by
a ferry which consisted of a fiatboat large
enough to hold a car, and a small motor
boat of barely sufficient capacity to drag
the fiatboat around the bends and over the
sand bars in the river. In order to cross the
river, which there is less than a hundred
feet wide, it was necessary to go about a
half mile down stream because of the ero-
sion of the banks. Once in the stream the
current of the river — say, three miles an
hour — carried the ferryboat along at a
greater speed than the motor boat could
maintain. Time and again the ferryboat
would bump into the river banks, first on
one side, then on the other, and would, in
turn, bump against the stern of the little
motor boat and knock off the rudder. Even
after the ferryboat drifted out of sight, we
who were left behind for the second trip
could hear the ferryman nailing the rudder
on his disabled boat.

We lost several hours of valuable day-
light while waiting for the ferryman to re-
place dead batteries with live ones. As the
slimt twilight deepened we ran up a slight
incline through a strip of pine woods, mak-
ing all haste compatible witii the innumer-
able curves in the trail and the proximity of
pine stumps, and found ourselves on the
great Indian Prairie. This comprises a
large part of an immense region lying west
of Lake Okeechobee, north of the Caloosa-
hatchee River, and east of Peace River. The
prairie is high and dry all the year round
and is uninterrupted, except by a single
stream, Fisheating Creek, one of the larger
feeders of Okeechobee. Up to a short time
ago it was practically uninhabited, except
by wandering Indians. At the present time
a half dozen or more " — ports," " — dales,"
" — monts," " — burgs," and even "—Cities"
have been put on the map, and a railroad
bisects the region, — so, farewell to its natu-
ral features.

In order to save time, we decided to cross
the prairie that night, and we certainly had
a weird ride. The trail at times was dis-
tinct, but at other times almost blind. Al-
though the prairie was a dead level, the
optical illusion created in the darkness was
that of running down hill and jumping off
the earth. We had some obstructions to
progress in the form of forks in the trail
which would, we knew, either come together
farther on or diverge indefinitely and thus
lead to some other part of the state. At
each fork, the four in our party would hold
a council, and in each instance consult the
stars. The stars always put us on the right
trail, and toward midnight, after passing
several half-discerned Indian camps, we saw
a few faint lights of human habitation ap-
pear, and finally we reached the recently
established colony of Palmdale on Fisheat-
ing Creek, or, in Seminole, "Thlathtopopka-
hatchee."

We did not hesitate to disturb the peace-
ful slumbers of the inhabitants, who were as
glad to see us as we were to see them, which
fact they showed in a substantial manner by
arising from their slumbers and preparing
a midnight meal. After a few hours' rest
we made an early start for Labelle, which
is an old settlement situated at about the
head of natural navigation on the Caloosa-
hatchee River.

The Indian Prairie extends nearly or quite

,4 BOTANICAL EXCURSION TO THE BIG CYPRESS

491

to the Caloosahatehee. Unfortunately, a
road had been laid out to connect Palmdale
with Labelle. It is well we decided to stop
at Palmdale until daylight, for although
we could travel the almost trackless prairie
in the dark with ease, we could barely tra-
verse this new road in broad daylight. The
deep sand had become very loose, and it
took more than the engine to get the cars
over a good many miles of the road.

Just south of Palmdale we crossed Fish-
eating Creek, which is an exceedingly pic-
turesque stream meandering through the al-
most uninhabited prairie, between banks
either exposed to the sun, or clothed with
shrubs and bright-colored asters or ham-
mocks of oaks, ash, and maple, which in
some places give way to groves of palmettos
that often lean far over the water's edge.
After leaving the hammocks which border
the creek we drove out on the prairie again,
and few trees came into view for a distance
of about eighteen miles, until the hammocks
bordering the Caloosahatehee appeared.

Perhaps the most interesting creature on
these prairies was the burrowing owl. This
bird had honeycombed the prairie in many
places with its burrows. These tunnels,
often six to eight feet long, are about a foot
beneath the surface of the sand. At one
end is an opening approximately six inches
in diameter, while at the other end is a nest.
The old owls were so tame that one could
almost pick them up, and often they would
sit perfectly quiet while the automobiles
passed them at a distance of not more than
two feet.1

On this same prairie many interesting

1 Out of curiosity we decided to dig into one
of the burrows. Starting at the opening, we began
by lifting the sand out very carefully. Of a sud-
den we were startled by the rattle of a rattle-
snake. After proceeding a few inches farther
we heard two rattlesnakes ; before going much
farther into the burrow a third rattlesnake be-
gan to rattle. The digging became more exciting
as we worked farther in and as the snakes rattled
more loudly. When we neared the end of the
burrow we cautioned one another to be careful
not to get our hands too close to the snakes.

This seemed to be an excellent opportunity to
get good photographs of living rattlesnakes. Con-
sequently the camera was set up and everything
prepared for the opening of the end of the bur-
row. As there was no woody growth on the
prairie the question of getting sticks with which
to fight the snakes arose. After considerable
search several surveyor's stakes were found, and
with these we prepared nooses for capturing the
serpents alive. With extreme caution we ap-
proached the end of the burrow; the Bnakes
began to rattle more viciously. Finally the
sand was removed from the top of the end of

plants were observed and collected. Milk-
weeds were represented by species of As-
clepias and Asclepiodora, while more con-
spicuous was the purple water willow (Di-
li nthera crassifolia). Low milkworts (Poly-
galas) with white and yellow flowers were
prominent in the landscape, and clumps of
the native beardtongue (Pentstemon multi-
florus) towered above all the other her-
baceous plants. There a white-flowered
heliotrope replaced the common yellow-
flowered heliotrope of the region lying east
of the Okeechobee basin and the Everglades.

After contending with the sand for
several hours we reached the Caloosahatehee
River and came to the town of Labelle,
where we did not delay, but went directly
up the Caloosahatehee several miles to Fort
Thompson. There we found a number of
magnificent live oaks around the old bar-
racks which date from the period of the
Seminole wars. After making a number of
photographs in that region we returned to
Labelle and at once started down the south
bank of the Caloosahatehee River for Fort
Myers.

We now left the prairies behind and en-
tered the flatwoods, where the arboreous
vegetation is made up almost entirely of
pine trees. Peninsular Florida, especially the
southern part, lacks what is ordinarily
understood as altitudes, in fact, most of it is
decidedly flat. It might well be called a
large sand bar. Notwithstanding this dis-
advantage, it reveals an astonishing number
of surprises in the matter of diversity.
The Big Cypress is one of the larger sur-
prises. Its area is about half that of the

the burrow, and to our surprise we found four
young owls, three large and one small, but — no
snakes !

It was the three larger owls that were making
the noise of a rattlesnake, and imitating it so
well that all of us who had had personal ex-
perience with rattlesnakes were deceived. We
decided that this experience proved that the stories
we used to hear of owls, prairie dogs, and rattle-
snakes living peacefully together in the same bur-
row were fantastic. Of course, a rattlesnake
might enter an owl's burrow, 'either to seek
shelter or food ; but it is a difficult matter for
any one well acquainted with the habits of
rattlesnakes to believe that a husky rattler would
be considerate and restrain his appetite, with
such a tempting morsel as a young owl or young
prairie dog lying about in his den. (For further
notes on this subject see: The American Natural-
ist, Vol. XLI, pp. 725-726; Vol. XLIII, pp. 754-
55; Birds of the World, pp. 536-37.) After
photographing owls instead of rattlesnakes, we re-
placed them in their nest and rebuilt their burrow,
as well as we could, by making a roof of brush
over which we replaced the sand.

492

NATURAL HISTORY

Palms and pine trees are often a favorite ref-
uge for wild turkey and deer. A flock of turkeys
took refuge in this particular grove just as we
suddenly rounded a sharp curve in the trail.
In the Big Cypress there may be prairies so ex-
tensive that woody vegetation can be seen merely
as a dark line along the distant horizon, or again
we may see at one time associations of palms and
pines, pure pine woods, solid broad-leaved ham-
mocks, cypress heads, and combinations of cypress
head and hammock

Everglades, and although it abuts directly
on the western side of them, it has but
little in common with them. Instead of
I icing a vast prairie-marsh like the Ever-
glades, the Big Cypress exhibits a variety
of conditions and plant associations. There
are pinelands, prairie, sloughs, cypress
heads, hardwood hammocks, palmetto ham-
mocks, and lakes.

Early in the afternoon we were prepared
to strike into the wilderness. After leaving
Fort Myers, roads disappeared and we took
to mere trails through the pine woods in a
southeasterly direction. As we proceeded,
strange plants and strange birds began to
appear. White terrestrial orchids (Gymna-
dcniopsis nivea) and single-flowered spider
lilies (Hymevocallis Jiumilis) dotted the dry
prairies, while uliginous creepers with vari-
ous colored flowers formed encircling
mats about all the shallow ponds. Ponds
and pools were the favorite feeding places
for the wood ibis, the white ibis, cranes, and
herons. The hammocks hid many flocks of
wild turkeys in their depths.

For some distance outside of Fort
Myers we traveled through unbroken pine
woods. As we went on, the pine trees be-
came more scattered and areas of prairie
came into view. Farther on, the prairie
began to increase and the pines appeared
only here and there as isolated colonies. A
little farther on cypress trees appeared, and
we were really in the Big Cypress. Here,
too, the cabbage palm was much in evidence,
and in some places it formed hammocks of
almost pure growth. As we proceeded, the
prairies grew larger and the cypress grew
less, until there was open prairie in all di-
rections almost as far as the eye could see.
Then the hammocks clothing the Okaloa-
coochee Slough appeared in the distance as
a mere line on the horizon. It is said that
the Seminole word "Okaloacoochee" signifies
"boggy-slough." Consequently the usually
associated word "slough" is really super-
fluous.

As we approached the slough we observed
immense flocks of ibis collecting at their
rookery for the night. The confused sounds
they made as they flew over the tops of the
tall trees could be heard for a distance of a
mile. The sight of the great flocks of
ibis and the racket of their croaks or
squawks as they collected in their rookery
we shall long remember.

A BOTANICAL EXCURSION TO THE BIG CYPRESS

49!

We drove into a small hammock within
half a mile of the slough and prepared to
camp there for the night. Many interest-
ing plants were collected on the prairies
near the slough before darkness drove us
back to camp. Indian plantains (Mesa-
denia), foxgloves (Agalinis), and helio-
tropes (Heliotropium) grew nearly every-
where. Fully as interesting as the native
plants was the climbing black-eyed Susan
(Thunbergia alata), which we found ex-
tensively naturalized on the prairie near the
Okaloacoochee. The plants now growing
there may be the descendants of specimens
introduced and cultivated in gardens the
Seminoles maintained there fully a century
ago.

The following morning we broke camp
about daybreak and proceeded to cross the
slough. We parked our cars in its midst on
the very spot where, it is said, more than
sixty years ago Lieutenant Harsuff's com-
pany of engineers had their sanguinary clash
with Chief Billy Bowlegs — after they had
destroyed the old chief's garden just to
"see old Billy cut up."

The larger trees of this hammock con-
sist of the bald cypress or river cypress
(Taxodium distichum) . It was a favorite
spot for the Indians to obtain logs for mak-
ing their dugout canoes. In the rainy sea-
son there is commonly about six feet of
water in the slough. After the rainy season
the water table is naturally lowered by
seepage. The waters, evidently, find their
way directly into the Everglade basin, and
directly or indirectly into the Gulf of
Mexico. In the dry season most of the
slough can be traversed on foot. It was the
custom of the Indians to go to the slough in
the dry season, cut down the trees they
selected for making the canoes, and then
wait for the wet season and high water
to float the logs out toward the western
coast.

We went down the slough afoot just as
the thousands of birds in the rookery were
awakening. The birds mostly represented
several species of ibis, and were present by
the hundreds and thousands on the large
cypress trees. In fact, they were so crowded
on some of the giant cypresses that they
were continually falling off for want of
sufficient room to etand. As a consequence
of not having been much disturbed by man,
they were so tame that one could walk

In the Okaloacoochee Slough dead trees as well
as living serve as part of the ibis rookery, for
the birds are so numerous that any available
space is used. Their nests are rude cradles of
sticks in the trees or on ledges of rock. During
the day the birds leave the rookery, traveling in
more or less definite groups or companies. This
photograph was taken in the morning, after the
greater number of the birds had departed

j*2**WMy

V -'." •7k!*/--,

Il

^ri^1

V

&*1 i

■^

C'.'Vrf'

; V

A NATURAL AMBUSCADE

From such beautiful coverts — perhaps from this very spot— commands of the United States Army fought the
Indians during the Seminole wars. A riotous growth of shrubs whose stems are intertwined with woody vines
form an almost impenetrable thicket extending back to a wood of river cypress in the lower part of the slough.
The hammock floor is a mass of ferns and small herbs ; Boston and sword ferns in particular are prevalent. There
are at least fifty other kinds of ferns— many of them epiphytic — which display the greatest possible variety in
structure and contour
494

A BOTANICAL EXCURSION TO THE BIG CYPRESS

495

toward them, set up a camera, and photo-
graph them at short range.

There was water in the lower parts of the
slough, but none was visible, for the sur-
face was completely covered with a soft
carpet of various small aquatics. These
were distributed in patches of beautiful
shades of green. In the higher parts of the
slough ferns and flowering plants grew in
about equal profusion and remarkable
luxuriance. The growth reminded me of
that in the hammocks of the eastern shore
of Lake Okeechobee.1 The large, straplike
leaves of the spider lily and the paddle-
like leaves of the golden club or bog torches
(Orontium) were very conspicuous. The
leaves of the golden club here at its most
southern known station were fully three feet
long, while the fruiting spadices lying
around on the ground were thrice the size
of any that I have ever observed at the
north. The lizard's-tail (Sanrurus) was also
there in great abundance.

Thus these typically northern plants, the
lizard's-tail and golden club, are there inti-
mately associated with such typically south-
ern plants as the water hyacinth and the
water lettuce. Other southern elements rep-
resented are the Boston fern (Nephrolepis
exaltata) and the wild coffee (Psychotria
undata \.

After making a collection of all the
plants observed and photographing the more
interesting views, we returned to our cars,
crossed the slough, and set out over the
prairie in the direction of Eocky Lake,
which lies in an uncharted spot in the
Big Cypress between the Okaloacoodiee
Slough and the Everglades. As we pro-
ceeded, palmetto hammocks, hardwood ham-
mocks, and cypress heads became more nu-
merous on the prairie. At last we came to
the hammock surrounding Rocky Lake,
which is known to the Seminoles as Okee-
hy-yot-lochee, a word said to mean "wide-
open-water," where we camped for lunch,
and made collections of the plants. This
lake is contained in a rock basin several
acres in extent. It is said that it is fully
seventy-five feet deep, and abounds in fish
and alligators. Of course, it would be some-
what of an exaggeration to say that one
could walk across the lake on the alligators'

1 See Journal of the New York Botanical Gar-
den, Vol. XV. pp. 69-79; Vol. XIX, pp. 279-
290. The American Museum Journal, Vol.
XVIII, pp. 684-700.

backs; but they were more numerous than
1 have ever seen them elsewhere.

After lunch we set out for the ruins of an
Indian mission- which some years before
had been established near the site of the
one-time Fort Shackleford, and then aban-
doned. After leaving Rocky Lake the trail
wound in anil out between hammocks and
cypress heads until finally more open prairie
was reached.

When we arrived at the Seminole mis-
sion we were now not more than four miles
from the western edge of the Everglades.
A unique specimen of the cabbage tree was
observed — a five-fingered object, with^fre
branches of about equal length arising from
the trunk, just above the surface of The
ground and all in one plane. Probably no-
where is this duplicated. Many interesting
plants were found in the vicinity, especially
several loosestrifes (Lythrum), and a false
indigo (A mot pha ) which is apparently
different from any known species.

2 The old Seminole mission thirty-five miles
beyond Imniokalee was established about 1910—11
through the instrumentality of William Crane
Gray, then Bishop of southern Florida, for the
Protestant Kin-.,, pal Church, the work being
undertaken by I>r. William J. Godden, of Green-
wich, England, who happened then to be touring
the United states. Dr. Godden, a man of high
connections and attainments, soon won the love of
both red and white men. Originally, he started a
small hospital and social center for the Seminoles
at a point about seventy miles back from Fort
Myers, near the historic site of old Fort Shackle-
ford. He called this lirst settlement Glade Cross —
because of it-- proximity to the Everglades and the
argi white cross he mounted against a cabbage
palm. But when a couple of red patients died
in the hospital l'O more Seminoles could be in-
duced to come near the place. The mission was
thereupon transferred to the lonely outpost called
Boat Landing on the edge of the Everglades, at
that time the head of all the canoe trails of the
region. It was not long, though, before the par-
tial drainage of the Everglades dried the canoe
trails, and Boat Landing ceased to be a port
of call, or any port at all. So the doctor once
more moved his mission, this time to about the
center of the present Seminole Reservation, five or
six miles from his former locations, right in the
heart of the Big Cypress, where he hoped to
establish an experimental farm. He put up a
number of buildings — a store, a dispensary, var-
ious shelters. He employed the Seminoles to dig
a couple of miles of drainage ditches about the
place. He himself worked far harder than any-
one else — without pay, mostly alone, always de-
voted, perfectly kind — while his people in Eng-
land urged him to return to them. He died at
the mission, suddenly, presumably of heart fail-
ure, in 1914. And now Glade Cross is jungle
again; only a few broken canoes mark the site of
Boat Landing; and the last site of all, still called
"Godden's Mission." is merely a weedy, haunted
ruin. The doctor's body was buried at Jm-
mokalee, a Seminole word which signifies "My
Home." — Perley Poore Sheehan.

<

CO

S3*

*^ >~~~ r—

LU

X

o-^> >><&

h

Oj fe a 3

S c o *

3 2°

Z

2 - '5

<w ;>*

£

the
on]

ice,

o

O 9 S <8

cc

»i§

DC

<c jS § o

D

o p,

a ->iT3 p.

CO

a " a s

O 5

Q I.

Z

.a . h
39o

= ^5;

SB »

a h

Sfe-H

^ c-2

i So

it

g-O CD

■2 © £
cp o

ft ti 03

a oj o
.9 i **

O a>

CO

r-> O

tr

^ ® a

UJ

<:

O

?"5io

u.

^■Efe

£•&

LL

a.^_ o

O

"S ° w

w « —

£

o

Q

■2 •< o

03

<

0) > M

LU

£'JC

C3

o S o a

— J w CJ CD

<D «w

■S .- o

as

a>

5 3,0
o bi 2

ft°

a g'

•- ft 03
^«

C EC m-

a — a
>».,. o

1 1 1

oj.2^
IP

J

2 5 a

cs a S
° fe.2

a 05

^ X ft
-2^2 o

h S-. <D

a >>»
p » >

c3 o ft
5* U CS

^"^•S

' :.*

(JL,-- '"•*

*^3f — .«^*^

I A

JHHHyr

E-< ftS

500

NATURAL HISTORY

Wild orange trees, some with sour fruits,
others with sweet, occur in the hammocks
of the Big Cypress. Of course, some of
these are the remnants of trees planted by
the Seminoles; but others may be derived
from ancestors planted there by the abo-
rigines of that region or by the Spanish
adventurers themselves.

The cypress of the region outside of the
large sloughs was the pond cypress (Taxo-
<l in in ascendens). The prairies were showy
flower gardens. Several species of Polyg-
ala, several of Sabbatia, three or four
kinds of terrestrial orchids, and a number of
other conspicuous plants, both monocotyle-
dons and dicotyledons, often covered acres
in extent. A yellow-flowered bladderwort
grew copiously in extensive patches in the
dry white sand! Many rare and little-
known plants were collected for future
study.

Eocky Lake proved to be the lunch sta-
tion. "While in a temporary camp near the
shore the writer rescued two animals from
living graves. On two different occasions,
while going to the lake for a drink of
water, he was startled by agonizing cries.
In the first instance, a large water moccasin
had caught a mocking bird and was at-
tempting to swallow it. In the second in-
stance, another moccasin had caught a
frog which he was trying to slip down his
throat. In each case the victim went free
and, it is to be hoped, survived.

After recrossing the Okaloacoochee
Slough, instead of retracing our former
course we turned more to the westward and
headed for the colony of Immokalee. After
passing through stretches of forest and
prairie we came in view of the scattered
houses of the settlement. This colony, situ-
ated about thirty miles in a direct line from
Fort Myers, comprises a general store and
a few dwelling houses. We reached Fort
Myers shortly after sunset, and early the
following morning started up the Caloosa-
hatchee Eiver by the same course we had
taken several days before. Numerous stops

were made along the way for collecting
plants and taking photographs. Palmdalc,
where we took the trail over the great
Indian Prairie, was reached early in the
afternoon. The herbaceous vegetation and
magnificent palmetto hammocks not vis-
ible in the dark gave an entirely different
impression of the prairie region. Some of
the same genera of plants were common to
both the Indian Prairie and the Big Cypress
but the species were usually different. The
Caloosahatchee Eiver is evidently a natural
boundary between different floral regions.
The most striking feature in the vegetation
of this prairie, however, is the cabbage tree.
This palm grows in small clumps and also
forms hammocks from one to many acres in
extent, surpassing in luxuriance any growth
of it I had seen previously.

After the usual bumping of banks and
sand bars the ferry landed us on the oppo-
site shore of the Kissimmee Eiver whence
we at once set out over a trail which seemed
to have endless windings, but which finally
brought us to Okeechobee City. From there,
after a night's rest, we journeyed to Fort
Pierce, collecting as we found favorable
places in the pine woods and in the swamps,
and next day we started on the final stage
of our return trip to Miami. The city was
reached without further incident, except
the passing survey of a large hammock on
a high sand dune along Saint Lucie Sound
or Lower Indian Eiver, which has already
been partly described1 and which has been
designated for thorough exploration.

This preliminary survey deeply impressed
upon us the wonderful natural history of
that little-known region. Our time was
limited and the region was large, but some
day, before drainage and other depreda-
tions of civilization, not to mention vandal
ism, have removed the bloom from that still
unspoiled garden, we hope to make another
and longer visit to the land of the Big
Cypress.

1 Journal of the New York Botanical Garden,
Vol. XIX, pp. 76-77.

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL
GARDEN— No. 230

"VERONICA" IN NORTH AND SOUTH

AMERICA

By FRANCIS \Y. PEXXELL

NEW YORK

Reprinted, without change of paging, from Rhodora 23.

LJBHARY

NEW YORK

SOTAMCAi.

OAKLibii

"VERONICA" IN NORTH AND SOUTH AMERICA.
Francis W. Pennell

This study is the outgrowth of several attempts to revise our
knowledge of the species of " Veronica" growing in different portions
of the Western Hemisphere. Whether in our own "Local Flora,"
in the Rocky Mountain or the Southeastern States, or in Colombia
and Ecuador, certain wide-ranging species were encountered, and the
effort to verify the nomenclature in many instances took the reviewer
into problems of the identity of Old World allies. On these accounts
it has seemed best to consider in one study the plants of this group
in both North and South America, and also to include known natural-
ized species.

Of all the tribes of the Scrophulariaceae mentioned by von Wett-
stein in his great revision of the family in " Die Natiirlichen Pflanzen-
familien," that of the Digitaleae, to which Veronica is assigned, prob-
ably has least coherence. The plants at least should agree in having
the antero-lateral lobes of the corolla external in aestivation and in
not being parasites. I know of no offenders against the latter crite-
rion. But, because they possess not only posterior corolla-lobes
external but also form characteristic glands on the fruit or in the
leaves, I have recently transferred the two lowland Tropical genera
Capraria and Scoparia to the essentially Tropical tribe Gratioleae.
This restricts the Digitaleae to a more likely distribution through
the Temperate or Arctic zones and the cooler zones of Tropical
mountains. I can not further analyze the tribe here, except to say
that Digitalis itself, through possessing styles distinct at apex and a

4 Rhodora [January

leaves occur in the perennial species, while the extreme stages of
capsule-lobing and flattening, of few and large seeds, and of alter-
nate leaves are in the relatively few annual sorts. Also, as is the
general habit in allied genera, the original Veronicellas were surely
erect herbs.

In Euveronica the stem continues indefinitely as a vegetative axis,
its leaves opposite throughout, while the inflorescence is localized
in specialized axillary racemes. This is the inflorescence of Hebe,
but the diversity of capsule-structure tells us that such localization
must have originated independently in these two groups. Else-
where in this tribe I know it only in the Chinese Botryoplcuron Hems-
ley, which seems to be separated by little else from Calorhabdos
Bentham. The flowers in the racemes of Euveronica are alternate
as they are in all inflorescences of Veronicella. This accords with
the view that these racemes are reduced branches, and not formed by
the forking of originally simple pedicels. No stages suggesting the
latter alternative are known to exist, and the fact that remote axil-
lary flowers occur in Veronicella only in some profoundly modified
annual species makes such a derivation very improbable. The
species of Euveronica are all perennial, and such an erect-growing
species as V. lati folia, which has a scarcely notched capsule, shows a
close approximation in habit to the most primitive group of Veroni-
cella, including species such as V. maritima and V. mcxicana.

A few words need be said concerning age and distribution of vari-
ous species. Contrary to expectation and certain widely advanced
theories, it is not those species whose structures proclaim them as
most ancient which are necessarily most widely dispersed. Struc-
turally none of our species can make better claim to antiquity than
Veronica mexicana, yet this species occurs only in a limited area in
the mountains of northern Mexico. Obviously it has no close kindred
in our flora, and so must, I believe, be considered a relict. Again,
V. Copelandii, our only other species definitely retaining the pos-
terior sepal, is known from but a few mountains in California. On
the other hand, the group of Veronica alpina and V. Wormskjoldii
has a wide range through Temperate North America and Eurasia,
although it has become divided geographically into several well-
marked species. But it is the obviously non-primitive species which
have prospered most and have become or are becoming very wide-
spread. These possess peculiar skill in taking advantage of natural,

1921] Pennell, — "Veronica" in North and South America 5

or recently of man-made, methods of dispersal. Of species natu-
rally distributed, Veronica serpyllifolia humifusa and V. Anagallis-
aquatica have become nearly cosmopolitan within their respective
climates and environments, while carried by human agency and to-
day aggressive weeds in our land are Veronica serpyllifolia, V. pere-
grina and its variety xalapensis, V. arvensis, V. persica, and V. offici-
nalis. To this list, otherwise of Palaearctic origin, America has
contributed Veronica peregrina, thus showing that the New World
may develop sufficiently aggressive plants.

I must thank the curators of the United States National Museum,
New York Botanical Garden and Philadelphia Academy of Natural
Sciences who have placed their rich collections at my disposal. Also
I have seen specimens of certain species from the Gray Herbarium.
Probably nearly 4000 sheets have been examined in the present study.
It has been with hesitation but I trust to the clearing of the problems
involved, that I have ventured so freely into Eurasian botany.

Key to Genera and Subgenera.

Capsule dehiscing loculicidally, its walls and septum
thin. Herbs, the stem dying with the leaves.
Corolla white or pinkish, its lobes much shorter
than the tube. Capsule acute, longer than
wide, not flattened, dehiscing by short apical
slits. Seeds slightly reticulate. Plant 10-20
din. tall, with leaves in whorls of four or five.
Main stem terminating in an inflorescence. . . I. Yeronicastrum
Corolla blue or white, its lobes nearly as long as or
usually longer than the tube. Capsule aeu-
tish to deeply notched, as broad as or broader
than long, more or less flattened contrary to
the septum, dehiscing by longer slits which
extend at times even to the base. Seeds not
reticulate. Plants lower, with leaves, at least
the lowermost, opposite, or very rarely in

whorls of three or four II. Veronica

Main stem terminating in an inflorescence, its
flowers remote and axillary or densely
crowded, in all cases the upper bract-leaves

alternate Subgenus 1. Veronicella

Main stem never terminating in an inflorescence,
the leases opposite throughout and the

flowers all in axillary racemes Subgenus 2. Euveronica

Capsule dehiscing septicidally, the thick septum
splitting and each carpel opening distally by a
median slit through the septal wall. Leaves
opposite throughout, and flowers all in axillary
racemes. Shrubs or small trees, the coriaceous
leaves in falling leaving conspicuous scars III. Hebe

6 Rhodora [Januab?

I. VERONICASTRUM [Heister] Fabr.

Veronicastrum [Heister] Fabr., Enum. Meth. PI. Hort. Helmstad.
111. 1759. (Checked only in second edition, 205. 1763.) Type
species, Veronica virginica L.
1. Veronicastrum virginicum (L.) Farwell.

Veronica virginica L., Sp. PI. 9. 1753. "Habitat in Virginia."
Plant grown in the Clifford Garden in Holland, and carefully de-
scribed by Linne, Hort. Cliff. 7. 1737. No pubescence on leaf
mentioned.

Veronicastrum album Moench, Meth. PI. Hort. Marburg. 437.

1794. " Veronica virginica Linn." Grown in the

Marburg Garden, Germany.

Calistachya alba Raf. in Med. Repos. New York. II. 5: 352. 1808.
Based on Veronica virginica L. Type species of Calistachya Raf.,
not Callistachys Vent., 1804.

Veronica quinqucfolia Stokes, Bot. Mat. Med. 1: 28. 1812. "In

a garden V. virginica L." Evidently name chosen

as more appropriate than "virginica." Two varieties, or actually
forms, alba and incarnata, are listed.

Leptandra virginica (L.) Nutt., Gen. N. Am. PI. 1: 7. 1818.
Type species of Leptandra Nutt.

Eustachya alba Raf., Cat. Ky. 14. 1824. Based upon Veronica
virginica L. Eustachya Raf. in Am. Mo. Mag. 4: 190. 1819, was a
new name for Calistachya Raf. Name antedated by Eustachys Desv.,
1810.

Leptandra virginica purpurea Eaton, Man. Bot. ed. V. 275. 1829.

" " Described as with "flowers purple." Credited to

Pursh, who however assigned his color variety no name. According
to Pursh, Fl. Am. Sept. 10. 1814: "On the mountains of Virginia
I observed a . . . . variety with purple flowers."

Leptandra purpurea Raf., Med. Fl. 2: 20. 1830. "Confined to
the savannas of the South and the West [of the United States]."
Carefully described; a purplish-flowered, single-spiked plant with
sessile leaves in whorls of three. Three varieties, or actually forms,
named.

Leptandra alba Raf., 1. c. 21. 1830. "The most common species,
being found all over the United States." Described with white
flowers and semi-petiolate leaves in whorls of usually five. Several
varieties, or actually forms, named.

Leptandra villosa Raf., 1. c. 21. 1830. "Mr. Schriveinitz [Schwei-
nitz] has found it in North Carolina." Careful description of the
form with leaf-blades pubescent beneath. This may be considered
a forma villosa (Raf.) Pennell, comb. nov.

Eustachya oppositifolia Raf., New Fl. Am. 2: 21. 1837. "Mts.
Apalaches of Virginia." Apparently an opposite-leaved virginica,
a form which I have never seen.

1921] Pennell, — "Veronica" in North and South America

Calistachya mrginica laneeolata Farwell in Mich. Acad. Sci. Rep.
17: 176. 1915. "Farwell No. 1165, July 18, 1891, from Ypsilanti
[Michigan]." A narrow-leaved form.

Veronicastrum virginicum (L.) Farwell, Drugg. Circ. 61: 231. 1917.

Meadows and open woods, from Vermont, southern Ontario and
Minnesota to Georgia and eastern Texas. Southward forma villosa
is more prevalent.

Variable, but certainly one species. Varying in number of leaves
in whorl (five, four, rarely three, or in oppositifolia even two), in leaf-
form from lanceolate to nearly ovate, pubescent or nearly or quite
glabrous below (forma villosa with whole under surface velvety), in
inflorescence being of one or several racemes, and in color of corolla
and filaments, varying from white through pinkish to lighter shades
of violet-purple.

II. VERONICA L., Subgenus 1. VERONICELLA (Fabr.)

Veronicella [Heister] Fabr., Enum. Meth. PI. Hort. Helmstad. ed.
II. 205. 1765. Type species, Veronica hederaefolia L., of Europe.
A. Perennials, from subterranean stems (rootstocks) .
Only the upper leaf-axils flower-bearing, so that
inflorescence is formed of definite racemes.
B. Capsules only slightly flattened, even the lowermost
on pedicels shorter than the capsule-length:
inflorescence appearing a spike. Style two or
three times the length of the capsule. Leaf-
blades lanceolate. Plants 4-10 dm. tall.
Petals broadly oblong. Leaves opposite or in
threes, rarely in fours, 3-15 cm. long, the
blades acute to acuminate, dentate-serrate
to sharply and somewhat doubly serrate.

Plant 4-10 dm. tall 1. V. maritima.

Petals oblong-lanceolate. Leaves always oppo-
site, 3-6 cm. long, the blades acutish, crenate or

crenately serrate. Plant 2-4 dm. tall 2. V. spicata.

R'. Capsules strongly flattened, the lowermost on pedi-
cels nearly or quite as long as the capsule-length;
inflorescence obviously a raceme. Style slightly,
if at all, longer than the capsule.
C. Capsule as long as or longer than wide, less deeply
or not notched. Corolla violet-blue, rarely
nearly white, glabrous within. Leaf-blades
obtuse, acutish to acuminate. Stems erect,
nearly or quite from the base.
Calyx 5-parted, the posterior lobe over half the
length of the others. Leaf-blades lanceo-
late, acuminate, irregularly dentate.

Plants 5-6 dm. tall 3. V. mexicana.

Calyx 4-parted, or with a rudiment of the fifth
lobe. Leaf-blades ovate to oblanceolate,
obtuse to acutish, entire to serrate. Plants
less than 3 dm. tall.

Rhodora [Janttart

Style longer than the capsule. Filaments
equaling or exceeding the corolla.
Calyx-lobes unequal, the anterior longer.
Leaf-blades entire.
Leaf-blades oblong-elliptic, hirsute-pubes-
cent, acute. Sepals five, the pos-
terior much the smallest, the others
slightly unequal. Capsule scarcely

notched. Corolla 5 mm. long 4. V. Copelandii

Leaf-blades elliptic-oval, glabrous or gla-
brate, obtuse to acutish. Sepals four,
decidedly unequal. Capsule strongly

notched. Corolla 5-7 mm. long 5. V. Cusickii.

Style shorter than or nearly as long as the
capsule. Filaments shorter than the cor-
olla. Calyx-lobes of uniform length.
Leaf-blades, at least the lower, crenate
to serrate.
Capsule rounded or acutish, not notched.
Style nearly as long as the capsule.
Sepals canescent, not ciliate, nearly
equaling the capsule. Corolla twice
as long as the sepals. Stem-leaves
elliptic-oblanceolate. Plant 1 dm.
tall or less, the inflorescence minutely

pubescent 6. V.fruticans.

Capsule slightly notched. Style less than
half the length of the capsule. Sepals
conspicuously ciliate, one-half to two-
thirds the length of the capsule. Cor-
olla less than twice as long as the se-
pals. Stem-leaves elliptic to ovate.
Plants mostly 1-3 dm. tall, the inflor-
escence hirsute-pubescent.
Capsule glabrous. Sepals glabrous on
back, ciliate on margins with non-
glandular hairs. Plant usually 1-2

dm. tall 7. V. alpina.

Capsule pubescent. Sepals usually pi-
lose on back as well as margins.
Leaf-blades serrate, the largest nearly
cordate at base. Capsule and
sepals with hairs which have at-
tenuate non-glandular tips, the
sepals less pilose or glabrous on
back. Plant usually 1-1.5 dm.
tall, with pedicels becoming 8-11

mm. long 8. V. Stelleri.

Leaf-blades crenate-serrate to nearly
entire, rounded at base. Capsule
and sepals with hairs which have
rounded glandular tips, the sepals
densely pilose on back. Plant
usually 1.5-3 dm. tall, with pedi-
cels 2-5(-10) mm. long.
Corolla mostly 6-7 mm. long. Ped-
icels mostly 5-10 mm. long.
Leaf-blades mostly ovate, fre-
quently serrate 9a. V. W or mskjoldii nutans.

19211 Pennell, — "Veronica" in North and South America 9

Corolla mostly 4-6 mm. long. Ped-
icels mostly 2-5 mm. long.
Leaf -blades mostly oblong-ov-
ate, rarely evidently serrate. .9. V. Wormskjoldii.
C. Capsule obviously wider than long, notched one-
fourth length. Corolla white or bluish, with
deep-blue lines on the posterior side, the tube
pubescent within. Style nearly as long as the
capsule. Leaf-blades ovate-oblong or oval,
obtuse, obscurely crenate. Stems extensively
repent, ascending at apex.
Stem throughout and pedicels minutely pubes-
cent with upcurved hairs. Corolla about
2 mm. long, white or whitish, with blue
lines on posterior side. Capsule mostly

3-4 mm. wide 10. V. serpyllifoha

Stem distally and pedicels finely pubescent with
longer mostly spreading hairs. Larger
throughout, the corolla mostly about 3 mm.
long, pale-bluish with deeper blue lines on
posterior side. Capsule mostly 4-5 mm.

wide 10a. V. serpyllifolia humifusa.

A\ Annuals, fibrous-rooted, but without subterranean
stems. Most leaf-axils flower-bearing, so that
inflorescence appears to be of "axillary" flowers.
B. Pedicels shorter than the lanceolate to linear sepals.
Capsule strongly flattened. Seeds many, less
than 1 mm. long, flat, smooth or nearly so.
Plants erect.
Leaf-blades, excepting the lowermost, sessile,
those of the lower stem-leaves oblanceolate,
nearly entire to dentate. Corolla whitish
throughout. Capsule greenish, notched, the
minute style hidden between the capsule-
lobes. Plant glabrous or with minute usu-
ally gland-tipped hairs.

Plant glabrous 11. V. peregrina.

Plant pubescent with short gland-tipped hairs,
which are usually present even on the cap-
sule 11a. V. peregrina xalapensis.

Leaf-blades petioled, or the upper nearly sessile,
those of the lower stem-leaves ovate, crenate-
serrate. Corolla deep violet-blue. Capsule
yellowish-brown, pubescent with slightly
gland-tipped hairs, strongly flattened, notched
nearly or about one-third length, the longer
style reaching about to the capsule-lobes.
Plant pubescent with white glandless or ob-
scurely gland-tipped hairs 12. V. arvensis.

B'. Pedicels longer than the ovate sepals. Capsule rela-
tively turgid. Seeds few, 1.3-3 mm. long, con-
vex-arched, roughened. Plants repent.
Leaf-blades ovate, serrate to dentate. Sepals
shortly ciliate. Capsule slightly flattened,
deeply notched, pubescent. Seeds 1.3-1.5
mm. long, brown.
Capsule-lobes united at least two-thirds their
length. Leaf-blades dentate, truncate or
cordate at base. Stem finely pubescent
with glandless hairs.

10 Rhcdora [January

Petals not exceeding the ovate sepals. Cap-
sule-lobes rounded, the most distal point
of each about midway between the style
and the lateral margin. Style shorter
than the capsule. ^
Leaf-blades oblong-ovate, crenate-serrate.
Capsule 4-5 mm. wide, with a narrow
notch about one-third depth of cap-
sule; the stout style (less than 1 mm.
long) about equaling the capsule-
lobes 13. V. agreslis.

Leaf-blades ovate, dentate with rounded
teeth. Capsule 4 mm. wide, less
deeply and narrowly notched; the
slender style (1-1.5 mm. long) surpass-
ing the capsule-lobes 14. V. polita.

Petals much exceeding the narrowly ovate
sepals. Capsule-lobes acutish in pro-
file, the most distal point of each near
the lateral margin. Style as long as the

capsule 15. V. persica.

Capsule-lobes united only at base. Leaf-blades
serrate, narrowed at base. Stem pubes-
cent with gland-tipped hairs 16. V. biloba.

Leaf-blades broadly and shallowly cordate, 3-5-
lobed, the lobes rounded and entire. Sepals
broadly ovate, conspicuously ciliate. Cap-
sule turgid, scarcely notched at apex, gla-
brous. Seeds 2.5-3 mm. long, blackish . . 17. V. hederaefolia.

1. Veronica maritima L.

Veronica maritima L., Sp. PI. 10. 1753. "Habitat in maritimis
Europae macris apricis." According to Linne, Fl. Lapp. 5. 1737:
"Ad fines Alpium Lapponicarum iuxta mare septentrionale saepius
conspicitur, in toto itinere nullibi coposior visa est, quam in maritimis
Tornoensibus. " Type carefully described by Sir. J. E. Smith, and
specimen from Tornea, collected by C. P. Laestadius, seen in Herb.
New York Botanical Garden. This is a form with long-acuminate,
sharply serrate leaves.

Veronica longifolia L., 1. c. 10. 1753. "Habitat in Tataria,
Austria, Svecia." Diagnosis quoted from Linne, Hort. Ups. 7.
1748, where this plant is more fully described and is cited: "Habitat
in Tataria." Sir J. E. Smith, in Rees Cyclop. 37: Art. Veronica,
no. 10, 1819, discussing the specimens in the Linnean Herbarium,
carefully contrasts these two Linnean species. V. longifolia he dis-
tinguishes by its leaves less deeply and doubly serrate, on shorter
petioles, and calyx shorter (not longer) than the tube of the corolla,
its lobes broad, ovate and nearly equal (not unequal and narrow).
All which contrasts certain plants; however each character varies,
and it seems difficult or impossible to distinguish these as species.

Roadsides and waste land, from the Magdalen Islands and Prince
Edward Island to Quebec, Connecticut and central New York.
Introduced from northern Eurasia.

1921] Pennell, — "Veronica" in North and South America 11

What is here termed Veronica maritima presents remarkable vari-
ability, and whether it constitutes one polymorphic species, one
species with varieties, or a group of closely related species, must be
decided by field-study in the lands in which it is native. Until
Old World students arrive at a much more definite consensus of
opinion, there seems to be slight profit in our attempting further
analysis and identification of the large number of named variants of
this group. But to show the range of variation seen in American
material I present this doubtless artificial outline of forms:

Corolla (of at least largest flowers) 5.5-7 mm. long. Filaments

much exceeding the corolla. Leaves opposite or in threes, 6-15

cm. long, long-acuminate; sharply serrate.

Leaf-blades beneath pubescent over most of surface. The most

prevalent form northward, Magdalen Islands and Nova

Scotia to Massachusetts and northern New York Forma A

Leaf-blades beneath glabrous or slightly pubescent on veins. In
Nova Scotia and Massachusetts. Evidently a mere variant

of Forma A Forma B

Corolla 4-5.5 mm. long.

Leaf-blades beneath pubescent over entire surface and usually also
above. Filaments nearly twice as long as the corolla.
Leaves in threes or fours, 8-10 cm. long, the blades linear-
lanceolate, long-acuminate, sharply serrate. Buckfield
and Cliff Island, Maine. Probably the typical V. mari-
tima L. As in Forma A, but smaller-flowered Forma C

Leaves opposite, 3-5 cm. long, the blades oblong-lanceolate,
acute, dentate-serrate. In Quebec, Massachusetts, and

central New York Forma D

Leaf-blades beneath glabrous, or slightly pubescent on the veins,
lanceolate, or oblong-lanceolate.
Leaves opposite, or very rarely in threes, 3-10 cm. long, the
blades acuminate or acute, dentate-serrate or somewhat
sharply serrate. Filaments usually only slightly longer
than the corolla. The most prevalent form southward,
and possibly a distinct species. On Prince Edward Is-
land; from Maine to Vermont and Connecticut; in

Martinique Forma E

Leaves opposite, 5-8 cm. long, the blades obtuse or acutish,
crenate-serrate. Filaments much exceeding the corolla.
An anomalous form, perhaps a hybrid containing some
Veronica spicata ancestry. Elmira, New York Forma F

Perhaps even the little-understood Veronica spuria L. is to be in-
cluded in this aggregate species. If so, as this name has precedence
of position, according to the American Code1 it must be adopted.
Veronica spuria L., Sp. PI. 10. 1753 ("Habitat in Europa australiore,

1 Priority of position within a work, or as in this case on a single page, affords an
unfortunate rule to follow, because it does not indicate any time-precedence in the
author's mind. Perhaps a better principle would be to select, among 'species' actually
seen by the author, that earliest known by him. In the case above, Linne knew
Veronica spuria and lon'Jfolia from 1748, but V. maritima from 1737.

12 Rhodora [January

Sibiria"), is really adopted from Linne's Hortus Upsalensis, where
the plant is more fully described, contrasted with V. maritima, and
stated to be from Sibiria. A plant with stems and under surface of
leaves glabrous, the latter with acute (not acuminate) serratures,
from Siberia, would appear to be the same as certain specimens seen
in Herb. New York Botanical Garden from Altai and Manchuria.
These all have very short petioles, a feature especially stressed by
Sir J. E. Smith in his re-description of the Linnean plant, and so
appear to be in contrast with any form seen of V. maritima L. It
will be noticed that this interpretation of Veronica spuria L., is not
that of Ledebour, Fl. Ross. 3: 231. 1846, and others, who hold
the name for a plant with leaves narrowed at base.

2. Veronica spicata L.

Veronica spicata L., Sp. PI. 10. 1753. "Habitat in Europae
borealis campis." A specimen was in the Linnean Herbarium in
1753, and the plant is included in Flora Suecica from Sweden. Sev-
eral specimens from Sweden seen in Herb. New York Botanical
Garden.

Roadside, Stockholm, northern New York. Introduced from
northern Eurasia.

3. Veronica mexicana S. Wats.

Veronica mexicana S. Wats, in Proc. Am. Acad. 23: 281. 1888.
"On cool damp bluffs of streams in the Sierra Madre, Chihuahua,
C. G. Pringle (n. 1349), Sept., 1887." Isotype2 seen in Herb. Col-
umbia University at New York Botanical Garden.

Mountain slopes, at altitudes of 1950 to 2400 meters; Sierra Madre
of southern Chihuahua and Durango.

4. Veronica Copelandii Eastw.

Veronica Copelandii Eastw. in Bot. Gaz. 41 : 288. /. 2. 1906. " Col-
lected on Mount Eddy [California] at an elevation of 2500m by
Dr. Edwin Bingham Copeland, August 18, 1903, distribution of C.
E. Baker, 1903, no. 3931." Isotype, collected on Mount Eddy,
Siskiyou County, and distributed by C. F. Baker, seen in Herb.
New York Botanical Garden.

Alpine slopes, known only from the mountains of Siskiyou, and
perhaps adjoining counties in northern California.

5. Veronica Cusickii A. Gray

Veronica Cusickii A. Gray, Syn. Fl. N. Am. »: 288. 1878. "Al-
pine region of the Blue Mountains, W. Oregon. W. C. Cusick."

2 The word "Isotype" is used to designate a specimen of the original collection,
other than the type itself. See Torreya 19: 13. 1919.

1921] Pennell, — "Veronica" in North and South America 13

Veronica Allenii Greenm. in Bot. Gaz. 25: 263. 1898. "Col-
lected by 0. D. Allen along Paradise river on Mt. Rainier [Washing-
ton], altitude 1700m, August 20, 1897, no. 95a." Isotype seen in
Herb. New York Botanical Garden. Differs, as stated by Green-
man, " in its smaller flowers, the white corolla, and less exserted sta-
mens and style." As stated by Macbride and Payson, "typical
V. Cusickii is common on Mt. Rainier," so that it seems probable
that Allenii is better considered as an albino form, forma Allenii
(Greenm.) Pennell, comb. nov. However the single collection
known differs from V. Cusickii by the following contrast: corolla
3-4 mm. long, white (not 5-6 mm. long, blue), and sepals less unequal.
The plant should be re-collected and studied.

Veronica Cusickii Allenii (Greenm.) Macbr. & Pays, in Contrib.
Gray Herb. II. 49: 67. 1917. I should not consider an albino state,
occurring with its species, as of rank higher than forma.

Mountain slopes, Cascade and Olympic mountains of Washington,
eastward to Coeur d'Alene Mountains of northern Idaho and Blue
Mountains of northeastern Oregon.

6. Veronica fruticans Jacq.

Veronica fruticans Jacq., Enum. Stirp. Vind. 2,200. 1762. " Cres-
cit copiose in herbidis saxosisque montium Schneeberg, Schneealbl,
Gans. &c. [Austria]." Description of calyx as covered with a very
light pubescence, of the corolla as larger than V. alpina (by which
name V. jnimila Allioni is intended) and more blue, indicate that
this name belongs to the plant now discussed rather than to V. fruti-
culosa L. The Greenland plant has been known by the later
name Veronica saxatilis Scop.

East Greenland (Lange), and on Disco Island, West Greenland.
Through the mountains of western Eurasia, Scandinavia, Scottish
Highlands and Alps.

Veronica fruticulosa L., Sp. PI. ed. II. 15. 1762. ("Habitat
in Alpibus Austriae, Helvetiae, Pyrenaeis".) From Linne's brief
description, the description and plate of Haller cited (Stirp. Helv.
1: 532. pi. 9. 1742), the south European range assigned, and the
identification by Sir. J. E. Smith (in Rees Cyclop. 37: Art. Veronica,
no. 20) who had Linne's specimen before him, this name must be held
for the pink-flowered, larger, slightly glandular-pubescent, longer- and
at times dentate-leaved plant of the Alps and Pyrenees. LinntVs
description of the calyx as glabrous is apparently inaccurate, as this
is somewhat glandular-pubescent. Scopoli, in his Flora Carniolae,
ed. II. 1: 11 and 19. 1772, well contrasts J'cronica fruticulosa and
V. fruticans, although describing both as new species from Carniola.

14 Rhodora [January

The former, his V. frutescens, is a plant of lower and subalpine sta-
tions, while the latter, his V. saxatilis, is truly alpine. For contrast-
ing illustrations see Reichenbach, Ic. Fl. Germ, et Helv. 20: pi.
1717. 1862.

7. Veronica alpina L.

Veronica alpina L., Sp. PI. 11. 1753. "Habitat in alpibus Eur-
opae." Based primarily upon the plant described in Linne, Fl.
Suec. 5. no. 13. 1745, where the locality is stated: "Habitat in
Alpibus Lapponicis monte Wallewari." This in turn is based upon
Linne, Fl. Lapp. 7. no. 7. pi. 9. f. J,. \131, where Linne's own Lap-
land plant is well described and illustrated. Obviously the name
must be given to the species of northern Europe now considered,
Linne using the term "alps" as applicable to any high mountain.

Veronica alpina corymbosa Hornem., Fl. Dan. fasc. 33: 3. pi. 1921.
1829. "Auf der Insel Disco in Greenland. Gefunden von Capi-
tain-Lieutenant Holb(J>ll." Figured as with an abbreviated, but
obviously young raceme. Specimen from Disco Island seen in Herb.
New York Botanical Garden.

Open slopes, East Greenland. Also in Scandinavia, and the
Highlands of Scotland.

Under this name have long been included two species which may
be distinguished as follows:

Capsule glabrous. Sepals glabrous on back, eiliate on margins,
apparently but little shorter than the corolla. Plant usually
1-2 dm. tall, usually little branched at base V. alpina

Capsule pubescent with glandless hairs. Sepals pilose on back
as well as margins, much shorter than the corolla. Plant
usually .5-1 dm. tall, usually much branched at base V. pumila

Veronica pumila Allioni, Fl. Pedem. 1: 75. pi. 22. f. 5. 1785
("In saxosis summae alpis Albergian dictae"), is stated by Allioni
to differ from "Veronica alpina" of the Italian Alps, in its leaves
'not crenate, but dentate, rugose and more acute.' Individual
variants of the South European "alpina" answer this characteri-
zation, and Bertolini, in his Flora Italica 1: 89. 1833, assures us
that he has obtained specimens proving this to be but a state. Fre-
quently the leaves of variants are dentate and more acute.

This is the species known as "Veronica alpina" through southern
Europe, the Pyrenees, Cevennes, and Alps, and as var. lamocarpa
in northern Europe, Scandinavia, and the Highlands of Scotland.
Wahlenberg, in his Fl. Carpat. Princip. 5. 1814, called this Veronica
alpina australis, and the true "alpina" V. alpina lapponica.

8. Veronica Stelleri Willd.

Veronica Stelleri Willd.; Link, ffe-b. 1^: 40. 1820. "In Herbar.
[Willdenow bei Berlin] aus Kamtscha|^i von Pallas gesandt. " Accord-

1921] Pennell, — "Veronica" in North and South America 15

ing to Chamisso and Schlechtendahl, in Linnaea 2: 557. 1827:

" Veronica Stelleri Pallas in Herb. Willd. n. 192 quam

e Camtschatca et e Curilis Pallas habuit, in Unalaschka insula Aleu-
torum legimus frequentem." By them very fully described, and
contrasted with their V. alpina unalaschkensis. Specimen from
"Mts. of Unalaska, 2000 ft.," collected by A. Kellogg no. 295, seen
in United States National Herbarium.

On the Aleutian and Pribiloff Islands, Alaska.
9. Veronica Wormskjoldii Roem. & Schult.

Veronica Wormskjoldii R. & S., Syst. 1: 101. 1817. "V. villosa
Wormskjold .... In Gronlandia." Evidently the species
now considered.

Veronica alpina unalaschkensis C. & S. in Linnaea 2: 556. 1827.
" Legimus [Chamisso et Eschscholtz] in montosis insulae Unalaschka
Aleutorum." Collections from Unalaska made by C. F. Baker
4988, W. L. Jepson 86, 135, C. H. Merriam in 1891, and C. V. Piper
4527, seen.

Veronica alpina Wormskjoldii (R. & S.) Hook, in Bot. Mag. 57:
pi. 2975. 1830.

Veronica mollis Raf., New Fl. Am. 4: 38. 1838. "From Origon,
seen alive in gardens." Apparently this species is intended, but the
flowers are described as "pale blue" and "large."

Veronica alpina villosa (Wormskj.) Lange, Consp. Fl. Groenl. 261.
1887. "(V. villosa Wormskj. mscr.) . . . . V. Gr. [West
Greenland]: Avangnardlek 62° 25' (Hoist.)."

Moist, grassy ledges and meadows, West Greenland, northern
Labrador, Gaspe County, Quebec, Hudson Bay, and Alaska, south-
ward, in the east on Mt. Katahdin, Maine,3 and the White Mountains,
New Hampshire, in the west through the Rocky Mountains to north-
ern New Mexico, the San Francisco Mountains of Arizona and the
Ruby Mountains of Nevada, and through the Cascade Mountains
and Sierra Nevada to California.

There appears to be a tendency for plants of the Pacific ranges
from Alaska to California to have styles slightly longer, usually J^
to }/% the length of the capsule, rather than \\ to 34- Northward is
the following more pronounced variant.

9a. Veronica Wormskjoldii nutans (Bong.) Pennell, comb. nov.

Veronica nutans Bong, in Mem. Acad. Petersb. 2: 157. 1833.
"Dr. Mertens a . . . cueillies a l'ile de Sitcha. " An old speci-
men in Herb. Columbia University labeled simply " Veronica nutans
Bong. Sitcha," may be an isotype.

3 Reported by Fernald, in Rhodora 3: 176. 1901 (as V. alpina L.).

1(5 Rhodora [January

Along the Alaskan coast from Sitka to Kodiak Island, while north-
ward and on the mountains inland typical Wormskjoldii occurs.

10. Veronica serpyllifolia L.

Veronica serpyllifolia L., Sp. PI. 12. 1753. "Habitat in Europa
& America septentrionali ad vias, agros." Specimen in the Linnean
Herbarium, and plant cited in the Flora Suecica as occurring "in
pascuis sterilioribus riguis frequens," are evidently of the species now
considered.

Meadows, barrens and open woodland, from Newfoundland and
Ontario to Minnesota, South Carolina and Missouri, mostly com-
mon; British Columbia; Costa Rica; Jamaica; Venezuela. Intro-
duced from western Eurasia, or perhaps also native, in which case
our plant, which is not montane, would appear to have been indepen-
dently derived from the wide-spread mountain variety, humifusa.
10a. Veronica serpyllifolia humifusa (Dickson) Vahl.

Veronica humifusa Dickson in Trans. Linn. Soc. 2: 288. 1794.

"I found [it] upon very high mountains, and under wet shady rocks
[Scotland. James Dickson in 1789]." Description not intended to
apply to the variety now considered, but to a depressed form of it.
Also is inaccurate (as stated by me in Torreya 19: 166. 1919) in
calling for a plant with leaves often in threes and fours, a condition
which I have not observed within this species. However this must
be a form of serpyllifolia, and this name has long been current in
British floras for denoting an alpine more pubescent depressed variety
of that species. Surely the depressed habit must prove ecologic,
but, as understood long ago by Sir J. E. Smith (Fl. Brit. 1: 19.
1800), there is a hirtous V. serpyllifolia in the upland, "in montosis."
Four specimens in Herb. Columbia University, collected along streams
in the Clava Mountains, Forfarshire, Scotland, show well this variety.
The stems are but 5 cm. long, ascending or even erect at apex, and
above are pubescent with spreading hairs. That the plants are but
dwarves of this wide-spread variety is confirmed by their obviously
large corollas. The plants are so dwarfed that, due to the crowding
of the pairs, the leaves might seem whorled.4 American plants
from high altitudes become likewise dwarf and spreading.

Veronica neglecta F. W. Schmidt, Fl. Boem. 1: 12. 1794. De-
scription not seen, but in Roemer & Schultes, Syst. Veg. 1: 102.
1817, we are informed that neglecta is "hirsuta, pilis brevibus con-
fertis," while Koch, Syn. Fl. Germ, et Helv. 529. 1837, terms it a
"forma maior, fol. ovatis." This combined description surely
indicates our plant.

4 Prof. Fernald has suggested that Dickson intended to describe his plant as bear-
ing three or four pairs of leaves.

1921] Pennell, — "Veronica" in North and South America 17

Veronica serpyllifolia humifusa (Dickson) Vahl, Enum. PL 1: 65.
1805.

Veronica ruderalis Vahl, 1. c. 66. 1805. "Habitat in ruderatis
versuris et humidis locis frigidis Peruviae." Re-naming, with a re-
arranged description, of the plant called by Ruiz and Pavon (Fl.
Peruv. et Chil. 1: 6. 1798) "Veronica serpillifolia" and obtained
by them "in ruderatis, versuris et humidis locis frigidis Pillao [Peru]."
Description distinctive of the variety now considered.

Veronica serpyllifolia neomexicana Cockerell in Am. Nat. 40: 872.
1906. " I found it at the top of the Las Vegas Range in New Mexico,
at 11,000 feet, June 28, 1902." Isotype seen in Herb. New York
Botanical Garden.

Veronica funesta Macbr. & Pays, in Contrib. Gray Herb. II. 49:
68. 1917. "Oregon: Swan Lake Valley, June 21, 1896, Elmer I.
Applegate, no. 424 (Type, Gray Herb.)." Type, collected "along
mountain streams," seen in Gray Herbarium. The filaments are
obviously shorter than in V. Cusickii A. Gray, and the (immature)
capsule is wider than long.

Alpine meadows, reaching sea-level northward, from Labrador to
Alaska, south, eastward to northern Maine and Vermont, westward
through all high ranges of Canada and the United States, at scattered
stations in Mexico5 (Popocatepetl and Ixtaccihuatl), and through the
Andes from Colombia to Bolivia. Through the mountains of Eur-
asia, from Scandinavia, Scotland and the Pyrenees to the Himalayas.
Very wide-ranging, and certainly the parent of the species, V. serpyl-
lifolia.

In Eurasian botany this pubescent larger-flowered plant of moun-
tains has repeatedly been distinguished from Veronica serpyllifolia,
specifically, varietally, or as but a mountain-form of that species.
Among names proposed for it are: Veronica serpyllifolia pubescens
Spenner, Fl. Frib. 351. 1826, from Germany; V. serpyllifolia bo-
realis Laestad. in Nov. Act. Soc. Ups. 11: 211. 1839, from Sweden;
V. serpyllifolia major Baumg., Enum. Stirp. Transsilv. 1: 20. 1816,
from Transsilvania; V. serpyllifolia major Schur., Enum. PI. Trans-
silv. 500. 1866, also from Transsilvania (name apparently inde-
pendently chosen; plant well-described); and V. serpyllifolia al-
pina Hook., Brit. Fl. 4. 1830, from Scotland. Veronica font an a
Willd.; Link, Jahrb. I3: 41. 1820, is a name which has been used in
Alaskan botany.

6 To be expected on all high Cordilleras of Mexico arjd Central America.

18 Rhodora [January

11. Veronica peregrina L.

Veronica peregrina L., Sp. PI. 14. 1753. "Habitat in Europae
hortis, arvisque. " Diagnosis quoted from Linne, Fl. Suec. 6. no.
15. 1745, where we are told that the plant "habitat in cultis &
terra nuda Upsaliae, rarissima apud nos hodie plana, olim forte
copiosior evasura. " A plant of 'cultivated fields and bare earth,'
known from a single locality in Sweden, and there ' formerly abundant
but now very rare,' would seem to have been an adventive. That
this was Linne's opinion is shown by the specific name chosen, mean-
ing "foreign."

Veronica caroliniana Walt., Fl. Carol. 61. 1788. Type doubtless
from lower South Carolina. The radical leaves are described as
subincised, cauline subserrate. Evidently this is a pronounced phase
of the 'romana' type, discussed below, and it is well-interpreted by
such a specimen as House 3179 from Clemson College, South Carol-
ina.

Veronica carnulosa Lam., Ency. Meth., Illust. 1: 47. 1791. "Ex
Europa & America septentr. in arvis."

Moist soil, river-banks, gardens and cultivated fields, usually
appearing as a weed; wide-spread through eastern North America
from New Brunswick to Iowa, Florida and Texas; also seen from
British Columbia, Alaska, New Mexico, Oregon, Bermuda and Ja-
maica. Westward passes into the yet more widely ranging variety
xalapensis.

In the Species Plantarum, 1753, Linne twice described this species,
once from plants known living to him, as Veronica peregrina, and
once from a specimen from southern Europe in his herbarium as
V. romana. V. peregrina was supposed to possess leaves lanceolate-
linear and very entire, while V. romana had these oblong and sub-
dentate. The leaves of this species vary from one state to the other,
and on each individual the larger leaves tend to the "romana" type.
As romana has priority of position on page 14 of the Species Plant-
arum, and its diagnosis is descriptive of a specimen in the Linnean
Herbarium, a claim might be made that this name should be used
for our plant. However the citations in the synonomy of V. romana
all pertain to another species, later separated by Linne as Veronica
acinifolia (Sp. PI. ed. II. 19. 1762), and the specific name "romana"
was adopted from " Veronica minima, clinopodii minoris, folio glabro,
romana. Bocc. mus. 2. p. 29. t. 102." As the Linnean diagnosis
would also include Boccone's plant, which was clearly illustrated
in the latter's Museo di Piante Rare della . . . Italia . . .
19. pi. 102. 1697, I think we should hold romana for this species,
placing acinifolia in its synonomy.

1921] Pennell, — "Veronica" in North and South America 19

11a. Veronica peregrina xalapensis (H.B.K.) Pennell

Veronica xalapensis H.B.K., Nov. Gen. et Sp. 2: 389. 1818.
" Crescit in Regno Mexicano prope Xalapa (alt. 630 hex.) in nemor-
ibus Liquidambaris Styracifluae [Humboldt & Bonpland]."

Veronica chillensis H. B. K, 1. c. 390. 1818. " Crescit in cultis
Regni Quitensis prope Chillo, alt. 1340 hex. [Humboldt & Bonpland]."
Described as differing from V. xalapensis in having stem repent,
leaves wider (oblong-spatulate instead of oblong), and calyx-lobes
narrower (lanceolate instead of oblong), at length reflexed. All
these are points of normal variation in this variety, excepting that
the plant is never truly repent. In the full description the word
"repentes" is followed by the truer statement " adscendentes. "

Veronica peregrina xalaj)ensis (H.B.K.) Pennell in Torreya 19:
167. 1919.

Environment as in Veronica peregrina, with which over an exten-
sive area transitional forms occur; wide-spread and usually common
through western North and South America from Alaska and Yukon
to Chile and Argentina, in the Tropical portions of its range found
only on the upper Cordilleras; eastward, in the United States fre-
quent nearly to the Mississippi River, and sporadically eastward,
probably as an introduction, to New England; also, probably also
introduced, in Brazil.

12. Veronica arvensis L.

Veronica arvensis L., Sp. PI. 13. 1753. "Habitat in Europae
arvis, cultis." Diagnosis quoted from Linne, Fl. Suec. 6. no. 16.
1745, where we are told that the plant occurs in Sweden "in agris
ruderatis cultis frequens." Our American introduced plant agrees
well with the description of this.

Gardens and fields, or in dry woods, on cliffs and talus slopes,
mostly common from Newfoundland to Iowa, Georgia and Okla-
homa; southern Alaska to Oregon; Bermuda; Jamaica; Argentina.
Introduced from Eurasia.6

13. Veronica agrestis L.

Veronica agrestis L., Sp. PI. 13. 1753. "Habitat in Europae
cultis, arvis." Diagnosis quoted from Linne, Fl. Suec. 6. no. 17.
1745, where the plant is said to occur in Sweden "in agris, areis,

6 This species frequently grows in such "native" environments that the question
of its being indigenous to the Northeast has been raised. But its weed-like character,
ensuring its early introduction, and the fact that its American range is not so great
as would be expected of such a species if native, leads me to think it introduced.
See also Fernald in Rhodoha 2: 137. 1900. In the case of this and Veronica officinalis
L., below, the burden of proof is on those who would claim them indigenous to both
hemispheres.

20 Rhodora [January

cultis." According to the statement of Fries, Novit. Fl. Suec. 65.
1819, the Linnean agrestis is identical with his own V. versicolor,
being the only one of Fries' segregates common or previously known
in Sweden. Two specimens from Sweden seen in Herb. New York
Botanical Garden, one collected at Scania by N. H. Nilsson in 1881.

Roadsides and rocky places, Newfoundland to Quebec and New
Brunswick; on ballast at New York City and Philadelphia; Ber-
muda. Introduced from central and northern Europe.

14. Veronica polita Fries

Veronica polita Fries, Novit. Fl. Suec. 63. 1819. "Ubique in
arvis Scaniae [Sweden]." In the second edition of the Novitiae we
are told that this is the only known station in Sweden. Well de-
scribed, and in the second edition contrasted with the V. agrestis
L. (V. versicolor Fries), both of which species of Fries were soon after
illustrated in Reichenbach's Kupfersammlung kritischer Gewachse,
plates 246 and 277 respectively. I am unable to maintain Fries'
contrast as to pubescence and the veining of the sepals. For further
discussion of this name and of the uncertain identity of the older
Veronica didyma Tenore, Prod. Fl. Nap. 6. 1811, of Italy, see
Ernst Lehmann in Bull. Herb. Boiss. II. 8: 237-244. 1908. Ten-
ore's mention of leaves profoundly crenate suggests polita, which is
the species of this group commonest in southern Italy.

Veronica crenulata Sesse & Mocino, Fl. Mex. 5. 1892. "Habitat
in montibus Sancti Eremi PP. Carmelitarum [Mexico. Mocino &
Sesse]." Description appears to be of the species now considered,
which is well-established in Mexico. Not V. crenulata Hoffm.,
1803.

Ballast, roadsides and gardens, occasional from New York to Ohio,
Florida and Texas; central Mexico; Argentina. Introduced from
central and southern Eurasia.

15. Veronica persica Poir.

Veronica persica Poir., Encyc. Meth., Bot. 8: 542. 1808. "Croit
dans la Perse. On la cultive au Jardin des Plantes de Paris (V. v.)."
Apparently this is our species, but description differs from prevalent
forms in stating the leaves to be very obtuse and ordinarily slightly
longer than the pedicels and the corolla to be shorter than calyx.
However these states are easily reconcilable to certain stages or
forms, and our species is a plant well-known from Persia. Lehmann
doubtfully identifies persica as this species, but Lacaita (in Jour. Bot.
56: 55. 1918), after examination of Poiret's type, declares these
identical.

Veronica precox Raf., Atl. Jour. 79. 1832. "Grown in the [Bar-
tram's Botanic] Garden [near Philadelphia] from seeds received
from a place unknown." Not V. praecox All., 1789.

1921] Pennell, — "Veronica" in North and South America 21

Veronica diffusa Raf., New Fl. Am. 4: 38. 1838. "Native of
-, naturalized on the Schuylkill near Philadelphia."

Re-naming of V. "precox Raf.

Veronica rotundifolia Sesse & Mocino, Fl. Mex. 5. 1892. "Habi-
tat in Eremo P. P. Carmelitarum [Mexico, Mocino & Sesse]." De-
scription apparently of the species now considered, although it may be
that I have transposed the application of this name and V. crenulata
S. & M. Not V. rotundifolia Ruiz & Pavon, 1798.

Fields and roadsides, occasional, or westward locally common,
through Temperate North America, from Newfoundland and southern
Alaska, south to Georgia, Texas and California; Mexico; Jamaica;
Colombia; Chile. Introduced from southern Eurasia.

Our plant has also been known as Veronica Tournefortii C. C.
Gmel., V. Buxbaumii Tenore, and V. byzantina (Smith) B.S.P.
The two last are subsequent names, dating as species from 1811 and
1888 respectively. The original description of V. Tournefortii C.
C. Gmel., Fl. Bad. 1: 39. 1805, was composite, based upon a plant
escaped from the botanic garden to fields near Carlsruhe, Baden,
and upon a specimen brought by Tournefort from the Levant, which
had recently been described as V. filiformis Smith (in Trans. Linn. Soc.
1: 195. 1791.). The former element was probably our species, but
the name Tournefortii should be applied to Tournefort's plant, and
this is the basis of V. filiformis, a distinct though related species.
Moreover the name was antedated by Veronica Tournefortii F. W.
Schmidt, Fl. Boem. 7. 1793. (Description not seen, but the publica-
tion of this name verified by Lacaita in his discussion of this whole
problem in Jour. Bot. 55: 271-276. 1917.)

16. Veronica biloba L.

Veronica biloba L., Mant. 172. 1771. "Habitat inter Cappado-
ciae segetes. D. Schreber. "

Collected at Yonkers, New York, by E. P. Bicknell; also at
Logan, Utah, by C. P. Smith, 1604 and 2167, and by him com-
mented upon (under the name of V. campylopoda Boiss.) and illu-
strated in comparison with V. persica Poir., in Muhlenbergia 6: 61.
1910.

Veronica campylopoda Boiss., Diagn. PI. Nov. 4: 80. 1844, dis-
tinguished from V. biloba as having its leaves and sepals narrower,
the former hardly denticulate to entire above, its pedicels recurved,
its seeds strongly rugulose and its style longer, half the length of the
capsule, seems not to be definitely separable by any of these charac-

22 Rhodora [Januart

ters. Sir J. D. Hooker, Fl. Brit. India 4: 295. 1884, assures us that
the plant lacks distinctness — " I cannot distinguish it as a variety
even" — and in confirmation of his statement that "the seeds vary
much in depth of pitting," it may be mentioned that Reichenbach's
plate cited by Boissier as illustrative of true biloba shows seeds deeply
rugulose. Our plants seem quite intermediate, with the leaf-breadth
and leaf-serration of biloba, but with the pedicels tending slightly
to recurve and with the seeds and style as described for campylopoda.
They match well a specimen of Boissier's collected at Roscheya,
Syria, May, 1846, and named by him Veronica campylopoda.

17. Veronica hederaefolia L.

Veronica hederaefolia L., Sp. PI. 13. 1753. "Habitat in Europae
ruderatis. " Diagnosis quoted from Linne, Fl. Suec. 7. no. 18.
1745, where the plant is said to occur "in Scania [Sweden] campestri
in ruderatis ad urbes & pagos. " Specimen from Sweden, from her-
barium of Per Larson, seen in Herb. Columbia University.

(?) Veronica reniformis Raf. in Med. Repos. New York 5: 360.
1808. "In New Jersey [C. S. Rafinesque in 1803-4]." I am unable
to be certain of the identity of this from the short description : " stem
procumbent, branched ['branded'], leaves sessile, reniform, hairy,
entire, flowers axillar, solitary." It is possibly V. hederaefolia L.,
in which case the petioles must have been so short as to be unnoticed,
or V. arvensis L., with unusually obscure serration of leaf, or else
some introduced species not since reported from America.

Orchards and roadsides, occasional from New York to North
Carolina. Introduced from Europe.

(To be continued.)

"VERONICA" IN NORTH AND SOUTH AMERICA
Francis W. Pennell

(Continued from p. 22.) ■

II. VERONICA L., Subgenus 2. EUVERONICA Pennell

Veronica L., Sp. PI. 9. 1753. Type species, V. officinalis L.,7 of
Europe.

A. Capsule pubescent. Stems, pedicels, leaves and sepals
pubescent. Leaf-blades oval or ovate, crenate-
serrate to dentate. Plants of dry soil.
Leaf-blades dentate, cordate or truncate at base.
Sepals 3.5-5 mm. long, linear-lanceolate, ex-
ceeding the capsule. Capsule with hairs not
glandular nor dark-jointed. Plant ascending
or erect.
Sepals unequal, the longest 4-5 mm. long. Cap-
sule slightly notched. Style 4-5 mm. long,
longer than the capsule. Leaf-blades sessile
or nearly so. Racemes over 10-flowered, the
pedicels more than 1 mm. long.
Corolla 7-8 mm. long, violet-blue, the
largest lobes ovate. Anterior sepals much
exceeding the posterior. Capsule longer
than wide. Racemes 30-60-flowered, the
pedicels scarcely exceeding their bracts.
Leaf-blades coarsely dentate. Stem erect,

4-8 dm. tall 18. V. latifolia.

Corolla 5-6 mm. long, paler violet-blue, the
largest lobes nearly orbicular. Anterior
sepals slightly exceeding the posterior.
Capsule wider than long. Racemes 10-
20-flowered, the pedicels much exceeding
their bracts. Leaf-blades crenately den-
tate. Stem ascending, 1-3 dm. tall.... 19. V.Chamaedrys.
Sepals equal or nearly so, 3.5-4 mm. long. Cap-
sule deeply notched, wider than long. Style

7 Selected, among the several species common to both Linne and Tournefort,
which answer Linne's generic characterization in Genera Plantarum, ed. V. 10. 1754,
and are native to Linne's country Sweden, because of its officinal nature. This
species had a long historic right to the name "Veronica."

30

Rhodora

[February

.5 mm. long, much shorter than the capsule.
Leaf-blades shortly petioled. Racemes 5-
10-flowered, the pedicels less than 1 mm.

long 20. V. javanica.

Leaf-blades oval, crenate-serrate, narrowed to a
petiolar base. Sepals 2-3 mm. long, oblong-
lanceolate, shorter than the capsule. Capsule
as wide or wider than long, notched, with hairs
dark-jointed and some of them glandular.
Style 2.5-3.5 mm. long, shorter than the cap-
sule. Plants repent, ascending at apex.
Corolla 8-9 mm. long, violet-blue. Capsule 5 mm.
long, as wide as long, its lobes rounded, the
most distal point of each midway between
the style and the lateral margin. Racemes
3-5-flowered, the pedicels longer than the
capsule. Leaf -blades about 2 cm. long, cre-
nate-serrate with very low teeth, hirsute
above, glabrous beneath, conspicuously cili-

ate. Stem less than .5 dm. long. 21. V. grandiflora.

Corolla 3-4 mm. long, pale-lavender, with lavender
blue lines on the posterior side. Capsule
3-4 mm. long, wider than long, its lobes with
the most distal point of each near the lateral
margin. Racemes 20-30-flowered, the pedi-
cels shorter than the sepals or capsules. Leaf-
blades 2-5 cm. long, crenate-serrate with
prominent teeth, pubescent on both surfaces
but not obviously ciliate. Stem extensively

repent, 2 dm. long or longer 22. V. officinalis.

A.'. Capsule glabrous or with a few minute gland-tipped
hairs. Stems, pedicels, leaves and sepals glabrous,
or very rarely pubescent. Leaf-blades oblong-
ovate to linear, finely serrate to entire. Aquatics.
B. Capsule not conspicuously wider than long, and
scarcely or not two-lobed. Sepals nearly or quite
equaling the capsule, slightly unequal, the an-
terior longer. Leaf -blades oblong-ovate to lance-
olate, obtuse to acuminate, serrate to crenate-ser-
rate. Stem glabrous or pubescent with minute
gland-tipped hairs. Racemes usually of more
than 10 flowers, the relatively stout pedicels as-
cending-spreading.
Leaf-blades all petioled, prevailingly ovate-ob-
long, acutish to obtuse. Racemes usually
10-25-flowered, the pedicels 5-13 mm. long.
Plants mainly emersed, glabrous throughout.
Capsule slightly wider than long, notched.
Leaf -blades oblong-oval, widest at or above the
middle, narrowed at base, mostly broadly
rounded at apex. Style 1.5-2 mm. long,
obviously shorter than the capsule. Plant-
extensively repent, ascending at apex. . . .23. V. Beccabunga.
Leaf-blades lanceolate to ovate, widest at or
near the base, mostly acute or acutish at
apex. Style 2-3 mm. long, scarcely shorter
than the capsule. Plant repent only at
base, soon ascending-erect 24. V. americana.

1921] Pennell,— "Veronica" in North and South America 31

Leaf-blades, at least the upper on the flowering-
stems, sessile and clasping, obtuse to acumi-
nate. Racemes 15-60-flowered, the pedicels
3-8 mm. long. Plants of deeper water, usu-
ally mostly submersed.
Capsule 2.5-4 mm. long. Style 1.3-2 mm. long.
Cauline leaf-blades acute to acuminate.
Sepals acute to acuminate. Capsule scarcely
wider than long, not or slightly notched.
Leaf-blades serrate with close teeth
(four or more to 1 cm.). Racemes usu-
ally 30-60-flowered, with pedicels 4-8
mm. long.
Stem distally, rachis and pedicels glabrous.
Capsule globose-ovoid, acutish or
rounded, not or scarcely emarginate.
Style 1.5-2 mm. long. Leaf -blades
oblong-ovate, mostly broadest about
the middle, the lower usually obviously
narrowed at base or petioled.
Sepals acuminate, 4-5 mm. long. Cap-
sule 3-4 mm. long. Pedicels 5-8

mm. long 25. V. Anagallis-aquatica.

Sepals acute, 2.5-4 mm. long. Capsule
2.5-3 mm. long. Pedicels 3-5 mm.

long 25a. V. Anagallis-aquatica Brittonii.

Stem distally, rachis and pedicels finely
pubescent with minute gland-tipped
hairs. Capsule nearly globose, slightly
emarginate. Style 1.3-1.5 mm. long.
Leaf-blades lanceolate, mostly broad-
est near the base, all obviously clasp-
ed . ing 26. V. glandifera.

Sepals obtuse to acutish. Capsule obviously
wider than long, evidently notched.
Leaf-blades crenate-serrate with remote
teeth (two to three to 1 cm.), lanceolate,
all clasping. Racemes usually 15-30-
flowered, the pedicels 3-6 mm. long.
Stem distally, rachis and pedicels glabrous.

Style 1.4-1.6 mm. long 27. V. catenata.

Stem distally, rachis and pedicels finely
pubescent with minute gland-tipped
hairs. Style 1.5-1.8 mm. long

27a. V. catenata glandulosa.
Capsule 2 mm. long, globose, not or scarcely
notched. Style .7-1 mm. long. Cauline
leaf-blades oblong-lanceolate, obtuse or
obtusish. Stem distally, rachis and pedi-
cels usually pubescent with minute gland-

_. _. . tipped hairs 28. F. undulala.

B . Capsule much wider than long, strongly two-lobed.
Sepals shorter than .the capsule, equal. Leaf-
blades linear or lanceolate, remotely setaceous-
toothed or entire. Stem glabrous or pubescent
with glandless hairs. Racemes 5-20-flowered,
the filiform pedicels reflexing in fruit 29. V. scutellata.

32 Rhodora [February

18. Veronica latifolia L.

Veronica latifolia L., Sp. PI. 13. 1753. "Habitat in Helvetia,
Bithynia. " The identity of this species has been much disputed, on
one side being such statements as Bentham in DC, Prod. 10: 469.
1846, who considers it to be a broad-leaved form of V. Teucrium L.,
on the other Kerner in Oesterr. Bot. Zeitschr. 23: 367-369. 1875,
who makes a strong plea for its identification as V . urticaefolia Jacq.,
Fl. Austr. 1: 37. pi. 59. 1773. Certainly some of the synonyms
cited by Linne appear to be V. urticaefolia, a species very readily
distinguished by its slender stem, thin smooth leaves which are
sharply serrate and long-acuminate, and its shorter racemes, on the
slender pedicels of which are borne the short sepals and small pinkish
corollas. Linne's description, in the use of the words "foliis rugosis
dentatis, " certainly does not describe urticaefolia, and moreover one
can scarcely believe that he would have omitted characterization
of the leaf-acumination. Sir J. E. Smith, in Rees Cyclop. 37: Art.
Veronica, no. 58, describes the Linnean specimen and emphatically
asserts its kinship to V. Teucrium L., not to urticaefolia Jacq. In
the absence of citation to other specimens studied by Linne, the
specimen of the Linnean Herbarium should stand as type. Sir J.
E. Smith carefully contrasts this with V. Teucrium L., but study of
the varying leaf-form of the latter confirms Bentham's view as to
their identity.

Veronica Teucrium L., Sp. PI. ed. II. 16. 1762. "Habitat in
Germania. " Linne possessed no specimen of this in his herbarium,
which readily explains his describing as new a narrower-leaved form
of this species than his own V. latifolia. The specific name is derived
from "Teucrii IV tertia species Clus. hist. 1 p. 349." L'Ecluse,
Rar. PI. Hist. 349. 1601, figures and briefly describes a plant, which
is an ovate-leaved form of the species, and says that it grows "in
herbosis collium jugis [Pannoniae .... Austriae Moraviae
. . . . & Bohemiae]." Linne's description of his plant as with
leaves "ovatis rugosis dentatis" shows the similarity of this to his
own latifolia.

Roadsides, pastures and waste land, New Hampshire to Ontario,
New Jersey and Ohio; introduced from Europe.

19. Veronica Chamaedrys L.

Veronica Chamaedrys L., Sp. PI. 13. 1753. "Habitat in Europae
pratis." The diagnosis is essentially taken from Linne, Fl. Suec.
5, no. 12. 1745, where the plant is stated to occur in Sweden "in
pratis ubique." Evidently this is the species now considered.
Several specimens from Sweden seen, one collected by Dr. W. A.
Murrill at Upsala, July, 1902, being probably a topotype.

Roadsides and meadows, occasional from Prince Edward Island to
Ontario, New Jersey and Ohio. Introduced from Europe.

1921] Pennell, — "Veronica" in North and South America 33

20. Veronica javanica Blume.

Veronica javanica Blume, Bijdr. Fl. Nederl. Ind. 742. 1826.
"Crescit in cacumine Sederato et ad cataractas fluvii Tjikundul
montis Gede [Java]." The brief original description, especially in
the phrase "spicis axillaribus, " would seem to denote the plant here
considered. I have followed Sir J. D. Hooker, Fl. Brit. Ind. 4: 296.
1884, in adopting this name, as the only named specimen which I
have for comparison, Griffith 3921 from East Himalaya distributed
by Kew Gardens as "Veronica Maddenii Edg.," is evidently this
species. There is also a previously unnamed specimen, in Herb.
New York Botanical Garden, from the Liu Kiu Islands.

Petropolis, Brazil, collected by J. Ball in 1882. Introduced from

the Oriental Region.

21. Veronica grandiflora J. Gaertn.

Veronica grandiflora J. Gaertn. in Novi Comm. Acad. Petrop. 14:
531. pi. IS, f. 1. 1770. " Kamtschatkam pro patria sua ....
in pratis alpinis . . . ., referente Stellero, copiose nascitur. "
A full description, and a carefully drawn illustration, make the
application of this name unmistakable, although the capsule is
described as smooth (the word "laevis" however, not the word
"glaber"). Apparently this was accidentally renamed by the
younger Linne (Suppl. 83. 1781), who says of it: " Veronica kamt-
chatica Gaertner Act. petropol. Habitat in Kamtschatka. " Speci-
mens, L. Stejneger 106, etc., seen from Bering Island, along the coast
of Kamchatka.

Western Aleutian Islands (Kiska and Attu Islands). Also in
Kamchatka.

Similar to, but much larger than, Veronica aphylla L., Sp. PI. 11.
1753, of the Alps of Europe; differs by having its stems frequently
1 dm. long, its peduncles longer, its leaves 2.5-4 cm. long (not 1-2
cm. long), obovate and more acute, its corollas 8-9 mm. long (not
5 mm. long), and its style 8-9 mm. long, exserted, probably as long
as the capsule (not 4 mm. long and only one-half to two-thirds length
of capsule).

22. Veronica officinalis L.

Veronica officinalis L., Sp. PI. 11. 1753. "Habitat in Europae
sylvestribus sterilibus." Refers to Linne, Mat. Med. 4, no. 11.
1749; then to Linne, Fl. Suec. 4, no. 8. 1745, where the plant is
said to occur in Sweden "frequens in sylvis praesertim exustis,"
and its medical uses are mentioned. The Linnean specimens are
more fully described by Sir. J. E. Smith in Rees Cyclop. 37: Art.
Veronica, no. 53. 1819. Specimen in Herb. New York Botanical
Garden, collected at Upsala, Sweden, July, 1902, by Dr. W. A.
Murrill, is probably a topotype.

34 Klmdora [February

Fields, barrens and open woods, mostly common; from Newfound-
land and Michigan to North Carolina and Tennessee. Apparently
introduced from Eurasia, although usually in seemingly native
habitats.
23. Veronica Beccabunga L.

Veronica Beccabunga L., Sp. PI. 12. 1753. "Habitat in Euro-pa
ad rivulos." Diagnosis quoted from Linne, Fl. Suec. 5. no. 11.
171."), where it is stated that in Sweden the plant "habitat in fossis,
rivulis, scaturiginibus passim," and that it is the "Beccabungae
Herbs Conserva, Aqua" of tlie Pharmacopoeia.

Running brooks, ditches and wet fields, well established in Quebec,
also at Rochester, New York and Perth Amboy, New Jersey. In-
troduced from Eurasia, where this species is as wide-spread as on
this continent is the following near relative.
21. Veronica americana Sehwein.

Veronica Beccabunga americana Raf., Med. Fl. 2: 109. pi. 94.
1830. "Grows from Canada to Virginia and Kentucky, near waters,
brooks, &c. " Well described, and contrast given with V. Becca-
bunga as understood by Rafinesque. Apparently this plant was
independently redescribed under this name by Torrey in Fl. New
York 2: 41. 1843, whose type I have seen in Herb. Columbia Uni-
versity.

Veronica americana Sehwein.; Benth. in DC. Prod. 10: 168.

1840. "Veronica americana (Sehweinitz! mss.) In

America boreali a Canada et Carolina usque ad flum. Oregon et in

ins, Sitcha (v. s.)" Specimen seen in Herb.

Academy of Natural Sciences of Philadelphia, labeled "Bethl."
I = Bethlehem, Pennsylvania], collected by Sehweinitz. may be an
isotype. Well contrasted with V. Beccabunga L., instancing leaf-
form and more erect habit.

Veronica americana hirsuta Coleman, Cat. Fl. PI. S. Michigan L'7,
1S74. "Southern peninsula of Michigan." Described as "plant
quite large, 24 to 30 inches high, very hirsute. " I have never seen
a pubescent form of this species, and Coleman's specimen, if extant,
should be studied.

I'cronica americana crassula Rydb. in Mem. New York Bot. Gard.
1 : 353. 1900. " In bogs, at an altitude of 2000-2500 m. Montana :
Little Belt Pass, 1896, Flodman, 77S (type)." Type seen in Herb.
New York Botanical Garden. This represents the dwarfed alpine
state of the species, which may better be considered a forma.

I'cronica oxylobtda Greene, Pittonia 5: 113. 1903. "Type speci-
mens from Golden City, Colorado, collected by myself in 1871."
Supposed to be distinguished by "its entire or subentire foliage and
the longer and almost acute capsules," features of variability within
this species.

1921] Pennell, — "Veronica" in North and South America 35

Veronica crenatifolia Greene, 1. c. 114. 1903. "The type . . .
is Baker, Earle and Tracy's n. 33, from along the Mancos River
in southern Colorado, 22 June, 1898." Isotypes seen in Herb. New
York Botanical Garden and U. S. National Herbarium. Apparently
supposed to be distinguished by its smaller size and crenate leaves,
variations frequent in V. americana.

Swamps, springs and woodland rills, from Newfoundland, Ontario
and Alaska, south, eastward to South Carolina and Tennessee, west-
ward to Chihuahua, California, and the Valley of Mexico; also on the
Commander Islands on the western side of Bering Sea. Generally
common over this wide area.

Veronica americana appears to be only inconstantly distinguish-
able from V. Beccabunga by its leaf -form and more erect habit. The
leaf is mostly narrower, widest near the truncately rounded or sub-
cordate base, narrowing to the acute or obtuse apex, and borne on
frequently shorter pedicels. The capsule-shape is the same, nearly
globose, flattened and emarginate at apex, the corolla, sepals and
pedicels are of about the same length as in that species, but the last
are usually more slender. The styles are longer and usually more
slender in americana. The leaves vary from serrate through crenate
to nearly or quite entire.

25. Veronica Anagallis-aquatica L.

Veronica Anagallis-aquatica L., Sp. PI. 12. 1753. "Habitat in
Europa ad fossas. " Description quoted from Linne, Fl. Suec. 5,
no. 10. 1745, where the plant is stated to occur in Sweden "in fossis
ad vias & paludes Uplandiae, Scaniae &c. " Described with leaves
serrate, and with citations to Tournefort and Bauhin who both term
the leaves oblong. The Swedish plant is well described by Nyman,
Utkast Sv. Vaxt. Naturh. Sver. Fanerog. 164. 1867, who tells us
that its leaves are lanceolate or oval-lanceolate, pointed, and its
capsules are rounded, very shallowly notched. All which indicates
the present broad-leaved plant with scarcely or not notched capsules,
not another plant of northwestern Europe which has elongate acumi-
nate leaves, and capsules decidedly notched, as broad as or broader than
long. Our plant has the lower leaves and those of autumnal shoots
narrowed or petioled at the base, a condition mentioned in such exact
descriptions as Hayek, Fl. Steiermark 2: 168. 1912; also the short
round form of these autumnal leaves is mentioned in Villars, Hist.
PI. Dauphine 2: 14. 1787.

Veronica lepida Phil, in Anal. Univ. Chile 91: 110. 1895. "Habi-
tat ad Vicum Cartajena (haud procul a Valparaiso [Chile]), Februario,
1895 lecta. " Described because the petioled lower leaves were
noticed.

36 Rhodora [February

Veronica micromera Wooton & Standley in Contrib. U. S. Nat.
Herb. 16: 174. 1913. " Type in the U. S. National Herbarium, no.
686250, collected along ditches about Shiprock, on the Navajo
Reservation [New Mexico], July 25, 1911, by Paul C. Standley (no.
7283). Altitude 1,425 meters." Type seen in U. S. National
Herbarium. A dwarf form, with small leaves which are more ob-
viously narrowed at base.

Slow-flowing streams, wide-spread through North and South
America; specimens seen from Michigan, Utah, New Mexico, Ariz-
ona, Argentina and Chile. Also of wide occurrence in Eurasia;
specimens seen from Germany, Switzerland, Italy, Albania, Algeria
and Syria. Of this critical species-group this is the most widely
dispersed and probably the original element.

25a. Veronica Anagallis-aquatica Brittonii (Porter) Pennell,

comb. nov.

Veronica Anagallis latifolia Britton in Bull. Torr. Bot. Club 12:

49. 1885. " In the latter part of September, 1883,

near Mahwah, Bergen Co., New Jersey, I noticed [it] in a small
stream . . . ." Type seen in Herb. Columbia University.
Not V. Anagallis latifolia Schultz, Prod. Fl. Stargard. Suppl. 3.
1819 (which is V. Anagallis-aquatica L.).

Veronica Brittonii Porter; Pennell in Torreya 19: 168. 1919.
"Type, base of Marble Hill, above Phillipsburg, New Jersey, col-
lected in flower and fruit June 24, 1892, T. C. Porter; in herbarium
Columbia University at the New York Botanical Garden."

Slow-flowing streams, western Connecticut to northern Pennsyl-
vania. For list of localities see Torroya 19: 170. 1919.
Perhaps not worthy of even varietal distinction.

26. Veronica glandifera Pennell

Veronica perfoliata Raf., New Fl. Am. 4: 37. 1838. "Florida."
Description almost certainly of the plant now considered, which
however is not authentically known from so far south. The clasp-
ing opposite leaves of V. glandifera, till closely seen, appear connate.
Not V. perfoliata R. Br.', 1810.

Veronica glandifera Pennell in Torreya 19: 170. 1919. "Type,
vicinity of Suffolk, Nansemond County, Virginia, collected in flower
and fruit, May 27, 1893, N. L. Britton and J. K. Small; in herbarium
Columbia University at the New York Botanical Garden."

Slow-flowing streams, in limestone, Virginia and Ohio to North
Carolina and Tennessee. Perhaps intergrades with V. catenata
glandxdosa.

The petioled leaves of late-summer shoots are well shown on speci-
mens of Bruce Fink 262 from Oxford, Ohio, collected August 8, 1908.

1921] Pennell, — "Veronica" in North and South America 37

While in pubescence this species parallels Palaearctic derivatives
of Veronica Anagallis-aquatica L., I am unable to place our plant of
eastern North America as of the same species as any of these. Such
species are: V. anagalloides Guss., PI. Rar. Sic. 5. pi. 3. 1829, which
has a capsule decidedly longer than wide, and not or scarcely emargi-
nate; V. oxycarpa Boiss., Diagn. I. 7: 44. 1846, with acute capsule
and leaves narrowed at base; and V. salina Schur, Enum. PI. Trans-
silv. 492. 1866, very similar to V. anagalloides.

27. Veronica catenata Pennell, sp. nov.

Flowering stem 1-3 dm. long, glabrous throughout. Leaves
lanceolate, acute or acutish, crenate to nearly entire, 3-5 cm. long,
1 cm. wide, all clasping, when submersed elongating and reaching
12 cm. long and 2 cm. wide. Racemes axillary to the upper leaves,
6-12 cm. long, 15-25-flowered. Bracts narrowly lanceolate, 4-5 mm.
long. Pedicels 3-5 mm. long, glabrous. Sepals 3-3.5 mm. long,
lance-ovate, obtusish. Corolla-lobes pale-blue. Style 1.2-1.7 mm.
long. Capsule 3 mm. long, 3.5 mm. wide, broad-globose, decidedly
emarginate. Seeds 0.5 mm. long, yellow-brown.

Type, Hot Springs, South Dakota, collected in flower and fruit
June 16, 1892, P. A. Rydberg 926, in Herb. New York Botanical
Garden. Named from the chain-like aspect of the long racemes of
short-pedicelled flowers.

Slow-flowing, streams, plains, from North Dakota and Saskatch-
ewan to Kansas and New Mexico, southward west to Nevada and
southern California.
27a. Veronica catenata glandulosa (Farwell) Pennell, comb. nov.

Veronica Anagallis-aquatica glandulosa Farwell in Rep. Mich.
Acad. Sci. 19: 249. 1917. "Zoo Park, near Royal Oak [Michigan],
[Farwell] No. 4323, July 13, 1916." Not V. Anagallis-aquatica
glandulosa Schur, Enum. PI. Transsilv. 492. 1866. Description
inadequate, but apparently of the plant now considered. As this
is a small plant and moreover is the only glandular-pubescent " Ana-
gallis-aquatica" known from Michigan, I apply the name to this.

Slow-flowing streams, western New York to Minnesota, South
Dakota, Kentucky and Oklahoma; also in western Massachusetts
and in southeastern and southern Pennsylvania. Probably inter-
grades with V. glandifera, and for the latter, in Torreya 19: 170,
[ have mistaken plants of our New York " Local Flora. "

28. Veronica undtjlata Wall.

Veronica undulata Wall.; Roxb., Fl. Ind. 1: 147. 1820. "Dis-
covered in the Turraye [India] by Mr. W. Jack. " Specimen in Herb.
Columbia University, labeled "Nepal Wallich," may be an isotype.

38 Rhodora [February

Ballast, Portland, Oregon, and Mobile, Alabama. Introduced
from southeastern Asia, where it occurs from northern India through
southern China, and in Japan.

Occasionally nearly or quite glabrous, but then readily distinguished
by the small size of the capsule and style.

29. Veronica scutellata L.

Veronica scutellata L., Sp. PI. 12. 1753. "Habitat in Europae
inundatis." Diagnosis quoted from Linne, Fl. Suec. 4. no. 9. 1745,
where the plant is said to grow in Sweden "in locis per hyemem
inundatis frequens." Evidently the plant now considered.

Veronica uliginosa Raf. in Am. Mo. Mag. 2: 175. 1818. "Ver-
onica scutellata Pursh . . . . Fl. Am. Sept. 1: 11." In his
Fl. Am. Sept. 11. 1814, Pursh states of "Veronica scutellata" that
"the American plant has longer leaves than any of the European
specimens I have seen," a condition not verified by the material at
hand to-day though Pursh's statement evidently misled Rafinesque
into assuming for it specific distinctness.

Veronica connata Raf., Med. Fl. 2: 110. 1830. "In west Ken-
tucky. " Very briefly characterized and leaves said to be " connate; "
surely they were merely cordate-clasping and opposite.

Meadows and swales, Newfoundland and Yukon to Virginia,
Indiana, Wyoming, and California.

Occasionally occurs in a form more or less pubescent throughout,
forma villosa (Schumacher) Pennell [Veronica scutellata villosa Schu-
macher, Enum. PI. Saell. 1: 7. 1801; also V. scutellata pilosa Vahl,
Enum. PI. 1: 70. 1805; V. scutellata pubescens Koch, Syn. Fl.
Germ, et Helv. 524. 1837.]. This occurs sporadically occasional
throughout the range of the species.

III. HEBE Commerson

Hebe Comerson; [Juss., Gen. PI. 105. 1789, generic diagnosis
only;] J. F. Gmelin, Syst. Nat. 2: 27. 1791. Type species, Hebe
magellanica J. F. Gmel.

Leaves lanceolate, acuminate to a small blunt tip, entire, not
revolute, 6-7 cm. long. Racemes 5-13 cm. long, many-
flowered. Sepals acuminate, 2-3 mm. long. Corolla
with the oblong lobes shorter than or but little longer
than the narrow tube. Stamens and style longer than
the corolla and conspicuously exserted, the latter
slender, 5-6 mm. long. Capsule oval in outline, rela-
tively thin-walled. Stem minutely pubescent when
young, especially between and proximad to bases of
leaves, becoming glabrate; bark slightly wrinkled in
drying.

1921] Pennell, — " Veronica " in North and South America 39

Racemes 13 cm. long; racbis, pedicels and lanceolate sepals
finely pubescent. Corolla 5 mm. long, its lobes
slightly shorter than the tube. Leaves attenuate-
acuminate. Internodes on flowering shoots about 3

cm. long 1. H. salicifolia

Racemes 5-7 cm. long; rachis, pedicels and lance-ovate
sepals puberulent. Corolla not seen. Leaves narrow-
ing to a blunt tip. Internodes on flowering shoots

less than 1 cm. long 2. H. blanda

Leaves elliptic-oval, apiculate, the margin revolute, callose,
and at times obscurely crenate, 2-3 cm. long. Racemes
2 cm. long, few-flowered. Sepals acute to obtuse, 4 mm.
long. Corolla 8 mm. long, the broadly ovate lobes much
longer than the broad tube. Stamens not longer than
the corolla, the stout style 4 mm. long. Capsule elliptic-
oval in outline, thick-walled. Stem densely and per-
sistently pubescent with pale hairs on side between and
proximad to bases of leaves, below leaf-bases reddish,
glabrous and shining; bark much wrinkled in drying. . . 3. H. elliptica

1. Hebe salicifolia (Forst.) Pennell, comb. nov.

Veronica salicifolia Forst., Fl. Ins. Austr. Prod. 3. 1786. " [Noua
Zeelandia, G. Forster]." Several specimens from New Zealand seen,
and one collected by A. H. Cockayne 8041, and labeled " Veronica
salicifolia Forst. Typical South Island form," shows precisely the
slender finely pubescent pedicels, small flowers, and acuminate,
almost attenuate leaves of our plant. Type species of genus Panoxis
Raf., Med. Fl. 2: 109. 1830.

Veronica Fonkii Phil, in Linnaea 29: 110. 1857-8. "En las
playas y barrancas de Chonos, ' in litore et valleculis, legit . . .
Dr. Fr. Fonk." Specimen in Herb. Columbia University, labeled
" Veronica Fonki Ph. Chonos, legit Philippi, com. am Treviranus
1864," is doubtless an isotype. This seems to be the same as the
plant of New Zealand.

Chonos, Chile. Also in South Island, New Zealand.

2. Hebe blanda (Cheesem.) Pennell, comb. nov.

Veronica amabilis blanda Cheesem., Man. New Zealand Fl. 506.
1906. "Port Chalmers [Otago, South Island, New Zealand] Petrie!

" Specimen in Herb. New York Botanical Garden,

collected at Anita Bay, Otago (where it forms "a considerable part
of the 'coastal Scrub'") appears to be exactly our plant, and to agree
with Cheeseman's variety.

Southern Patagonia. Also in South Island, New Zealand.

3. Hebe elliptica (Forst.) Pennell, comb. nov.

Veronica elliptica Forst., Fl. Ins. Austr. Prod. 3. 1786. "[Noua
Zeelandia, G. Forster]," Several specimens from New Zealand
seen, two from Port Otway and Tuesday Bay respectively, agreeing
exactly with our plant. Also a specimen from the Auckland Islands,
Wilkes Expedition, is quite the same.

40 Rhodora [February

Veronica decussata [Soland. in] Ait., Hort. Kew. 1 : 20, 1789. " Nat.
of Falkland Islands. Introd. 1776, by John Fothergill." Described
as with bracteoles on pedicels, an appearance probably caused by
the terminal bractlets of the raceme appearing, while the bud of the
rachis is suppressed.

Hebe magellanica J. F. Gmel., Syst. Nat. 2: 27. 1791. Based
upon Hebe Juss., Gen. PI. 105. 1789, where the name is attributed
to Commerson and the plant said to be from Magellan. Evidently
collected by Commerson at the Straits of Magellan in 1767-8.

Veronica Simpsonii Phil, in Anal. Univ. Chile 1873: 26. 1873.
" Enrique Simpson trajo de las orillas del rio Aysen, en Patagonia. "
The careful description of the branch, leaves, fruiting inflorescence,
capsules and seeds appears to denote the species now considered.

Southern Patagonia and Falkland Islands. Also in the Auckland
Islands and South Island of New Zealand.

Nomina Excludenda.

Veronica caroliniana Poir., Encyc. Meth., Bot. 8: 520. 1808.
"Communiquee par M. Bosc, qui l'a recueillie dans la Caroline."
This is Cynoctonum Mitreola (L.) Britton, of the Loganiaceae. Not
V. caroliniana Walt., 1788.

Veronica cinerea Raf ., New Fl. Am. 4 : 39. 1838. " From Origon. "
Description of plant as " cinereous villose, leaves alternate ....
flowers spicate very dense sessile .... stamens very long"
appears to denote some species of Synthyris.

Veronica fluminensis Veil., Fl. Flum. 11. 1825; Icones 1: pi. 25.
1827. "Abunde provenit locis umbrosis ad vias maritimas Regii
Praedii Sanctae Crucis [Brazil]." Description and illustration show
a plant of the Acanthaceae.

Veronica litoralis Veil., Fl. Flum. 10. 1825; Icones 1: pi. 24-
1827. " Silvis maritimis Regii Praedii Sanctae Crucis [Brazil] prope
litus, ad loca arenosa habitat." Description and illustration show a
plant of the Acanthaceae.

Veronica marilandica L., Sp. PI. 14. 1753. "Habitat in Vir-
ginia. " According to B. D. Jackson (in Proc. Linn. Soc. 14. Suppl. :
150. 1912), Linne transferred his specimen bearing this name to
Polypremum procumbens L. Both his description and that in Gro-
novius' Fl. Virg. 4. 1739, indicate this plant of the Loganiaceae.
I cannot locate the reference which is erroneously cited as : " Veronica
marilandica Murr. Comm. Gotting. 11: t. 3. 1782."

Veronica missurica Raf. in Am. Monthly Mag. 3: 175. 1818.
New name for Veronica reniformis Pursh, which was a species of
Synthyris. See below.

Veronica Purshii G. Don, Gen. Hist. Dichl. PI. 4: 573. 1838.
" Native on the banks of the Missouri. V. reniformis Pursh . . .,
but not of Rafin. " A species of Synthyris. See below.

1921] Pennal, — "Veronica" in North and South America 41

Veronica reniformis Pursh, Fl. Am. Sept. 1 : 10. 1814. " Collected
by Messrs. Lewis and Clark in boggy soil, on the banks of the Miss-
ouri . . . . v. s. in Herb. Lewis." Type was apparently a
plant collected on Hungry Creek, in what is now Montana, June 26,
1806, and an isotype of this in the Herbarium of the Academy of
Natural Sciences of Philadelphia was determined by Robinson and
Greenman [in Proc. Acad. Nat. Sci. Phila. 1898: 39. 1898] as
Synthyris reniformis major Hook. Pursh's description is inaccurate,
but I think must certainly apply to this collection which is the
species, S. major (Hook.) Heller.

Veronica rotundifolia Ruiz & Pa von, Fl. Peruv. et Chil. 1 : 6. 1798.
"Habitat copiose in Peruviae uliginosis ad Pillao vicum." This is a
species of Sibthorpia.

Veronica sparsiflora Raf., Atl. Jour. 79. 1832. Described from
a plant in the Bartram Botanical Garden, Philadelphia, Pa., which
was said to have been "native of Arkansas or Texas, received from
Prof. Nuttall." I know of no American species at all fitting this
description: "stem erect, simple round solid, leaves opposite sessile
cuneate oblong entire obtuse. Raceme terminal lax very long,
flowers scattered, bracts linear oblong obtuse, pedicels filiform.

Capsules bilobed subcompressed. Annual Stem 1 or

2 feet high. Flowers vernal purpurescent handsome. Corolla

rotate, segments of the calix unequal oblong, obtuse "

Is it a foreign species, or not a Veronica?

New York Botanical Garden.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 231

TYPES OF FLOWERS AND INTERSEXES

IN GRAPES WITH REFERENCE TO

FRUIT DEVELOPMENT

By A. B. STOUT

NEW YORK
1921

Technical Bulletin No. 82, New York Agricultural Experiment Station, Geneva, N. Y.

TECHNICAL BULLETIN No. 82

JANUARY, 1921

m

[ark % grkuliitral f xpmuttni Iftate.

GENEVA, N. Y.

TYPES OF FLOWERS AND INTERSEXES IN GRAPES WITH
REFERENCE TO FRUIT DEVELOPMENT

A. B. STOUT

LIBRARY

NEW YORK

BOTANICAL

(JARUfifll

PUBLISHED BY THE STATION

BOARD OF CONTROL

Governor Nathan L. Miller, Albany.
Commissioner George E. Hogue, Albany.
Irving Rouse, Rochester.
Frank M. Bradley, Barker.
Charles C. Sackett, Canandaigua.

Charles R. Mellen, Geneva.
John B. Mulford, Lodi.
C. Fred Boshart, Lowville.
Peter G. Ten Eyck, Albany.

OFFICERS OF THE BOARD

Charles R. Mellen,

President.

STATION STAFF

William O'Hanlon,

Secretary and Treasurer.

Whitman H. Jordan, Sc.D., LL.D., Director.

fALBERT R. Mann, A.M.,

Agricultural Economist.
George W. Churchill, Agriculturist.
Reginald C. Collison, M.S.,

Chief in Research (Agronomy).
"j"T. Ltttleton Lyon, Ph.D.,

Chemist (Agronomy).
James E. Mensching, M.S.,

Associate in Research (Agronomy).
James D. Harlan, B.S.,

Assistant in Research (Agronomy).
William P. Wheeler,

Chief in Research (Animal Industry).
Robert S. Breed, Ph.D.,

Chief in Research (Bacteriology).
Harold J. Conn, Ph.D.,

Chief in Research (Soil Bacteriology).
fWiLLiAM A. Stocking, Jr., M.S.A.,

Bacteriologist.
George J. Hucker, M.A.,

Assistant in Research (Bacteriology).
Fred C. Stewart, M.S.,

Chief in Research (Botany).
fDoNALD Reddick, Ph.D., Botanist.
Walter O. Gloyer, M.A.,

Associate in Research (Botany).
Mancel T. Munn, M.S.,

Associate- Botanist.
Elizabeth F. Hopkins, A.B.,

Assistant Bo'.anist.
Lucius L. Van Slyke, Ph.D.,

Chief in Research (Chemistry).
Richard F. Keeler, M.S.,

Associate in Research (Chemistry).
Rudolph J. Anderson, Ph.D.,

Chief in Research (Biochemistry).
fLEONARD A. Maynard, Ph.D.,

Biochemist.
Walter L. Kulp, M.S.,

Assistant in Research (Biochemistry).
Arthur W. Clark, B.S.,

Associate Chemist.
Morgan P. Sweeney, A.M.,
William F. Walsh, B.S.,

Assistant Chemists.

Millard G. Moore, B.S.,
Nathan F. True, A.B.,

Assistant Chemists.
George A. Smith,

Chief in Research (Dairying).
James D. Luckett, M.S.A.,

Editor and Librarian.
Laura G. Collison, A.B.,

Assistant Editor and Librarian.
Percival J. Parrott, M.A.,

Chief in Research (Entomology).
fGLENN W. Herrick, B.S.A.,

Entomologist.
Hugh Glasgow, Ph.D.,
*Fred Z. Hartzell, M.A., (Fredonia),
Associates in Research (Entomology).
Clarence R. Pmprs, B.S.,
Guy F. MacLeod, B.S.,

Assistants in Research (Entomology).
Ulysses P. Hedrick, Sc.D.,

Chief in Research (Horticulture).
fRoLLiNs A. Emerson, Sc.D.,

Geneticist.
TWilliam H. Chandler, Ph.D.,

Pomologist.
*Fred E. Gladwin, B.S. (Fredonia),
Orrin M. Taylor,
George H. Howe, B.S.A.,
Richard Wellington, M.S.,

Associates in Research (Horticulture).
Theodore E. Gaty, B.S.,
Thomas O. Sprague, B.S.,
Harold B. Tukey, M.S.,

Assistants in Research (Horticulture).
James S. Lawson, Phm.B.,

Museum Preparator.
Jessie A. Sperry, Director's Secretary.
Frank E. Newton,

WlLLARD F. PATCHIN,

Lena G. Curtis,
Mae M. Melvin,
Maude L. Hogan,

K. LORAINE HORTON,

Clerks and Stenographers.
Elizabeth Jones

Computer and Mailing Clerk.

Address all correspondence, not to individual members of the .-taff, but to the
New York Agricultural Experiment Station, Geneva, N. Y.

The Bulletins published by the Station will be sent free to any farmer applying
for them.

♦Connected with Grape Culture Investigations.

tMembera of the faculty of the New York State College of Agriculture alEliated with this
Station.

Technical Bulletin No. 82

TYPES OF FLOWERS AND INTERSEXES IN

GRAPES WITH REFERENCE TO FRUIT

DEVELOPMENT*

A. B. STOUT f

INTRODUCTION

The types of flowers in grapes have long aroused interest among
both the growers and the breeders of grapes who are concerned not
only with the quality and the productiveness of seeded varieties,
but with the development of the seedless or partially seedless varieties.
Breeders of grapes are also finding it desirable, if not necessary, to
determine as fully as possible how the development of the desirable
distribution of the sexes in the flowers may be regulated, controlled,
or influenced by breeding and by selection of parentage.

It is well recognized that the production of fruit in any variety
and in any seedling plant depends, first of all, on the development of
stamens and pistils and their power to function as sex organs in
fertilization and in fruit and seed formation. This is well illustrated
by the flower types of the best varieties of grapes now in general
cultivation.

Self-fruitful varieties like the Delaware, Niagara, Winchell, and
Concord have flowers in which both stamens and pistil are well
developed and highly functional inter se. The stamens are en ct
or upright (Figs. 13-21) and their anthers contain pollen much of
which is viable and capable of functioning in fertilization. Both
kinds of sex organs are present and functional inter se in the same
flower. The plants are perfect hermaphrodites. They set fruit to
self-pollination.

Other varieties like Brighton, Lindley, Barry, and Massasoit
have flowers whose stamens are recurved and more or less aborted
and the pollen which they contain is much shrivelled and defective,
and is entirely or almost entirely incapable of functioning in fertili-

* This is the first report of work done in cooperation with the New York Botanical
Garden.

t Director of the Laboratories, New York Botanical Garden.

3

zation (Figs. 32-39). Such varieties are slightly or even not at all
fruitful if left to themselves for self-pollination. The pistils, however,
are well formed and are capable of forming fruits containing seeds
if properly cross-pollinated with good pollen. Since the stamens
are more or less impotent these plants are called imperfect herma-
phrodites. Such varieties have been called " self -sterile " and the
pollen has been spoken of as " impotent " or " lacking in affinity "
with pistils. There is, however, very general agreement among
those who have investigated the condition of the pollen (Beach
1902, Booth 1902, Dorsey 1914, and Detjen 1917)1 that the sterility
is due to the poor condition of the pollen which is thus unable to function
at all.

The so-called " self-sterility " in these varieties of grapes is, there-
fore, very different from the " self-sterility " of certain varieties of
plums, apples, pears, and cherries, in which the flowers are perfect
but the viable pollen is capable of functioning only in certain cross-
fertilizations with other varieties or other individuals. In the
grape the sterility, it appears, is due solely to imperfect development
of stamens and pollen. In the plum the pollen is physiologically
incompatible with the female elements of the same plant, of plants
of the same clonal variety, and also of certain other plants of diverse
origin. In progenies grown from the seed of plants having this type
of sterility (Stout 1918. 1920) all degrees of self- and cross-compati-
bilities and incompatibilities are often in evidence.

Studies made of grapes indicate that the imperfections in the
stamens very generally, but not always, involve a recurving of
filaments which gives a very different flower from that of the perfect
hermaphrodite with its erect stamens.

Among all wild species of American grapes and in seedlings of
cultivated sorts another type of flower, the purely staminate, has long
been recognized. In this the pistil is much aborted and entirely
functionless or is even lacking, but the stamens are well developed
and yield much viable pollen. A plant with such flowers is highly
fertile as a male parent in cross-pollination but is itself fruitless.
It appears that wild species of American grapes consist only of stami-
nate plants and imperfect hermaphrodites except perhaps for rather
rare cases when individuals are found bearing some or even all
perfect flowers.

1 See Literature Cited, p. 15.

Still another type of flower is seen in the grapes that produce
seedless or nearly seedless fruits. Here the pistils are capable of
developing into fruits but the ovules, if present, do not develop into
seeds. According to Muller-Thurgau (1898), the fruits of the
seedless raisins develop under the stimulus of pollen-tube growth.
Pollination causes development of fruit but does not result in seed
formation. Such varieties as Sultana and Sultanina, therefore, are
fully seedless when grown under ample opportunity for both self-
and cross-pollination. This seems clearly to indicate that in the
fully seedless varieties no ovules capable of developing to full
maturity are present. Nothing definite seems to be known regarding
the exact origin of these seedless varieties and no seedless grape
has yet become of commercial value in the eastern United States.
The types of flowers which they possess have, however, been observed
among seedlings and are to be considered in judging the value of the
various flower types appearing in seedlings.

It seems certain that the flowers borne by the greater number of
seedlings obtained in breeding grapes can be classed broadly as (1)
staminate, (2) perfect hermaphrodite, and (3) imperfect herma-
phrodite. No purely female types are known. For describing the
general heredity of sex in grapes, such a grouping is useful (Hedrick
and Anthony 1915, Detjen 1917). It has, however, been recognized
that a sharp distinction between these types does not exist. As
Booth (1902) remarks " the whole path is marked by transitional
forms; thus there are no distinct classes of self-sterile and self-fertile
grapes, but all gradations exist from one extreme to the other."
It is in the study and selection of these intergrading intersexual forms
that the possibility of finding types with the degree or kind of femaleness
which exists in the seedless types seems most promising.

In continuing the studies of the inheritance of sex in grapes in
progress at the New York Agricultural Experiment Station, and
especially with reference to the production of the flower types that
give seedless fruits, it has seemed advisable to describe in more
detail the variations in the development of the flowers. A general
survey has revealed at least one new type of flower here designated
as having crinkled stamens, a wide range of variations in the length
of stamens among flowers classed as perfect hermaphrodites, and
various intermediates between the typically upright and the reflexed

8

one chews the pulp. The fruits produced in 1920 ranged from
7 to 18 thirty-seconds of an inch in diameter. They are of good size
for seedless sorts, but smaller than the fruits of the Concord, and
as shown in Plate III the bunches are well filled.

A typical flower from one of these plants is shown in Fig. 6. The
stamens are unusually long, the anthers are well developed, and at
least 95 per cent of the pollen germinated in each of two tests on the
three strengths of sugar-agar media used.

These plants bearing seedless grapes arc strongly male, but their
seedlessness is, it seems to me, due to what wc may properly call a
weak grade of femaleness.

Figure 7 shows a flower of a plant bearing near-seedless or partially
seedless fruits. The stamens are erect and well developed tho
only of medium length. In tests for germination about 50 per cent
of the pollen produced tubes. As is the case in Fig. 6, pistils of good
size are present and from the appearance of the flower alone the plant
would be considered as a perfect hermaphrodite. However, of the
fruits examined, 27 were entirely seedless and ranged from 9 to 13
thirty-seconds of an inch in diameter; 8 fruits contained one soft
seed each and ranged from 12 to 16 thirty-seconds of an inch in
diameter; 52 fruits contained one well-developed seed each and
were from 12 to 18 thirty-seconds of an inch in diameter; while only
5 fruits contained as many as two seeds each and these were 14, 15,
16, 17, and 18 thirty-seconds of an inch in diameter, respectively.
The fruits are hence low in number of seeds, many are seedless,
and none of the 1920 crop had more than two seeds. The fruits were
of good quality and the bunches were uniformly well filled.
(See Plate IV.)

This plant is a seedling from Triumph crossed with Delaware
both of which are seeded grapes. A sister seedling bore fruits
ranging from 15 to 19 thirty-seconds of an inch in diameter but none
were seedless, altho no fruit contained more than two seeds.

A seedling derived from the cross Triumph x Dutchess has pro-
duced fruits ranging from 10 to 20 thirty-seconds of an inch in
diameter. Many of the smaller fruits are seedless and of those
examined none had more than two seeds.

All of these plants bearing seedless or near-seedless fruits have
flowers with erect stamens with filaments of medium or long lengths
and pistils of medium size. Descriptive records of the Station for

EXPLANATION OF PLATES

The flowers here illustrated are shown about four times their
natural size. The original drawings for these were made to the
same scale but there were some differences in the reduction for
the various plates. Fig. 10 is enlarged about 110 times and Figs.
25, 37, and 39 about 350 times. The photos of Plates III and IV
were taken natural size and are here reproduced as taken.

Plate I

All figures are of staminate types of fruitless plants. Rudiments
of pistils are present in Figs. 3, 4, and 5, but there is no stigmatic
surface in Fig. 3. All are from seedlings: Fig. 1 of a plant of
unknown parentage, Fig. 2 of America selfed, Fig. 3 of the cross
Vitis Berlandieri x V. riparia, Fig. 4 of the cross Solonis x Riparia
No. 616, and Fig. 5 of Croton selfed.

Plate I

Plate II

Fig. 6. Typical flower of plant producing seedless fruits, a seed-
ling of Concord Seedless.

Fig. 7. Flower of near-seedless plant, a seedling of Triumph
x Delaware.

Figs. 8 and 9. Two flowers one with and one without pistil, borne
in same cluster on a seedling of Solonis x Riparia. Fig. 10. Shows
few potent and viable pollen among many shriveled grains from
these flowers (x 110).

Figs. 11 and 12. Two flowers, both highly staminate: one with
rudimentary pistil, other with pistil functional in fruit and seed
production. From a seedling of Dutchess x Seibel No. 2.

"a

£3

8 & ® 1**

10

® 0 Q

0ow

11

Plate III. — Cluster of Seedless Fruits borne by Seedling of the Concord
Seedless. Flower of this Plant shown in Fig. 6

Plate IV. — Cluster of Fruits, Many Seedless and none with more than two
Seeds, borne by a Seedling of Triumph x Delaware. Flower shown in
Fig. 7.

Plate V

All perfect hermaphrodites with erect stamens. Highly functional
both as males and females.

Fig. 13 of Wilder; Fig. 14 of Cottage; Fig. 15 of Worden; Fig. 16
of Concord; Fig. 17 of Seedling of Hexamer (selfed); Fig. 18 of Mus-
cat Hamburg; Fig. 19 of Blauer Portugieser; Fig. 20 of Kensington;
and Fig. 21 of seedling of Lindley x Niagara.

13

14

15

16

17

18

19

Plate V

Plate VI

Figs. 22 and 23 are flowers with crinkled stamens; Fig. 24 is a
stamen 20 times natural size; and Fig. 25 the shrivelled pollen
magnified about 350 times. For ancestry of the plants bearing these
see page 11.

Fig. 26. Flower witfi crinkled stamens from a second generation
seedling of Salem x Worden.

Fig. 27. Flower with spreading stamens from seedling of Janes-
ville selfed.

Fig. 28. Flower with strongly spreading stamens from seedling
of Janesville selfed.

Fig. 29. Showing slightly recurving stamens of seedling of Clinton
x Diana.

22

24

25

26

27

28

29

Plate VI

Plate VII

Fig. 30. Recurving or semi-reflexed stamens of seedling of Janes-
ville selfed.

Fig. 31 flower just before, and Fig. 32 flower just after calyx is
shed showing short stamens that become reflexed. (From Barry.)

Fig. 33. Flower of seedling of Noah selfed.

Fig. 34 flower of Aminia, Fig. 35 of Brighton, Fig. 36 of Mas-
sasoit, and Fig. 38 of Merrimac, all with strongly reflexed stamens.

Fig. 37 pollen of Massasoit, and Fig. 39 of Merrimac enlarged 350
times.

30

31

32

33

34

35

36

Plate VII

flowers of such sorts as Hubbard Seedless and the Sultanina (the
Thompson Seedless, a raisin grape widely cultivated on the Pacific
Coast) indicate that these also have flowers of this general type. In
respect to maleness the sex of these plants is highly developed, but
their seedlessness shows that femaleness is weak and but partially
developed. In the case of plants that produce seedless and seeded
fruits in the same cluster there is indication that femaleness is
irregular in its development and may appear in various grades among
flowers of a single cluster altho such a mixing may not be apparent
in the general appearance of the flowers, and not especially con-
spicuous in the general appearance of the clusters of fruit. (See
Plate IV.)

Noll (1902) has called fruits which are produced without polli-
nation parthenocarpic. The available evidence indicates that this is
a relatively rare phenomenon. Such types should probably be
classed quite differently than those in which pollination is required
for any setting of fruit. They may perhaps be regarded as more
strongly female or at least as strongly fruitful. There may be some
question as to whether parthenocarpy is a phenomenon of vegetative
or of reproductive vigor. However, the numerous grades of seed-
lessness which are found show, it seems to me, that when pollination
is required seedlessness is to be regarded as a stage in the series of
intersexes.

PLANTS WITH TWO OR MORE FLOWER TYPES INTERMIXED

In some plants it is plainly evident from an examination of the
flowers that such extreme types as the purely staminate and what
resembles the perfect hermaphrodite are intermingled in the same
cluster. The latter may,, however, give no fruit, seedless fruits,
seeded fruits, or various mixtures of these. Figures 8 with 9 and
11 with 12, show this condition.

The plant bearing the flowers shown in Figs. 8 and 9 produced
no fruit in 1920 which suggests that the pistils that are formed are
functionless. The erect stamens are of good size and the anthers
dehisce normally but only about 5 per cent of the pollen germinated.
The grains that did not germinate were with few exceptions shriveled
and aborted (Fig. 10).

The plant bearing the flowers shown in Figs. 11 and 12 produced
a " light " crop in 1920, the clusters were decidedly " loose," the

10

berries irregular in size, and some of them wove nearly seen'
Such crops are characteristic of many seedlings and also of some
of the commercial varieties. A minding of flowers in which there
are various grades of development in the pistils is at. least one of
the causes of such incomplete fruiting. Purely staminate flowers
and flowers with funetionless pistils are fruitless, and other grades
are seedless, near seedless, or strongly seeded. Irregular fruiting
from year to year may. in some cases, be due in a large degree to
variation in the proportions of the various types of flowers present.

PERFECT HERMAPHRODITES WITH UPRIGHT STAMENS

Marked differences in length of filaments, sue of anthers, and -
and shape of pistils are to be seen among flowers of this general
class. The flowers: illustrated in Figs. 13 to 20 are from well-known
varieties. All of these sorts are highly productive of fruits that
contain seeds. Ahho the number I - jeds characteristic of the
variety may differ somewhat, all 8 - female and the fruit-

fulness depends on fertih and on the development of seeds.

As shown in the figures, the stamens of a variety may be so short
that they scarcely reach above the stigma Kensington. Fig. 20) and
in another variety they may be at least twice that length Cottage.
Fig. 14'. All grades between the extremes of length are to be seen
among the varieties and the seedlings usually class I -ether as
having upright stamens. In all the well-known cultivated varieties
of this class. Europe: - D as American, so far as tested by
the writer, the pollen is remarkably viable with at least To per cent
of all grains germinating in tests. In many seedlings, however, the
pollen is mostly shrivelled and not viable in the germination tests,
and the presence and general vigor of erect stamens is not a sure
indication that the stamens are functional.

IMPERFECT HERMAPHRODITE WITH ERECT CRINKLED STAMENS

In these the stamen^ - ing but the rilanients are variously

-:ed or crinkled in irregular and incomplete spirals. See F:gs.
22 to 26 The anthers appear on general examination, to be
perfectly formed and to dehisce normally Fig. 24 . Repeated
- sts for germination of the pollen were made for several plants
bearing these flowers but no germination was obtained. So fa
observed all pollen grains are irregularly shrivelled ^Fig. 25).

11

The crinkled-stamen type of flower was found in each one of 34
hui the length of filaments and the <\<^v<'M of crinkling
differed somewhat as La shown by comparing Figs. 22 and 23. In
1920 all but 2 of the 34 plants produced seeded fruit and of tb
bore a " full " crop, 9 a " medium " crop, 9a" light " crop, and 7
a "very light" crop. With the exception of two plant-, these
seedlings were able to function as females in fruit and seed pro-
duction.

The ancestry of this series of 34 plants is well known for three
generations. First, Winchell was crossed with Diamond (both
classed as hermaphrodites with upright stamens), and Station
Seedling No. 123 with upright stamens was obtained. This seedling
was used as a pollen parent on Worden, also with upright stamens,
and three seedlings, all with upright stamens, Xos. 931, 033, and 934,
were obtained. The 34 plants derived from selling seedling No. 933
were all of the type with crinkled stamens. Of the progeny obtained
by selfing the sister plants, 931 and 934, 24 have bloomed and
have only the normal erect stamens of the perfect hermaphrodite.
As to length, the stamens of 1 were; short, of 20 medium, and of 3
long.

It appears that the plants with this crinkled type of stamen are
not able to function as males. Functionally their sex is quite like
that of the imperfect hermaphrodites with reflexed stamens. One
might consider that in these there is a combination of the " reflexed "
character of filament with the " upright." The appearance of this
type of flower in all of 34 sister plants after three generations of
parentage with normal upright stamens suggests the phenomenon
of the so-called mass mutation.

Another plant was found bearing flowers with crinkled but less
erect stamens (Fig. 26). In this case the filaments may be con-
sidered as tending to be recurved. This plant is a seedling derived
from selfing Station Seedling No. 4574 (having normal upright
stamens), which was an offspring of Salem (reflexed) crossed with
Worsen (upright). In this case the reflexed character was pre
in one grandparent.

Whether the crinkled character of stamen is to be considered as
due to a recombination of hereditary units, as an intermediate

12

between reflexed and upright, or as a mutation, it is a new type
of flower to be recognized in classifying the flowers of grapes.

FLOWER TYPES WITH SPREADING OR SEMI-REFLEXED STAMENS

Frequently among the seedlings there are plants with flowers
whose stamens have nearly straight or slightly recurved filaments
which at the time of complete anthesis stand at a decided angle
from the axis of the pistil. In general these may be considered as
intermediate between erect and recurved. Flowers of this sort
from four plants are shown in Figs. 27 to 30. In these the anthers
are well formed and dehisce quite normally, but the pollen was of
irregular sizes, more or less shrivelled, and not viable in germination
tests. These plants are able, therefore, to function only as females.

FLOWER TYPES WITH RECURVED STAMENS

The general character of this sort of flower is shown in Figs. 31
to 38, all but one of which are from well-known varieties long
recognized as fully or very decidedly self-fruitless. The stamens
may be strongly but simply recurved as seen in the flower of Barry
(Fig. 32), extremely recurved and curled as in Massasoit (Fig. 36),
or of various gradations between such conditions. The filaments
are sometimes very slender. The anthers are in various grades of
development; and in Barry (Figs. 31 and 32) thay are small in size,
many appear to contain no pollen, and some do not dehisce. In other
varieties the anthers are uniformly larger and dehisce well and a
part of the pollen appears to be normal (Fig. 37).

The writer made extensive tests for germination of pollen both
in varieties and in seedlings having reflexed stamens. In many
cases all of the grains were irregularly shrivelled and devoid of granu-
lar contents. In other cases some of the grains became spherical
on the sugar-agar media and appeared to have granular contents
(Fig. 37), but in no case did any pollen from recurved stamens
germinate. It should be noted, however, that seeds have evidently
been obtained at the Geneva Station from selfing and crossing several
of these varieties. Possibly a few of the grains are capable of germi-
nating on pistils but incapable of germinating on media that give
excellent germination of the pollen from staminate types and from
hermaphrodites with normal upright stamens.

13

DISCUSSION AND CONCLUSION

A summary of the variations in the flowers of cultivated varieties
of grapes and of seedlings from them shows that there are many
grades in the relative development of pistils and stamens.

Complete loss of femaleness is seen in only a few cases, but various
grades of rudimentary pistils are present that are incapable of yield-
ing fruits of any kind. This evidence suggests that it is a weak
grade of femaleness which is responsible for the development of
seedless fruits. In near-seedless sorts, femaleness is of still another
grade but below that of high seed production.

A complete morphological loss of stamens has not been found. The
least male plants are perhaps among those with recurved stamens,
small sterile anthers, and only aborted pollen. Such plants are
able to function as females only. Plants with erect crinkled stamens
and some plants with normal upright filaments are likewise unable
to function as males because of impotent pollen. In the extremes
there is complete loss of one or the other sex, at least functionally.
Between these extremes there are various grades of relative develop-
ment and functioning ability of pistils and stamens. In the per-
fect hermaphrodites both of these organs are highly developed and
functional, and it is these types that include all of the most important
and productive of the seeded grapes.

It appears that, in individual plants, the stamens are more uni-
form than the pistils; at least they appear to be rather uniform
in respect to length of filaments and general character of anthers.
The intermixing of flowers of several types in the same cluster seems
to involve especially variations in the condition of the pistils. The
most obvious of these are mixtures of staminate flowers and of
perfect flowers that are capable of development into fruits with
seeds. The type of pistil that produces seedless fruits may also be
present.

Variations in the character of pistils in flowers from year to year
have been observed. (See especially Detjen, 1917 for plant named
" Hope.") There is some evidence that the fruit production of the
self-fruitless varieties can be influenced by girdling and bending
(Beach, 1902), but whether such treatment changes the character
of the stamens by rendering them more potent as the results would
perhaps suggest is not determined.

14

The variations in the morphological development of stamens and
pistils and in their ability to function as sex organs are to be described
as phenomena of intersexualism. The abortion of pollen in reflexed,
in crinkled, and in other stamens is seen in flowers that are highly
pistillate and functional as females. Conversely extreme loss of
femaleness is seen in staminate types in which maleness is highly
developed. In these extremes the loss of sex is one-sided. Such
one-sided impotence is characteristic of intersexualism as distinct
from the sterility of hybridity which tends to affect both sexes in
the individual quite the same.

Cases of sterility in grapes due to hybridity are reported by Detjen
(1919) in Fi hybrids between Vitis vinifera (Malaga) and V. rotundi-
folia in which flowers of the external form of the perfect hermaphro-
dites are mostly or entirely sterile and impotent in both stamens
and pistils. Such sterility appears to be largely absent in hybrids
between V. vinifera and the more northern species of native American
grapes. Many varieties known to have originated thus have per-
fect flowers with both sets of sex organs highly functional. A sys-
tematic study of known hybrids between V. vinifera and native
species from which such a variety as Concord originated is needed
to determine to what extent, if at all, sterility from hybridity is
present along with intersexualism.

The results of the breeding work already obtained at the Station
indicate clearly that the use of seedless and near-seedless plants
as male parents in crosses with varieties that are strongly female
(perfect and imperfect hermaphrodites) gives progeny that are
strongly female and seed producing. The Fi offspring of many
crosses of standard seeded varieties with Hubbard Seedless have
all been strongly pistillate, yielding seeded fruit. Weak femaleness
(seen in seedless fruits) is in this case dominated or swamped by the
strong femaleness of the seed parent. However, the seeded char-
acter of Fi individuals is no index of the variation in intersexes
that may appear in later generations in which the segregation of
at least some plants bearing seedless fruits may be expected. The
use of other seedless sorts in such crosses may, however, give different
results.

The most effective course in breeding for the development of seed-
less sorts is suggested by the conditions of intersexualism. Most
individuals and varieties producing seedless or near-seedless fruits

15

are strongly staminate. The former can be used as male parents
on the latter, which do produce a few viable seeds. Plants strongly-
male and seedless can be crossed with plants strongly male but
weakly female and near-seedless and, also, the self-fertilized progeny
of the latter may be obtained. In this way families weak in female-
ness may undoubtedly be obtained in which a considerable number
of individuals will produce seedless fruits.

LITERATURE CITED

Beach, S. A.

1902 Potency of the pollen of self-sterile grapes. N. Y.
Agr. Exp. Sta. Bui. No. 228.

Booth, N. O.

1902 A study of grape pollen. N. Y. Agr. Exp. Sta. Bui.
No. 224.

Detjen, L. R.

1917 Inheritance of sex in Vitis rotundifolia. N. Car. Agr
Exp. Sta. Tech. Bui. No. 12.

1919 Some Fi Hybrids of Vitis rotundifolia with related
species and genera. N. Car. Agr. Exp. Sta. Tech.
Bui. No. 18.

Dorsey, M. J.

1912 Variation in the floral structures of Vitis. Bui. Torrey
Bot. Club 39, 37~o 2.

1914 Pollen development in the grapo with special reference
to sterility. Minn. Agr. Exp. Sta. Bid. No. 144.

Engleman, G.

1883 Bushberg Catalogue. 3rd Edition.

Hedrick, U. P. and Anthony, R. D.

1915 Inheritance of certain characters of grapes. Jour.
Agr. Research 4,315-330.

Miiller-Thurgau, Herman.

1898 Abhangigkeit der Ausbildung der Traubenbeeren und
einiger anderer Friichte von der Entwicklung der Samen.
Landto. Jahrb. der Schweiz 1898. Abs. in Bot. Centr
77, 135-138.

16

Noll, F.

1902 Fruchtbildung ohne vorausgegangene Bestaubung (Par-
thenocarpie) bei der Gurke. Sitzungsberich. der Nieder-
rhein. gesellsch. f. Natur-u. Heilkunde zu Bonn 1902, 1-13.

Stout, A. B.

1918 Fertility in Cichorium Intybus: Self-compatibility and

self-incompatibility among the offspring of self-fertile

lines of descent. Jour. Genetics 7, 71-108.

1920 Further experimental studies on self-incompatibility in
hermaphrodite plants. Jour. Genetics 9,85-129.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 232

NOTES ON ROSACEAE-X1I1

PER AXEL RYDBERG

NEW YORK
1921

Reprinted, without change of paging, from The Bulletin of the Torrey Botanical
Club 48: 159-172. August 1, 1921.

[From the Bulletin or the Torfey Botanical Club, a«: i-:--*:* 5 Aogast 1921]

Notes on Rosaceae — XIII utitH

Per Axel Rydberg

Roses of the Columbia region

In this region I here include the states of Oregon and Washing-
ton, together with British Columbia and northern Idaho. M
important are the roses of the Cascade Mountains. * In the north-
ern part many of the species of the Rocky Mountains have in-
vaded the region, not only on the western slopes of the Rockies
but also in the Cascades: in the Siskiyou Mountains of southwest-
ern Oregon there are also found some Californian species: and in
eastern Oregon some of those belonging to the Great Basin are
found. Since my revision in the Xorth American Flora we have
received in exchange a collection from British Columbia, from
the Geological Survey of Canada; and Professor J. K. Henry of
the University- at Vancouver has sent in another collection for
determination. Also. I have seen collections from southeastern
Oregon made by Miss Eastwood and Dr. Rehder for the Arnold
Arboretum.

Key to the groups of species

a much exserted. about equaling the ,ies

adnate; sepals reflexed. deciduous:
climbing, with curved pri . L Si

? not exserted. or only slightly so: stigmas forming a head

closing the mouth of the hypanthium.
Pistils numero _ he upper part

hypanthium persistent.

he Bulletin 48: 141-15$. pL 3 _

159

100

Rydberg: Notes ox Rosaceae

Stem climbing, with scattered prickles, rarely with
intermixed bristles; sepals more or less lobed; foliage
glandular-punctate, sweet-scented.
Stem not climbing, at least the young shoots bristly;
prickles infra-stipular or lacking; sepals entire or the
outer sometimes with one or two lobes; foliage not
very sweet-scented.
Pistils few; styles deciduous with the upper part of the
hypanthium, which falls off like a ring.

II. Caninae.

III. ClNNAMOMEAE.

IV. Gymnocarpae.

I. SYNSTYLAE

Stipules pectinately lobed and glandular-ciliate; corymb

many-flowered.
Stipules merely serrate; corymb one- to few-flowered.

1. R. midli 'flora.

2. R. arvensis.

I. Rosa multiflora Thunb.

See my notes in the preceding paper of this series.* The

species has been reported as a ballast plant at one station in

Washington.

2. Rosa arvensis Huds.
I

The following specimens were sent to me by J. C. Nelson,
principal of the high school at Salem, Oregon, for determination.
In the accompanying letter Mr. Nelson wrote among other things:
"He [the collector] reports this form as common in the vicinity of
Vancouver [Washington], apparently fully spontaneous. The
flowers were always single, and the petals of a wonderful shade of
pearly white." The specimens apparently belong to the so-called
Ayrshire Rose, which is by some regarded as a form of R. arvensis,
by others as a hybrid of the same. The leaflets are larger and
more pointed and the sepals more inclined to be lobed than in the
wild English form of that species.

Washington: Vancouver, R. V. Bradshaw 105 j.

II. CANINAE

Leaflets suborbicular or broadly oval, mostly rounded at the
apex; hypanthium in fruit obovoid or broadly ellipsoid,
abruptly contracted at the apex; sepals tardily deciduous
or persistent.

Leaflets ovate or oval, acute or short-acuminate; hypanthium
in fruit narrowly ellipsoid, tapering at both ends; sepals
early deciduous; styles glabrous or nearly so.

3. R. rul

4. R. micrantha.

* Bull. Torrey Club 47: 47- 1920.

Rydberg: Notes ox Rosaceae

161

3. Rosa rubiginosa L.

The European sweet brier, often cultivated, has established
itself at several places in Oregon and Washington. It is very
variable and the naturalized specimens show also much variation.
One of chese many forms, which is rather more glandular than
usual, was mistaken for a rative species and published by Dr.
E. L. Greene as new, under the name R. Walpoleana. The
author placed it in the Gymxocarpae, a group to which it has no
relation.

4. Rosa micraxtha Borrer

See Bull. Torrey Club 47: 49. 1920. This species also is
naturalized in Oregon and Washington.

III. CINNAMOMIAE

Infra-stipular prickles not present; branches unarmed or
bristly, not prickly.
Stem densely bristly even in age; flowers solitary.

Hypanthium decidedly pear-shaped or ellipsoid, acute
at the base, with a distinct neck at the top.
Sepals rarely more than i cm. long; bark yellowish

green.
Sepals 2-3 cm. long; bark brown.

Leaflets obovate, sparingly hairy beneath.
Leaflets elliptic, densely pubescent beneath.
Hypanthium subglobose, almost without a neck.
Stem unarmed or when young covered with more or less
deciduous bristles; flowers corymbose.
Fruit ellipsoid; leaflets serrate.
Fruit subglobose; leaflets crenate.
Infra-stipular prickles present.

Flowers mostly solitary; petals usually 2.5 cm. long or
more (except in No. 11); hypanthium in fruit 12-20
mm. thick.
Prickles straight or nearly so.
Hypanthium densely prickly.
Hypanthium not prickly or rarely slightly so.

Leaflets more or less double-toothed, more or
less glandular-granuliferous beneath; stip-
ules and rachis glandular.
Pedicels and calyx densely bristly but hypan-
thium glabrous; leaflets oval, rarely 2 cm.
long; prickles weak; petals 15 mm. long.
Pedicels and calyx not bristly, sometimes
slightly glandular-hispid.

s-

K.

collaris.

6.

R.

Butleri.

7-

R.

acicularis.

S.

R.

Bourgeauiana

2.

R.

Pringlei.

9-

R.

anacantha.

10. R. MacDougalii.

11. R. yainacensis.

162

Rydberg: Notes on Rosaceae

Leaves very thin, pale and slightly
glandular-puberulent, but not muri-
culate beneath. 12. R. Nutkana.

Leaves thicker, dark green on both sides,
conspicuously glandular-muricate be-
neath. 13. R. muriculata.
Leaflets simple-toothed, puberulent beneath,
but not glandular-granuliferous or -muricate;

rachis rarely glandular. 14. R. Spaldingii.

Prickles more or less curved.

Prickles very large, flat, pubescent; twigs densely

glandular and pubescent. 15- R- Durandii.

Prickles slender, glabrous, not conspicuously
flattened; twigs not conspicuously glandular,
mostly glabrous.
Sepals not glandular-hispid; leaflets 2-4 cm.
long, not conspicuously glandular double-
toothed. 16. R. Columbiana.
Sepals conspicuously glandular-hispid; leaflets
1-2 cm. long, conspicuously glandular double-
toothed. 17- R- myriadenia.
Flowers mostly corymbose, if solitary the petals 2 cm.
long or less; fruit rarely more than 1 cm. in diameter.
Prickles more or less curved.

Leaflets softly villous beneath; prickles stout,

flattened; hypanthium with a neck. 18. R. califomica.

Leaflets very finely puberulent and somewhat

glandular-pruinose beneath; prickles not

flattened; hypanthium without a distinct

neck.

Leaflets with simple teeth, not copiously

glandular-pruinose beneath. 19. R. puberulenta.

Leaflets with double gland-tipped teeth, con-
spicuously glandular-puberulent beneath. 20. R. delitescens.
Prickles straight or nearly so.

Hypanthium normally neither prickly or bristly.
Hypanthium globose; neck usually obsolete.
Sepals lanceolate, with long, caudate-
attenuate or sometimes foliaceous
tips, more than 1 cm. long.
Stipules, petiole, and rachis copiously
glandular; leaflets double-toothed,
with gland-tipped teeth, densely
glandular-pruinose or muriculate
beneath.
Leaflets orbicular or rounded oval. 13. R. muriculata.
Leaflets oblong to oval. 22. R. Fendleri.

Stipules, petiole, and rachis not glan-
dular or the stipules slightly glan-
dular-dentate.

Rydberg: Notes on Rosaceae

163

Leaflets glabrous or nearly so.

Leaflets broadly oval, 3-5 cm.

long.
Leaflets obovate or elliptic, rarely
3 cm. long.
Leaflets more or less pubescent be-
neath.
Leaflets villous or pilose beneath,

rounded or broadly oval.
Leaflets finely puberulent be-
neath.
Leaves obovate, more or less
pale or glaucous beneath;
prickles comparatively long,
spreading.
Leaflets elliptic or oval,
equally green on both
sides; plant with few
short ascending prickles.
Sepals decidedly glandular;

leaflets thin and soft.
Sepals not glandular, or
slightly so on the mar-
gins; leaflets rather thick.
Sepals ovate, less than 1 cm. long, not con-
spicuously caudate; leaflets simple-toothed;
puberulent beneath.
Hypanthium elongate with a distinct neck;
leaves finely puberulent beneath.
Leaflets light green, coarsely toothed; hy-
panthium in fruit pyriform.
Leaflets dark green above, finely toothed;
hypanthium in fruit ellipsoid.
Plant unarmed or nearly so; fruit 1 mm.

broad.
Plant with slender prickles; fruit 6-8 mm.
broad.
Hypanthium densely bristly or prickly; leaflets
orbicular in outline.
Leaflets subsessile.
Leaflets with petiolules, 1-4 mm. long.

21. R. rivalis.
23. R. Woodsii.

25. R. myriantha.

24. R. Maconnii.

26. R. pisocarpa.

27. R. ultramonlana.

28. R. Covillei.

31. R. pyrifera.

32. R. Pringlei.

33. R. Eastwoodiae.

29. R. spithamea.

30. R. adenocarpa.

5. Rosa collaris Rydberg

The following specimens are doubtfully referred to this species:
Washington: Fort Colville, 1880, S. Watson 121.

164 Rydberg: Notes on Rosaceae

6. Rosa Butleri Rydberg

This species belongs to the Rocky Mountain region and was
originally described from northwestern Montana. The following
specimens from British Columbia are referred here:

British Columbia: Camloops, May 24, 1912, A. Thorpe.

7. Rosa acicularis Lindl.

See Bull. Torrey Club. 47: 56. 1920. This is a common
species throughout the northern Rockies and the Cascade Moun-
tains.

8. Rosa Bourgeauiana Crepin

See notes in Bull. Torrey Club 47: 57. 1920. In the Columbia
region, this species has been found only north of the Canadian
boundary.

British Columbia: Okanogan Landing, Golden, Pense; Skeena
River, /. K. Henry.

9. Rosa anacantha Greene

This species is related to the eastern R. blanda, but differs in
the crenate instead of serrate leaflets and the smaller fruit. It is
known only from the type locality, in salt marshes near Tacoma,
Washington.

10. Rosa MacDougalii Holz.

This is regarded by many as a form of R. Niitkana. It differs
not only in the hispid fruit but also in the fact that the teeth of
the leaflets are seldom double as they are in R. Niitkana and the
leaflets are rarely glandular-granuliferous. Neither could it be
referred as a form of R. Spaldingii, for in the latter the leaflets are
pubescent beneath. The following specimens belong here:

Idaho: Landing-Cuprum road 1901, Cusick 2533, 2352;
Farmington Landing, 1892, Sandberg, MacDougal & Heller 572,
Heller 3256.

Oregon: Rye River Valley, Leiberg 4410.

Washington: Pullman, 1896, Elmer 72, 74.

11. Rosa yainacensis Greene
Until lately this species was known only from the type locality.
The habit, long prickles, and large fruit suggest R. Niitkana, but

Rydberg: Notes on Rosaceae 165

it is unique in the group on account of its densely glandular-
bristly pedicels. It evidently is a good species unless of hybrid
origin. It might have been produced by R. Nutkana and R.
gymnocarpa.

Oregon: Yainax Indian Reservation, Mrs. Austin.

California: Crescent City, Del Norte County, 1912, Eastwood
2270.

12. Rosa Nutkana Presl

Although an easily distinguished species (except from the two
next following species), it has been mistaken for R. fraxinifolia
(i.e. R. blanda) and R. Woodsii. It has also been named R. caryo-
carpa Dougl. and R. Lyalliana Crepin; but these names have not
been published, except in synonymy.

13. Rosa muriculata Greene

This is closely related to R. Nutkana and is perhaps not distinct.
It differs in the thicker and smaller leaves, densely glandular-
muricate beneath, and in the often corymbose inflorescence.
As it is more common on the Vancouver Island than the pre-
ceding species, there may be a possibility that it is the original
R. Nutkana Presl. In such a case the plant known as that species
would be without a published name.

14. Rosa Spaldingii Crepin

This has been confused with R. Nutkana but differs in the
simple-toothed leaflets, which are decidedly pubescent but scarcely
at all glandular-granuliferous beneath. Dr. Watson merged it in
R. Nutkana, and this fact probably influenced Crepin to withdraw
his species. The plant had been recognized before: Borrer in
Hooker's Flora included it in R. cinnamomea, which it approaches
more than any other American species does, differing principally
in the longer and straight prickles; Nuttall recognized it as a
species, R. megacarpa, but this name was published only as a
synonym in Torrey & Gray's Flora; in the meantime Rafinesque
had published it as R. macrocarpa Nutt. It is common through
the northern Rockies as well as the Cascades.

166 Rydberg: Notes on Rosaceae

15. Rosa Durandii Crepin

This was based on E. Hall 146, which was referred to Rosa
kamtschatica by Dr. Gray. Dr. Watson in his monograph reduced
it to a synonym of R. Nutkana. Crepin, probably influenced by
Watson's treatment, later retracted his species. In my opinion
R. Durandii is perfectly distinct from R. Nutkana. The leaves
are thicker, more hairy than in that species, and rarely double-
toothed; the prickles are not like those of R. Nutkana, but curved,
short, strongly flattened and pubescent. As long as only Hall's
specimens were known, the specimens might have been regarded
as freaks, but specimens almost exactly like these have since been
collected.

Oregon: E. Hall 146; Brownville, 1895, Canby; near Spring-
field, Coville & Applegate 1076.

British Columbia. Queen Charlotte's Island, Dawson

Si 44 (?)•

Idaho: Lower Priest River, Leiberg 288 j. This specimen is
very peculiar, nearly unarmed, but with the pubescence of the
leaves and twigs of R. Durandii.

16. Rosa Columbiana Rydberg

This also is related to R. Nutkana but differs in the strongly
curved prickles. It is a rather local plant, the following specimens
belonging here:

Idaho: Little Potlatch River, 1892, Sandberg, MacDougal &
Heller 381; Kooteney County, Sandberg.

Oregon: Forest Grove, 1893 and 1894, Lloyd.

17. Rosa myriadenia Greene
This species has the habit, toothing, and glandular pubescence
of R. muriculata, but the leaflets are smaller and the prickles are
decidedly curved; these characters would place it in the key
next to R. columbiana. From this it differs in the smaller and
more pubescent leaves. It is known only from the type locality,
Huckleberry Mountains, Jackson County, Oregon.

18. Rosa califorxica Cham. & Schlecht.
This Californian species has been collected in the extreme
southern part of Oregon.

Rydberg: Notes on Rosaceae 167

19. Rosa puberulexta Rydberg
This species belongs to the Great Basin and is related to

R. arizonica and R. neomexicana. The following specimens are

referred here, though somewhat doubtfully:

Washington: Wilson Creek, Sandberg & Leiberg; Rattlesnake

Mountains, Cotton 469.

Oregon: Antelope Creek, Applegate 2399; Cold Spring,

Coville & Applegate 1122; Wallowa, Samson & Pearson 78.

20. Rosa delitescens Greene
This is related to R. californica but characterized by the
double-serrate leaflets, glandular-pruinose beneath and with gland-
tipped teeth, and by the subglobose fruit. It is known only from
the type locality in the Siskiyou Mountains, Oregon.

21. Rosa rivalis Eastwood

This species belongs to the Californian flora (see notes, Bull.
Torrey Club 44: 74. 1917) but the following specimen has been
collected in Oregon:

Oregon: Cold Spring, Crook County, 1898, Coville & Apple-
gate 131.

22. Rosa Fendleri Crepin

23. Rosa Woodsii Lindl.

24. Rosa Macounii Greene

These three species belong to the Rocky Mountain region but
extend west into the Cascade Mountains. They will be treated in
a subsequent paper.

25. Rosa myriantha Carr.

This is primarily a Californian rose (see notes, Bull. Torrey
Club 44: 75, 76. 1917), but the following specimens belong here:

Oregon: Barlow Gate, Lloyd; lower Albina, Portland, Sheldon
10659; Wimmer, Hammond 119, in part.

26. Rosa pisocarpa A. Gray

27. Rosa ultramontana (S. Wats.) Heller
These two species have been treated previously, under the

Roses of California and Nevada (see Bull. Torrey Club 44: 77, 78.
191 7). Though the former has its best development in the

168 Rydberg: Notes on Rosaceae

Columbia region, especially west of the Cascades and the latter
in the northern part of the Great Basin, which includes eastern
Oregon, both extend north into British Columbia.

28. Rosa Covillei Greene

This species and R. Bolanderi are characterized among the
cinnamon roses by the short ovate, short-acuminate sepals.
R. Covillei is known only from the type locality, near Naylor,
Klamath County, Oregon.

29. Rosa spithamaea S. Wats.

This species of northern California has been collected at one
locality in the Columbia region : —
Oregon: Wimmer, Hammond 120.

30. Rosa adenocarpa Greene

Notwithstanding Dr. Greene's remark, "despite all its peculiar
characteristics, a genuine member of this gymnocarpa group,"
I am compelled to exclude it. The upper part of the hypanthium
and calyx show no sign of being deciduous, the sepals are not
those short ones of the R. gymnocarpa allies, and the hypanthium
is bristly. These characters and the dwarf herbaceous stems indi-
cate that the plant is related to R. spithamaea, where I have
placed it. It is known only from the type locality, Mt. Gray-
back, southwestern Oregon.

31. Rosa pyrifera Rydberg

This species is related on one hand to R. Macounii and on the
other to R. ultramontana and R. pisicarpa, but differs from them
all in the pear-shaped hypanthium, which has a more distinct
neck. It is common in the northern Rockies, but the following
specimens are from the Columbia region : —

Washington : Clarks Spring, Kraeger 47.

Oregon: Powder River Mountains, 1896, Piper; Horse
Creek Canyon, Sheldon 8138; Thompson Creek, Brown 86;
Ashland, Walpole 375; Tygh Valley, Walpole 336; Cow Creek,
Coville 1162; Blue Mountains, Cusick 16Q7.

Rydberg: Notes on Rosaceae 169

32. Rosa Pringlei Rydberg

See notes, Bull. Torrey Club 44: 79. 191 7. The following
specimens are from the region :

Oregon: Wallowa River, 1897, Sheldon 8687 (?).

Washington: Klickitat County, 1885, Suksdorf.

British Columbia: Armstrong Pass, 1912, E. Wilson; Elgin,
Miss Edstrom.

33. Rosa Eastwoodiae Rydberg

This is related to R. Pringlei but the plant has smaller leaves,
smaller flowers and fruit, and is usually prickly. The fruit is
only 6-8 mm. in diameter. To this belong: —

California: Sisson, Siskiyou County, 1912, Eastwood 2100,
mainly.

Oregon: Waldo, Josephine County, 1912, Eastwood 2168.

IV. GYMNOCARPAE

Leaves glabrous beneath.

Flowers mostly 2-4 together, leafy-bracted; upper stipules
broadly dilated; leaflets with broadly ovate teeth, in-
distinctly reticulate beneath; fruit 6-8 mm. broad. 37. R. dasypoda.
Flowers usually solitary, not leafy-bracted; stipules nar-
row; leaflets with lanceolate or ovate lanceolate
teeth; fruit 6 mm. in diameter or less.
Leaves not very thin, pale and indistinctly reticulate

beneath; sepals 1-1.5 cm. long. 35. R. leucopsis.

Leaves very thin, scarcely paler on the lower surface,

which is distinctly reticulate with semi-pellucid

veinlets; sepals less than 1 cm. long.

Leaflets 5-7, usually more than 1.5 cm. long; teeth

ovate-lanceolate in outline; terminal leaflet

usually rounded at the base. 34. R. gymnocar pa.

Leaflets 7-9, usually less than 1.5 cm. long, the
terminal one usually acute or cuneate at base;
teeth lanceolate in outline. 36. R. prionota.

Leaves pubescent beneath. 38. R. Bridgesii.

34. Rosa gymnocapra Nutt.

This species is one of the most common roses in the Columbia
region (see my notes, Bull. Torrey Club 44: 82. 1917). As
stated there Dr. Greene described twelve species belonging to the
group and some of these will be discussed here.

Rosa glancodermis Greene. Dr. Greene pays especial attention

170 Rydberg: Notes on Rosaceae

to the ashy gray bark and the petioluled leaflets, characters
which I think are not distinctive. In the typical R. gymnocarpa
the bark is often light in color on old stems, and many specimens
from the range of the typical R. gymnocarpa have more or less
petioluled leaflets, such as Jones 1088 and R. S. Williams 876 from
Montana, Sheldon 8521 from Oregon, and Allen 72 from Washing-
ton. I therefore regard this as a synonym of R. gymnocarpa.

Rosa abietorum Greene, I take as an unusually bristly R.
gymnocarpa. I have been unable to find any constant character.

Rosa amplifolia Greene. I have not been able to distinguish
this from R. gymnocarpa. Dr. Greene's remarks, "the leaves
are so very large and have so much the color, texture and pattern
of R. acicularis, that but for the small solitary flowers this would
have passed readily with many for that species," are at least con-
fusing. R. acicularis has rather thick, elliptic leaflets, decidedly
pubescent beneath. Dr. Greene must have had an erroneous
idea of R. acicularis.

35. Rosa leucopsis Greene

This is so close to R. gymnocarpa that I hesitated to admit it
as a species. In R. gymnocarpa and R. prionota the leaflets are
very thin and reticulate with semi-pellucid veins. In R. leucopsis
the leaves are not so thin and the veins are faint and not at all
pellucid. Otherwise the three plants are very similar. The
species was described from fruiting material; the plant becomes
paler in age.

Rosa Helleri Greene is evidently the same plant in flowers and
shows only slight variation from the type of R. leucopsis. The
Idaho specimens are mostly like those of the type of R. Helleri,
but two numbers of Sandberg, MacDougal & Heller are almost
identical with the type of R. leucopsis.

36. Rosa prionota Greene

This has been discussed before (see Bull. Torrey Club 44:
32. 1917).

Rosa piscatoria Greene I regard as an unusually strong and
bristly plant of R. prionota.

Rosa apiculata Greene was so named on account of the form

Rydberg: Notes on Rosaceae 171

of the fruit, "which is elongated, fully twice as long as broad,
ending in a narrow necklike apiculation." The type sheet bears
eight fruits. Of these only four are twice as long as broad, three
are ellipsoid, but not as long, and one is nearly globose, slightly
pear-shaped; four of them (three of the longer and one of the
short-ellipsoid ones) have a conspicuous neck; one of the short-
ellipsoid ones has no neck and the other three a short one. Miss
Cooley's specimens from Vancouver Island, which from the locality
should belong to this species and which agree with it in all respects
except the fruit, has three fruits. One of these is elongated-
ellipsoid, one somewhat pear-shaped and one subglobose; all with
obsolete neck. They cannot be distinguished from R. prionota.

To R. prionota belongs the following specimens from the
Columbia region:

Washington: Whiddley Island, Saunders (type of R. apicu-
lata) .

British Columbia: Nanaino, Miss Cooley.

36. Rosa dasypoda Greene

See notes, Bull. Torrey Club 44: 83. 191 8. To this species
belong the following specimens:

Oregon: Wallowa County, Sheldon 8813 and several other
numbers; Siskiyou Mountains, Applegate 2251; without locality,
E. Hall 143; Toledo Canyon, Rusby; Wimmer, Hammond 117.

Washington : Chehalis County, Heller 3897.

British Columbia: Chilliwack Valley, Macoun 79841, 24748;
between Kettle and Columbia River, Macoun 64008, 64007;
Esquimalt, Macoun 79840.

37. Rosa Bridgesii Crepin

See Bull. Torrey Club 44: 83. 191 7. To this species belong
the following: —

Oregon: between Union Creek and Whiskey Creek, Applegate
2622; Huckleberry Mountains, Coville 1434.

HYBRIDS
The following hybrids have been recorded from this region: —

Rosa acicularis X Fendleri
British Columbia: Golden, /. K. Henry.

172 Rydberg: Notes on Rosaceae

Rosa acicularis X muriculata
British Columbia: Narramantha, Lake Okanogan, /. K.
Henry.

Rosa Bourgeauiana X muricata

British Columbia: Hazelton, Skeena River 1915, /. K.
Henry.

Rosa gymnocarpa X Nutkana

British Columbia: Crescent 191 5, J. K. Henry.

Rosa Nutkana X pisocarpa
British Columbia: Savannas, 1892, F. E. Lloyd.

Rosa pisocarpa X Spaldingii
Eastern Oregon: 1900, Cusick 2418.

Rosa gymnocarpa X pisocarpa
Oregon: Siskiyou Mountains, 1904, Render.

Rosa gymnocarpa X muriculata
British Columbia: Cameron Lake, Vancouver Island, 1912,
/. K. Henry.

Rosa gymnocarpa X Woodsii

British Columbia: Skeena River 191 5, /. K. Henry.

New York Botanical Garden

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 233

A REARRANGEMENT OF THE

BOLIVIAN SPECIES OF CENTROPOGON

AND SIPHOCAMPYLUS

H. A. GLEASON

NEW YORK
1921

Reprinted, without change of paging, from Bulletin of the Toeeey Botanical Club
48: 189-201. September 26, 1921.

[From the Bulletin o» the Torrey Botanical Club, 48: 18^-201. 26 September 1921 ]

A rearrangement of the Bolivian species of Centropogon and
Siphocampylus

H. A. Gleasont

Of forty species enumerated in the following article, six have
been known to science for from fifty to ninety years, while thirty-
four bear names not over thirty years old. Modern investigation
of the two genera begins with Zahlbruckner, who in 1891* de-
scribed three new species from the collections of Mandon. In
1892 Britton published in his enumerationf of the South American
plants collected by Rusby a list of fifteen species and two varieties
of these genera, of which eight species and both varieties were
described as new. This was followed in 1896 by Rusby 's paper J on
the Bolivian collections of Miguel Bang, in which four species
were mentioned, including three described as new. Zahlbruckner
in i897§ monographed the Lobeliaceae of Bolivia and recognized
nine species of Centropogon, of which two were published as new,
and twenty species and three varieties of Siphocampylus, of which
eight species and the three varieties were regarded as new, making
a total of thirty-four. Rusby listed || several species of both
genera and described one as new in 1907, and in 1912 described**

*Ann. K. K. Naturhist. Hofmus. Wien 6: 432-444. 1891.

f Bull. Torrey Club 19: 371-374. 1892.

% Mem. Torrey Club 6: 72-74. 1896.

§ Bull. Torrey Club 24: 371-385. 1897.

|| Bull. N. Y. Bot. Gard. 4: 400-404. 1907.

** Bull. N. Y. Bot. Gard. 8: 121-123. 1912.

[The Bulletin for June (48: 159-188) was issued August 1, 1921.]

189

190 Gleason: Centropogon and Siphocampylus

three new species of Siphocampylus and one of Centropogon from
the collections of Williams. Lauterbach published* a single
variety in 1910, and in 1913 Zahlbruckner added three speciesf
from the collections of Herzog.

Not all of these various species and varieties still stand just
as they were published, but it is worthy of note that the studies
in which they appeared were so exhaustive of the material at
hand and subsequent collections have been so scanty that the
present paper adds but one new species, although a second, recog-
nized by Zahlbruckner and given a manuscript name by him, is
here published. It is nevertheless fair to assume that more
extended collection in Bolivia will bring to light probably a score
or more new species.

Zahlbruckner's work was based largely on the same material
as that used by Britton and Rusby, with additional specimens from
certain European herbaria. The types of Britton 's and of Rusby's
species, and either the types or isotypes of most of Zahlbruckner's
species, are in the collections at the New York Botanical Garden.

The four cited articles by Britton and Rusby include mere
enumerations of the species with detached descriptions of their
new species and varieties. Zahlbruckner gives in his monograph
somewhat more detailed descriptions of the new species and
frequent notes on the supposed affiliation of the others, and pre-
faces each genus with a key to the species included. His keys,
however, are in general based on minor characters. While
obviously related species are in most cases keyed out together,
the characters used in distinguishing them are not of fundamental
importance, give little idea of the relationship of the species, and
can not be successfully used for the species of other South American
countries.

Centropogon and Siphocampylus present the usual lobeliaceous
structure. The leaves are almost always alternate, usually ample
in size, and with a wide variation in shape, texture, and pubescence.
The inflorescence is in a strict sense a terminal raceme, which by
shortening of the internodes and suppression of the bracteal leaves
becomes a terminal corymbose or subumbellate cluster, or by

* Buchtien, Contr. Flora Bolivia I: 187. 1910.
t Med. Rijks Herb. 19: 49~53- 1913-

Gleason: Centropogon and Siphocampylus 191

elongation of the internodes and better development of the sub-
tending leaves is reduced to a series of solitary axillary flowers.
The hypanthium, varying from depressed hemispheric to cylindric,
bears five sepals, ranging from a size approximating the length
of the corolla-tube to almost suppressed. The large corolla is
usually brightly colored, with five equal or unequal, long or short,
spreading, erect, or depressed lobes. The filaments usually surpass
the corolla, and the anther-tube is frequently hirsute; in almost
all species the two anterior anthers bear an appendage or tuft of
hairs at the apex. The foliage is in some cases glabrous, but in
most species is pubescent to tomentose with simple, branched
or stellate hairs.

The distinctions between Centropogon and Siphocampylus are
difficult and in herbarium material frequently obscure or lacking;
as a result, sheets of what is obviously one species have been
classified by Zahlbruckner himself in both genera. In Centropogon
the fruit is a dry or leathery indehiscent berry. This flattens out
in pressing and is thin enough to show the impression of the small
seeds within. In Siphocampylus the fruit is a stiff, firm-walled
capsule, but in many or even most herbarium sheets available
mature fruit is not at hand. In the former genus, also, the summit
of the ovary is described as truncate, that is, as almost completely
adnate to the hypanthium, while in the latter it is stated to be
conical, with a free distal portion, and adnate to the hypanthium
only at its base. This condition does obtain in the ripened
fruit, but can not always be demonstrated in flower. Between
the baccate and the capsular species there is extensive parallelism ;
so extensive that a doubt may legitimately arise whether the
characters of the fruit are really of generic value.

Because of this parallelism, the species of the two genera have
not been kept separate in this paper but have been keyed out side
by side, using the shape of the corolla as the primary distinction
between groups. Three such groups have been distinguished,,
of which one includes only species of Centropogon (in the usual
sense), the second only species of Siphocampylus, and the third
species of both genera. To obviate the necessity of descriptions,
the analytical key has been made more detailed than necessary
for the mere separation of the species.

192 Gleason: Centropogon and Siphocampylus

In both genera, the species show unusually clear lines of
demarcation, and are separated by characters of the hypanthium,
corolla, anther-tube, inflorescence, foliage, and pubescence.
But since these characters show little correlated variation, the
delimitation of species-groups and the determination of intragen-
eric relations become a matter of some difficulty, which is
increased rather than lessened by extending the study to other
South American countries.

Both genera find their center of distribution with the greatest
number of species in the Andean region of northwestern South
America from Colombia to Bolivia. Other South American and
all North American forms are to be referred to this center for
their geographical origin.

Conspectus of the species

I. Corolla relatively short and broad, but large, the tube equaling or little exceeding
the lobes, thick and firm in texture, white, yellowish, ochroleucous, or dull
purple; corolla-lobes all depressed, the upper somewhat exceeding the lower;
sepals ample, equaling or longer than the hypanthium; flowers solitary, on long
axillary peduncles.
A. Hypanthium turbinate, broader than high when pressed, prominently ribbed
anther-tube glabrous, or pubescent only on the connectives; corollas white
or cinereous, the lobes equaling or exceeding the tube in length; fruit cap-
sular, so far as known.
i. Anther-tube glabrous, except the terminal brush.

a. Sepals triangular-subulate, 4 mm. long, much shorter than the hypan-

thium.

1. Siphocampylus tunarensis A. Zahlb. Bull. Torrey Club 24:

376. 1897.
The specimen in the herbarium of the New York Botanical
Garden is fragmentary and does not permit the verification of
Zahlbruckner's detailed description, but is sufficient to demon-
strate its close relation to the following three species. Fruit has
not been seen.

b. Sepals linear, equaling or exceeding the hypanthium in length.

2. Siphocampylus Rusbyanus Britton, Bull. Torrey Club 19:

372. 1892.
Sepals 10 mm. long by 2 mm. wide; leaf-blades ovate-elliptic,
5-9 cm. long, acute or obtuse, rugose above, reticulate beneath,
abruptly narrowed into a petiole 10-15 mm. long.

Gleason: Centropogon and Siphocampylus 193

3. Siphocampylus Vatkeanus A. Zahlb. Bull. Torrey Club 24:

377- 1897-
Sepals little exceeding the hypanthium; leaf-blades oblong-
lanceolate, 6-10 cm. long, 2-3 cm. wide, acuminate, tomentose on
the reticulate veins beneath, obtuse or rounded at base, on petioles
1-2 cm. long.

2. Anther-tube white-woolly.

4. Siphocampylus radiatus Rusby, Mem. Torrey Club 6: 73.

1896.
Sepals 20-25 mm. long, 3-4 mm. wide, somewhat dilated
below; leaf-blades lanceolate, 15-25 cm. long, acuminate, flat
above, not conspicuously reticulate, gradually narrowed below to
an indefinite petiole or sessile base.

4a. Siphocampylus radiatus minor A. Zahlb. Bull. Torrey
Club 24: 376. 1897.
Plant smaller in all its parts; pedicels surpassing the leaves.

B. Hypanthium short-cylindric, higher than broad when pressed, or depressed-
hemispheric; corolla-lobes usually shorter than the tube; anther-tube densely
woolly; fruit baccate, so far as known.

I. Sepals linear or narrowly lanceolate, four to ten times as long as broad, longer
than the short-cylindric hypanthium, equaling or exceeding the corolla-tube,
with rounded sinuses nearly or quite as broad as the sepals; leaf-blades
large, elliptic, 15-25 cm. long.

5. Cextropogon incanus (Britton) A. Zahlb. Bull. Torrey

Club 24: 374. 1897.
Siphocampylus incanus Britton, Bull. Torrey Club 19: 373.

1892.
Petals and sepals densely floccose with yellowish gray hairs;
filaments little exserted, the anther-tube barely surpassing the
corolla; leaf-blades 5-10 cm. long, abruptly acuminate, floccose
below and somewhat so above.

6. Centropogon Brittonianus A. Zahlb. Bull. Torrey Club

24: 373. 1897.
Siphocampylus giganteus latifolius Britton, Bull. Torrey Club

I9:373- 1892.
Petals and sepals thinly pubescent; filaments long (2 cm. or
more) -exserted; leaf-blades 4-7 cm. broad, gradually acuminate,
thinly tomentose on the veins beneath, puberulent above.

192 Gleason: Centropogon and Siphocampylus

In both genera, the species show unusually clear lines of
demarcation, and are separated by characters of the hypanthium,
corolla, anther-tube, inflorescence, foliage, and pubescence.
But since these characters show little correlated variation, the
delimitation of species-groups and the determination of intragen-
eric relations become a matter of some difficulty, which is
increased rather than lessened by extending the study to other
South American countries.

Both genera find their center of distribution with the greatest
number of species in the Andean region of northwestern South
America from Colombia to Bolivia. Other South American and
all North American forms are to be referred to this center for
their geographical origin.

Conspectus of the species

I. Corolla relatively short and broad, but large, the tube equaling or little exceeding
the lobes, thick and firm in texture, white, yellowish, ochroleucous, or dull
purple; corolla-lobes all depressed, the upper somewhat exceeding the lower;
sepals ample, equaling or longer than the hypanthium; flowers solitary, on long
axillary peduncles.
A. Hypanthium turbinate, broader than high when pressed, prominently ribbed
anther-tube glabrous, or pubescent only on the connectives; corollas white
or cinereous, the lobes equaling or exceeding the tube in length; fruit cap-
sular, so far as known.
i. Anther-tube glabrous, except the terminal brush.

a. Sepals triangular-subulate, 4 mm. long, much shorter than the hypan-

thium.

1. Siphocampylus tunarensis A. Zahlb. Bull. Torrey Club 24:

376. 1897.
The specimen in the herbarium of the New York Botanical
Garden is fragmentary and does not permit the verification of
Zahlbruckner's detailed description, but is sufficient to demon-
strate its close relation to the following three species. Fruit has
not been seen.

b. Sepals linear, equaling or exceeding the hypanthium in length.

2. Siphocampylus Rusbyanus Britton, Bull. Torrey Club 19:

372. 1892.
Sepals 10 mm. long by 2 mm. wide; leaf-blades ovate-elliptic,
5-9 cm. long, acute or obtuse, rugose above, reticulate beneath,
abruptly narrowed into a petiole 10-15 mm. long.

Gleason: Centropogon and Siphocampylus 193

3. Siphocampylus Vatkeanus A. Zahlb. Bull. Torrey Club 24:

377. 1897.
Sepals little exceeding the hypanthium; leaf-blades oblong-
lanceolate, 6-10 cm. long, 2-3 cm. wide, acuminate, tomentose on
the reticulate veins beneath, obtuse or rounded at base, on petioles
1-2 cm. long.

2. Anther-tube white-woolly.

4. Siphocampylus radiatus Rusby, Mem. Torrey Club 6: 73.

1896.
Sepals 20-25 mm. long, 3-4 mm. wide, somewhat dilated
below; leaf-blades lanceolate, 15-25 cm. long, acuminate, flat
above, not conspicuously reticulate, gradually narrowed below to
an indefinite petiole or sessile base.

\a. Siphocampylus radiatus minor A. Zahlb. Bull. Torrey
Club 24: 376. 1897.
Plant smaller in all its parts; pedicels surpassing the leaves.

B. Hypanthium short-cylincliic, higher than broad when pressed, or depressed-
hemispheric; corolla-lobes usually shorter than the tube; anther-tube densely
woolly; fruit baccate, so far as known.

1. Sepals linear or narrowly lanceolate, four to ten times as long as broad, longer
than the short-cylindric hypanthium, equaling or exceeding the corolla-tube,
with rounded sinuses nearly or quite as broad as the sepals; leaf-blades
large, elliptic, 15-25 cm. long.

5. Centropogon incanus (Britton) A. Zahlb. Bull. Torrey

Club 24: 374. 1897.
Siphocampylus incanus Britton, Bull. Torrey Club 19: 373.

1892.
Petals and sepals densely floccose with yellowish gray hairs;
filaments little exserted, the anther- tube barely surpassing the
corolla; leaf-blades 5-10 cm. long, abruptly acuminate, floccose
below and somewhat so above.

6. Centropogon Brittonianus A. Zahlb. Bull. Torrey Club

24: 373. 1897.
Siphocampylus giganteus latifolius Britton, Bull. Torrey Club

i9:373- 1892.
Petals and sepals thinly pubescent; filaments long (2 cm. or
more)-exserted; leaf-blades 4-7 cm. broad, gradually acuminate,
thinly tomentose on the veins beneath, puberulent above.

194 Gleason: Centropogox and Siphocampylus

6a. Centropogon Brittonianus brevidentatus Zahlb. & Rech-
inger, Med. Rijks Herb. 19: 51. IQI3-
The authors state that the variety differs from the species in
sepals only 10-12 mm. long and in much narrower and longer
corolla-lobes.

2. Sepals broadly ovate to oblong, less than four times as long as wide, mostly
shorter than the corolla-tube and not much exceeding the hypanthium, with
narrow acute sinuses; hypanthium depressed-hemispheric or somewhat
urceolate.

a. Sepals broadly ovate-triangular, herbaceous.

7. Siphocampylus tunicatus A. Zahlb. in herb.
Hypanthium and foliage nearly glabrous; leaf-blades broadly

elliptic to obovate, 20-25 cm. long, 7-13 cm. wide, thin, abruptly
acuminate; additional characters as in the key.

Type collected April, 1892, in Bolivia, at an altitude of 3000
m., by Otto Kuntze, and deposited in the herbarium of the New
York Botanical Garden. A second sheet from the same collector
comes from Santa Rosa, Bolivia, and is possibly a part of the
same collection. The fruit is unknown but the floral characters
show unmistakably the close relation of the species to the two
following, Centropogon Mandonis and C. gloriosus, and it is
quite probable that it will eventually be transferred to the genus
Centropogon.

b. Sepals oblong or oblong-ovate, thick and firm, serrulate.

8. Centropogon Mandonis A. Zahlb. Ann. K. K. Naturhist.

Hofmus. Wien 6: 438. 1891.
Corolla-tube about twice as long as the erect, narrowly oblong
sepals; filaments much surpassing the corolla; anther-tube hirsute
with tawny hairs; leaves acute.

9. Centropogon gloriosus (Britton) A. Zahlb. Bull. Torrey

Club 24: 373. 1897.
Siphocampylus gloriosus Britton, Bull. Torrey Club 19: 373.

1892.
Corolla-tube one third to one half longer than the oblong-ovate,
frequently spreading or renexed sepals; filaments about equaling
the corolla; anther- tube hirsute with purple hairs; leaves long-
acuminate.

Gleason: Centropogon and Siphocampylus 195

II. Corolla tubular, slender, straight or somewhat curved, usually constricted above
its base, little if any wider at the throat than at the base; corolla-lobes linear or
linear-deltoid and regularly tapering from base to apex, all erect or slightly
spreading, the lateral and ventral fissures usually progressively deeper than
the dorsal; filaments equaling or barely exceeding the dorsal petals; the two
ventral anthers with a terminal brush of separate hairs; fruit capsular, so far
as known.
A. Hypanthium well developed and the ovarj- consequently mainly inferior.

i. Flowers in leafy corymbs, leafy racemes, or solitary in the axils of normal
foliage leaves.
a. Anther-tube glabrous, except for the terminal brush, or with a few
scattered hairs only.
i. Flowers axillary, subtended by normal foliage leaves.

* Sepals linear, exceeding the hypanthium; leaves much longer
than wide,
f Peduncles well developed, equaling or surpassing the
flowers, which exceed the subtending leaves.
% Leaves in whorls of three.

io. Siphocampylus Orbigniaxus A. DC; DeCandolle, Prodr.
7: 405. 1839.

Siphocampylus volubilis Britton, Bull. Torrey Club 19: 372.
1892.

Leaf-blades ovate to ovate-oblong, sharply and irregularly
dentate; sepals filiform, much exceeding the hypanthium; corolla
scarlet.

The wealth of herbarium material indicates that this is the
commonest Bolivian species of the two genera. It is distinguished
at once by its verticillate leaves.

XX Leaves alternate, oblong or elliptic to linear.
§ Corolla yellow or yellowish; leaf-blades linear to narrowly lanceolate, subentire,
undulate, or with minute spinulose teeth.

11. Siphocampylus Kuntzeanus A. Zahlb. Bull. Torrey Club

24:378. 1897.
Anther-tube surpassing the corolla; leaves neatly linear,
crowded, conduplicate.

12. Siphocampylus aureus Rusby, Mem. Torrey Club 6: 72.

1896.
Siphocampylus aureus latior A. Zahlb. Bull. Torrey Club 24:

378. 1897.
Anther-tube not exserted; leaves linear-lanceolate, membra-
nous, flat.

§§ Corolla red.
|| Leaf-blades nairowly oblong-linear, narrowed below to an indefinite petiole, thick,
remotely denticulate.

196 Gleason: Centropogon and Siphocampylus

13. Siphocampylus Williamsii Rusby, Bull. N. Y. Bot. Gard.

8: 122. 1912.

|| || Leaf-blades lanceolate to ovate-lanceolate, petioled, membranous, sharply and
irregularly spinulose, 8-12 cm. long.

14. Siphocampylus boliviensis A. Zahlb. Ann. K. K. Natur-

hist. Hofmus. Wien 6: 443. 1891.
Sepals equaling or barely exceeding the subglobose hypanthium ;
stems stout; leaves firm, brownish and scabrously pubescent be-
neath.

15. Siphocampylus dubius A. Zahlb. Bull. Torrey Club 24:

385- i897-
Sepals longer than the depressed-hemispheric hypanthium;
stems slender; leaves thin, green and puberulent on the veins
beneath.

ft Peduncles short, equaling or shorter than the flowers, which do not surpass the
subtending leaves; hypanthium turbinate; sepals linear, refiexed; leaf-blades
oblong, short-petioled; stems climbing.

16. Siphocampylus membranaceus Britton, Bull. Torrey Club 19 :

372. 1892.
Leaves thin, membranous, sharply spinulose-dentate, veiny;
peduncles equaling the corolla.

17. Siphocampylus oblongifolius Rusby, Mem. Torrey Club

6:73. 1896.
Leaves firm, minutely and remotely spinulose; peduncles
shorter than the corolla.

** Sepals narrowly triangular, shorter than the hypanthium.

18. Siphocampylus correoides A. Zahlb. Bull. Torrey Club

24: 382. 1897.
Foliage and flowers tomentulose; leaf-blades broadly ovate,
undulate and spinulose-denticulate, 15-25 mm. long.

ii. Flowers in terminal corymbs or much abbreviated racemes; hypanthium turbinate
or somewhat urceolate; sepals about equaling the hypanthium, distant, with
rounded sinuses; leaf-blades sharply and saliently serrate, of an ovate type,
firm, dull-green.

19. Siphocampylus corymbiferus Pohl. Plant. Brasil. 2: 112.

pi. 175. 1831.
Lobelia corymbifera Presl, Prodr. Monogr. Lobel. 37. 1836.
Siphocampylus gracilis glabris Britton, Bull. Torrey Club 19:

374. 1892.
Leaves glabrous.

Gleason: Centropogon and Siphocampylus 197

20. Siphocampylus gracilis Britton, Bull. Torrey Club 19: 374.

1892.
Siphocamplyus corymbiferus gracilis A. Zahlb. Bull. Torrey

Club 24: 384. 1897.
Leaves softly pubescent.

iii. Flowers in crowded leafy racemes.

21. Siphocampylos tupaeformis A. Zahlb. Ann. K. K. Natur-

hist. Hofmus. Wien 6: 440. 1891.
Hypanthium hemispheric or somewhat urceolate, ribless or
faintly nerved; flowers 40-45 mm. long, approximately equaling
the ascending, straight or gently curved pedicels; leaves softly
pubescent beneath.

b. Anther-tube conspicuously woolly in five strongly marked lines; leaf-blades lanceo-
late or broader, rounded at base; sepals much longer than the hypanthium,
which is strongly ribbed at maturity.

* Leaf-blades long-acuminate, ovate-lanceolate to lance-oblong, sharply and
irregularly spinulose-denticulate.

22. Siphocampylus argutus A. Zahlb. Bull. Torrey Club 24:

383. 1897.
Leaf-blades thin, glabrous beneath, not prominently reticulate;
petioles straight or curved.

23. Siphocampylus bilabiatus A. Zahlb. Bull. Torrey Club

24:382. 1897.
Siphocampylus bilabiatus glabratus Lauterbach; Buchtien,

Contr. Flora Bolivia 1: 187. 1910.
Leaf-blades firm, pubescent beneath, prominently reticulate;
petioles sigmoid.

** Leaf-blades ovate-oblong, 5-8 cm. long, broadly rounded or subcordate at
base, obtuse or rounded at apex, minutely and remotely denticulate.

24. Siphocampylus subcordatus Rusby, Bull. N. Y. Bot.

Gard. 8: 121. 1912.

2. Flowers in terminal, leafless, bracted racemes; pedicels flexuous in anthesis, 1 cm
long or less; hypanthium turbinate, strongly ribbed or subangulate, somewhat
constricted at the throat, equaling or exceeding the relatively short sepals;
calyx-sinuses broad.

a. Hypanthium glabrous; leaf-blades falcate, conduplicate.

198 Gleason: Centropogon and Siphocampylus

25. Siphocampylus angustiflorus Schlecht.; Lechler, PI.

Lechler. Peruv. 264Q.
Specimens of this species have not been seen: the distinction
between it and the following are taken from Zahlbruckner.

b. Hypanthium pubescent; leaf-blades flat.

26. Siphocampylus flagelliformis A. Zahlb. Bull. Torrey

Club 24: 380. 1897.
Siphocampylus angustiflorus Britton, Bull. Torrey Club 19:

372. 1892; not S. angustiflorus Schlecht.
Leaf-blades lanceolate or narrowly elliptic-ovate, acuminate,
40-60 mm. long, 15-22 mm. wide, sharply serrulate, glabrous and
conspicuously veiny beneath.

27. Siphocampylus altiscandens sp. nov.

Stems puberulent, climbing 40 feet high on trees; leaves rather
crowded, the blades flat, firm in texture, ovate-oblong, acute or
subacuminate, entire or remotely and minutely serrulate, almost
glabrous above, minutely puberulent beneath, the larger ones
60-75 mm- l°ng by 25~35 mm- wide; inflorescence secund, the
pedicels 10 mm. long or less; flowers deflexed; hypanthium thinly
pubescent, turbinate or obconic, 4 mm. high; sepals triangular,
1 mm. long, with recurved tips; corolla rose-color, 30-35 mm. long,
its lobes linear, erect or spreading; filaments about equaling the
petals; anthers glabrous below, all more or less hirsute at the tip.

Type, Bang 2553, collected at Uchimachi, Corvico, Bolivia,
July 20, 1894, and deposited in the herbarium of Columbia Uni-
versity. His field notes state that it grows in damp forests and
that only two plants were seen, from which forty specimens were
collected. It is sharply distinguished from the preceding species
by its leaf characters.

13. Hypanthium very small in proportion to the flower and the ovary mainly free;
corolla scarlet, 6-7 cm. long; leaf-blades of a lanceolate or ovate type; sepals
usually serrulate; stems climbing: anther-tube glabrous.
I. Sepals linear to linear-oblong, prominently reflexed; leaves sharply acuminate.

28. Siphocampylus elegans Planch. Flore des Serres 6: 19.

1850.
Leaf-blades thick and firm, elongate-lanceolate, rounded or
truncate at base.

Siphocampylus elegans boliviensis A. Zahlb. Bull. Torrey

Club 24: 381. 1897.
The single specimen examined, Mandon 498, does not agree

Gleason: Centropogon and Siphocampylus 199

perfectly with Planchon's figure and may prove, specifically
distinct.

29. Siphocampylus reflexus Rusby, Bull N. Y. Bot. Gard. 4:

403. 1907.
Siphocampylus elegans cordatus A. Zahlb. Bull. Torrey Club

24:381. 1897.
Leaf-blades thin, ovate-lanceolate, with a long narrow acumi-
natum, distinctly cordate at base.

2. Sepals ovate-oblong, erect or somewhat spreading; leaves acute to subacumi-
nate.

30. Siphocampylus andinus Britton, Bull. Torrey Club 19:

373. 1892.

III. Corolla tubular, stout, conspicuously ventra'.ly curved, little or not at all con-
stricted at base, distinctly ampliate toward the throat, corolla-lobes with
broad bases, more or less oblique, and curved or directed anteriorly, the two
upper usually distinctly larger and broader than the three lower, the lower
fissures little if any deeper than the dorsal; fruit baccate, so far as known.

A. Ventral anthers terminating in a tuft of loose hairs, dorsal anthers sparsely

pilose; hypanthium distinctly urceolate; foliage and flowers strongly to-
mentulose; petals all about equal, broadly and obliquely triangular, acute;
filaments long-exserted; leaves petioled, the blades oblong-lanceolate, acute.

31. Centropogon unduavexsis (Britton) A. Zahlb. Bull.

Torrey Club 24: 374. 1897.
Siphocampylus unduavensis Britton, Bull Torrey Club 19:
373. 1892.

B. Ventral anthers terminating in a stiff scale composed of united hairs; hypan-

thium hemispheric to campanulate; foliage and flowers no* tomentulose;
at least the much smaller lower petals terminating in linear or subulate
strongly curved appendages.
1. Sepals linear to niangular, with broad flat sinuses; leaf-blades narrowly to
broadly oblong or obovate.
* Leaf-blades broadly oblong, almost entire.

32. Centropogon yungasensis Britton. Bull. Torrey Club 19:

371. 1892.
Leaves firm, glabrous; sepals triangular, distinctly shorter
than the hypanthium.

** Leaf-blades narrowly oblong denticulate or crenate.

33. Centropogon aggregatus (Rusby) comb. nov.
Siphocampylus aggregatus Rusby, Bull. N. Y. Bot. Gard. 8:

122. 1912.

200 Gleason: Centropogon and Siphocampylus

Filaments tomentose toward the summit; anther-tube pubes-
cent in lines; sepals equaling the hypanthium or a little shorter;
leaves crenate.

34. Centropogon surinamensis (L.) Presl, Prodr. Monogr.

Lobel. 48. 1836.

Lobelia surinamensis L. Sp. PI. ed. 2, 1320. 1763.

Filaments glabrous or with scattered hairs; anther-tube
sparsely hirsute or glabrous; sepals serrate, distinctly exceeding
the hypanthium: leaves denticulate.

2. Sepals linear-triangular to ovate, with narrow acute sinuses; leaf-blades oblong-
lanceolate to lanceolate, usually sharply denticulate; flowers in terminal corym-
bose clusters.
* Sepals narrowly linear-triangular, entire, much longer than the hypanthium.

35. Centropogon amplifolius Vatke, Linnaea 38: 716. 1874.

** Sepals triangular-ovate to broadly ovate.

36. Centropogon Bangii A. Zahlb. Bull. Torrey Club 24:

372. 1897.
Sepals triangular-ovate, entire, shorter than the hypanthium.

37. Centropogon roseus Rusby, Bull. N. Y. Bot. Gard. 8:

123. 1912.
Sepals broadly ovate, foliaceous, denticulate.

IV. Species of uncertain position, specimens of which have not been seen.

38. Centropogon magnificus Zahlb. & Rechinger, Med. Rijks

Herb. 19: 50. 1913.
Evidently a member of the first group of this conspectus;
compared by the authors with C. Mandonis Zahlb., from which it
differs in its fewer flowers on longer pedicels, straight-edged sepals,
denser pubescence on the corolla, obviously lignified twigs, and
stiffer rough leaves.

39. Centropogon Herzogi Zahlb. & Rechinger, Med. Rijks

Herb. 19: 49. 1913.
Evidently a member of the second section of this conspectus,
as indicated by the description of the corolla; flowers axillary,
hypanthium globose or subglobose, sepals short, acute, broadly
triangular, anther-tube glabrous, except the terminal tuft.

Gleason: Centropogon and Siphocampylus 201

40. Centropogon cardinalis Zahlb. & Rechinger, Med. Rijks,
Herb. 19: 51. 1913.

Evidently a member of the third group of this conspectus,
possibly related to C Bangii Zahlb. ; inflorescence a short terminal
raceme, hypanthium globose, sepals short, triangular, entire,
anther-tube glabrous, except the terminal appendage, leaves
obovate-elliptic.

New York Botanical Garden

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 23-1

STERILITY AND FERTILITY
IN HEMEROCALLIS

By A. B. STOUT

NEW YORK
1921

Reprinted, without change of paging, from Torreya, 21: 57-62. July-August, 1921

[Reprinted from Torreya, Vol. 21, No. 4, July-August, 1921.]

STERILITY AND FERTILITY IN SPECIES OF
HEMEROCALLIS

For a number of years problems of fertility and sterility in
numerous species of plants that freely and naturally propagate
by vegetative means have been under investigation by the writer.
In these studies several species of Hemerocallis have received
considerable attention, and a brief preliminary report of the re-
sults obtained with them has been given (Journal N. Y. Bot.
Garden 20: 104-105, May, 1919). Certain aspects of the re-
search await the blooming of seedlings that are now being
grown and the testing of wild plants of several species which
it is hoped can be obtained from their native home in the orient.
But the results already obtained, and in part published, sup-
plement and to some degree extend the observations recorded
in a recent number of the Torreya (21 : 12-13, Jan. and Feb.
1921) and for this reason may be briefly summarized for the
readers of this journal.

It is to be noted that the double-flowered form of Hemero-
callis fulva reported in Torreya (18: 242) and referred to later
(Torreya 21: 13) is undoubtedly an old and well known sort.
A double-flowered variety of this species is reported by Thun-
berg in his Flora Japonica published in 1784 and there identi-
fied as the double-flowered plant which Kaempfer (Amoen.
Exot. 1712) thought was an Iris. At the present time two
double-flowered varieties are recognized (Bailey, Cyclopaedia of
Horticulture) as belonging to H. fulva. One of these, var.
Kwanso, is illustrated in color in Gartenflora in 1866 (plate
500) and there said to have been introduced into Europe by von
Siebold. This is evidently the double-flowered form most
widely found in cultivation in Europe and America. The other
variety (flora plena) is illustrated in color in Flora des Serres

57

58

( 1 869-1870) and there called H. disticha var. flora plena
(disticha is now considered as a synonym of fulva). The
colored plate shows this to be quite different from var. Kwanso
in appearance. The writer has never seen this variety. A
double-flowered variety has also long been known in the species
H. Dumortierii.

A thorough search of the literature indicates that no one has
ever reported fruit on the single-flowered type of H. fulva.
This orange-colored day lily is widely distributed over Europe and
America. Its complete failure to produce fruit and seeds has
often been noted. Only one variety of it (var. maculata) ap-
pears to have been involved (probably as a pollen parent) in
the production of hybrids.

In the writer's experiments with this species many intra-spe-
cific pollinations have been made between plants obtained from
such widely different sources as Wisconsin, Michigan, New
York, Vermont and England with complete failure in every
case. The ovaries of flowers thus pollinated do not start to
enlarge, and about J2 hours after the flowers open the entire
flower falls leaving only spurs as shown at a in Fig. 4.

But the pollen of this species used in controlled crossing on
H. flava has given pods (Fig, 2) with seeds and the hybrids
resulting are now being grown. The reciprocal cross between
these two species failed to yield mature pods. Pollen of H.
-fulva on H. minor has given seed but no germination was se-
cured.

Pollen of H. Thunbergii and of H. aurantiaca has been used on
many flowers of H. fulva. Usually the pods begin to form and
seeds start to develop with some of them, but as a rule the
pods fall when about one third mature (& in Fig. 4). In a few
instances, however, mature pods with ripe seeds (Fig. 5) have
been secured, but no germination has yet been obtained in such
seeds. The reciprocals of these crosses likewise produce seed
rarely. From the results of crossing H. fulva with H. flava., H.
aurantiaca and H. Thunbergii it appears that its pollen and ovules
are potent and are able to function in certain relations, but that
the compatibility in these combinations is of a weak grade.

59

The literature gives conflicting reports regarding seed pro-
duction in H. flava. Some investigators have reported plants
of it to be self-fertile, others have reported the plants they have
studied to be self-sterile. Both self-compatible and self-incom-
patible plants have been found among plants of this species
grown in the New York Botanical Garden. Such conditions
are often seen in a species in which self-incompatibility is pres-
ent, especially if the species is propagated by seed (Cichorium
Intybus, Nicotiana Forgetiana, Eschscholtzia californica,
Brassica pekinensis, Brassica chinensis, and others). The most
highly self-compatible plants produce pods in abundance, but
in them are many shrivelled ovules in which fertilization may
not have occurred and seeds in various stages of embryo abor-
tion together with seeds that are fully matured and viable (Fig.
3). This condition is also specially characteristic of plants that
are not fully self-compatible.

A third species, H. Thunbergii, has in the author's experience
proved to be only feebly self -compatible. Very many carefully
made self-pollinations fail (see 6. 7 and 8). but many pods do
mature and these contain some seeds which will germinate.
All the plants of this species which are growing in the New
York Botanical Garden have behaved quite the same, but these
may have all descended from a single parent through vegeta-
tive propagation. A wide range of self-compatibility may be
exhibited by the seedlings which are to be tested as soon as
they bloom.

The type of sterility in these species is, undoubtedly, that
of physiological incompatibility operating between the organs
concerned in sexual reproduction. The readiness with which
these species propagate from pieces of the roots and by rhizomes
has practically eliminated the use of seeds in commercial propaga-
tion. Such a method tends to perpetuate the grade of self-com-
patibility of the original plant which was used. It is possible that
the plants of the single-flowered type of fuk'a now growing in
America and Europe may have all come by vegetative propa-
gation from a single plant which happened to be fully self-in-

60

compatible. According to Clusius (Plantarum Historia, p.
137) this species was commonly in cultivation in middle
Europe as early as 1601. Since then its cultivation has been
extended over large areas of Europe and America, and in many
sections it has escaped from cultivation and is spreading widely,
purelv by vegetative means of propagation.

It can be predicted with confidence that a search in the region
where H. fulva is native and wild will reveal plants that are
producing seed or at least strains that will prove compatible
with the self-incompatible strain now found in the United
States. Focke showed that such a condition as this existed in
Lilium- bulbifcrum. After failing for years to get seed by sell-
ing and crossing plants of this species obtained from various
parts of Germany, he obtained wild plants from the native
habitat in Tyrol and these he found compatible with strains
that previously failed to produce seed.

It has very generally been held that the seed sterility of such
plants as Hemerocallis fulva, Lilium bulbifcrum, Lilium ti-
gr'mum, etc., is "correlative." That is, the vegetative organs of
propagation are conceived to divert and utilize the available
food so that the embryos in seeds are virtually starved to
death during stages in development, or perhaps organs are
so poorly nourished that they do not function previous to fer-
tilization. But evidence is increasing to the effect that seed
production in these plants is relative and depends on whether
fertilizations are compatible, quite as is the case in numerous
species of plants that are naturally propagated only by seeas.
The experimental proof of this is sometimes difficult to obtain
in the plants that are propagated vegetatively.

When self-compatible and self-incompatible plums are tounn
and the latter prove to be highly seed-producing in certain
crosses, as is the case with Hemerocallis flava, the evidence of
incompatibilities is clear. The American strain of Hemero-
callis fulva has sex organs that do function to some extent in
certain inter-specific crosses and will, undoubtedly, produce
abundant seed when it can be tested with stocks from a dif-

Explanation" of Plate
i. Pod of a plant of H. flava; the result of self-pollination showing thai
the plant is self-compatible.

2. Pod on same plant as i ; the result of controlled cross-pollination with
pollen of H. fulva.

3. Seeds from such a pod as shown in 1 and 2 ; some ovules become mere
rudiments of seeds and evidently are not fertilized ; some embryos die during
the development of seeds; some seeds develop fully and are viable.

4. Flowering branch of H. fulva near close of period of bloom, (a) Spur
left when flowers fall. (b) Pod 10 days old, from cross with pollen of
H. Tluinbcrgii, but becoming wrinkled and about to fall. Occasionally such
pods contain one or two partly developed seeds.

5. Mature pod of H. fulva from cross-pollination with H. Thunbergii,
Such pods are rare. As far as known to the writer this is the first time the
fruit of this species has been illustrated.

6. 7 and 8. All from a single plant of H. Thunbergii. All flowers carefully
self-pollinated. Some pods (a) maturing, and yielding a few viable seeds:
some pods (b) becoming much shrivelled ; no good pods on branch shown al
x. Results characteristic of feebly self-compatible plants of this species.

62

ferent seed source. But to obtain these, plants from widely dif-
ferent geographical sections or even wild plants from the native
habitat may need to be secured.

A. B. Stout.

New York Botanical Garden.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 235

PHYTOGEOGRAPHICAL NOTES ON
THE ROCKY MOUNTAIN REGION

X. GRASSLANDS AND OTHER OPEN

FORMATIONS OE THE MONTANE

ZONE OF THE SOUTHERN

ROCKIES

P. A. RYDBERG

NEW YORK
1922

Reprinted without change of paging, from Bulletin op the Toebey Botanical Cldb
48:315-326. February 28, 1922

[From the Bulletin of the Torrby Botanical Club 48: 315-326. 4 March, 1922.

Phytogeographical notes on the Rocky Mountain region

X. Grasslands and other open formations of the

Montane Zone of the Southern Rockies

Per Axel Rydberg

In a previous article* I have discussed the wooded formations
of the Montane Zone. In this article the remaining formations
are taken up. In these, the plants of the plains and foothills
make up a larger proportion of the vegetation than in the wooded
formations. The river valleys, meadows, and hog backs have a
flora much more like that of similar habitats in the foothill region
and even on the plains. The transcontinental elements and those
from the prairie region are much more numerous.

The formations have been arranged so that the hydrophytic
are first taken up and then the mesophytic, the most xerophytic
being left until the last. The species followed by a dagger, "f",
are confined to the Southern Rockies.

A. AQUATIC FORMATIONS

Lakes are not very common in the Montane Zone of the
Southern Rockies. In the higher parts they are mostly glacial,
fed by the snow fields, and contain little vegetation. The brooks
are swift and rarely form lakes or ponds in the higher regions.
Most of the lakes are found in the lower river valleys and " Parks "

* Bull. Torrey Club 47: 441-454. 1920.
[The Bulletin for November (48: 285-314) was issued January 19, 1922.]

315

316

Rydberg: Phytogeographical notes

of Colorado, and here we find most of the aquatic vegetation.
The majority of the phaneorgamic plants are found also in the
plains and prairie regions. To these belong a score of species of
Potamogeton which here are omitted. Besides these the following
species are found:

i. Transcontinental and eastern

Sparganhim angustijolium
Sparganium minimum
Triglochin palustris
Triglochin maritima
Alisma brevipes
Sagittaria latifolia
Phragmites Phragmites
Catabrosa aquatica
Panicularia grandis
Panicularia borealis
Panicularia septentrionalis
Eleocharis palustris
Scirpus validus
Lemna trisulca
Lemna gibba
Lemna minor

Crunocallis Chamissonis

Persicaria coccinea
Batrachium trichophyllum
Batrachium flaccidum
Batrachium Drouetii
Ranunculus Purshii
Sisymbrium Nasturtium-

aquaticum
Callitriche palustris
Callitriche autumnalis
Hippuris vulgaris
Sium cicutifolium
Menyanthes trifoliata
Veronica americana
Utricularia vulgaris
Utricularia minor

Western

Nymphaea polysepala

3. Endemic
Panicularia pauciflora Cardamine injausta\

Agrostis depressa\ Ranunculus intertextus\

Cardamine cordijolia\

B. SEDGE BOGS
The sedge bogs are found where the drainage is poor or in low
places around lakes, ponds, and streams. In most cases the most
numerous plants are species of Car ex, but many grasses also are
intermixed, and in some places they are predominant. Of these
Alopecurus aristulatus and Agrostis alba are the most frequent and,
in the lower part of the zone, species of Calamagrostis. The com-
position of the bog vegetation consists of the following plants:—

Rydberg: Phytogeographical notes

317

i. Continental and eastern

Deschampsia coespitosa
Phalaris arundinacea
Calamagrostis elongata
Calamagrostis canadensis
Calamagrostis hyperborea
Beckmannia enicaejormis
Panicularia nervata
Alopecurus aristulatus
Eriophorum angustifolium
Eriophorum gracile
Carex gynocrates
Carex diandra
Carex disperma
Carex canescens
Carex brunnescens
Carex leptalea
Carex aurea
Carex paupercida
Carex Buxbaumii
Carex aquatilis

Poa interior
Agrostis humilis
Agrostis grandis
Agrostis asperifolia
Agrostis idahoensis
Muhlenbergia comata
Carex simulata
Carex athrostachya
Carex tenuirostris
Carex Raynoldsii
Carex Kelloggii
J uncus Mertensianus
Iris missouriensis
Limnorchis viridiflora
Limnorchis borealis
Rumex hymenosepalus
Alsine strictiflora
Alsine obtnsa

Carex lanuginosa
Carex viridula
Carex rostrata
Carex retrorsa
Rumex occidentalis
Rumex mexicanus
Ranunculus sceleratus
Cardamine pennsylvanica
Parnassia parviflora
Argentina Anserina
Geum rivale
Viola palustris
Veronica Wormskjoldii
Elephantella groenlandica
Gnaphalium uliginosum
Senecio pauciflorus
Lycopodium annotinum
Equisetum arvense
Equisetum pratense
Equisetum variegatum

Western

Alsine calycantha
Ranuncidus alismaefolius
Ranunculus Eschscholtzii
Argentina argentea
Vicia oregana
Vicia sparsiflora
Epilobium brevistylum
Epilobium occidentale
Dodecatheon pauciflorum
Castilleja exilis
Mimulus Lewisii
Galium subtriflorum
Aster occidentalis
Aster Burkei
Erigeron salsuginosus
Gnaphalium sidphurescens
Arnica longijolia

318

Rydberg: Phytogeographical notes

A lopecurus occidentalis
Graphephorum Shearii\
Panicularia pauciflora
Agrostis Rossae
Agrostis melaleuca
Agrostis Bakeri\
Carex pseudoscirpoidea
Carex ebenea\
Juncus truncatus
Juncus Hallii
Sisyrinchium alpestre\
Rumex praecox\
Rumex subalpinus\
Rumex densiflorus^
Ranunculus inamoenus
Ranunculus acriforme
Ranunculus micropetalus\
Caliha rotundijolia\
Aconitum porrectum\
Aconitum columbianum
Clementsia rhodantha
Vicia dissitifolia^
Epilobium ovatijolium\
Epilobium americanum

3. Endemic

Epilobium stramineum
Epilobium Drummondii
Epilobium wyomingense
Angelica pinnata
Primula incana
Dodecatheon philoscia\
Dodecatheon radicatum\
Anthopogon thermalis
Swertia scopulina
Myosotis alpestris
Castilleja sulphurea
Castilleja wyomingensis
Orthocarpus purpureo-albus
Erigeron minor
Erigeron lonchophyllus
Arnica rhizomata
Arnica folio sa
Senecio cymbalarioides
Senecio nephrophyllus
Senecio crassulus
Senecio semiamplexicaulis
Senecio rapifolius
Senecio dispar
Senecio perplexans

C. SPRINGY PLACES

The springs and the wet places in their vicinity usually have
a flora of their own. This consists of hydrophytes, growing either
in the water itself or on the muddy shores. As a rule the plants
are small in size.

1. Eastern and transcontinental

Eleocharis acicularis Parnassia parvifiora

Lemna trisulca Tillaeastrum aquaticum

Lemna minor Callitriche palustris

Lemna gibba Callitriche autumnalis

Sisymbrium Nasturtium- Epilobium adenocaulon

aquaticum Myosotis alpestris

Rydberg: Phytogeographical note

319

Muhlenbergia filiformis
Juncus Mertensianus
Crunocallis Chamissonis
Parnassia fimbriata

Agrostis depressa^
Cardamine cordifolia
Cardamine infausta]
Radicula alpina
Micranthes rhomboidea

Western

Micranthes arguta
Androsace filiformis
Mimidus Langsdorfii

Endemic

Micranthes arnoglossa
Oxypolis Fendleri\
Primula incana
Pleurogyne fontana
Mimidus puberulus

D. SANDY RIVER BANKS

The flora of the sandy river banks and sand-flats along the
water courses is very meager. Besides some of the species which
are enumerated below, under the heading "Sand-draws," there
are found the following which require more moisture: — -

Eleocharis acicularis
Scirpus pauciflorns
Muhlenbergia Wolfii]
Sagina saginoides
Alsinopsis macrantha\
Myosurus aristidatus
Ranunculus reptans
Halerpestes Cymbalaria
Radicula alpina

Tillaeastrum aquaticum*
Gilia calcarea\
A ndrosace filiformis
Androsace diffusa]
A ndrosace subulifera\
Mimidus moschatus
Mimidus puberulus
Limosella aquatica*
Erigeron minor

Subularia aquatica*

E. ALLUVIAL RIVER BANKS
The formations on the alluvial river banks and river flats are
of three kinds. Along the smaller brooks, especially in the upper
portion of the zone, they resemble much those of the Subalpine
Zone, the most important plants being species of Mertensia,
Epilobium and Juncoides. In the lower part of the zone Urtica
gracilis and various composites constitute the greater part of the
vegetation. At middle altitudes species of Delphinium and Aco-
nitum often take their places. Very few of the plants are trans-
continental or eastern.

* Growing mostly in shallow water with sandy bottom.

320

Rydberg Phytogeographical notes

i. Transcontinental

Juncoides parviflorum
Juncoides intermedium
Juncoides spicatum
Urtica gracilis

Amaranthus Powellii
Delphinium midtiflorum
Epilobium occidentale

Delphinium occidentale
Delphinium reticidatum
Delphinium robustum\
Delphinium elongatum\
Delphinium ramosum\
Delphinium cucullatum\
Aconitum insigne
Aconitum lutescens
Aconitum Bakeri\
Epilobium Palmeri\
Epilobium rubescens\
Epilobium stramineum]
Mertensia ciliata
Mertensia Leonardi\

2.

Epilobium adenocaulon
Veronica Wormskjoldii
Artemisia biennis

Western

Rudbeckia occidentalis
Senecio triangularis

3. Endemic

Mertensia brevistyla]
Scrophidaria occidentalis
Rudbeckia ampla
Rudbeckia montana^
Helianthclla quinquenervis
Cirsium Parryi\
Cirsium scopulorum\
Cirsium Centaureae^
Cirsium Eatonii
Cirsium coloradense]
Cirsium oreophilum\
Cirsium foliosum
Cirsium, griseum\

F. MEADOWS

By meadows I mean here the more moist meackws of the richer
bottom lands. There are also grasslands which bear hay-making
grasses in the dryer portions of the valleys, which I have included
in the Dry Valley Formation. The grasses of the wet meadows
consist mostly of species of the tribes Agrostideae and Festuceae.
Those of the latter are mostly of the tribe Hordeae and are found
on the plains also. The list of grasses is given in an earlier article*
and is here omitted.

* Bull. Torrey Club 43: 635-636. 191 5.

Rydberg: Phytogeographical notes

321

i . Eastern

Carex stenophylla
Carex lanuginosa
Carex interior
Juncus Vaseyi
Juncus bufonius
Polygonum ramosissimum
Capnodes aureum

Juncus longistylis
Juncus confusus
Juncodes comosum
Anticlea elegans
Polygonum sawatchense
Polygonum Watsonii
Potentilla Bakeri
Potentilla Nuttallii
Potentilla diver sifolia
Drymocallis glandulosa
Lupinus argenteus
Astragalus striatus

Juncus Jonesii
Carex pseudoscirpoidea
Veratrum tenuipetalum\
Calochortus Gunnisoni
Calochortus Nuttallii
Calochortus acuminatus
Polygonum Engelmannii
Claytonia rosea
Thalictrum venulosum
Thalictrum megacarpum
Arabis rhodantha
Potentilla filipes
Potentilla juncunda
Potentilla pidcherrima
Potentilla plattensis
Potentilla propinqua

AND TRANSCONTINENTAL

Arabis ovata
Draba nemorosa
Vicia trifida
Viola nephrophylla
Viola septentrionalis
Viola pedatifida
Artemisia biennis

, Western

Astragalus goniatus
Viola venosa
Viola vallicola
Oenothera Hookeri
Lavauxia flava
Collomia linearis
Orthocarpus luteus
Antennaria corymbosa
Arnica Rydbergii
A mica Jul gens
Agoseris laciniata

. Endemic

Potentilla Nelsoniana
Fragaria pauciflora
Drymocallis convallaria
Trifolium Rydbergii
Trijolium Kingii'\
Astragalus sulphur escens
Aragallus albiflorus
Lathyrus brachycalyx\
Viola bellidifolia
Primula incana
Anthopogon thermalis
Castilleja brunnescens\
Plantago Tweedyi
Campamda Parryi
Erigeron consobrinus\
Gnaphalium exilifolium\

324

Rydberg: Phytogeographical notes

i. Eastern and transcontinental

Calamagrostis purpurascens Carex Halleri

Poa crocata
Festuca rubra
Carex concinna

Carex Rossii

Ibidium porrifolium
Allium Geyeri
Eriogonum ovalifolium
Polygonum sawatchense
Eurotia lanata
Cerastium strictum
Draba nitida
Sieversia grisea
Sieversia ciliata
Collomia linearis

Ibidium strictum
Draba nemorosa

2. Western

Phacelia sericea
Lappula floribunda
Campanula petiolata
Aster apricus
Macronema suffritticosum
Antennaria pulcherrima
Antennaria anaphaloides
Balsamorrhiza sagittata
Wyethia amplexicaulis
Arnica Parryi
Arnica mollis

Muhlenbergia (3 species) f
Poa (8 species) f
Festuca ingrata
Festuca saximontana
Festuca (3 other species) f
Allium Brandegei
Allium recurvatum
Erythronium parvifiorum\
Erythronium utahense
Cerastium (4 species)
Draba (3 species)
Lesauerella curvipes\
Lithophragma australis]
Potentilla (5 species)
Pachylophus macroglottis
Pachylophus hirsutus

3. Endemic

Amarella scopulorum
Dasystephana affinis
Dasystephana Forwoodii
Tessaranthium (5 species) f
Phlox Kelseyi
Mertensia (7 species)!
— Valeriana (4 species) f
Solidago (3 species)
Aster andinus
Erigeron viscidus\
Townsendia strigosa]
Helianthella uniflora
Tetraneuris leptoclada
Tetraneuris lanigera
Senecio Fendleri

Rydberg: Phytogeographical notes

325

K. HOG-BACKS AND DRY RIDGES

The flora of the hog backs in the Montane Zone, especially in
its upper part, resembles that of the Subalpine Zone; in the lower
regions there are many of the species of the dry plains added. Be-
sides these there are a few restricted to the Montane Zone; nearly
all of these are endemics. The flora consist of the following:

i. Eastern or transcontinental
Juniperus sibirica Bouteloua gracilis

Carex praticola Koeleria gracilis

Carex stenophylla Poa crocata

Calamagrostis purpurascens

Lewisia rediviva
Oreobroma pygmaea
Oreobroma nevadensis
Draba lutea

Lesquerella alpina
Eriogonum (6 species) f
Physaria vikdifera
Physaria acutifolia '
Trifolium nanum
Trifolium dasyphyllum\
Kentrophyta aculeata
Kentrophyta Wolfii\
Aragallus minor
Phlox depressa
Phlox Kelseyi

2. Western

Sedum stenosepalum
Petrophytum caespitosum
Leptodactylon purgens

3. Endemic

Chrysopsis pumila
Townsendia strigosa]
Townsendia montana\
Erigeron ursinus
Antennaria rosidata\
Tetraneuris lanigera\
Tetraneuris brevifolia
Senecio Purshianus
Senecio Fendleri\
Senecio werneriaefolius]

SAND HILLS AND SAND-HILL DRAWS
While the best developed sand and sand-hill draws are found
in the region of the Great Plains east of the Rockies, and in the
desert plains of the Great Basin west of the same; they are found
also in the Submontane and Montane Zones. In the latter the
sand-draws are found mostly in the bottoms of the canyons and the

326

Rydberg: Phytogeographical notes

sandhills in the so-called "Parks" of Colorado and in some of the
broader river valleys. Most of the species characteristic of both
are also found on the Great Plains and are such as have migrated
along the rivers to higher altitudes. The following species are
components of the flora. None of these are transcontinental or
common to the Rockies and the Canadian # Zone. A few are
common to the Rocky Mountains and the Sierra Nevada Region,
but most of them are endemics, either of the Great Plains or the
Great Basin, or both.

Eriocoma hymenoides
Muhlenbergia pungens
Stipa speciosa
Oryzopsis Webberi
Sitanion elymoides
Arenaria confusa
Arenaria polycaulos
Peritoma. serrulatum
Epilobium paniculatum

Gayophytum intermedium
Gayophytum ramosissimum
Gayophytum racemosum
Nuttallia midtiflora
Nuttallia densa
Nuttallia speciosa
Cryptantha Torreyana
Lithospermum mtdtiflorum
Helianthus petiolaris

M. ROCK SLIDES

The rock slide flora of the upper Montane Zone resembles that
of the Subalpine, which has already been described.* In lower
altitudes the following species appear, all of which are endemic to
the Southern Rockies except Viola biflora, which is found also in
Europe.

Elymus ambiguus\
Aquilegia saximonana\
Aragallus Hallii\
Limnobotrya montigena\
Viola biflora^
Pseudopteryxia anisata~\
Pseudopteryxia aletijolia\

New York Botanical Garden.

Polemonium confertum^
Polemonium mellitumf
Polemonium Brandegei\
Pentstemon stenosepalus\
Senecio Fendleri\
Senecio amplectens\
Senecio canovirens\

* Bull. Torrey Club 44: 453. 1917.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 236

STUDIES OF WEST INDIAN PLANTS -X

NATHANIEL LORD BRITTON

NEW YORK
1922

Reprinted, without change of paging, from Bulletin of the Torbey Botanical
Club 48: 327-343. February 28, 1922.

[From the Bulletin of the Torrby Botanical Club, 48: 327~343- 4 March, 1922.]

LIBRA RT

NEW V<

Studies of West Indian plants— X

Nathaniel Lord Britton

60. UNDESCRIBED SPECIES FROM TRINIDAD

Eleocharis savannarum sp. nov.

Rootstocks very slender, elongated; culms filiform, weak,
smooth, 6-15 cm. long, the upper sheath membranous, its mouth
oblique; spikelet ovoid, 3-4 mm. long; scales oblong or ovate-
oblong, about 2 mm. long, obtuse, nearly white with a greenish
midvein; achene trigonous, obovoid, about 0.5 mm. long, truncate;
tubercle low, nearly flat, apiculate; bristles none.

Moist hole on the O'Meara Savanna, Trinidad {Britton 2491).

Eleocharis oropuchensis sp. nov.

Roots fibrous, finely filiform; culms finely filiform, densely
tufted, weak, 5 cm. long or less, the upper sheath membranous,
oblique at the summit. Spikelets terminating culms, and sessile
at the base of the plant; compressed, 2-3 mm. long, ovate, about 6-
flowered ; their scales ovate-oblong, pale or brown with pale mar-
gins and keel, subdistichous, blunt, 1.5 mm. long; style 3-cleft;
achene trigonous, smooth, pale, about 0.5 mm. long, about as long
as the 3 or 4 bristles; tubercle conic, one fourth as long as the
achene.

In mud in sunny, grassy situations, Trinidad; type from Oro-
puche Lagoon (Britton, Hazen and Freeman 1155, March 29, 1920).

I am indebted to Mr. N. E. Brown for comparing this little
plant with West Indian and South American species in the Kew
Herbarium, where he was unable to match it. He remarks upon
its unusual character of having both terminal and basal spikelets,
and compares it with Chaetocyperus Jamesoni Steud. from Guaya-
quil, Ecuador (Jameson 36Q), pointing out important differences,
however.

Rynchospora aripoensis sp. nov.

Perennial by short horizontal rootstocks; culms filiform, tufted,
smooth, erect, 2-4 dm. high, longer than the filiform leaves.
Spikelets few, 2 mm. long, ovoid, acute, i-fruited, sessile in 1 or 2
small clusters subtended by a filiform bract 1-3 cm. long; scales

327

328 Britton: Studies of West Indian plants

ovate, brown, acute; bristles none; achene obovate-elliptic, I mm.
long, smooth, light brown; tubercle compressed-conic, acute, one
third to one half as long as the achene.

Grassy plain, Aripo Savanna, Trinidad (Britton 2934). Per-
haps most nearly related to R. Chapmani M. A. Curtis of the
eastern United States.

Bromelia aurea sp. nov.

Leaves numerous, rigid, linear, long-attenuate, 6-8 dm. long,
about 3 cm. wide, armed with distant curved prickles 4-5 mm.
long. Scape rather slender, shorter than the basal leaves, bearing
several prickle-armed small leaves; inflorescence brownish-floccose,
about 3 dm. long; bracts lanceolate, membranous, acuminate,
2-4 cm. long; bractlets ovate, membranous, strongly nerved,
mucronate, about 1 cm. long; flowers 2 to several together, the
clusters 2-3 cm. apart; sepals similar to the bractlets, about 1 cm.
long; petals linear, bright yellow, about 3 cm. long and 2 mm. wide.

Wooded hillside, near western end of Monos Island, Trinidad
(Britton, Button and Brown 2736). In flower Apiil 4, 1921.
Related to B. chrysantha Jacq. of Venezuela.

Aechmea porteoides sp. nov.

Leaves firm in texture, linear with a somewhat broadened
base, about 8 dm. long, 4-7 cm. wide, sharply acute, the margins
armed with very numerous, approximate, nearly black, slightly
curved spinules 4 mm. long or less; inflorescence paniculate, as
long as the leaves or longer, its ultimate branches spreading, about
6 cm. long, stellate-pubescent, slender, few-flowered; bracteoles
subulate-acicular, 4-6 mm. long; flowers blue, about 20 mm. long;
sepals striate, about I cm. long, with a terminal spinule 2-3 mm.
long; ovary about 1 cm. long, oblong.

On the ground in mountain forests, Trinidad. Type from
Mount Tocuche (Britton, Hazen and Mendelson 1342). In flower
April 5, 1920. Related to A. Fendleri of Venezuela.

Tillandsia viscidula sp. nov.

Basal leaves tufted, linear, 2.5-3.5 dm. long, flat, gradually
narrowed upward, abruptly contracted at the apex and short-
acuminate, the base expanded and dark-blotched. Stem 3-5
dm. high, erect, bearing leaves similar to the basal ones, but smaller,
the upper ones 7 cm. long or less; panicle 3-5 dm. long, few- to
several-branched, viscid; flowers yellow, distant, sessile, about 2

Britton: Studies of West Indian plants 329

cm. long, at first appressed, later spreading, about as long as the
bracts.

On trees, Trinidad. Type from Moruga (Britton and B>oad-
way 24 jo). Related to T. aloifolia Hook.

Alpinia silvicola sp. nov.

Rootstocks rather stout, scaly. Sterile stem 2-3 m. high;
leafy; leaves oblong, thin, glabrous, closely many-veined, 3-5 dm.
long, 6-10 cm. wide, the apex acuminate, the base narrowed, the
petioles 1-2.5 cm. long, the sheaths striate; lower leaves reduced
to thin sheathing scales. Fertile stems about 3 dm. high, pubes-
cent above, with a few, narrow scales 3-4 cm. long; spike dense,
several- to many-flowered, 7-10 cm. long; bracts 1- flowered,
shorter than the flowers; calyx about 1.5 mm. long, pubescent,
its lobes broad; corolla yellow, about 2 cm. long; fruit oblong,
about 3 cm. long, its juice blue-black.

Forests of the northern mountain range, Trinidad. Type from
Mount Tocuche (Britton, Hazen and Mendelson 1301).

Calathea trinitensis sp. nov.

Leaves erect, oblong to oblong-lanceolate, glabrous, the blade
up to 1 m. long and 3 dm. wide, rather shorter than the slender
petiole, rather abruptly narrowed at the base, the apex abruptly
tipped, the midvein prominent, the innumerable lateral veins close
together. Scape glabrous, about 7 dm. high; spike dense, about
2 dm. long; bracts oblique, many- veined, 3-4 cm. long, their
spreading tips acute; flowers yellow, about 3 cm. long, the seg-
ments linear, parallel-veined, acute.

Forest, heights of Aripo, Trinidad (Britton and Freeman 2360).
In flower March 16, 1921.

Ficus ierensis sp. nov.

A tree up to 10 m. high or higher, glabrous. Leaves broadly
elliptic or elliptic-orbicular, subcoriaceous, pinnately 5- to 9-veined
on each side of the rather prominent midvein, rounded at the
apex, cordate at the base, 10-20 cm. long, the stout petiole one
fourth to one third as long as the blade; fruiting peduncles
slender, about 1 cm. long; fruit globose, 12-18 mm. in diameter;
bracts 2 or 3, broad, rounded, nearly as long as the fruit; ostiolum
sunken, about 2 mm. in diameter.

Hillsides in relatively dry districts, Trinidad. Type from
North Post Road (Britton, Hazen and Mendelson 774). Similar

330 Britton: Studies of West Indian plants

to F. crassinervia of Hispaniola, to which species it has been

referred, and also related to F. Urbaniana Warburg, of the Lesser

Antilles.

Ficus arimensis sp. nov.

Twigs stout. Leaf-blades firm in texture, obovate, about 2
dm. long and twice as long as wide, distantly pinnately veined,
the base subcuneate, the apex rounded and short-cuspidate; fruit
subglobose, sessile, about 8 mm. in diameter (immature).

Arima, Trinidad (/. Dannouse). Collected in 1905.

Ficus Mendelsonii sp. nov.

A tree up to 16 m. high. Leaves elliptic to obovate-elliptic,
the blade thin, smooth, 10-15 cm. long, about twice as long as
wide, distantly pinnately veined, triple-veined just above the base,
the base rounded or obtuse, the apex abruptly short-acuminate;
petioles slender, .2-6 cm. long; peduncles slender, 1-2 cm. long;
fruit globose, pale, 10-15 mm. in diameter, the ostiolum deeply
concave; basal bracts 2, triangular-ovate, about 2 mm. long.

Forests in moist or wet districts, Trinidad. Type from the
northern hills between North Post and Maqueripe {Britton, Hazen
and Mendelson 8yg).

Phoradendron chaguaramasanum Trelease sp. nov.

Scarcely forked, the moderate branches with basal cataphyls
only, androgynous?. Internodes short (2 X 10-30 mm.), at first
papillately roughened, quadrangular and somewhat 4-winged, little
flattened. Cataphyls a single pair, nearly basal, deeply notched.
Leaves spatulate-oblong, obtuse to subtruncate, scarcely 1 X 2.5
cm., cuneately subpetioled for 5 mm., fleshy, drying yellow, I- or
obscurely 3-nerved from the base. Spikes solitary, very short
(about 5 mm.), with 2 or 3 very short characteristically 4-fiowered
joints; peduncle about 1 mm. long; scales ciliolate. Immature
fruit subglobose, scarcely 2 mm. in diameter, reddish, verrucose;
sepals yellow, erect, not meeting.

Chaguaramas, Trinidad {Britton 2718, the type, April 4, 1921).

Phoradendron caerulescens Trelease sp. nov.

Pseudodichotomous, the moderate branches with basal cata-
phyls only, androgynous?. Internodes short (2-3 X 10-20 mm.),
smooth, glossy, quadrate, the upper ancipitally dilated to a width
of 4 mm. below the nodes. Cataphyls a single pair, basal, glossy
brown, tubular-bifid. Leaves round-elliptical, submucronately

Britton: Studies of West Indian plants 331

acuminate, 1.5 X 2-2.5 cm-> cuneately subpetioled for about 5
mm., cartilaginous-margined, about 5-nerved, at first delicately
blue-glaucous. Spikes solitary, short (scarcely 15 mm.) with 3 or
4 short 4-flowered joints; peduncle I mm. long; scales slightly
ciliolate. Fruit (immature) small, globose, deeply immersed, ver-
rucose: sepals inflexed.

Chacachacare Island, Trinidad (Britton and Hazen 1726, April
13, 1920), the type; also from the same locality, on Capparis
(Britton, Freeman and Watts 2701, 2708, 1921).

Of the Emarginatae, but with inflexed sepals, and in this, as
well as in its flattened twigs, related to P. Ottonis Eichler of Vene-
zuela.

Seguiera ierensis sp. nov.

A climber with slender twigs and branches, the prickles 2-25
mm. long. Leaves oblong or oblong-lanceolate, 6-16 cm. long,
acute or acuminate at the apex, narrowed or obtuse at the base,
or the upper ones smaller and obtuse, the petioles 2-7 mm. long;
panicles many-flowered, 5-9 cm. long, puberulent; pedicels 2-3
mm. long; perianth-segments unequal, elliptic to obovate, 3-3.6
mm. long, rounded, concave; stamens about 20 with filiform
filaments and linear anthers.

McBean Estate, Carapachaima, Trinidad, April 30, 1918
(type). Forest, Ortoire River, Guayaguayare Road (Britton,
Freeman and Nowell 2527, barren).

Seguiera cordata sp. nov.

Twigs slender; prickles stiff, nearly straight, about 8 mm.
long. Leaves broadly ovate, subcoriaceous, glabrous, 14 cm.
long or less, cordate or subcordate at the base, acute at the apex,
or small ones obtuse at both ends, the venation prominent beneath,
impressed above, the stout petioles 6-8 mm. long; panicles many-
flowered, about 6 cm. long; bractlets ovate, 1 mm. long, about as
long as the pedicels or a little shorter; sepals 1— 1.5 mm. long.

Lobajos near Erin, Trinidad (Trinidad Herbarium Q122, W. E.
Broadzvay, May 14, 1920).

Chrysobalanus savannarum sp. nov.

An irregularly branching shrub 0.5-4 m- high, the glabrous twigs
shnder. Leaves obovate to elliptic-obovate, 1.5-3.5 cm- l°n»>
strongly reticulate-veined above, faintly veined beneath, obtuse
or acutish at the apex, narrowed at the base, the petioles 1.5-2

332 Britton: Studies of West Indian plants

mm. long; flowers few, in small axillary clusters not longer than
the leaves, or these terminating short leafy twigs; pedicels short,
puberulent; calyx silky-pubescent, its lobes short, broad, obtuse;
petals cuneate-spatulate, rounded at the apex, 3-3.5 mm. long;
drupe oblong or narrowly oblong-obovoid, obtuse, about 8 mm.
long and 4 mm. in diameter.

Grassy plain, Aripo Savanna, Trinidad {Britton, Hazen and
Freeman 2005, April 21, 1920).

Acacia quadricostata sp. nov.

A woody vine, up to 8 m. long, the old stems 4-ribbed or 4-
flanged, about 8 cm. thick near the base, the twigs 4-angled, armed
with recurved prickles about 3 mm. long. Leaves 2-pinnate, 6-17
cm. long; stipules wanting; petiole slender, bearing a sessile,
circular gland; rachis glabrous, sometimes bearing a few minute
prickles; rachilla very slender, angular; leaflets 35 pairs or fewer,
sessile, linear-lanceolate, acutish, 6-8 mm. long, about 1 mm.
wide, truncately inequilateral at the base, the midvein somewhat
eccentric, the lateral veins few, glabrous, except for a characteristic
line of yellowish hairs on the underside of the midvein near the
base; legume linear-oblong, flat, densely puberulent, 5-9 cm.
long, nearly 2 cm. wide, narrowed at the base with a short stipe,
acute and short-tipped; seeds orbicular, dark brown, about 7 mm.
in diameter.

Hillside, Chacachacare Island, Trinidad {Britton, Freeman and
Watts 2685), in fruit April 3, 1921. The prickles are persistent
upon the old stems and branches.

Erythrina pallida Britton and Rose sp. nov.

A small tree, 4 meters high ; young growth puberulent; second
year growth glabrate with shining gray bark; spines at base of
leaves stout, reflexed, 6-8 mm. long; leaves large; rachis 2 cm.
long or more, glabrous; leaflets 3, broadly ovate, sometimes
obliquely .so, acuminate, 12-15 cm. long, green above, very pale
beneath, glabrous on both sides at least when mature; inflores-
cence subsessile, paniculate, 1-2 dm. long; pedicels stout, about
1 cm. long; calyx 1.5 cm. long, nearly truncate, except a small
protuberance on the lower side; corolla salmon-colored, narrow,

7 cm. long; fruit with a long slender stipe 3 cm. long; tipped with
a rigid persistent style, 2 cm. long, moniliform, 8-io-seeded; seed

8 mm. long, 6 mm. broad.

Hillsides and roadsides, Trinidad. Type from roadside near
Carenage {Britton 2656), April 4, 1921. Commonly planted as
a fence tree in Trinidad.

Britton: Studies of West Indian plants 333

Elaphrium trinitensis Rose sp. nov.

A gnarled tree, 5 meters high, fragrant; branches glabrous,
often short and spur-like, with brownish bark; leaves simple,
borne at the end of stubby branches, the petiole not winged, 12
mm. long or less, the blade broadly ovate, 2.5 cm. long or less,
entire, glabrous; flowers solitary or in small panicles sometimes
only 3- or 4-flowered; pedicels 3 to 5 mm. long, glabrous; fruit
3-angled, glabrous, 8 to 10 mm. long; nutlet white, 3-angled.

Western end of Monos Island, Trinidad {Britton, Britton and
Brown 2739), April 4, 1921.

Phyllanthus graminicola sp. nov.

Annual, with fibrous roots, glabrous; stem slender, erect,
becoming much branched, 1-3 dm. high, the branches almost
filiform. Leaves oblong-elliptic, thin, faintly pinnately veined,
4-10 mm. long, the apex obtuse, the base more or less narrowed,
the petiole about 1 mm. long; stipules minute; flowers nearly ses-
sile; calyx of the pistillate flowers deeply 6-parted, the linear
segments at length about 1 mm. long; fruit depressed, 1.5 mm.
in diameter; seeds about 0.6 mm. long.

Moist grassy situations at low elevations, Trinidad. Type
from grassy roadside, Carenage (Britton and Hazen 12). In fruit
February 25, 1920. Related to P. carolinensis Walt, of the eastern
United States, which has larger fruit and seeds and broader
pistillate calyx-segments.

Clusia tocuchensis sp. nov.

A tree about 18 m. high, the twigs rather stout. Leaves
borne at the ends of the twigs, obovate, coriaceous, glabrous, about
10 cm. long, 4-5 cm. wide, rounded at the apex, cuneate at the
base, rather finely pinnately veined, the short stout petiole only
about 5 mm. long; fruit globose, 10-15 mm. in diameter; carpels
about 7; styles stout, 3-4 mm. long; stigmas flat, obliquely
oblong, the center depressed.

Forest, Mount Tocuche, Trinidad (Britton, Hazen and Mendel-
son 1247). In fruit April 3-5, 1920.

Terminalia nyssaefolia sp. nov.

A tree up to 20 m. high, the slender twigs glabrous. Leaves
chartaceous, obovate, 7-12 cm. long, 4-6 cm. wide, entire,
abruptly short-acuminate at the apex, cuneate at the base, strongly

334 Brtton: Studies of West Indian plants

reticulate-veined, glabrous and bright green above, pale green,
dull and strigose-pubescent on the veins beneath, the strigose and
cilia te petioles 5-1 o mm. long; fruit compressed, 2-winged, broader
than long, 2-3.5 cm- broad, 1.5-2.5 cm. long, notched at the apex,
at base abruptly contracted into a stalk about 2 mm. long, the
thin striate wings about one half as wide as the seed-bearing part.
Coastal woods, Manzanilla, Trinidad (Britton 2177). In fruit
(fruit fallen) March 9, 192 1.

Combretum trinitense sp. nov.

Vine-like, woody, the stems up to 6 m. long, the branches long
and slender, the twigs glandular. Leaves elliptic or elliptic-
lanceolate, chartaceous, 8-12 cm. long, 5 cm. wide or less, the
midvein prominent beneath, impressed above, the lateral veins
about 7 on each side of the midvein, slender, curved upwardly,
the upper surface glabrous, reticulate-veined, the under surface
impressed-glandular, the glandular petioles 6-7 mm. long; fruiting
racemes 6-10 cm. long; fruit oblong, 10-15 mm. long, the four thin
wings glandular, the pedicels about 2 mm. long.

Hillside thicket, Chacachacare Island, Trinidad {Britton,
Freeman and Watts 26gg). In fruit April 3, 1921.

Myrcia arimensis sp. nov.

A small tree, the slender young twigs rather densely strigose.
Leaves elliptic to elliptic-ovate, chartaceous, 4-6 cm. long, reticu-
late-veined, punctate, dark green and shining above, pale green
beneath, the apex acute, the base narrowed, the strigose or glabrate
petioles 3-5 mm. long; panicles several- to many- flowered, 5-7 cm.
long, their slender branches spreading; flowers nearly or quite
sessile; calyx-lobes rounded; immature fruit globose, about 5 mm.
in diameter.

Arima, Trinidad (/. Dannouse, Feb. 10, 1905).

Eugenia Baileyi sp. nov.

A tree up to 12 m. high or higher, the twigs terete, rather
slender. Leaves coriaceous, glabrous, broadly elliptic, 12-16 cm.
long, 9-12 cm. wide, rounded at the apex, obtuse at the base,
strongly pinnately veined, abundantly punctate, the stout petioles
1-1.5 cm. long; fruit lateral, oblong or ovoid-oblong, about 2 cm.
long and I cm. thick, pale, puberulent, nearly sessile.

Forest, Morne Bleu, Trinidad (Britton, Freeman and Bailey
2246). In fruit March 13, 192 1.

Britton: Studies of West Indian plants 335

Hydrocotyle Hazenii Rose sp. nov.

A delicate creeping plant, rooting at the nodes; petioles slender,
2 to 5 cm. long, pubescent and more pronounced just below the
blade; blade nearly orbicular, 2.5 cm. in diameter or less, glabrous
and paler beneath, somewhat hairy or glabrate above, the sinus
usually narrow, the margins doubly crenate, the crenations low
and broad; peduncle slender, hairy above, longer than the petiole;
flowers in a small compact umbel ; pedicels 1 to 2 mm. long.

Among wet rocks, Maracas Waterfall, Trinidad {Britton, Hazen
and Mendelson 1660, April 10, 1920).

Psammisia recurvata sp. nov.

Stems branched, about 5 m. long. Leaves coriaceous, glabrous,
elliptic-ovate, 10-20 cm. long, 7-10 cm. wide, pinnately 7-veined
from near the base, with numerous curved transverse veinlets, the
venation impressed above, prominent beneath, the apex acute,
the base narrowed, the stout petiole 1-2 cm. long; flowers several
in short axillary racemes ; bractlets ovate, acute, about 2 mm. long;
pedicels stout, recurved, glabrous^, 2-3 cm. long; ovary subglobose,
glabrous; calyx subcampanulate, persistent, about 6 mm. long in
fruit, its lobes rounded, apiculate, their margins thickened ; corolla-
tube about 4 cm. long, red, the short limb white; fruit globose,
about 1 cm. in diameter, many-seeded.

Forest bank, near summit of Mount Tocuche, Trinidad
{Britton, Hazen and Mendelson 12Q4).

Specimens of this plant were mixed by Grisebach with those
of another and the complex described by him (Fl. Br. W. I. 143)
as Thibaudia latijolia Griseb., subsequently referred by Bentham
and Hooker to Vaccinium; it would appear that the name latijolia
should go with this other species, which, however, is not a good
Vaccinium.

Sophoclesia trinitensis sp. nov.

Pendent from forest trees; stem slender, sparingly branched,
short-villous when young, 1-4 m. long. Leaves subcoriaceous,
glabrous or nearly so, ovate or some of them ovate-lanceolate, 1.5-
3.5 cm. long, from about one half to two thirds as wide as long,
triplinerved with a pair of delicate veins at the base, this venation
rather distinct beneath, obscure above, the apex bluntly pointed,
the base rounded or subtruncate, the villous petiole about 2 mm.
long; peduncles filiform, glabrous, somewhat shorter than the
leaves; ovary globose, glabrous or with a few hairs; corolla white,

330 Britton: Studies of West Indian plants

glabrous, about 5 mm. long; fruit blue or mauve, glabrous, about
5 mm. in diameter.

Forests, in wet districts, Trinidad. Type from Mount Tocuche
{Britton, Hazen and Mendelson 13 16). In flower and fruit April
3, 1920.

Sophodesia major (Griseb.) Benth. & Hook., also of Trinidad,
differs in its ovate-orbicular rounded, subcordate leaves and
densely pilose ovary.

Cavendishia Urichiana sp. now

A much-branched, woody vine, up to 6 m. long, glabrous
throughout, the branches slender. Leaves oblong-lanceolate to
oblong-elliptic, somewhat fleshy, dark green above, bright green
beneath, chartaceous in drying, triple-nerved, 10-15 cm. long,
3-5 cm. wide, the nerves rather prominent beneath, slightly
impressed above, the apex acuminate, the base narrowed, the
petiole 8-10 mm. long, rather stout; flowers fascicled on the twigs
or in the leaf-axils, bracteolate, the bractlets ovate, acute, 2-3
mm. long; pedicels slender, 8-20 mm. long; calyx-limb with 5
short acute teeth, expanded above the ovary; corolla-tube sub-
cylindric, about 2 cm. long, scarlet, the limb 4 mm. long, white,
with 5 ovate-lanceolate teeth; stamens shorter than the corolla;
style slender, exserted; immature fruit subglobose, crowned by
the calyx-limb.

Climbing on forest trees, Heights of Aripo, Trinidad {Britton
and Freeman 2364, March 16, 192 1). Specimens were brought
from the same place a few days earlier by Mr. F. W. Urich, Ento-
mologist of the Trinidad Department of Agriculture, and he
guided us to it. The plant from which the type specimens were
taken is wonderfully elegant when in bloom, displaying long
wands of the scarlet, white-tipped flowers, and is locally called
"clove-plant."

Diospyros ierensis sp. nov.

A tree, 12 m. high or higher, the twigs and leaves glabrous.
Leaves chartaceous, oblong to elliptic, 10-15 cm. long, 7 cm. wide
or less, bluntly short-acuminate at the apex, narrowed or obtuse
at the base, reticulate-veined, the venation prominent beneath,
the stout petiole 10-14 mm. long; fruiting calyx 2-2.5 crn- broad,
nearly flat, wrinkled, 4-lobed, the lobes broad and short; fruit
subglobose, about 4 cm. in diameter (not quite mature) ; seeds
about 8.

Britton: Studies of West Indian plants 337

Forest near the summit of Mount Tocuche, Trinidad {Britton,
Hazen and Mendelson 1243). In fruit April 5, 1920.

Specimens collected by J. Dannouse at Guanapo, Trinidad,
showing imperfect flowers about 1 cm. wide {Trinidad Herb. 6415)
are probably referable to this species.

Chrysoprrvllum (?) minutiflorum sp. no v.

A tree about 15 m. high, the young leaf-buds puberulent,
otherwise glabrous. Leaves thin-chartaceous, oblong-lanceolate,
13 cm. long or less, 2.5-5 cm- wide, long-acuminate at the apex,
narrowed at the base, delicately pinnately and reticulate-veined,
the rather slender petioles 10-15 mm. long; flowers green, several
to many in sessile axillary fascicles; bractlets minute; pedicels
very slender, thickened upward, 5-8 mm. long; calyx about 1.5
mm. long, its 3 or 4 lobes rounded; corolla-segments 3 or 4,
minute; stamens 3 or 4.

Forest, Arima, Trinidad {Britton, Britton and Brown 2403).
In young flower March 18, 192 1.

The flowers are too young to enable the generic status of this
tree to be certainly determined. Mr. N. E. Brown has identified
specimens as the same as Crueger's no. 247 in the Kew Herbarium,
from Trinidad, which is a leafy branch and a detached fruit.
The leaves are also the same as those of Trinidad Herbarium
no. 1454, also leafy shoots with the remains of a fruit which was,
apparently, about 2 cm. long, attached, and with some flowers
even younger than those of our 2403.

Evolvulus bocasanus sp. nov.

Perennial, shrubby, branched, 3-4 dm. high, the branches
nearly erect or ascending, strigose. Leaves lanceolate to oblong,
2-5 cm. long, 5-15 mm. wide, strigose on both sides, acute or
obtuse at apex, obtuse at base, pinnately few-veined, the veins
nearly parallel, the petiole 1-2 mm. long; peduncles filiform,
axillary, 3 cm. long or less, i-few-flowered ; bracts lanceolate,
acuminate, 2-3 mm. long; pedicels 2-10 mm. long; calyx strigose,
3-4 mm. long, 5-lobed to about the middle, the lobes lanceolate,
acute; corolla rotate, bright blue, 8-15 mm. broad.

Hillsides, Bocas Islands, Trinidad; type from Chacachacare
{Britton, Freeman and Watts 2674). In flower April 3, 1921.
This is the species recorded from Trinidad by Grisebach as Evol-
vulus alsinoides L., at least in part.

338 Britton: Studies of West Indian plants

Solanum ierense sp. nov.

A shrub about 2 m. high, the slender unarmed branches diver-
gent, glabrous, the young twigs stellate-pubescent. Leaves thin,
ovate or elliptic-ovate, 5 cm. long or less, delicately pinnately
few- veined, sparingly stellate-pubescent above, densely stellate-
pubescent beneath, the apex acute or obtuse, the base obtuse or
narrowed, the slender petioles 4-10 mm. long; flowers solitary or
geminate (rarely 3) in the axils on nearly filiform pubescent
peduncles 2.5 cm. long or less; calyx stellate-pubescent, 5-7-cleft,
the lobes linear-lanceolate, acuminate; corolla white, 5-6-cleft, its
lobes linear-lanceolate, acuminate, 10 mm. long; stamens 5 or 6,
all alike, the anthers somewhat attenuate, 8 mm. long, the fila-
ments 1-2 mm. long; immature berry "globose, about 7 mm. in
diameter.

Hillside, Chacachacare, Trinidad {Britton, Freeman and Watts
2706). In flower and young fruit April 3, 1921.

Solanum Hazenii sp. nov.

A tree up to about 5 m. high with a woody trunk about 1.5
dm. in diameter. Leaves broadly elliptic, rather thin and flaccid,
10-20 cm. long, 8-15 cm. wide, strongly pinnately veined on the
under side, the apex short-acuminate, the base obtuse, the upper
surface finely stellate when young, glabrous when old, the under
surface persistently stellate-tomentulose, the stout, stellate-tomer-
tulose petioles 3 or 4 cm. long; inflorescence stalked, about
equalling the leaves, many-flowered, tomentulose; pedicels short;
flower-buds subglobose, rounded, tomentulose; calyx tomentulose,
deeply lobed, 5 mm. long, the lobes obtuse; corolla white, 20
mm. broad, 5-lobed to below the middle, the lobes broadly ovate,
acute; stamens all alike, the anthers 3 mm. long, about twice as
long as the filaments; berry globose, black, about 8 mm. in diam-
eter.

Hillsides, northwestern parts of Trinidad mainland and on the
adjacent Bocas Islands. Type from Saddle Road {Britton and
Hazen 156).

Solanum capillipes sp. nov

A shrub, about 2 m. high, the slender young branches pilose,
the older branches glabrous. Leaves oblong-elliptic to elliptic-
lanceolate, membranous, all alike, pinnately veined, glabrous,
somewhat darker green above than beneath, 10-15 cm. long,
3-6 cm. wide, acuminate, the slightly inequilateral base narrowed,
the petioles about 1 cm. long; racemes lateral, few-flowered,

Britton: Studies of West Indian plants 339

glabrous; peduncles very slender, 2.5 cm. long or less; pedicels
filiform, spreading, 1-3 cm. long; calyx obconic, about 5 mm.
long; corolla white, deeply lobed, 5-6 mm. broad; stamens all
alike, the oblong anthers obtuse, 1.5 mm. long, nearly sessile;
fruit globose, 6-7 mm. in diameter.

Forests in wet districts, southern Trinidad. Type from forest,
Ortoire River, Guayaguayare Road {Britton, Freeman & Nowell

2521).

Codonanthe (?) triplinervia sp. nov.

A somewhat fleshy, epiphytic woody vine, 1-2 m. long, the
young twigs sparingly pubescent. Leaves ovate or ovate-lanceo-
late, 8 cm. long or less, 2-3.5 cm- wide, glabrous, rather thin,
triplinerved, the apex acuminate, the base narrowed or obtuse,
the slender, sparingly pubescent petioles 6-12 mm. long; flowers
solitary or 2-4 together at the nodes; bractlets linear, pubescent,
acute, 6 mm. long or less; pedicels slender, pubescent, about as
long as the bractlets; calyx 10-12 mm. long, sparingly pubescent
below, deeply cleft, its segments linear-oblong, acute; corolla
3-3.5 cm. long, white, the throat yellow within, its tube slightly
bent just above the gibbous base, about 1 cm. long, the subcam-
panulate throat about 1.5 cm. long, the somewhat spreading
limb with rounded, translucent lobes.

On forest trees, Trinidad; type from Ortoire River, Guaya-
guayare Road {Britton, Freeman and Nowell 2543). In flower
March 25, 1921.

Sabicea trinitensis Standi, sp. nov.

Vine, the stems about 2 m. long, pubescent with dense
short ascending rufescent hairs; stipules 5-6 mm. long, rounded-
ovate, obtuse or subacute, finely ciliolate but otherwise glabrous;
petioles 8-14 mm. long, densely hirtellous with ascending hairs;
leaf-blades elliptic or oblong-elliptic, 9-10.5 cm. long, 4.5-6 cm.
wide, acute or subacute, abruptly contracted at base, hispidulous
above along the costa, elsewhere scaberulous, appressed-pilose be-
neath along the veins, glabrate elsewhere; flowers few, sessile in
the leaf-axils,' subtended by two green bracts similar to the stip-
ules; hypanthium about 3 mm. long; glabrous or with a few ap-
pressed white hairs; calyx-lobes linear, 2-3 mm. long, green,
plane, obscurely ciliolate; corolla appressed-hirsute, the tube 6-7
mm. long, the lobes lanceolate, acuminate, 3 mm. long.

Type in the U. S. National Herbarium, No. 1,059,334, c°l-
lected in a thicket of the O'Meara Savanna, Trinidad, March 22,
1 92 1 {Britton 248 q).

340 Britton: Studies of West Indian plants

The proposed species is most nearly related to S. hirsnta ad-
pressa Wernham, which has been reported from Trinidad. It
differs essentially, however, in the short narrow calyx-lobes and
in the short scant pubescence of the upper leaf-surface.

61. UNDESCRIBED SPECIES FROM JAMAICA
Pilea Maxoni sp. nov.

Stem trailing, freely rooting at the nodes, 3-6 dm. long or
longer, rather densely pubescent, the pubescent branches erect,
6-12 cm. high or higher. Stipules ovate-orbicular, rounded, sub-
membranous, 3-6 mm. long, sometimes broader than long; leaf-
pairs nearly equal; leaves ovate, regularly crenate, 1-3 cm. long,
acute, acuminate, or some of the smaller ones obtuse at the apex,
obliquely obtuse or rounded at the base, 3-nerved from above the
base, sparingly pubescent and reticulate-veined beneath, the upper
surface nearly veinless, with long, flat scattered hairs and some
short ones, the linear glochides largely marginal only; petioles
slender, pubescent, as long as the blades or shorter; pistillate
inflorescence paniculate, slender-peduncled, about 6 cm. long, the
staminate shorter; sepals lanceolate or oblong-lanceolate.

Rocky woods in the Cockpit Country, Jamaica. Type, Maxon
& Killip 1555, from near Mocho, above Catadupa, April 3, 1920;
previously collected near Troy, June 28, 1904 {Maxon 2834), our
specimen of this number barren.

Zanthoxylum Harrisii P. Wilson sp. nov.

An unarmed tree 15-18 m. high, with grayish-brown branches.
Leaves equally pinnate, 2-3.5 dm. long, the petioles and rachis
terete, glabrous or nearly so; leaflets 8-12, alternate or sometimes
opposite, short-petioluled, elliptic-lanceolate to elliptic, 7-15 cm.
long, 4-5 cm. broad, coriaceous, acuminate at the apex, inequi-
lateral at the base, entire, lustrous above, the midvein impressed,
paler beneath, the midvein prominent; inflorescence terminal or
lateral in the axils of the upper leaves, the branches densely
puberulent; staminate flowers: calyx lobes 5, triangular-ovate;
petals 5, lanceolate, 3 mm. long, 1.5 mm. broad, stamens 5,
exserted.

St. George's, Portland, Jamaica, March 30, 1918 {Harris

12878).

Salvia clarendonensis sp. nov.

A spreading shrub with weak stems and branches 2-3 m. long,
the twigs, petioles and inflorescence puberulent. Leaves slender-

Brittox: Studies of West Indian plants 341

petioled, membranous, oblong-lanceolate, serrulate, glabrous
above, puberulent beneath, acuminate at the apex, narrowed or
obtuse at the base, 5-15 cm. long; racemes slender, 7-10 cm. long;
bracts linear-lanceolate, the lower ones sometimes 1 cm. long;
pedicels very short; calyx about 5 mm. long, its ovate cuspidate
teeth shorter than the tube; corolla rose-purple, 1-1.2 cm. long.
Peckham woods, Upper Clarendon, Jamaica (Harris 12787).

Gesneria jamaicensis sp. nov.

A glabrous shrub, up to 3.3 m. high, the young twigs and
leaves resiniferous, the branches terete. Leaves oblong to elliptic,
subcoriaceous, 7-12 cm. long, 2.5-5 cm. wide, acute or acuminate
at the apex, closely serrate except near the entire, acute or acutish
base, pinnately veined, the rather stout petioles 5-10 mm. long;
peduncles axillary, slender, 1 -flowered, about as long as the
leaves; calyx-tube obconic or turbinate, 6-8 mm. long; calyx-
lobes linear-lanceolate, bluntish, 8-10 mm. long, I mm. wide;
corolla crimson or rose-pink, its tube subcylindric, 8-10 mm. long,
its spreading lobes orbicular, strongly veined, erose, about 4 mm.
broad; stamens included; filaments distinct, about 8 mm. long;
anthers nodding; staminodium a little shorter than the filaments;
style glabrous, 2-lobed; capsule broadly turbinate, about 8 mm.
long.

Ipswich, St. Elizabeth, Jamaica {Harris 12509, type); Mul-
grave, St. Elizabeth, Jamaica (Harris 12374).

Nearest related to G. Harrisii Urban, which has a sulphur-
yellow corolla with oval lobes, the calyx-lobes filiform, the leaves
crenate-dentate.

62. UNDESCRIBED SPECIES FROM CUBA

Scleria motemboensis sp. nov.

Annual with fibrous roots; culms solitary or few together,
slender, erect, villous, about 3 dm. high. Leaves narrowly linear,
villous, 1-1.5 mm. wide; inflorescence glomerate-spicate ; glom-
erules of 2 or 3 sessile spikelets; bracts linear, long-ciliate, 3-4
mm. long; achene white, shining, 1.5 mm. in diameter, apiculate,
faintly longitudinally striate, with 4 pores at each side of the sub-
trigonous base.

Small lagoon, Sabana de Motembo, Santa Clara, Cuba (Leon
and Loustalot 9405). Nearest to 5. Lindleyana Clarke of South
America.

342 Britton: Studies of West Indian plants

Ouratea savannarum Britton & Wilson sp. nov.

A glabrous shrub 5 m. high, the slender twigs grayish-brown,
often somewhat flexuose. Leaves coriaceous, elliptic-lanceolate
to elliptic-ovate, 4-8 cm. long, 2-3.8 cm. broad, acuminate at the
apex, rounded at the base, faintly pinnately nerved, the margin
spinose-serrate; petioles short, stout, 3-4 mm. long; inflorescence
terminal, 6-1 1 cm. long; pedicels slender, I-I.5 cm. long; buds
ovoid; sepals oblong-lanceolate to oblong-elliptic, 7-7.5 mm. long,
3.2-3.5 mm. broad, obtuse at the apex; petals obovate, 9 mm.
long, 7 mm. broad, crenulate; anthers oblong-lanceolate or oblong,
2.5 mm. long, subsessile; style 6-7 mm. long; fruit unknown.

Sabana de San Marcos, Santa Clara, Cuba (Leon 9205).

Banara Brittonii Roig sp. nov.

A slender shrub about 4 m. high, the young twigs densely
pubescent. Leaves oblong to oblong-lanceolate or ovate-lanceo-
late, 4-7 cm. long, 1 .5-3 cm. wide, acute at the apex, more or less
inequilateral and rounded at the base, glabrous and shining above,
reticulate-veined and densely pubescent beneath with short,
whitish, mostly appressed hairs, the margin revolute, serrulate;
petioles 5-8 mm. long, slender, with one or two orbicular glands
at the summit; inflorescence terminal, inclined or pendulous, 5-7
cm. long, the branches puberulent, pedicels slender, 3-5 mm. long;
sepals oblong-elliptic to oblong-ovate, 2.5 mm. long, obtuse at
the apex, the margin ciliate; petals oval to suborbicular, about as
long as the sepals, yellow; stamens indefinite; anthers rounded;
style 1.5 mm. long; fruit globose (immature?), 4 mm. in diameter,
becoming black in drying.

Rocky limestone soil, Cayo Mono, inside the swamp near
San Pedro, Isle of Pines (M. Cremata, May 17, 1920).

Psidium Loustalotii Britton & Wilson sp. nov.

A small shrub, with short, grayish, often spine-like twigs.
Leaves elliptic to oval, 2-4 mm. long, 1.5-3 mm- broad, rounded
at both ends, dark green and more or less minutely hispidulous
above, paler and strigillose beneath, coriaceous, the short petiole
0.5 mm. long or less; pedicels axillary, solitary, 2 mm. long;
young fruit ellipsoid, 5 mm. long, 4 mm. in diameter, black.

Sabana de Motembo, Santa Clara, Cuba (Leon & Loustalot

9394)-

Jacquinia Roigii P. Wilson sp. nov.

Shrub or tree?; twigs and branches glabrous. Leaves obovate,
4-8 cm. long, 1.4-2.7 cm. broad, strongly spine-tipped at the apex,

Brittox: Studies of West Indian plants 343

cuneate at the base, 3-nerved, in whorls of 4 or more with distinct
internodes between the whorls; pedicels 3.5 cm. long; sepals
long-ciliate; fruit ellipsoid, 1.7 cm. long, 1.1 cm. broad.
Canete, Oriente, Cuba (J. T. Roig 69).

Tabebuia saxicola sp. nov.

A much-branched shrub, about 3 m. high, the short, stiff
twigs white-lepidote. Leaves simple, entire, oblong or oblong-
lanceolate, 2-4 cm. long, 5-1 1 mm. wide, acute at the apex, nar-
rowed at the base, the upper surface pale green, loosely lepidote
and with very obscure venation, the under-surface reticulate-
veined and densely white-lepidote, the petioles 1-3 mm. long;
flowers solitary or 2 together at the ends of short twigs; peduncles
1 cm. long or less; calyx narrowly campanulate, lepidote, persis-
tent, about 7 mm. long, irregularly toothed; corolla 2-2.5 cm-
long; capsule 4-5 cm. long, about 8 mm. thick.

Rocky top of highest mogote near Sagua la Grande, Santa
Clara, Cuba {Leon and Loustalot 9477). In both flower and fruit
August 12, 1920.

63. A NEW TOURNEFORTIA
Tournefortia barbadensis X. E. Brown

Stem probably scrambling, with the young parts minutely ad-
pressed-puberulous. Leaves very spreading; petiole about 3
lines long; blade 1^-2^ in. long and 6-10 lines broad, lanceo-
late, gradually tapering from about the middle to an acute apex,
subacute or slightly rounded at the base, glabrous above, thinly
and microscopically puberulous beneath. Cymes lateral below
the ends of the branches, 2-3^ in. in diameter, on peduncles 3-5
lines long, lax, with the primary and sometimes the secondary
forkings subhorizontally diverging, then branching in a zigzag
manner, minutely adpressed-puberulous; branches about 3 lines
apart and ^3 in. long, very slender, variously curved, with the
small slender flowers about 1 line apart. Pedicels ^4-^ line
long. Calyx very small, about yi line long, lobed almost to the
base; lobes subulate or deltoid-subulate. Corolla minutely pu-
berulous outside, with a rather slender tube about \]A, line long,
swollen at the upper part, and filiform lobes 1 line long. Ovary
and style glabrous.

Barbados: Middle School, Christ Church, Bovell & Freeman
404, and without precise locality, Lane 428 (in Herb. Kew), type

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 237

CYCLIC MANIFESTATION OF STERILITY

IN BRASSICA PEKINENSIS AND

B. CHINENSIS

By
A. B. STOUT

NEW YORK

1922

Reprinted without change of paging, from
The Botanical Gazette, Vol. LXXIII, No. 2, pp. 1 10-13:

CYCLIC MANIFESTATION OF STERILITY IN BRASSICA
PEKINENSIS AND B. CHINENSIS

A. B. Stout
(with SEVEN figures)

The transition from asexual or vegetative growth to the con-
dition of flower and fruit production in hermaphrodite plants is
to be recognized as a most fundamental aspect of sexuality.
Furthermore, the inter-relations that exist between vegetative and
reproductive vigor and the influence of the former on the latter
are reflected and exhibited in certain phenomena of sterility.

It is now certain that vegetative vigor and the internal inter-
relations incident to it may limit reproductive vigor and sexuality.
The limitation from these causes may take place in two ways: (i)
they may interfere with or influence the morphological development
of flowers or other reproductive organs, and (2) they may affect the
functioning powers of organs that are fully formed. If these influ-
ences are marked, one or more types of sterility may appear.

Only recently observations have indicated that, at least in
some cases, the compatibilities and the fertility of the sex organs
may vary rather definitely within the cycle of vegetative and
reproductive development characteristic of the particular species.
A phenomenon of this sort is reported by East and Park (4),
who found that in the few plants which are self-compatible in
certain species and hybrids of Nicotiana, the self-compatibility
develops only at the end of the flowering period. Cases of cross-
compatibility appearing only at the end of the period of bloom are
reported also. A very decided case of the development of self-
compatibility only at the close of the period of bloom was observed
by the writer, in a plant of Lythrum Salicaria, and reported at the
annual meeting of the Botanical Society of America for 191 7.
These observations suggested that new evidence on the old problem
of the relation between vegetative vigor and reproductive vigor,
as expressed in the formation of flowers and the functioning of the

Botanical Gazette, vol. 73] [no

I922] STOUT— STERILITY in

parts in seed formation, may be obtained by experimental means
from a study of the fluctuations in fertility that are to be seen in
those feebly self-compatible individuals which are to be found in
species in which general sexual incompatibilities are strongly
developed.

A subsequent report of a more detailed study on this problem
(Stout ii), however, showed that in Verbascum phoeniceum,
Eschscholtzia calif omica, and Cichorium Intybus the various grades
of self-compatibility operate very uniformly throughout the entire
period of bloom, and that there are in the feebly self-compatible
plants of these species no specially marked tendencies to self-
compatibility at any definite phase of the blooming period. It was
also found that in Nicotiana Forgetiana Hyb. Hort. and in Lytkrum
Salicaria end-bloom self-compatibility develops as an infrequent
individual variation rather than as a phenomenon characteristic of
the self-compatible plants. In these species there is no cyclic pro-
duction of fruits and seeds which would indicate a general relation
between vegetative vigor and the development of self-compatibility.

Such a cyclic occurrence of self-compatibility was found,
however, and reported for Brassica pekinensis, and it was noted
that the highest degree of self-compatibility attained by any given
plant appeared very uniformly during the period of mid-bloom.
Further studies with this species have since been made which show
this to be the rule for all those individuals that are self-compatible
in any degree. Similar behavior has also been found in cultures of
Brassica chinensis and in hybrids between this species and B.
pekinensis. So far as known to the writer, this is the most uniform
and definite case of a general and definite periodicity in the modifi-
cation of sexual compatibilities within a blooming period. In
these species, also, flower abortion appears in the transition of
vegetative to reproductive vigor, exhibiting an influence of vege-
tative vigor on the morphological development of flowers.

Material and methods

Several strains of the "head" sorts of Brassica pekinensis,
commonly known as Chinese cabbage or Pe-tsai, and one strain of
the loose-leaved or headless sort (the Nanking variety) were grown

112 BOTANICAL GAZETTE' [February

from seeds furnished by the Office of Foreign Seed and Plant
Introduction of the United States Department of Agriculture.
The seeds of the strain of B. chinensis which have been grown were
obtained from China by a Chinese student at Columbia University
for the gardener in charge of the greenhouses belonging to the
University, and the writer obtained seeds from the first lot of plants
there grown. From controlled cross-pollinations between plants
of the two species, seeds were obtained and plants of an Fj hybrid
progeny were grown.

The greater number of plants have been grown in pots in a
greenhouse and brought into bloom during the winter and spring
before they could suffer from the heat of summer. When thus
grown, plants of the head varieties of B. pekinensis form a rather
loose headlike rosette, much smaller and less compact than when
grown under field culture, after which they shoot up into flower.
Plants of the Nanking variety of B. pekinensis and plants of B.
chinensis do not form a head even under the best of field culture;
a very loose rosette of leaves develops, and this grades up into the
leaves of the flowering stem. Pot grown plants of these species
closely resembled field grown plants except that they were smaller.

A few plants of all strains have been grown to full maturity
in the garden, both as spring and autumn crops. Such plants were
larger than the pot grown plants and more flowers were produced,
but their behavior in respect to fertility and sterility was identical
with that of plants grown in the greenhouse. Special effort was
taken to make controlled self-pollinations throughout the entire
period of blooming.1 Numerous plants have bloomed alone or in
isolation from other species of Brassica, both in the greenhouses and
in the field, and hand pollinations were made from one to four days
apart as long as the plants bloomed. The plants not grown in iso-
lation were "bagged." Flowering branches were inclosed in glassine
paper bags on or within a day or two following the opening of the
first flowers; the plants were visited at least twice a week (at first

1 During the winter season of 1919-1920 the writer was greatly aided by the
voluntary assistance of Mrs. Mortimer J. Fox. Miss Hester M. Rusk has assisted
in the research and made many of the pollinations of the plants grown in 1920-1921.
Through this efficient assistance and cooperation the many pollinations necessary to
the research were accomplished.

i922] STOUT— STERILITY 113

many were visited daily), and pollen from fully dehiscent stamens
in liberal amounts was placed on pistils of all freshly opened
flowers. After the cyclic modification of self-compatibility was
recognized, frequent cross-pollinations were made to test the
functional power of pistils and stamens during the periods of self-
incompatibility preceding and following the period of mid-bloom.
The potency of the pollen has been studied by germination tests,
and a cytological study of the phenomena of pollen tube growth
and fertilization is under way.

Sterility in Brassica pekinensis and B. chinensis

Three distinct and quite different types of sterility are in
evidence during the period of bloom in both these species.

I. One type is to be classed in general with impotence (Stout
12), but here two very distinct types of impotence may be observed.
These may be described as (1) flower abortion of the first flowers,
and (2) arrested development of the last flowers that start to form.

II. In some plants of both species, axial proliferations develop
from the pistils of many flowers, and the pistils of such flowers are
functionless in respect to fruit production.

III. Among the flowers that open fully and are capable of
functioning in certain relations, various grades of incompatibilities
are in evidence, and self-compatibility whenever present is most
strong during the period of mid-bloom.

I. Impotence

(1) Flower abortion. — Frequently in Brassica pekinensis the
first flowers on the main stalk and often also the first flowers on
laterals are completely aborted. The flower buds remain small and
do not open, but soon become dead and black. A rather character-
istic case of flower abortion is shown in fig. 1, which is of a pot
grown plant of B. pekinensis blooming in February. Nearly forty
of the first flowers of the main branch aborted, after which the
flowers were completely formed, normal in appearance, and fully
capable of functioning in certain fertilizations. The uppermost
lateral branch coming into bloom later than the main axis had only
five flowers that aborted, and the next lower lateral had no aborted

114

BOTANICAL GAZETTE

[FEBRUARY

flowers. On such plants the lateral branches which come into
bloom at the time when the main branch is producing normal

flowers as a rule have normal
flowers from the first (figs. 1-3).
This coordination between
flowers opening simultaneously
on different branches as to kind
of development is very marked.
At first view, this abortion of
flowers appears to resemble the
blasting of flowers which fre-
quently occurs in all sorts of
plants as the direct effect of un-
favorable environmental influ-
ences, but here the phenomenon
is due primarily to internal
conditions. As grown in the
various cultures, the plants
came into bloom at various
times, some were producing
mid-bloom and potent flowers,
while other plants by them and
just coming into bloom showed
flower abortion. The abortion,
therefore, is essentially self-
induced.

Flower abortion of the first
flowers is the rule among plants
of the varieties of B. pekinensis
which form leafy heads and
which are grown in the field
under conditions which favor
the development of heads. In
such plants, if left to bloom,
the flowering branches are at
first inclosed within the head.
At the time when the flowering branches first come to the light,

Fig. 1. — Typical case of flower abor-
tion in plant of Brassica pekinensis; about
forty of first flowers on main axis abor-
ted; there are five such flowers on upper-
most lateral and none on second lateral,
showing correlation in morphological
character according to time of blooming.

1922]

STOUT— STERILITY

"5

they are somewhat blanched and tender, and the first flowers are
already aborted. This condition of itself suggests that, in this

**

-D

O w

a- 1.

1

c«

<u

— c
_. '-3 'St

£ 18.

Sz o _

I 2

rt p -G

S <= S
^ o rt

O T3 S

_£ .2 c "S
w 3 2 j?

> ~ o O

~o -r

= 5 >

*3 f °

"2 & «

Hi O S?

case, temperature and light conditions are important factors which
are concerned in the abortion of the first flowers.

116 BOTANICAL GAZETTE

[FEBRUARY

Flower abortion, however, is quite pronounced in many plants
of B. pekmensis grown in pots and in which the head is scarcely

5 £

15 °

E o

I S
.5 ea

o ,9

4) O)

<c c

.3 -o

to rt

.9 .S

£ o

o o

= -b

1 ■§

.5 « ^ V ^ ^ V * rt

1 S ^j. N N «V ** Or) 0r> o J-

developed. In the Nanking variety in which the leaves form only
a loose rosette, about one-third of the plants grown have had some

i922] STOUT— STERILITY 117

aborted flowers, but usually only a few of the first flowers abort.
Flower abortion also appears in many plants of B. chinensis which
have very loose rosettes of leaves.

In these species flower abortion occurs as a transitional stage
between a period of vigorous vegetative vigor and a period of flower
formation and seed production. The plants which exhibit abortion
are not able to pass at once into complete reproductive activity
in producing potent flowers. The amount of abortion is greatest
in the varieties of B. pekinensis in which vegetative vigor is most
marked and in which excessive vegetative growth can readily be
induced by good cultural conditions and which have been selected
and bred for this feature. Flower abortion occurs in numerous
plants of these sorts that are grown in pots, as it does in many
plants of the loose-leaved kinds, but it apparently tends to be less
marked in these.

Flower abortion is here undoubtedly correlated with the degree
of vegetative vigor. It is not merely due to a stifling of flowers
from simple direct injury because of inclosure within a head,
however, but to a constitutional feature of which the formation
of a leafy head or rosette is an extreme expression. In this sense
the abortion of flowers is self -induced and to some degree hereditary.
Usually the transition from aborted flowers to apparently normal
flowers is sudden and complete (fig. 1). Sometimes, however, the
first flowers to appear after the aborted ones, or the first flowers
when there are no aborted ones, are poorly developed, are plainly
immature and undersized, and especially in B. chinensis there may
be premature opening.

(2) Arrested development of last flowers. — At least some
of the last flowers which begin to form remain immature and func-
tionless. In the first of such flowers the corollas wither quickly
and may become dry and papery without falling. Then the flowers
become smaller in size and more incomplete in development until
at last they are mere stubs of tissue. Usually from six to ten
flowers in these various stages of arrested development may be
counted at the tip of each branch. On short, lower, lateral branches
and on secondary or later laterals all of the flowers may fail to de-
velop fully. This condition is shown in the illustrations. The dis-
tribution of flowers that fail to develop is indicated in figs. 2 and 3.

Ii8 BOTANICAL GAZETTE [February

This type of sterility, of course, is very common in all sorts of
flowering plants, and is clearly associated with old age and death
of the entire plants or of the individual flowering branches. In
these species of Brassica it is unusually conspicuous, and begins to
develop when growth has ceased and parts of the plant, especially
the basal leaves, are dying or even dead and falling from the plant.
Flowers that have aborted or developed poorly at the beginning
of the period of bloom, and those in which development is arrested,
are all functionless. Their failure to produce fruit is entirely
independent of any sort of fertilization. It is clearly due to
impotence.

II. Proliferation

In a few plants of several strains of both Brassica chinensis
and B. pekinensis, noticeable axial proliferations develop. The
axis anlage inclosed within the carpels of the pistil grows and
branches until it bursts open the pistil. The pedicel of the flower
enlarges; the proliferated branch may become several inches long
and bear as many as twenty-five flowers, many of which are able
to function in seed production. Proliferation may be regarded as
the sterilization of a pistil by vegetative growth of the tissue
beneath and within it. In the end it is the expression of a tendency
to vegetative vigor which culminates in the production of many
more pistils and stamens.

Although proliferation is often irregular in its distribution, it is
most frequent during the earlier portion of the period of bloom.
Frequently it is most highly developed in the first flowers of plants
which show little or no flower abortion, but it often does appear
later. The last flowers of those which open normally as a rule are
free from proliferations. This abnormality is certainly to be
regarded as an expression of excess vegetative vigor, as a result of
which the axis about which flower parts are grouped resumes
active vegetative growth. The stamens in many of the flowers
whose axes proliferate seem to be normal, but the pistils are not
productive of fruit.

Another type of excessive vegetative vigor is seen in the develop-
ment of green leaves at the base of each flower, giving a leafy
inflorescence. This has only been observed in a few plants, and

IQ22] STOUT— STERILITY Iig

its possible relation to the production of flowers and to their
impotence has not been determined.

III. Physiological incompatibility

During the phase when the flowers are completely developed,
many flowers are produced that are capable of producing pods and
seeds. In general the plants produce such flowers in abundance,
in succession for a period of about twenty days, and with con-
tinuous and rather rapid elongation of branches (cf . fig. i with 4 and
6). A free and indiscriminate functioning of the organs in seed
production, however, is decidedly limited by incompatibilities in
fertilization.

Self-incompatible plants. — Plants may be completely self-
incompatible throughout, as was the case for the plant shown in
fig. 4. The first six flowers on the main branch aborted, but the
very first flowers on the three uppermost laterals were normal.
A few flowers at the ends of the branches failed to develop. In all,
about seventy flowers on the main branch, forty on each of the first
and second laterals, and fifty on the third lateral were capable of
functioning. Three lower branches, which bore together about
one hundred normal flowers and were like the third lateral in general
appearance, were not included in the photograph. This plant grew
in isolation in a greenhouse, and self-pollinations were made by
hand at least three times a week throughout the period of bloom.
At least two hundred flowers were carefully self-pollinated, but
not a pod resulted. The pods which were formed on this plant
were all from compatible cross-pollinations. Six fine large pods
near the base of the main stem were all from flowers that opened
rather early; the two first flowers on the first lateral yielded fine
pods to a cross; and large pods containing viable seeds were
obtained by crossing some of the very last flowers to open normally
on the main stem and on the first and second laterals. The stamens
were apparently normal throughout the time when flowers opened
normally; pollen from many stamens examined at different times
was found to be plump and normal in appearance, and the use of
pollen in certain crosses covering the entire period of normal
bloom resulted in seeds. Such tests have been made repeatedly on

120

BOTANICAL GAZETTE

[FEBRUARY

numerous plants with results as noted, which show that the failure
to set seed to self-pollination is due to a sexual incompatibility
between reproductive elements that are capable of functioning in

certain other relations.

A total of 13 7 1 plants of
these two species of Brassica
and hybrids between them have
been tested at the time this is

Fig. 4. — Plant of B. pckincnsis, com-
pletely self-incompatible but producing
good pods containing viable seeds to com-
patible crosses at any time while flowers are
fully developed.

Fig. 5. — Feebly self-compatible plant
of B. chinensis; first two flowers and
last to open normally on main axis highly
fertile in compatible cross.

written, and of these 653 were found to be completely self-
incompatible. Plants were thus classed when no pods developed
to selfing at any time throughout the entire period of bloom.
There were, however, several grades to be seen in respect to the
length of time the pistils remained attached to the plant. In the
plant shown in fig. 4, with few exceptions, the pistils of flowers

1922]

STOUT—STERILITY

selfed fell soon after the petals had fallen. In other plants, and
especially plants of B. chinensis, nearly all the pistils of selfed
flowers remained attached only during the period of mid-bloom.

Self-compatible plants. — A total of 718 plants of the various
cultures grown were self-compatible in some degree. For the

Fig. 6. — Plant of B. pekinensis, highly self-compatible during period of mid-
bloom; no flower abortion; first flowers to bloom on laterals were self-compatible,
showing correlation with main branch in physiological character according to time
of bloom.

purpose of a general classification the self-compatibility was judged
as feeble, medium, and strong, but there were many grades within
each class with no sharp distinctions between them. The weakest
grade includes the cases in which, most typically, a few small pods
containing only aborted seeds developed. In some cases pods of
good size were formed, as is shown in fig. 5, but the seeds were all

122

BOTANICAL GAZETTE

[FEBRUARY

shriveled and not viable. Such plants were classed as feebly
self-compatible. Plants whose self-compatibility was classed as

medium produced some viable
seeds. The number of pods, the
number of shriveled seeds, and
to some extent the number of
viable seeds varied greatly
among plants thus grouped.
The plants classed as strongly
self-compatible produced nu-
merous pods, and the total
number of viable seeds was
high. In these also the num-
ber of pods, their size, and the
numbers of viable and of shriv-
eled seeds varied greatly. The
various grades of self-compati-
bility were seen among sister
plants that were as nearly iden-
tical as is possible in regard to
vegetative vigor, number of
branches and flowers produced,
and as to calendar dates for
period of blooming.

It was readily recognized
that the self-compatibility of
such plants was most strong
during the period of mid-bloom,
and that previous to and fol-
lowing this period there was
complete self-incompatibility.
A highly self-compatible plant
grown in isolation and carefully
self-pollinated from day to day
appeared at the end of the
fruiting period, as shown in figs. 6 and 7. Repeated tests by
crossings showed that in these, as in the case of the self-incompatible

Fig. 7. — Plant of B. chinensis; no flower
abortion; plant highly self-compatible;
showing cycle of self-compatibility with
climax at time of mid-bloom.

i922] STOUT— STERILITY 123

plants, the flowers that opened normally during the time of self-
incompatibility were functional in compatible crosses. The re-
sults of a test of this sort are shown in fig. 5. The cyclic develop-
ment of self-compatibility with its coordination among the various
branches of an individual according to time of bloom is shown in
figs. 2 and 3.

There is evidence from other species (Sirks 10) that various
grades of cross-compatibility may exist between the individuals of
the same race or species, that the group relations may be variable
in different cultures of the same species or race (East and Park 4),
and that in general cross-incompatibilities appear with much
the same irregularity in heredity and in expression as do self-
incompatibilities.

The writer's studies with these plants have been chiefly con-
cerned with self-compatibility. In the species of Brassica studied
the cross-relations have not been studied sufficiently to state with
certainty whether the grades of cross-incompatibility undergo
cyclic changes like those of self-incompatibility, but perhaps it
may be assumed that certain of the weaker grades of cross-
incompatibility do thus operate.

Heredity of mid-bloom self-compatibility in pedigreed
lines of descent

In the first or "parent" series grown of both Brassica pekinensis
and B. chinensis, of a total of 253 plants there were 21 plants that
produced viable seeds to self-pollination during the period of
mid-bloom. From such seeds pedigreed progenies were grown
through two generations, to test the inheritance of self-compatibility
and to determine the result of repeated selection for this character.
A summary of the records for the various series and families grown
to date is presented in table I. In these records the first series of
plants grown are given arbitrary numbers. The number of a series
with that of the self-compatible plant used as a seed parent is
employed in designating the series of succeeding generations.
Thus the line of descent and the relationship of the various series
of sister plants are fully indicated.

In one series of the selfed Fx generation of B. pekinensis (series
1-2) comprising 88 plants, 24 were strongly self-compatible. In

124

BOTANICAL GAZETTE

[FEBRUARY

TABLE I
Records of self-compatibility for families of Brassicapekinensis, of B.c hi nensis,

AND OF HYBRIDS BETWEEN THESE SPECIES

Ancestry, generation, and self-compatibility of parent

Brassica pekinensis

P series i, seeds of S.P.I.

seeds of S.P.I.

seeds of S.P.I.

seeds of S.P.I.

seeds of S.P.I.
P series 34, seeds of S.P.I.
P series 15, seeds of S.P.I.

P series
P series
P series
P series

no. 44892
no. 44935
no. 44291
no. 44312
no. 44292
no. 38783
no. 45187

Brassica chinensis
P series 8

F, series 8-1, parent strongly self-compatible
Fz series 8-5, parent medium self-compatible
Fx series 8-6, parent strongly self-compatible
F! series 8-15, parent strongly self-compatible

series 8-1-13, parent strongly self-compatible

series 8-5-32, parent medium self-compatible ....
series 8-5-39, parent medium self-compatible ....
series 8-15-5, parent strongly self-compatible . . . .
series 8-1, no. 39X8-1 no. 32 (medium X medium)
series 8-5, no. 39X8-5 no. 41 (medium X self-sterile)

Fs

F

F
F
F
F,

F2 series 8-5, no. 30X8-5 no. 17 (both self-sterile

F2 series 8-5, no. 27X8-5 no. 32 (self-sterile X medium)

Brassica chinensisXB. pekinensis

Fx 8-5, no. 23X2 no. 3 (both self- incompatible) .
F, 8-5, no. 19X3 no. 16 (both self-incompatible)
F, 8-5, no. 18X3-1 no. 63 (medium X self-sterile)

Record for progeny

7

So

Fi series 1-2, parent strongly self-compatible
Fi series 2-1, parent strongly self-compatible
Fi series 3-1, parent medium self-compatible

Fi series 1-2-9, parent strongly self-compatible 36

46
60

19

16

£~

114
61

F2 series 1-2-18, parent strongly self-compatible
F2 series 1-2-29, parent strongly self-compatible
F2 series 1-2-38, parent strongly self-compatible
F2 series 3-1-1, parent strongly self-compatible
F2 series 3-1-9, parent strongly self-compatible
F2 series 3-1-32, parent strongly self-compatible 18

19

in
48

57
122

19

34

5°

tn a

H

- -

s.s
z

4-1
4i
28

40

7

7

7

3°

23

15

15

7

19

Self-compatible

4^

14

I

15

0

1

2

0

O

0

O

0

8

3

12

1

2

0

2

1

7

1

1

0

1

0

1

0

2

0

0

0

0

0

0

0

1

0

0

0

2

0

19

5

i922] STOUT— STERILITY 125

the next generation of this family, the progeny of four strongly
self-compatible plants, 275 plants were grown. Of these only 44
were completely self-incompatible. There were, however, only 6
plants as highly self-compatible as the immediate seed parents.
The majority of the plants (182 in number) were feebly self-
compatible and did not produce any viable seeds to selling. This
family, however, was somewhat more highly self-compatible than
was the family derived from plant no. 1 of series 3, in the second
generation of which only 2 plants out of 53 produced viable seeds
to selfing.

Selection for the highest grades of self-compatibility in B.
chinensis was also carried through the second generation. In the
Fx generation, 26 out of 338 plants produced viable seeds to selfing.
In the F2, 5 out of 88 plants produced such seeds. The proportion
of self-compatible plants was low and remained about the same,
not being appreciably increased or decreased in the second genera-
tion. No plants classed as highly self-compatible were found in
the F2, but this may have been due to the proportionally smaller
number of plants grown in this generation.

Four series comprising 86 plants were grown from seeds obtained
by crossing certain plants of the Fj. Of these, 25 were feebly self-
compatible and one plant produced viable seeds. The Ft hybrids
between the two species exhibited the three types of sterility
characteristic of the parent species. There was no indication of
a general impotence of both sex organs (pistils and stamens) such
as often results from hybridity. During the time when flowers
opened normally, branches left to open pollination produced pods
and viable seeds, and about 100 plants of this generation grown
in the field and left to open cross-pollination produced pods in
abundance.

As to mid-bloom self-compatibility, the Fx hybrids were like
the pure bred parents. Relatively few were highly self-compatible.
In one series, derived from crossing a plant of a medium grade
self-compatibility and one completely self-incompatible, of 50
plants, 12 were fully self-incompatible, 14 were feebly self-
compatible, 19 were self-compatible of medium grade, and 5 were

I26 BOTANICAL GAZETTE [February

highly self-compatible. Two series were grown from parents that
were self -incompatible. In one of these all of the 19 progeny were
self -incompatible; in the other series of 34 plants 27 were fully
self-incompatible, 5 were feebly self-compatible, and 2 were self-
compatible of medium grade during the period of mid-bloom.

Summary. — The results obtained in these various pedigreed
cultures show that self-compatibility is a character which is not
directly hereditary. Self-compatibility occurs sporadically in a few
members of these prevailing self-incompatible species. This
character does not breed true. Selection for self-compatibility
does not immediately lead to the establishment of self-compatible
races. Neither is self-compatibility nor self-incompatibility domi-
nant in crosses. There is some indication, however, that certain
races may be secured in which the mode of distribution in respect
to self-compatibility is higher than in others.

Discussion and conclusion

The strains of Brassica pekinensis and B. chinensis studied
were previously selected and bred for excessive leafy growth rather
than for fruit and seed production, yet they are reproduced exclu-
sively by seeds. The vegetative vigor is not in the least utilized
in the development of parts which may propagate the plants
vegetatively. In their habit of growth and bloom, the stage of
sexual reproduction in these plants quickly follows a period of
remarkably vigorous vegetative development, hence these species
are favorable material in which to study the correlative relations
of the asexual or the vegetative phase to the sexual or reproductive
(by seeds) phase in the complete life cycle.

The two types of sterility, impotence (including flower abortion
and arrested development), and proliferation, or the destruction
of a pistil by vegetative growth, as they occur in B. pekinensis and
B. chinensis, are both phenomena associated with the formation
of floral organs. The other type of sterility, physiological incom-
patibility or relative sterility which is present, is concerned with
the physiological inter-relations of the sex organs in the various
processes of fertilization.

These three types of sterility develop and operate in these two
species and in their hybrids in intimate correlation with the cyclic

i922] STOUT— STERILITY 127

alternation of vegetative and reproductive vigor. Flower abortion
occurs normally as a transitional stage between the formation of
green leaves and the production of functional sporophylls. Those
plants which exhibit flower abortion are not able to pass directly
from producing green leaves or leaves with branches at the nodes to
the production of flowers, and flower abortion occurs as a transitional
stage. The abortion of flowers appears in the phase where vege-
tative vigor is waning, but before reproductive vigor is fully in
evidence. There is also a marked agreement among the various
branches of a plant as to the grade of development reached at any
one date of blooming (figs. 1-3), which indicates a definite relation
between the condition causing flower abortion (and also normal
flower formation) and a condition of the plant as a whole. These
phenomena, therefore, have many aspects characteristic of physio-
logical correlation.

The arrested development of flowers at the ends of branches
after a period of vigorous blooming of the plant is obviously due
to an extreme waning of vigor and the approaching death of the plant
as a whole, and of course is a phenomenon prevalent in all sorts of
plants. Axial proliferation from the pistils is to be considered as a
resumption of vegetative growth after the differentiation of the
pistils has been accomplished.

Turning to the functional relations of the sex organs in these
two species of Brassica, at least to the compatibility in self-
fertilization, it is seen that they also exhibit a periodicity on their
occurrence which forms a very definite cycle. A total of 718 plants
that were self-compatible to some degree have now been observed
in these two species and in hybrids between them. With the excep-
tion of a few individuals in which pods developed irregularly, the
maximum of self-compatibility was reached during the mid-bloom
of the plant (figs. 2,3, 5-7). Previous to and following this period,
the self-compatibility grades into complete self-incompatibility or
into a much weaker grade of self-compatibility. Furthermore, the
climax of self-compatibility is remarkably coordinated among the
different branches according to the time of blooming quite as is
the earlier development of flower abortion.

The remarkably uniform development of self-compatibility
during the time of mid-bloom in Brassica chinensis, B. pekinensis,

128 BOTANICAL GAZETTE [February

and hybrids between them, is convincing evidence that the functions
of fertilization are here operating in a cycle of intensities. The
period of mid-bloom may be considered as the time when conditions
are most favorable for fertilization. The cross-fertilizations which
are highly effective both previous to and following the maximum
for self-compatibility are hence to be considered as indicating a
different and possibly a stronger grade of sexual relation. It
seems conclusive that, judged by the functional relations in fertili-
zation, the physiological properties of the sex organs in these
plants vary in a rather definite cycle.

It is clear that self-compatibility as contrasted with certain
grades of cross-compatibility in these species of Brassica is limited
to a specific period following the transition from vegetative to
reproductive activity and limited by the waning senility of the plant
as a whole. Self-compatibility appears coincidently with the
climax of the reproductive activity.

Sexual reproduction itself is generally characterized as a
phenomenon of maturity (Coulter 3). The differentiations of
sex as indicated by anatomical features and by the physiological
compatibilities are perhaps to be considered as a smaller cycle
operating within the larger alternation of vegetative and repro-
ductive phases and subject to the same biogenetic regulation.

In the flowering plants especially, there is great diversity
among species in the relative development of their vegetative and
reproductive habits and in the inter-relations between these two
phases. Perhaps the most universal of the biogenetic conditions
incident to the transition from the vegetative to the reproductive
phase is that change in nutrition which leads to the accumulation
of carbon compounds. This is an internal condition that arises in
the plant as a whole in the course of maturity, in contrast with
relative excess of nitrogenous material that is characteristic of the
vegetative stage. The decided influence of nutritive relations in
regulating development and in influencing fruitfulness has recently
been discussed by Kraus and Kraybill (7), who have emphasized
the fact that a well-balanced development, especially in regard to
fruitfulness of fruit-bearing plants, is associated with a proper
balance between nitrogen and carbohydrate metabolism.

i922] STOUT— STERILITY I2Q

It is not to be considered, however, that a single simple change
in nutrition is the sole biogenetic factor regulating the appearance
of maturity and its attending morphogenesis of flowers. In flower-
ing plants, such as the species of Brassica whose sterilities are
reported in this paper, there is progressive differentiation of parts
in reference to metabolic activities which is most obvious in respect
to the manufacture, distribution, and consumption of food. It has
been shown that there are also special stimulating and inhibiting
influences which in a decided manner regulate and correlate
development. That these influences may be substantive and
special (but metabolic) and different from food materials was
postulated by Sachs (8) in one of the latest of his papers; that some
influences are stimulative and correlative in the sense of nervelike
impulses or even electrical stimuli have repeatedly been shown in
studies of the nature of transmission and excitation in phenomena
of dominance and control in correlative growth and development
(Child i).

It is to be noted that the complete life cycle of flowering plants
involves two periods of vegetative vigor and maturity; one for the
sporophyte and one for the gametophyte. The former culminates
in the production of spores and the latter in the production of
gametes. The generations are antithetic. In its length of life,
vigor of vegetative growth, and reproductive power (number of
gametes), the gametophytic phase has become relatively weak and
highly specialized. In the sporophyte great vegetative vigor is
correlated with great reproductive vigor in the production of spores
(which are, however, in themselves asexual) and in the nurture of
the gametophyte and the embryo. Sex differentiation in the great
group of flowering plants has been pushed back during the progress
of evolution into the sporophytic stage of the entire cycle, and here
sexuality now culminates in seed formation in which the nutrition
of the embryo is a most important factor. Sexual reproduction in
these higher plants has become more and more inter-related with
the vegetative phase of the sporophyte and subject to its internal
and biogenetic regulation.

The decided influence of such regulation is seen in the fact that
in the great group of hermaphrodite plants, the whole trend of the

130 BOTANICAL GAZETTE [February

morphological and physiological differentiations constituting sexu-
ality is initiated in the morphogenesis of flowers. The cells of
pistils and stamens are not only alike in their preformed genetic
composition, but they are identical in this particular with the cells
that entered into the preceding vegetative structures. Correns
(2) has noted that the regeneration from sister cells of the egg and
sperm (the archegonial and antheridial cells) in certain monoecious
mosses shows that, at least in hermaphrodites and monoecious
plants, maleness and femaleness are carried equally by both male
and female gametes. The male gametophytes and their most
highly specialized male cells are male only because of a temporary
suppression of femaleness. Likewise the femaleness of egg cells is a
temporary and one-sided expression of cells carrying both sex
potencies. The various expressions of maleness and femaleness
even in the sex generation, at least in hermaphrodite plants,
according to Correns, are "phenotypic" or biogenetic expressions
independent of any qualitative differentiation in the component
units of the germ plasm. The expressions of the so-called factors
for sex or the so-called inhibitors of one or the other sex are hence
independent of corresponding differentiations in germ plasm which
may have arisen during sporogenesis. The expression of sex,
therefore, is on the same basis as are the somatic differentiations
that arise among the various parts of the individual. It hence
becomes a most fundamental biological problem to consider and
to determine as far as possible what conditions determine these
differences in the level of the so-called ''physiological gradient."

Maturity, with its transition from the vegetative to the repro-
ductive phase, whether giving homologous or antithetic alternation
or a continuation of either, occurs in cycle after cycle with remark-
able uniformity. This emphasizes the phylogenetic or hereditary
aspect of particular phases of the development. One may assume
a " gene " or a "factor " for maturity, and assume that it is gradually
awakened from a dormant condition to the exercise of its influence
at a particular time and in a particular group of cells. One may
further assume that the loss of such a gene would throw a line of
progeny into a condition of perpetual immaturity, so that flowers
or other reproductive organs could never be formed. The evidence

i922] STOUT— STERILITY 13 1

is decidedly against such a view. The loss of maturity, as seen
especially in the complete failure of flower formation, has very
universally been shown to be due to the indirect influence of such
external factors as light, heat, and nutrition on the metabolism
and attending correlations in the organism (see numerous papers
by Mobius, Vochting, Klebs, Sachs, and Goebel, and recent
papers by Garner and Allard 5 and 6, and by Setchell 9).

That species or strains showing flower abortion and physiological
incompatibility are different genetically from others that do not
show such sterilities is obvious. That these types of sterility are
more completely hereditary in some species than in others is clear.
That these characteristics are not definitely and directly represented
as such in the germ plasm by hereditary units is very evident
from the results of genetical studies. Self-compatibility and self-
incompatibility especially are not found to be alternative conditions
in tests by crossing or in line breeding; the heredity is decid-
edly irregular and sporadic even when compatibilities are not
cyclic in their appearance as they are in Brassica chinensis and
B. pekinensis.

The various types of sterility seen in these species of Brassica
decidedly indicate a mutually limiting relationship between vege-
tative and reproductive vigor. Their irregular inheritance, their
appearance at definite periods in the cycle of development of the
plant as a whole, and especially the cyclic manifestation of self-
compatibility, indicate that the morphological and physiological
differentiations of sex are regulated and determined by those
internal and biogenetic processes which in general determine the
cycle of growth, development, and maturity in the life of the
individual.

New York Botanical Garden
Bronx Park, N.Y.

LITERATURE CITED

1. Child, C. M., Certain aspects of the problem of physiological correlation.
Amer. Jour. Bot. 8:286-295. 1921.

2. Correns, C, Die geschlechtliche Tendenz der Keimzellen gemischt-
geschlechtiger Prlanzen. Zeitschr. Bot. 12:40-60. 1920.

3. Coulter, John M., Evolution of sex in plants. Chicago. 1914.

132 BOTANICAL GAZETTE [February

4. East, E. M., and Park, J. B., Studies on self-sterility. I. The behavior
of self-sterile plants. Genetics 2:505-609. 191 7.

5. Garner, W. W., and Allard, H. A., Effect of the relative length of day
and night and other factors of the environment on growth and reproduction
in plants. Jour. Agric. Research 18:553-606. 1920.

6. , Flowering and fruiting of plants as controlled by the length of day.

Year Book U.S. Dept. Agric. 1920 (pp. 377-400).

7. Kraus, E. J., and Kraybill, H. R., Vegetation and reproduction with
special reference to the tomato. Oregon Exp. Sta. Bull. 149. 1918.

8. Sachs, J. von, Physiologische Notizen. VII. tjber Wachsthumperioden
und Bildungsreize. Flora 77:217-253.

9. Setchell, W. A., Geographical distribution of the marine spermaphytes.
Bull. Torr. Bot. Club 47:563-579. 1920.

10. Sirks, M. J., Sterilite, auto-incompatibilite, et differenciation sexuelle
physiologique. Arch. Neerland. Sci. Exactes et Naturelles, B. 3:205-234.
1917.

11. Stout, A. B., Further experimental studies on self-incompatibility in
hermaphrodite plants. Jour. Genetics 9:85-129. 1920.

12. , Self- and cross-pollinations in Cichoriam Intybus with reference to

sterility. Mem. N.Y. Bot. Gard. 6:333-454. 1916.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN -No. 238

STUDIES OF WEST INDIAN PLANTS-X1

NATHANIEL LORD BRITTON

NEW YORK
1923

Reprinted, without change of paging, from Bulletin of the Torres Botanical
Olub 50: 35-56. January, 1923.

Studies of West Indian plants — XI ,OJ<K

Nathaniel Lord Britton *■'''

64. UNDESCRIBED SPECIES FROM CUBA
Juniperus saxicola Britton & Wilson, sp. nov.

A densely branched small tree, 3-8 m. high. Leaves all
subulate-acicular, subimbricate, somewhat spreading, 4-7 mm.
long, about 1 mm. broad at the base, lustrous; fruit blue, broadly
ellipsoid or subglobose, 5 mm. long, 3-4 mm. in diameter.

On rocks, crest of Sierra Maestra, Oriente {Leon ioyg8, type) ;
also collected on Farallones de Regino, Estribo Turquino {Leon
1 1024).

Myrica cacuminis Britton & Wilson, sp. nov.

A shrub, 3-4 m. high, much branched, the slender twigs
densely pilose. Leaves elliptic to suborbicular, 7-18 mm. long,
short-petioled, sharply few-toothed or entire, coriaceous, pin-
nately few-veined, pubescent on both sides, the apex rounded,
the base rounded or obtuse; flowers and fruit not seen.

Pico Turquino, Oriente {Leon 10073, type; Bucher ig).

Pilea ovalifolia Britton & Wilson, sp. nov.

Stems creeping, rather stout, rooting at the nodes. Leaves
chartaceous, elliptic, obtuse or rounded at the apex, rather
coarsely crenate, pinnately 3-nerved, glabrous, those of each
pair often unequal in size; larger leaves up to 5.5 cm. long, 3.4
cm. wide, with petioles 1.5-3 cm- long; smaller leaves 2.5-4 cm»
long, 1.5-2.6 cm. wide, the petioles 1-2 cm. long; upper leaf-
surface densely covered with minute linear raphides, the lower
surface papillose; pistillate inflorescence longer than the leaves,
peduncled, its branches spreading; achenes ovate, about 0.9
mm. long.

Sierra Maestra, Oriente {Leon 10767).

Pilea membranacea Britton & Wilson, sp. nov.

Stem slender, puberulent. Leaves membranous, lanceolate
jy^ to ovate, acuminate, coarsely toothed, 3-nerved, glabrous, those
cvj of each pair unequal in size; larger leaves up to 5 cm. long, 1.8
^2. cm. wide, with slender petioles 0.7-1 cm. long; smaller leaves
G^ 1-2 cm. long, 0.6-1 cm. wide, the petioles 2-4 mm. long; upper
, — 1 leaf-surfaces covered with minute white linear raphids, those of

5 35

36 Britton: Studies of West Indian plants — XI

the lower surface rather inconspicuous; pistillate inflorescence
slender, longer than the leaves; achenes elliptic, I mm. long.

Sierra Maestra, Oriente {Leon 10815).

Pilea micromeriaefolia Britton & Wilson, sp. nov.

Stems elongate, woody, procumbent, the branches erect or
ascending, 4-angled, scaly encrusted. Leaves in whorls of.3's
or 4's, lanceolate, 0.8-1 cm. long, about 2 mm. wide, acute at
the apex, obtuse at the base, glabrous, the petioles 1 mm. long
or less; upper leaf-surfaces covered with linear raphides, those
on the lower surface coarser; inflorescence not seen.

Brecha de Regino, Estribo Turquino, Oriente {Leon 11 008).

Pilea yarensis Britton & Wilson, sp. nov.

An erect monoecious herb, 2-3.5 dm. high. Leaves mem-
branous, broadly elliptic, acuminate at the apex, acute at the
base, crenate-dentate, 3-nerved, glabrous above, often slightly
pubescent on the midvein beneath, those of each pair unequal in
size; larger leaves up to 15 cm. long, 6 cm. wide, with slender
petioles sometimes 3.5 cm. long; smaller leaves up to 11 cm.
long, 5 cm. wide, with petioles 1-2 cm. long; upper leaf-surface
papillose, densely covered with linear raphides, those of the
lower surface often indistinct; staminate inflorescence equalling
or longer than the leaves, the peduncles sometimes 1 dm. long,
the branches spreading, the perianth about 1.5 mm. long;
pistillate inflorescence equalling or shorter than the leaves;
achenes elliptic, 0.8 mm. long.

Bank of the Yara River, Sierra Maestra, Oriente {Leon 10736).

Pilea ermitensis Britton, sp. nov.

Stems slender, densely covered with linear raphides. Leaves
oblong, oblong-lanceolate or elliptic, membranous, acute, entire,
3-nerved, glabrous, those of each pair unequal in size; larger
leaves up to 3 cm. long, 6-7 mm. wide, with petioles 3 mm. long;
smaller leaves 1-1.7 cm. long, 4-5 mm. wide, the petioles short;
raphides of the upper leaf-surface minute, linear-filiform, those
of the lower surface scattered, more conspicuous and stouter;
inflorescence often shorter than the petioles, the flowers borne
in globose heads 1.5-2 mm. in diameter; achenes elliptic-ovate,
about 0.5 mm. long.

La Ermita, Oriente {Hioram 4921).

Pilea crenata Britton & Wilson, sp. nov.

Stem creeping, often rooting at the nodes, the branches erect
or ascending, pilose. Leaves chartaceous, oval to suborbicular,

Britton: Studies of West Indian plants — XI 37

rounded at the apex, rounded or acute at the base, crenulate,
3-nerved, glabrous above, pilose on the nerves beneath, those of
each pair often unequal in size; larger leaves up to 3 cm. long,
2.5 cm. wide, with slender petioles 1.5-2 cm. long; smaller leaves
1-2 cm. long, 1-2 cm. wide, the petioles 0.7-1 cm. long; upper
and lower leaf-surfaces densely covered with linear raphides;
pistillate inflorescence as long as or shorter than the leaves:
achenes about 0.6 mm. long.

Sierra Maestra, Oriente {Leon 10766, type; Leon, Clement and
Roca 9872).

Coccolobis monticola Britton, sp. nov.

Branches slender, glabrous; young twigs pubescent. Ocreae
cylindric, 1 cm. long or less; leaves coriaceous, light green,
3-6 cm. long, reticulate-veined on both sides, shining above,
dull and black-dotted beneath, the apex acute, the base obliquely
obtuse, the petioles about 2 mm. long; racemes very slender,
glabrous, densely many-flowered; ocreolae subtruncate, about
0.5 mm. long; pedicels glabrous, about 1 mm. long; calyx about
1 mm. long.

Sierra Maestra, Oriente {Leon 10713).

Coccolobis saxicola Britton, sp. nov.
A shrub about 3 m. high, glabrous throughout, the twigs
short, stout, more or less tortuous. Ocreae subtruncate, 4-6
mm. long; leaves coriaceous, elliptic to obovate, 9 cm. long or
less, the veins somewhat impressed above, prominent beneath,
reticulate-veined on both sides, the apex abruptly short-acu-
minate, the base narrowed or obtuse, the petioles 5-15 mm.
long; fruiting racemes slender, solitary, 6-9 cm. long; pedicels
numerous, about 1.5 mm. long; ocreolae truncate, less than 1
mm. long, subpapillose, the sepals ovate.

On rocky crest in woods, Loma del Gato, Cobre Range of
Sierra Maestra, Oriente {Leon, Clement and Roca 10167).

Talauma (?) orbiculata Britton & Wilson, sp. nov.

Tree 6-8 m. tall, with glabrous branches. Leaves orbicular,
8-12 cm. in diameter, rounded or subtruncate at the apex,
rounded or truncate at the base, glabrous and prominently
reticulate-veined on both sides, the midrib flat above, prominent
beneath, the petioles rather slender, 3-6 cm. long, flat or shal-
lowly grooved above, glabrous; sepals broadly elliptic, 2.6 cm.
long, 2-2.3 cm. broad; petals thick, oblong or elliptic, about 2
cm. long, 0.8-1 cm. wide; carpels tomentulose.

Loma de Quintin, Nagua, Oriente {Leon 10955).

38 Britton: Studies of West Indian plants — XI

Persea anomala Britton & Wilson, sp. nov.

A shrub or tree, sometimes io m. tall, with appressed-pubes-
cent twigs. Leaves lanceolate to ovate or elliptic, 4-10 cm.
long or less, 1.8-5.5 cm- broad, rounded, obtuse, or acutish at
the apex, rounded or somewhat acutish at the base, glabrous and
reticulate-veined on both sides, the petioles slender, 1-1.8 cm.
long; branches of the inflorescence pubescent with appressed
hairs; calyx pubescent, greenish, its lobes unequal, the outer
broadly ovate, about 1 mm. long, 0.7 mm. broad, the inner
elliptic or broadly ovate, 1.7-2 mm. long, 1.5-1.7 mm. broad;
berry subglobose, 11 mm. long, black.

Sierra Maestra, Oriente (Leon 10707, type; 10975; 10976;
10079; 1 1057).

Persea similis Britton & Wilson, sp. nov.
Twigs appressed-pubescent. Leaves elliptic, 4-7 cm. long,
acute or short-acuminate at the apex, acute at the base, dark
green and glabrous above, the midvein impressed, paler beneath
and pubescent with short appressed hairs, the lateral veins
rather indistinct on both sides; petioles 3-4 mm. long; in-
florescence pubescent with appressed brownish hairs; calyx-
lobes elliptic to oval, 1.8 mm. long, about 1 mm. broad, ap-
pressed-pubescent.

Sierra Maestra, Oriente (Leon 11058).

Nectandra reticularis Britton & Wilson, sp. nov.

A shrub with pubescent twigs. Leaves elliptic-lanceolate,
3.5-6.8 cm. long, 1-2.3 cm- broad, acute to acuminate at the
apex, acute at the base, glabrous, dark green, shining and
reticulate-veined above, paler and reticulate-veined beneath;
petioles 7 mm. long or less; inflorescence long-peduncled, the
peduncle puberulent; flowers short-pedicelled ; calyx white, its
lobes elliptic, 1.5 mm. long, about 1 mm. broad, puberulent;
drupe ellipsoid, about 1. 1 cm. long, 6 mm. in diameter.

Sierra Maestra, Oriente (Leon 10746, type; 10958).

Rubus turquinensis Rydberg, sp. nov.

Stem terete, 1-2 m. high, densely fuscous, villous-tomentose
and with numerous stalked glands, armed with somewhat com-
pressed curved pubescent prickles 4-8 mm. long; leaves of the
branches 3-foliolate, those of the main stem not seen; petiole and
ribs of the leaflets villous-tomentose, glandular, and slightly
prickly; petiolule of the terminal leaflet 7-15 mm. long, those of
the lateral leaflets 1-4 mm. long; leaflets lanceolate, long-acu-
minate, mostly rounded at the base, finely and closely serrate,

Britton: Studies of West Indian plants — XI 39

3-10 cm. long, dark-green, sparingly pubescent and somewhat
glandular-granuliferous above, paler and somewhat fuscous and
densely short-villous, almost subvelutinous, beneath; racemes
3-6 cm. long, 5-10-flowered; peduncles and pedicels densely
pubescent and glandular, slightly prickly; sepals lanceolate,
acuminate, 5-6 mm. long, glandular and pubescent without,
grayish-tomentose within; berry about I cm. long and 7-9 mm.
thick; drupelets many, glabrous.

Type collected on the top of Pico Turquino, Oriente, Cuba,
July 23, 1922, Leon 10755 (New York Botanical Garden); also
at an altitude of 1600 m. on the same mountain, 10756.

R. turquinensis is a member of the Adenotrichi, differing from
R. costaricanus and R. adenotrichos in the narrower lanceolate
long-acuminate, instead of broadly ovate, abruptly acuminate
leaflets; from R. Vera-Crucis in the more dense pubescence and
the texture of the leaves. It is more closely related to R.
irasuensis and R. miser. In habit, leaf-form, and pubescence it
resembles the former, but the inflorescence is mostly simple, the
sepals narrower, and the drupelets perfectly glabrous; neither
have any 5-foliolate leaves been seen. From R. miser it differs
in the shorter sepals and the thicker leaflets, with shorter petioles,
more numerous and more prominent lateral veins, and much
closer serrations.

Polygala scabridula Blake, sp. nov.

Virgate leafy shrub 2-3 m. high, the branches erect, his-
pidulous. Petioles hispidulous, 2 mm. long; leaf blades obo-
vate, 1. 3-1. 8 cm. long, 9-12 mm. wide, broadly rounded at apex,
rounded or cuneate-rounded at base, coriaceous, scabrid-his-
pidulous above with persistent ascending hairs, similarly pu-
bescent beneath; racemes axillary, the axis few-flowered, about
1 mm. long, the peduncle obsolescent; pedicels glabrous, 1.2
mm. long; flowers immature; sepals deltoid-ovate, ciliolate,
obtuse, i.i mm. long; wings similar, 1.4 mm. long and wide;
keel ciliate, otherwise subglabrous, 2.5 mm. long; capsule sub-
quadrate, rather broadly margined, slightly wider at the apex,
ciliolate, otherwise glabrous, 6 mm. long, 8.5 mm. wide, lobed
for about 1/3 its length; seed ellipsoid, sparsely pubescent, 4
mm. long; aril orange, 1.8 mm. deep, with repand lateral margin,
the dorsal lobe 1 mm. long.

Sabana de Motembo, Santa Clara, August 9-10, 1920 {Leon
and Loustalot 9333).

This species is related to Polygala portoricensis (Britton)
Blake, of Porto Rico, which has larger emarginulate leaves,
and a smaller, less deeply lobed capsule.

40 Britton: Studies of West Indian plants — XI

Poly gala rhynchosperma Blake, sp. now

Slender erect annual, 14-25 cm. high, simple or few-branched
above, evenly but not densely puberulous with short, thickish,
incurved or subappressed eglandular hairs. Leaves alternate
except for 1 or 2 basal whorls, rather crowded, short-petioled,
the blades linear, 6-14 mm. long, 0.5-1.3 mm. wide, acute,
cuspidulate, i-nerved, sparsely puberulous like the stem, erect;
peduncles 5-15 mm. long; racemes slender, cylindric, gradually
narrowed to an obtuse apex, loose below, 2-9 cm. long, the axis
becoming 10.5 cm. long and more, puberulous like the stem;
bracts oval or obovate, caducous, 1.5 mm. long, puberulous at
base, with rather short abrupt caudate purplish tips; pedicels
glabrous, 1 mm. long, clavellate; sepals pinkish white with green-
ish center, glabrous, the upper oval-ovate, obtuse, 1.5 mm.
long, the 2 lower oblong, obtuse, 1.2-1.5 mm. long, sometimes
obscurely 1 -denticulate on each side near apex; wings pinkish
(when dried), obovate, 3 mm. long, 1.5 mm. wide, rounded or
obtuse, scarcely clawed, 3-nerved; upper petals slightly shorter
than keel, the free portion obliquely ovate, obtuse, 3-nerved;
keel 3-3.2 mm. long, slightly surpassing the wings, the crest on
each side of a deeply trifid lamella and a 2- or 3-fid or 2-parted
lobe with linear segments; capsule pendulous at maturity,
slightly surpassing the wings, elliptic-oblong, emarginate, 3.3
mm. long, 1.5 mm. wide; seeds subcylindric, 2.2 mm. long,
densely short-pilose, distinctly rostrate at base (beak conic,
0.4 mm. long); aril fastened to upper half of beak, 0.7 mm.
long, the 2 lobes rhombic, acute or obtuse, appressed; stigma
with short blunt retrorse lower lobe and substipitate penicillate
upper lobe.

Sabana del Cerro, near Zarzal, Oriente, July, 1922 {Leon loggo;
type No. 1,049,981, U. S. National Herbarium).

Among North American species Polygala rhynchosperma is
nearest P. paniculata L., which is densely stipitate-glandular
and has shorter wings and an erostrate seed. The Venezuelan
P. Funkii Chod., known to me only from Chodat's description
and figures, is evidently closely related, having similar bracts,
seed, and stigma. It is described as glabrous, with acute lance-
elliptic sepals, acute wings, and a crest composed on each side
of an acute triangular entire lamina and an undivided linear
lobe.

Polygala ambigens Blake, sp. now
Stems usually solitary from a slender but evidently perennial
root, simple or sparsely branched, 3.5-6 dm. high, very slender,
glabrous throughout or sparsely incurved-puberulous on the

Britton: Studies of West Indian plants — XI 41

younger parts; leaves alternate, linear or linear-filiform, 9-25
mm. long, 0.5-1.5 mm. wide, cuspidulate, short-petioled, 1-
nerved, the margins often inflexed; peduncles 5-10 mm. long;
racemes loosely many-flowered, the axis sometimes becoming
25 cm. long; bracts subulate, 1.5 mm. long, caducous; pedicels
glabrous, 1.5-2 mm. long; sepals bearing 3-6 pairs of stipitate
glands, sometimes also ciliolate with eglandular hairs toward
apex, the upper sepal 2 mm. long, the lower sepals (connate
nearly to apex) 1.5 mm. long; wings magenta-purple, broadly
and obliquely cuneate-obovate, 3.8-4.5 mm. long, 2.8-3.2 mm.
wide, subsessille or short-clawed, emarginate at apex, glabrous,
about 5-veined, the veins loosely anastomosing, not at all
prominulous; upper petals magenta-purple; keel yellow-tipped,
becoming 4 mm. long; capsule 3.8 mm. long, 2.4 mm. wide,
slightly shorter than the mature wings; seed densely pilose,
2.8 mm. long; aril sparsely pubescent on the sides, 0.8 mm. high.

Sabana del Cerro, near Cerro Pelado, between Zarzal and
Nagua, Oriente, July, 1922 {Leon 10860; type No. 1,049,889,
U. S. National Herbarium).

Nearly related to Poly gala cubensis Chod., which is an annual
with shorter sepals (1. 1-1.3 mm- l°ng)> and somewhat smaller
wings shorter than the capsule. In its slender but evidently
perennial root resembling P. Wrightii Chod., which is densely
incurved-puberulous and has the sepals ciliate with eglandular
hairs. Much more material of the Cuban representatives of
this group (section Adenotricha of subgenus Hebeclada) must be
accumulated before the specific characters can be established
satisfactorily.

Maytenus saxicola Britton & Wilson, sp. nov.

A glabrous shrub, 4 m. high, with rather stiff, somewhat
flexuose twigs. Leaves elliptic to elliptic-ovate, rigid-coriaceous,
3-6 cm. long, 1.5-3.3 cm. broad, obtuse or rounded at the apex,
a cute at the base, dark green above, paler beneath, the midvein
elevated on both sides, the lateral veins often rather indistinct,
the margin minutely crenulate; petioles 5 mm. long or less;
pedicels slender, up to 7 mm. long; capsule globose-obovoid,
8 mm. long, 7-8 mm. in diameter.

On rocks, Sierra Maestra, Oriente {Leon, Clement and Roca
10221).

Rhamnidium bicolor Britton & Wilson, sp. now

A shrub, with slender puberulent twigs. Leaves opposite, el-
liptic or ovate-elliptic, 8 cm. long or less, repand-crenate, mem-

42 Britton: Studies of West Indian plants — XI

branous, the apex acuminate, the base rounded or subtruncate,
the venation ascending, densely and finely reticulate-veined on
both sides, the upper surface green and glabrate, the under side
whitish puberulent, the petioles 6-IO mm. long; stipules obliquely
linear-lanceolate, nearly as long as the petioles; inflorescence
axillary, glabrous; peduncles nearly as long as the petioles;
flowers subglomerate; pedicels very short; sepals triangular-ovate,
2.5 mm. long, acute; glandular-dotted; petals orbicular-obovate,
1.5 mm. long, about 1.5 mm. broad.

High Sierra Maestra, Oriente {Leon 11025). Other specimens
from the Sierra Maestra {Leon 10052), in foliage only, have
entire leaves, somewhat larger, and may represent another
species.

Ouratea xolismaefolia Britton & Wilson, sp. nov.
A glabrous shrub, with slender terete twigs. Leaves oblong,
elliptic, or elliptic-obovate, coriaceous, 3 cm. long or less, the
midvein prominent, the lateral venation delicate and nearly
obscure, the apex rounded or emarginate, the base narrowed or
cuneate, the petioles 1-2 mm. long; pedicels slender, 5-6 mm.
long, sepals elliptic to oval, 3 mm. long; petals obovate, 4-4.5
mm. long, 3 mm. broad, rounded or truncate at the apex; anthers
3 mm. long.

High Sierra Maestra, Oriente {Leon iogn).

Taonabo monticola Britton & Wilson, sp. nov.
A glabrous shrub, 1 m. or more tall, with light brown twigs.
Leaves elliptic-obovate, rigid-coriaceous, 2-5 cm. long, 1.3-2.5
cm. broad, rounded at the apex, acute at the base, indistinctly
veined, the midvein impressed above; petioles 4 mm. long;
peduncles 2-3 cm. long; sepals 5, the outer oval to suborbicular,
5.5-6 mm. long, 5-5.5 mm. broad, glandular-ciliate, the inner
4.5-5 mm. broad, rigidly and sharply acute, without marginal
cilia; fruit (immature?) ovoid, 9.5 mm. in diameter.

Sierra Maestra, Oriente {Leon 10839, type; 10837; *i*93)-

Taonabo Leonis Britton & Wilson, sp. nov.

A glabrous shrub, with light-gray twigs. Leaves obovate,
2.5-5 cm. long, 1.4-2.7 cm. broad, rounded or truncate and
often slightly emarginate at the apex, acute at the base, indis-
tinctly veined, dark green above, paler beneath, the petioles 2-3
mm. long; peduncles 3-7 mm. long, reflexed; sepals 5, the outer
oval, 4.7-5 mm. long, 4 mm. broad, rounded at the apex, glandu-
lar-ciliate, the inner elliptic, 3-3.5 mm. long, acute, without
glandular cilia.

Vicinity of Pico Turquino, Oriente {Leon 10910).

Britton: Studies of West Indian plants — XI 43

Haemocharis angustifolia Britton & Wilson, sp. nov.

A tree 20-30 m. tall, with slender grayish brown twigs, pu-
bescent with strigillose hairs. Leaves oblong or oblong-obovate,
2-5.5 cm. long, 0.5-1.3 cm. broad, obtuse or rounded at the
apex, acute at the base, light green and glabrous above, paler
beneath and pubescent; flowers about 3 cm. in diameter; sepals
oval or suborbicular, 4.5-6 mm. long, 4-6 mm. broad, pubescent
on the back with appressed hairs; petals obovate, 1.2-1.3 cm.
long, 6-7 mm. broad, rounded at the apex, pubescent; filaments
filiform, glabrous; ovary villous.

Sierra Maestra, Oriente {Leon, 11072).

Begonia lomensis Britton & Wilson, sp. nov.

Suffruticose, the stems 3-4.5 dm. high, rusty pilose. Leaves
obliquely cordate-ovate to obliquely suborbicular, 1.5-4 cm-
long, 1-2.7 cm. wide, acute, glabrous above at maturity, rusty
pilose or nearly glabrous beneath, coarsely toothed, the petioles
4-20 mm. long, pilose; stipules lanceolate, acuminate, entire,
4-6 mm. long, 1.5-2 mm. broad; inflorescence long-peduncled;
staminate flowers; pedicels slender, up to 7.5 mm. long, glabrous;
sepals 2, orbicular, 7-8 mm. long, 8-9 mm. wide; petals 2, obo-
vate, 4 mm. long, 3 mm. wide; stamens 20 or more; capsule
long-pedicelled, the body 8-9 mm. long, 5-6 mm. wide, the
wing about 9 mm. broad, obtuse or acutish.

Farallones de la Loma Regino, Oriente (Leon ioyji).

Ginoria montana Britton & Wilson, sp. nov.

A glabrous unarmed shrub, with slender ascending branches.
Leaves lanceolate, coriaceous, 2.5-4.3 cm. long, 0.7-1.7 cm.
broad, obtuse at the apex, acute at the base, faintly reticulate-
veined; petioles slender, 2 mm. long; cymes sessile or nearly so;
flowers immature, long-pedicelled; calyx-lobes spreading, appen-
daged below the sinuses; stamens about 20.

Sierra Maestra, Oriente (Leon 11000).

Psidium (?) cacuminis Britton & Wilson, sp. nov.

Glabrous, with straight slender twigs. Leaves oblong.thick-
coriaceous, 4-7 cm. long, reticulate-veined, sparingly punctate,
bright green and shining above, pale green beneath, the apex
rounded or obtuse, the base narrowed, the stout petioles about
6 mm. long; fruit oblong, solitary and nearly sessile in upper
axils, nearly 2 cm. long.

Pico Turquino, Oriente (Leon 10740).

44 Britton: Studies of West Indian plants — XI

Calyptranthes montana Britton & Wilson, sp. nov.

A much-branched shrub, with slender glabrous branches, the
twigs rusty-brown and somewhat pubescent when young.
Leaves lanceolate, coriaceous, 9-30 mm. long, 6-1 1 mm. broad,
obtusely acuminate at the apex, acute at the base, rusty brown
and pubescent when young, glabrous at maturity, the lateral
veins rather prominent on both sides; petioles 0.9 mm. long or
less; inflorescence 1-3-flowered, the peduncles filiform, 7 mm.
long or less; buds globose-ovoid, 1.5 mm. in diameter.

Sierra Maestra, Oriente {Leon 10805).

Eugenia cati Britton & Wilson, sp. nov.

A glabrous shrub or small tree, the twigs slender. Leaves
elliptic to elliptic-obovate, chartaceous, 7 cm. long or less,
rather sparingly punctate, dark green above, pale beneath, not
very conspicuously veined, the apex rounded or obtuse, the base
mostly narrowed, the petioles 3-7 mm. long; flowers glomerate
in the axils; calyx-lobes suborbicular, about 3 mm. broad.

In woods, Loma del Gato, Cobre Range of Sierra Maestra,
Oriente {Leon, Clement and Roca 10045, type; 10588).

Eugenia ginoriaefolia Britton & Wilson, sp. nov.

Glabrous; twigs slender. Leaves ovate to elliptic, char-
taceous, 3-7 cm. long, dark green and shining above, rather
prominently veined, pale beneath, densely punctate, the apex
obtuse, the base obtuse or narrowed, the petioles only 3 mm.
long or less; flowers glomerate at the nodes of twigs, sessile;
calyx-lobes about 2 mm. broad, densely glandular, glabrous;
petals elliptic to oval, about 3 mm. long, 1.8-2.2 mm. broad,
obtuse or acutish at the apex; filaments filiform, glabrous, the
anthers elliptic.

Sierra Maestra, Oriente {Leon 10812).

Eugenia lomensis Britton & Wilson, sp. nov.

Shrub 2-3 m. high, with slender pubescent twigs. Leaves
lanceolate or elliptic, coriaceous, 3.5-7 cm. long, 1.7-3 cm-
broad, acuminate at the apex, acute at the base, glabrous or
nearly so above, the midrib and lateral veins impressed, pubes-
cent beneath, the midvein prominent; petio'es 2.5 mm. long or
ess ; fruit subglobose, 9 mm. in diameter, black, slender-peduncled,
solitary or clustered in the axils of the leaves.

Foothills of the Sierra Maestra, Oriente {Leon ioyg4, type;
10062).

Britton: Studies of West Indian plants — XI 45

Eugenia brevipetiolata Britton & Wilson, sp. nov.

A glabrous shrub, with rather thick brownish gray twigs.
Leaves oval to suborbicular, 5-8 cm. long, 4-6.5 cm. broad,
rounded at both ends, dark yellowish green above, paler beneath,
the midrib impressed above, prominent beneath, the lateral
veins slender, slightly elevated on both surfaces; petioles stout,
2-3 mm. long; buds subglobose, 6 mm. in diameter, solitary or
few together in the axils of the leaves, glabrous.

Sierra Maestra, Oriente {Leon 11046).

Eugenia rigidula Britton & Wilson, sp. nov.

Shrub 1 m. tall, with slender brownish pubescent twigs.
Leaves lanceolate to ovate-lanceolate, rigid-coriaceous, 1-2.8
cm. long, 4-1 1 mm. broad, acute at the apex, rounded at the
base, sparingly pubescent on both sides when young, glabrous at
maturity, the petioles 1 mm. long or less ; young fruits subglobose,
4-5 mm. in diameter, pubescent with appressed whitish hairs.

Edge of woods, Sierra Maestra, Oriente {Leon, Clement and
Roca 10634).

Anamomis reticulata Britton & Wilson, sp. nov.

Shrub or tree, the young twigs slender, brown, puberulent,
the branches gray. Leaves obovate, coriaceous, 3-4.5 cm. long,
1-2.5 cm- broad, rounded at the apex, cuneate at the base,
yellowish-green above, paler beneath, sparingly pubescent on
both sides when young with short appressed hairs, glabrous at
maturity, reticulate-veined; branches of the inflorescence ap-
pressed-pubescent ; flowers undeveloped.

Sierra Maestra, Oriente {Leon 10068).

Graffenrieda rufescens Britton & Wilson, sp. nov.

Shrub 3 m. tall, the branches, twigs, calyx and under surface
of the leaves densely scurfy with brownish scales. Leaves
elliptic, 4.5-10 cm. long, 2.4-4 cm. wide, acuminate at the apex,
acute at the base, 3-nerved, dark green and glabrous above, the
petioles 1-2 cm. long; inflorescence much-branched; flowers
short-pedicelled ; calyx-tube 3-3.5 mm. long, 8-ribbed, its lobes
semiorbicular, I mm. long or less; petals suborbicular or orbicu-
ular-obovoid, 3.5 mm. long, 3-3.5 mm. wide, yellow.

Sierra Maestra, Oriente {Leon 11047).

Miconia plumieriaefolia Britton & Wilson, sp. nov.

Shrub, the twigs, petioles and branches of the inflorescence
clothed with a dense brownish white stellate scurf. Leaves

46 Britton: Studies of West Indian plants — XI

3-nerved, oblong or oblong-lanceolate, coriaceous, 1 0.5-1 5 cm.
long, 2.5-3.8 cm. broad, green and glabrous or nearly so above,
densely ferruginous-pubescent beneath with short stellate hairs,
reticulate-veined, the lateral veins 2-3 mm. distant, nearly at
right angles to the midrib, the petioles 1.5-2 cm. long; branches
of the inflorescence loosely articulate; pedicels 2-4.5 mm. long;
flowers undeveloped; calyx-lobes shortly triangular-ovate.

Sierra Maestra, Oriente {Leon 10925).

Rapanea microphylla Britton & Wilson, sp. nov.

A glabrous shrub, with short and slender twigs. Leaves
obovate or oblong-obovate, coriaceous, 8-20 mm. long, pin-
nately few-veined, punctate, the apex rounded, the base connate,
the petioles 2-3 mm. long; fruits nearly sessile, subglobose,
shining, about 3 mm. in diameter.

Pico Turquino, Sierra Maestra, Oriente {Leon 107 14).

Symplocos Leonis Britton & Wilson, sp. nov.

Shrub, the twigs dark brown, strigillose. Leaves elliptic,
subcoriaceous, 4-8.5 cm. long, 2-4 cm. broad, rounded or ob-
tusely acute at the apex, acute at the base, dark green, glabrous
and indistinctly reticulate-veined above, the midrib impressed,
paler, reticulate-veined and pubescent with appressed hairs
especially on the midvein beneath, or nearly glabrous; petioles
about 5 mm. long; inflorescence pubescent with ferruginous hairs;
calyx-lobes elliptic, 1 .4 mm. long, 0.9 mm. broad ; petals 5, broad-
ly oval, 3-3.5 mm. long, 2.5 mm. broad, glabrous; filaments dis-
tinct nearly to the base; fruit ellipsoid, 8.5 mm. long, 5.5 mm. in
diameter, pubescent with scattered appressed hairs.

Sierra Maestra, Oriente {Leon 10738, type; 10907).

Plumiera montana Britton & Wilson, sp. nov.

A tree up to 10 m. high or higher, glabrous throughout.
Leaves oblanceolate, 15 cm. long or less, membranous when dry,
the lateral venation rather widely spreading, the apex rounded,
the base cuneate, the slender petioles 1-3 cm. long; flowers
several or numerous; pedicels rather stout, 1.5 cm. long or less;
calyx about 5 mm. long; corolla (white?) its slender tube about
2 cm. long, about one-half as long as the narrowly oblanceolate,
rounded lobes; pod about 12 cm. long.

Sierra Maestra, Oriente. Type from banks of Yara River
{Leon 11060).

Brixton: Studies of West Indian plants — XI 47

Marsdenia bicolor Britton & Wilson, sp. nov.

A slender vine, with puberulent stems. Leaves oblong-
elliptic to elliptic, 5-6.5 cm. long, 1.5-3.2 cm. broad, green above,
whitish beneath, glabrous on both surfaces, the lateral veins
indistinct; calyx-lobes ovate, 4 mm. long, about 2 mm. broad,
acute or acuminate, slightly pubescent; corolla 1.5 cm. long,
1 .5-1 .8 cm. wide, glabrous, its lobes triangular, 6 mm. long,
5.5-6 mm. wide.

Sierra Maestra, Oriente {Leon 10787, type; 10788).

Sebesten Leonis Britton & Wilson, sp. nov.

A tree up to 15 m. high, the young twigs sparingly pubescent,
soon glabrous and gray. Leaves thick-coriaceous, glabrous or
the short and stout petioles sparingly pubescent, shining, elliptic,
oblong or obovate-elliptic, 5-9 cm. long, the venation impressed
above, prominent beneath, the apex rounded, the base obtuse or
rounded, the petioles only 8 mm. long or less, the upper surface
dark green, the under surface light green; flowers several in
terminal clusters; calyx subcylindric, nearly 2 cm. long, densely
puberulent, its teeth very short and broad; corolla funnelform,
its tube about 3 cm. long, enlarged upward, the widely spreading
limb about 4 cm. broad, about 6-lobed; fruit ovoid, about 2 cm.
long, its summit beaked.

Loma de Naranjo, Sierra Maestra, Oriente, at 600 m. elevation
{Leon 11021).

Varronia longipedunculata Britton & Wilson, sp. nov.

A shrub with slender branches, the young twigs puberulent.
Leaves elliptic or oblong, chartaceous, 7 cm. long or less, scabrate
above, puberulent on the veins beneath, entire or sparingly
repand, the apex obtuse or acute, the base mostly narrowed, the
puberulent petioles 6-12 mm. long; peduncles solitary, slender,
puberulent, longer than the leaves; flowers several in a terminal
globose head about 1.5 cm. in diameter; calyx-lobes triangular,
1.5 mm. long, acute; corolla-lobes broadly obovate, 7.5 mm.
long, 4.5 mm. broad.

Sierra Maestra, Oriente {Leon 10064).

Salvia scopulorum Britton & Wilson, sp. nov.

A shrub, about 3 m. high, the branches densely pubescent with
short spreading hairs. Leaves ovate, coriaceous, crenulate, 3-5
cm. long, 1.5-3 cm. broad, acute to abruptly short-acuminate at
the apex, rounded and often more or less subcordate at the base,
dark green above, paler beneath, pubescent on both sides, the

48 Britton: Studies of West Indian plants — XI

petioles 12 mm. long or less; calyx 10-11 mm. long, finely pu-
bescent; corolla red, 1.8 cm. long or more, pubescent.

Farallones de Regino, Estribo Turquino, Oriente, 1600 m. alt.
{Leon iioio).

Rondeletia naguensis Britton & Wilson, sp. nov.

A shrub about 2 m. high, the young twigs and inflorescence
densely strigose. Leaves elliptic, subcoriaceous, 4-10 cm. long,
sparingly strigose on the veins beneath when young, soon
glabrous, the apex obtuse or acute, the base obtuse or narrowed,
the venation inconspicuous, the strigose petioles 6-12 mm. long;
flowers in peduncled heads, the peduncles 2-4 cm. long; bractlets
obovate, about 3 mm. long; calyx-lobes oblanceolate, 3.5-4 mm.
long, densely pubescent with appressed hairs; capsule subglobose,
5.5 mm. in diameter, appressed-pilose.

Banks of a stream, Loma de Quintin, Nagua, Oriente {Leon

10971).

Rondeletia calophylla Standley, sp. nov.

Branchlets slender, pilose-strigose; stipules deltoid-acuminate,
1.5 to 2.5 mm. long, erect, persistent; petioles stout, 2 mm. long,
strigillose; leaf blades ovate, sometimes broadly so, 2 to 3.5 cm.
long, 1 to 2.5 cm. wide, acute or sometimes obtuse, broadly round-
ed or subcordate at base, thick-coriaceous, dark green above,
lustrous, scaberulous or glabrate, the venation prominulous-
reticulate, much paler beneath and brownish, glabrate or
scaberulous along the nerves, the costa prominent, the lateral
nerves few and distant; inflorescence terminal on the main
branches and on leafy lateral twigs, few-flowered, lax, the
flowers pedicellate; calyx globose, densely whitish-strigillose;
calyx lobes 5, oblong or linear-oblong, 2 mm. long, obtuse, gla-
brate, suberect; corolla densely white-strigillose outside, the
tube 4 mm. long, the 5 lobes rounded, 2 mm. long; capsule
subglobose, 3-4 mm. broad, glabrate; seeds broadly winged.

Type (in the U. S. National Herbarium) collected in the Sierra
Maestra, Oriente, July, 1922, {Leon 10741).

In the key to the species of Rondeletia in the North American
Flora, this plant runs at once to the group Odoratae, but it is
not closely related to any of the species placed there, or to any
other Cuban species previously known to the writer.

Guettarda cryptantha Standley, sp. nov.

Shrub, the branches grayish, the branchlets slender, densely
pilose with subappressed fulvous hairs; stipules 8-10 mm. long,

Britton: Studies of West Indian plants — XI 49

oblong-lanceolate, long-acuminate, appressed-pilose; leaves op-
posite, the petioles stout, 2-3 mm. long, appressed-pilose, the
blades elliptic or ovate-elliptic, 2.5-6 cm. long, 1.2-3 cm- wide,
rounded at base, obtuse or rounded at apex, coriaceous, dark-
green above, lustrous, at first pilose with very slender whitish
appressed hairs but later glabrate, the venation depressed, paler
and brownish beneath, thinly scaberulous-pilosulous and with
longer subappressed hairs along the veins, the costa stout and
prominent, the lateral veins 5-7 pairs, subarcuate.anastamosing,
remote from the revolute margin; inflorescence few-flowered,
closely sessile, terminal and axillary, subtended by large, densely
pilose, oval bracts; hypanthium densely sericeous-pilose with
long whitish hairs, the calyx 5-7 mm. long, cleft into 3 or 4
spatulate-obovate, obtuse, densely sericeous lobes; corolla
brownish-red, the tube stout, 10-12 mm. long, densely pilose
with long whitish ascending hairs, the 4 lobes rounded-oval, 4
mm. long, glabrous within.

Type collected on Loma del Gato, Cobre Range of Sierra
Maestra, Oriente, altitude 900-1000 meters {Leon, Clement and
Roca 10301; herbarium of the New York Botanical Garden).

Guettarda cobrensis Standley, sp. nov.

Shirub 3 meters high, the branches dark reddish brown, the
branchlets thick, densely covered with minute appressed fulvous
hairs, the internodes short; stipules oblong-ovate, 4 mm. long
or more, acuminate, thin, brown, pubescent like the branchlets
and with few long subappressed hairs; leaves opposite, the
petioles stout, 4-6 mm. long, minutely and densely appressed-
pilose, the blades oval or broadly oval, 3-6 cm. long, 2-4.5 cm-
wide, rounded at apex, deeply cordate at base, thick-coriaceous,
yellowish-green, concolorous, often somewhat bullate, glabrous
above, the venation depressed, beneath bearing a few short
appressed hairs along the coarse prominent costa, but elsewhere
glabrous, the lateral veins stout and prominent, 5 or 6 on each
side, subarcuate, ascending at an angle of 45-60°, the intermediate
veins inconspicuous, the margin subrevolute; peduncles about I
cm. long, pubescent like the branchlets, the flowers usually 3,
sessile; fruit globose, 1 cm. in diameter, minutely tomentulose.

Type collected at edge of woods, Loma del Gato, Cobre Range
of Sierra Maestra, Oriente, altitude 800 meters (Leon, Clement
and Roca 10271; herbarium of the New York Botanical Garden).

Related to G. crassipes Britton, in which the leaves are densely
long-pilose beneath along the costa.

50 Britton: Studies of West Indian plants — XI

Stenostomum apiculatum Britton Sc Standley, sp. nov.

Branchlets thick, resinous, densely leafy, sericeous-strigose;
stipules united, persistent, the sheath 5 mm. long, densely stri-
gose; petioles stout, 5-12 mm. long, densely strigillose; leaf
blades elliptic, 4.5-7 cm. long, 2.5-4 cm- wide, obtuse or acute at
base, obtuse at apex and apiculate-acuminate, coriaceous, green
above, lustrous, glabrous, the costa depressed, paler beneath
and brownish, sparsely strigillose, the veins prominent, the
lateral ones about 9 pairs, arcuate, distant, the margin plane or
subrevolute; peduncles slender, about 3 cm. long, the 2 branches
2-2.5 cm- l°ng» io-12-flowered; calyx and hypanthium 1-1.5
mm. long, glabrate, the calyx with very short, obtuse teeth;
corolla 7 mm. long, densely strigillose outside; filaments very
short, the anthers oblong-linear, equaling the corolla tube;
stigma 4-lobate.

Type (in the U. S. National Herbarium) collected in the Sierra
Maestra, Oriente, July, 1922 (Leon, 10806.)

Psychotria Leonis Britton & Wilson, sp. nov.

Shrub, the twigs and branches of the inflorescence puberulent.
Leaves elliptic to oval, 3-5 cm. long, 1.5-3 cm. broad, acute to
short-acuminate at the apex, acute at the base, glabrous above,
pubescent beneath when young with short appressed hairs, the
primary veins indistinct above, prominent beneath; petioles
I.3-I.5 cm. long, puberulent; panicles few-flowered; calyx
funnel-shaped, shallowly toothed, the teeth triangular, acute;
buds narrowly ellipsoid, puberulent; mature flowers and fruits
not seen.

Sierra Maestra, Oriente (Leon 10780).

Lobelia cacuminis Britton & Wilson, sp. nov.
Suffrutescent, the stems 3-4 m. high, nearly glabrous. Leaves
chartaceous, oblanceolate, 7-13 cm. long, 2.5-4.5 cm- wide,
acuminate at the apex, cuneate at the base, rather coarsely
dentate, the petioles 1.5-3 cm. long; midrib flat above, prominent
beneath , the lateral veins ascending ; inflorescence many-flowered ;
pedicels slender, 10-12 mm. long, pubescent; calyx-lobes nar-
rowly lanceolate, 3-4 mm. long; mature corolla not seen; young
fruits 1 cm. in diameter.

Pico Turquino, Oriente (Leon 10862).

Senecio Leonis Britton & Wilson, sp. nov.

Vine, the slender stems glabrous. Leaves ovate, triangular-
ovate or occasionally lanceolate, 1.5-4.5 cm. long, 0.9-2.7 cm.
wide, often with one or more prominent teeth at or above the

Britton: Studies of West Indies plants — XI 51

base, palmately-veined, glabrous on both sides; petioles 0.5-1.5
cm. long; heads long-peduncled, the peduncles pubescent;
involucre subcylindric-campanulate, 5-6 mm. high, its bracts
oblong, glabrous, their tips acute or acuminate; pappus white,
4 mm. long; achene about I mm. long.

Climbing on shrubs, Loma del Sabicu, Oriente {Leon 10751).

Chaptalia montana Britton, sp. now

Leaves oblanceolate-spatulate, thin, 4-7 cm. long, lyrate-
pinnatifid, the terminal lobe large, elliptic-obovate, obtuse,
apiculate, the lateral ones few, small, rounded, the upper surface
green and glabrate, the under grayish-lanate; scape filiform,
1.5-2.5 dm. long; head erect; involucre subcampanulate, its
bracts linear, acuminate, lanate below, the inner ones about
7 mm. long; achenes fusiform, 5.5-6 mm. long, the slender beak
longer than the body; pappus white.

Sierra Maestra, Oriente {Leon 10802).

65. UNDESCRIBED SPECIES FROM TRINIDAD
Bactris savannarum Britton, sp. nov.

Trunk slender, up to 10 m. high, erect, copiously armed, like
the leaf-rachis, with nearly white, deflexed spines 1-5 cm. long,
which turn darker in age. Leaves about I m. long or longer,
petioled, the numerous, unarmed segments 2-4 dm. long, about
4 cm. wide or less, light green, paler beneath than above, the
base attenuate, the apex acuminate or acute, some of the upper
ones of young leaves confluent; spathe narrowly oblong, 2-3
dm. long-acuminate, densely appressed-bristly, the bristles
rather soft; inflorescence about as long as the spathe, its axis and
slender branches densely brown-villous; flowers numerous, nearly
white, about 4 mm. long; fruit orange, depressed-globose, small
for the genus, about 8 mm. in diameter, vertically striate;
fruiting calyx nearly flat, about 5 mm. broad, its lobes rounded,
striate.

Aripo Savanna and its forest borders, Trinidad {Britton,
Broadway and Hazen 260, type; Trinidad Herbarium 9850,
Broadway) ; Professor L. H. Bailey has kindly sent us two
photographs of this palm taken by him in 1921.

Pilea aripoensis Britton, sp. nov.

Stems filiform, 4-7 cm. long. Leaves membranous, spatulate,
entire, i-nerved, 12 mm. long or less, the apex rounded, the base
cuneate, the filiform petiole somewhat shorter than the blade,

52 Britton: Studies of West Indian plants — XI

the raphides linear, undulate, transverse, prominent upon the
upper surface.

Hanging from wet dark banks, Heights of Aripo, Trinidad
(Trinidad Herbarium 10,008, Broadway).

Inga ierensis Britton, sp. nov.

A tree up to 16 m. high, the petioles, leaf -surfaces and ped-
uncles strigillose. Petioles slender, 2-3 cm. long; rachis-glands
suborbicular, semiglobose; leaflets in 2 pairs, ovate-elliptic, char-
taceous, shining, 7-10 cm. long, pinnately veined, the apex
rather abruptly short-acuminate, the oblique base narrowed
or obtuse, the strigulose petiolules 2-3 mm. long; peduncles
about 6 cm. long or less; legume broadly linear, compressed,
glabrous, shining, up to 10 cm. long, about 2 cm. wide, finely
striate, abruptly short-tipped, the margins continuous or deeply
constricted.

Forests, northern mountain range, Trinidad. Type from the
Maracas Valley {Britton and Hazen 1627). In fruit April io,
1920.

Serjania ierensis Britton, sp. nov.

A vine up to 7 m. long, the branches slender, glabrous. Leaves
about 6 cm. long, biternate, the short, slender petioles 1.5 cm.
long or less, sparingly pubescent; leaflets coriaceous, glabrous,
4 cm. long or less, serrate, finely reticulate-veined, shining,
those of the terminal segment acute or acuminate, those of the
lateral segments rounded or obtuse; peduncles about as ong as
the leaves, glabrous; tendrils very slender, 2-4 cm. long; panicles
puberulent nearly as long as the peduncles; samaras glabrous,
about 15 mm. long, the wing finely reticulate-veined, about 4
times as long as the body and only partially enclosing it.

Dry thickets, Trinidad. Type from Palo Seco (Britton and
Hazen 1172). In fruit March 30, 1920.

Clusia intertexta Britton, sp. nov.

Forming impenetrable masses with interlocking aerial roots.
Leaves orbicular-obovate, 10-14 cm- l°ng> thick-coriaceous,
rounded, the base subcuneate, the lateral veins many, about 4
mm. apart, prominent on the upper surface, the very stout
petioles 1 cm. long or less; inflorescence several-flowered, its
branches stout; bracts 2-3 cm. long; buds globose, glabrous;
fruit (immature) oblong-ovoid, beaked, about 2.5 cm. long;
stigmas 6.

Forest, Heights of Aripo, Trinidad (Trinidad Herbarium 9786,
Broadway), January 10-26, 1922.

Britton: Studies of West Indian plants — XI 53

Clusia aripoensis Britton, sp. nov.

A tree, with slender twigs. Leaves narrowly obovate, sub-
coriaceous, 7-8 cm. long, the apex obtuse, the base cuneate, the
rather closely pinnate venation prominent on the under side,
almost wholly obscure above, the rather slender petioles 1-2
cm. long, stigmas 5; young fruit obovoid, 2.8 cm. long, 2 cm. in
diameter.

Forest, Heights of Aripo Trinidad (Trinidad Herbarium 9785,
Broadway), January 10-26, 1922.

Stylogyne tenuifolia Britton, sp. nov.

A tree up to about 10 m. in height. Leaves elliptic, mem-
branous, 8-20 cm. long, glabrous, the base narrowed, the apex
rather abruptly acuminate, the midvein impressed above, prom-
inent beneath, the stout petioles 5-12 mm. long; inflorescence
lateral and terminal, 5 cm. long or less, short-paniculate; bract-
lets oblong-lanceolate, 2-3 mm. long, deciduous; pedicels nearly
filiform, 5-10 mm. long; calyx-segments broadly ovate, rounded,
about 2 mm. long; corolla about twice as long as the calyx; fruit
globular, white, 6-7 mm. in diamater.

Forests at higher elevations in the northern mountain range,
Trinidad. Type from Mt. Tocuche (Britton, Hazen and Mendle-
son 1341) ; collected also by Mr. Broadway on the Heights of
Aripo (Trinidad Herbarium, 9837, 9840).

Prestonia Fendleri N. E. Brown, sp. nov.

Stem climbing, the flowering parts with internodes 4.5-12 cm.
long, 1-2 mm. thick, thinly puberulous with minute curved hairs
on the youngest parts, becoming glabrous and brown with age.
Leaves with petioles 6-15 mm. long and blades 5-10 cm. long
and 3-7 cm. broad, elliptic, subequally and broadly rounded at
both ends, with a small abrupt point 1-2 mm. long at the apex,
with about 5 ascending curved primary veins on each side of the
midrib, inarching at their tips and forming loops near the
margin. Racemes axillary, the axis 1-2 cm. long in the example
seen, bearing 15 or more flowers on the upper half, flowerless
below, thinly and minutely puberulous. Bracts minute, less
than 1 mm. long. Pedicels 12-15 mm. long, slender, puberulous.
Sepals 5 mm. long and 1.5-2 mm. broad, lanceolate, acute, thinly
puberulous. Corolla pubescent at the top of the tube inside,
otherwise glabrous; tube 15-17 mm. long and about 4 mm. in
diameter, cylindric, with the mouth, where the lobes pass from
it, raised into an erect ring I mm. high; inside this ring arise 5
linear erect exserted processes (coronal-lobes) 4 mm. long; lobes

54 Britton: Studies of West Indian plants — XI

about 12 mm. long and 7 mm. broad, somewhat rhomboid, or
obliquely subtruncate. Stamens inserted near the top of the
tube, glabrous; filaments 2 mm. long; anthers sagittate, 4.5 mm.
long, exserted. Ovary surrounded by an obtusely 5-lobed fleshy
disk.

Trinidad, without precise locality, Fendler 628.

Cordia ierensis Britton, sp. nov.

A tree up to 10 m. high, with slender, finely appressed-pu-
bescent twigs. Leaf-blades elliptic to ovate, thin, 2 dm. long
or less, scabrate-puberulent, the apex abruptly acuminate, the
base obliquely obtuse or narrowed, the petioles 2 cm. long or
less; panicle about 8 cm. broad, broader than long, many-
flowered, its branches slender, densely appressed-pubescent ;
flowers sessile, the buds obovoid, rounded, densely appressed-
pubescent, about 2 mm. long.

Forests, Trinidad. Type from Morne Bleu (Britlon, Freeman
and Bailey, 22ff). In bud March 13, 1921. This is, as deter-
mined by Mr. N. E. Brown at Kew, the Cordia panicularis of
Grisebach, not of Rudge.

Aegiphila trinitensis Britton, sp. nov.

A shrub, with slender, glabrous branches, the bark nearly
white, the young shoots brown, short, minutely puberulent.
Leaves short-petioled, the blades ovate, thin, small for the genus,
4 cm. long or less, the apex acute or acuminate, the base nar-
rowed; panicles small, short-peduncled, strigose, several-many-
flowered, 2-3 cm. broad; calyx obconic, truncate, puberulent;
corolla white, about 8 mm. long.

Hillsides and thickets in moist districts, Trinidad. Type
from St. Ann's (Mrs. W. E. Broadway) . In flower May 10, 1919.
This is the species recorded by Grisebach from Trinidad, as

Aegiphila laevis.

Anguria elliptica Britton, sp. nov.
A slender, glabrous, high-climbing vine, with very slender
tendrils 15 cm. long or less. Petioles about 2 cm. long; leaf-
blades simple, oblong-elliptic, thin, entire, about 8 cm. long,
pinnately veined, the apex abruptly acute, the base obtuse;
peduncles slender, longer than the leaves; staminate flowers
scarlet, spicate; spikes about 3 cm. long; calyx-lobes subulate,
1 mm. long; corolla-lobes spatulate-obovate or obovate, 4 mm.
long, 2-2.6 mm. broad, rounded or subtruncate and mucronate
at the apex, the margin minutely ciliate; anthers 4.5 mm. long,
straight.

Britton: Studies of West Indian plants— XI 55

Forests in wet or moist districts, Trinidad ; type from Arima
(Britton 2084). In flower March 4, 1921.

66. UNDESCRIBED SPECIES FROM PORTO RICO
Scleria doradoensis Britton, sp. nov.

Perennial? Culms slender, rather stiff, glabrous, erect, tufted,
3-4 dm. high. Leaves glabrous, scabrate, shorter than the culm,
about 2 mm. wide, the apex attenuate; inflorescence terminal,
glomerate-spicate, 5-7 cm. long; glomerules 6-8, few-flowered,
not nodding; bracts about 3 mm. long, brown-ciliate; achene
white, subglobose-obovoid, very obtusely trigonous, verrucose-
tuberculate, 1.5 mm. long, 12-porose underneath just above the
trigonous base; hypogynium none.

White sand near Dorado [Britton, Britton and Brown 7064).

Scleria Stevensiana Britton, sp. nov.

Perennial? rootstocks short or none; culms very slender,
tufted, slightly scabrous above, weak, 4-5 dm. long. Leaves
soft, scabrous-margined, 4 mm. wide or less, panicles 2 or 3,
axillary, the lower stalked, the upper one nearly sessile, about
3 cm. long; bracts narrowly linear-subulate, 2.5 cm. long or less;
spikelets several or numerous; achene depressed-globose, about
1.5 mm. in diameter, apiculate, glabrous, shining, faintly re-
ticulated, nearly white, with 3 vertical brown bands; hypogynium
lobes lanceolate, acuminate, glabrous.

Lares {Stevens and Hess 4944).

Lagenocarpus portoricensis Britton, sp. nov.

Culm stout, erect, trigonous, glabrous, about 8 dm. high.
Leaves 8-12 mm. wide, nearly as long as the culm, smooth,
long-attenuate, with numerous transverse veinlets; panicle
narrow, puberulent, about 3 dm. long and 5 cm. thick, the lower
half of staminate spikelets on nearly erect branches, the upper
half of pistillate; bractlets aristate; staminate spikelets about 4
mm. long, short-stalked; scales subtending the achene obovate,
subtruncate, ciliolate, mucronate; achene, with its beak, ovoid,
3 mm. long, brown with three darker blotches, the beak about
as long as the body.

Wet sandy situations between Manati and Vega Baja (Under-
wood and Griggs 946, type) ; collected also by Sintenis near
Dorado. Recorded by C. B. Clarke and by Urban as L. tremulus
Nees.

56 Britton: Studies of West Indian plants — XI

Rynchospora Blauneri Britton, sp. nov.

Perennal by short rootstocks; culms filiform, weak, clustered,
glabrous, 3-5 dm. long; leaves filiform, glabrous, shorter than
the culms; spikelets narrowly oblong, 3 mm. long, few together
in 2 or 3 distant compact clusters subtended by a filiform bract
3-5 cm. long; scales ovate-oblong, acute, light brown; bristles
none; achene obovate, light brown with darker brown margins,
smooth, about 1 mm. long; tubercle conic, acute, about one-
third as long as the achene.

Sierra de Luqiullo (Blauner 247, in herbarium of Columbia
University, collected in 1852-53). Referred by Clarke to R.
divergens.

Rynchospora luquillensis Britton, sp. nov.

Perennial by rootstocks; culms very slender, smooth, erect,
about 2 dm. high. Leaves narrowly linear, about 1 mm. wide,
flat, smooth, shorter than the culm; inflorescence a small ter-
minal cluster of few spikelets; spikelets narrowly oblong, about
4 mm. long; scales oblong, brownish, acute; bristles about 6,
retrorsely barbed, a little longer than the achene; style elongated ;
achene narrowly obovate, smooth, brown, about 2 mm. long;
tubercle subulate, about as long as the achene.

Sierra de Luquillo {Brother Hioram 364.)

67. AN UNDESCRIBED SIPHOCAMPYLUS FROM HAYTI
Siphocampylus pinnatisectus Gleason, sp. nov.

Stem herbaceous, erect, slender, glabrous. Leaves alternate,
broadly oblong in general outline, 5-7 cm. long, 3-3.5 cm. wide,
narrowed at base into a petiole, glabrous, deeply pinnatisect
into a terminal and 5-7 pairs of lateral linear lobes, each lobe
and the rhachis about 2 mm. wide, with 2-4 sharp salient teeth
or the smaller entire, sharply acute or submucronate, with a
single midvein and faint lateral veinlets; flowers in the upper
axils, few in number, appearing subcorymbose by the shortened
internodes, on minutely puberulent pedicels 1 cm. long; hypan-
thium obconic, acute at base, 3 mm. high; calyx-lobes narrowly
linear-triangular, acuminate, erect, separated by narrow acute
sinuses corolla red, about 25 mm. long, narrowly tubular, some-
what curved ventrally, constricted at base, gradually enlarged
upward, 3 mm. in diameter at the throat, its lobes all depressed,
3-4 mm. long.

Type, Nash and Taylor 1701, collected on an open sunny
hillside, between La Brande and Mt. Balance, Hayti, at altitude
of 3150 ft., August 15, 1905.

Siphocampylus pinnatisectus differs from all other West Indian
members of the genus in its deeply pinnatisect leaves.

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL
GARDEN— No. 239

ALTERNATION OF SEXES AND INTERMITTENT PRO-

DUCTION OF FRUIT IN THE SPIDER

FLOWER (CLEOME SPINOSA)

A. B. STOUT

NEW YORK

[Reprinted, without change of paging, from American Journal of Botany
10: 57-66. fig. i + pl. 6. Feb. 1923]

'Recruited from the American Journal of Botany, X: 57-66, February, 1923.]

LIBRARY

NEW YORK

BOTANICAL

GARDEN

ALTERNATION OF SEXES AND INTERMITTENT

PRODUCTION OF FRUIT IN THE SPIDER

FLOWER (CLEOME SPINOSA)1

A. B. Stout

(Received for publication April 6, 1922)

Irregularities in the formation of reproductive organs, such as are seen
in the phenomena of intersexualism in both plants and animals, have two
points of special interest. First, they involve a particular type of sterility
of various grades and degrees of expression, which in plants often affects
the production of fruit and seeds and becomes a matter of practical im-
portance in respect' to crop production and in the breeding of various
economic plants. A second point of interest is in the bearing which the
phenomena of intersexualism have on questions of sex differentiation, the
alternation of sex, and the evolutionary tendencies in reproduction.

In its general significance, several points regarding sterility from inter-
sexualism are clear. In plants it tends to the alternative development of
one or the other kind of sex organs, giving, in comparison to the fundamental
condition of hermaphroditism, a one-sided sterility. There is incomplete
development or abortion of one or the other of the sex organs which is
discriminating and which results in alternative development, with, however,
many grades in the relative development. Thus, in plants, the so-called
"sterile" intersexes are, in general, individuals that are predominantly male
and often highly functional as such. These individuals are sterile only in
the sense that they are fruitless. Also the so-called "self-sterile " individuals
and varieties of plants, as is well shown in the cultivated grapes in which
sterility from intersexualism is well marked, are predominantly female and
able to function feebly or not at all as males. They are productive of fruit
only when properly pollinated from male or hermaphroditic individuals.
Very seldom, if ever, is complete sexual impotence for a plant as a whole
seen as a condition of intersexualism, as is frequently the case in sterility
from hybridity.

But in many cases of intersexualism in animals, to which attention has
recently been especially directed, the complete sterility of individuals is
very frequent. Here, however, the condition arises in dioecious forms and
involves the partial change of an organ from one sex to the other after

1 Contribution from the New York Botanical Garden, No. 239.

57

58 AMERICAN JOURNAL OF BOTANY [Vol. 10,

differentiation has been partly achieved. This often results in a more or
less complete sexual impotence or sterilization, a condition which has
naturally been very generally regarded as abnormal and pathological. This
is, however, not the case in those species of plants which are prevailingly
dioecious or monoecious, for here, as well shown in the muskmelons, there
is a tendency to produce flowers that are fully functional as hermaphrodites.

Whether, however, intersexualism results in complete sterility, as it
frequently does in dioecious animals, or in one-sided sterility, as is the rule
in hermaphroditic plants, the physiological basis for these variations in sex
is to be regarded as most fundamental in the determination and expression
of sex.

It is, furthermore, to be recognized that the mixture of sexes, with
blending and changes in the character of the organs, often results in a wide
range of variation in the morphological character of the different sex organs
produced by a single individual. In many plants, the flowers on a single
individual may be staminate, pistillate, and hermaphroditic, with also many
intergrading types, thus exhibiting many grades of sexual impotence with
marked differences in the ability to produce fruit.

These cases of partial variability in sex are of special interest, for here
the various conditions of alternative impotence with corresponding irregu-
larities in the production of fruit are all seen among the flowers of a single
individual. In such cases there is also opportunity to observe whether the
variations are irregular and sporadic or whether they are related to a
definite period in development or are otherwise periodic. It is with special
reference to these questions that the changes in the character of the flowers
of Cleome spinosa L. are here reported as decidedly alternative and re-
peatedly cyclic, resulting in the intermittent production of fruit.

Observations on Cleome spinosa

This species is most favorable material for a study of variation in the
sex of flowers in relation to the development of the plant as a whole. It
has long been known as having mixed flowers, yet the species has not
become dioecious. All the individuals of the species are apparently quite
alike in respect to the general range of variations in the sex of the flowers.
The species is a quick-growing herbaceous annual. The first flowers open
on the main raceme when the plant is relatively small — about two feet tall
— and while the lateral branches are scarcely visible. The main raceme
continues to elongate, producing flowers daily, often for a period of from
eight to twelve weeks. Meanwhile a dozen or more lateral branches develop,
and these may in turn branch. All the branches grow rapidly and produce
flowers in abundance. When autumn arrives, well-grown plants are five
or more feet tall with a spread of branches of as many feet in diameter.
There has been a long period of bloom, often covering as many as ninety
days, and this has been for the most part coincident with the period of

Feb., 1923]

STOUT

ALTERNATION OF SEXES

59

rapid vegetative development. By far the greater amount of the vegetative
growth of the plant takes place after the blooming begins. The first fruits
are ripe and shed their seeds when the plant is only about half grown.
The period of vegetative vigor overlaps that of the flowering and repro-
ductive vigor in a decided degree and to an extent seldom seen in plants.
Only during the last few days of bloom do vegetative growth and vigor
noticeably wane.

Fig. 1. Main raceme of a plant of Cleome spinosa at the close of the blooming period,
showing the intermittent production of fruit. This raceme was about 2>h feet long. The
pods on the lower half have dehisced.

At the close of the growth of the plant, about September 15th to October
5th as grown at the New York Botanical Garden, the main branch of a
plant from the point where the flowers are first produced appears as shown

60 AMERICAN JOURNAL OF BOTANY [Vol. 10,

in figure I, the raceme being about three feet in length. Lateral branches
are similar but frequently somewhat shorter. It is to be noted that the
pods are in groups separated by sections of the stem upon which no fruit
was formed. Fruit formation is therefore decidedly intermittent.

This habit of producing fruit intermittently was observed by the writer
in groups of plants grown in ornamental planting in the Botanical Garden
during previous years. For the purpose of making special observations on
the conditions involved in the intermittent production of fruit, a crop of
128 plants was grown in 1921. These plants were examined frequently
throughout the entire period of bloom, and records were taken for each
individual plant as to the character of the flowers opening at a particular
time. At the end of the season, observations on the distribution of fruit
in regard to the record for the flowers was made. Controlled self- and
cross-pollinations were made on many plants.

Every one of the 128 plants produced many pods and the seeds were
numerous, but without exception there was decided intermittency in the
production of fruit. On several plants there was considerable irregularity
in the distribution of pods, but for most plants the pods were in several
groups quite as shown in figure I.

The study of the flowers from day to day together with the results of
controlled pollination showed that the intermittent production of fruit is
due to repeated cyclic changes in the morphological character of the flowers,
which in the course of the cycles give many grades of intersexes. The
flowers of any individual plant varied from perfect or fully hermaphroditic
flowers to flowers that were functional only as males or only as females,
with also innumerable intergrades as to the relative abortion of pistils or
stamens. As a rule, however, the loss of sex is decidedly one-sided. When
the flowers are hermaphroditic or are female, fruit is produced provided
pollination is accomplished; when the flowers are male only, no fruit is
produced. The plants pass through alternating periods when the flowers
are predominantly hermaphroditic or are female, during which they are
productive of fruit, to periods when the flowers are predominantly or only
male and fruitless.

The sex character of the flowers, therefore, varies in cycles, which varia-
tion makes the intermittent production of fruit a necessary result. The
main raceme shown in figure 1 bloomed for a period of 107 days, and on
it were produced about 250 flowers. During this time there were for this par-
ticular raceme five periods when hermaphroditic and female flowers were
produced, with intervening periods when the flowers were staminate only.

In selecting material to illustrate these changes in sex, flowers opening
on the same raceme at the same time were taken, the selection being made
at a time when the variation was marked. It is, however, seldom that the
variation on any one date represents the complete range observed for a
plant during a complete cycle. As is shown in the figures, the sex organs

Feb., 1923] STOUT ALTERNATION OF SEXES 6l

(pistils and stamens) when fully developed are large and conspicuous, and
it is easy to observe variations in the degree of their development. The
petals were removed from the flowers shown in Plate VI.

The three flowers shown in figure 2, Plate VI, were situated in a raceme
in the succession shown and illustrate the range seen for the plant on
the particular date when the photo was made; all the flowers were fully
and very uniformly male; but the pistils were either normal and functional
as in a, decidedly aborted and functionless as in c, or less conspicuously
aborted as in b. On the particular date the flowers were varying in respect
to femaleness. On other dates, however, maleness was quite as variable.

Maleness is well developed in all the flowers shown in figure 3, although
the filaments vary in length and there is one stamen fully aborted in b and
one in c. The pistils vary from the highly developed and functional as at
a to the extremely aborted as in d. Figure 4 represents flowers of the same
plant a few days later, showing extreme abortion in stamens of some flowers
and some variation in the size of pistils, although all pistils were functional.

The pistils of the various flowers shown in figure 5 are either well devel-
oped (a, b, and c), or decidedly aborted (d, e, and/). The development of
stamens is very irregular, and the extremes are seen for a single flower in
the various grades as to length of filament and size, development, and
dehiscence of anthers. Such irregularities as these are frequently seen, and
for numerous plants the condition was more or less present throughout the
period of bloom, with, however, no pod production for the flowers that had
only aborted pistils.

Figure 6 shows two typical flowers of a plant on a date when the flowers
could function only as females.

The four flowers shown in figure 7 show grades of abortion in both
pistils and stamens and illustrate very well how the abortion tends to be
one-sided, affecting first stamens and then pistils.

The many controlled pollinations that were made revealed that there
was in these plants no limitation to fertility through physiological incom-
patibilities in fertilization. Every plant was highly productive of seed
whenever pollen of dehiscing stamens was used on well-developed pistils
either in self- or in cross-pollinations. Rudimentary pistils always failed to
set seed. Pistils over 2 cm. in length usually produced seed.

Examination of pollen and tests for germination were made of pollen
from all sorts of anthers. In large, well-developed anthers, 95% or more
of the pollen grains appeared to be normal, and on a sugar-agar medium as
many as 80% often germinated, producing tubes as long as 750 n. In such
rudimentary stamens as at d, figure 4, only a few shriveled, empty, partly
developed pollen grains were present which did not even swell up when
placed in water. In the large-sized but indehiscent anthers of short stamens
as in c, figure 4, varying percentages of the pollen appeared to be normal,
but in no case did the pollen of such indehiscent anthers germinate when

62 AMERICAN JOURNAL OF BOTANY [Vol. 10,

removed and placed on the same medium which gave good germination for
the pollen of normally dehiscing stamens. There were many anthers that
were partly dehiscent, that is, they opened to some extent, and the pollen
thus shed was often viable in tests and productive of seed in controlled
pollinations.

As a rule, the loss of sex for individual flowers was one-sided. When
the pistil was rudimentary the stamens tended to be well developed as at
c, figure 2. When the stamens were aborted the pistil was usually of good
size as at c, figure 4. Occasionally, however, the pistil in flowers with
aborted stamens was undersized as at d, figure 4, but cases of complete or
extreme abortion of all stamens and of the pistil in the same flower were
not observed.

The entire lot of plants were grown throughout under very uniform
conditions which favored continued vegetative vigor, and only one genera-
tion has been critically studied. Development under conditions which
affect differently the vegetative vigor and the length of the growing period
may influence and possibly decidedly change the behavior in respect to
cyclic changes, quite as such conditions are known to affect the sex of certain
plants, particularly of A risaema triphyllum (Pickett, 1915; Schaffner, 1922),
from year to year. Definite evidence regarding the direct or indirect
influences of environment and the somewhat synchronous changes of sex in
the spider flower remains to be obtained.

At the close of the season, when the vigor of plants perceptibly wanes,
all parts of the flower, corolla, pistils, and stamens alike, are uniformly
undersized. Whether the last flowers that are produced on a plant that
reaches old age before being killed by freezing temperatures are as a rule
predominantly pistillate or staminate was not determined with certainty.
On many such plants the last flowers were decidedly weak in maleness, but
for other plants such flowers were decidedly male or bisexual.

Summary

In the cultures of Cleome spinosa grown for this observational study
there was wide variation in the morphological character of the flowers in
regard to the relative development of the two kinds of sex organs. The
entire range of variations was seen among the flowers of a single plant,
giving bisexual flowers, flowers that were functional only as males or as
females, and many intergrading types. The variation from one extreme to
the other was repeatedly cyclic, which condition results in the intermittent
production of fruit.

All of the 128 plants grown under special observation were quite similar;
all exhibited the extreme ranges of flower forms or intersexes; in all the
production of fruit was more or less intermittent; none was exclusively
staminate or pistillate.

feb., 1923] stout alternation of sexes 63

Discussion

The alternative loss of maleness and femaleness in the flowers of Cleome
spinosa and the recurring periodic changes in the sex of the flowers are to
be regarded as phenomena of internal and biogenetic regulation closely
related to those influences which determine the development of the plant
as a whole.

It is, of course, well recognized that in plants as contrasted with animals
there is continually the formation of really new organs from a persistent
embryonic complex of cells and that this continues until the maturity and
death of the plant as a whole. Internal and biogenetic relations of corre-
lation and self -regulation, operating independently or in response to external
influences, are hence repeated successively in determining the character of
the new organs in the same fashion as they operate once for all in the animal.
When there is in addition a long flowering period which overlaps and is
coincident with the period of the most vigorous vegetative development,
as is the condition in this robust annual Cleome spinosa, the conditions are
most favorable for a study of the factors influencing the differentiation
of sex.

The fact that the loss of sex organs in the flowers of the spider flower is
very decidedly one-sided and qualitative is of special significance. When
the stamens are aborted the pistil is as a rule functional, and in many cases
it is well developed ; when the pistil is aborted the stamens are often highly
developed and functional. Here, as is the rule in plants, intersexualism
does not lead to sterility of the plant or of a flower as a whole. Not a
flower was found in which the pistil and all the stamens were extremely
aborted, and rather rarely was the abortion of one sex accompanied by
the decided abortion of the other sex in the same flower. Abortion
of pistils was frequently accompanied by irregular abortion among the
various stamens of a flower, but the same irregularity in maleness was also
seen for flowers in which there was no abortion of the pistil (see the flowers
of fig. 4). While the expression of sex in at least half of the flowers of a
plant is decidedly one-sided and alternative, it is not mutually exclusive,
for on every plant many bisexual flowers are produced.

It should be noted that the influences operate primarily and almost
discriminatingly on the organs of sex. The pedicels, sepals, and petals are
often uniformly well developed for all the types of flowers ; but undersized
flowers were to be seen (c of fig. 3, and d of fig. 4) in which the flower as a
whole is undersized. Such cases, if more general, would suggest a direct
relation to waning vigor and decreased food supply such as may be con-
sidered to be the direct cause of undersized flowers and of loss of sex in
gynomonoecious forms at the end of a period of bloom. That the conditions
are more intricate in Cleome is evident, for in a marked degree the extreme
variations in sex are independent of any other visible change and the various
grades of intersexes are present from the beginning of bloom.

64 AMERICAN JOURNAL OF BOTANY [Vol. 10,

Furthermore, the influences that affect the sex of a single flower often
extend to groups of flowers. Thus there is a period of maleness, which is
followed by a period of femaleness or of bisexualism, and this in turn is
followed by maleness. Flowers in the same condition as to sex are grouped
along the raceme. There is a series of cyclic changes all occurring during
the period of continuous bloom.

These qualitative changes in sex in flowers of Cleome spinosa do not
involve the transformation of organs of one sex into organs of the other sex
after differentiation has begun, as is the case in many of the intersexes
reported in animals (Goldschmidt and Poppelbaum, 1914; Goldschmidt,
1916; Banta, 1916; Lillie, 1917; Sexton and Huxley, 1921). Here the
change is accomplished, as it is in dioecious plants, by the abortion of one
or the other kind of sex organs. The relative position of each in the flower
as a whole is maintained, but the differentiation giving male and female
flowers (along with bisexual flowers) is as complete as is seen in many species
of dioecious plants. The differential determination of sex in repeated
cyclic alternative changes as they occur in Cleome spinosa shows to what
degree the internal correlative differentiations in development may be
extended to the organs of sex after the plant as a whole has passed from the
exclusively vegetative to the reproductive stage. At the time of the transi-
tion to the reproductive stage, the change is not necessarily complete and
discontinuous, nor are the flowers produced in succession necessarily of the
same grades of sex. Even when the flowers appear to be morphologically
the same there may be a decided cyclic change in their physiological char-
acter, as is the case with Brassica chinensis and B. pekinensis (Stout, 1922).
The contrast between these species of Brassica and Cleome spinosa illustrates
well the different types of sterility that may develop in plants and the
different expressions of cyclic regulation of them. In these Brassicas there
is frequently rather decided abortion of flowers at the time of transition
from vegetative to reproductive organs; in Cleome no indication of such
abortion is present, the first flowers to appear being often fully developed
as hermaphrodites. In the Brassicas there is a somewhat extended period
of flower formation with flowers all morphologically bisexual — but in which
the physiological relations in fertilization may vary in a very definite and
single cycle; in Cleome spinosa there is no variation in the physiological
nature of stamens and pistils that are at all functional in so far as these
may be tested by the relations of fertilization, but there is the cyclic alter-
nation in the morphological development of the organs of sex. This
comparison illustrates two rather widely different expressions of sex in its
relation to fertility and sterility.

The conditions in Cleome spinosa favor the view that, as held by Yam-
polsky (1920), there is a general tendency away from hermaphroditism
toward dioecism among the higher plants. In the persistence of perfect
flowers in greater or less numbers along with those which are more or less

Feb., 1923] STOUT — ALTERNATION OF SEXES 65

purely staminate or pistillate, Cleome spinosa is like most species which are
in the transition stages toward dioecism. The alternate appearance of
male, female, and hermaphroditic flowers in a raceme of course favors cross-
ing, and when this alternation tends to be synchronous on all the branches
of a plant, selfing is largely prevented except in the case of the hermaph-
roditic flowers.

In the spider flower, with its long flowering period and its alternation of
maleness and femaleness in the racemes, it is evident that practically the
whole vegetative feeding power of every plant is drawn upon for seed
production. The conditions are markedly different, and we may consider
them more highly adaptive to the demands of reproduction, than is the
case in strictly dioecious plants in which seed production is confined to one
of each pair of male and female plants. We may, perhaps, characterize the
sex conditions in Cleome spinosa as effecting a sort of super -dioecism in that
the conditions favor both reproduction and crossing for each individual.

Certain points regarding the determination of sex in the flowers of
Cleome spinosa are clear. The conditions illustrate well the fact that the
morphological differentiations of sex are fundamentally an extension of the
phenomena of somatic differentiations. The expressions of differential
qualities in leaves, stems, and flowers, with further differentiation of calyx,
corolla, pistil, and stamens, with still further differentiations of tissue within
each, are all recognized as one-sided, qualitative, and alternative expressions
in protoplasmic units that are alike and which still remain alike in funda-
mental constitution. Even the physically qualitative division of germ
plasm in the reduction divisions is found in regeneration experiments and
in parthenogenesis not to be a direct and absolute condition in the alter-
nation of generations. The theory of sex chromosomes decidedly fails in
general application to plants, and even in animals, where its application
seems most marked, sex is often intergrading and reversible, showing that
there is alternative expression rather than alternative inheritance.

In Cleome spinosa it is evident that there are rather special and perhaps
very specific stimulating and inhibiting influences which regulate the
development of the sex organs. Whether these influences are substantive
or more of the nature of stimuli, their action is cyclic and decidedly alter-
native. The results clearly show that sex of flowers is determined progres-
sively as they are formed in response to regulation by internal biogenetic
conditions.

New York Botanical Garden

LITERATURE CITED
Banta, A. M. 1916. Sex intergrades in a species of Crustacea. Proc. Nat. Acad. Sci.

2: 578-583-
Goldschmidt, R., and Poppelbaum, H. 1914. Erblichkeitsstudien an Schmetterlingen II.
Zeitschr. Ind. Abst.- Vererbungsl. 11: 280-316.

66 AMERICAN JOURNAL OF BOTANY [Vol. 10,

Goldschmidt, R. 1916. Experimental intersexuality and the sex-problem. Amer. Nat.

50: 705-718.
Lillie, F. R. 1917. The free-martin: a study of the action of sex hormones in the foetal

life of cattle. Jour. Exp. Zool. 23: 371-452.
Pickett, F. L. 1915. A contribution to our knowledge of Arisaema triphyllum. Mem.

Torr. Bot. Club 16: 1-55.
Schaffner, J. H. 1922. Control of the sexual state in Arisaema triphyllum and Arisaema

Dracontium. Amer. Jour. Bot. 9: 72-78.
Sexton, E. W., and Huxley, J. S. 1921. Intersexes in Gammarus chevreuxi and related

forms. Jour. Mar. Biol. Assn. United Kingdom 12: 506-556.
Stout, A. B. 1922. Cyclic manifestation of sterility in Brassica pekinensis and B. chinen-

sis. Bot. Gaz. 73: 1 10-132.
Yampolsky, C. 1920. The occurrence and inheritance of sex intergradation in plants-

Amer. Jour. Bot. 7: 21-38.

EXPLANATION OF PLATE VI

Each group of flowers is from the same raceme on the same date. Petals have been
removed. The scale in inches shows the reduction from natural size.

Fig. 2. Three sister flowers uniform as to maleness but highly variable in femaleness.

Fig. 3. The stamens of the four flowers vary as to length; one stamen in b and one
in c are aborted; all others are highly functional. Pistils are functional in a, b, and c, but
aborted in d.

Fig. 4. Four flowers from same plant as those of figure 3, several days later. Pistils
vary as to length but all are functional. Stamens all excellent in a, two much aborted and
indehiscent in b, all indehiscent in c, and all much aborted and containing no pollen in d.

Fig. 5. Pistils variable; in a, b, and c, fully functional; in d, e, and /, rather aborted.
Stamens highly variable in each flower irrespective of the condition of the pistil.

Fig. 6. All stamens much aborted or rudimentary. Pistils somewhat undersized
but functional.

Fig. 7. Abortion of stamens only, as at a, or of pistil, as at c and d, with a flower (b)
fully hermaphroditic; all in bloom at the same time on the same raceme. Illustrates well
the marked one-sided abortion of sex organs.

American Journal of Botany.

Volume X, Plate VI.

a 'b cd abcde'/1

H 5

2 b

Stout: Alternation of Sexes

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN

No. 240

STERILITY IN LILIES

A. B. STOUT

NEW YORK
1923

Reprinted, without change of paging, from The Journal of Heredity, 13:309-373. 25 Ap 1923

STERILITY IN LILIES

A. B. Stout
Nczv York Botanical Garden, New York Citv

THE tendency of many species of
lilies to produce seed irregularly,
sparingly, or even not at all, is
well exemplified in the old familiar
Tiger Lily of our flower gardens. This
species has been in cultivation in Eu-
rope since 1804 ; it thrives and blooms
abundantly over a wide area ; it has
been observed in cultivation and also
apparently wild in its native home
( China and Japan ) by persons inter-
ested in discovering whether fruits
were formed. Yet the author is aware
of only two references to the pods and
seeds of this species, and the accuracy
of these is somewhat doubtful. On
the other hand, its fruitlessness has
repeatedly been observed wherever it
has been grown. The Tiger Lily is one
of the most striking examples that can
be cited of a plant which blooms in
abundance but is propagated only by
vegetative means.

The Law of Compsnsation

A ready explanation has long been
offered by gardeners and botanists for
the condition of sterility found in such
plants as the lilies. It has been gener-
ally believed that they are physically
unable to develop seeds because the
bulbs and bulblets use the available
food. The tiny embryos were assumed
to be virtually starved to death. This
explanation of the condition has seemed
so obvious and so correct that it has
received the sanction of the most crit-
ical authorities, and is stated in Dar-
win's law of compensation and in
Goebel's law of correlation. But this"
simple and apparently very satisfactory
explanation we now know to be entirely
wrong, at least in its application to the
condition of sterility in the lilies.
r> Obviously the best evidence of the

truth of this statement is the fact that
these usually fruitless plants have been
shown to be fully able to produce
capsules and seeds. Such evidence is
in most cases readily obtained.

Sterility Due to Incompatibility of
Pollen

For a period of about fifteen years
Focke persisted in attempts to obtain
seed of L. bulbiferum. He secured
bulbs from various parts of northern
and western Germany and made many
cross-pollinations of their flowers, but
obtained at best only a few poor cap-
sules. Finally, bulbs were secured
from Switzerland and with the use of
pollen of plants grown from them pods
and seeds in abundance were matured
on the hitherto fruitless plants.1

The writer has had much the same
experience with other species of lilies.
A fine large cluster of L. croceum, all
descended from a single bulb, growing
in the New York Botanical Garden,
completely failed to set seed for eight
consecutive years. The numerous
flowers that appeared each year
were self- and cross-pollinated, but
there was never even a slight en-
largement of any of the ovaries. In
1 92 1, plants of L. clcgans bloomed
at the same time as the cluster of
L. croceum, and their pollen was
used in pollinations. Fine large pods
then developed, and these contained
many viable seeds. In 1922 these
plants of L. croceum also yielded
pods and seed by cross-pollina-
tion with another strain of the same
species. Hence these plants were fully
able to mature fruit, and their pollen
was likewise able to function in cer-
tain reciprocal crosses.

1 Several papers by W. O. Focke bearing on phenomena of self-incompatibility were pub-
M lished between the years 1887 and 1893.
^ 369

370

The Journal of Heredity

SELF-STERILE, BUT PRODUCING PODS WHEN CROSSED

Figure 17. This plant of Lilium spcciosum is fully self-incompatible, but yields fine
pods to pollen from certain other plants of the same species, which presumably indicates
their origin from different seeds. This, of course, does not insure compatibility, as plants
known to be of distinct seed-origin are often incompatible. Self-incompatibility is also
met with when large plantings of fruit trees of the same variety are made. Unless two
or more distinct varieties of different seed origin are inter-planted, no fruit is produced.

Stout: Sterility in Lilies

371

SEED OF THE SEEDLESS TIGER LILY

Figure 18. While the Tiger Lily has thus far failed to yield pods to intra-specific pol-
lination, many have been obtained to pollen of L. maximowiczii, as is proved by the two
fine pods on the plant at the right. To the pollen of L. zvarleyensc only imperfect pods
were developed, as shown in the two upturned but partly developed pods on the left.
The Tiger Lily is evidently able to produce seed and its usual sterility is due to physiological
incompatibility of the pollen used, and is not in any way connected with the fact that
the plant habitually reproduces by means of bulbs.

In an effort to discover compatible
fertilizations that would produce seed
on the Tiger Lily, bulbs of this species
have been secured from widely sepa-
rated localities, some even coming di-
rectly from apparently wild plants in
China. All plants grown from these
have, so far as tested, failed to yield
fruits either by self-pollination or by
cross-pollination between the different
strains. Capsules also failed to form
when pollen of such species as L. can-
adcnsc, L. supurbum, L. hcnryi, L.
speciosum, L. auratum, and L. hiim-
boldtii was used. To pollen of L. zvar-
leyensc, the ovaries of the Tiger Lily
began to enlarge and become upturned
(see Fig. 18), but the capsules did not
develop fully and yielded no seeds.
But when pollen of L. maximowiczii
was used fine pods were readily ob-
tained (see Fig. 18). Potted plants.

plants grown in the open, old plants
of vigorous vegetative growth, and
plants from bulblets blooming for the
first time have all produced fine pods
from the flowers thus pollinated. Yet
on these same plants complete failure
resulted from all attempts to produce
seed by self-pollination.

It is clear that these usually sterile
plants are able to form fruit and seeds
and that their pollen is potent, provided
there are compatible relations in fer-
tilization. The bulb and the bulblet-
forming habits of growth, and of veg-
etative propagation do not render seed
production impossible. The sterility is
that of physiological incompatibility in
fertilization, a well-marked type which
is also seen in numerous species that
naturally reproduce exclusively by
seeds, and is by no means peculiar to
plants that reproduce vegetatively.

372

The Journal of Heredity

In the course of experimental studies
with species of Lilium now in progress
at the New York Botanical Garden, con-
trolled pollinations have been made for
plants of twenty-eight species. In
every species thus far studied seed
production is decidedly limited by in-
compatibility in fertilization.

The phenomena of incompatibility
are well illustrated by the results ob-
tained with L. regale, a splendid and
beautiful lily recently introduced from
China. Of the ten plants tested, nine
were self-incompatible and one was
self -compatible. Cross pollination be-
tween plants sometimes succeeded and
sometimes failed. The same sort of
results were also obtained with plants
of L. speciosum, L. candidum, L.
henry i, L. parvum, L. tenuifolium, L.
humboldtii, L. kelloggii, L. zvarleyense,
L. longiflorum, and L. auratum.

No capsules have yet been obtained
on plants of L. hansonii, L. parry i, L.
maximowiczii, and L. choice donicum,
but in all cases except that of L. han-
sonii, only a few plants have been
tested, and these were all obtained from
a single source.

Cross-Incompatibility

It is to be noted that crosses between
different plants of a species in which
there is self -compatibility are also fre-
quently incompatible. One is, however,
not at all certain that any two plants
of a variety of lily are really of differ-
ent seed origin, for they both may be
the daughter bulbs of a single parent
plant, and thus belong to a single
clone. If the original plant grown
from seed was entirely self -compatible,
then one may expect the members of
the clone to fail to set seed in crosses.
The condition is similar to that en-
countered in fruit-growing, when an
entire orchard of a self-incompatible
variety of apple or cherry fails to set
fruit unless proper cross-pollination
with another variety is provided for.
Without doubt it is often the extensive
vegetative propagation of a single self-
incompatible plant that makes it diffi-
cult to secure compatible crosses from

apparently different plants of certain
varieties of the day lilies and the true
lilies.

Seedlings of L. longiflorum have
been grown to the flowering age in the
course of the studies here reported.
Many of these were self-incompatible,
a few set seed with their own pollen,
and many crosses with other plants
failed. Such ranges of variation in
regard to the compatibilities of sister
plants of seed origin are very common
in many species.

Other Types of Sterility

There are other types of sterility in
the lilies besides that of incompatibility.
In certain hybrid varieties. L. bate-
manniae, for example, there appears to
be complete impotence, a type of ster-
ility that is very characteristic of hy-
brids ; the spores, both in pistils and
stamens, are aborted and functionless
in any relation. A one-sided abortion
characteristic of intersexes has also
been noted in at least one species.

How to Obtain Seed

The principal condition limiting seed
production in the lilies is that of in-
compatibility. When seed is desired
from plants that are yielding no seeds
because of this condition, it is simply
necessary to obtain other stocks of the
same species and interplant so that
cross-pollination may be effected by
insects, or by hand. If the new stock
is cross-compatible, seed will readily be
produced, but if it is not, seed will not
be formed until stock which is compat-
ible is secured. Usually compatible
clones may readily be obtained, and
often a single planting of commercial
bulbs contains such a mixture of stocks,
but for certain species and especially
for L. tigrinum, L. hansonii, and L.
candidum, it seems somewhat difficult
to secure compatible fertilization.

One of the principal sources of
failure in establishing a planting of
lilies in one's garden is the poor phys-
ical condition of bulbs when received.
Especially is this true when bulbs are
imported. In many species the dormant

Stout: Sterility in Lilies

373

RESULTS OF POLLINATION ARE QUICKLY EVIDENT

Figure 19. Both of these plants of L. longiflorum are self-incompatible and four
crosses with other species likewise failed. The three growing (pods are from other more
successful crosses. The photograph was taken about two weeks after the last pollination
was made.

period is very short or is entirely ab-
sent. With the death of the top there
is usually continued development of
the daughter bulbs, and growth of
feeding roots. Without doubt the best
means of securing healthy and vigorous
bulbs, of many species at least, is to
grow them locally either from seed or
by vegetative propagation, and to trans-
plant so that the bulbs are only out of
the ground for a short time, when
they are most nearly dormant.

Such excellent varieties as the Easter
Lily, the Madonna Lily, Henry's Lily,
the Lance-leaved Lily (L. speciosum),
and the Regal Lily are readily grown
from seed. Frequently blooms may be

had on seedlings in two years — cer-
tainly they may be had in three years.
The growing and breeding of lilies
from seed presents many interesting
possibilities, not only to the commercial
growers of bulbs but to the amateur
gardener as well.

In lilies the bulb habit of growth
does not in the least affect seed pro-
duction. To obtain seed one must
either grow self-compatible plants or
grow strains that are cross-compatible.
When this is done, seed in abundance
may readily be obtained except in the
relatively few cases of impotence from
intersexes or hybridity.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 241

NOTES ON ROSACEAE-XIV

PER AXEL RYDBERG

NEW YORK

1923

Reprinted, without change of paging, from Bulletin op the Torrey Botanical
Olob 50: 61-71. February 23, 1923.

Lffi*A*T

Notes on Rosaceae — XIV

Per Axel Rydberg

Roses of the Prairies and Plains
The area here treated comprises the states between the
Mississippi River and the Rockies and north of Oklahoma, with
the exception of Missouri, in which so many southern plants
are found and in which the Ozark Mountains have a different
flora from the plains. As a key to the groups has already been
published in a preceding paper,* I shall here omit the same,
and I shall omit also the general discussion of any species
already treated there, only citing specimens collected in the
region now under consideration.

I. SYNSTYLAE

Leaflets glabrous or slightly pubescent on the veins be-
neath, dark green and shiny above. I. R. setigera.
Leaflets velutinous-pubescent beneath, rather dull above.

2. R. rubifolia.

i. Rosa setigera Michx.
Kansas: Cherokee County, 1896, Hitchcock 676.

2. Rosa rubifolia R. Br.
Kansas: Manhattan, 1892, Clothier.

II. CANINAE

3. Rosa rubiginosa L.
Kansas: Riley County, 1896, /. B. Norton 677; 1897,
Clothier 1107.

*See Bull. Torrey Club 45: 45-66. 1920.
H

61

62 Rydberg: Notes on Rosaceae — XIV

III. CAROLINAE

Leaflets not subcoriaceous.

Infrastipular prickles decidedly curved, short, stout. 4. R. palustris.

Infrastipular prickles straight (rarely slightly curved),
slender.
Leaflets decidedly pubescent beneath. 5. R. Lyoni.

Leaflets glabrous or pubescent only on the veins be-
neath.
Leaflets not glandular-dentate, the teeth usually

simple. 6. R. Carolina.

Leaflets glandular-dentate, often with double

teeth. 7. R. serrulata.

Leaflets subcoriaceous, densely pubescent beneath. 8. R.rudiuscula.

4. Rosa palustris Marsh.
This has been reported only in the Mississippi and Minnesota
River valleys, where the more pubescent western form has been
found.

5. Rosa Lyoni Pursh
Kansas: Cherokee County, Hitchcock 978.
Iowa: Iowa City, 1916, Wylie.

Also reported from Minnesota.

6. Rosa Carolina L.

This has been reported from Minnesota and Kansas, but
no specimens have been seen. It is found also in Wisconsin
and Missouri.

7. Rosa serrulata Raf.

This has been reported from Iowa, but no specimens are
now c'.t hand.

8. Rosa rudiuscula Greene

This belongs to the Carolina and not to the Cinnamomea
group, having the hypanthium bristly and with the achenes
inserted in the bottom only. The infrastipular prickles are
small and scarcely stronger than the bristles. It also has spread-
ing and deciduous sepals; and the leaflets are thicker and firmer
than in R. Carolina and R. Lyoni. The flowers are usually
corymbose and the plant has the general aspect of R. suffulta.
Dr. Greene in the original publications cited several specimens
that do not belong to the species.

Rydberg: Notes on Rosaceae — XIV

63

Missouri: Vale, 1906, Bush 3916; Little Blue, Bush 208
(Type); Dodson, Jackson County, Bush 160; railroad north
of Morley, 1893, Eggert; Grain Valley, 6998.

Iowa: Amaxa, 1909, M. P. Somes 3825.

Indiana: White County, southeast of Wolcott, Deam 32712;
Laporte County, 32395.

IV. CINNAMOMIAE

Infrastipular prickles wanting; branches unarmed
or bristly, not prickly.
Inflorescence corymbose, terminating the stem
(or rarely th^ branches); plant mostly
suffruticose; stem very bristly, usually
dying back to the ground (except in
R. polyanthema and R. Bushii); leaflets
9-11.
Leaves glabrous or nearly so.

Leaflets obovate, pale and glaucous on

both sides.
Leaflets elliptic or oval, dark-green above.
Leaflets mostly 2-5 cm. long, not
glaucous beneath, acute at the apex;
plant 3 dm. high or more; flowers
several.
Leaflets rarely more than 2 cm. long,
mostly rounded at the apex, some-
what glaucous beneath; plant 1-2
dm. high; flowers 1-3.
Leaflets densely pubescent, especially beneath.
Plant 3-6 dm. high, usually dying back to
the ground.
Sepal erect in fruit; plant not glaucous.
Upper stipules and bracts not
densely glandular; leaflets 1.5-4
cm. long; plant 3-5 dm. high.
Upper stipules and bracts densely
glandular; leaflets seldom more
than 1.5 cm. long; plant 1-3 dm.
high.
Sepals reflexed in fruit; plant glaucous.
Plant shrubby, not dying back; branches
from the upper axils soon overtopping
the inflorescence.
Inflorescence of few, either solitary or corymbose,
flowers ending lateral branches; plant
shrubby leaflets; 5-7 (or rarely 9 on the
shoots).

9. R. subglauca.

10. R. arkansana.

n. R. Lunellii.

12.

R. suffulta.

13.

R. alcea.

14-

R. conjuvcta

15-

R. polyanthema

64

Rydberg: Notes on Rosaceae — XIV

1 6. R. Engelmanni.

17. R.acicularis.

18. R. Bourgeauiana.

19.
20.

R. blanda.
R. subblanda.

Stem densely bristly even in age; stipules,
petioles, and rachis often glandular-
granuliferous; flowers solitary.
Hypanthium decidedly pear-shaped or
elliptic, acute at the base, with a
distinct neck at the top.
Leaflets conspicuously glandular-granu-
liferous, but scarcely at all hairy
beneath, rounded -oval, often double-
toothed.
Leaflets densely pubescent only rarely
slightly glandular beneath, mostly
elliptic, simple-toothed.
Hypanthium subglobose, almost without
a neck; leaflets glandular-granuliferous
as well as pubescent beneath.
Stem unarmed or when young slightly covered
with more or less deciduous prickles;
flowers solitary or few.
Leaflets decidedly and finely pubescent

beneath.
Leaflets glabrous on both sides, glossy,
lnfrastipular prickles present.

Hypanthium globose; neck usually obsolete.
Stipules, petiole, and rachis copiously glandu-
lar; leaflets often double-toothed with
gland-tipped teeth.
Stipules, petiole, and rachis not conspicuously
glandular, the first ones sometimes
slightly gland-toothed.
Leaflets glabrous or nearly so.
Leaflets finely puberulent beneath.

Young shoots armed with stout flat
prickles; fruit about 15 mm. broad.
Young shoots merely bristly; fruit
rarely more than 10 mm. broad.
Hypanthium elongate, pyriform or elliptic, with
a distinct neck.

9. Rosa subglauca Rydberg

This is related to R. arkansana Porter and R. suffulta Greene;
from the former it differs in the decidedly obovate leaflets,
rounded at the apex and more coarsely toothed ; from the latter
in the glabrous leaves; and from both in the glaucous foliage.

Saskatchewan: Six miles east of Battle River, August 15,
1906, Macoun & Herriot 70942, 70943-, Breakmore, July 18,
1906, 70941; Brandon, 12389.

21. R. Fendleri.

22. R. Woodsii.

23. R. terrens.

24. R. Macounii.

25. R. pyrifera.

Rydberg: Notes on Rosaceae — XIV 65

A specimen from Decatur County, Iowa, collected by T.
J. and M. F. L. Fitzpatrick may also belong here.

10. Rosa arkansana Porter

This species was described from the Arkansas Canon near
Canon City. The type is a glabrous plant and so described.
The more common R. suffalta of the plains is densely pubescent.
This was mistaken for R. arkansana by Watson, who has been
followed by most American botanists. R. arkansana is found
in the prairie region as well, but is much more local. It seems
to be confined to the river valleys, while R. suffulta prefers the
prairies and plains. R. Rydbergii Greene is nothing but R..
arkansana. The peculiar hue of the upper surface of the leaves,
so much emphasized by Dr. Greene, is due to some disease,
probably bacterial. Some of the leaves or some part of them
have retained their natural color on the upper surface. Best
regarded it as a variety of R. blanda.

Minnesota: Winona, 1882, Hasse Herbarium.

North Dakota: Fargo, Bolley 12Q, 130.

Kansas: Comance County, 1897, Carleton 254.

11. Rosa Lunellii Greene

This is closely related to R. arkansana and perhaps only a
depauperate form of the same. It is, however, a much smaller
plant and more glaucous, and the leaflets are more rounded
at the apex.

Manitoba: Brandon, 1887, Foivler; 1898, E. S. Thompson.

North Dakota: Leeds, 1907 and 1908, Lunell.

South Dakota: Hermosa, 1892, Rydberg 678a.

Nebraska: Long Pine, 1893, W. R. Dudley.

12. Rosa suffulta Greene
The first record of this species is in Hooker's Flora Boreali-
Americana, where Borrer referred it to a variety of R. stricta.
Crepin made it a variety of R. blanda and Watson confused it
with R. arkansana. Dr. Greene was the first to see that it was
not R. arkansana and proposed the name R. pratincola, over-
looking the fact that this name had been used before. Dis-
covering this fact twelve years later, he substituted the name
R. heliophila. C. K. Schneider discovered the same fact a year
later, but not knowing of Greene's last name, proposed R.

66 Rydberg: Notes on Rosaceae — XIV

arkansanoides . All these names are superfluous, however, for
the plant is the same as R. suffulta Greene. Besides these names
several varietal names have been proposed. R. dulcissima
Lunell is apparently based on a more luxuriant and less bristly
form of this species.

13. Rosa alcea Greene

This species stands somewhat in the same relationship to
R. suffulta as R. Lunellii does to R. arkansana. The plant
is described as having prickly hypanthium, but this is not
always the case. The upper part of the plant, especially the
upper stipules and the bracts, are densely glandular, the glands
often extend to the pedicels and hypanthium, the stalks of the
glands sometimes become indurate and hence the hypanthium
is prickly in fruit. This is sometimes also the case in R. suffulta,
which form was described as R. pratincola setulosa by Cockerell.

Saskatchewan: Milk River Ridge, 1895, Macoun 10541;
Moose Jaw, Spreadboro 10625 ;a\so Macoun 1053 g; Regina 1903,
Fowler; Indian Head, 1; Prince Albert, Macoun 12767 (?).

South Dakota: Dead wood, Can 150.

14. Rosa conjuncta Rydberg

This is related to R. suffulta and R. subglabra on the one
hand, and to R. virginiana on the other. It differs from the
last in the lack of glandular hairs on the hypanthium; in the
pale green leaves with sharper toothing; in the absence of
infrastipular spines; and in the insertion of the achenes, which
is on the inside walls as well as in the bottom of the hypanthium.
From R. subglauca it differs in the leaves pubescent beneath
and in the reflexed sepals. It resembles R. suffulta in habit and
pubescence, but the leaflets are more acute at the apex and
cuneate at the base and glaucous, and the sepals are reflexed
or spreading after anthesis. As the plant has been confused
with R. suffulta and R. arkansana, this latter character caused
the error in Britton's Manual, in which R. arkansana (which
then included R. suffulta) was described as having spreading
sepals. I am responsible for the error. R. conjuncta might be
a hybrid of R. suffulta and R. Lyoni or some other species of the
Carolina group, but none of the species of that group or of the
Cinnamomea group (except the high northern R. subglauca) has

Rydberg: Notes on Rosaceae — XIV 67

the glaucous foliage of R. conjuncta. Hence it must be regarded
as a distinct species.

15. Rosa polyanthema Lunell
This species resembles R. suffulta in the leaf-form and
inflorescence, but the plant is shrubby, the upper branches
soon overtopping the inflorescence. It is known only from the
type locality and vicinity.

16. Rosa Engelmanni S. Wats.

In describing this species Dr. Watson included some eastern
specimens which belong to R. acicularis. This has caused much
confusion. Some botanists have regarded R. Engelmanni as a
variety of R. acicularis, some as a mere synonym. R. Engelmanni
is a purely Rocky Mountain species. The most eastern station
from which the writer has seen the same is in the Black Hills
of South Dakota. The specimens referred to it from Minnesota,
Michigan and northern New York belong to R. acicularis.
In R. Engelmanni the leaflets are much more rounded, without
pubescence, but distinctly glandular-granuliferous beneath,
usually double-toothed and with gland-tipped teeth. In R.
acicularis on the contrary the leaflets are inclined to be elliptic,
conspicuously pubescent but scarcely glandular beneath and
with simple teeth.

Most of the specimens seen of R. Engelmanni are from Colo-
rado and a few are from southern Wyoming. The only specimens
from the region here treated are the following, but even these
are not from the prairies and plains but from the mountains or
wooded hills.

South Dakota: Spearfish Canon, Black Hills, Murdock
4136', Rochford, Rydberg 676 (in part); Custer, Rydberg 676
(in part).

17. Rosa acicularis Lindl.

Within the region the species is found only in the northern
wooded part. Among other specimens may be mentioned:
Minnesota: Sandy Lake, Sandberg 793.

18. Rosa Bourgeauiana Crepin

Within the region it is found only in the northeastern wooded
region.

68 Rydberg: Notes on Rosaceae — XIV

Minnesota: Duluth, Rydberg 8004; St. Louis River,
Carlton County, Sandberg 130; Two Harbors, E. P. Sheldon.

19. Rosa blanda Ait.

This species is confined within the region to the hardwood
groves in the northern part. I can see no difference between
the specimens from North Dakota, on which R. gratiosa Lunell
was based, and those from the Eastern States.

Minnesota: Lake Minnetonka, Sandberg, 798, Thompson,
Carlton County, Sandberg 397; Minneapolis, 1895, E. P. Sheldon;
Lake Itasca, G. B. Alton.

North Dakota: near St. John, Rolette County, 1912,
Lunell (R. gratiosa); Pleasant Lake, Benson County, 1912,
Lunell; Turtle Mountains, 1912, Lunell; Bottineau, 1890, Waldron.

Manitoba: Oak River, Macoun & Herriot 70937; six miles
east of Forest, Macoun & Herriot 70940.

20. Rosa subblanda Rydberg

It is questionable if this species should be included among
the prairie roses, as the following specimens are referred to it
with considerable doubt.

Iowa: Grinell, 1877, M. E. Jones.

21. Rosa Fendleri Crepin

This species is related to R. Woodsii and often confused
with it. It is distinguished by the glandular stipules, petioles,
and leaf-rachis. Watson distinguished it from R. Woodsii by
the entire instead of lobed sepals. In both species they are
either entire or lobed. R. neomexiana, R. arizonica, R. puberu-
lenta and R. granulifera have been confused with it, but these
all have curved prickles, while in R. Fendleri they are longer
and more slender, straight or nearly so. R. Fendleri is a common
species in the Rockies, ranging from Montana to northern
Mexico. It extends eastward also into the plains and prairies
and was redescribed from North Dakota by Lunell as R. poetica.

Minnesota: Montevideo, Moyer 581.

North Dakota: Bismark, 1912, Lunell.

South Dakota: Hermosa, Rydberg 677.

Nebraska: McCalligan Canon, Deuel County, Rydberg 100.

Rydberg: Notes on Rosaceae — XIV 69

22. Rosa Woodsii Lindl.

This was published in 1820 in Lindley's Monograph. Five
years later Lindley published an illustration in the Botanical
Register of what he supposed to be the same, but evidently he
was mistaken. In the Botanical Register (pi. gy6) Lindley gave
the following remarks:

It was subsequently named and published by the writer of these remarks
. . . but the specimens which were examined for the purpose, were so im-
perfect that, upon a comparison of the characters ascribed to the species with
fresh specimens, they were ascertained to be materially erroneous; the stipulae,
which were stated to possess the remarkable peculiarity of being convo-
lute like those of R. Carolina, proving to be, in fact, like those of R.
lucida.

But in M. de Candolle's Prodromus a new character is proposed for
this plant. M. Seringe, by whom the article Rosa was prepared, had an
opportunity of examining specimens in De Candolle 's Herbarium. And yet
our original error is still retained by Mr. Seringe, who has added to it more
than one of his own. He defines the leaflets to be shining, while in fact they
are the reverse; the sepals to be naked, which are covered with glands; and
the lower pair of leaflets to be placed at a distance from the others, and fringed
with glands, a peculiarity which we believe does not exist. *

We cannot dismiss this subject without expressing our regret that the
general brilliancy of M. de Candolle's Prodromus should be tarnished by
an article so inaccurately compiled as the genus Rosa is, in the 2d volume of
that work.

These cutting remarks of Lindley 's were wholly unwarranted,
for Seringe did not assign any new characters that were not
found in Lindley's original publication, and it was the latter
himself that assigned new characters. Let Us recite a few lines
from Lindley's own description in his Monograph, page 22.

Leaves without pubescence; stipules very narrow and acute, convolute
and fringed with glands . . . Leaflets 7-9, shaped like those of R.
rubella, shining, flat, simple, acute, paler beneath . . . Fruit naked,
ovate, with short, connivent, entire sepals which are free from glands as is
the peduncle.

From this it is evident that Rosa Woodsii of the Botanical
Register is not the same as the original one described in Lindley 's
Monograph. This carelessness on Lindley's part has caused a
great deal of confusion, and it is hard to know what the
original R. Woodsii was. Some have suggested R. humilis, but
as the pedicels, hypanthium and sepals were without glands
and the latter connivent, this suggestion is far from the truth.

70 Rydberg: Notes on Rosaceae — XIV

Others have suggested R. blanda Ait., but the true R. blanda is
a boreal plant and not found on the Missouri, and the leaves
are dull and pubescent beneath. The R. Woodsii of the Botanical
Register might sooner be a form of that species. The only
species that agrees with the description of the original R. Woodsii
is the one that Torrey called R.foliosa leicocarpa and in my flora
of Colorado I called R. Macounii, the same as Greene has des-
cribed as R. Sandbergii and Lunell as R. deserta. I also think
R. fimbriatula Greene belongs here. It is a shrub belonging to
the Rocky Mountain region but extends eastward to the Missouri
River. Watson referred it partly to R. Woodsii, partly to R.
Fendleri. Watson assigned also new characters to R. Woodsii,
viz. lobed sepals. Notice that Lindley originally described them
as "entire." The lobing or not-lobing of the sepals is a character
of no value in the Cinnamomea group. In other groups as for
instance, the Carolina and Canina groups, it is a fairly
reliable character. Rosa Maximiiliani Nees belongs to this
species.

North Dakota: Pleasant Lake, Benson County, 1912,
Lunell (R. deserta Lunell); Little Missouri River, Moyer 702.

South Dakota: Missouri River, north of White River,
Hayden 254; White River, Stearns.

Nebraska: Cheyenne County, Rydberg 101.

Kansas: Rawlins County, Hitchcock gj8a.

23. Rosa terrens Lunell
This species differs from all the species of the Cinnamomiae
of eastern North America in the stout prickles especially on
the shoots. The prickles are fully as stout as those of the sweet
brier. It is only known from the type collection.

24. Rosa Macounii Greene
It has been shown that the pubescent plant illustrated by
Lindley in the Botanical Register under the name R. Woodsii
has had a very confused nomenclatorial history. Watson
thought it was the original R. Woodsii and stated that it was
the same as R. Maximiliani Nees. At the time when I noticed
the discrepancy in Lindley 's treatments, I thought that Watson 's
statement was correct and adopted the name R. Maximiliani
for the present species. The name was used in that sense in my

Rydberg: Notes on Rosaceae — XIV 71

Flora of Colorado. Later I have had access to Prince Maxi-
milian's Reise and found that the plant described there was the
glabrous plant or the true R. Woodsii. In my Flora of Colorado
I made another mistake, adopting the name R. Macounii
for the glabrous plant, depending upon the impression I had
received from Greene's rather vague description. After having
seen Greene 's type, though this does not represent the common
form of the species as I know it, I came to the conclusion that
the name R. Macounii should be adopted for the species here
treated, being the oldest available name. R. grosseserrata E.
Nelson was a mixture, but the type evidently belongs to the
large-leaved form of this species not uncommon in the Central
Rockies. R. subunda Lunell represents a depauperate form and
R. naiadum the more common form of the plains and prairie
region.

Minnesota: Montevideo, Moyer 372.

North Dakota: Kuhn, La Mouse County, Brenckle;
Minot, Lake Ibsen, Jamestown, Tower, and Butte, Lunell
(without numbers).

South Dakota: Hermosa and Hotsprings (Black Hills),
Rydberg 677; Wolf Creek, Visher 2i8q; Bear Creek, Visher
2032; Mobridge, Moyer 691, Big Stone Lake, 542

Nebraska: Chadron, /. M. Bates; Lawrence Fork, Rydberg
102; Banner County, Rydberg 98; Hay Springs, MacDougal 102;
Nattick, Thomas County, Rydberg 1848.

25. Rosa pyrifera Rydberg
This is a species of the Northern Rocky Mountains, first
mentioned by Watson in a note under R. Fendleri* It differs in
the pear-shaped fruit and the more numerous, rather corymbose
flowers. A few rather dubious specimens have been collected in
the plains and prairie region. These have fruits which are
inclined to be ellipsoid, rather than pyriform. Among them
may be noticed.

South Dakota: Mobridge, Moyer 688.
*Proc. Am. Acad. 20: 345. 1885.

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN

No. 242

THE VEGETATIONAL HISTORY OF THE
MIDDLE WEST

HENRY ALLAN GLEASON

NEW YORK
1923

[Reprinted, without change of paging, from Annals of the Association of American Geographers
Volume XII: pp. 39-85. February, 1923

Annals of the Association of American Geographers
Volume XII, pp. 39-85

THE VEGETATIONAL HISTORY OF THE MIDDLE WEST
HENRY ALLAN GLEASON

CONTENTS Page

Part. I. Introduction

The Dynamic Viewpoint 40

General Control of 'Distribution 40

Vegetational History and Human History 40

Outline 41

Part II. The General Nature of Plant Migrations.

Invasion, Migration, Ecesis 42

Efficiency of Plant Dispersal 42

Migration and Succession 43

Rate of Migration 44

The Time Factor and Migration 45

Significance of the Margin of Range 47

Advancing and Retreating Migrations in the Middle West 47

Effects of Changes in Physiological Requirements 47

Migrations as Related to Continuity of Suitable Habitats 48

The Interpretation of Isolation 49

Adjustments in Distribution 50

General Results of Continued Migrations 50

Summary of Part II 51

Part III. The Evidence of Plant Migrations.

Historical Evidence of Migrations 52

Succession as Evidence of Migration 52

Determining Migration Centers 53

The Five Elements in the Flora of the Middle West 54

Glacial vs. Modern Climate 54

Relic Colonies 55

Summary of Part III 56

Part IV" The Development and History of Vegetation in the Middle West.

Preglacial 57

Early Glacial Stages 59

The Mississippi Embayment 59

Northern Limits of Southern Plants 60

The Migration of the Conifers 61

The Migration of Deciduous Forests 61

Wisconsin Glacial Period 62

Distribution of Plant Life During the Wisconsin 63

Climate in the Early Post-Wisconsin 66

Tundra and Coniferous Forest Migration 67

Early Migration of the Prairie Flora 70

Early Migrations of the Deciduous Forests 73

The Southern Migration 74

The Northern Migration 74

The Combined Effects 75

Migration and Evolution 75

Prairie and Boreal Relics 76

Development of the Prairie-Grass Formation 77

Extent of Forest Migration in the Prehistoric Period 78

Second Period of Prairie Dominance 80

Second Period of Forest Advance 82

Coastal Flora of the Great Lakes 83

Summary of Part IV and Conclusion 84

' 39

NEW Y

BO! AN1<
40 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

Part I

The Dynamic Viewpoint. — Many years have elapsed since Engler
published his famous and fundamental "Leitende Ideen," which he
had adopted as a basis for his discussion1 [of the Tertiary and tPledsr
tocene development of the vegetation of the earth. Here, in thirty-six
concise paragraphs, he summarized the general principles controlling
the distribution of species and of species-groups. The first two of them
serve as a summary for all :

"1. Die gegenwartige Verbreitung der Pflanzen ist nicht bios bedingt
durch die jetzt auf der Erde herrschenden klimatischen Bedingungen
und die Bodenverhaltnisse.

" 2. Ein wahres Verstandniss der Verbreitung der Pflanzen ist nur
dann moglich, wenn man die allmalige Entwicklung derselben zu
ermitteln sucht."

Engler was apparently inspired to his studies of vegetational history
and development, at least in part, by the classic essays of Asa Gray.2'3'*
His lead was followed by numerous European botanists, who, assisted
by the co-operation of palaeontologists and glacial geologists, have in
the past two decades contributed greatly to the knowledge of the phyto-
geographical history of central and western Europe. Later American
botanists, on the other hand, neglected such lines of investigation almost
completely and their phytogeographical activities were confined, as is
natural in relatively new regions, to descriptions of floras and the de-
termination of modern distribution. It remained for Adams5'6 to
revive developmental biogeography in America and to show that the
distribution of a species can be satisfactorily explained only through
the location of its origin and the nature of its subsequent migrations.

General Control of Distribution. — The distribution of plants,
as shown by Engler's principles quoted above, depends, in general
terms, on modern environment and earlier developmental history. Both

i Engler, Adolph. Versuch einer Entwieklungsgeschichte der Pflanzemvelt, insbe-
sondere der Florengebiete, seit der Tertiarperiode. Leipzig, 1879.

2 Gray, Asa. The flora of Japan. Scientific Papers 2: 125-141. (1859) 1889.

3 Gray, Asa. Sequoia and its history. Scientific Papers 2: 142-173. (1872)
1889.

4 Gray, Asa. Forest geography and archaeology. Scientific Papers 2: 204-233.
(1878) 1889.

e Adams, C. C. Southeastern United States as a center of geographical distribu-
tion of flora and fauna. Biol. Bull. 3: 115-131. 1902.

e Adams, C. C. Postglacial origin and migrations of the life of the northeastern
United States. Journ, of Geog. 1: 303-310, 352-357. 1902.

THE GENERAL NATURE OF PLANT MIGRATIONS 41

of these are intimately concerned with migration, the latter factor por-
traying its progress and the former its limitation. Neither factor alone
can account completely and satisfactorily for the present range of any
species. This is the resultant of its range at any period in the past
and its opportunities for subsequent migration; one exposure in the
long film of an activity which has been continued since the origin of
the species. If logs in the glacial drift show that Pinus Strobus
formerly existed in central Illinois and experiment proves that it can
thrive there now, historical evidence must be used to show why it left :
modern conditions may show whether and why it is not returning.

Vegetational History and Human History. — The history of the
vegetation of the Middle West, as of every other portion of our con-
tinent, is a history of repeated migrations of diverse floristic elements,
arriving in the region from various directions, persisting there for
various lengths of time, and finally retreating under the pressure of
environmental changes which made their position no longer tenable.
This history has been complicated by the extinction of old and the
differentiation of new species, by the mingling of species of originally
different history, and by the persistence of isolated relics of early
migrations. In all these respects, vegetational history exhibits a strik-
ing parallelism to human history. Thus Europe has witnessed during
the last two millenniums the extinction or absorption of the early
Iberians, the persistence of the Basques, the arrival and retreat of the
Moors and Turks, and an enormous evolution in culture. Furthermore,
the underlying causes of vegetational and human history are similar,
even though the human species exhibits an exceedingly complex rela-
tion to its environment. The geography of man has long been recog-
nized as merely the present condition of his history, and the relations
of both to natural conditions and processes have been well summarized.7
Unfortunately, no such summary exists for plants.

Outline. — It has therefore been necessary to preface this history of
plant life in the Middle West, presented in Part IV., by a discussion
of the conditions and nature of plant migrations in Part II., and by a
resume of the evidence used in determining past migrations in Part III.

Part II. The General Nature of Plant Migrations
A great body of literature dealing with the migration of plants has
appeared in the last century and a half. This has been so well sum-
marized bv Clements8 that it needs no further discussion here. Most

i Semple, Ellen C. Influences of geographic environment. New York, 1911.
s Clements, Frederic E. The development and structure of vegetation. Studies
in the vegetation of the state III. Botanical Survey of Nebraska VII. Lincoln, 1904.

42 II. A. GLEASON VEGETATIONAL HISTOEY OF MIDDLE WEST

of it deals with the migration of individual species, and Clements'
exposition of the whole subject is without doubt the best general treat-
ment extant.

Invasion, Migration, Ecesis. — Following Goeze, he considers the
general movement of species or associations under the term invasion,
and recognizes that this consists of two distinct processes. One of
these, to which he restricts the term migration, involves merely the
movement of the plant by means of its reproductive bodies or vegeta-
tive processes. The mechanics of such movement, the methods of seed
dispersal or vegetative propagation, while exceedingly diverse, are
already well understood and require no discussion here. The second
process, which he terms ecesis, includes the actual establishment of the
plant in a new location. Ecesis depends entirely upon the environ-
mental conditions of the new location in their relation to the physio-
logical requirements of the seedlings. The details of these relations
are for most plants entirely and for all plants mostly unknown, and
are 1o be determined by the methods of experimental ecology. Migra-
tion, in Clements' use of the term, is completely unsuccessful in the
failure of ecesis, and ecesis, to be effective and complete the proce=s
of invasion, must be carried to the maturity of the plant and the pro-
duction of a second generation of disseminules. The distinction be-
tween the two processes is fundamental, and Clements' clear analysis
has done much to systematize the knowledge of the subject.

The term invasion, which Clements applies to the whole phenomenon
of movement, is in many cases scarcely appropriate. His own defini-
tion and the general meaning of the word both imply a movement of
plants into an area not before occupied by them, that is, an extension
of their range. As a matter of fact, the migration of plants is almost
or quite as frequently out of an area previously occupied as into a
new one, and then involves a retreat or restriction of their range, a
result to which the term invasion can not be properly applied. The
term migration, in the sense of Clements, is distinctly a process of one
generation. Ecesis involves the development of a second generation
of individuals, but scarcely applies to the whole population of a species.
Yet these processes, participated in by a whole population and con-
tinued regularly through successive generations, may lead to broad
results far more inclusive than those chiefly considered by Clements.
For the purposes of this article, the migration of plants is understood
as any general movement by which the range of a species is changed.

Efficiency of Plant Dispersal. — The efficiency of plant dispersal
is well understood. Each normal individual produces one or many

MIGRATION AND SUCCESSION 43

crops of seeds or other reproductive bodies and these are dispersed
thoroughly over the surrounding territory to an extent dependent upon
their structure and their environment. At the margin of the present
range of a species, whether that margin is the boundary of an associa-
tion in which it occurs or the general margin of its whole range, seeds
are annually dispersed in all directions, inward through its present
range and outward and beyond it as well. Examples of this are well
known and need not be cited. Except in comparatively infrequent
cases, where the ability to produce seeds or other reproductive bodies
is lost, the potentiality of continuous and extensive migration is con-
stantly present. The slow rate of actual normal migration is not due
to inefficiency of the migratory devices of the plant, nor, except as
mentioned above, to deficient seed production, but to the inhibiting
effect of environmental factors in the area beyond the present distri-
bution of the species. This is particularly true toward the center of
the range. Every swamp is annually and abundantly planted with
seeds of upland species, while the swamp plants are simultaneously
distributed over the upland, and in neither case do any of them grow,
except under the most unusual circumstances.9 Even at the margin
of the specific range, where reduced or intermittent seed production
might be expected, Griggs has shown that many species are able to
propagate themselves freely.10

There is for each species a general limit of distance over which its
seeds may be dispersed. The number of seeds arriving on any area
diminishes with increasing distance from the parent plant ; in fact, for
small distances, the number seems to vary inversely as the square of
the distance. Since only a small portion of the seeds germinate and
grow to maturity, most young plants of the second generation are located
relatively close to the parent.

Migration and Succession. — Actual migration, therefore, requires
a change in the existing environment for its inception. (The abnormal
type of migration, due to the acquisition of new means of dispersal
by which previous environmental barriers are crossed, is mentioned in
a later paragraph.) A migratory advance postulates the development
of individuals in territory not previously occupied and requires an
environmental change of sufficient extent not merely to permit the
growth of the migrants, but to permit it against the competition of
the previous inhabitants. A migratory retreat requires a detrimental

» Brewster, William. Occurrence of the skunk cabbage in an unusual place.
Khoflora 11: 63, 64. 1909.

io Griggs, Robert F. Observations on the behavior of some species at the edges
of their ranges. Bull. Torrey Club 41 : 25-49. 1914.

44 H. A. GLEASON VEGETATIONAE HISTORY OF MIDDLE WEST

change, either in physical conditions or in their control by plant life
or in parasites and enemies, sufficient to prevent the establishment of
new plants in the immediate vicinity of the parent, so that the species
disappears from the habitat with the death of existing individuals.
Succession results when migration is so complete and is shared by indi-
viduals of so many species that the nature of the vegetation is funda-
mentally changed. The close relation between succession and migration
have been noted elsewhere,11 and the nature of the causal environmental
changes have been discussed in detail by Clements.12 Instances are
of course well known in which species migrate into an association and
become an integral part of it without causing succession. Thus in
parts of New Jersey Daucus Carota has entered the hemlock-maple
forest, and in northern Michigan Rumex Acetosella has become widely
distributed through the aspen association. Such a result may be
termed taxonomic succession, as distinguished from the ecologic suc-
cession previously mentioned.

Kate of Migration. — The simple production, movement, and estab-
lishment of reproductive bodies, constituting together the method of
migration, is not the only factor concerned in the process. The rate
of migration depends upon one or the other of two conditions: the
rate of environmental change and the migration-capacity of the plant.
The latter condition may in turn be resolved into two components ; the
mobility of the plant, depending on the structure of its migrating parts
and its ability to utilize the necessary environmental agencies, and the
length of its juvenile period, which must be completed before a new
crop of disseminules is produced to continue the march. Migration-
capacity may be formulated in general terms as mobility divided by
length of juvenile period. Hence those plants which combine excellent
structural devices for dispersal with a short vegetative cycle, such as
the dandelion (Tara&acum), migrate more rapidly than others with
less mobility, as lamb's-quarters (Chenopodium) , or with a long vege-
tative period, as the maple (Acer), while slowest of all are those plants
which lack both favoring features, as the oak (Quercus). The actual
rate of migration is fixed by the slowest of the two general conditions.
If environmental change is slow, migrating species may keep even
with it and continually scatter their seeds forward into territory where
they can not yet grow. If environmental change is rapid, the slower
species may not be able to keep even with it and consequently follow

11 Gleason, H. A. The structure and development of the plant association. Bull.
Torrey Club 44: 463-481. 1917.

12 Clements, Frederic E. Plant succession, an analysis of the development of
vegetation. Carnegie Inst. Washington, Publ. 242. 1916.

THE TIME FACTOR AND MIGRATION 45

at a distance far behind other more mobile forms. Thus in the slow
extinction of a pond by the deposition of peat, there is no break in
the time-continuity of the contracting zones of vegetation around it,
since mobility exceeds the rate of environmental change. But if a
pond is drained artificially, producing in a single year an effect similar
to that of a century of peat formation, the more mobile species enter
first on the freshly exposed area and some years may elapse before all
the remainder have appeared. So also the sudden cessation of prairie
fires on parts of the prairies led to the immediate advance of the
forest, and the more mobile species composed the bulk of the early
invaders. In general, migration connected with physiographic suc-
cession keeps pace with the environmental changes, so that associations
are well marked and the correlation between zonation and succession
is prominent. Even such a slow environmental change as the advance
or retreat of a continental glacier may have been too rapid for some
species, while certainly slow enough for others. Numerous species
were doubtless destroyed completely with the advance of the glaciers
and others during their retreat, from inability to keep pace in their
migration with the changing conditions of the environment.

The most rapid rate of migration is found in the anthropochorous
species which utilize human activities for their dispersal. Such not
only move with extreme rapidity, but by the nature of their dispersal
are able to cross barriers quite impassable for other species. Our own
country has been occupied during the past century by hundreds of
such species which have crossed an ocean to reach us and have spread
hundreds of miles over the land in a very short time. Yet most of
these same species are stopped completely by an area of undisturbed
natural vegetation, which they are unable to colonize.13

The Time Factor and Migration. — The present status of migra-
tion depends upon its rate, as discussed in the preceding paragraph,
and on the elapsed time since migration began. The more rapidly
moving species may be at the limit of their present potential range and
wait upon further environmental change, while the slower ones may be
far from their actual limit, although progressing toward it at their
best speed. The condition is well illustrated in the belts of forest
which parallel the rivers of Kansas and Nebraska, in which, in general,
the more mobile species have extended farthest to the west,14 It is
also prominent throughout the prairie region of the Middle West, where

is Gleason, H. A. & F. T. McFarland. The introduced vegetation of the vicinity
of Douglas Lake, Michigan. Bull. Torrey Club 41: 511-521. 1914.

i* Kellogg, R. S. Forest belts of western Kansas and Nebraska. U. S. Dep.
Agr. Forest Bull. No. 66. 1905.

46 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

there is a continuous movement of the more mobile forest species into
the land originally occupied by prairies.

The time factor naturally begins with the origin of the species con-
cerned. Those of relatively recent origin have had little time in
which either to extend or to contract their original ranges. Since
major environmental changes are usually slow in their accomplish-
ment, recent species have in general had also relatively little oppor-
tunity to use their migratory abilities. Such species are therefore
restricted more closely to the site of their origin.

It is probably true, however, that many of our species have not
had a single point or even a limited area of origin. The bulk of later
Tertiary plants, as far as palaeontological evidence indicates, are of
genera still existing, and many of the comparatively few known species
of Pleistocene plants are either identical with or closely similar to
existing species. Such evidence deals especially with woody plants,
but what is true of them is doubtless true of herbaceous plants as well.
The distinction between such Pleistocene plants and their modern
representatives may depend largely on a break in the record, on a
period from which fossils are lacking. Probably if a complete series
of specimens were at hand, showing comprehensively the maples of the
eastern states, for example, from the Pliocene to the present time, it
would be seen that some of the earlier forms are absolutely continuous
with our present species and that the slight morphological distinctions
between them are only the result of continuous slow variation through-
out the centuries. According to this view, many modern species had
no localized origin and are not the offshoot of any parent, but represent
the mass development of a species, which, under our present taxonornic
ideas, came to a stop at the beginning of ,a break in our geological
record of it and reappeared as a new species at the beginning of our
next experience with it. During the long history of such a species,
most of which is unknown to us, it may have migrated repeatedly in
various directions, occupied larger or smaller areas, been separated
into disjunct regions and again united, or one part of its population
exterminated, and its range as we see it today represents merely the
present condition of this long development. We are never sure of
the prehistoric stages in this history, through the imperfections in the
fossil record, but in certain cases we can arrive at some idea about
them by circumstantial evidence.

Such a consideration does not account for the multiplication of
species. Whether that is caused by mutation, natural selection, or
hybridization, it seems certain that their perpetuation to the present
time has depended greatly upon their ability to migrate as changes
of environment demanded.

ADVANCING AND RETREATING MIGRATIONS 47

Significance of the Margin of Range. — The margin of the range
of a species, then, does not always represent the boundary of the ter-
ritory in which the species can or will live under present conditions,
but merely the distance which it has travelled so far in its march to
this goal. Instances are not lacking, of course, in which this migration
has been completed and further extensions or retractions of the range
wait upon future climatic changes. As one illustration, the distribu-
tion of Pinus ponderosa along the eastern foothills of the Rocky
Mountains may be cited, where it has been clearly shown15 that the
species is now at the margin of its potential range. On a smaller
scale, a sharply marked [boundary between two associations indicates
completed migration in that restricted area of the species on both sides
of the boundary to the limits permitted by the present environment.

Advancing and Retreating Migrations in the Middle West. —
Both advancing and retreating migrations are now in progress in the
states of the Middle West. In most cases these movements are too
slow to come under direct observation, and reported extensions of
ranges are generally due to incomplete or faulty early observations.
Still the rapid spread northward of Prosopis glandulosa from Texas
across Oklahoma into Kansas,16 the westward migration of trees along
the rivers of Nebraska,14'17 and the recent discovery of young trees
of Quercus palustris in southeastern Nebraska18 may be cited as in-
stances of visible migration. In general, migration is shown chiefly
by the role of the species in succession. If, at or near the edge of its
range, it participates in succession as an invading species, it is advanc-
ing; if it io not able to hold its own in competition, if the associations
of which it is a member are regularly succeeded by others of which
it is not a part, a retreating migration is indicated. This is illustrated
in Michigan and Wisconsin by the regular succession of coniferous asso-
ciations by deciduous forests, indicating the retreat of the conifers and
the advance of the deciduous trees.19

Effects of Changes in Physiological Requirements. — There is
no reason why changes in the physiological requirements of a species

ij Dodds, Gideon S., Francis Ramaley, & W. W. Robbins. Studies in mesa and
foothill vegetation, I. Univ. Colorado Studies 6: no. 1, 11^9. 1908.

is Bray, William L. Distribution and adaptation of the vegetation of Texas. Bull.
Univ. Texas 82. 1906.

it Bessey, C. E. Are the trees advancing or retreating upon the Nebraska plains?
Science 1L, 10: 768-770. 1899.

is Pool, Eaymond J. Pin oak in Nebraska. Torreya 20 : 50-52, 1920.

19 Whitford, H. N. The genetic development of the forests of northern Michigan:
a study in physiographic ecology. Bot. Gaz. 31: 289-325. 1901.

48 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

might not have the same effect as a change of environment in causing
migration. It is well known that plant breeders have developed new
races of cultivated plants which thrive in climates different from that
of the original home of the species. Possibly the rapid spread of
American Opuntias in parts of Australia may be explained in this
way. Under natural conditions, such changes in physiological require-
ments are usually correlated with changes in structure, and the latter
are used as a basis for taxonomic differentiation. Whenever such
evolutionary changes appear, new areas may be opened to the new
species and eventually occupied by them. So such genera as Eupa-
iorium and Vernonia, originating probably in northern South America,
have migrated northward into the cold-temperate zone, as evolution has
changed their physiological requirements and morphological structure.
So have bromeliads migrated away from the humid tropical forests,
entered the deserts of southern Mexico, and established themselves as
epiphytes on the cacti ;20 so have the cacti migrated out of the desert
region, where they probably originated, and established themselves
as epiphytes in jthe tropical forest.

Migrations as Related to Continuity of Suitable Habitats. —
The range of every species is discontinuous. Throughout its whole area
only particular habitats are occupied, leaving others to different species
and associations. In each habitat, the space is shared with other
associated species. These phenomena are 'in general of geological
rather than geographical interest, but are still of importance in their
relation to the direction of migration. Within the habitat, so far as
the association is continuous, even the least mobile species may migrate
freely and ultimately attain a uniform distribution.11 But from one
habitat to another, the means of dispersal must be sufficient to carry
the species across the intervening gap, which becomes a barrier to the
less mobile forms. During the long course of time, accidents of dis-
persal may carry many species across a barrier, but accidents are
probably not as efficient as frequently supposed. Mink Grove and
Lynn Grove, two isolated areas of forest in Champaign County, Illinois,
had large trees in them when first observed by white men nearly a
century ago, and their age is doubtless 150 years at a minimum. Al-
though these groves lie but a few miles from the nearest strips of forest
along streams, in which species of Hicoria and Quercas are abundant,
not an individual of these genera has reached them. Extensive migra-
tions of plants, proceeding at a relatively rapid rate, therefore take

20MacDougal, D. T. Botanical features of North American deserts. Carnegie
Inst. Washington, Publ. 99. 1908.

THE INTREPRETATION OF ISOLATION 49

place along routes where suitable habitats are nearly continuous, such
as river valleys for mesophytic species or interfluvial uplands for
xerophytic forms. There is every reason to believe that the post-
glacial migration of the deciduous forests into the Middle West fol-
lowed the river courses almost exclusively. From these earlier migra-
tions along the line of least resistance, a slower type of migration
gradually extends the range out upon other habitats. This follows
regularly the physiographic development of an area, as Cowles has
shown,21 and may also be induced by the reaction of an association
upon the environment in the margin of a contiguous association beyond
it.11 So the forests of the Middle West, entering the region along the
nearly continuous habitats of the river courses, where they migrate
rapidly, have also spread at right angles to them, but more slowly, until
in some cases those of different river systems have become united.

Each important change in the climate or the configuration of the
land, due to geological causes, has created new migration routes and
destroyed old ones, interrupting migrations then in process and initiat-
ing new ones, which may progress in diiferent directions and be par-
ticipated in by plants of different origin. There is evidence of several
such migrations in the Middle West, which are discussed in Part IV.

The Interpretation of Isolation. — Actual geographical isolation
of a species, in the ordinary sense, is effective only when an outlying
habitat is separated from the general range by a distance greater than
the migration-capacity of the plant. Instances of this are too well
known to need citation. Only two explanations are possible for such a
condition. Either the species has had two separate and independent
origins, or else the vicissitudes of migration have brought it about.
Even if the possibility of the former alternative is admitted, the latter
is the more plausible in the majority of cases. It implies that the two
disjunct bodies of the species were at some time one, either in a ter-
ritory from which both portions have migrated and which is no longer
occupied, or through the existence of connecting links between the two
which have now disappeared. Both cases have been caused by a, retreat-
ing migration.

Environmental changes which permit the advance or compel the
retreat of one species are quite likely to affect others similarly, since
associated species demand in general the same environment. Migra-
tion is therefore concerned with numerous species simultaneously.

21 Cowles, H. C. The physiographic ecology of Chicago and vicinity; a study of
the origin, development, and classification of plant societies. Bot. Caz. 31 : 73-108,
145-182. 1901.

50 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

Whole floras migrate together, the hardiest and most mobile species
first, the others in their train. They establish successional series in
the region which they enter, and build up new associations in the new
territory. These are analogous to the associations of their original
range, but differ in the absence of the slower moving species and in the
presence of laggards from the retreating flora.

Just as portions of a single species may be isolated during its migra-
tions, so may areas of one type of vegetation be completely surrounded
by an advancing flora and left isolated from the main body. These
are known as relic colonies. The cause of such isolation is the failure,
up to the present time, of sufficient environmental change to cause their
extinction, or, in ecological terms, it is the relative slowness of the suc-
cession in certain habitats unfavorable to most of the advancing species.
These habitats remain with their original vegetation upon them, while
elsewhere it is displaced. Such relic colonies are of common occurrence
wherever migrations are recent. Every movement of vegetation has cer-
tainly left relic colonies behind it, but their perpetuation to the present
time depends absolutely on the continuation of an environment not
only favorable to the relic species but distinctly unfavorable to the
invaders. !

Adjustments in Distribution. — The distribution of a species and
the distribution of vegetation must both be regarded as subject to con-
tinual adjustment. Boundaries extend and contract, associations
advance and retreat with every geographical fluctuation of the environ-
ment. Even within the limit of a single association, every minor en-
vironmental variation, whether geographical or seasonal, is marked
by the entrance or disappearance of some species, or by changes in the
numerical proportion of the others. This may account for the condi-
tions recently reported on Mt. Marcy,22 where twenty-one species
formerly reported by Peck have apparently disappeared and seven.
" rather conspicuous plants, which such a careful botanist could scarcely
have missed, have apparently come into the flora."

General Results of Continued Migrations. — Obviously the
general result of continued migrations is the mingling of floras of
diverse origin, while great climatic changes or important geological
episodes tend toward the segregation of floras and the consequent
initiation of new migrations. Regions affected by Pleistocene glacia-

22 Adams, C. C, Geo. P. Burns, T. L. Hankinson, Barrington Moore, & Norman
Taylor. Plants and animals of Mount Marcy, New York. Ecology 1 : 71-94,
204-233, 274-288. 1920.

THE EVIDENCE OF PLANT MIGEATIONS 51

tion, such as the area here discussed, fall in this group, and in them
floras of different origin are still relatively distinct and adjustments
of range are actively in progress. On the other hand, in regions which
have enjoyed relatively uniform environmental conditions for a long
period of time, migrations are more generally completed and vegeta-
tion is in more nearly a static condition, except in those successional
series which are related to' continuous physiographic processes such as
base-leveling. This is now the case in parts of the tropics, and was
formerly the case over wide areas of the North Temperate zone in the
later Cretaceous and earlier Tertiary periods, when the flora and vegeta-
tion of North America and Eurasia were remarkably uniform.

Summary of Paet II. — 1. Migration of a species depends on an
environmental change within or beyond its range,

2. The rate of migration depends on the rate of environmental
change or the migration-capacity of the plant.

3. The present status of migration, as shown by the range of the
species, depends on its rate and on the time available for it.

4. The present range of a species is not necessarily final, even with-
out further environmental change.

5. Migration proceeds most rapidly along routes with nearly con-
tinuous habitats.

6. Species of similar environmental demands migrate together.

7. Isolated areas of a species or of vegetation are to be interpreted
as results of retreating migration.

8. Long continued uniformity of environment leads to floristic uni-
formity; recent environmental changes to floristic segregation and
to new migrations.

Part III. The Evidence of Plant Migrations

The entire distance which may have been travelled by the native
plants of the Middle West in their migrations since the close of the
last glaciation seldom exceeds 500 miles. Since the time available
for this movement may be 10,000 to 40,0G$:*years, this requires an
average annual migration of only 66 to 264 feet. In the case of trees,
a juvenile period of 20 years would still necessitate a movement during
one generation of not more than a mile. The discovery of migrations
still in progress is accordingly removed from direct observation, except
in the case of a few species, and any conclusions must depend entirely
upon historical or indirect evidence.

52 H. A. GLEASON VEGETATIONAL HISTOKY OF MIDDLE WEST

Historical Evidence of Migrations. — Historical evidence may
occasionally be used with considerable success, especially where civiliza-
tion has existed for a long period of time and written records preserve
trustworthy accounts of some of the original features of the vegeta-
tion. Thus Graebner23 has been able to demonstrate that certain
heaths of northern Germany were originally beech forests. Tansley21
has also made frequent use of historical records. For the Middle West,
such data cover little more than a single century and are correspond-
ingly restricted in value. There are, however, numerous early books
of travel which give fairly accurate accounts of the plant life. Their
descriptions of prairie fires, of the former extent of prairies, and of
the oak openings are in many cases excellent and may probably be
relied upon completely. They generally lack detail and seldom mention
many species by name, but are nevertheless better for phytogeographical
purposes than the floras compiled by early botanists. In almost every
county seat copies of the original land survey maps may be consulted.
On many of them various features of the vegetation are shown, such
as the occurrence of swamps, prairies, and barrens. In some cases the
nature of the forest may be approximated by reference to the accom-
panying field books, since the surveyors' monuments were usually
located with relation to certain described trees.

Succession as Evidence of Migration. — For migrations still in
progress, the most valuable line of contemporary evidence is found in
the successional relations of the species toward the margin of their
ranges. Even there, succession can not generally be observed directly,
but numerous other considerations offer convincing evidence of the na-
ture of the process, such as the relation of the associations to physioV
graphic development, the reaction of the plant cover on the environment,
the presence and relative abundance of seedlings, or the relative age of
individuals. These have been used so freely by ecologists and such
a body of accepted data has been published that they need no dis-
cussion here. Nen iheless, a timely warning against the too general
application of the, Wiciples of succession has been sounded by Har-
vey25 and deserves careful consideration. When, at or near the margin
of a range, a spec1'"' or tthe members of a flora of uniform geographical
distribution pari* * ates in succession as invading species, they are

23 Graebner, P. Die Heide Norddeutschlands und die sich anschliessenden
Formationen in biogischer Betrachtung. Die Vegetation der Erde, vol. 5, Leipzig,
1901.

24Tansley, A. G. Types of British vegetation. Cambridge, 1911.

25 Harvey, Leroy H. Some phytogeographical observations in Lake County,
Michigan. Mich. Acad. Sci. Rep. 21: 213-217. 1920.

DETERMINING MIGRATION CENTERS 53

extending their range. Similar behavior toward the center of the range
may indicate only a general closing-up process, representing the later
slower stages of a migration and not implying any general extension
of range in the ordinary sense of the word.

The known successional tendencies of the three major types of vegeta-
tion of the Middle West have frequently been described, especially
the tendency of the deciduous forests to succeed the prairies toward
the west and the coniferous forests toward the north. Project this
tendency into the future and it points to a still wider extension of this
type of forest, with a corresponding restriction of prairie and coni-
ferous forest. Project it into the past and it reveals a former condi-
tion of restricted deciduous forest and of a larger extent of prairie
and coniferous forest. Obviously, such reasoning must not be carried
too far, and its results must always be checked by the existence of relic
species or relic colonies, as discussed in a following paragraph.

The stratigraphic sequence of fossils has been used with great suc-
cess as one line of evidence in Europe.26 While similar evidence has
already yielded important results in America, the necessary data for
extensive generalizations are usually lacking or deficient. The careful
examination of the structure of peat has scarcely more than begun in
this country, and the pioneer work of Dachnowski27 may lead to impor-
tant results.

Geological events must have necessitated certain migrations and by
their very nature give a general idea of them. Thus it is incredible
that beech-maple forests could have occupied northern Michigan during
the maximum advance of the Wisconsin glaciers, which reached some
400 miles farther south. A northward movement of these trees and
their associated species to their present location must have taken place
after the retreat of the ice, but the details of this migration must be
discovered by other lines of evidence.

Determining Migration Centers. — Adams5 has advanced certain
criteria for determining the migration-center of a ' >als, which are in
many cases equally applicable to plants. They ar^ th two omissions,
as follows : • t

1. Location of greatest differentiation of a type.*? ,

2. Location of dominance or great abundance or Individuals.

3. Location of synthetic or closely related forms.

2« Lewis, F. J. The plant remains in the Scottish peat mosses. Part I. The
Scottish southern uplands. Trans. Roy. Soc. Edinburgh 41: part 3, 699-723. 1905.

27 Dachnowski, Alfred P. Peat deposits and their evidence of climatic changes.
Bot. Gaz. 72: 57-89. 1921.

54 H. A. GLEASON VEGETATIONAL HISTOEY OF MIDDLE WEST

4. Location of maximum size of individuals.

5. Continuity and convergence of lines of dispersal.

6. Location of least dependence on a restricted habitat.

7. Continuity and direction of individual variations or modifica-
tions radiating from the center of origin along the highways of dispersal.

8. Direction indicated by biogeographical affinities.

The Five Elements in the Floea of the Middle West. — Ap-
plying his principles to the vegetation of the Middle West, it is found
to be composed of five elements, not of equal importance, centering
respectively in the southern Appalachian Mountains, in the southern
Coastal Plain and the Mississippi Embayment, in the Ozark Moun-
tains, in the plains of Kansas and Nebraska, and in Canada east of the
Great Lakes. Four of these lie beyond the glaciated region far enough
to have suffered little or not at all from the advance of the ice, and
there is no reason to believe that they did not occupy these positions
continuously through the later glacial period. The last one is in the
glaciated region, and must have been developed since the retreat of the
ice from a temporary refuge farther south. For the first four, we are
probably safe in assuming that all subsequent migrations of their floral
elements have been on lines essentially radial, either from or toward
these centers, as ranges may have extended or retreated. For the fifth,
we are undoubtedly correct in assuming a general northward shift of
the whole flora following the ice retreat.

Glacial vs. Modeen Climate. — The close similarity between the
interglacial and the modern floras, so far as evidence is at hand, indi-
cates that the glacial climate differed from the modern not so much
in kind as in degree. It may be expected accordingly that migrations
of the deciduous forests took place first in a northward direction from
their Appalachian center and later westward into the Middle West,
and that any eastward migration of a prairie flora took place toward
the north of our area, where there is still a marked distinction between
winter and summer rainfall and a generally smaller total amount,
rather than at the south where rainfall is more abundant and more
evenly distributed throughout the year.

Further evidence may be drawn from the topography of the region.
Large and sharply marked moraines may indicate a rapid advance of
the glacial ice. Wide stream valleys leading from the moraines, with
great outwash deposits, may indicate rapid melting of the ice under
relatively high temperatures, and consequently in the adjacent region
a flora and vegetation correlated with such climate. Such evidence has
been used chiefly in connection with the Wisconsin glaciation, where

RELIC COLONIES

55

it seems that the temperature at the time of maximum ice advance
was high and probably not unlike that of the present time. The imma-
ture development of the drainage systems over the Wisconsin drift,
particularly toward the west, may also be due not merely to youth
but also to an early postglacial climate drier than that of the pres-
ent day.

If a considerable number of species are now confined to an area
immediately beyond a moraine and seldom or never occur within it,
their distribution may be explained in several ways. First and most
obviously, present conditions of soil and climate may exclude them.
But if the general vegetation and climate of the two regions are similar
and if habitats apparently identical, as judged by the vegetation, exist
on both sides of the moraine, this explanation is less plausible. Sec-
ondly, their migration up to the present time may have extended
merely to the moraine, which they will later cross. But it is difficult
to understand why the boundary of the range of numerous species
should be so nicely adjusted to the moraine. The third explanation,
which is here accepted, is that they occupied this area before or during
the formation of the moraine, have occupied it ever since, and are not
at present extending their range.

Relic Colonies. — Lastly, and most important of all, past migra-
tions may be judged by relic colonies. Unfortunately, they can be used
only for the more recent movements, since the relics of earlier migra-
tions have been completely destroyed. Many such colonies still exist,
isolated by recent migrations, or did exist long enough to be recorded
before being sacrificed to agriculture, and afford the most valuable
evidence of the past range of the floras which they represent. They
even give valuable clues to migrations of still earlier date, during
which the flora of the colonies entered the region, and therefore show
at once an early advance followed by a later retreat.

Although the probability of extreme migrations, except in the case
of very mobile species, is very slight, the time available is very long
and has certainly been sufficient for the improbable to happen many
times. Therefore, in general, isolated stations of a single mobile species
shed little light on its past history. But isolated colonies of several
or many species, provided with diverse methods of dispersal, represent
a different condition. It is entirely beyond the limits of probability
or coincidence to presume that a dozen or more species, normally grow-
ing together in one section of the country, should also be found together
in another remote section beyond the normal limits of dispersal of any
of them. Isolated colonies of this character are common phenomena
in the Middle West and can have but one significance, that at some time
in the past these colonies were more numerous and separated by narrow

56 II. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

intervals within, the normal migration-capacity of the species, or were
entirely continuous over the whole area. Stated in more general terms,
the present occurrence of relic colonies indicates the past extent of
the vegetation of which they are a type. For the same reason, isolated
stations of a single relatively immobile species indicate a former more
nearly continuous distribution, and evidence from this source is of
greater importance when the species in question is regularly associated
with some peculiar or unusual habitat.

A combination of evidence from successional tendencies and relic
colonies indicates both the direction and the extent of prehistoric migra-
tions. The full extent may not be indicated, because the isolated
colonies may have disappeared completely from the more remote parts
of their original range. In such cases isolated stations of single species
may give some not entirely untrustworthy idea of it. The farther
back in time plant migrations are traced, the fewer relic colonies may
be expected and the more dependence must be placed on individual
species.

No one of these various lines of evidence is sufficient in itself to
build up a history of vegetational movements in the Middle West.
All of them must be used together and the deductions from all woven
into a history which accounts for present conditions satisfactorily and
is plausible in itself. The development of this history must be com-
menced with the present, where the latest migrations are shown by
successional relations. The extent of these recent events may be esti-
mated and checked by relic colonies, which in turn point to earlier
migrations. These must be checked by glacial topography, by infer-
ences concerning glacial climates, and by the distribution of plants
along the moraines. Still farther back, plant distribution and migra-
tions must be deduced from fossils, from the location of the moraines,
from the known physiological requirements of the plants, and through
the use of Adams' criteria.

The vegetational history which follows in Part IV has been con-
structed in this way. Observed conditions and recorded vegetational
changes during the historical period have given evidence on migrations
then in progress and permitted deductions as to those of the later
prehistoric period. Conclusions as to the preceding stages have been
reached through use of the various lines of evidence outlined above
and carried back with decreasing detail and assurance to the advance
of the Illinoian glaciers.

Summary of Part III. — 1. Migrations in progress now or in the
recent past are indicated by historical evidence or by observation of
successions.

PREGLACIAL HISTOEY

57

2. Former migrations are indicated by the occurrence and distribu-
tion of relic colonies and species, by ecological and taxonomic evidence
as stated by Adams, by glacial history and topography, by fossils, and
by inferences as to former climate.

3. Evidence from all available sources must be combined to build a
plausible and possible chain of events, leading on to and culminating
in the known present distribution of plants.

4. Conclusions as to early stages of plant distribution and migra-
tion are reached in reverse order of their occurrence, that is, from
present to past. The accuracy of each conclusion depends on the
accuracy with which the conditions of the following stage have been
interpreted. Each earlier stage must therefore be discussed in less
detail and with greater probability of error.

Part IV. The Development and History of Vegetation in the

Middle West

Preglacial. — From the close of the Carboniferous Period to the
Lower Cretaceous or Comanchean Period, the area now comprised
between the Missouri Valley and the Appalachian Mountain system
was continuously land. The pteridophytic and gymnospermous flora
which occupied it can be partly reconstructed from the fossils of those
ages, but its migrations are unknown. During the Lower Cretaceous
the first representatives of the angiosperms must have arrived in our
area, since the fossil remains of such plants are preserved in con-
temporary deposits still farther west, and through the Upper Cretaceous
angiosperms constituted the dominant vegetation. It is not probable
that herbs were prominently developed; angiosperms consisted chiefly
of woody plants of families and genera still in existence in the area
and of other groups now primarily or exclusively tropical. This flora
is undoubtedly continuous with the forest flora of the Tertiary, usually
designated the arctotertiary flora, and through it with the forests of
the present.

While there may have been differentiation of floristic types in the
Cretaceous, correlated with local climatic conditions, the first of the
great vegetational segregations which is still of importance in our
region began at the close of the Cretaceous with the uplift of the
Cordilleran complex of mountains.28 Intercepting the moisture-laden
winds from the Pacific and restricting the rainfall of the lands imme-
diately east of them to moisture derived from the Gulf of Mexico,
the elevation of these mountains led to the development of semiarid
conditions over the Great Plains, which soon had an effect on the

28 Harvey, Leroy H., Floral succession in the prairie-grass formation in south-
eastern South Dakota. Bot. Gaz. 46: 81-108. 1908.

58 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

character of the vegetation. The result was the grassland type which
still prevails in that region.

This result naturally required thousands of years for its accomplish-
ment, but the modus operandi may be summarized as involving five
processes: (1) the disappearance of the arborescent flora; (2) the
great increase in the number of individuals of the herbaceous species ;
(3) competition for space among the herbs, leading to the eventual
dominance of the grasses, as the group best suited by growth-form and
ecological requirements to the new conditions; (4) the development by
evolution of new and characteristic species and genera, and (5) the
immigration, usually accompanied by specific or generic evolution, of
other species from the Sonoran deserts at the south and the Great Basin
deserts at the west.

How far toward the east the prairie vegetation may have extended
in Tertiary time and how many advances and retreats it may have
made in response to climatic variation are unknown. We may believe,
however, that the present climatic center of the Prairie Province in
western Kansas and Nebraska and eastern Colorado has been occupied
by this vegetation continually since its origin, and that amoeba-like
arms have been pushed out many times in many directions and with-
drawn again. It is also probable that the arctotertiary forests have
continuously occupied the Ozark uplift, since it still harbors many
old species, although not so many as the Appalachian uplift of the
same latitude, from which it was isolated during the Tertiary by the
oceanic waters of the Mississippi Embayment. This isolation and
the proximity of the Ozarkian region to prairie and Sonoran floras on
the west have led to a considerable differentiation between the Ozarkian
and Appalachian forest centers, as has already been noted in Part III.

Fossil evidence indicates that the arctotertiary flora included both
angiosperms and gymnosperms and had little or no latitudinal differ-
entiation, at least as far north as 70°. The second great floristic devel-
opment affecting our region was the general segregation of these two
groups into a northern flora, with gymnosperms predominating, and a
southern flora, in which angiosperms were dominant. This probably
began only with the approach of the first glacial period and may not
have been completed until the ice age was well under way. The
processes were essentially the same as those outlined above for the
prairie vegetation, but in this case involved also the permanent dis-
appearance from both floras of a number of species belonging to genera
whose modern representatives are tropical. The former presence of
Ficus, Artocarpus, Sabal, and other genera of similar climatic require-
ments in northern latitudes has frequently been taken to indicate a
tropical or subtropical climate in those regions. The writer sees no

EARLY GLACIAL STAGES 59

reason for believing that such genera could not have produced species
adapted to a temperate climate then, just as Phoradendron, Diospyros,
and Tripsacum have at present, so that the actual shift of climate north-
ward may have been only a few hundred miles.

These three floristic and vegetational types have since maintained
their identity, and the history of the vegetation of the Middle West is
concerned almost exclusively with them. A fourth floristic element,
the Coastal Plain flora of the southeastern states, was likewise segre-
gated from the arctotertiary forests, probably during the late Tertiary
or Pleistocene, and enters the southern portion of our region, but has
probably never occupied a more extensive territory than at present.
The Coastal Plain flora has been considerably modified by the immi-
gration, accompanied by evolution, of numerous tropical forms, enter-
ing the region via Texas or the Florida peninsula. Some of these have
extended still farther north and have invaded other floras as well.

Early Glacial Stages. — There is no present evidence concerning
the migrations of vegetation during the jSTebraskan and Kansan glacial
advances, at least one of which entered our area, or during the corre-
sponding Aftonian and Yarmouth interglaeial stages, and the earliest
evidence at hand deals with its location during the Illinoian period at
the time of maximum advance of the ice.

The southern boundary of Illinoian glaciation extends in a generally
southwesterly direction across Ohio and Indiana and lies on the north-
ern slope of the Ozark uplift in southern Illinois. It then turns north-
ward, following in a general way the present course of the Mississippi
river to the Wisconsin border, displacing the river westward in eastern
Iowa, but passing to the east of the driftless area in northwestern
Illinois. Its deposits in Wisconsin are almost wholly covered by
those of a later period and its further extent in the Middle West is
unknown.

The pre-Illinoian flora of the territory covered by this ice sheet must
have been destroyed or have migrated south, west, or in both directions.
The immediate problems are to ascertain the distance beyond the glacial
boundary to which it was forced and the location of the extraglacial
floras during the time of maximum ice advance. Several features of
present distribution seem to cast some light on the questions.

The Mississippi Embayment. — The Mississippi embayment of south-
ern Indiana, Illinois, and Missouri is occupied by a large number of
species which reach here their northern limits in the extensive flood-
plains of the Ohio, Wabash, and Mississippi rivers, without migrating
farther to the north along these floodplains into glaciated territory.

60 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

Noteworthy among them are Taxodium distichum* Rhamnus caro-
liniana, Gleditsia aquatica, Nyssa, aquatica, Ilex decidua, Catalpa
speciosa, Fraxinus profunda, Celtis mississippiensis, Quercus Phellos,
Leitneria floridana, and a host of other shrubs and herbaceous species.
Soil conditions can not be cited as the cause of this range, because the
soils in which these plants grow is all alluvial, washed down by the
streams from the glaciated region, and parallels them without any
essential difference for miles within the glacial boundary. Neither
can climate be cited, since two of them, Taxodium distichum and
Catalpa speciosa, are commonly, and others occasionally, cultivated
200 to 300 miles north of their natural limits.

Northern Limits of Southern Plants. — The rocky hills of the Ozark
uplift in southern Illinois also carry a number of southern plants
which reach here their northern limits, such as Pinus echinata^, Bato-
dendron arboreum, Ulmus alata, Azalea nudifiora, Bumeli-a hjcioides,
B. lanuginosa, and many others. These also have failed to utilize the
extensive strip of similar habitats along the rocky bluffs of the Missis-
sippi river for further migrations into the glaciated region. Still other
species have interesting ranges correlated in some way with the glacial
boundary. Thus Heuchera parviflora occupies isolated areas along the
boundary only, with the exception of its stations in the lower Allegheny
mountains. Trichomanes Boschianum preserves an outpost in southern
Illinois, where it has recently been discovered by Cowles. Sidlivantia
Sullivantii occurs only along the glacial boundary from Ohio to Illinois
and reappears in the driftless area 300 miles to the north. Saxifraga
Forbesii is strictly limited to two stations along the glacial boundary
in southern Illinois and southeastern Missouri. Phlox Stellaria is
chiefly confined to certain limestone cliffs along the glacial boundary
in Illinois and not far beyond it in Kentucky.

Undoubtedly there are species which reach their northern limit
near the glacial boundary because of climatic conditions. This might
be true of Phoradendron flavescens, whose host trees are abundant north
of the boundary, but which has not crossed itself. But with numerous
species with coincident range margins at this line, and several species
with isolated stations near it, an easier explanation is that they re-
mained here during the glacial period and have not migrated north-
ward since then, indicating that the northern boundary of the arcto-
tertiary forests lay parallel with and close to the ice margin.

There is at present no evidence whether this forest flora at that
time extended westward toward the Ozarks, leaving northern Missouri

* A single colony aceurs in Knox County, Indiana, within the glaciated area.

THE MIGRATION OF DECIDUOUS FORESTS 61

and Iowa to a prairie flora of western affinity, or northward along the
western boundary of the ice fields, restricting the prairie to a location
farther west. Extensive forest beds overlie the Kansan drift in north-
eastern Iowa, as shown by McGee,29 who states that more wood is pre-
served in these beds than at present grows in the area, but the time-
position of these forests may be either Yarmouth, Illinoian, or San-
gamon.

The Migration of the Conifers. — The conifers of the present north-
eastern forests, or their ancestral prototypes, migrated southward before
the ice and must have been reduced to a narrow strip between the ice
margin and the deciduous forests. There is no evidence whether this
flora extended west as far as southern Illinois or existed in the un-
glaciated region of northwestern Illinois. It is noteworthy that Pinus
Strobus does not occur on the exposures of St. Peter's sandstone near
the confluence of the Mississippi and Illinois rivers, but its absence
proves nothing. The only species of essentially boreal affinity now
existing in southern Illinois are Sulliva-ntia Sullivantii and Saxifraga
Forbesii, and it is not necessary to presume that they were ever accom-
panied by coniferous forests. Such forests may have existed anywhere
along the glacial margin, or may have been restricted to some espe-
cially favorable area, such as the re-entering angles in Ohio and Indiana,
or to the mountains of Pennsylvania and adjacent states. The promi-
nent development of vegetation of this affinity in the southern Alle-
gheny mountains may have arisen during a later glacial advance or
may have persisted since the Illinoian.

The Migration of Deciduous Forests. — During the long Sangamon
interglacial stage which followed the retreat of the Illinoian glaciers,
deciduous forests undoubtedly migrated to the north and probably also
to the west. The full extent of this migration can at present only be
surmised, since it must be ascertained wholly by fossil evidence, and
this is in most cases lacking. Nevertheless, logs of both angiospermous
and coniferous species are found frequently in the soil layer between
the Illinoian drift and the superposed Wisconsin soil in various parts
of Illinois. Leverett30 has noted that the coniferous remains must not
be considered evidence for the maintenance of such forests throughout
the interglacial stage, since they probably were deposited at the end
of it, just before the approach of another glacier. The notable dis-
coveries in the vicinity of Toronto, summarized by Chamberlin and

29 McGee, W. J. The pleistocene history of northeastern Iowa. U. S. Geol. Surv.
Ann. Rep. IH : 189-599. 1891.

so Leverett, Frank. The Illinois glacial lobe. U. S. Geo. Surv, Mon. 38. 1899.

62 H. A. GLEASOX VEGETATIOXAL HISTORY OF MIDDLE WEST

Salisbury,31 indicate that deciduous trees may have migrated during
the Sangamon stage to distances some 500 miles beyond their pres-
ent range.

There must have been other vegetational movements during the
Iowan glaciation and the succeeding Peorian interglacial stage, but
no evidence concerning them has been found, and none is expected
from the region under discussion, except from fossils which may even-
tually be discovered in the Peorian soils. The Peorian stage may have
had a relatively warm and arid climate, since during this time extensive
deposits of loess were made over wide areas in the Middle West. This
type of climate may have persisted into or even through the Wisconsin
glaciation.

Wiscoxsix Glacial Period. — The close of the Peorian interglacial
stage saw the advance of the Wisconsin glaciers, which as the last
glacial invasion have left the greatest impress on the nature and loca-
tion of the present vegetation. Originating in the Labradorian center,
the ice sheet flowed in a southwesterly direction into our area, wholly
covering the state of Michigan and partially covering Ohio, Indiana,
Illinois, and Wisconsin. Toward the periphery, the ice sheet was
divided into more or less well-marked lobes, with re-entering angles
between them. The ice margin along the eastern boundary of Ohio
reached nearly to the northern extremity of West Virginia, extended
across Ohio in a generally southwesterly direction to a point not far
north of Cincinnati, westward across Indiana in about latitude 39° 30',
westward into Illinois, and thence northward on about longitude 89°
to latitude 45° in Wisconsin. Here the boundary again turned west,
and another lobe extended south into the north-central portion of Iowa.
Throughout this area there is an extensive development of moraines,
the chronological relations of which have been discussed in admirable
detail by Leverett80'82-33 and Alden.34 These moraines mark succes-
sive stages in the retreat of the ice, and it is to be hoped that the
application of De Geer's methods to the region will eventually yield
an approximate time-scale for their history, an event which will be of
the highest importance in developing a clear idea of the consequent
migrations of vegetation. At the same time, a continued study of the

3i Chamberlin, T. C. & E. D. Salisbury. Geology. New York, 1907

32 Leverett, Frank. Glacial formations and drainage features of the Erie and
Ohio basins. U. S. Geol. <Surv. Mon. 41. 1902.

33 Leverett, Frank, & Frank B. Taylor. The pleistocene of Indiana and Michigan
and the history of the Great Lakes. U. S. Geo\. Surv. Mon. 53. 1915.

34 Alden, Wm. C. The quaternary geology of southeastern Wisconsin, with a
chapter on the older rock formations. U. S. Geol. Surv. Prof. Paper 106. 1918.

THE WISCONSIN

63

peat deposits27 is certain to yield precise information concerning the
nature of the flora.

Outside the margin of the Wisconsin ice lay the five floristic groups
of plants already mentioned, the prairie flora to the west, the Ozarkian
forests to the southwest, the coastal plain flora of the Mississippi
Embayment to the south, and the deciduous forests of the Appalachian
and Piedmont regions to the southeast, while the coniferous forests of
the present northeastern states were limited to a belt paralleling the
glacial margin. There is no reason to believe that the general space
relation of these five groups has ever been altered, although they must
have suffered many fluctuations in their distribution and extent dur-
ing the preceding glacial advances.

With the retreat of the ice, the new glacial soil was thrown open
to migration and speedily covered by plants. Adams has given a
picturesque account6 in general terms of the waves of vegetation which
swept on to the north a3 the ice margin retreated. But postglacial
migration was by no means as simple as he has described it, particularly
because of the opportunities for migration in easterly and westerly
directions as well as toward the north. It is also obvious that the
advantage in migration lies with the floristic type located nearest to
the new land. Hence it is important to locate the five floristic types
as accurately as possible for the time of maximum advance of the
Wisconsin ice.

Distribution of Plant Life During the Wisconsin. — When it
is considered that the past distribution of plants can in general be
interpreted only through circumstantial evidence, it becomes hazardous
to venture any opinion on the actual details of plant distribution during
this period. Nevertheless certain modern conditions are of interest
in this connection.

The glacial drift of southern Illinois beyond the Shelbyville moraine,
which marks the southern boundary of the Wisconsin glaciation, is
characterized by a large number of more or less xerophytic species.
Among these may be mentioned Ambrosia bidentata, Ascyrum hyperi-
coides, Chamaecrista nictitans, Crotalaria sagittalis, Crotonopsis
linearis, Diodia teres, Diospyros virginwana, Galium pilosum, Par-
sonsia petiolata, Passiflora lutea, Plantago aristata, Plantago virginiana,
Quercus marylandica, and Quercus stellata. Few of these are native
anywhere within the Wisconsin drift limits, although several of the
weedy species have within recent years migrated northward and
are now found rather uniformly throughout Illinois. They may be
described collectively as xerophytic selections from a southern flora.
The same region is also marked by a few rather peculiar xerophytic

64 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

selections from a western or southwestern flora, not found elsewhere
in the state, such as Geoprumnon mexicanum and Megapterium mis-
souriense, but it must be noted that these are much fewer than the
southern element, and also much fewer than species of the same western
element in western Illinois.

Some of the most noteworthy mesophytic trees are nearly or quite
absent from the same region, such as Fagus grandifolia, Cynoxylon
floridum, Liriodendron Tulipifera, and Magnolia acuminata, although
some of them extend much farther north in almost the same longitude
and all of them, crossing Illinois at its southern end, extend westward
as far as the Ozark mountains.

The Ozarkian flora at the present time barely reaches Illinois. A
few species occur on the rocky hills of the Ozark uplift, such as Soli-
dago Drummondii and Solidago Radula; Trillium viride occurs north-
east of St. Louis, but in general the flora is poorly developed east of
the Mississippi, although strongly marked only a few miles west of
it. Even such a mobile species as Yemonia Baldwini, common in
the vicinity of St. Louis, is scarcely known on the Illinois side. There
is no evidence as to the date when this flora appeared in Illinois, and
no reason to believe that it ever extended farther east or took a more
prominent part in the plant life of the region.

In western Illinois, beyond the Wisconsin drift, north of the low
range of hills which divided the Illinoian till into two portions (see
Leverett's map30), and mostly west of the Illinois river, there are a
number of intensely xerophytic western species, such as Bouteloua
hirsuta, Bouteloua oligostachya, Schedonnardus paniculatus, Opuntia
fmgilisj Mentzelia oligospermia, Lesquerella argentea, and Cristatella
Jamesii. They are not merely xerophytes, but are limited to peculiar
and extreme habitats. They have not been reported east of the ter-
minal moraine, even from similar habitats on the sand dunes of Lake
Michigan. Schedonnardus paniculatus, with a single station in Illi-
nois, and Cristatella Jamesii, with two known stations in Illinois and
one in eastern Iowa, present excellent examples of discontinuous dis-
tribution, since their nearest, stations to the west are 300 to 500 miles
away. Such distances are much beyond the normal range of their
migration and these eastern stations must be regarded as the relics
of a former nearly continuous range across eastern Nebraska and
Iowa. This in turn implies a climate some time in the past much
drier than at present. There are also in this part of the state a few
xerophytic southern plants, such as Quercus marylandica and Croton-
opsis elliptica, but their successional relations indicate a later arrival
in the region.

THE WISCONSIN 65

It has been indicated that coniferous plants during the Illinoian
glaciation must have been confined to a narrow strip paralleling and
adjacent to the ice margin. The same conditions must have obtained
during the Wisconsin. The lack of relics of this flora along the Shelby-
ville and Bloomington moraines is striking. From Edgar County,
Illinois, where the terminal moraine enters the state from Indiana,
west and north to La Salle County, almost no boreal relics occur.
The absence of pines may well be explained by the soil conditions, but
the swamps and bayous of the rivers are also without Larix and Thuja
and any of the ericads of the peat bogs of the north; even Caltha
palustris and Spathyema foetida are almost unknown along the glacial
boundary until well towards the northern end of the state, although
most of these plants are found freely along the moraine in similar
latitudes in Ohio and some in Indiana. At the north they reappear:
Abies balsamea, occurs in northeastern Iowa, and Pinus Strobas is
abundant from southwestern Wisconsin northward. Even Primula
Mistassinica occurs in northwestern Illinois.

All of these facts of modern distribution may be explained by
postulating a glacial climate during the Wisconsin considerably drier
than at present and not much different in temperature, so that the
vegetation of extra-glacial Illinois assumed a xerophytic aspect. Under
this view, we may assume that the Ohio valley in southern Indiana
and Illinois was occupied by its present forest flora, possibly not so
luxuriantly developed; the Illinoian drift to the north of it by a
xerophytic forest of southeastern affinity with a slight admixture,
decreasing toward the west, of a prairie element; western Illinois
was exclusively prairie, of a type similar to that now prevailing pos-
sibly 400 miles farther west. A narrow and interrupted strip of
coniferous forest followed the glacial boundary, especially in places
where greater topographic relief afforded better shelter. Toward the
east, across Indiana and Ohio, the strip became broader and included
more species. Toward the north it broadened out again in the drift-
less area in the shelter of the deep rocky ravines, with the additional
protection of the projecting Des Moines lobe of ice extending south-
ward to the west of them. It is doubtful if any conifers or associated
species occurred west of this lobe. Tundra vegetation, less affected
by the environmental conditions, grew on the thin soil overlying the
ice back of the glacial margin. In general, the climatic conditions
and vegetation may have been shifted in this latitude 300 to 400 miles
cast of their present location. The presence of an old flora in the
Ozarks indicates that this shift did not extend much farther toward
the south.

66 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

There is also some geological evidence of a climate during Wis-
consin time well suited to a xerophytic vegetation. Various rivers of
the region, notably the Wisconsin, the Green, and the Illinois, which
rise within the Wisconsin drift and flow out to the west or southwest,
occupy, beyond the Wisconsin terminal moraine valleys entirely out of
proportion to the size of the present rivers and choked with immense
quantities of glacial outwash. This may indicate unusually rapid
melting, due to a mild (and presumably also a dry) climate.

During this period, forest belts of the more mesophytic species may
have existed along the Mississippi, Wabash, Illinois, Missouri, Cedar,
and Des Moines rivers in the eastern portion of this semi-arid region,
just as they do today in a similar climate in Nebraska and Kansas.

Climate in the Early Post-Wisconsin. — Whether or not the
conclusion is accepted that a mild and semi-arid climate existed in the
Peorian and persisted through the Wisconsin stages, it seems almost
certain that such a climate characterized at least a portion of the
time involved in the post-Wisconsin glacial retreat. Palaeontological
and geological evidence for such a condition is scanty, and is concisely
summarized by various authors in a recent volume.35 Alden 35a believes
" that during the deposition of the post-Wisconsin loess the climate iu
the northern interior may have been somewhat drier than at present,
but was not greatly different," Knowlton says 35e " There is some little
palaeobotanical support for the contention that there was a slightly
warmer period following the close of the glacial epoch." Tyrrell con-
cludes3^ that the glacial climate of the Canadian northwest was " suc-
ceeded by a dry continental climate, under neither of which conditions
was a forest growth possible." Basing his opinion on the occurrence
of certain animal remains, Hay believes 35d " that after the retreat of
the ice-sheet a warmer period ensued," at the culmination of which
" the region along the southern shores of Lakes Ontario, Erie, and
Michigan enjoyed a climate similar to that now prevailing in Tennessee

as Die Veriinderungen des Klimas seit dem Maximum der letzen Eiszeit. Eine
Sammlung von Berichten . . . herausgegeben von dem Exekutivkomitee des 11.
Internationalen Geologenkongresses. Stockholm, 1910. Including:

35a Alden, Wm. C. Certain geological phenomena indicative of climatic conditions
in North America since the maximum of the latest glaciation. 353-363.

ssd Hay, O. P. On the changes of climate following the disappearance of the
Wisconsin ice sheet. 371-374.

35e Knowlton, F. H. The climate of North America in late glacial and subsequent
post-glacial time. 367-369.

35f Tyrrell, J. B. Changes of climate in northwestern Canada since the glacial
period. 389-391.

TUNDRA AND CONIFEROUS FORESTS 67

and Arkansas." Similar opinions have also been expressed by Cole-
man, Matthew, Dawson, Leverett, Upham, and Davis.

The same period was early recognized in Scandinavia by Blytta6
and has been generally accepted by most European phytogeographers.
Of the numerous papers on the subject published in the same volume,
those of Andersson 35b"c seem especially conclusive. He decides that
during the early postglacial stages the climate of Scandinavia had
winters similar to those of the present, while the summers were longer
and about 2.5° C. warmer. Later the climate became wetter, while
remaining equally warm, while for the modern period there is evidence
of a slow lowering of the temperature. It is also worthy of note that
Andersson lays great stress on the value of relic colonies as evidence
of past distribution.

The term xerothermic period has long been used by Europeans for
this period of mild and drier climate, and may well be extended to the
same type of climate, probably contemporaneous, in America.

A xerothermic period, occurring in comparatively recent times, and
without any subsequent geological episodes to modify greatly the trend
of plant migrations, would certainly have left an impress on the dis-
tribution of vegetation which would still be visible. Among the effects
which might be expected is a great extension of the prairie flora toward
the east, taking advantage of the favorable climate. This should now
be evidenced by relic prairie colonies and by isolated stations of western
species at the east and by a deficiency of hydrophytic and mesophytic
boreal relics at the west. Both of these results are actually demon-
strable at the present time, as will be shown below. A third effect
should be seen in the migrations of the deciduous forests from the
southeastern center, in which the more mesophytic species would be
limited at first to a northward migration only, passing to the east
of the area affected by xerothermic conditions, while the more xerophy-
tic species would migrate both north and west, accompanied by other
species of edaphically moist floodplains, such as now form the western-
most groves of Kansas and Nebraska. Present distribution indicates
that this was also a fact, although it is less clearly evident at the
present time.

Tundra and Coniferous Forest Migration. — The northern
boundary of the tundra vegetation overlying the glacial ice extended
toward the north with the retreat of the ice. At its southern boundary,

35b Andersson, Gunnar. Das spiitquartare Klima, erne zusammenfassende Ueber-
sieht. xii-lvi.

35c Andersson, Gunnar. Swedish climate in the late-quaternary period. 247-294.

36 Blytt, Axel. Dia Theorie der wechseluden kontinentalen und insurlaren
Klimate. Engler's Bot. Jahrb. 2 : 1-50. 1881.

68 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

the tundra persisted until destroyed by the warmer climate or by the
succession of the coniferous forests. In general, tundra withstands
a climate slightly warmer than its optimum better than the competition
of forest species, which quickly destroy it by shading.22 As in all
such retreating migrations, relic colonies were isolated, surrounded by
the advancing forests, and persisted for a longer or shorter time until
finally completely overgrown by the spruces and firs. Naturally those
endured the longest which occupied habitats least favorable to forest
growth. We may presume that the sand dunes along the shores of
the postglacial lakes and such isolated rock outcrops and cliffs as
Starved Rock in Illinois and the Dells of the Wisconsin River were
for a long time occupied by tundra plants, just as they are now occu-
pied by relics of the coniferous forests. The xerothermic period and
ordinary successional processes have long since destroyed the last vestige
of tundra vegetation in these latitudes, and the southernmost species
are now found scarcely farther south than Lake Superior. On Isle
Royale37 associations still exist which bear a general resemblance to
the arctic tundra and contain a few species of distinctly northern dis-
tribution. As the tundra migrated northward into cooler and moister
regions, and the cooling influence of the ice was felt farther beyond
its margin, the belt of tundra broadened, that is, it migrated toward
the north faster than it retreated from the south, until it now occupies
an area some hundreds of miles wide.

The boreal coniferous forests which had occupied a narrow or inter-
rupted strip along the glacial margin during the Wisconsin stage also
migrated northward following the retreat of the glaciers, surrounding
the relic colonies of tundra and eventually replacing them by succession.
The pioneers in the movement were probably then as now the xerophytic
species advancing along the rock outcrops and till of the upland, and
the bog species proceeding along the drainage lines and the glacial
lakes. These forests also advanced toward the north faster than they
retreated from the south, and consequently occupied a strip of increas-
ing breadth. This led to their present dominance over a large area
north and east of the Great Lakes and to their temporary dominance,
which is still continued in many habitats, in northern Minnesota, Wis-
consin, and Michigan. Their northward movement brought the coni-
fers into climates progressively cooler and relatively moister. This led
not only to a greater number and size of individuals, but also per-
mitted a northwestward migration into northern Minnesota and
Manitoba. This northwestern prolongation was eventually extended,

37 Gleason, H. A. The ecological relations of the invertebrate fauna of the Isle
ftoyale, Michigan. Eep. Mich. Geol. Survey 1908: 57-78. 1909.

.TUNDRA AND CONIFEROUS FORESTS 69

with the complete disappearance of the land ice, as far as Alaska.
Since the morainal strip of boreal vegetation was, in the Middle West
at least, always narrow, frequently interrupted, and probably com-
pletely absent about the Des Moines lobe, and composed chiefly of a
xerophytic selection of species, the western end of the forest belt was
composed, during its advance to its present location, of a comparatively
meager flora. Each mile of advance to the northward saw the widen-
ing of the belt and the junction of portions previously isolated, and
consequently permitted a westward migration of species from the
eastern states, which were little affected by the xerothermic climate.
This westward migration has by no means been completed and is
probably even now in progress through the forests of Quebec and
Ontario, but the boreal flora of the upper Lake region is still poor in
comparison with that of Quebec, New Brunswick, and the New Eng-
land states. The northwestern extension across Manitoba to Alaska
shows a progressive diminution in the number of species concerned,
as graphically shown by Transeau's38 map.

When these forests reached the eastern foothills of the Kocky Moun-
tains in Alberta, they were again in contact with the closely related
species of the Pacific Conifer Province, from which they had long been
separated. It will be interesting to learn from future studies what
exchange of species now took place ; whether widely distributed species
common to both regions, such as Populus tremuloides, had been in
both since their segregation ; whether the numerous Pacific plants of
the St. Lawrence valley are preglacial members of the northeastern
flora, or migrated eastward at this comparatively recent stage. It is
not too much to expect that all of these matters are possible of solution.

In this migration of boreal plants, numerous relic colonies were
left behind. Many of them have since disappeared before the encroach-
ment of surrounding vegetation or the changes in climate, but thousands
still exist. As is generally the case with relic colonies, they occupy
for the most part extreme habitats, either xerophytic rock hills and
sand dunes, clothed with Pinus Strobus and Pinus Banksiana, or bogs,
characterized by Larix laricina and Thuja occidentalis. It is especially
noteworthy that in Ohio, where the influence of xerothermic conditions
was reduced by the eastern position, the hydrophytic colonies persist
farther south than the xerophytic. In Illinois the condition is re-
versed, and neither of the two hydrophytic trees is reported from Iowa.
Tamarack bogs are found in Illinois only along the borders of deep
lakes in the immediate vicinity of Lake Michigan, while Pinus Strobus

38 Transeau, Edgar N. Forest centers of eastern America. Amer. Nat. 39 : 875-
889. 1905.

70 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

occurs not far from the Mississippi and in one county in southeastern
Iowa. The occurrence of boreal relics in the driftless area of north-
eastern Iowa has already been noted.

Ponds were abundant on the Wisconsin drift in Illinois half a
century ago, and were notable for the complete lack of such hardy
boreal plants as Comarum palustre, Dulichium arundinaceum, and
Menyanthes trifoliata. They were mostly shallow and easily drained
and have long since disappeared. They were also characterized by
the absence of deep peat deposits, indicating that they had never been
occupied by boreal vegetation and were of comparatively recent origin.
The carefully prepared reports of the Illinois soil survey present
interesting figures concerning them. Ten counties of central Illinois
contain only 1,986 acres of deep peat, or 0.04% of their area, and
of this 1,779 acres lie in the floodplain of the Illinois River. Three
counties of north-central and northern Illinois contain 0.15% of deep
peat, one county in the Kankakee valley 0.83%, while three counties
of northeastern Illinois contain 4.67%. Even in southern Michigan,
far within the boundary of the Wisconsin deposits, only the deeper
lakes are bordered by tamarack bogs. This lack of hydrophytic boreal
relics toward the southwestern angle of the Wisconsin glaciation is
best explained by the assumption of the xerothermic period, as already
noted, during which hydrophytic habitats were obliterated toward the
west, except in the deepest depressions or in local areas affected by
subsurface water,39 thereby restricting the relic colonies chiefly to the
xerophytic types.

Early Migration of the Prairie Flora. — An advance of the
prairie vegetation toward the east and northeast followed immediately
behind the coniferous forests, displacing the rearguard of the forest
by successional processes. There is now no place in the Middle West
where grassland is succeeding forest, and it becomes difficult to picture
the detailed steps by which such succession proceeded. In bogs, as
the climate grew warmer, and the ingress of water was reduced with
increasing distance from melting ice, the gradual drying may have
inhibited the growth of seedlings and permitted the entrance of prairie
species, This process is now seen in modified form in southern Michi-
gan, where partial drainage of tamarack bogs leads to the displace-
ment of the usual bog shrubs by Dasiphora fruticosa. On uplands,
the exposure of the marginal trees to warm and dry winds during the
summer, with consequent injury through excessive transpiration, may

39 Gates, F. C. A bog in central Illinois. Torreya 11: 205-211. 1911.

EARLY MIGRATION OF THE PRAIRIE FLORA 71

have had the same effect,40 while severe winter-killing among young
plants may have resulted from deficient snow cover.41'42. Certainly the
fewest boreal relic colonies are now found in southern Wisconsin and
southeastern Minnesota,43 where such atmospheric conditions are still
more or less in effect.* Geographically the advance of the prairies was
favored by the slow withdrawal of the eastern ice northward and the
rapid retreat of the western ice northeastward, thereby opening to
invasion first Illinois and then Indiana. A slight change in the nature
of the ice retreat might have affected the future vegetational develop-
ment of the region very greatly.

The distance to which the prairie vegetation migrated northward is
not definitely known. Very likely it reached in central Michigan as
far as the jack pine plains, which still contain numerous prairie species,
and if this is true must have reached similar or even higher latitudes
in Wisconsin. There is no present reason to believe that prairies were
developed on the north shore of Lake Erie. Deciduous forests in
Minnesota now occupy a narrow strip between the prairie and the
coniferous forests, and it is probable that their entrance was about
equally at the expense of the two earlier types of vegetation. If this
is true, the prairies did not extend much beyond their present range
in that state.

The eastern migration of the prairie proceeded as a wedge-shaped
extension between the coniferous vegetation at the north and the
deciduous forests at the south and reached limits considerably beyond
the eastern margin of modern continuous prairies. Numerous relic
colonies formerly occurred, before they were destroyed by agriculture,
in eastern Indiana, northwestern Ohio, and southern Michigan44 (see
also Cooper's description of an oak opening in his novel of the same
name). Vegetation closely simulating typical prairie exists along the
Scioto River near Columbus, Ohio. Bonser45 has described a marshy
tract near Sandusky, Ohio, which still contains some western species,
notably Vernonia fasciculata. This so-called prairie has doubtless

* Contrast maps in references 37 and 42.

40 Shimek, B. The prairies. Bull. Lab. Xat. Hist. Univ. Iowa 62 : 169-240. 1911.

4i Gates, F. C. The relation of snow cover to winter killing in Chamaedaphne
calyculata. Torreya 12: 257-262. 1912.

42 Gates, F. C. Winter as a factor in the xerophily of certain evergreen ericads
Bot. Gaz. 57: 445-489. 1914.

43 Livingston, B. E. A study of the relation between summer evaporation inten-
sity and the centers of plant distribution in the United States. Plant World 14:
205-222. 1911.

44Gleason, H. A. A prairie near Ann Arbor, Michigan. Ehodora 19: 163-165.

1917.

45 Bonser, Thomas A. Ecological study of Big Spring Prairie, Wyandot County,
Ohio. Ohio Acad. Stei. Special Paper 7. 1903.

72 H. A. GLEASON VEGETATIONAL HISTOKY OF MIDDLE WEST

preserved its hydrophytic environment since an early period, being
watered by springs, and consequently contains a considerable number
of boreal species, notably Betula pumila. Opuntia humifusa, occurs
near the same place. ,Species of distinctly western affinities do not
occur on Presque Isle, near Erie, Pennsylvania,46 and the abundant
development of boreal species and the typically Alleghenian aspect
of the flora in northeastern Ohio make it reasonably certain that
extensive prairies were not developed east of Cleveland.

The unglaciated areas of southern Ohio, Indiana, and Illinois, and
the Ozark region of southern Missouri are now populated with a forest
flora that presents indications of great antiquity, and in their greater
topographical relief offer better opportunities for forest species to
endure unfavorable atmospheric conditions. It seems reasonably cer-
tain that the prairies did not encroach on this area at any time. Prob-
ably the southern boundary of the glaciated area marked the division
between forest and prairie, and probably forest belts followed even
then the courses of the rivers northward well into the prairie region.
The forest belts of the Mississippi may have been derived from the
southern species of the Mississippi embayment; at the present time
several of them follow the course of this river northward, such as
Gleditsia aquatica and Hicoria Pecan. No evidence is now at hand
concerning the possible extent of the prairie into the Ozark region
of southwestern Missouri and northwestern Arkansas, What rela-
tion the xerothermic period may have had to the barrens of Tennessee,
the prairie belt of central Alabama, the prairies of eastern Arkansas,
and the coastal prairies of Louisiana must be left to the deductions
of observers who have had personal experience with the vegetation of
these areas.

In further support of the idea of a xerothermic period, it need only
be recalled that the discontinuous distribution of such plants as Opuntia,
fragilis and 0. humifusaj Sporobolus heterolepis, Cristatella Jamesii,
and Callirrhoe triangulata can be adequately explained only by assum-
ing a former period of climate sufficiently drier than the present to
permit their continuous migration. This, with the existence of relic
prairie colonies at the east and the prevailingly xerophytic nature of
the relic boreal colonies at the west, seems to be sufficient evidence
of the actuality of a post-Wisconsin xerothermic period.

46 Jennings, Otto E. A botanical survey of Presque Isle, Erie County, Pennsyl-
vania. Annals Carnegie Mus. 5: 289-421. 1909.

EARLY MIGRATIONS OF THE DECIDUOUS FORESTS

73

The eastward advance of the prairies was either accompanied or
followed by some specific evolution among the species participating,
as indicated by such plants as Phymosia remota" Synthyris Bullii, and
Tetraneuris herbacea. These plants are now confined to the eastern
arm of the Prairie Province, but in each case have their nearest related
species much farther west. i

There is no reason to suppose that the xerothermic period came to
a sudden close, which would require an equally sudden geological
event in explanation. In fact, it may have been most pronounced
during the Wisconsin glaciation and diminished in intensity ever since,
but not sufficiently to check the rapid advance of the prairies behind
the retreating coniferous forests. Neither can the time of maximum
advance of the prairies be correlated at present with the postglacial

stages of the Great Lakes.

i

Early Migrations of the Deciduous Forests. — At some time an
amelioration of the climate began. This change was probably less
connected with temperature, which has doubtless changed but little in
our area since the Wisconsin glaciation, than with rainfall and atmos-
pheric humidity. The total rainfall probably increased considerably,
but the increase fell chiefly during the winter months and changed
the climate from one of summer rains, such as is now characteristic
of the prairies farther west, to one of fairly equable rainfall. This
change first made itself felt in the east and gradually progressed
toward the west. At k,the present time, each hundred miles west of
the Wabash Kiver along the fortieth parallel shows a marked diminu-
tion in winter rainfall.

The effect .of this climatic change was the retardation and eventual
cessation of the old successional relation between coniferous forest
and prairie, soon followed by a break in the equilibrium along the
southern and eastern margins of the prairies, leading to an advance
of the deciduous forests northward and westward.

Two general routes were followed in this forest migration, a north-
ward and .westward route from the forests along the southern edge of
the prairies and a northward and then northwestward route from those
lying east of the easternmost extension of the prairies. Many species
were common to >oth regions and participated in both migrations ;
others were localized and show by their present distribution that they
entered the Middle West by only one of these routes. Both took ,place
simultaneously, and in some places both migrations eventually met,

47 Clute, W. N. The rarest American Plant. Amer. Botanist 26: 127-129. 1920.

74 II. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

so that races of similar ultimate ancestry but of different recent history
are again growing together.

The Southern Migration. — The southern migration was participated
in chiefly by oaks ..and hickories of the drier uplands, and by oaks,
elms, ashes, walnuts, maples, hackberry, cottonwood, honey locust,
coffee tree, buckeye, and sycamore of the lowlands. Following as
closely behind the prairies as environmental change would permit,
these plants established two successional series, a xerarch series on
the uplands, leading to the establishment of an oak-hickory forest, and
a hydrarch series along the stream valleys, leading to the develop-
ment of marginal belts of hydrophytic or mesophytic forest. The
hydrarch series migrated more rapidly up the stream courses, since
the species concerned could take advantage of local conditions of favor-
able moisture, while the xerarch series found its most favorable route
along the broken topography of the river bluffs, between the prairies
above and the hydrarch forests below. *lThe two types of forest thus
advanced upon the prairies together along the river courses leading
southward into the Ohio valley and eastward into the Mississippi, and
the present outposts of this migration are still to be seen in Kansas,
Nebraska, the Dakotas, and Manitoba on the west and in Michigan
and Wisconsin on the north.

Some of the mesophytes were left far behind in this advance, and
the beech, sugar maple, tulip tree, and hasswood are not particularly
important members of the climax forest or even totally absent from
Illinois west. In such cases a temporary climax developed in ravines
and on floodplains and was composed chiefly of a selection of species
persisting from the earlier successional stages. The slower trees have
been migrating, nevertheless, and with the greater success the farther
east. Beech is common in the Ohio valley: it has scarcely moved up
the Mississippi, but has followed the Wabash well into Indiana, and
the Miami, Muskingum, and Scioto until Ohio is occupied by it com-
pletely. There is reason to believe48 that large areas of forest were
developed on the uplands in which the most mesophytic tree was the
red oak, Quercus rubra.

The Northern Migration. — The northern migration proceeded from
the deciduous forests which had had a glacial center of preservation
in the Allegheny Mountains from Pennsylvania southward. These
species, now constituting the well known Alleghenian element in the
flora of New York and New England, have been during and since the

48Gleason, H. A. An isolated prairie grove and its phvtogeographical significance.
Bot. Gaz. 53: 38-49. 1912.

MIGRATION AND EVOLUTION 75

glacial period more or less mixed with, boreal elements, and this condi-
tion has been increased by their postglacial migration into a country
from which the coniferous forests were retreating and leaving numer-
ous relics behind, without an intervening prairie stage. Moving slowly
northward behind the glaciers, which occupied the Ontario basin long
after migration routes farther west were open, a portion of the species
concerned finally turned to the west and northwest and entered Michi-
gan, Wisconsin, and Minnesota, while others, such as Quercus Prinus,
moved scarcely west of the Appalachian region.

It has not been possible so far to form an opinion as to the par-
ticular postglacial stage of the Great Lakes when this westward turn
took place. It may have been in a narrow strip between the prairies
and the coniferous forests, following the former to the north and
broadening out only after reaching relatively high latitudes, in which
case it could have occurred at an early postglacial stage, or it may be
a more recent movement, even as late as the Xipissing stage, in which
case it might have passed to the north of Lake Erie and entered
Michigan by crossing the Detroit and St. Mary's rivers.

The Combined Effects. — The two migrations together may be com-
pared to a vast U, with its base in the southern Alleghenies, one side
extending as far west as the Ozarks, and the other passing northwest
to Minnesota. The northern migrants have shown little tendency to
move southward, especially toward the western part of their route,
while the southern ones have moved steadily northward until their
outposts have entered the area of the northern arm in Michigan and
Wisconsin, where oak, sassafras, and sycamore occur in the same area
with sugar maple, beech, and white pine. The base of the U has
gradually closed over Ohio and part of Indiana, chiefly through immi-
gration from the south. As a result there is little difference between
the forests of southern Michigan and those of any part of glaciated
Indiana. The valley of the Grand River in central Michigan is a
well marked division line between forests of the two types, according
to observations of B. E. Quick, and above it the deciduous forests are
derived mostly from the northern migration.

Migration and Evolution. — The forest migration was accompanied
by specific evolution also, and various species have been described from
the forest region of the Middle West. In most cases their probable
evolution is in doubt, from lack of careful study, but Dr. F. W. Pennell
has kindly supplied certain cases among the Scrophulariaceae in which
the ancestry of the species or varieties seems reasonably certain. Thus
Agalinis paupercula of the glaciated region is considered as derived

76 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

from A. purpurea; Aureolaria Pedicularui (L.) Raf. (Dasy stoma
Pedicularia) has given rise to the variety ambigens (Fernald) of the
Middle West. In other cases, the parent species, existing through the
glacial period south of the ice margin, has followed both routes of
migration and in so doing has become segregated into a pair of closely
related forms, one of which took the southern route leading northward
from the Ohio valley and the other the northern route along the Alle-
ghenies and the Great Lakes. Thus Trillium hivale, T. declinatum,t
and Cynoglossum virginianum of the southern portion of the Middle
West are paired with Trillium undulatum, T. cernuum, and C. boreale
respectively of the northern portion.

Changes in the prairie flora took place at the same time. Distinctly
western species withdrew entirely from the eastern extension or, like
Cristatella. Jamesii and others already mentioned, left relics behind in
especially favorable xerophytic habitats. Relic colonies were isolated
throughout the region, many of which still persist. These were sub-
ject to slow but persistent succession by the forests ; many must have
been obliterated, while others were greatly reduced in size.

Prairie and Boreal Relics. — The nature and location of these prairie
relics and of the boreal relics in the same area may yet give a clue
for the better understanding of the period and route of the northern
migration of deciduous forests. There is in general a great difference
in environment between forested and unforested habitats, such as a
prairie and either a coniferous or a deciduous forest. Differences also
exist between the two types of forest, but these are of less importance
and to many species of little significance. Boreal relics, therefore,
have not commonly persisted in the prairie, but remained in definite
colonies. Similarly, prairie relics have been confined largely to colonies
and have not persisted as scattered plants within the invading deciduous
forests.49 So, wherever the deciduous forests have succeeded the coni-
ferous forests directly, numerous boreal species persisting to the present
time bear evidence of the fact. These are of common occurrence in
southeastern Michigan and indicate that the prairies were not uni-
versal there, but are rare or lacking in Illinois and Indiana, showing
that the coniferous forests had disappeared before the advent of the
deciduous trees from the south. Relic colonies, as has been pointed
out before, remain in habitats not necessarily favorable to themselves,
but unfavorable to the invaders. Consequently, prairie and boreal
relics sometimes occupy the same station, into which they have been

o Vestal, Arthur G. Local inclusions of prairie within forest. Trans. Illinois
Acad. Sci. 11: 122-126. 1918.

THE PRAIRIE-GRASS FORMATION 77

forced by the invasion of the deciduous forests. Such conditions are
visible at the Dells of the Wisconsin, the dunes at the head of Lake
Michigan, and in various places in southern and southwestern Michi-
gan. The conclusion from both conditions is that in the western and
central part of our area successions between the coniferous forests and
the prairie were mostly completed before the arrival of deciduous trees,
while toward the east the deciduous forests arrived before the prairies
or during their advance upon the boreal associations.

Development of the Prairie-grass Formation. The prairie flora
also changed considerably during this advance of the deciduous forests.
Scores of species, which had been inhabitants of sand dunes, rocky soil
and barrens, or the more xerophytic upland woods, migrated faster than
the trees and became firmly established in the prairie associations. The
only restrictions on this movement were that the plants should be herbs,
dying to the ground every year and thereby escaping the effect of the
arid xerothermic winter, and adapted to compete for space with grasses,
which were the dominant plants of the prairies. Among these emigrants
were numerous southeastern grasses which, having the same vegetative
form, mingled easily with the western species and became dominant
with them.

As the xerothermic period finally drew to a close, the western species
were at a disadvantage. They were unable to withstand a large increase
in moisture, while the species of eastern derivation could exist with a
much gTeater rainfall. The latter accordingly became more and more
abundant, until they practically dominated the prairies. In fact, the
four most important grasses of the Illinois prairies, Andropogon furca-
tus, A. scoparius, Sorghastrum nutans, and Spartina Michauxiana, are
all of eastern origin. The western species became reduced in number of
individuals or restricted to extreme habitats where edaphic conditions
compensated for the increased rainfall.

Here we have the origin of the flora of the eastern arm of the Prairie
Province, early recognized by Pound and Clements as distinct from that
of the western plains, and designated by them the prairie-grass forma-
tion. Just as the original western prairies were segregated from the
arctotertiary forests by climatic differentiation, so the xerothermic
period produced an analogous vegetation thousands of years later from
the same stock. In the later case, however, the process was one of select-
ive migration, rather than of extermination and evolution, and most of
the southeastern species now characteristic of the prairie-grass formation
still live also in the forested region of the southeastern states. Never-
theless, the movement was doubtless accompanied by evolution in many
instances: Asclepias Meadii, a typically prairie species, seems to be

78 H. A. GKLEASON VEGETATIONAL HISTORY OF MIDDLE WEST*

a derivative of Asclepias amplexicaulis of a more eastern range; a vari-
etal evolution from Agalinis termifolia, illustrates the same geograph-
ical relation, according to Dr. F. W. Pennell, and other similar cases
might be cited.

The numerous ponds of the eastern arm of the Prairie Province were
also formed at this time, in response to increasing rainfall in a region
physiographically immature. They were of necessity colonized almost
completely by species of southeastern origin, since true hydrophytes
were not ^ound in the western vegetation.

Extent of Foeest Migration in the Prehistoric Period. The
northern arm of the forest advance lay chiefly or wholly in a region not
reached by .the prairies. There is no reason to believe that it has
suffered any retardation to the present time, except as caused within
the past century by commercial developments in lumbering and agri-
culture.

The southern migration, on the other hand, was quite different in
nature, due to a less favorable climate and to the character of the vegeta-
tion encountered. It has already been noted that it followed the water
courses, most of which, as tributaries of the Mississippi, offer convenient
highways leading from the center of forest preservation into the area
of xerothermic prairies. It has also been noted that the longitudinal
advance following the streams was more rapid than the lateral advance
at right angles to them. The forests therefore soon assumed the form
of long branching strips, following the rivers from their mouths toward
their sources. Through the slow lateral advance these strips were
widened and eventually became confluent across the interfluvial prairies,
frequently isolating portions of the prairie in the process.

The nature of the forest advance was determined by the location of
sufficient environmental change to permit the migration. Since the
unfavorable environment was chiefly due to conditions of rainfall and
atmospheric humidity, the advance followed the lines of topographic
relief, where some shelter was offered from the drying winds and the
subsurface water stood at a greater height, and the water courses where
edaphic moisture compensated for low rainfall. The rolling surface of
the moraines offered favorable conditions for trees of the xerarch series,
and they were soon colonized by forests of oak and hickory, reaching the
moraines by the intersecting streams and spreading across from one
stream valley to another. On the other hand, the wide alluvial bottom
lands of the larger rivers, notably the Missouri, Mississippi, and Illinois,
seem to have resisted forest invasion, and on them the forests were
limited to relatively narrow strips along the channel and the abandoned
oxbows, alternating with strips of prairie. This condition, recorded

FOKEST MIGRATION IN PREHISTORIC PERIOD 79

by numerous early observers, is so unlike our modern experience with
floodplain vegetation in the Middle West that it at once stimulates
inquiry into the circumstances of more recent forest development.

In a forest advance along the bluffs of a river flowing from north to
south through a region of prevailingly westerly winds, the greatest pro-
tection against atmospheric conditions will be found on the western lee
side of the river, rather than on the windward eastern side. Actual
conditions throughout the Middle West, however, show an unexpected
distribution, in that by far the largest areas of forest were originally
found on the eastern side of such streams. This has been discussed in
an earlier article,50 in which the explanation was offered that the irregu-
lar distribution is due to prairie fires, which, driven before a westerly
wind, have gradually destroyed much of the forests on the western side
but have had little effect on those to the east of the stream. The period
of great forest migration must therefore have closed with the advent of
the Indian and the prairie fire.

The extent of forest advance before that time can only be approxi-
mated by a study of the present distribution of morainal forests and of
the relative amounts on different sides of streams. Planimeter measure-
ments have been made of maps showing forest distribution in two parts
of Iowa and in three counties in Illinois and have given results of
remarkable but not necessarily significant uniformity. In Shimek's51
map of the Lake Okoboji region, Iowa, 26.5% only of the forests are
on the west side of the lakes, aud 73.5% on the east side. McGee's29
map of northeastern Iowa shows for the Wapsipinicon Kiver 27.9% on
the west and 72.1% on the east, and for the Cedar River 29.7% on the
west and 70,3% on the east. Tracings of the original land surveys
of Champaign, Piatt, and Moultrie Counties, Illinois, show forests on
the west side of the streams amounting to 31%, 31.6%, and 29.8%, and
on the east side 67.3%, 67.4%, and 70.2% respectively. In Champaign
County 1.7% and in Piatt County 1% of the forest area is in isolated
groves apart from any river system. If we may assume that forest was
originally developed equally on both sides of the streams, the total area
would be increased by amounts ranging from 36% of the present area
in Piatt County to 44% along the Wapsipinicon River. The lakes of
the Okoboji region would have additional forest on the west side amount-
ing to 47% of the present total, and if they were carried completely
around the lakes the increase would be still greater.

soGleason, H. A. The relation of forest distribution and prairie fires in the
Middle West. Torreya 13: 173-181. 1913.

si Shimek, B. The plant geography of the Lake Okoboji region. Bull. State Univ.
Iowa. Bulletins from the laboratories of natural history 72. 1915.

80 II. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

The extent of migration up the streams beyond the present limit of
forest can be considered with less accuracy. The original land surveys
frequently show small areas of forest isolated from the main body along
the head waters of a stream. Also, the upstream end of the forest belts
is regularly wide and blunt. Modern observation shows that in the
same area establishment of new forest is rapid as far upstream as there
is a differentiation of upland, bluff, and floodplain. If the interruped
belts were made continuous and were extended upstream according to
the topography, there appears to be room in Champaign County for at
least 28 more miles of forest belt, comprising probably 14 square miles.
There is also evidence that the moraines were forested wholly or in part
in prehistoric time. Fragments of the forests still persist in sheltered
places in Champaign County, and well developed morainal forests
occupy considerable areas of McLean County. In Champaign County,
about 50 square miles of land is morainal and may have been forested.

The land survey maps indicate 44,846 acres of forest in Champaign
County, and this figure is apparently somewhat too small, since the
surveyors frequently referred to trees up to two feet in diameter stand-
ing on land designated on the map as prairie. Increasing this amount
by 64 square miles makes a possible total of nearly 86,000 acres of
prehistoric forests, almost twice the present extent, and increases the
proportion of forest, from 7.1% to 13.5% of the total area. While the
accuracy of this estimate can not be proved, we may safely conclude
that the forested area was at one time greatly larger than its present
extent through Illinois and Iowa, and doubtless also through the prairie
region of Wisconsin and Indiana.

Second Period of Prairie Dominance. The advance of the forest
would in time have given a continuous forest covering to the Middle
West, except on those limited habitats where tree growth is impossible
for edaphic reasons. The forest would have extended west to the
climatic limit of tree growth and tongues would have projected beyond
the principal area along the habitats where special environmental con-
ditions compensated for the unfavorable climate. This advance came to
a sudden conclusion with the arrival of the American Indian and the
consequent introduction of prairie fires. This stage in the history of
the Middle West has been discussed in an earlier paper50 and needs
only brief recapitulation here.

Prairie fires were set annually by the Indians in the autumn months
to drive game from the open prairies into the forest, where it was more
easily stalked. Sweeping eastward before the prevailingly westerly
winds of that season, the fires destroyed seedling trees at the west margin
of the forest, preventing further advance in that direction. It is doubt-

SECOND PERIOD OF PEAIKIE DOMINANCE 81

ful if they penetrated far into the forest, but by destroying the under-
growth and killing the more susceptible species, they gradually reduced
the forest to the open parklike condition known as oak openings. Here
the herbaceous vegetation was composed largely of prairie grasses, which
furnished additional material for fires and led to the annual extension
of the prairie through the eventual death of the remaining trees. The
forest was thus slowly pushed back to the river bluffs, where the greater
topographic diversity favored its persistence, or, where the bluffs were
low and with gradual slopes, was completely destroyed on the west side
of the streams. Fires even crossed the smaller streams, attacked the
forests of the eastern side, and by burning through them isolated areas
of forest toward the headwaters of the stream from the main body below.
This condition is shown in several places on the original land surveys
of central Illinois, and examination of them indicates that the separa-
tion was always in places of low relief, leaving forests isolated where the
bluffs were steeper or higher.

On the eastern side of the forest, protected against fires driven by
the prevailing winds, encroachment by the prairie was less rapid and
doubtless in many cases negligible. It is probable that the forest
advance was completely stopped in some places, but may have continued
slowly in others.

Along the moraines, which are generally characterized by a gently
rolling topography, the forests were destroyed completely, except in
those few favored situations where ponds or swamps on the west side
acted as an effective barrier to the fires.

Another effect of prairie fires was the production of the so-called
barrens, areas characterized by a sparse growth of "scrub" oak
(apparently Quercus velutina in most cases), hazel, and wild plum.
The exact cause of the barrens is unknown, but descriptions from early
literature and accounts of personal observers in Illinois, Indiana, and
Iowa, agree that these three species of plants were the principal, if not
the only, woody species present, and that they seldom exceeded four or
five feet in height. They seem to represent a late stage in the degenera-
tion of the forest, where seedlings were established and the young plants
burned to the ground frequently without completely destroying the root
system. This developed further each year, reached a large size, and
sent up sprouts annually which persisted a year or two longer before the
next fire again killed them to the ground.

During this same period, a few new isolated groves were established
on the prairies, but always in areas well protected from the fires. Two
such have been examined in Champaign County, Illinois, are mentioned
before in this paper, and described in more detail elsewhere50. They

82 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

are composed exclusively of mobile species and have been in existence
such a short time that they have not yet changed the nature of the
prairie soil52.

Second Period of Forest Advance. — Settlement of the forested
regions of the Middle West began about the opening of the nineteenth
century, progressed steadily westward, and reached the Missouri Kiver
in fifty years. Actual settlement of the prairies was long avoided, vast
areas were still untouched at the time of the Civil War, and prairie
fires did not cease being a menace in parts of Illinois until 1860 and in
Iowa until somewhat later. As soon as fires ceased, the advance of the
forest was renewed, and at a rate probably more rapid than the original,
since considerable improvement of climate may have taken place since
the first advance was stopped by fires.

Early literature contains many accounts of this spread of the forested
area, which was rapid enough to attract the attention of travellers. It
seems to have progressed in several different ways. The lateral advance
of the forest at right angles to the streams and the longitudinal advance
along the streams continued as before. Willows, cottonwoods, and some
hydrophytic shrubs moved rapidly up the rivers and creeks, and were
followed by elms, maples, and ashes. Thirty years after the first prairie
settlements were made in Macon and Moultrie counties, Illinois, these
fringing forests had extended one or two miles up the smaller streams.
Isolated colonies of mobile species, especially willow, Salix humilis, and
wild plum, Prunus americana, were established well out on the prairie,
grew into thickets, and frequently received further additions of forest
species, carried in by the birds which visited them or by wind. Large
areas of barrens were converted into forest as by magic, when the fires
that had maintained them were stopped and the oak sprouts became
trees.

The total afforestation during this period was considerable and in some
cases almost unbelievable, and may best be indicated by a few examples.
Thus the driftless area of northwestern Illinois and southwestern
Wisconsin, which now gives the impression of having been heavily
forested, was 80-90% prairie a century ago,53 and was described by
Keating54 as presenting "the waved appearance of a somewhat ruffled

62 Hopkins, C. G., J. G. Mosier, E. Van Alstine, & F. W. Garrett. Champaign
County soils. 111. Agr. Exp. Sta. Soil Report 18. 1918.

ss Schoekel, Bernard H. History of development of Jo Daviess County. Illinois
State Geol. Surv. Bull. 26. 1916.

s* Keating, William H. Narrative of an expedition to the source of St. Peter's
River . . . performed in the year 1823 . . . under the command of Stephen H. Long.
Philadelphia, 1824.

COASTAL FLORA OF THE GREAT LAKES 83

ocean; it is covered with a dry short grass," while the higher knobs,
now concealed by the forest, rose from the midst of the prairie and
were visible thirty miles. A little later we find farmers near St. Louis
complaining that the rapid growth of trees had seriously restricted
the natural pastures.55 The Mississippi and Illinois rivers at that
time ,had long strips of prairie in their floodplains, which are now
completely forested, except where under cultivation. The dune region
at the head of Lake Michigan, which is now well forested with oak, was
then "treeless, except for a few stunted pines." Forested areas in
Indiana were developed, within the memory of men now living, from
barrens over which a man on horseback had been visible for six miles
(personal communication from Mr. C. C. Deam). The rapid develop-
ment of forests on the sand dunes of the Illinois River has been noted
elsewhere.56

This great forest advance of the nineteenth century proceeded by
ordinary successional processes and affected hundreds of square miles.
It ceased only with the complete occupation of the prairie land for agri-
culture. Since then, that is, during the last 50 years, forest succession
has been limited to sand dunes, roadsides, and railway lines. Every
roadside in the Middle West is a potential forest and would speedily
become so in fact except for constant effort in keeping the thickets and
young trees cut down.

Coastal Flora of the Great Lakes. — The extensive sandy beaches
of the Great Lakes are characterized by a number of species which do
not occur or are rare elsewhere in the Middle West. The geographical
origin of many of these species is indicated by their distribution else-
where or by the range of their nearest relatives. Thus 7ns lacustris
is closely related to Iris cristata,; Solidago Houglitonii to S. Biddellii
and S. ohioensis, and both were probably derived by immigration from
the south, accompanied by specific evolution. Cirsium Pitcheri may
have a similar origin. Corispermum hyssopifolium, for similar reasons,
appears to have entered the region from the west. Besides these, there
is a considerable group of species which are distributed along the
northern shores of the Atlantic, sometimes in Europe as well as Amer-
ica, are not ordinarily found inland, but appear along the shore of
the Great Lakes. This group includes Ammophila arenaria, Elymus
arenarius, J uncus ba^ticus, Halerpestes Cymbalaria, Cakile edentula,
Prunus pumila, Lathyrus maritimus, Chamaesyce polygonifolia, Hud-

65 Peck, J. M. A gazetteer of Illinois. Philadelphia, 1837.

ee Gleason, H. A. The vegetation of the inland sand deposits of Illinois. Bull.
Illinois State Lab. Nat. Hist. 9: 21-174. 1910.

84 H. A. GLEASON VEGETATIONAL HISTORY OF MIDDLE WEST

sonia tomentosa,, and Tanacetum liuronense. Two other species, Salix
syrticola and Solidago Gillmani, are endemic to the Great Lakes, but
the distribution of their nearest relatives indicates that they are of the
same ultimate geographical origin.

The most plausible explanation for the occurrence of these species so
far inland from their maritime range is that they migrated along the
shores of the Champlain Sea during the Algonkian or Nipissing stages
of the Great Lakes, and entered the lake region along the Kirkfield or
North Bay outlets.33 This explanation agrees with the present center
of abundance of the lake shore species near the north end of Lake Huron
and, since the ice fields were not far away during this period, with their
prevailingly northern maritime distribution.*

Summary of Part IV and Conclusion. — 1. The principal vegeta-
tional and floristic elements of the Middle West were differentiated
during the Tertiary Period and have continuously maintained their
present relative position.

2. Each glacial advance has induced extensive migrations toward the
south and west, but in them the coniferous and arctic floras have been
greatly narrowed and the deciduous forests have lived uninterruptedly
in the Ohio valley.

3. Extensive readjustments of distribution have taken place in each
interglacial stage. During the Sangamon, most or all of the Middle
West was probably occupied by deciduous forests.

4. Climate in the Middle West beyond the Wisconsin glaciation was
semi-arid during the time of maximum ice advance, thereby restricting
the boreal floras greatly toward the west.

5. A xerothermic period followed the retreat of the Wisconsin ice,
and caused the eastward migration of prairies as far as Ohio, succeed-
ing the coniferous forests as the latter migrated north.

6. A gradual increase in rainfall was accompanied by a westward
migration of deciduous forests. This was composed of two elements,
one moving from the Ohio valley and one from the Allegheny Moun-
tains. At the close of this period, forests occupied a much greater
proportion of the Middle West than at present.

7. With the arrival of the Indian and prairie fires, prairies again
began to encroach on the forests and reduced the forested area to its
condition at the beginning of the last century.

8. During the nineteenth century extensive afforestation took place,
continuing until the land was placed under cultivation.

* Since this paper went to press, D. C. Peattie has discussed this subject in
more detail in Rhodora 24: 57-70, 80-88. 1922.

SUMMARY AND CONCLUSION 85

The writer has no doubt that further study of the problem of glacial
and postglacial vegetational history, based on additional facts of modern
distribution, on better knowledge of plant migration and the ecological
requirements of species, and on new evidence from fossils, and carried
on by investigators with different viewpoints and different personal
experiences, will lead to many important modifications of the outline
here presented. The field is large, almost untouched, and of com-
pelling interest, and it is hoped that it will attract the attention of
students not only in the region here discussed, but elsewhere through
the country as well.

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 242

24*

EVOLUTION AND GEOGRAPHICAL DISTRIBUTION OF
THE GENUS VERNONIA IN NORTH AMERICA

By Henry Allan Gleason

New York
1923

[Reprinted from the American Journal of Botany, 10: 187-202. April, 1923]

[Reprinted from the American Journal of Botany, X: 187-202, April, 1923.]

EVOLUTION AND GEOGRAPHICAL DISTRIBUTION OF THE
GENUS VERNONIA IN NORTH AMERICA

Henry Allan Gleason
(Received for publication July 13, 1922)

The genus Yernonia, with its vast assemblage of over 500 species, ranges
through the western hemisphere from Argentina to Manitoba, occupying
a region of great climatic variation and habitats of great ecological diversity.
In preparing manuscript for the treatment of the genus in the "North
American Flora," 123 species have been recognized north of Colombia and
Trinidad. Within this number a few stand comparatively isolated from all
the others, while many are so closely related in form and structure and so
similar in distribution that they must be closely akin genetically. Over
30 species-groups may be distinguished in this way. Within these minor
groups evidences of specific evolution correlated with geographic distribu-
tion are frequently seen, while most of the groups, considered each as a
whole, present strong evidence in favor of their relation to, and probable
origin from, each other. It is therefore possible to build up a general
scheme of evolution and migration within the genus in which the two lines
of evidence, structural and geographic, complement and support each other.

There can be little doubt that the ancestral home of the genus, as far
as North American species are concerned, is tropical South America. This
is shown by the presence there of a large number of species of greater struc-
tural diversity than exist on the North American continent, and also by the
fact that many South American species are of a structure which clearly
indicates their primitive nature.

Within the genus as a whole, the more fundamental structural differ-
ences, which have been used in the division of Yernonia into its many
sections and subsections, relate chie.'y to the structure of the achenes, the
pappus, and the involucral scales. Nothing can now be said concerning
the possible evolution of these groups. Of those distinguished by Bentham
and Hooker in "Genera Plantarum " and accepted by Hofman in "Die
nati:rlichen Pflanzenfamilien," four have reached North America.

1. The section Stengelia, of the East Indies, with veiny, foliaceous in-
volucral scales, is represented by a single species, V. anthelmintic a (L.)
Willd., sparingly introduced into a few islands of the West Indies. Certain
Mexican species which bear a superficial resemblance to this section appear
to belong rather to the section Lepidaploa, and their similarity to Stengelia
is better explained by convergent evolution.

2. The section Tephrodes.of the paleotropical region, with terete achenes,
is represented by a single species, V. cinerea (L.) Less., widely introduced

187

188 AMERICAN JOURNAL OF BOTANY [Vol. 10,

as a weed in tropical America and recently reported from extreme southern
Florida. Its further migration through the agencies of commerce is to be
expected.

3. The section Stenocephalum, with coriaceous, spine-tipped involucral
scales, represented by several species of tropical South America, has a single
little-known and recently discovered species, V. jncunda Gleason, in the
mountains of southern Mexico. If we assume from the negative evidence
at hand that the section is actually absent from Central America, we may
infer an early migration of the section northward, followed by extinction
in Central America and the isolation of the single species in Chiapas.

4. The section Lepidaploa, with membranous involucral scales and
ribbed achenes, includes 120 species of North America and many more in
South America.

Within this large section there is still well-marked evidence of evolu-
tionary development in structure, illustrated most plainly by the character
of the inflorescence. One type of inflorescence may be assumed a priori
as the most primitive, and from it by successive small changes all the other
types may be derived. Since the center of evolution and migration for the
genus is considered to be tropical South America, where this primitive
type is largely developed, and since the succeeding stages in the modifi-
cation of the inflorescence occur progressively farther to the north, structure
and distribution complement each other, and it may be assumed with little
hesitation that migration and structural evolution have proceeded simulta-
neously; that the tropical species, while not necessarily the oldest in time,
are at least the most primitive in structure, and that the outlying species
of the temperate part of North America are both young in age and advanced
in evolution. It can not now be stated whether there is a similar correlation
between structure and distribution among the species of South America.

In the section Lepidaploa, the inflorescence is either a scorpioid cyme
or some other type of cluster obviously derived from it by certain apparent
modifications in the original structure. In such an inflorescence each
head is morphologically terminal; a lateral branch, arising at the first node
below the involucre, terminates in a second head and bears another lateral
branch which behaves in the same way. There is thus produced a more or
less elongate sympodial axis, morphologically indeterminate in its develop-
ment, and with its series of truly terminal heads apparently lateral and
secLind along it. Since the successive lateral branches arise from nodes,
which are normally marked by leaves, it may be assumed at once that the
leafy scorpioid cyme is the primitive inflorescence, while those species in
which the bracteal leaves are suppressed stand relatively higher in the scale
of evolution.

Each segment of such an inflorescence then consists of a basal internode
with a leaf at its summit and a head beyond it. The structure of the brac-
teal leaves varies greatly, but in general they maintain the form and pu-

Apr., 1923] GLEASON — VERNONIA IN NORTH AMERICA 189

bescence of the cauline leaves and differ from them chiefly in size. The head
is pediceled if separated from the bract by an obvious internode, and the
inflorescence is then a scorpioid raceme. If this internode is reduced, the
head is sessile and the cluster becomes a scorpioid spike. The clusters may
be straight or flexuous, long or short, crowded or loose, with heads ranging
from 2 to 25 in number. As a result of the straightening of the sympodial
axis, the heads appear lateral and are usually placed about 900 around the
axis from the bracteal leaf.

The leafy scorpioid cyme is found in 57 species, ranging throughout the
West Indies and on the continent extending north into southern Mexico.
Species with leafless cymes, 63 in number, occur commonly on the continent
from Panama to New England. From this region four have crossed the
narrow gulf east of Yucatan and entered Cuba; one has reached the Bahamas
from Florida; one is endemic to St. Vincent, and another reaches Trinidad
and the neighboring islands. In general, the distribution of the latter group
is continental, of the former Antillean.

Perhaps the simplest type of the primitive leafy inflorescence is found
in the species-group Graciles, in which the cymes are stemlike and quite
undifferentiated from the truly vegetative portion of the stem, with
bracteal leaves closely resembling the cauline in size and shape. Species
of this group are almost entirely South American, ranging, according to
Ekman, from Colombia to eastern Brazil. One species only, V. gracilis
H. B. K. var. tomentosa Ekman, of Bequia, occurs in our range.

A second group, with almost equally simple inflorescence, composed
of long, irregular, branching cymes, with long internodes and lea: ike bracts,
is the Argyropappae, of tropical South America, Central America, and
southern Mexico. The South American origin of the group may be assumed.
From there, V. remotiflora Rich, has been introduced into St. Thomas;
V. acilepis Benth. is endemic to Costa Rica, and V. argyropappa Buek
extends from Peru to Mexico. Two offshoots of the latter have arisen in
Mexico, V. hirsutivena Gleason in Yucatan and V. ctenophora Gleason in
Campeche, differing in minor structural details.

It will be observed that these two groups, simplest in structure, are
distributed primarily in South America and that only a part of their species
reach North America, although among these are three endemic species and
one endemic variety.

A third group of similar primitive structure as to inflorescence is the
Schiedeanae, of Central America and southern Mexico. While its members
differ sharply from the preceding group in their large heads, the peculiar
development and specialization of the involucral scales, and the absence
of foliar resin dots, they retain the simple cymes and broad, heavy leaves,
and may possibly be derived from it. V. vernicosa Klatt, with narrow
acuminate scales, appears to be the simplest and is endemic to Costa Rica.
V. Seemanniana Steetz follows in Costa Rica, with broad, obtuse scales,

190 AMERICAN JOURNAL OF BOTANY [Vol. 10,

and the greatest modification is found in V. Schiedeana Less., ranging from
Honduras to Vera Cruz, with involucral scales broadly dilated at the tip.
The progressive specialization in structure, correlated with increasing dis-
tance to the north, is here clearly shown.

There now follow seven species-groups with 33 species, all West Indian,
all clearly related, and all exhibiting a remarkable correlation ,betwee
structure and distribution.

The most primitive of these, from which the other six are directly or
indirectly derived, is the Arborescentes, ranging from the Windward Islands
to Jamaica. The wide range and primitive structure probably indicate an
early arrival in the region. The most primitive species, V. icosantha DC,
has stems bearing leaves of normal size to the apex and terminating in a
single sessile head. At the base of this head the two primary cymes arise;
they are straight, elongate, with prominent internodes, sparingly branched
or simple, and bear numerous heads. The chief distinction in the inflo-
rescence between it and the Graciles is the regular presence of paired primary
cymes. In V. sericea L.C.Rich., of the Virgin Islands and Porto Rico,
the cymes are shorter and more frequently branched. V. borinquensis
Urban, of Porto Rico, has exceedingly fiexuous, many-headed, freely
branched cymes, the branches invariably arising at the base of a head. V.
arborescens (L.) Sw., of Jamaica, has numerous frequently congested cymes
and reduced bracteal leaves. V. permollis Gleason, of Jamaica, completes
the group, with congested cymes and an unusual development of foliar pu-
bescence. The general tendency of the group is toward the production of
cyme-branches and supernumerary cymes, making a congested inflorescence
in which the bracteal leaves are reduced.

The Longifoliae, a group of three species, is related through V. longi-
folia Pers., of the Lesser Antilles,, to V. icosantha. Superficially the two
species are much alike, but the inflorescence in the former shows a distinct
difference. The primary cymes are short, compact, divergently spreading
at an angle of 60-900, crowded, bearing only 2-5 heads on short internodes
with bracteal leaves considerably smaller than the heads. Secondary cymes
arise just below the primary in the upper leaf axils. They are essentially
leafless for the first 2-5 cm., and then bear toward the summit either the
usual crowded heads or a terminal head subtended by two short cymes.
This whole inflorescence terminates completely the growth of this portion
of the stem, but during the next vegetative season new branches appear
from the next lower axils in order, grow out at a divergent angle, soon
surpass the old cymes of the previous season, and at the next blooming
season bear their cymes in turn. The plant has therefore a method of
continuing its vegetative growth beyond one season, and as a result reaches
a considerably larger size. V. Shaferi Gleason, of Montserrat, is closely
similar, and represents an island endemic. V. albicanlis Pers., of the Virgin
Islands and Porto Rico, preserves the same inflorescence but differs in its
obtuse or broadly rounded leaves.

Apr., 1923] GLEASON — VERNONIA IN NORTH AMERICA I9I

It is a comparatively short distance across open water to the north of
Porto Rico, the home of V. albicaulis, to the southern islands of the Bahamas.
In these southernmost islands occurs V. bahamensis Griseb., the most
primitive member of the species-group Bahamenses. The fundamental
difference between this group and the Longifoliae is again found in the in-
florescence. Here the cymes, after the flowering period, continue their
e ngation into the vegetative shoots of the next season. Not every cyme
necessarily elongates, but there are regularly 2-4 such branches at the apex
of each^ year's growth. Toward their base, paired scars mark the location
of former heads and bracteal leaves, while above them scars in spiral arrange-
ment indicate the former position of foliage leaves. All these Bahaman
species are therefore bushy, widely spreading, freely branched shrubs.
It is particularly worthy of note that they all have broad obtuse to retuse
leaves; that V. bahamensis, the species most nearly resembling V. albi-
caulis in leaf form, is the species of the southernmost islands, and that the
particular specimen in herbaria which most closely approximates the leaves
of the Porto Rican plant in size was collected on South Caicos Island, al-
most the extreme southeastern island of the group. V. arbuscula Less,
and V. obcordata Gleason occur farther to the northwest in the Bahamas.
V. complicate Griseb., of eastern Cuba, differs only in minor characters.
It is difficult to imagine a more striking instance of correlation between
structure and distribution than is presented by this group in its relation
to the Longifoliae.

The last three species-groups illustrate the following course of develop-
ment in the inflorescence:

1. The cyme is a special branch with reduced bracteal leaves and
elongate axis.

2. The cyme and leafy branches differ merely in position, and the
inflorescence is compact.

3. The cyme becomes the leafy branch at the conclusion of the
blooming season, and the inflorescence is compact and reduced.

The fourth species-group of the seven, the Racemosae, includes five
species of Hispaniola and Cuba. They are probably derived from V.
sericea of the Arborescentes, which is located near by in Porto Rico; V.
racemosa Delp. was considered by Ekman a variety of V. sericea, and, like
it, most of the species have leaves pubescent on the lower surface. In this
group the two upper primary cymes are short, with only 2-5 heads. Below
them, every leaf axil for a considerable distance down the stem produces
similar short lateral cymes. The whole makes an elongate subcylindric
inflorescence, quite different from the broad, spreading type of the preceding
groups. Secondary vegetative branches apparently do not exist. Within
the group, evolution is seen in the progressive reduction of the leaf surface,
of the cymes, and of the number of flowers in the head. While V. racemosa
of Hispaniola has lanceolate leaves, and cymes of 2-5 many-flowered heads,

192

AMERICAN JOURNAL OF BOTANY [Vol. 10,

the next three, one of Hispaniola and two of Cuba, have progressively nar-
rower and more revolute leaves and smaller and fewer heads. The group
culminates in a fifth species, V. corallophila Gleason of Cuba, with linear
leaves revolute to the midvein, n-flowered heads, and i-headed cymes
which appear as single axillary heads.

A fifth species-group, the Gnaphaliifoliae, also appears to be derived
from V. sericea or some species similar to it. There is the usual terminal
head, subtended by two primary cymes, and numerous other cymes arise
from the upper axils. They are usually flexuous, spreading or ascending,
and only occasionally branched. Such structures point unmistakably to
an origin within the Arborescentes, with which they share many structural
features and from which they are indeed rather weakly separated. The
three species are all Cuban. In this group axillary branches do not con-
tinue the vegetative growth, but the whole herbaceous stem dies and is
replaced by new growth from the perennial base.

All groups heretofore described in this general series have acute or
acuminate involucral scales. The sixth, the Acuminatae of Jamaica, have
obtuse scales, and also differ from the Arborescentes in their resinous-dotted,
non-papillose leaves and in their flattened and twisted pappus bristles.
Nevertheless, the simplest species of this group bears a general resemblance
to V. arborescens; has been placed adjacent to that species by Ekman, and
may have been derived from it. I fail completely to see any resemblance
between this group and the Fruticosae, as has been claimed by Ekman. V.
acuminata Less, is the common species of lower elevations. V. pluvial is
Gleason, the high-mountain derivative, presents an inflorescence of short,
much congested cymes, aggregated in subcapitate clusters.

The seventh and last group of this series, the Fruticosae, includes one
species of Hispaniola and eight of Cuba, particularly of the mountains of
the eastern part. Many of these are poorly known, some by a single col-
lection, and the number of species may easily be subject to increase or de-
crease as further collections are accumulated for study. From the inflores-
cence standpoint, they exhibit the simplest scorpioid cymes to be found in the
West Indies. They are mostly straggling vinelike plants with indeter-
minate growth. At some distance above the base the main axis ends in a
terminal head, while immediately beneath it a lateral branch, diverging
at a small angle, continues the sympodial axis and bears heads in the same
way. The heads are separated by internodes about equal to those of the
sterile section of the stem in length, they are subtended by bracteal leaves
which in almost all species are virtually indistinguishable from the cauline
in size and shape, and furthermore the cyme axis is frequently prolonged
after flowering into a leafy, sterile stem. In their leaf habit and papillose
pubescence they approach V. gnaphaliijolia; in other features they have no
near relatives in the West Indies and apparently none in South America.
Nevertheless, they offer no new structural features to separate them from

Apr., 1923] GLEASON VERNONIA IN NORTH AMERICA I93

the preceding species-groups. In the lack of sufficient material, the evo-
lution within the group can not now be discussed.

It is not necessary to presume that only one ancestral stock of Vernonia
migrated into the West Indies. The seven species-groups just described,
constituting probably one evolutionary stock, have spread over the whole
region and developed into many species. Other stocks may also have
immigrated from South America, been isolated in certain islands, and
developed endemic species. Certainly two species-groups now exist whose
relations can not be explained, and which should probably be considered
as entirely distinct evolutionary lines. These are the Buxifoliae and the
Sagraeanae.

The Buxifoliae include three species of the mountains of Hispaniola.
They are characterized by glabrous achenes, heads in. subcapitate clusters,
and an unusually large number of involucral scales, arranged in a beauti-
fully spiral imbrication.

The Sagraeanae include ten species, nine in Cuba and one in Hispaniola,
with an outlying variety in Dominica, characterized by large glabrous
achenes and usually by large many-flowered heads. Ekman would relate
the group to the Bolivian V. robusta Rusby, which differs in achenes, hispid
in the furrows, and in the number of setae of the pappus, about 25, instead
of 40-70; also to the Bolivian V.obtusata Less. (V. subacuminata Hieron.)
which has densely hirsute achenes. There is a superficial resemblance to
these Bolivian plants in their heavy, rugose, reticulately veined leaves, and
to V. robusta also in their large heads. On the ground that specialized
involucral scales, few-flowered heads, and rigid, coriaceous, or tomentose
leaves are characters which indicate an evolutionary advance, V. Sagraeana
DC. and V. viminalis Gleason may be regarded as the most primitive species,
and V. Wrightii Sch.-Bip. and V. purpurata Gleason as the most advanced.

We have now disposed of all leafy-bracted scorpioid species of North
America except two, V. yunquensis Gleason and V. segregata Gleason. These
Cuban species are poorly known and the former is represented in herbaria
only by the type specimen. While each of them exhibits certain points of
resemblance to other West Indian species, it is not possible to draw any
conclusions as to their relationships.

The general affinities of the 57 species of the leafy-bracted groups may
be summarized by the diagram (fig. 1), from which it may be seen that
without exception the more advanced groups lie progressively farther from
South America, that no group is common to the West Indies and the con-
tinent of North America, except as introduced, and that, with very few
exceptions, the more advanced species of each group also lie farther away
from the center of origin, either in horizontal or in altitudinal distance.

The 63 species in which the bracteal leaves are suppressed show certain
fundamental differences among themselves in the structure of the inflores-
cence, as a result of which five well-marked evolutionary stages may be

194

AMERICAN JOURNAL OF BOTANY

[Vol. 10,

distinguished. In the first of these the cymose structure is obvious, each
cyme is more or less elongate with secund heads, and branches occur at
such intervals that the scorpioid structure is not obscured. In the second,
secondary branches are developed at the bases of a great many heads, so
that three successive nodes without branches rarely occur. The result
is a large branching cluster which bears little superficial resemblance to the

CUBA

Yunquensis
<i ?

segregate
f ?

Sagraeanae
« ?

,Frutioosae

Baharaenees

BAHAMA ISLANDS \

HISPANIOLA

Buxifoliae
< ?

PORTO RICO

VIRGIN IDS

.albidaulla

LESSER ANTILLES

L > i\g i f o 1 i a e

CENTRAL- AMERICA

SOUTH AMERICA

Fig. i. Migration and evolution of the leafy-bracted Vernoniae of North America.
Solid lines show distribution by their location, migration by the direction of the arrow.
Dotted lines show probable connection by evolution between species-groups.

simple scorpioid cyme, although undoubtedly derived from it. In the
third, the heads are suppressed at those nodes where secondary branches
are developed. Since these appear at virtually every node, only terminal
heads are produced on the cymes, and the whole cluster appears to be
dichotomously branched. The fourth stage represents a much greater
step forward. Here the basal internodes of the inflorescence are much
shortened or almost suppressed, while the number of heads is greatly re-
duced. Since the terminal internodes retain a normal length, the whole
inflorescence appears subumbellate. While in the first three stages new
vegetative branches may arise from below the inflorescence, so that the
stem may live several years and reach a large size, in the fourth type, as
well as in the fifth, the appearance of the inflorescence prevents further growth

Apr., 1923]

GLEASON VERNONIA IN NORTH AMERICA

195

of the stem, and new vegetative parts appear only from the perennial base
of the plant. In the fifth and last stage, only a few terminal and subtermi-
nal heads retain the umbellate arrangement, while from the upper stem
axils similar clusters arise, producing a broad, flat-topped or hemispheric
cluster with all the heads peduncled. These five stages are shown diagram-
matically in figure 2.

Fig. 2. Modifications of the inflorescence in the bractless Vernoniae of North America
A. Stage 1, Vernonia scorpioides, lateral branch with two terminal cymes. B. Stage 2,
canescens, lateral portion of the inflorescence. C. Stage 3, havanensis, portion of a terminal
inflorescence. D. Stage 4, Karvinskiana, terminal inflorescence, with a few primary branches
omitted. E. Stage 5, texana, complete terminal inflorescence. All figures diagrammatic
as to position of branches or cymes, but accurate as to character of branching and pro-
portion.

Other evidences of evolution appear within the last two stages, leading
to the segregation of several species-groups.

The five stages are again well correlated with their geographical dis-
tribution. The first occurs in South America and is represented in our
region by a single species of St. Vincent, V. pallescens Gleason, and by two
which extend across the Isthmus of Panama into southern Central America,
V. scorpioides (Lam.) Pers. and V. brachiata Benth. The second includes

I96 AMERICAN JOURNAL OF BOTANY [Vol. 10,

several species of northwestern South America, some of which extend into
Central America also, and several others ranging as far north as southern
Mexico. Species of the third stage have crossed the narrow channel from
Yucatan and are now limited to western Cuba; the fourth is confined to
Mexico, and the fifth has four species in northern Mexico and 31 in the
United States and the Bahamas.

The species-group Stellares, representing the second stage, includes the
commonest species of the mountains of Colombia, Central America, and
southern Mexico. The Colombian species V. canescens H.B.K. and V.
mollis H.B.K. retain the primitive character of acuminate involucral
scales; the former also extends north into Mexico, and the latter is doubt-
fully admitted into the North American flora. Of the remaining five with
acute to rounded scales, V. patens H.B.K. occurs in both continents, while
the others are strictly North American. The most advanced morphologi-
cally is V. morelana Gleason, which alone does not occur south of Mexico.

The third stage includes the species-group Menthaefoliae, undoubtedly
closely related to the Stellares, but now isolated in western Cuba and the
Isle of Pines, except for a few specimens from central and eastern Cuba as
well, where they may have been recently introduced.

The fourth stage includes three well-marked species-groups, which
are nevertheless closely related. The Umbelliformes include 9 species,
mostly quite closely related and in some cases separated with difficulty.
The simplest species (and the commonest in herbaria) have small heads,
with short involucres and seldom more than 15 flowers; the more advanced
have larger heads and taller involucres. One of these, V. Conzattii Robin-
son, with its abruptly rounded and mucronate involucral scales, marks a
transition to the group Mexicanae, with three species in the higher moun-
tains of southern Mexico. Here the scales are extraordinarily specialized,
being 3-8 mm. wide, loosely spreading, at least at the tip, and prominently
reticulately veined. The two closely related species of the Alamanianae
have also large scales but lack the reticulate venation. The general evolu-
tionary tendency of the series is apparently toward large heads and special-
ized scales, and this is correlated geographically with an ascent to higher
levels in the mountains.

Passing now to the 35 species of the fifth stage of evolution, as shown
by the inflorescence, we find the most primitive members in the Texanae,
a group of four species, three in northern Mexico and one in Texas. Since
the inflorescence has already passed to the paniculate stage, equally char-
acteristic of the other species of the United States, evidence for the primitive
character of the group must be sought in other characters. The leaves in
all four species are more or less pitted beneath and the outer pappus bristles
are poorly differentiated from the inner in width, both of which features
occur also in the Umbelliformes. The most important primitive character,
however, lies in the involucre, and has not been mentioned before because

Apr., 1923] GLEASOX VERXOXIA IN XORTH AMERICA 197

it is shared by virtually all the groups hitherto discussed. Here the scales
are relatively few in number and poorly imbricated. The inner scales are
progressively more exposed than the middle and outer ones, contrasting
plainly with the numerous regularly imbricated scales of most other species
of the United States. Of the four species in the group, V. texana (A. Gray)
Small, is best known and occurs in Texas, Louisiana, and Arkansas.

From the area of the Texanae, migration accompanied by specific
evolution has proceeded in two directions, northward through the prairie
region and eastward along the coastal plain. In each direction one or more
of the primitive structures have been lost, until in Michigan and Massachu-
setts they have disappeared completely.

In northern Texas occurs the group Lindheimerianae of three species,
two of which are suspected to be hybrids. V. Lindheimeriana Gray &
Engelm., which is undoubtedly a good species, retains the primitive invo-
lucre and narrow outer bristles of the Texanae, from which it seems to be
derived, and differs chiefly in its tomentose leaves and scales.

The Fasciculatae, extending from Texas and New Mexico northward
and eastward to Manitoba and Ohio, retain the pitted leaves and present
to a still greater degree the narrow, undifferentiated outer bristles of the
pappus. They have lost the primitive involucre and have developed long
heads with numerous scales imbricated with great regularity. It is note-
worthy that the more southern species, as V. marginata (Torr.) Raf., still
show a tendency toward acumination of the scales, as in Texanae, which
character is lost to a large extent in V. fasciculata Michx., ranging from
Nebraska to Ohio, and completely in V. corymbosa Schw., distributed along
the Red River of the North in Minnesota, the Dakotas, and Manitoba,
where it marks the extreme northern limit of the genus.

The peculiar local species, V. Lettermanni Engelm., of Arkansas and
adjacent Oklahoma, bears a strong superficial resemblance to V. fasciculata
and is possibly an evolutionary development from it. It retains the gla-
brous leaves with pitted lower surfaces and the congested heads with closely
imbricated scales, like the latter species, but has broader, well differentiated
outer pappus bristles.

The group Interiores takes its name from V. interior Small, which is
undoubtedly the basic species. Common in central and northern Texas,
where it overlaps the range of the ancestral Texanae, it extends north to
Nebraska, and thence east to the Mississippi River. The involucre in this
species has only partially lost its primitive structure; the outer pappus is
narrow but nevertheless plainly differentiated; the leaves are broad, without
pits, and characterized by multilocular hairs forming a more or less tomen-
tose pubescence. V. Baldwini Torr. is an Ozarkian derivative, with broader
outer pappus bristles, and with the acuminate involucra! scales recurved
at the tip and pubescent on the inner face. The species is probably of
recent origin, and specimens from the overlapping ranges of V. Baldi^uu

I98 AMERICAN JOURNAL OF BOTANY [Vol. 10,

and V. interior are frequently intermediate in character. V. aborigina
Gleason, known so far only from the original collection in southeastern
Oklahoma, appears to be a giant form of V. Baldwini. It retains most of
the morphological characters of that species, but is much larger in all its
dimensions, with about twice as many flowers in each head. V. missurica
Raf., the last of the group, has the widest distribution of the four, ranging
from Texas, where it is not particularly common, north and east to Michi-
gan, and becoming exceedingly abundant in Iowa, Illinois, and Indiana.
It is characterized by larger, more compact inflorescence, fully differentiated
pappus, and regularly imbricated involucre. Many specimens from the
southern part of its range retain the sharply acute, relatively narrow invo-
lucral scales of V. interior, while those from farther east have fewer, broader,
and obtuse or apiculate scales. The species also occurs to a limited extent
and with slightly different structure along the Gulf Coast as far east as

Alabama.

The origin of two other western species is in doubt. V. Bolleana Sch.-
Bip., of northwestern Mexico, seems to bear no close relation to any other
known species. V. crinita Raf., of the Ozarkian region, is characterized by
filiform involucral scales, and may represent an extreme development from
the Interiores.

The eastward migration along the coastal plain from Texas led to the
present development of seventeen species. They are not easily divided
into distinct species-groups, a feature possibly indicative of recent immi-
gration and evolution. The most primitive group is the Angustifoliae,
ranging from Louisiana east to the Atlantic, thence north to the Carolinas
and south into Florida and the Bahamas. The group retains the primitive
involucre, narrow leaves, and low stature of the Texanae, and the type
species of that group was originally described as a variety of V. angustifolia
Michx. Some of the species have an inflorescence approaching umbelliform
and rather suggestive of V. liatroides DC. or other species of northern
Mexico. V. angustifolia has the widest distribution, almost coextensive
with that of the group. The other four, each of restricted distribution and
lacking the acuminate scales of the simpler species, seem to represent recent
evolutionary developments. Of these, V. Blodgettii Small, in southern
Florida, marks the re-entrance of the group into the tropics, and leads to the
closely related V. insularis Gleason of the nearby Bahamas.

The group Pulchellae is obviously closely related to the Angustifoliae, as
shown by narrow leaves and general vegetative habit, but differs in the
absence of resin glands on the achenes and in the prolongation of the in-
volucral scales into filiform appendages. The three species, V. pulchella
Small, V. recurva Gleason, and V. scaberrima Nutt., are all of limited dis-
tribution in the coastal plain of Georgia and the Carolinas.

The species-group Glaucae lies generally to the north of the Angusti-
foliae and has probably been derived from it. Here the heads are larger,

Apr., 1923] GLEASON VERNONIA IN NORTH AMERICA 199

the pappus is tawny or almost white, and the involucral scales are long-
acuminate or almost filiform. The leaves are large in proportion to the
height of the stem, and the greatest expanse of foliar surface is toward the
base of the stem. While this feature is apparent in V. glauca (L.) Willd.,
an Alleghenian species ranging northward to Pennsylvania, it is still further
developed in V. acaiilis (Walt.) Gleason and V. georgiana Bartlett, two
coastal plain species with distinctly basal leaves.

Two other species with prolonged filiform scales constitute the group
Noveboracenses. V. noveboracensis (L.) Michx. has attained a wide dis-
tribution over the Piedmont region of the eastern states from Mississippi
to Massachusetts, occasionally invading the coastal plain also. There it has
given rise to a localized species, V. Harperi Gleason, characterized by larger
heads with more numerous flowers.

V. gigantea (Walt.) Britton is closely related to V. concinna Gleason,
of the Angustifoliae, and like that species is confined to the southeastern
portion of the coastal plain.

The last species-group of the southeastern states is the Altissimae. V.
ovalijolia T. & G. is a variable species of the southeastern coastal plain, and
appears to be the most primitive species of the group in their evolution
from the Angustifoliae. Although some of its variants approach V. con-
cinna, it is generally distinguished by the broader, regularly imbricated
involucre and the broad leaves. V. flaccidifolia Small is a well-marked
species of the southern Appalachian region. V. altissima Nutt., the last
species of the group, has a wide distribution from Georgia and Alabama
north and west to New York and Missouri. Typical forms of the species
avoid the coastal plain and are characteristic of the woodlands of the central
states, but the variety laxa Gleason occurs along the Gulf Coast. In its
western extension V. altissima comes in contact with several species of the
western migration route, and many intermediate forms occur which are
probably to be considered as hybrids.

Considering the 35 species with paniculate inflorescence as a whole, we
see that the species with primitive involucre invariably lie far to the south
or southwest, and that those with the broadest and most obtuse scales, as
well as those with the most filiform scales, lie always well to the north or
northeast. It is also worthy of note that only seven of the 35 have attained
a wide distribution, while the other 28 occupy small or localized ranges.
These seven are V. fasciculata Michx., V. interior Small, V.missurica Raf.,
V. angustijolia Michx., V. glauca (L.) Willd., V. noveboracensis (L.) Michx.,
and V. altissima Nutt., representing six species-groups. This fact alone
may indicate the comparatively recent immigration and incomplete evo-
lution of the genus in the northern portion of its range.

The general relations of the 63 species of the genus in which the bracteal
leaves are suppressed is exhibited in a diagram (fig. 3).

In the evolution of the genus in North America, no general plan or

200

AMERICAN JOURNAL OF BOTAXV

[Voi. 10,

tendency is apparent, except in the one feature of inflorescence. In every
case it is found that species structurally aberrant from the general type

CAM ADA

Prairie
Region

Fasc iculatae

Central

— ^ Forests

Interiores

Altissimae

— ^ Ozark Mta.
Lottermanni ^

Alleghenian Region

Coastal Plain

Advanced Involuc re

Llndheimer

ianae

Primitive Involucjre
Angustifollac \

UNITED

STATES

\Gigantea

MEXICO

Bolleana

t

Inflorescence Stage ^

CUBA

Menthaefoliae

CENTRAL AMERICA

Inflorescence Stage 1

St .VINCENT

pallascens

SOUTH AMERICA I Soorpoldes

FlG. 3. .Migration and evolution of the bractless \'ernoniae of Xorth America. Solid
lines show distribution by their location, migration by the direction of the arrow. Dotted
lines show probable connection between species-group.

occupy outlying ranges or peculiar habitats. The evolution in structure
may be summarized as follows:

1. From elongate cymes to short, compact, freely branching cymes
or capitate clusters.

2. From small, acute or acuminate involucral scales to broad, blunt,
or veiny scales, or to narrow, prolonged, or filiform scales.

3. From a medium number (13-21) of flowers in each head to a large
number (55-89, or even more), or to a reduced number (as low as 5).

4. From lanceolate, acuminate leaves to linear, 1 -nerved, or revolute
leaves, or to blunt, broad, rigid or coriaceous leaves.

In the leafy-bracted forms, comprising 57 species, there is a general

Apr., 1923] GLEASOX VERNONIA IN NORTH AMERICA 201

tendency toward the congestion of the inflorescence by repeated branching or
toward its reduction by shortening the cymes. The former is most apparent
in the species-group Arborescentes, the latter in the Racemosae and Acu-
minatae. There is a great reduction in the number of flowers in the outlying
members of the Acuminatae, Sagraeanae, and Racemosae. Specialization
of the involucre by the development of broad, blunt scales occurs in the
Schiedeanae, and of narrow, prolonged scales in the Sagraeanae, while the
Buxifoliae have increased the number and the regularity of imbrication
of the scales. Leaves have shown a tendency to become broad and blunt
from the Arborescentes through the Longifoliae and into the Bahamenses,
or narrow, one-nerved, and revolute from the Arborescentes into the Race-
mosae. Montane species have been developed in the Acuminatae and the
Buxifoliae, and in both cases are characterized by crowded, few-flowered
heads and by small and broad leaves.

Among the 63 species with bractless cymes there is less diversification
in structure, except in the inflorescence, which has already been discussed.
In the Stellares there is a gradual progression from narrow, acuminate
involucral scales to short and blunt ones. A similar tendency occurs among
the Interiores and Altissimae, and reaches a climax in the series from the
Umbelliformes to the Alamanianae and Mexicanae, with their highly spe-
cialized, broad or veiny scales. On the other hand, there is a notable ten-
dency toward prolongation of the scales in V. crinita and in the Glaucae,
Pulchellae, and Xoveboracenses. Excepting the Stellares, all these groups
show likewise a tendency to larger heads, reaching a maximum in the Mexi-
canae, Alamanianae, and V. crinila. Forms with unusually small heads
rarely occur, and are most characteristic of the single species^ V. gigantea.
Two groups only have developed montane forms, the Alamanianae and Mex-
icanae, and in their unusually large, many-flowered heads differ remarkably
from the montane forms of the West Indies.

Neither is there any correlation between structure and habitat. The
variation between the montane species of the West Indies and Mexico has
already been mentioned. The relatively arid conditions of the Bahamas
are reflected in the thick, firm leaves of the Bahamenses and V. insularis,
but those of the former, with broad-leaved ancestors, are broad and blunt,
while the latter, originating from the Angustifoliae of the Gulf States,
preserves the linear leaves. The Racemosae, of arid situations in Cuba and
Hispaniola, and V. texana, likewise a xerophyte, have narrow leaves, but
the hydrophytic V. fasciculata has narrow leaves also, while the xerophytic
V. Baldwini has broad leaves. The xerophytic Bahamenses have assumed
the form of bushy shrubs, while V. texana has remained an herb, although
growing in a region where the shrub form is common.

Three processes seem to have been concerned in the general history of the
genus, by which it has reached its present distribution and differentiation.
Physiological evolution, scarcely indicated by structure, has enabled the

202 AMERICAN JOURNAL OF BOTANY

genus to migrate into new environments beyond its original home; migration
has brought it to its present distribution and is doubtless still continuing:
structural evolution, favored by geographic isolation, has differentiated
the present species, but is very little correlated with their physiology or
ecology, although proceeding simultaneously with their migration. A
single structural tendency appears to be general and possibly orthogenetic,
that of the shortening and branching of the cymes.

Summary

i. Three sections of the genus are represented in North America by a
single species each; one section is represented by 120 species.

2. In this section the chief differentiation of groups rests on the structure
of the inflorescence; minor differentiation of species-groups is based on the
achenes, the involucre, the pappus, and the character of the pubescence.

3. Of the two subsections, one is chiefly Antillean, the other continental,
while both are developed in continental South America.

4. Characters which are held to represent primitive conditions in one
group may indicate advanced evolution in another, and such characters
have no apparent correlation with environment.

5. In every case, those groups which appear to be the simplest in mor-
phological structure occur to the south, while the more complex groups
appear progressively farther to the north. In most groups the same state-
ment holds for the individual species.

6. The geographical arrangement of the species-groups and species
follows well-known migration routes and supports the conclusion that
evolution and migration have proceeded together.

New York Botanical Garden

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN

No. 244

NOTES ON FABACEAE-I

PER AXEL RYDBERG

NEW YORK

1923

Reprinted without change of paging from Bulletin of the Torrey Botanical Club
50: 177-187. May 25, 1923.

LIBRARY
NEVV YORK
BOTANICAL

Notes on Fabaceae — I

Per Axel Rydberg
HOMALOBUS Nutt.

For several years I have worked on genera of the F~abaceae
for the North American Flora. Two parts containing the tribe
Psoraleae have already been published, and the manuscript for
the Indigoferae and a large part of the Galegeae is ready. My
work has taken me to Astragalus and related genera. I have
heard a rumor that M. E. Jones has been working on an illustrat-
ed revision of Astragalus, and I wish that this had been published
so that I could have availed myself of it in my studies. It
would perhaps have helped me to avoid some of the errors which
are so likely to creep in.

There is little room for critical notes in the Flora, and yet
notes of this character are necessary for a proper understanding
of my disposition of genera and for a just appreciation of the
amount of work that I have done on the group. It is my
intention, therefore, to publish a series of preliminary notes
in order to thresh out any unavoidable criticism, which might
come from certain quarters, before my final monograph is
published. If I should criticise any fellow botanist in this
undertaking I do it in order to bring out the facts.

Just before the present paper was completed there appeared
an article by J. F. Macbride, entitled: A revision of Astragalus,
subgenus Homalobus, in the Rocky Mountains.* Now it may
perhaps seem superfluous for me to go into details and publish
another paper on the same subject so soon. So far as generic
and specific limits are concerned, 1 do not think that either
Macbride or Jones can ever agree with me. It surprises me,
however, that so many of my species, even some that had been
sneered at by Jones, have been accepted by Macbride. That
he should have reduced some of my species to varieties I might
have expected, since I do not see where to draw lines between
species and varieties and prefer a binomial name to a trinomial

* Contr. Gray Herb. II. 65: 28-39. 1922.

177

178 Rydberg: Notes on Fabaceae — I

for the sake of convenience. He has not treated me more
harshly in this respect than he has others; for he has reduced
to varieties or synonyms the species of other botanists, such as
Astragalus lingulatus Sheld., A. exiliflorus A. Nels., Homolobus
caespitosus Nutt., H. canescens Nutt., H. brachycarpus Nutt.,
A. amphidoxus Blankinship, H. dispar Nutt., H. juhceus Nutt.,
A. junciformis A. Nels., A. diversif alius Gray, A. strigosns
Coult. & Fish., A. Palliseri Gray, and A. acerbus Sheld. In
this way he has reduced half of Nuttall's species, all of A.
Nelson's and all but one of Sheldon's. Is such a reduction
warranted?

I once criticised Brand for his work on the Polemoniaceae,
because he did not try to find what the types really represented.
May I not now have the right to criticise Macbride for the
same laxities? In his paper on Homalobus he reduced my H.
microcarpus, H. proximus, and H. Salidae, although he admitted
that he had not seen the types; he reduced also my H. dementis,
II. stipitalus and H. uniflorus, without including the types
among the specimens cited, thus implying that he did not have
them before him. The same can be said about Sheldon's
Astragalus lingulatus and A. acerbus. Except in the case of A.
hngulatus the types of all these species are readily accessible to
anyone who may wish to see them.

As I have not found it necessary to change essentially my
grouping of the species from that published in my Flora of the
Rocky Mountains, I shall keep the same groups as there. I
shall also omit the keys except in cases where the species are
from the Pacific coast. Keys would be superfluous, since my
own are found in my Flora of the Rocky Mountains and since
Macbride has given a good one in the paper cited above.

A. Caespitosi

1 . Homalobus caespitosus Nutt. This is the most common
and best known of the group, and its nomenclature has not
been much confused. Occasionally the leaves are two- or three-
foliolate instead of simple. H. canescens Nutt. is only a more
silvery form of the species. Gray, who merged Homolobus
in Astragalus, recognized the identity of the two, retaining
the specific name caespitosus, although there was an earlier

Kydberg: Notes on Fabaceae — I 179

Astragalus caespitosus Pallas. On this account Sheldon sub-
stituted the name A. spatulatus. In the meantime O. Kuntze
had transferred all species of Astragalus to Tragacantha. M. E.
Jones* has made A. caespitosus a variety of A. simplicifolius
(Nutt.) A. Gray, probably because he did not understand the
latter species, which is discussed below. Macbride has followed
Jones.

2. Homalobus brachycarpus Nutt. Gray, Watson, Shel-
don, Jones, and Macbride have regarded this as a synonym of
H. caespitosus. Nuttall distinguished it by its shorter, broader
and more abruptly acute pod and by its smaller size. I take
it to be the same as the plant described and figured by Watson
in King's Report under the name Astragalus simplicifolius Gray,
although the leaves of Nuttall's specimens are narrower than
those figured by Watson. Evidently Jones has followed Watson
in his interpretation, for he has shown the close relationship
between what he called A . simplicifolius and A . caespitosus, and
afterwards went so far as to make the latter a variety of A.
simplicifolius. The plant described and figured by Watson,
however, is not the same as Phaca simplicifolia Nutt.; if it is
not H. brachycarpus it is without, a name. The duplicate of
Nuttall's type of H. brachycarpus in the Torrey Herbarium
strongly resembles Watson's plant but is somewhat taller and
has narrower leaves. That in the Gray Herbarium is different
and is exactly like the type of H. canescens; it does not agree
with the description of H. brachycarpus. The following speci-
mens belong to the latter: —

Wyoming: hills on the Platte, Nuttall; northwestern Wyo-
ming, Jones 68; Big Horn Mountains, Tweedy 136; Laramie
Plains, 1897, Osterhout; Fort Steele, Tweedy 4193; Pine Bluffs,
A. Nelson 2877; Natrona, A. Nelson 142; Laramie, A. Nelson
31, 7289; Fort Bridger, Lauderdale.

Utah: Unitah Mountains, Goodding 1285; Wasatch Moun-
tains, Watson 289.

3. Homalobus simplicifolius (Nutt.) Rydb. There are
several points in the original description of Nuttall's Phaca
simplicifolia that do not fit Watson 's plant, such as the following :
"scapes 1 -2-rlowered, scarcely exserted;" "legumes glabrous,

* Proc. Calif. Acad. II. 5: 647. 1895.

180 Rydberg: Notes on Fabaceae — I

coriaceous, sessile, somewhat triquetrous, pointed, scarcely
longer than the calyx;" "flowers ochroleucous." In Watson's
plant the scape is three- to seven flowered, usually as long as, or
longer than, the leaves; the pod is decidedly flat; and the corolla
is purple. All the characters given by Nuttall fit A. exiliflcrus
A. Nelson, but I do not think that the two are identical. There
is, however, a plant of southern Wyoming and northern Colorado
that agrees fully with Nuttall 's type. This resembles Nelson's
species but is smaller and has shorter leaves. In this respect
it approaches H. br achy car pus, being distinguished from that
species by its narrower and more pungent leaves and by its
shorter, unmottled, and rather turgid pod, thicker nearer the
lower suture than the upper, hence Nuttall 's expression, "some-
what triquetrous.'1 Nuttall 's specimen in the Gray Herbarium
has leaves only, so that Watson had no opportunity to compare
the pods. The following specimens belong to this species: —

Wyoming: Rocky Mountains, Nuttall; Fremont; Ham's
Fork, C. C. Curtis ; Green River, Jones; Hanna, Payson & Payson
1694.

Colorado: Canon City, Brandegee.

4. Homalobus exilifolius (A. Nels.) Rydb. See notes
under tne preceding species. The following specimens belong
here : —

Wyoming: Freezeout Hill, Elias Nelson 4493 ; Fort Steele,
Tweedy 4194.

5. Homalobus lingulatus (Sheld.) Rydb. This is known
only from the original collection, and its description is much
like that of the preceding species, the only essential difference
being the glabrous calyx. The mature fruit is unknown. When
the plant is better understood it may prove to be the same as
H. exilifolius, in which case the species should bear Sheldon's
specific name, which antedates Nelson's by five years.

6. Homalobus uniflorus Rydb. It is possible that this
may represent an aberrant form of H. br achy car pus, i.e. Astrag-
alus simplicifolius, as understood by Watson. * The pod, how-
ever, is different, being lanceolate in outline, more tapering at
the apex, 6-7 mm. long and 2-3 mm. wide. It is, therefore,

Rydberg: Notes on Fabaceae — I 181

more like that of the true //. simplicifolius* The following
specimens may be cited : —

Wyoming: Evanston, A. Nelson 2QJI (flower); Carter, M.
E. Jones (fruit).

B. Tenelli

7. Homalobus tenellus (Pursh) Britton. The description
of Astragalus tenellus Pursh, dating from 18 14, was drawn
(according to Pursh himself) from flowering specimens of the
present species and the fruit of some other Astragalus, both of
which had been collected by Lewis. When Pursh found that
his description had been based on a mixture, he redescribed the
flowering specimens, associating them with a fruiting plant
collected by Bradbury on the Upper Missouri and giving the
species the new name Ervum multiflorum , on account of its
Ervum-Yike pods. A fragment of Bradbury 's plant in the Torrey
Herbarium is clearly identical with the flowering specimens
collected by Lewis: hence there is no doubt that Astragalus
tenellus (so far as the flowering specimens are concerned) and
Ervum multiflorum are the same.

8. Homalobus dispar Nutt. Nuttall first described this
species in 1818 under the name Orobus dispar, from specimens
that he collected at Fort Mandan. To those who consider H.
dispar and H. tenellus identical, this note will be of no interest.
It is here presented, however, for the benefit of those who are
willing to admit their distinctness as species or varieties. Since
Nuttall and Bradbury were members of the same expedition
and often collected together, one might perhaps assume that
Orobus dispar and Ervum multiflorum were the same. They are
not. Nuttall 's specimen in the Torrey Herbarium agrees with
his description in having broader leaflets than H. tenellus and
broader short-stipitate pods. Of Phaca nigrescens Hook.,
which was described in 1830 from specimens collected by
Richardson, there is likewise a duplicate in the Torrey Herbar-
ium. This belongs to H. dispar, while a specimen collected by
Nuttall on the Platte River, which was named Homalobus
nigrescens by him and Astragalus multiflorus by Hooker, is
like A . tenellus Pursh. Macbride states correctly that " Nuttall's

* See also Bull. Torrey Club 34: 49. 1917.

182

Rydberg: Notes on Fabaceae — I

specimen in the Gray Herbarium has narrowly linear leaflets
and the stipe of even very immature pods is quite as long as the
calyx." These statements do fit Nuttall's specimens of "H.
nigrescens" in the Torrey Herbarium but not his description of
H. dispar. The specimens in question were collected on the
Missouri, but I do not know the date; there is no evidence that
they came from Fort Mandan.

9. Homalobus stipitatus Rydb. Although H. tenellus is
much more common than H. dispar, the two species have nearly
the same range; H. stipitatus is confined to the northeastern
part of this range. All the specimens in the New York Botanical
Garden Herbarium are cited under the original description and
need not be repeated here.

10. Homalobus strigulosus Rydb. Macbride has reduced
this to Astragalus tenellus forma strigulosus. An additional
specimen is here recorded: —

Utah: Jugtown, Jcnes 5400.

11. Homalobus Standleyi sp. nov.

Astragalus tenellus var. dementis Macbride, Contr. Gray Herb.

11. 65: 35 (in part). 1922.

A perennial, with a caespitose caudex; stems 2-3 dm. high,
sparingly strigose; stipules deltoid, about 2 mm. long, connate;
leaves 3-5 cm. long; leaflets nine to fifteen, elliptic or oblong,
5-10 mm. long, 2-3 mm. wide, glabrous above, strigose beneath;
racemes lax, 4-7 cm. long, including the short peduncle; bracts
lance-subulate, 1-2 mm. long; calyx strigose, the tube 2 mm.
long, the teeth subulate, fully 1 mm. long; corolla white or
tinged with purple, the keel purple-tipped; banner 6 mm. long,
obovate, slightly retuse; wings 5 mm. long; keel-petals 5 mm.
long, with a rounded apex; pod about 1 cm. long, 3 mm. broad,
strigose, elliptic-oblong, acute at each end, short-stipitate, the
stipe shorter than the calyx-tube.

Type collected at Ponchuelo Creek, New Mexico, July 4,
1908, Standley 4181 (herbarium of the New York Botanical
Garden).

This is closely related to H. strigulosus but differs in the
white or purple-tinged (not ochroleucous) corolla, with a de-
cidedly purple-tipped keel, and in the narrower pods with much
shorter stipes. Macbride thought that this belonged to H.
dementis Rydb. and, on account of the similarity of the pod

Rydberg: Notes on Fabaceae — I 183

to that of //. strigulosus, made H. dementis a variety of Astrag-
alus tenellus. He had evidently not seen the type or a duplicate
of the type of H. dementis, but had seen specimens of Baker
489 from Marshall Pass, Colorado, which evidently belong to
H. dementis. In our specimen of this number the pods are
very immature, but a closer examination would have shown
Macbride that Baker's plant and Standley 4181 were not the
same. In the type of H. dementis the pods are half broader
than in H. Standleyi, distinctly black-hairy and tapering at
both ends, but the stipe if any is only a fraction of a millimeter
long, i. e. the pod is subsessile; the corolla also is much larger.
Baker 489 also has black-hairy pods.

12. Homalobus acerbus (Sheld.) Rydb. Macbride re-
duced this species to Astragalus tenellus forma acerbus, but had
evidently seen no specimens of it. M. E. Jones has referred
the type of A. acerbus in the Columbia Herbarium to A. ivinga-
tanus and expressed his views in print.* as follows: "Astragalus
acerbus seems to be identical with A. Dodgeanus Jones, and the
latter is not surely separable from A. wingatensis \wingatanus\."
In H. tenellus and its closer relatives the corolla is ochroleucous
(not a very important character); the racemes, together with
the peduncles, seldom overtop the leaves to any extent; the
pods are decidedly veiny; and the plants are inclined to blacken
in drying. In H. acerbus, on the other hand, the corolla is
white or purple-tinged; the racemes are twice to four times as
long as the leaves; the venation of the pods is indistinct; and
the plants show no inclination to blacken. The following
specimens belong here: —

Colorado: Glenwood Springs, 1893, Saunders; DeBeque,
Osterhout 4282; Grand Junction, 1895, M. E. Jones.

13. Homalobus Dodgeanus (M. E. Jones) Rydb. At
present I have no authentic specimens of this species before me
but, according to my memory, it is very closely related to H.
acerbus, and Jones, as shown above, regards the two as identical.
In fact Jones's own specimens, determined by him as Astragalus
Dodgeanus, from Grand Junction, collected May 22, 1895,
represent H. acerbus and are here listed under that species.

* Proc. Calif. Acad. II. 5: 636. 1895.

184 Rydberg: Notes on Fabaceae — I

The type came from Thompson 's Springs, Utah. The decidedly
oblique pods and black-hairy calyx of //. Dodgeanus would
make the species distinct enough, if these features should prove
constant; otherwise the two species should perhaps be united.
In this case Jones's specific name should be retained.

14. Homalobus Clementis Rydb. See my notes under
H. Standleyi above. Notwithstanding Macbride's remarks and
his reduction of this species to a variety of Astragalus tenellus,
I still consider it nearest to //. wingatanus and worthy of recogni-
tion. In both //. wingatanus and the allied H. lancearius the
pod is acute at each end, as in H. Clementis, and a minute stipe
is sometimes produced. In //. lancearius, in fact, this stipe
may be fully as long as in H. Clementis. The following speci-
mens may be cited: —

Colorado: Saugre de Cristo, Clements; Marshall Pass,
Baker 489; without locality, C. H. Hall.

15. Homalobus wingatanus (S. Wats.) Rydb. To the
specimens listed by Macbride the following may be added: —

Colorado: Mancus, Eastwood.

Utah: Monticello, Rydberg & Garrett 915 1, 9228.

16. Homalobus lancearius (A. Gray) Rydb. This species
is still known with certainty only from the type collection:
Beaver Dam, Arizona, 1877, Palmer 114. A flowering specimen
from Allen Canyon in southeastern Utah, Rydberg & Garrett
9309, resembles the type very much but is doubtful in the
absence of fruit. In the original description of Astragalus
lancearius, Gray called attention to the fact that the species
strongly resembled A.filipes in habit and in flowers, but that the
pod was perfectly sessile. On this account it would have to
be placed in the Tenelli group, probably near H. wingatanus
and H. acerbus. Jones* claimed that the species was the same
as Astragalus recurvus Greene, which on account of its completely
two-celled fruit is a species of Hamosa. Later hef retracted his
statement, adding that A. lancearius was "a good species but
near A. Rusbyi Greene." The first part of his conclusion is
correct but the latter part is far from the mark.

17. Homalobus episcopus (S. Wats.) Rydb. Under Astrag-
alus Coltoni Jones, Macbride makes the following statement:

* Proc. Calif. Acad. II. 5: 636. 1895.
t Contr. West. Bot. 8: 12. 1898.

Rydberg: Notes on Fabaceae — I 185

"Rydberg's reduction ... of this species to A. episcopus,
a plant with strictly sessile pods, is not understandable." The
explanation of this reduction is as follows. In the herbarium
of Columbia University there is a sheet labelled: "Coll. United
States Department of Agriculture. Astragalus Episcopus, S.
Watson, n. sp. Locality Utah. Collector, Capt. Bishop, 1872."
The specimens on the sheet are more or less broken up into
pieces but the habit is very similar to that of A. Coltoni, as
represented by our specimens collected at Castle Gate by Jones.
The plants are in flower, but the color of the corolla can not
be determined. There is also a broken immature pod, showing
a stipe nearly 1 cm. long. On the same sheet is a pocket con-
taining two pods; one of these is mature and of exactly the same
shape as the pods of A. Coltoni (except that the stipe is broken
off), while the other is very young but shows a stipe 7 mm. long.
On the pocket is written in what I take to be Watson's hand,
"Astragalus episcopus, Wats. S. Utah. Capt. Bishop, 1873.
pods. " These specimens in the Columbia University Herbarium
clearly belong to the same species as A. Coltoni. Maebride's
statement that the pod of A. episcopus is "sessile" does not
agree either with the type specimen or with Watson 's description,
which reads: "pod narrowly oblanceolate, compressed, \1/^
inches long, il/2 lines broad, acute at each end, very shortly
stipitate, reflexed." Watson's description, however, agrees
in every respect with the Columbia University specimens,
except the words "very shortly;" for the stipe is rather long.
As Watson also describes the flowers as being "purple or yellow-
ish," he might have had a mixture, and if the specimens in the
Columbia Herbarium were included in his conception of the
species, this was certainly the case. The specimen in the Gray
Herbarium is also fragmentary but shows a plant related to H.
lancearius ; its pod, however, is larger, 3 cm. long, and 5-6 mm.
broad, tapering at both ends and very shortly stipitate, and the
corolla is larger, being fully 12 mm. long.

18. Homalobus vexilliflexus (Sheld.) Rydb. This was
originally described as Astragalus pauciflorus Hook.; since, how-
ever, there is an older A. pauciflorus Pallas., Hooker's specific
name is not tenable in Astragalus and hence not in Homalobus.
As the plant is well known no specimens need be cited.

1 86 Rydberg: Notes on Fabaceae — I

//. miser (Dougl.) Rydb. should be withdrawn, at least
provisionally. My conception of it was partly based on
Gray's interpretation of the species, when he included under it
Lyall 7, a specimen in flower only. Jones and Piper have referred
this specimen to Astragalus microcystis A. Gray, a species of
Phaca. I do not feel certain that their identification is correct
or that Lyall's specimen really represents Astragalus miser
Dougl. Since Gray placed the species among the Homalobi
and since I at that time identified certain slender specimens
in fruit with Lyall's plant, I admitted H. miser as valid. I now
think that the true A. miser is still to be identified.

C. BOURGOVIANI

This group bears the name of Debiles in my Flora of the
Rocky Mountains. It is desirable to change the name, however,
since Astragalus debilis (Nutt.) A. Gray, as will be shown below,
is an imperfectly known species.

19. Homalobus Bourgovii (A. Gray) Rydb. The habit
and the black-hairy calyx and pod make this plant resemble a
good deal Astragalus alpinus L. and Phaca elegans Hook., both
now belonging to other genera. Except for the somewhat
ridged valves of the pod the plant is a true Homalobus. In
the alpine regions of the Canadian Rockies there grows a low,
almost acaulescent form, with minute leaflets. This looks very
unlike the ordinary form, but all grades between the two ex-
tremes are found.

20. Homalobus retusus sp. nov.
A perennial, with a slender caespitose rootstock; stems
decumbent or ascending, 1-2 dm. long; stipules lanceolate or
the lower ovate or triangular, connate, 1-2 mm. long; leaves
5-8 cm. long; leaflets nine to eleven, oblong to oval, often retuse
or emarginate, 5-12 mm. long, 2-4 mm. wide, glabrous above,
sparingly strigose beneath; peduncles 6-10 cm. long, slender;
racemes short, five- to ten-flowered, bracts lanceolate, 2 mm.
long; calyx black-hairy, the tube 2 mm. long, the teeth subulate,
1 mm. long; corolla purplish; banner 8-9 mm. long, obovate;
wings 7-8 mm. long, the blade obliquely oblanceolate, with a
long basal auricle; keel-petals slightly shorter, with a rounded
apex; pod black-hairy, sessile, 8 mm. long, 3.5 mm. wide.

Type collected at Fort Selkirk, Yukon, July 20, 1899,
Tarleton 128, a & b (herbarium of the New York Botanical
Garden). The following specimens in flower may belong here: —

Rydberg: Notes on Fabaceae — I 187

Yukon: Ranch Valley, Gorman 1082; White Horse, Macoun
58429.

When I transferred Astragalus debilis {Phaca debilis Nutt.)
to Homalobus, I thought that the specimens above cited be-
longed to it and drew the characters of the pod from them. I
have since found that they represent a species close to H.
Bourgovii. M. E. Jones identifies Phaca debilis, with Astragalus
Bodini Sheld., which is a true Phaca related to P. microcystis.
A duplicate of the type of P. debilis, which is in the Torrey
Herbarium, somewhat resembles P. Bodini in habit, but the
calyx is not black-hairy and the leaflets are emarginate. Mac-
bride admits Astragalus debilis as a member of the subgenus
Homalobus, citing the following two species from Colorado:
North Park, Osterhout 3, and Upper Platte, Geyer j. The
latter I have not seen but Osterhout 's plant clearly belongs
to P. Bodini. Osterhout had labelled it "Astragalus debilis,"
apparently following Jones's interpretation. Until the fruit
is discovered, I fear that P. debilis can not be fully interpreted.

Astragalus grallator S. Wats, is regarded by M. E. Jones as
an abnormal form of A. Haydenianus, and I am inclined to
agree to this view. In this case the species should be removed
from Homalobus. The specimen of A. grallator in the Torrey
Herbarium, so named by Watson, is a similar abnormal form
of some other species.

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN

Xo. 245

THE TAXONOMIC AND MORPHOLOGIC

STATUS OF
OPHIOGLOSSUM ALLENI LESQUEREUX

ARTHUR HOLLICK

NEW YORK

1923

Keprinted, without change of paging, from Bulletin op the Torrey Botanical Clod
50: 207-213. pi, 10-12. 1923.

LIBRARY

NEW YORK

BOTANICAL

GARDEN

The taxonomic and morphologic status of
Ophioglossum Alleni Lesquereux
Arthur Hollick
(with plates 10-12)

About fifty years ago Lesquereux described, without any
accompanying illustration, an imperfect fossil plant specimen
from the Miocene Tertiary shales of Florissant, Colorado, under
the name Ophioglossum Alleni* The description was as follows:

Leaf, elliptical, narrowed by a curve to the acute base; shorter and broader
than in O. vulgatum L., of our time, with the same areolation.

The leaf is about 3 cent, long (point broken), a little more than 2 cent,
broad, marked in the middle by the remnant of a fruiting pedicel . . .

Subsequently the same author redescribed and figured the
specimen under the name Salvinia Alleni,\ and remarked:
"By its form, its areolation, its size, all its characters, indeed,
it is remarkably similar to Salvinia reticulata Heer. "f His
amended description was as follows:

Leaves oval, rounded in narrowing to the base; lateral veins, none
visible, areolae large, irregularly square or equilateral, inordinately distrib-
uted.

Leaf about three and a half centimeters long, twenty-two millimeters
broad, of a thin substance, with a thick middle nerve and irregularly quadrate
meshes, formed of very distinct black nervilles, the primary ones more or
less in right angle to the middle nerve, with oblique, generally parallel vein-
lets between them ....

Later he listed and figured twro other, more perfect specimens, §
with the following brief comment:

The species is common and has been obtained in large well-preserved
specimens by the different collectors. The leaves are merely variable in size,
obtuse or slightly emarginate at the apex, topped by the point of the ex-
current nerve.

* U. S. Geol. Survey Terr., Sixth Ann. Rept. 1872: 371. 1873.

f U. S. Geol. Survey Terr. Rept. vol. 7 (The Tertiary Flora): 65. pi. 5, /. //.
1878. Reproduced on Plate io, fig. i.

% Flora Tertiaria Helvetiae 3: 156. pi. 145, f. 16. 1859. Reproduced on
Plate io, fig. 4.

\ U. S. Geol. Survey Terr. Rept. vol. 8 (The Cretaceous and Tertiary
Floras): 136. pi. 21, f. 10, 11. 1883. Reproduced on Plate io, figs. 2, 3.

207

208 Hollick: Status of Ophioglossum Alleni L.

In 1894 I had occasion to examine the figures and kiescriptions
of this and other fossil species described under, or referred to,
the genus Salvinia, and in a paper on the subject I referred
Lesquereux' species to the genus Tmesipteris* and incidentally
compared it with Salvinia reticulata (Ettingshausen) Heer,
above mentioned, which Heer had suggested was, apparently,
specifically identical with a fragmentary leaf or pod originally
described by Ettingshausen under the genus Dalbergia;\ but
the resemblance to a pod did not impress me at the time.

In 1913, Professor T. D. A. Cockerell,* in a discussion of the
fauna and flora of the shales in which the species was originally
discovered and from which the specimens subsequently collected
were obtained, remarked:

The so-called Tmesipteris alleni (Lx.) Hollick, although common, can
not be referred to any genus known to those who have examined it. It has
nothing whatever to do with Tmesipteris, nor does it belong to Salvinia or
Ophioglossum, as placed by Lesquereux. It may be known for the present

as Carpolithes alleni (Lx.).

Thus it was given its fourth generic appellation — one that
suggested similarity in appearance to a carpel or capsule or
pod-like organism of some kind.

In 1919 Florin, § in a paper on fossil Salvinias, expressed an
opinion similar to that expressed by Cockerell — to the effect
that the taxonomic status of Lesquereux' species was too un-
certain to warrant its reference to any well-defined genus — and
relegated it to the comprehensive fossil genus Phyllites, under the
name P. alleni (Lesquereux) Florin (loc. cit. p. 254). The fifth
generic name was thus applied in connection with the species — a
name, apparently, meant to indicate that the author intended to
suggest that it probably represented a foliar organ.

* Fossil Salvinias, including description of a new species. Bull. Torrey
Club 21: 253-257. pL 205. 1894.

f Dalbergia reticulata Ettingshausen, Beitrag zur Kenntniss der Fossilen
Flora von Tokay. Sitzb. K. Akad. Wiss. [Wien.] Math.-Naturwiss. CI. II4:
813. pi. 4, f. 5. 1853. Reproduced on Plate 10. fig. 5.

J The Fauna of the Florissant (Colorado) Shales. Am. Jour. Sci. 36:
498-500. 19 1 3.

§ Eine Uebersicht der Fossilen Salvinia- Ar ten, mit besonderer Beriick-
sichtigung eines Fundes von Salvinia Formosa Heer in Terti/ir Japans. Bull.
Geol. Inst. Upsala 16: 249-260. pi. 11. 19 19.

Hollick: Status of Ophioglossum Alleni L. 200

Recently Professor Cockerell* called attention to the fact that,
in connection with certain specimens of the species, there
could be detected a peculiar protuberance or body, apparently
attached to the midrib, near the middle, in regard to one of
which he remarked (loc. cit. pp. 211-212):

It looks like a small berry, with the contents extruded. . . Probably! lie
riddle would never have been solved but for the discovery of a younger
specimen . . . This . . . shows that we have to do neither with a
leaf, frond nor pod, but with a cladode. Attached to the midrib is an indis-
tinct mass, presumably a thin bract, upon which can be seen a dark object

which seems to agree very closely with the flower of Ruscus
The genus Ruscus [has] . . . lanceolate to ovate sharply pointed cladodes
At first sight the venation seems quite different from that of I he
fossil, hm if we imagine the Ruscus cladode br< adened and abbreviated until the
principal veins are nearly or quite transverse instead of longitudinal, the cor-
respondence is exact.]

Based upon this course of reasoning a new genus was evolved,
and Lesquereux' multi-generic species became Brachyruscus
Alleni (Lesquereux) Cockerell (loc. cit. p. 212) — the sixth
binomial under which it was described and discus: ed. Inciden-
tally it may also be noted that this last change of name definitely
transferred the species from the Pteridophyta to the Spermato-
phyta and changed it from a frond, a leaf, or a fruit to a cladode.
The question whether or not this last change of name should
be accepted as the final word in regard to its probable taxonomic
and morphologic status did not, however, appear to be answered
conclusively or satisfactorily, based as it was on a frank appeal
to the imagination to supply the necessary evidence. Also
the naively worded footnote on page 212 appeared to be more
or less of the nature of a challenge. In any event the statements
and conclusions set forth in the paper certainly called for a
critical examination of all of the available facts and their faithful
presentation, without any attempt to connect them with any
preconceived theory and without any appeal to the imagination
to supply any features in the specimens that might be poorly
preserved or lacking.

During a recent visit to the United States National Museum
I was enabled to examine some fifty specimens of the species in

* A new genus of fossil Liliaceae. Bull. Torrey Club 49: 211-213. /. 7.
1922.

t The italicising is mine.

210 Hollick: Status of Ophioglossum Alleni L.

question, all of them from the type locality at Florissant,
Colorado. Among these I had no difficulty in finding a dozen
or more that showed, either distinctly or indistinctly, the
presence of some kind of a body, as described by Professor
Cockerell. Nine were selected for illustrating the err -al char-
acters, all of which were photographed and are .^produced,
natural size, on Plate io, figs. 6-ii, and on Plate ii, figs. 1-3;
and the latter three were also, enlarged so as to show the char-
acters more clearly. The three enlargments are reproduced
on Plate ii, figs. ia-3a. The salient characters that they
may be seen to reveal are: (1) a system of reticulate nervation,
with the nerves of uniform rank throughout; and (2) either an
indistinct thickening or a well defined, flattened spheroidal
body (occasionally two), located toward the middle of the
foliaceous organ.

The general appearance of the specimens, however, strongly
suggests that they represent a flattened pod, or a detached
carpel of a pod, rather than a foliaceous organ, to which are
attached either immature, or imperfect, or mature and perfect
seeds. In certain of the specimens, as in those represented on
Plate ii, figs, i, ia, and 3, 3a, the seedlike bodies are appar-
ently in the superior position and are impressed on the exposed
surface of the specimens. In others, as represented on Plate
1 1, figs. 2, 2a, they appear to occupy the inferior position and
are more or less masked by the overlying tissue. The superior
position is also indicated in the specimens depicted on Plate
10. figs. 10, 11, and the inferior position in Figs. 6-9 on the
same plate. The latter figures also appear to represent more or
less immature or imperfect specimens; and in the specimen
represented by Fig. 8 the overlying tissue near the middle was
carefully chipped away from what appeared to be some kind
of a body beneath, but nothing was revealed except some
obscurely defined ridge and furrow markings that appear to repre-
sent a slight thickening or expansion of the median nerve or rib.
On Plate 12, fig. i, is shown a pressed capsule of Staphylea
trifolia Linnaeus, and in Fig. 2 the exterior of a detached carpel
of Koelr enter ia bipinnata Franchet. In each may be seen the
surface inequalities due to the underlying seeds. Fig. 3 repre-
sents the exterior of a carpel of Koelreuteria paniculata Laxman ;
Fig. 4, the interior with seeds attached to the alate expansion

Hollick: Status of Ophioglossum Alleni L. 211

of the midrib; Fig. 5, the latter flattened by pressure; Fig. 6,
the same with seeds detached. Fig. 7 is a reproduction of
Cockerell's figures illustrating his description of Brachyruscus,
and Fig. 8, a drawing of a cladode of Ruscus aculeatus Linnaeus,
with infloresepce attached. A comparison between the last
two figures hardly seems to suggest, even with the assistance
of an earnest effort of the imagination, that there could be any
near relationship between the specimens that they represent;
and, in any event, it is evident that, if the description and
figures of Brachyruscus correctly describe and depict the speci-
men selected by the author as the type of the genus, Brachy-
ruscus Alleni Cockerell can not possibly be either congeneric
or morphologically identical with Ophioglossum Alleni Les-
quereux. The author has, however, provided for this contingency
to a certain extent, in the footnote previously mentioned,
which should be read in this connection.

Incidentally, on Plate 12, fig. 9, is a drawing of a bract of
Dobinia vulgaris Hamilton, showing the character of the nerva-
tion and a fructification attached to the midvein, in order that
it may perhaps be recognized and accepted, in preference to
the cladode of Ruscus, as the morophological equivalent and
a possible family relative of Brachyruscus.

In view of the facts here presented if anyone should deem it
necessary to coin a seventh generic name for Ophioglossum Alleni,
based upon its resemblance to capsules or carpels of species of
certain genera included in the Staphyleaceae and Sapindaceae,
the responsibility will rest on him, not on me. The already
established name, which is descriptive of what the species appears
to morphologically simulate and which is therefore properly ap-
plicable, is Carpolithes Alleni (Lesquereux) Cockerell.

Explanation of plates 10-12

Plate 10

Salvinia Alleni (Lesquereux) Lesquereux
[= Ophioglossum Alleni Lesquereux]

Fig. I. U. S. Geol. Survey Terr. Rept. Vol. 7 (The Tertiary Flora): pi.
$.f. a. 1878.

10 Fig. 2. Idem. Vol. 8 (The Cretaceous and Tertiary Floras): pi. 21, f.
1883.

Fig. 3. Idem. /. //.

212 Hollick: Status of Ophioglossum Alleni L.

Salvixia reticulata (Ettinghausen) Heer
F"ig. 4. Flora Tertiaria Helvetiae 3: pi. 145, f. 16. 1859.

Dalbergia reticulata Ettinghausen

Fig. 5. Sitzb. K. Akad. Wiss. [Wien] Math.-Naturwiss. CI. II4: pi. 4,
f. 5- 1853.

Salvixia Allexi (Lesquereux) Lesquereux

Reproductions of photographs, natural size, of specimens from the type
locality at Florissant, Colorado.

Fig. 6. Showing indications of immature or imperfectly developed
fructification. U. S. Nat. Mus. No. 36693.

Fig. 7. Idem. U. S. Nat. Mus. No. 36690.

Fig. 8. Idem., with overlying tissue removed. U. S. Nat. Mus. No.
36692.

Fig. 9. Showing indications of more maturely developed fructification.
U. S. Nat. Mus. No. 36694.

Fig. 10. Showing mature fructification. U. S. Nat. Mus. No. 36691.

Fig. 11. Idem. U. S. Nat. Mus. No. 36688.

Plate i i
Salvixia Allexi (Lesquereux) Lesquereux
Reproductions of photographs of specimens from the type locality at
Florissant, Colorado, showing mature fructifications.

Fig. 1. Natural size. U. S. Nat. Mus. No. 36695.

Fig. ia. Same specimen, X 2.

Fig. 2. Natural size, U. S. Nat. Mus. No. 36689.

Fig. 2a. Same specimen, X 2.

Fig. 3. Natural size. U. S. Nat. Mus. No. 50277.

Fig. 3a. Same specimen, X 2.

Plate 12

Staphylea trifolia Linnaeus

Fig. 1. Capsule, natural size, compressed, showing inequalities due

to included seeds; distal end upward. From a specimen in the herbarium of

the New York Botanical Garden, collected at Knoxville, Tennessee, in

April, 1878, by Albert Ruth, No. 358.

Koelreuteria bipixxata Franchet
Fig. 2. Carpel, natural size, compressed, showing outer surface with
inequalities due to included seed; distal end downward. From a specimen in
the herbarium of the New York Botanical Garden, collected in Central
China, 1885-88, by Angust Henry, No. 7591.

Koelreuteria paxiculata Laxman

Carpel, natural size, under varied conditions; distal ends downward.
From a tree in the grounds of the U. S. Department of Agriculture, Washington,
D. C.f collected in August, 1922, bv Arthur Hollick.

Fig. 3. Outer surface.

Hollick: Status of Ophioglossum Alleni L. 213

Fig. 4. Inner surface, with seeds attached to a late expansion of the
midrib.

Fig. 5. Same, flattened by pressure.

Fig. 6. Same, with seeds detached.

Brachyruscus Alleni (Lesquereux) Cockerell
Fig. 7. Bull. Torrey Club 49: /. 13, A-D. 1922.

Ruscus aculeatus Linnaeus
Fig. 8. Cladode, natural size, upper surface, showing inflorescence
attached to the median nerve; distal end upward. From a cultivated specimen
in the herbarium of the New York Botanical Garden, collected at Montpellier,
France.

Dobinea vulgaris Hamilton
Fig. 9. Bract, natural size, upper surface, showing fructification;
distal end upward. From a specimen in the herbarium of the New York
Botanical Garden, collected in Nepal, India, probably in 1821, by Nathaniel
Wallich.

Bull. Torrey Club

Volume 50, plate 10

ft
6

'«:.' Pfl

11

HOLLICK: STATUS OF OPHIOGLOSSUM ALLEN!

Bull. Torrev Club

Volume 50 plate i i

HOLLICK: STATUS OF OPHIOGLOSSUM ALLEN1

Bull. Tcrrey Club

Volume 50, plate 12

8 9

HOLLICK: STATUS OF OPHIOGLOSSUM ALLENI

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL
GARDEN— Xo. 246

STUDIES IN TROPICAL ASCOMYCETES-IL

AN INTERESTING XYLARIA

FROM PORTO RICO

BY

FRED J. SEAVER

NEW YORK
1923

Reprinted, without change of paging, from Bulletin of the Torrev
Botanical Club 50: 307-310 pi. 18. Sept. 1923.

LttJRAKY

NEW YOKK

BOTANICAL

GARBEN

Studies in tropical Ascomycetes — II. An interesting Xylaria
from Porto Rico

Fred J. Seaver

(WITH PLATE 1 8)

During a recent collecting trip to Porto Rico* the writer
had the opportunity of visiting one of the most interesting and
fruitful collecting grounds for fungi to be found on the island,
the Luquillo Mountains. This trip was made possible through
the efforts of Mr. E. M. Bruner, Forester of Porto Rico, who
placed at our disposal the Forest Ranger's cabins, located at an
elevation of about 1200 feet. This camp consisted of two cabins,
one of boards with metal roof and the other a rough building
with thatched roof. The cabins were provided with cots,
utensils, and charcoal stoves for cooking. Mr. Bruner acted
as host and s ipervised the house-keeping activities during our
stay here. The party consisted of Dr. and Mrs. Britton, Mr.
Bruner, Professor Dale of the University of Porto Rico, and the
writer.

The trip from San Juan to Mameyes was made by auto-
mobile and the ascent from this point to the Forester's camp
on foot or by horse. Arriving at the camp soon after noon on
Friday, February 23, the remainder of the afternoon was spent
in the vicinity of the camp site.

Saturday, February 24, an expedition was planned to the
top of El Yunque, for a long time supposed to be the highest
peak in Porto Rico but now known to be exceeded by several
others. The ascent was made much easier by reason of the trail
of broken stone which had been constructed under the super-
vision of Mr. Bruner for a greater part of the distance. Soon
after the expedition started, it began to rain, as it nearly always
does at this elevation, and a slow drizzle continued during the
remainder of the day, making the climb at places very slippery
_and disagreeable. We succeeded, however, in reaching the top
c '

* Jour, of the New York Bot. Card. 25: 99-101. 1923.

307

c^L

308 Seaver: Studies in tropical Ascomycetes

at an elevation of about 3500 feet. Here the soil and atmosphere
were saturated so much so that it was impossible to find a dis-
place where one could sit to rest or eat. The trunks of the trees
and even the leaves were loaded with mosses and hepatics, and
there seemed to be scarcely room for a fungus to grow, although
a number of specimens of Aschersonia in fruit (Ilypocrella) were
found and in some abundance.

The view, which must be remarkable under favorable con-
ditions, was entirely missed by us on account of the atmospheric
conditions. In spite of the fact that we were under the tropical
sun, had it not been obscured by the clouds and fog, a cold chill
wind swept over the mountain top which compelled us to look
for a sheltered spot where we could eat our lunch in compara-
tive comfort.

Not finding the conditions especially favorable for myco-
logical collecting at this elevation the writer preceded the party
down the trail in order to have time to hunt leisurely for things
of a mycological nature. Many interesting collections were
made, most of which have not been studied critically. One which
was of more than usual interest was a species of Xylaria which is
made the subject of the present paper. These specimens were
found opposite the overhanging rock about midway between
the summit and the Forester's cabin to the left of the trail going
down.

On account of the density of the forest and the general gloomi-
ness of the day it was difficult to detect the presence of minute
fungi. However, on picking up what appeared to be a dead
stick a seed the size of an ordinary marble was found attached
to its base. On closer examination the object, which proved
to be a large fungus, was at first thought to be a fine Cordyceps.
It finally turned out to be a Xylaria. On further search other
plants were found until fifty or a hundred specimens had been
collected, each growing from a seed of what proved to be that
of the mountain palm, Euterpe globosa Gaert. A number of
fungi, such as certain species of Sclerotinia, grow exclusively on
seeds. Also a number of species of Cordyceps grow on under-
ground fungi which have much the same general appearance,
but so far as the experience of the writer has gone this is the
only species of Xylaria encountered which has the general ap-
pearance of a Cordyceps with the habitat of the Sclerotinia.

Seaver: Studies in tropical Ascomycetes 309

The species has been found only in the one locality in Porto
Rico. The host on which the fungus preys, however, occurs
from Cuba to Hispaniola and Porto Rico and in the Lesser
Antilles from Saba to Grenada. Assuming that the fungus is
restricted to the one host, as it seems to be, its possible range
of distribution then is as indicated above. The only economic
bearing which the fungus could have is in connection with the
destruction of the seeds of the host. How important this can
be cannot be stated.

Our material from Porto Rico agrees quite closely, so far as
we can judge from the description, with Xylaria palmicola
Winter,* described from Brazil. No specimens of the South
American species have been seen and the host of that species
is not named. The fact that the host of the Porto Rican species
does not occur in South America, so far as we know, leads the
writer to regard ours as distinct. The species is dedicated to
Mr. Bruner and a description appended.

Xylaria Bruneriana, sp. nov.

Strom. ita growing upon and their mycelium tilling the seeds
of their hosl and apparently absorbing the greater part of their
contents, reaching a height of 10 12 cm. and a diameter of 2-4
mm., consisting of a sterile stem and a fertile club-shaped head,
the latter comprising aboul one fourth of the whole stroma and
considerably thicker; stem cylindrical, slightly tapering above
and becoming twisted (especially when dried); peril hecia sub-
globose, aboul 50a j. in diameter; asci subcylindric, eight-spored,
reaching a length of 140 ;j., theascus wall being very transparenl
and often almost invisible; spores unequal-sided, rather sharp
pointed, [6x6 7 y., often slightly flattened, hyaline, becoming
dark brown.

( )n fallen seeds of 1 he mountain palm, Jinlcrpc globosa ( Jaert.,
El Yunque, February 24, 1923.

•Grevillea 15: 89. [887.

Description of plate 18
Xylaria Bruxeriana Seaver

Fig. i. Several plants about natural size.

Fig. 2. Diagram of a section of the fertile head.

Fig. 3. Ascus with spores, drawn with the aid of the camera lucida.

Bull. Torrey Club

Volume 50, plate 18

XYLARIA BRUNERIANA Seaver

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN

No. 247

NOTES ON FABACEAE— II

PER AXEL RYDBERG

NEW YORK
1923

Reprinted, without change of paging, from Bulletin op the Torrey Botanical Club
50: 261-272. August 17, 1923.

LI«RaRY
NEW YORK
BOTANICAL

HARDEN

Notes on Fabaceae — II

Per Axel Rydberg

HOMALOBUS Nutt. (con.)

D. Campestres

This group differs from the Tenelli not only in the narrow-
strictly sessile pod but in the acute keel-petals, a character which
as far as I know separates it from all the other groups of
Homalobus.

21. Homalobus campestris Nutt. This was described
from specimens collected on the "Colorado of the West," mean-
ing some of the tributaries of the Green River. One specimen
of Nuttall's is in the Torrey Herbarium and is in flower. The
calyx is 5 mm. long, more or less black-hairy; the lobes are
subulate, decidedly longer than broad; nearly every leaf has one
or two pairs of lateral leaflets, and the terminal leaflet is repre-
sented by the flattened, somewhat falcate prolongation of the
rachis. The corolla is evidently ochroleucous, without a trace
of purple. Gray, in transferring this species to Astragalus,
evidently confused it with some other member of the group,
describing the leaves as having often nine to thirteen leaflets and
the corolla as being yellowish white, usually tinged with purple.
He also regarded II. tenuifolius Nutt. as a depauperate form
and evidently had in mind II. hylophilus Rydb., II. oblongifolius
Rydb. and their allies. As Astragalus campestris was preoc-
cupied, Greene substituted the name A. convallarius . Jones at
first made the plant A. serotinus campestris * but cited under this
name specimens belonging to other species, one of them, 601511,

*Proc. Calif. Acad. II. 5: 668. 1895.

[The Bulletin for July (50: 231-260 pi. 13, 14) was issued July 27, 1923.]

261

262 Rydberg: Notes on Fabaceae— II

being H. humilis. Later* he changed the name to A. decumbens
convallarius . Even at that time, however, he had a misconcep-
tion of certain species; for, on a later page of the same article,
he changed Astragalus junceus (Nutt.) Gray to A. diver sifolins
roborum and labeled with this name specimens belonging to
H. campestris.

Macbridef united A. campestris and A. junceus, regarding
"the relative length of the calyx-lobes and the presence or
absence of black pubescence" as "trivial variations" and of
"no practical taxonomic value." Perhaps they are so and per-
haps H. campestris, H. junceus and H. junciformis should be
only one species, but he has used one of these characters, the
longer calyx-lobes, together with the presence of lateral leaflets,
as a character to separate his A . campestris diver sijolius (Gray)
Macbr. from the typical form of the species. These characters
are just those that characterize the type.

Only two of the specimens cited by Macbride under the variety
diver sijolius are found in New York herbaria: the first of these
is Nuttall's, from the "sources of the Platte," Wyoming, which
represents the type collection of H. orthocarpus Nutt.; and the
second is Baker 242, from Cedar Edge, Colorado, which matches
the type of H. campestris in every respect. Evidently Macbride
has misinterpreted A. diver sijolius A. Gray (see below). Of
the specimens in the Gray Herbarium, Canby 84, from Helena,
Montana, is H. junceus; so also are Watson Q2 and Nelson &
Macbride 12Q2, both from Idaho. The rest represent the typical
H. campestris, except Macbride & Payson 3183, in flower only,
from Clyde, Idaho, which may be H. orthocarpus. In the
herbaria of the New York Botanical Garden and Columbia
University, the following specimens of H. campestris are found : —

Wyoming: Nuttall.

Utah: Glenwood, Ward 22J; Hooper, Leonard q8; Tropic,
Jones 5312k; Mandan, Muljord 118; Salt Lake City, Garrett
Q84; Pavant Mountains, Eggleston 1014Q.

Colorado: Cedar Edge, Baker 242; Como, Crandall; Middle
Park, Patterson; Glenwood Springs, Osterhout; White River,
Mrs. Dun.

* Contr. W. Bot. 10: 58. 1902.

f Contr. Gray Herb. II. 65: 35. 1922.

Rydberg: Notes on Fabaceae — II 263

22. Homalobus junceus Nutt. The duplicate of the type
in the Torrey Herbarium is in fruit but shows a white-hairy
calyx with very short calyx-lobes and leaves with no or very
small lateral leaflets. Sheldon thought that this and Astragalus
diversifolius represented but one species and applied the latter
name to it, as the earlier A. junceus was preoccupied by A.
junceus Ledeb. Since, however, A. diversifolius is a synonym
of H. orthocarpus, A. junceus, if transferred to Astragalus, would
be without a name. This species is much more common than
H. campestris. I shall therefore not enumerate any specimens.
Most of those cited by Macbride under A . campestris belong here.
Jones, as already noted, named the species A. diversifolius
roborum.

23. Homalobus junciformis (A. Nels.) Rydb. The status
of this species is somewhat problematical, since it has the short
calyx-lobes of H. junceus and the black-hairiness of H. cam-
pestris. It is stouter than either, with larger flowers and
broader leaflets. Standley regarded it as identical with H.
orthocarpus, but that species has smaller flowers and a much
broader pod. The mature pod of H. junciformis is unknown,
but the young pod suggests a similarity to that of H. junceus.
Jones* made H. junciformis a synonym of A. diversifolius,
evidently not being acquainted with the true A. diversifolius.
The only specimens of H. junciformis seen are: —

Wyoming: Fort Steele, A. Nelson 483Q; Point of Rocks, A.
Nelson 3081; Leucite Hills, A. Nelson 4668.

24. Homalobus orthocarpus Nutt. The duplicate of the
type in the Torrey Herbarium shows only the upper part of the
plant, but Nuttall states that the lower leaves bear several
leaflets, while the upper are more or less reduced to the rachis.
This fact furnished Gray the name Astragalus diversifolius,
when he found that A. orthocarpus was preoccupied by A. ortho-
carpus Boiss. In H. orthocarpus the calyx is white-hairy, the
lobes are deltoid and not much longer than broad, while the
pod is much broader than in the related species, in the duplicate
of the type being about 2 cm. long and 4 mm. wide. In the
herbarium of the New York Botanical Garden the only specimen
like the type is Goodding 1084 from Juab, Utah, which has been

* Contr. West. Bot. 10: 68. 1902.

264 Rydberg: Notes on Fabaceae — II

referred to A. episcopus Wats. According to Macbride A. epis-
copus has sessile pods; one might suppose from this that the
latter species was a synonym of H. orthocarpus; but it is not, since
it really has a short-stipitate and glabrous pod. There is also
a specimen collected at Salida, Colorado, Johnston & Hedgecock
634, which resembles Goodding's very much, but it is in flower
only and therefore doubtful. So also is Macbride & Pay son
3183 from Idaho, to which reference has already been made.

25. Homalobus decumbens Nutt. The specimen in the
Torrey Herbarium consists of two pieces, a small plant with five
attached branches arid a single loose branch; the former bears
one mature pod and two small racemes in bud; the separate
branch bears five immature pods. All the pods are minutely
pubescent and decidedly arcuate, and the mature pod is nearly
2.5 cm. long and 3 mm. wide; the calyx lobes are subulate and
less than half as long as the tube. The only specimens in the
herbarium of the New York Botanical Garden similar to the
type are Hall & Harbour 142, in part, from Colorado, and
Goodding 142Q from Wood's Creek, Wyoming. The pods of
the latter are very young and scarcely show any indication of
being arcuate. The specimens in the Gray Herbarium bearing
the same number have a better developed pod, and I would
refer them to H. microcarpus.

A specimen in flower, from Mammoth Hot Springs, Wyoming,
named by the collectors A. decumbens, viz., Aven & Elias Nelson
5649, resembles this species in habit but is more canescent.
Macbride has cited this specimen under A. divergens, but I
rather think that it belongs either here or under H. tenuifolius
Nutt.

26. Homalobus serotinus (A. Gray) Rydb. Gray de-
scribed Astragalus serotinus as having glabrous or minutely
pubescent pods, and Macbride says in his key, "pods glabrous
or nearly so." There are three sheets of Cooper's collection in
the New York herbaria, and in all three the pods are minutely
strigose. The species varies much in the width of the leaves.
When they are very narrow the plant resembles much in habit
II. campestris, but the tip of the keel is always purple and the
leaflets more numerous and glabrate above.

Rydberg: Notes on Rydberg — II 265

Washington: Okanogan River, Cooper; Peshasti, Sandberg
& Leiberg 473; Loomis, Griffiths & Cotton 343; without locality
Vasey 273.

Oregon: Wallowa County, Sheldon 8065.

Montana: Big Fork, Umbach 621.

Alberta: Tunnel Mountain, McCalla 2134.

British Columbia: Nicola Valley, Dawson 4269.

27. Homalobus strigosus (Coult. & Fish.) Rydb. This
resembles H. Palliseri in habit but differs in the pubescent pod
and longer calyx-lobes. Jones* claims that it is the same as
II. serotinus. It is true that the two resemble each other very
much in general appearance, but the calyx-lobes in H. strigosus
are almost twice as long and the leaves are cinereous on both
sides. Macbride regards it as a variety of H. serotinus.
Alberta: Vermillion Range, Moodie 1238, in part.
Montana: Park County, Tweedy; St. Ignatius Mission,
MacDougal 290; Columbia Falls, Williams 1003, Helena, Kelsey;
Tobacco Mountain, Butler; Deer Lodge River, Cooper.

Wyoming: Mammoth Hot Springs, Mearns 1329; Alcova,
Goodding 146;} Sheridan and Buffalo, Tweedy 3168; Big Horn,
Tweedy 2355.

Idaho: Beaver Canon, Watson 91* Howe, Macbride &
Payson 3106, \ in part; Mammoth Hot Springs, Nelson & Nelson
5648;} Salmon, Payson & Payson 1762.

28. Homalobus tenuifolius Nutt. The duplicate of Nut-
tail's type in the Torrey Herbarium is a small plant about 7
cm. high and bears two racemes, one with six flowers and the
other with a single pod. The calyx is 2 mm. long, including the
lobes, which are about 0.5 mm.; the corolla is about 4 mm. long,
and the pod, 14 mm. long and 3 mm. wide, is strigose; the leaflets
are narrowly linear-subulate. In the herbarium of the New
York Botanical Garden there is only one specimen in fruit like
it, Tweedy 259. The others are larger and would be referred to
H. decumbens if they were not so canescent and if the pods were
not straight. Gray, as already noted, thought that the species
was a depauperate form of H. campestris. Coulter and Nelson
seem to include it in that species, while Macbride united it with

* Proc. Calif. Acad. II. 5: 668. 1895.

t Referred by Macbride to A. divergens Blankinship.

266 Rydberg: Notes on Fabaceae — II

A. divergens Blankenship. It would have been more logical
to reduce it to a variety of A. serotinus. It is not more distinct
from that species than A . strigosus or A . Palliseri. None of the
specimens cited by Macbride and seen by me belong to H. tenui-
foliiis except Parry 81 and A. Nelson 9648. See under H. stri-
gosus and H. decumbens.

Wyoming: Rocky Mountains, Nuttall; Parry 81; Mount
Leidy, Tweedy 259; Medicine Bow, A. Nelson 9648.

Montana: Lima, Rydberg 2706; Red Rock, Shear 5043.

Colorado: Lake City, Pease; Hall & Harbour 142, in part.

29. Homalobus divergens (Blankinship) .Rydb. In habit
the plant resembles the preceding, but the leaflets are much
broader and shorter. I have not seen the type, but Blankinship 's
illustration (except the pod) represents the plant I have in mind.
In the original diagnosis the pod is described as stipitate and is
so figured. In all the specimens I have seen it is sessile.
Blankinship might have had a mixture containing pods of some
other species, as for instance the alpine form of H. Bourgovii.
Not having seen the type, it was not so strange that I, under
these circumstances, redescribed the plant as H. camporum*
This is the name it would bear, if by chance Astragalus divergens
should be different from our present conception.

Wyoming: Leckie, Merrell & Wilcox 537, 585; Steamboat
Mountain, A. Nelson 7085; Laramie Hills 198; Bird's Eye,
Nelson 9559.

Montana: Helena, Rydberg 2708.

Idaho: Beaver Canon, Rydberg 2707.

Utah: Alta, Jones i2io.\

30. Homalobus humilis Rydb. Macbride has transferred
this to Astragalus, under the name A. Carltonii, as there was an
older A. humilis Bieb.

Utah: Bullion Creek, Rydberg & Carlton 7147; Delano Peak,
7219 and 7219a; Mt. Ellen, Jones 5670; Panguitch Lake, 601571
(this with narrow leaflets).

31. Homalobus oblongifolius Rydb. Macbride regarded
this as a variety of H. hylophilus. As far as I know the plant

* Bull. Torrey Club 32: 666. 1906.

f This Macbride refers to A. Carltonii Macbride, i. e. to H. humilis, but the
pod is that of II. divergens and //. tenuifolius.

Rydberg: Notes on Fabaceae — II 267

is confined to Colorado. Macbride cites two specimens from
Wyoming, Nelson 869 and Payson 2068. The former, as repre-
sented in the Columbia University Herbarium is typical H. hylo-
philus, having glabrous pod and broad leaflets. The other
number I have not seen.

Colorado: Cerro Summit, Baker 409; Nederlands, Clokey
3284, Leadville, Jones, Teller, C. S. Sheldon 108; Marshall Pass
and Leadville, Jones; Pinkham Creek, Goodding 1472, Delta,
N. F. T. Nelson 181; without locality, Parry, in 1864.

32. Homalobus paucijugus Rydb. This resembles H.
tenuifolius in habit and pod but differs in the few leaflets. As
there is an older Astragalus paucijugus Schrenk, Macbride re-
named my species A. Garrethi.

33. Homalobus decurrens Rydb. This was also renamed
A. Rydbergii by Macbride, on account of A. decurrens Boiss.

Colorado: Golden, Jones 528; Greene; Estes Park, Oster-
hout 2222; Lake Eldora, Clokey 3201; Jefferson Countv, 3808;
Larimer County, OsterhontA; Stove Prairie and Como, Crandall;
Gray Back, Rydberg & Vreeland 5960; Tolland, Johnston &
Hedgecock 664; without locality, Hall & Harbour 142, in part;
Central City, Sccvell.

34. Homalobus microcarpus Rydb. This was reduced by
Macbride to a synonym of A. decumbens. There would have
been more reason for reducing it to, or for making it a variety of,
H. divergens, since the general habit, the form of the leaflets
and the shape of the pod approach that species, but the whole
plant is greener, more luxuriant and more erect. One specimen,
Watson 287 from the Uintah Mountains, Utah, may belong
here. It has the small pod and the flowers of H. microcarpus
but is taller, with very long peduncles, 10-14 cm. long, longer
leaves, up to 15 cm. long, and more numerous leaflets, the pod
is hairy with short black hairs. Macbride erroneously referred
the specimen to Astragalus serotinus var. Palliseri, which has
glabrous pods.

Colorado: east of Rabbit-Ear Range and on the forks of
Poudre River and Big South, Crandall; Como, Crandall & Cowen
131; North Park, C. F. Baker; Steamboat Springs, Osterhout
2774; Empire, Tweedy 5642; without locality, Hall & Harbour
143; Parry 435.

268 Rydberg: Notes on Fabaceae — II

Wyoming: Woods Creek, Goodding 1429.

35. Homalobus hylophilus Rydb. This is the most common
species of the group in the Northern Rockies, extending as far
south as northern Utah and the Black Hills, but not being found
in Colorado. Jones* claims that this is a shade form of A. de-
cumbens convallarius , due mainly to his misinterpretation of
A. convallarius , which latter was based on Homalobus campestris
Nutt.

36. Homalobus Palliseri (A. Gray) Rydb. Macbride re-
duced this to a variety, A. serotinus Palliseri (Gray) Macbride.
It is more closely related to H. hylophilus, having the same
glabrous pod, but the leaflets are much narrower and the calyx-
lobes shorter. There are, however, specimens of H. hylophilus
which have nearly as narrow leaflets as those of H. Palliseri.
The Montana specimens of H. Palliseri have longer pods and
in that respect approach H. hylophilus.

Alberta: Rocky Mountains, Bourgeau; headwaters of the
Saskatchewan, S. Brown 923; Banff, 5. Brown 80; VanBrunt 6;
Canby 41; Barber 105, 279.

British Columbia: Carbonate, Shaw 227, 241; Green, 3210;
Kettle River, /. M. Macoun 63152, 63756; Lake Osoyoos, 70440;
Canmore, /. Macoun 15; Maligne River, Spreadborough 19310;
Armstrong, E. Nelson 133.

Washington: Concomilli, Griffiths & Cotton 319; Ophir,
Elmer 528; Walla Walla, Brandegee.

Montana: Spanish Basin, Flodman 636; Bozeman, Wilcox
115; Old Hollowtop, Rydberg & Bessey 4489; Bridger Mountains,
4494; Old Sentinal, MacDougal 168, 172; Terminus, Watson 88.

37. Homalobus detritalis (M. E. Jones) Rydb. This has
been placed in the present group but probably does not belong
here. In general habit it resembles the Simplicifolii, but the
leaves have five leaflets. Whether it has the acute keel-petals
characteristic of the. Cam pestres or not I do not know, as I did
not look for that character when I saw the type some years ago.

E. Stenophylli

Pod glabrous, except the stipe.
Leaflets narrowly linear.

Contr. West. Bot. 10: 69. 1902.

Rydberg: Notes ox Fabaceae — II 269

Leaflets few, the terminal one wanting or continu-
ous with the rachis; calyx-lobes lanceolate, 1.5-2
mm. long. 38. H. Coltoni.

Leaflets many, the terminal one present ; calyx lobes
triangular, 1 mm. long or less.

Leaflets slightly strigose beneath or glabrous

throughout. 39- H. stenophyllus.

Leaflets strigose-cinereous on both sides.
Stipe of the pod fully half as long as the

body, ocrolla ochroleucous or white. 40. II. MacGregorii.

Stipe about one third as long as the body;
corolla purple. 41. H. canovirens.

Leaflets linear-oblong to oval, glabrous above,

strigose beneath. 42. II. Antiselli.

Pod strigose throughout.

Leaflets linear; raceme elongate lax.

Body of the pod 2-2.5011. long abruptly contracted

at the base, straight. 43. H.filtpes.

Body of the pod 2.5-3 cm- long, gradual!)' tapering

at each end, slightly arcuate. 44- TI. inversus.

Leaflets elliptic; raceme short dense. 45. .1. gavwtus.

38. Homalobus Coltoni (M. E. Jones) Rydb. comb. nov.
Astragalus Coltoni M. E. Jones, Zoe 2: 237. 1891.
Homalobus episcopus Rydb. Bull. Torrey Club 40: 53, in part.

191 3. Not A. episcopus S. Wats., 1875.

Astragalus Coltoni was based on specimens from Castle
Gate, Utah. It has the habit of //. junceus and H . campestris,
but the pod has a slender stipe, the corolla is purplish, the
wings are nearly as long as the banner and the keel is rounded
at the apex. It is therefore more closely related to H. steno-
phyllus. I had mistaken it for //. episcopus (see my notes
under that species on pages 184-5).

Utah: Castle Gate, Jones, in 1890 and 1894; southern Utah,
Bishop, in 1872.

New Mexico: Carrigo Mountains, Matthews, in 1892.

39. Homalobus stenophyllus (T. & G.) Rydb. This is the
best known species of the group and was the first one published,
the original name being Astragalus leptophyllus Nutt. As that
name was preoccupied, A. stenophyllus was afterwards substi-
tuted. Torrey & Gray suggested thai the species might belong
in the genus Homalobus. The type was collected in flower,
and the fruit remained unknown for some time. The type is

270 Rydberg: Notes on Fabaceae — II

perfectly glabrous, except for the more or less black-strigose
calyx, but sometimes the upper part of the stem and the lower
surface of the leaflets are slightly strigose. This is the case with
61. Watson 283, which was mistaken for S. filipes. Watson's
description of that species also applies to H. stenophyllus.

British Columbia: Nicola Valley, G. M. Dawson 4269,
in 1877; also J . Macoun, in 1889.

Oregon: Andrews, Griffiths & Morris 523; Hoover Creek,
Leiberg 140; Juniper Springs, Malheur County 2264; between
Dayville and Mitchell, Lawrence 1024; east of Heffner, 582 A;
Blue Mountains, 895; Burns, Peck 3029; without locality, Nivius.

Idaho:. Big Willow, Canyon County, Macbride 161; Mackay,
1414; House Creek, Nelson & Macbride 1761; Shoshone, 1165,
Martin, Macbride & Payson 3079; Reynolds Creek, Macbride
1029; Big Butte, Palmer 240.

Nevada: Quinn River Crossing, Griffiths & Morris IJ2;
Palisade, M. E. Jones, in 1882; Pah-Ute Mountains, S. Watson
283; Trucke Pass, Kenedy 1314; Alleghany Creek, Nelson &
Macbride 2166; Reno, Jones, in 1897; Montana headwaters of
the Missouri, Wyeth.

Washington: Coulee City, Henderson 2358.

40. Homalobus MacGregorii Rydberg, sp. nov.

A perennial, with a cespitose caudex; stems strigose, striate;
lower stipules triangular, scarious, connate, the upper ones more
herbaceous, distinct; leaves 5-8 cm. long; leaflets thirteen to
nineteen, linear, 10-15 mm. long, 1-2 mm. wide, strigose on
both sides; peduncles 1-2 dm. long; racemes lax, 5-12 cm. long;
bracts subulate, 2-3 mm. long, calyx strigose, the tube 4-5 mm.
long, the teeth triangular, 1 mm. long or less; corolla ochroleu-
cous or white; banner 12 mm. long, obovate; wings 11 mm. long,
the blade obliquely oblanceolate, equalling the claw; keel-
petals 10 mm. long, rounded at the apex, the blade broadly
lunate with a rounded basal auricle; pod stipitate, the stipe
12-15 mm. long, slightly pubescent, the body glabrous, linear-
elliptic, acute at each end, 2.5 cm. long, 5-6 mm. broad.

Type collected near Frazier Borax Mine, Mount Pinos,
Ventura County, California, June 12-14, 1908, Abrams & Mc-
Gregor 219 (herbarium of the New York Botanical Garden).

This species is related to H. stenophyllus but the stem and
leaves are strigose-canescent. It has been mistaken for H. An-

Rydberg: Notes ox Fabaceae — II 271

liselli but the leaflets are much narrower and pubescent on both
sides. Besides the type, the following specimens belong here.

Southern California: Kenworthy, II . M. Hall 1182; San
Bernadino Mountains, Parish Brothers 1281.

Lower California: Orcutt, in 1885; 856, in 1883.

41. Homalobus conovirens Rydb. Astragalus Coltoni mo-
abensis M. E. Jones. This is also related to H. stenophyllus,
but the leaves and stem are densely grayish strigose, the stipe
much shorter, the corolla dark bluish purple.

Utah : Western slope of La Sal Mountains, Rydberg & Garrett
S536, 8564; La Sal Mountains, Jones, in 1993.

Colorado: Naturita, Payson 336.

42. Homalobus Antiselli (A. Gray) Rydb. comb, now
Astragalus Antiselli A. Gray, Bot. Calif. 1 : 152. 1880.

This species differs from the three preceding in the broader
leaflets, glabrous above and strigose beneath. A. Hasseanus
Sheld.* is without doubt the same.

California : Santa Catalina Island , Kixford, in 1 9 1 4 ; Sulphur
Mountains, Abrams & McGregor 40; Santa Inez, Mrs. Curran,
in 1885; Ventura, Elmer 3616; Santa Barbara, S. F. P. [Peck-
ham ?]; Torrey Q2 bis; Buenaventura, Hasse, in 1888.

43. Homalobus filipes (Torr.) Heller. This is very closely
related to //. stenophyllus and distinguished only by the strigose
pod and strigose stem and leaves. The two species have been
confused and may not be specifically distinct.

Washington : Fort Okanogan, Pickering, in U. S. Exploring
Expedition; Grand Coulee, Griffiths & Cotton 440; Ritzville,
Sandberg & Leiberg i6j; Sprague, Sandberg & Leiberg, in 1893;
Ellenburg, Piper 2717; Yakima Region, Brandegee 732; Crab
Creek, Saksdorj 281.

Oregon: Alkali, T. Howell, in 1882.

Nevada: West Humboldt Mountains, Heller 10631.

44. Homalobus inversus (M. E. Jones) Rydb. comb, no v.
Astragalus inversus M. E. Jones, Zoe. 4: 276. 1893.

This is closely related to H. filipes, differing mostly in the
longer, more tapering and slightly curved pod.

* Minn. Bot. Stud. 1: 124. 1894.

272 Rydberg: Notes on Fabaceae — II

California: Susanville, Brandegee; same locality, Jones;
Lava beds, Modoc County, M. S. Baker, in 1893.

45. Homalobus gaviotus (Elmer) Rydb. comb. nov.
Astragalus gaviotus Elmer, Bot. Gaz. 39: 54. 1905.

This species is referred with some doubt to this group as the
flowers have not been seen. The pod is strigose, nearly straight
on the upper suture, the stipe is shorter than in the other species,
the raceme is short even in fruit and the leaflets broad.

California: Gaviota, Elmer 375 q; same locality, Eastwood 5$
(determined erroneously as Phaca canescens Nutt.).

CONTRIBUTIONS FROM THE NEW YORK
BOTANICAL GARDEN— No. 248

STUDIES OF LYTHRUM SALICARIA— I.
THE EFFICIENCY OF SELF-POLLINATION

By A. B. Stout

New York
1923

[Reprinted from the American Journal of Botany, 10: 440-449. October, 1923]

STUDIES OF LYTHRUM SALIC ARIA I. THE EFFICIENCY
OF SELF-POLLINATION

••I.
Reprinted from the American Journal of Botany, X: 440-449, October, 1923.

STUDIES OF LYTHRUM SALIC ARIA I. THE EFFICIENCY OF

SELF-POLLINATION

A. B. Stout
(Received for publication December 8, 1922)

The conditions that exist in species with trimorphic flowers permit the
investigation of the problems regarding the nature of sex-differentiation
and the degrees of compatibility between male and female organs under very
favorable circumstances. In these forms the morphological adaptations for
cross-pollination are often decidedly correlated with physiological incom-
patibilities which make cross-fertilization more certain by excluding the
functioning of the chance self-pollinations which occur.

In general it is to be recognized that sexual fusions are favored by
similarity of the gametes both in genetic constitution and in immediate
origin, and that such conditions as trimorphism and self-incompatibility
are to be regarded as, in a high degree, secondary and acquired. While
these conditions secure the advantages of bringing together gametes of
different origins and in greater or less degree of different genetic constitu-
tion, they decidedly limit and restrict free fertilization and full productivity.

The combination of morphological trimorphism with physiological in-
compatibilities as seen in species like Lythrum Salicaria may well be regarded
as the highest degree of specialization in sex-determination and fertilization
that exists in flowering plants. For this species there is the obvious mor-
phological differentiation giving three lengths of pistils and three sets of
stamens of lengths corresponding to those of the pistils, with the stamens
bearing pollen differentiated as to size, color, and nature of the reserve food
material in storage. For the individual, the flower of any particular plant
has a pistil of one of the three lengths and a set of stamens for each of the
other two lengths. This gives differentiation of forms as such, and in the
single plant there is the differentiation that gives two kinds of stamens.
Furthermore, this morphological differentiation is decidedly correlated with
physiological differentiation. The noteworthy researches of Darwin (1865,
1877) showed that there is marked or even complete sterility to (1) self-
pollinations, to (2) intra-form cross-pollinations, and to (3) the inter-
form cross-pollinations that are illegitimate (i.e., those that involve different
lengths of pistil and stamen). Seed-production was hence found to be
more or less limited to legitimate pollinations, which are necessarily crosses.

The specializations in these forms allow no doubts as to their significance
such as have often been raised in regard to the colors of entomophilous
flowers, for here the adaptations are morphological and depend directly

Oct., 1923 1 STOUT — STUDIES OF LYTHRUM SALICARIA 44I

on the agency of insects for their effectiveness in pollination. The relative
lengths of the different sets of stamens and of the pistils are unquestionably
provisions favoring crossing.

It is, however, obvious that such differentiations are not fundamental
expressions of sexual antithesis, since they are all reciprocal in any pair of
flowers. Any two plants of any two forms are cross-compatible or cross-
incompatible according to whether the cross is legitimate or illegitimate.
The differentiations, at least in respect to maleness, that in dimorphic
plants are seen in individuals as such and which appear to have definite
genetic value are here seen equally well in the two sets of stamens in a single
flower. All this emphasizes the fact that the conditions are secondary and
acquired in contrast to the more primitive condition of homomorphism and
a more general compatibility of gametes.

The generally accepted view has been that the differentiations in this
species are well established and very stable. The tendency has been to
emphasize, as did Darwin, the evidence that here there is adaptation favor-
ing crossing, and to pass the evidence, which has to some extent been noted,
that the adaptations are incomplete. It is to be recognized that such
evidence has a very direct bearing on questions of the origin of trimorphism,
of the nature of sex-differentiations, and of whether there is still opportunity
for further selection in the species either toward greater or toward less
restriction of fertilization. It is evidence along these lines that the writer
wishes to present in reports, of which this is the first, of investigations with
the species.

The Efficiency of Self-pollination for Plants Grown in Isolation

The writer's studies of Ly thrum Salicaria were begun in 191 7 in testing
the self-compatibility of plants grown in isolation from other plants of the
same or of related species. On such a plant hundreds of flowers open
daily during a rather extended period of time and insect visitors can go
from flower to flower, but with no chance, if the isolation is complete, of
bringing pollen from other plants of Lythrum Salicaria. A large number
of flowers are thus involved in the chance for self-fertilization (including
here autogamy and geitonogamy), and the results can be obtained for the
entire period of bloom. This test does not, however, determine the relative
fertility of a plant to pollen of its two sets of stamens, nor does it reveal
the need or the efficiency of particular species of insects in the self-pollination
of the various forms, which may indeed give results that are highly variable
from season to season or from year to year or according to location. How-
ever, if seed is produced there is positive evidence of self-compatibility, and
the negative results may be tested further by controlled pollinations.

Short-styled Plants Grown in Isolation. Two large, well-developed plants
several years old were dug from a mixed population growing at the New
York Botanical Garden. One (S no. 2) was grown in the garden of the

442 AMERICAN JOURNAL OF BOTANY [Vol. 10,

Department of Botany at Columbia University, under the observation of
Professor R. A. Harper, and one (S no. i) at the University of Missouri under
the care of Professor G. M. Reed. Both plants made vigorous growth and
bloomed profusely, but neither plant produced a single capsule during the
entire period of bloom of the season of 1917. In 1918 the plant at Missouri
suffered severely from drought, and its failure to set any seed that year is
not to be considered as adequate evidence of complete self-incompatibility.

The short-styled plant at Columbia University thrived and in the course
of its season of bloom in 1918 produced 17 pods. Ten of these contained
only mere rudiments of seeds, two contained one good seed (plump and
apparently fully developed) each, three had two good seeds each, one had
seven seeds, and one had eight. Sixteen of these 23 seeds germinated. In
the following year (1919) this plant bloomed more abundantly than in the
previous year and produced at least 5000 flowers, and during the entire
season 161 pods matured. The seeds in 100 of these pods were counted; the
number per pod ranged to 116, and the average was 23.67 (see table 1).
This plant grew poorly in 1920, when its roots were separated to make two
plants. In 192 1 these thrived, and there was abundant bloom but only
about 25 pods were produced. The irregular pod production by this plant
may involve one or more of several conditions; possibly in 1919 insects may
have brought pollen from a distance from plants of this species growing in
city parks, or the irregularity may involve the presence or absence of certain
insects that are most efficient in causing self-pollination.

Long-styled Plants Grown in Isolation. A long-styled plant (L no. 1)
was grown at Baraboo, Wisconsin, under the care of Mr. William Toole, Sr., a
well known pansy specialist. Plants of the variety rosea were also growing
in the nursery at some distance away, so that some of the seed produced by
the plant L no. 1 may have been due to cross-pollination by insects. But
another long-styled plant (L no. 2) was grown in what was certain isolation
at Pleasantville, N. Y., under the care of Dr. M. A. Howe. Capsules were
produced by both these plants. Of the 65 capsules on L no. 1 in 19 17, 16
contained no seeds, 17 contained one seed each, and the highest number of
seeds in any capsule was 17. In 191 7 a total of 53 capsules matured on the
plant L no. 2. As shown in table 1, the number of seeds per capsule for
this plant was also low, although all but three of the capsules contained
some fully matured seeds.

In 1918 both plants produced capsules quite as in 1917, but the capsules
on L no. 2 were somewhat more numerous than in the previous year though
still constituting a small proportion of the entire number of flowers.

The capsules produced by these two plants were distributed irregularly
but rather indiscriminately throughout the flowering branches. Self-
fertilization appeared to be effected with the same frequency and efficiency
throughout the flowering period. No very decided seasonal differences
appeared in respect to the proportional number of flowers which matured

Oct., 1923]

STOUT — STUDIES OF LYTHRUM SALICARIA

443

Table I . Self -compatibility in Lythriim Salicaria to insect- pollination, and cases typical for
the results obtained for controlled hand- pollinations

Isolation, Insect-
pollination

S. no. 1

S. no. 2. (1918).
S. no. 2. (1919).

L. no. 1

L. no. 2

M. no. 2

to1*) [lower i!

2

o

17
161

65

53

3 ro

Si

3

Distribution of Seed per Pod. Class Groups i^to 10

17 10
IOO O

65 16

53 3

3000 300 14

many 71 7

few 31 3

few 40 6

7

34 30

47 2

47 3

202 48

43 12
16 3

20 3

" total

Hand-pollination

M. no. 1. (1919)- • 233

M. 5-1 no. 13 103

M. j no. 7 4>S

M. 1 no. 48 35

M. 1 no. 31 33

M. / no. 56 1 60

M. 5-1 no. 25 32

M. 5-1 no. 35 32

M. / no. 30 37

M.J no. 57
L. 4 no. 2 .
L. / no. 66.
L. 2 no. 18.
L. 1 no. 8. .
L. 5-5 no. 38

? 1500 142 16

40

55
68
50
49

24
S. 5 no. 5 14

134 '34
o

8 8
6
4
9
9
5
9
44

7

6

4

44

0

2

0

2

0

7

0 1

22

0

6

0

4

0

79
28;

5
1

o
o
o
o
c
1

2

[6

1
10 1

1 2

2 ! I

4 5

7 6

2 1

3 2

4 o

1 5

0 2

1 2

01

1 1 2 3 1

000

O I 2

I I

8 3

1

o 1

o 1

12 5

1 22

0 |0 I

3-28

23.67

2.96

4-85
8.91

18.25

24.61

7.00
20.33
25.00

28.33
50.88

49-5o
59.66

63- 1 5
6.00
6.00
7-42
8.90
20.00
14.00

capsules that could be referred definitely to differences in insect visitations
or to the constitution of the plants.

Mid-styled Plants Grown in Isolation. A plant of this form (M no. 2)
was sent to Mr. H. L. Skavlem, who grew it at Carcajou Point, Lake
Koshkonong, Wisconsin. Mr. Skavlem states that by the middle of July
this plant was "over four feet high with an abundance of bloom which con-
tinued for about six weeks." The plant bloomed from July 14, 1918, until
the first week in September. There were about 35 well -developed main
flower spikes ranging from 3 to 14 inches in length, and the total number
of flowers produced was estimated at 8,000. This description would apply
as well to any well-developed plant of any of the three forms.

This mid-styled plant was decidedly self-compatible in this isolation
test. At least 3,000 capsules matured on it in 191 7. A total of 300 cap-
sules were examined and counts were made of the seeds present, with

444 AMERICAN JOURNAL OF BOTANY [Vol. 10,

results as summarized in table I. The number of seeds per capsule ranged
to 121 ; twenty-eight capsules had more than 30 seeds each, but 202 capsules
contained less than 11 seeds per capsule, and 14 contained only rudiments
of seeds. The average number of seeds per pod for those that had good
seeds was 8.91. In 1918 the self-compatibility of this plant was quite the
same as in 1917.

In making the counts of seeds in 300 capsules (191 7 crop) of this plant,
position was taken into account. Each branch involved in the counts
was divided into three sections of equal length, here designated as the
lower third, middle third, and upper third. In the lower third of all branches
there were 10 capsules with more than 30 seeds each. It is clear that the
average number of seeds per capsule was lower toward the top of the
branches of this plant, but smaller capsules and fewer seeds per capsule are,
as a rule, to be expected toward the apex. Otherwise the plant was rather
uniformly highly self -compatible, and capsules with seeds were produced
in a considerable proportion of the flowers subjected to open pollination
throughout the entire period of bloom.

Another mid-styled plant (M no. j) was grown in my own home garden.
This plant made a vigorous growth and produced in 191 7 at least 6,000
flowers. At the close of the season it was found that relatively few of the
flowers produced pods during the first two thirds of the period of bloom,
but that later nearly every flower produced a pod.

In 1918 it was planned to test experimentally the behavior of this plant,
and especially to determine if the marked difference in production of fruit
involved changes in the innate ability of the plant to produce fruit or indi-
cated merely some difference in insect-pollination. A large, well-developed
long-styled plant was planted by its side. The two began blooming only
one day apart. The long-styled plant was allowed to bloom, thus affording
opportunity for free cross-pollinations by insects between the two, until
the 17th of August, when the long-styled plant was cut down. During the
time that both were in bloom nearly all flowers that opened on the mid-styled
plant developed fine pods, showing that the feeble production of pods during
the early part of bloom as observed in the previous year when the plant was
in isolation was not due to an impotence of the pistils. During the rest of
the season its performance was quite as in the previous year.

In 19 19 the long-styled plant was kept cut down so that no flowers were
produced by it, and a series of guarded self-pollinations were made on the
plant M no. 1. Branches were enclosed in glassine paper bags. When-
ever pollen from short stamens was used for pollinations, the flowers to be
thus pollinated were opened early in the morning and the long stamens
were removed, and then pollen from the short stamens was used later in the
day when the anthers were dehiscing.

Legitimate cross-pollinations, using pollen of flowers brought in vials
from the experimental plots over a mile distant, were made on 22 flowers

Oct., 1923]

STOUT

STUDIES OF LYTHRUM SALICARIA

445

during the first 18 days of bloom. The plant was in bloom 55 days. The
results obtained for selfing are summarized in table 1 and are shown in more
detail in table 2, in which, to facilitate ready comparison, the data are
compiled for three periods.

Table 2. Record for controlled pollinations of mid- styled plant no. 1 in summer of igig

1st to 1 8th Day

19th to 36th Day

37th to 55th Day

Selfed with pollen of long
stamens
Failures — no pods

70

18

(0 to 117) av. 34

35
46

(2 to 81) av. 32

9

70

(3 to 82) av. 32

Pods produced

Seeds per pod — range and
average

Selfed with pollen of short
stamens
Failures

33
I

37

2

20

(12 to 176) av. 98

57

1

10

47
1

Pods

Seeds per pod

1

Legitimate cross-pollination

Failures

Pods

Seed-range and average . .

Of the 140 flowers hand-pollinated with pollen from short stamens, only
three produced pods yielding 1, 10, and 37 seeds respectively, but these
may have been due to chance pollination with pollen from the long stamens
at the time of their removal or with pollen of other flowers which were
enclosed in the same bag. The results indicate that the plant remained
decidedly if not completely self-incompatible to pollen of its own short
stamens throughout the entire period of bloom.

When pollen of long stamens was used in hand-pollinations, the propor-
tion of pods produced increased as the season advanced, and during the
last 18 days of bloom there were but 9 failures out of 79 flowers pollinated.
The results show conclusively that the self-compatibility in this plant
involves fertilizations from the pollen of long stamens, and also that this
compatibility actually increases toward the end of the period of bloom.
The increase in compatibility affects, however, the number of pods that are
formed rather than the number of seeds in a pod. The average number of
seeds in the pods that were produced was almost the same for all periods,
but the highest number of seeds in any pod was obtained during the first
iH days of bloom.

The 22 legitimate cros>e> made (hiring the first 18 days gave 20 pods
with seeds ran^in^ in number from 12 to 176, with an average of 98 seeds
per capsule. These results show conclusively, as do those of open cross-
pollinations during the early part of the period of bloom in the previous
year, that the pistils of the plant are highly potent during the period of
marked self-incompatibility. The decided change in fruit-production is

446 AMERICAN JOURNAL OF BOTANY [Vol. 10,

hence due to a change in the physiological relations between pistils and the
pollen of long stamens. Pollen from both short stamens and long stamens
was examined at frequent intervals throughout the entire period of bloom;
there was very little abortion, the pollen was successfully germinated in
artificial media, and the use of such pollen in legitimate crosses on several
dates during the first 30 days of bloom was almost invariably highly effective
in pod- and seed-production. There were in this plant no noticeable
evidences of impotence of stamens and anthers, such as are to be seen in
some plants of this species.

The branches left to open-pollination produced pods quite as in the year
191 7. Several of the main branches were selected and divided into thirds,
and seeds in all pods in the lower two thirds were counted; then an equal
number of pods from the many pods in the uppermost third were taken at
random for counts, the entire number, 142, being as near the number of
pods secured from the selfing by hand in which the pollen of long stamens
was used as was possible. As shown in table 1, the range for number of
seeds per pod and the average were higher for the lower two thirds than in
the last 'third, showing that the change in compatibility involves number of
pods formed rather than number of seeds per pod. Comparison shows that
the controlled pollinations in which pollen from long stamens was used were
somewhat more successful than free open-pollinations, when judged by the
average number of seeds produced.

A third mid-styled plant (M no. 3) was grown in isolation in the New
York Botanical Garden at a distance of about one mile from the location of
the plant M no. 1. This plant proved to be decidedly less self-compatible to
open self-pollination than were M no. 1 and M no. 2, but, as in M no. 2,
there was quite the same proportion of pods produced throughout the entire
season of bloom, no change in compatibility being evident as observed in
the plant M no. 1.

Summary. In these tests seeds were obtained to illegitimate self-
pollination of plants of all three forms. Mid-styled plants were most
highly self-compatible in respect to the number of pods produced. The
pods found were distributed irregularly over the entire period of bloom
except for one plant (M no. 1), in which it was found that there was an actual
change in the degree of self-compatibility to pollen from its own stamens.
The results from year to year have been very uniform for all plants, except
for the plant S no. 2. Its feeble production of pods in 1918, followed by
the production of 161 pods in 19 19, suggests that the plant is rather strongly
self-compatible provided insects make the proper pollinations. It may
readily be conceded that the kinds of insects that can most efficiently self-
pollinate long-styled and mid-styled plants are not the ones which best self
the short-styled plants.

Self-pollination does not appear to be uniformly as efficient in seed-
production as are certain legitimate cross-pollinations, at least for the one

Oct., 1923] STOUT STUDIES OF LYTHRUM SALICARIA 447

mid-styled plant M no. I (see table 2). Of the 22 flowers on it that were
crossed during the first 18 days of bloom there were only two failures,
the highest number of seeds for a capsule being 176 and the average 98.
Whether such pollination would be more effective than selfing during the
last part of the period of bloom was not tested.

Self-compatibility Tests by the Bagging Method

A total of about 600 plants have been grown in pedigreed cultures from
seed. A considerable number of these have been tested for self -com-
patibility in the following manner: branches were enclosed in glassine paper
bags, and pollinations of flowers opening within were made from day to day.
Long-styled plants and short-styled plants were selfed by using the pollen
from mid-length stamens, and the mid-styled by use of the pollen of long
stamens. In making pollinations, stamens with dehiscing anthers were
removed with sterile tweezers and brushed on pistils, leaving an abundance
of pollen. In cases of pseudo-proterogyny the fully protruding pistils of
partially opened flowers were likewise treated along with pistils of more
mature flowers. It appears that in the decidedly pseudo-proterogynous
flowers the pistils protrude long before they are receptive to any kind of
fertilization, and that highest seed production in compatible fertilizations
occurs when pollination is made at or about the time that petals open and
pollen is shed. On plants two or more years old, a total of as many as 500
or more flowers were often thus pollinated. On plants in the first year of
growth from seed the number thus selfed was often much less.

These tests are undoubtedly less adequate than tests in isolation for
revealing feeble grades of self-compatibility and in showing such changes in
self-compatibility as are seen in the plant M no. 1, but hand-pollinations
make certain that pollen in abundance is applied to the stigmas at the time
when they are judged to be most receptive.

The general results summarized for each form without reference to
lines of descent are as follows:

Fully Self- Feebly Self- Medium Self- Highly Self-

incompatible compatible compatible compatible

Mid-styled plants 64 20 21

Long-styled plants 83 14 o

Short-styled plants .... 22 1 o

An attempt has been made to grade the plants according to whether the
self-compatibility is feeble, medium, or strong, the judgment being based on
the proportion of selfed flowers that gave pods and the number of seeds
produced. Results typical for various grades are given in table 1. The
tests made show that many plants of the species are without doubt entirely
self-incompatible (M 5-1 no. 13 in table 1 for example), and that others are

448 AMERICAN JOURNAL OF BOTANY [Vol. 10.

highly self-compatible (31 I no. 57), with various intervening Y'
grouping of which is neither definite nor accurate. ^mens

In these tests mid-styled plants have been more highly sen -»tOom;
than plants of the other forms. This is true both in relative i. t in
produce fruit to selfing, and in the range to higher grades of jeveral

one plant, 31 1 no. 57, every flower that was selfed produced a \ -ctive

other pods were produced in other flowers that spontaneous!- Sf\\ iifle

enclosed in a glassine bag.

Of the 97 long-styled plants tested in hand-pollinations, only 14 produced
pods, and not one gave over 30 seeds in any pod. In all of ^hese the self-
compatibility was apparently of a weak grade.

Twenty-three short-styled plants were tested, and only one produced
seeds.

The results obtained in the controlled self-pollinations with these plants
agree in general with those obtained in isolation tests. A rather large
proportion of mid-styled plants are self-compatible in some degree, and
nearly half of the plants of this form produced pods containing viable seeds
to selfing, and a few were highly self-compatible. There has been no
difficulty in finding mid-styled plants to use as parents of self-fertilized
lines of progeny. Relatively few long-styled plants produced pods to self-
ing, and in all such plants the self-compatibility was feeble, few pods being
produced and these having few good seeds. Short-styled plants have as a
class been decidedly self -incompatible, and of the seedlings tested only one
has produced seeds to selfing. The high seed-production seen in the plant
5 no. 1 in 1919 was not duplicated by the plant in 1918 nor in 1920 and 192 1..
There has not been opportunity to test this plant by controlled hand-
pollinations as the plants grown from seed have been tested.

Summary

1. Many plants of Lythrum Salicaria are capable of producing capsules
and viable seeds to illegitimate self-pollination brought about either In-
controlled hand-pollination or by insect-pollination in the field. The
capacity for self-fertilization still lingers strongly in the species.

2. The proportion of self-compatible plants is greatest in the mid-
styled plants, in which also the highest grades of self-compatibility are to
be seen. Long-styled plants are, as a class, less self-compatible, and the
short-styled plants are still less so. The three forms appear to differ in the
capacity for self-compatibility.

3. There are wide variations in the degree of self-compatibility. In
the most highly self-compatible form, the mid-styled, there are all grada-
tions between complete self-incompatibility and the highest grade of
self-compatibility.

4. The variations in the physiological condition of the sex organs, as

Oct., .„. ] STOUT — STUDIES OF LYTHRUM SALICARIA 449

exhibited in selfing, suggests that wide variations may likewise be expected
rrna-fc^ t -ings even for those that are legitimate.

crossec ie noticeable case of end-bloom self-compatibility was found. This

the hi' ..mid-styled plant and involved only fertilizations from pollen of

Whetht .ilong stamens.

last par- physiological relations of the sex organs in plants of this tri-

jpecies exhibit quite the same range of variations as are seen in

ui'du. } amorphic species.

Conclusion

For the species Lythrum SaUcaria the evidence of wide variation in the
degree of self-incompatibility is definite. The physiological differentiations
of the sex organs are incompletely correlated with the apparent structural
adaptations for cross-pollination; they are not fixed, constant, and fully
achieved either in expression or in heredity, but are fluctuating and inter-
grading. They still present opportunity for further selection either toward
greater or toward less restriction of fertilization.

The persistence of self-compatibility in various degrees of expression,
and the apparent difference in respect to self-compatibility seen among the
various forms, present strong evidence that self-compatibility was the
antecedent condition in the species out of which the present complex of sex
relations is still evolving, just as the sets of styles and stamens of different
lengths have been developed out of an original homomorphic species.

New York Botanical Garden

LITERATURE CITED
Darwin, C. 1865. On thejsexual relations of the three forms of Lythrum SaUcaria. Jour.

Linn. Soc. Bot. 8 : 169-196.
, 1877. The different forms of flowers on plants of the same species.

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN"

No. 249

The Physiology of Incompatibilities

A. B. STOUT

NEW YORK
1923

Reprinted, without change of paging, from American Journal of Botany, 10: 450-461. N 102.'|

Reprinted from the American Journal of Botany, X: 459-461, November, 1923.

L1BKART

NEW YtJRK

BOTANICAL

UAKDeN

THE PHYSIOLOGY OF INCOMPATIBILITIES1

A. B. Stout

In general survey, physiological incompatibilities in fertilizations include
two groups of phenomena :

1st. There are the very general and characteristic failures of cross-
fertilization between different species, long considered as the most adequate
evidence of specific distinction ; and

2d. There are those physiological limitations to free and general
fertilization within species exhibited best in the failure of certain homo-
morphic hermaphrodites to self-fertilize, but also in the cross-incompati-
bilities among seed-grown individuals of the same species or race.

Certain aspects of the physiology of these incompatibilities are clear.
They exist and are in operation when the sex organs and sex elements are
in a condition for proper fertilization; the elements do not function in
certain relations but do in others; fertilization is discriminative.

It is not, therefore, a question of what brings the spores or gametes to
ripening, or of the mere production of those egg or stylar secretions or
chemotactic influences which make fertilization possible. It is rather a
question of a very special kind of development or physiological condition
which discriminates between fertilizations when they are otherwise possible.

Inter-specific incompatibilities are very generally considered to involve
species specificity. They are expressed in the interaction between egg
secretions and sperms, in mechanical and chemical resistance of the cortical
layers of eggs to the entrance of sperms, in the extrusion of sperm chromatin
after fertilization, in the death of the heterogenetic hybrid, or in the sexual
impotence of such hybrids. In all these ranges of expression the incom-
patibilities appear to be deep-seated and inherent in the physical and
chemical differences in the organization of species.

Turning now to intra-specific incompatibilities, there is apparently
a very different physical basis. Here there is self-incompatibility involving
the germ cells of a single individual. Here also there is cross-incompati-
bility between individuals of the same parentage. Such cases are frequent
among homomorphic hermaphrodites. They are so widely distributed in
the families of flowering plants that it would seem that there must be some
1 Read in the symposium on "Sterility in Plants," at the joint meeting of Section G
of the American Association for the Advancement of Science, the Botanical Society of
America, and the American Phytopathological Society, at Cambridge, December 27, 1922.

459

460 AMERICAN JOURNAL OF BOTANY [Vol. io,

fundamental principle operating in their origin and development. In
these higher plants there is the development of a nearly naked egg (free of a
decided membrane) imbedded in the tissue of the ovule, and a highly
specialized male gametophyte— the pollen tube — with its various growth
relations in the style. While fertilization in these cases involves a series of
physical and chemical relations, it is fundamentally an egg-and-sperm
reaction.

It is to be noted that studies of the physiology of pollen germination
have failed to reveal a specificity that is comparable or related to the com-
patibility or incompatibility found in homomorphic species. It has, how-
ever, long been known that in cases of self-incompatibility the pollen tubes
often make feeble growth in the style, and more recently it has been shown
that their rate of growth is not accelerated as the eggs are approached.
This condition apparently persists when incompatible pollen is mixed with
compatible pollen and the tubes of both grow side by side in the style.
This indicates that the reactions are decidedly discriminative and suggests
that they involve reactions of the pollen tubes to secretions of the eggs.
There is, however, evidence that in many grades of self-incompatibility the
injurious effects may be exhibited after what is apparently a successful
fertilization.

One aspect of the development of self-incompatibility in the hermaphro-
dite is clear. Every such individual is itself the result of a compatible
fertilization in which two cells fused, and the two elements of the resulting
diploid organization show themselves compatible throughout the somatic
life of the individual; but the two kinds of sex elements produced by it are
incompatible, and this is of course irrespective of their germ-plasm complex.
The incompatibility arises along with sex differentiation, which in seed
plants begins with the development of pistils and stamens and is independent
of any readjustments of the germ plasm in the reduction divisions.

The biogenetic nature of the development of self-incompatibilities is
further indicated by the wide variations which exist in their expression in
individuals. Between the extreme or alternative conditions there are many
intergradations, and the extreme conditions are reversible in a progeny.
This is the general rule of behavior in such pedigreed cultures as have been
critically tested in this particular. Cross-incompatibilities exhibit quite
the same ranges of expression, and here reciprocals may give directly
opposite results.

But there is also conclusive evidence in certain species of cyclic changes
in the self-compatibility of an individual. These are best seen in plants
which have a somewhat extended and continuous period of bloom. In
some species there is self-compatibility at the end of the period of bloom ;
in others the climax of self-compatibility is at the mid-period of the bloom,
and for certain perennials there is some evidence of changes from year to
year in relation to the age of the plant. It is to be noted that a mid-period

Nov., 1923] STOUT — PHYSIOLOGY OF INCOMPATIBILITIES 461

self-compatibility is preceded and followed by self-incompatibility. There is
alternative expression of extreme grades of compatibility and incom-
patibility in the series of flowers produced in succession on the same branch.
The functions of fertilization are in such cases operating in a cycle of
intensities.

It may here be reported that, in the species Brassica pekinensis, self-
incompatibility of a plant as a whole or of a family of plants may be de-
cidedly changed by a cultural treatment which reduces vegetative vigor.
In a family of this species grown for three generations, less than 10 percent of
the total of 326 plants were highly self-compatible and there was no hered-
itary effect of selection for self-compatibility. When a generation of this
family was grown in small pots with decided reduction of vegetative vigor,
of a total of 1,128 plants there were 734 (65%) that were highly self-com-
patible, and only 22 (less than 2%) were self -incompatible. Furthermore,
a large proportion of the former were self-compatible in the earliest flowers
that opened. The family was decidedly changed in regard to the number
of plants that were self -compatible, and in the individuals the characteristic
cycle was altered. Such results, together with the other behavior noted
above, indicate that there is a direct and very decided physiological correla-
tion between vegetative vigor and the functional properties of the organs
concerned with fertilization.

This is at least suggestive that the physiological conditions which
restrict and limit indiscriminate fertilization within species are not only
subject to internal regulation, but that in some cases at least they are corre-
lated with changes in vegetative vigor.

The situation gives hope that the cyclic expression of sexual affinities
and the development of extremes of compatibility may be so regulated
experimentally that the specific biogenetic factors and conditions operating
in this highly specialized differentiation may be determined.
New York Botanical Garden

CONTRIBUTIONS FROM THE NEW YORK BOTANICAL GARDEN

No. 250

Genera of North American Fabaceae — I

P. A. RYDBERG

Reprinted from American Journal of Botany, io: 485-498. N 1923

V YORK
A NIC At.
UAKDfiM

Reprinted from the American Journal of Botany, X: 485-498, November, 1923.

GENERA OF NORTH AMERICAN FABACEAE
I. TRIBE GALEGEAE

P. A. Rydberg
(Received for publication December 16, 1922)

The tribe Galegeae has been divided since Bentham's time into seven
subtribes. Of these one has already been raised to tribal rank, Psoralieae,
which is characterized by the glandular-punctate foliage and by the pods,
which are one- or few-seeded, usually indehiscent but rarely breaking open
irregularly across the middle, never valvate. Another subtribe should be
removed as a tribe Indigofereae, having three characters seldom found
elsewhere in Fabaceae and never combined in any of the tribes. These are
malpighiaceous hairs (appressed hairs, 2-branched with branches forming a
straight line, the hairs therefore apparently attached at their middle),
appendaged connectives in the anthers, and lateral spurs on the keel-petals.
It may be that Glycyrrhiza and one or two genera should be removed also
and form another tribe, on account of the indehiscent (though several-
seeded) glandular fruit and confluent anther cells.

After the removal of the two subtribes mentioned, the tribe may be
characterized as follows:

Herbs, shrubs, or trees, or woody vines, without glandular dotted
foliage (except in Glycyrrhiza). The leaves are odd-pinnate or abruptly
pinnate, with or without stipels, very rarely with malpighiaceous hairs.
The calyx is campanulate to tubular, 5-toothed or 5-lobed, rarely more or
less 2-tipped. The corolla is papilionaceous, with 5 petals, the two keel-
petals more or less adnate along the lower margins. The stamens are 10,
diadelphous or monadelphous, the connective without appendages. The
ovary is usually many-ovuled, more or less curved or bent upwards; the
style slender and the stigma small. The fruit is several- or many-seeded,
2-valved, usually dehiscent.

As stated before, Bentham and Hooker acknowledged 7 subtribes.
These authors were closely followed by Taubert in Engler and Prantl's
Natiirlichen Pfanzenfamilien. Taubert's key in translation reads as
follows, after the Indigofereae and Psoralieae have been removed:

a. Seeds strophiolate. Flowers in 2's in the leaf-axils or forming a

terminal raceme (compare Tephrosia and Fordia). c. Brongniartiinae.

/3. Seeds not strophiolate.

I. Inflorescence racemose, terminal, opposite the leaves, or on
the branches paniculate, more rarely in the axils of the
upper leaves or all peduncles or only the lower paired or
clustered in the axils. Stipe inside the staminal sheath
sometimes surrounded by a small cup-like disk. d. Tephrosiinae.

485

486 AMERICAN JOURNAL OF BOTANY [Vol. 10,

II. Inflorescence always axillary. Stipe without disk at the
base.
Pod flat, when inflated at least the endocarp flat, 2-valved. e. Robiniinae.
Pod inflated or bladdery, rarely flat, then 2-celled lengthwise.
Style bearded above; pod inflated or bladdery, not
dehiscent or opening only at the apex, more rarely 2-
valved, never with a longitudinal partition. /. Coluteinae.

Style naked, more rarely hairy like a brush around the
stigma; pod 2-celled or nearly 2-celled lengthwise,
seldom 1 -celled. g. Astragalinae.

From this key it is evident that the subtribes are very artificial and that
it has been very hard to find really good distinctive characters. It is also
evident that these characters have been drawn without taking into con-
sideration the American genera in full, in which they break down repeatedly.

While the strophiole is very well developed in the Brongniartiinae, in
fact better so than in any other group in the Fabaceae, strophioles are
present, although small, in several species of Cracca.

The Tephrosieae of Bentham (Tephrosiinae of Taubert) is in itself
not a natural division, for Barbiera is not at all closely related to the rest,
being distinguished by the long-clawed petals, the presence of 2 bractlets
beneath the tubular calyx, and the style bearded along the upper side. In
fact, it has no close relative in the whole tribe. The paired bractlets are
found in Sesban and related genera and in Diphysa, which are all included
in the Robinieae of Bentham (Robiniinae of Taubert).

Neither does Krauhnia seem to be closely related to Cracca or Galega,
but apparently forms a natural group with several Asiatic genera, especially
Millettia.

The position of the inflorescence, being either terminal or apparently
opposite the leaves, a character very important in the minds of both Ben-
tham and Taubert, seems to be of little value, for in many species of Galega
and a few of Cracca (see explanation under Cracca) the inflorescence is
axillary, just as in the Coluteinae and Astragalinae. The other character,
the presence of the disk at the base of the staminal sheath, is also unreliable,
for, while it is fairly well developed in Cracca and Krauhnia and their
relatives, it is not at all in Barbiera.

The Robinieae of Bentham (Robiniinae of Taubert) is distinguished
from the Coluteae and Astragaleae (Coluteinae and Astragalinae of Taubert)
by the one-celled, two-valved, flattened, not inflated pod, but in the genus
Diphysa, included in the first, the exocarp is strongly inflated, forming two
lateral elongate bladders, and in Homalobus and Kentrophyta, segregates
of Astragalus, the pod is decidedly compressed and fulfills all the characters
ascribed to the Robinieae. In Bentham and Hooker's Genera Plantarum
the following remarks are made: "In Astragalus paucis [legumenj planis-
simum est, sed membranaceum et perfecte septatum." This statement
applies to Hamosa Med., a segregate of Astragalus found both in the Old

Nov., 1923]

RYDBERG NORTH AMERICAN FABACEAE

487

Subtribe 3. Brongniartianae.

World and America, in which the pod is flat and 2-celled. In Atelephragma
Rydb., another segregate found only in America, the septum is incomplete,
while in Homalobus and Kentrophyta, both exclusively American, there is
not a vestige of a partition and the fruit is exactly like the fruit of Poitoea,
Coursetia, or Corynella, placed in the Robinieae.

' Both Bentham and Hooker, and Taubert seem to have ignored or else
knew imperfectly these segregates of Astragalus when characterizing their
subtribes. In Benthamantha, also placed in the latter subtribe, the pod is
also similar but with false transversal partitions. In Sesban and Dauben-
tonia, both referred to Robinieae, the pod is hardly compressed, in the
former terete or nearly so, in the latter 4-angled. There is therefore no
definite line between these three subtribes as constituted by Bentham and
Hooker and by Taubert, nor are the Brongniartiae or the Tephrosieae
very distinct.

The classification must therefore be remodeled along other lines. I
shall give here a tentative reclassification, which will probably be modified
when more study has been made on the Astragalinae.

Seeds strophiolate; embryo with a straight radicle;
flowers 1 or 2 in the axils of the leaves or in terminal
racemes or panicles; calyx subtended by a pair of
deciduous bractlets; pods flat; trees or shrubs.
Seeds not strophiolate (rarely slightly so in species of
Cracca) ; embryo mostly with an incurved radicle;
flowers mostly racemose.
Bractlets 2 under the calyx; pods internally with
more or less distinct false cross-partitions.
Calyx tubular, with long lobes; petals long-
clawed, their blades cuneate at the base, that
of the banner oblong, not spreading; pods flat,
many-seeded, dehiscent; shrubs with odd-
pinnate leaves.
Calyx campanulate, with short tooth-like lobes;
petals short-clawed, at least the blades of
the wings with basal auricles; blade of the
banner sub-orbicular or broadly ovate;
pods not flattened or if flat 2-seeded
(Glottidium).
Exocarp of the pod not inflated; hypanthium
obsolete, not differentiated from the calyx;
shrubs or trees, with abruptly pinnate
leaves.
Exocarp of the pod inflated, forming two
elongate bladders, one on each side of the
pod; hypanthium well developed and dif-
ferentiated from the calyx, obconic; shrubs
and trees, with odd-pinnate leaves.
Bractlets under the calyx wanting; pods usually
without cross-partitions, except in Bentha-
mantha, Sphinrtospermum, and Hebestigma.

Subtribe 4. Barbierianae.

Subtribe 5. Sesbanianae.

Subtribe 6. Diphysanae.

488

AMERICAN JOURNAL OF BOTANY

[Vol. io,

Base of the pistil or of its stipe usually surrounded
by a more or less developed saucer-shaped
disk within the staminal sheath; racemes
terminal, or terminal and in the upper
axils, or opposite the leaves, or if axillary,
the leaflets with parallel oblique lateral
veins.
Leaflets with parallel or indistinct lateral
veins; pods obliquely striate; leaves with-
out stipels; herbs or low shrubs.
Leaflets netted- veined; pods not obliquely
striate; leaves mostly with stipels; trees or
(ours) woody vines.
Base of the pistil or its stipe not surrounded by a
disk; racemes always axillary; leaflets not
with parallel oblique lateral veins.
Pods flat, or if terete with prominent sutures,
neither inflated nor with even a vestige
of a longitudinal partition.
Flowers in fascicles, on short branches
arising in the axils of fallen leaves;
leaves abruptly pinnate; banner usually
enveloping the other petals.
Flowers in axillary racemes; leaves odd-
pinnate, except in some species of
Coursetia and occasionally in Olneya;
banner mostly spreading.
Pods more or less inflated or completely or
partially 2-celled, by the intrusion of
one or both sutures.
Style bearded along the upper margin;

pods always I -celled and inflated.
Style glabrous, or bearded only around
the stigma.
Anther-cells not confluent; pods not
prickly; plant not glandular-dotted.
Anther-cells confluent at the apex;
pods in ours prickly; plant glandu-
lar-dotted.

Subtribe I. Craccanae.

Subtribe 2. Millettianae.

Subtribe 7. Corynellanae.

Subtribe 8. Robinianae.

Subtribe 9. Coluteanae.

Subtribe 10. Astragalanae.

Subtribe 11. Glycyrrhizanae.

Subtribe i. Craccanae

Herbs or shrubs, with alternate odd-pinnate leaves, persistent stipules,
but no stipels, the leaflets usually with parallel oblique lateral veins. The
flowers are racemose. The calyx is campanulate 5-lobed, but the upper
lobes are often more or less united. The corolla is truly papilionaceous,
the petals are more or less clawed, the banner is broad and spreading. The
stamens are monadelphous or diadelphous, inserted on a more or less de-
veloped campanulate disk. The style is glabrous throughout or more or
less hairy around the stigma. The legume is elongate, flat, 2-valved,
usually obliquely striate. The seeds are reniform or transversely oblong,
i.e., their longer axis parallel to the axis of the pod.

Nov., 1923] RYDBERG NORTH AMERICAN FABACEAE 489

The subtribe contains, besides the following three genera, a few confined
to the Old World. Cracca is cosmopolitan of warmer regions, Peteria is
endemic American, and Galega is Eurasian, introduced in the New World.

Stipules not spinescent; lateral veins of the leaflets prominent.

Upper filaments wholly united with the staminal sheath, forming a closed
tube; banner in ours glabrous. 1. Galega.

Upper filament free, at least at the base; banner strigose on the back. 2. Cracca.

Stipules spinescent; upper filament free; lateral veins of the leaflets obsolete. 3. Peteria.

i. Galega [Tourn.] L. Sp. PI. 714. 1753

Perennial herbs. The leaves are odd-pinnate, with semi-sagittate
stipules. The flowers are in axillary or terminal racemes with narrow bracts
and no bracteoles. The calyx has 5 subequal lobes. The corolla is white
or light blue; the banner is obovate-oblong, narrowed below into a very
short claw; the wings have an oblong blade with a prominent basal auricle,
and a longer claw, and are adherent to the keel at the middle; the keel-
petals are obtuse, more or less arcuate, longer than the wings, and united
nearly their whole length. The filaments are monodelphous, i.e., all united
into a sheath. The ovary is sessile, many-ovuled; the style glabrous; the
stigma small, terminal. The pod is linear, terete, 2-valved, sometimes
constricted between the seeds. Seeds are transversely oblong, without
strophiole.

Illustration: Plate XXXIII A. Galega officinalis L., X 2/3; 1. calyx,
2. banner, 3. wing, 4. keel-petal, 5. staminal sheath, 6. pistil, X 2; 7. pod,
X 1 ; 8. cross section of pod, 9. seed, X 2.

In the Species Plantarum, the genus Galega contained only one species,
Galega officinalis L., which therefore is the type.

Synonyms :

Callotropis G. Don. Gen. Syst. 2: 228. 1832. Type: C. tricolor (Hook.)
G. Don., based on Galega tricolor Hook., which is supposed to be the
same as G. officinalis L.

Accorombona Endl. Gen. 1427. 1841. This was a substitute for Callo-
tropis G. Don., not Calotropis R. Br. 1809. Hence the same type.

The genus consists of 4 or 5 species native of southern Europe and the
Orient. Of these, G. officinalis is sometimes cultivated as a forage plant
and in olden times was used in medicine. It has been found occasionally
in the western states from Kansas to Utah, as an escape from cultivation
or introduced incidentally among seeds. The genus is closely related to
Cracca, differing mainly in the monadelphous stamens. The racemes are
mostly axillary, and therefore the genus is, according to Bentham and
Hooker, anomalous in their subtribe Tephrosieae, but, as will be shown,
this abnormality is found even in species of Cracca.

2. Cracca L. Sp. PI. 752. 1753

Herbs, often woody below, or shrubs. The leaves are odd-pinnate, the
leaflets striate, with veins oblique to the midrib and parallel; the stipules

490 AMERICAN JOURNAL OF BOTANY [Vol. 10,

are setaceous, or broader and striate. The flowers are racemose, the
racemes are either terminal, Math or without smaller racemes in the upper
axils, or apparently opposite the leaves, or rarely axillary. The individual
flowers are usually in clusters of 2-6 at each node of the peduncle. The
calyx is campanulate, furnished with a more or less developed disk; the
lobes are five; either subequal or the lowest one longer, the upper two usually
more or less united. The petals are clawed; the blade of the banner is sub-
orbicular, more or less sericeous without; the blades of the wings are oblong
or obliquely obovate, slightly coherent with the keel, with a more or less
developed basal auricle; the keel-petals are more strongly lunate with a
larger basal auricle. The stamens are usually partly monadelphous, the
uppermost filament at first free from the staminal sheath at the base, adnate
to it at the middle, and often separating from it later. The ovary is sessile,
many-ovuled, the styles inflexed or incurved, somewhat horny at the base,
most commonly glabrous, the stigma terminal. The pod is sessile, flat,
2-valved, many-seeded.

Illustration: Plate XXXIII B. Cracca virginiana L., X 2/3; 1. calyx,
2. staminal sheath, 3. pistil, 4. banner, 5. wing, 6. keel-petal, 7. pod, X i;
8. pod in cross-section, 9. seed, X 2.

The genus was first established in Linnaeus' Flora Zeylanica 1747. The
first species described both in this book and in Species Plantarum is C. villosa,
which must be regarded as the type.

Synonyms :

Brissonia Neck. Elem. 3 : 36. 1790. No type was given, but the genus was
based on "some species of Galega L." [Necker's reference indicated
the 14th edition of Linnaeus' Systema Vegetabilium.] The first botanist
to assign species in the genus was Desvaux (Jour. Bot. 3: 78. 1814),
who proposed B. trapesicarpa, B. stipularis, and B. coronillaefolia, but
none of these are found in Linnaeus' work named above. De Candolle
(Prod. 2: 249. 1825) adopted Brissonia as a section under Tephrosia.
Of this section apparently T. toxicaria Pers. should be regarded as the
type.

Reineria Moench, Meth. Suppl. 44. 1802. Type: R. reflexa.

Thephrosia Pers. Syn. PI. 2: 328. 1807. It was based on 39 species, with-
out definite type. The first species is T. filifolia, but the type ought to
be sought in the second and larger division with pinnate leaves.

Kiesera Reinw. Syll. PI. Nov. 2: 11. 1828. Type: K. sericea Reinw.,
which is supposed to be the same as Tephrosia Candida DC.

Apodynomene E. Meyer, Conn. PI. Afr.Aust. in. 1835. Type: E. grandi-
flora (Pers.) E. Meyer, based on Tephrosia grandiflora Pers. This is
the first and best known species.

Catacline Edgew. Jour. As. Soc. Beng. 16: 12 14. 1847. Type: C. sericea
Edgew.

Macronyx Dalz. Hook. Kew Jour. 2 : 35. 1850. Type: M. strigosus, which
is close to if not identical with Tephrosia tenuis Wall. Perhaps this
should be excluded from the synonyms of Cracca.

Nov., 1923] RYDBERG — NORTH AMERICAN FABACEAE

491

Balboa Liebm. Vidensk. Meddel. 1853: 106. 1854. Type: Balboa

diversifolia Liebm., the only species.
Seemannantha Alef. Bonplandia 10: 264. 1862. This was a substitute

for Macronyx, and hence based on the same type.

Cracca is a large genus, probably containing about 150 species, found
in warmer regions of both hemispheres. In North America (including
Central America and the West Indies), there are 72 species, of which 65 are
native and 7 introduced.

Taubert divided the genus into 4 sections: Brissonia, Reineria, Pogo-
nostigma, and Requienia. Of these the last two, which are not represented
in America, should be removed as distinct genera, having 1 -seeded fruits.

The principal character by which Taubert distinguished the first two
sections was the relative width and length of the calyx lobes. Some species,
as for instance C. lencantlia, which has long and narrow calyx-lobes, really
belong to Brissonia instead of Reineria. A better distinction is the in-
florescence, which is monopodial in the former and sympodial in the latter.
In the monopodial inflorescence the terminal raceme is best developed, but
many of the species bear also secondary racemes in the upper axils. In
the sympodial inflorescence the terminal raceme develops first, but in the
uppermost leaf-axils a branch is produced which in its turn develops a
terminal raceme; this is repeated several times, and the racemes therefore
appear to be opposite the leaves. In a few species, as for instance C.
rhodantha, C. foliosa, C. vicioides, and C. Brandegei, the racemes are mostly
at the nodes, but neither opposite the leaves nor strictly axillary; they are
inserted in the axils but obliquely, i.e., not in the plane determined by the
stem and the rachis of the leaves. The monopodial or sympodial character
is not perfectly clear.

Several of the species are used as a fish-poison by the natives of the
region where they grow, others for poisoning arrows in Southern Africa;
still others furnish a blue dye, somewhat resembling indigo.

3. Peteria A. Gray. PI. Wright 1 : 50. 1852

Perennial herbs, somewhat woody at the base. The leaves are odd-
pinnate with spiny stipules. The flowers are usually in terminal racemes.
The calyx is cylindro-campanulate, gibbous at the base above; 5-lobed with
the upper two lobes united high up. The corolla is ochroleucous or nearly
white, the banner long-clawed with an oblong-obovate blade; the wings
have an obliquely oblong blade, slightly auricled at the base, and a slender
claw; the blades of the keel-petals are obliquely obovate, obtuse, with a
broad, rounded basal auricle. The stamens are diadelphous, with the upper
filament wholly free. The ovary is short-stipitate or sessile, many-ovuled,
the style with a horny base, inflexed. glabrous, except at the apex, where
there is a ring of hairs. The pod is linear, straight, compressed, 2-valved,
many-seeded, with thick sutures.

The genus was based on a single species, P. scoparia A. Gray, and con-

492 AMERICAN JOURNAL OF BOTANY [Vol. 10,

sists of three species of southwestern United States and northern and
central Mexico. In habit the species resemble some species of Astragalus
and its segregates, but the racemes are not axillary but terminal, or when
an axillary branch is developed become apparently opposite the leaves as
in Cracca. The spinescent stipules constitute also a distinctive character.
The genus is evidently related to Cracca, but the hair-tufts around the
stigma and the less distinct veining of the leaflets obscure the relationship.
Illustration: Plate XXXIII C. Peteria scoparia A. Gray, X 2/3;
1. calyx, X 2; 2. stamens, 3. pistil, 4. banner, 5. wing, 6. keel-petal, X 1;
7. pod of P. glandulosa, X 2/3; 8. cross section of the same, X 2; 9. seed,
X 2.

SUBTRIBE 2. MlLLETTIANAE

Trees or woody vines with alternate odd-pinnate leaves and persistent
stipules, usually also with stipels. The calyx is campanulate. 5- toothed,
but the lobes are often minute, or the upper 2 and the lower 3 more or less
united, forming an upper and a lower lip. The corolla has short-clawed
petals, the banner being broad, spreading or reflexed. The stamens are
monadelphous or diadelphous, the upper filament free at least at the base.
The pod is fiat, 2-valved, elongate, several-seeded. Seeds mostly reniform.

Besides the following genus which is represented by native species in
the eastern United States and eastern Asia, the subtribe consists of four or
five Asiatic and perhaps two African genera.

4. Kraunhia Raf. Med. Rep. N. Y. II, 5: 352. 1808

High-climbing woody vines wTith odd-pinnate leaves, small stipules and
stipels. The flowers are in terminal racemes, with deciduous bracts. The
calyx is more or less 2-lipped, the upper lip with 2 broad teeth united to near
the apex, the lower with 3 longer teeth. The corolla is blue or purple, rarely
white, the petals are subequal in length ; the banner has a suborbicular blade,
reflexed, and with 2 callosities or appendages, the claw is short; the blades of
the wings are obliquely obovate, falcate with a large basal auricle on the
upper edge and often a smaller one on the lower; the keel-petals are clawed,
united at the apex, the blade is lunate with a sharp basal auricle. The
stamens are diadelphous, the upper stamen is free or slightly adherent at
the middle. The ovary is stipitate, many-ovuled, glabrous; the style
inflexed, glabrous; the stigma small. The pod is elongate, flat, 2-valved.
The seeds are reniform, without strophiole.

Illustration: Plate XXXIII D. Kraunhia frutescens (L.) Greene,
X 2/3; 1. calyx, 2. stamens, 3. pistil, 4. banner, 5. wings, 6. keel-petals, X 1 ;
7. fruit, 8. the same in cross section, X 2/3.

The genus was established on Glycine frutescens L. without a diagnosis.

Synonyms :
Diplonyx Raf. Fl. Ludov. 101. 1817. Type: D. elegans Raf., which is

regarded as the same as Kraunhia frutescens (L.) Greene.
Thysanthus Ell. Jour. Acad. Phila. 1: 371. 1818. Type: T. frutescens

(L.) Ell., based on Glycine frutescens L.

Nov., 1923] RYDBERG NORTH AMERICAN FABACEAE 493

Wisteria Nutt. Gen. 2: 115. 1818. Type: W. speciosa Nutt., based on
Glycine frutescens L. Sprengel (Syst. 3 : 255. , 1826) corrected the spelling
of the name to Wistaria, as the genus had been dedicated to Dr. Wistar.
The genus consists of 5 or 6 species, of which two are native of eastern

and southern United States. It is evidently very closely related to the

large Asiatic genus Millettia, the species of which are mostly trees or shrubs;

only a few of them are climbing.

SUBTRIBE 3. BRONGNIARTIANAE

Trees or shrubs, with alternate odd-pinnate leaves, stipules, and some-
times stipels. The flowers are axillary or in terminal racemes or panicles.
The calyx is more or less 2-lipped, the tube short, the upper 2 lobes united
high up, the lower 3 lobes also somewhat united. The corolla has a broad
banner. The stamens are monadelphous or diadelphous. The pod is
usually elongate, flat, 2-valved, several-seeded. The seeds are erect, i.e.,
the longer axis of the seed is at right angles to the axis of the pod, with a
well developed strophiole.

The subtribe consists of the following two genera and two from Australia.
It is distinguished by the well developed strophiole. Its relationship is
probably with the Robinianae.

Calyx 5-lobed, the upper 2 lobes united two thirds their length, the

lower 3 usually free to near the base; stamens diadelphous. 5. Brongniartia.

Calyx 2-lipped, the upper 2 and the lower 3 lobes united to the apex;

stamens monadelphous. 6. Harpalyce.

5. Brongniartia H.B.K. Nov. Gen. & Sp. 6: 465. 1823

Shrubs or trees, with odd-pinnate leaves. Stipules are present, but
often caducous, the leaflets entire-margined, sometimes with minute stipels.
The flowers are normally axillary in small 1- -7-flowered clusters. In some
species, however, the upper floral leaves are reduced to the two stipules,
which resemble a pair of bracts, and the inflorescence becomes falseiy
racemose. The individual flowers are subtended by a pair of bractlets,
sometimes foliaceous. sometimes reduced to a pair of hair-tufts. The upper
two calyx-lobes are united high up, the lower only slightly at the base.
The corolla is red, brown, or purple; the banner is broad, its blade orbicular
or broadly obovate, short-clawed; the wings are obliquely oblanceolate or
obovate, more or less falcate, with a short fleshy claw and a rounded auricle;
the blades of the keel-petals are broadly lunate, with a fleshy claw, united
from the middle to the tip. The stamens are diadelphous, the alternate
ones shorter; the ovary is short-stipitate, the style incurved, glabrous,
the stigma minute. The pod is short-stipitate, flat, elongate, 2-valved,
usually several-seeded, slightly wing-margined on the upper suture, the
valves leathery.

Illustration: Plate XXXIV E. Brongniartia Benthamiana Hemsl.,
X 2/3; 1. calyx, 2. staminal sheath, 3. pistil, 4. banner, 5. wing, 6. keel-
petals, X 1; 7. pod, X 2/3; 8. seed of B. sericea Schlecht., X 1.

The genus was established on two species, B. mollis and B. podalyrioides ,
of which the first may be regarded as the type.

494 AMERICAN JOURNAL OF BOTANY [Vol. 10.

Synonyms:
Peraltea H.B.K. Nov. Gen. & Sp. 6: 469. 1823. Type: P. lupinoides

H.B.K., now known as Brongniartia lupinoides.
Megastegia G. Don, Gen. Syst. 2 : 468. 1832. Type: M. speciosa G. Don,

regarded as the same as B. thermoides.

The genus consists of 37 species, natives of Mexico and Central America,
and a few species from South America. One species, B. oligosperma Baill.,
is somewhat abnormal in the genus on account of its hairy few-seeded pods.

6. Harpalyce DC. Mem. Leg. 496. 1825
Trees or shrubs, with alternate odd-pinnate leaves and small stipules.
The leaflets are entire-margined, petioluled, sprinkled beneath with yellow
or orange glands or gland-like scales. The flowers are racemes The
calyx is 2-lipped, with the upper 2 and the lower 3 lobes united to the tip.
The banner is rounded or obovate, short-clawed, the wings are very irregular,
strongly curved, obtuse, the keel-petals more or less falcate, united to above
the middle but the tips free and obtuse. The stamens are monadelphous,
but the sheath is split to the base; the anthers are alternately longer and
shorter. The ovary is sessile, the style arcuate, glabrous, the stigma
minute, terminal. The pod is 2-valved, several-seeded, leathery or woody.

Illustration: Plate XXXIV F. Harpalyce Loeseneriana Taub.,
X 2/3; 1. calyx, 2. stamens, 3. pistil, 4. banner, 5. wing, 6. keel-petal, X I ;
7. pod, 8. cross section of the same, X 1 ; 9- banner of H. cubensis Griseb.,
10. wing, 11. keel-petal, 12. seed, X 2/3.

The genus was based on an unpublished illustration in Mocino and
Sesse's Flore de Mexique of Astragalus carnosus. In De Candolle's Pro-
dromus the type is given as H. formosa DC, based on the same.

The genus consists of 14 species, viz., 7 from Mexico, 1 from Guatemala,
3 from Cuba, and 3 from Brazil. One species, H. mexicana, is abnormal
in that the valves of the pod are woody. As the flowers of this species are
unknown, it may belong to some other genus. The rest can be divided
into three natural groups: (1) Cuban species, in which the petals are fleshy,
the keel much longer than the other petals, and the pod is small and narrow;
(2) Mexican species (including the one from Guatemala), in which the
petals are membranous, the keel is scarcely longer than the banner, and the
pod large and broad; (3) Brazilian species, similar to the Mexican but the
pod narrower and more or less divided internally by false cross-partitions
of spongiose tissue.

SUBTRIBE 4. BARBIERIANAE.

Shrubs, with alternate odd-pinnate leaves, narrow stipules and stipels.
The flowers are in axillary and terminal racemes, each subtended by a pair
of bractlets. The calyx is cylindric, 5-lobed, the lobes subequal. The
corolla is not truly papilionaceous, the petals with long slender claws ; the
blade of the banner is oblanceolate, not spreading; the wings have elliptic
blades, the keel-petals oblanceolate or oblong, obtuse blades, scarcely

Nov., 1923] RYDBERG — NORTH AMERICAN FABACEAE 495

falcate, and united above the middle. The stamens are diadelphous, the
sheath is straight; the ovary is sessile, the style nearly straight to near the
apex, hairy along the upper side above; stigma minute. The pod is flat,
straight, 2-valved, transversely septate within.

The subtribe consists of a single genus and a single species.

7. Barbieria DC. Mem. Leg. 241. 1825

The generic characters are included in the subtribal description.

Illustration: Plate XXXIV G. Barbieria pinnata (Pers.) Baill.,
X 2/3; 1. calyx, 2. stamens, 3. pistil, 4. banner, 5. wing, 6. keel-petal, 7. pod,
X 1 ; 8. cross section of the same; 9. seed, X 2.

The genus was based on Clitoria polyphylla Poir. or Galactia pinnata
Pers. Its distribution extends from Porto Rico to Cuba, southern Mexico,
Bolivia, and Brazil.

Subtribe 5. Sesbanianae

Trees, shrubs, or herbs with abruptly pinnate leaves and caducous
stipules but no stipels. The flowers are borne in axillary racemes. The
calyx is round-campanulate, fully as broad as high, with very short teeth.
The banner has a broad reflexed blade and a short claw; the keel-petals are
strongly arcuate, united at the middle, but the claws and tips are free.
The stamens are diadelphous; the staminal sheath is dilated below and the
upper filament knee-bent near the base. The pod is 2-valved but often
indehiscent, stipitate and beaked, compressed, usually with cross-partitions
between the seeds, but not disarticulate.

On account of the abruptly pinnate leaves this subtribe, as well as the
Corynellanae, is somewhat abnormal in the tribe but could not be trans-
Jerred to the tribe Vicieae, to which it does not show any relationship in
other ways. The structure of the flower is practically the same as in many
genera of the subtribe Robinianae; the pod, however, is, in most of the
genera, but slightly compressed and internally divided by false transverse
partitions.

The subtribe consists of the five following genera, which are thus dis-
tinguished.

Pods flattened, neither 4- winged nor 4-angled.

Pods many-seeded, linear, with thick margins; calyx not con-
spicuously oblique.
Pod not torulose.

Flowers middle-sized or small; banner suborbicular;
blades of the broad keel-petals with an auricle; seeds
subcylindric-oblong. 8. Sesban.

Flowers large; banner elliptic; blades of the rather
narrow keel-petals without a basal auricle; seeds reni-
form-oblong. 9. Agati.

Pod torulose; banner suborbicular, equaling the other
petals; blades of the broad keel-petals without a basal
auricle; seeds reniform-oblong. 10. Daubentoniopsis.

496 AMERICAN JOURNAL OF BOTANY [Vol. 10,

Pods 2-seeded, with thin margins, lance-elliptic; seeds oblong-
reniform enclosed in the inner membranous layer of the
valves; calyx decidedly oblique. 12. Glottidium.

Pod 4-angled, often 4-winged, not torulose. II. Daubentonia.

8. Sesban Adans. Fam. 2:327. 1763

Herbs or shrubs with abruptly pinnate leaves and numerous leaflets.
The flowers are borne in axillary racemes, with deciduous bracts, and a pair
of deciduous bractlets under the flowers. The corolla is yellow or the
banner dotted, streaked, or tinged with purple. The banner has often one
or two callosities at the base of the suborbicular reflexed blade; the wings
are short-clawed, the blades obliquely oblanceolate or oblong, with a basal
auricle; the keel-petals have long claws, the blades are strongly and broadly
lunate with a lateral auricle on the upper side. The pod is slender, terete
or slightly compressed, short-stipitate, many-seeded, with cross-partitions
between the seeds, 2-valved. Seeds cylindro-oblong, subtruncate at the
ends.

Illustration: Plate XXXIV H. Sesban Sesban (L.) Rydb., X 1/2;
I. calyx, 2. stamens, 3. pistil, X 2; 4. banner, 5. wing, 6. keel-petal, X 1;
7. pod, X 1/2; 8. cross section of pod; 9. seed, X 2.

The genus was established on Aeschynomene Sesban L. Scopoli (Introd.
308. 1777) changed the name to Sesbania, the latter being better Latin
form.

Synonyms :
Darwinia Raf. Fl. Ludov. 106. 1817. Type: D. exaltata Raf., which has

been known under the name Sesbania macrocarpa Muhl.
Monaplectra Raf. Fl. Ludov. 106. 18 17. This was proposed instead of

Darwinia in case the latter happened to be preoccupied.

The genus consists of perhaps 20 species of the warmer regions of both
hemispheres. It is represented in North America by 4 native and 3 intro-
duced species.

9. Agati Adans. Fam. 2:326. 1763

Small trees, having abruptly pinnate leaves, with many entire leaflets
and deciduous stipules. The flowers are large, borne in small axillary
racemes, and with two deciduous small bractlets subtending the calyx. The
calyx is in structure the same as in Sesban. The petals are comparatively
narrower, the banner is without callosities, its blade oval in outline, and
retuse at the apex; the blades of the wing- and keel-petals are obliquely
lanceolate-lunate, tapering at the base without any distinct auricle. The
staminal sheath and the pistil resemble also those of Sesban, except that
they are more gradually arcuate. The pod is the same as in that genus.

Illustration: Plate XXXV I. Agati grandiflora (L.) Desv., X 1/2;
1. stamens, 2. pistil, 3. banner, 4. wing, 5. keel-petal, X 1/2; 6. pod, X 1/4;
7. cross section of the same, X 1 ; 8. seed in position in the pod, X I.

The genus was based on Robinia grandiflora L.

Synonym :
Resupinaria Raf. Sylva Tell. 115. 1838. Type: R. grandiflora Raf. or

Robinia srandifiora L.

Nov., 1923I RYDBERG NORTH AMERICAN FABACEAE 497

The genus is represented by A. grandiftora (L.) Desv. and one or two
species closely related to it and perhaps not specifically distinct from it,
natives of southern Asia and northern Australia, and by A. tomentosa
(H. & A.) Nutt. from the Hawaiian Islands.

10. Daubentoniopsis Rydb., gen. nov.

Shrubs, having abruptly pinnate leaves, many entire caducous stipules,
caducous bracts and bractlets. The flowers are in axillary racemes. The
calyx is rounded campanulate, broader than high, its lobes very short.
The corolla is yellow, the banner suborbicular, retuse, reflexed, with a short
claw, without callosities; the wings are short-clawed, with obliquely oblong
blades, without a distinct basal auricle; the keel-petals are also clawed with
a lunate, nearly semicircular blade. The stamens are diadelphous, the
staminal sheath dilated at the base. The ovary is stipitate, glabrous, the
style arcuate, glabrous; the stigma minute. The pod is coriaceous, stipitate,
somewhat compressed, linear, several-seeded, decidedly constricted between
the seeds and with spongy transverse partitions. The seeds are oblong-
reniform, about twice as long as high.

Illustration: Plate XXXV /. Daubentoniopsis longifolia (Cav.)
Rydb., X 1/2; 1. calyx, X 2; 2. stamens, 3. pistil, 4. banner, 5. wing, 6.
keel-petal, X 1 ; 7. pod, 8. cross section of the same, X 2/3.

The genus is based on Aeschynomene longifolia Cav. Ic. 4: 8. 1797.
It is intermediate between Sesban and Daubentonia, having exactly the
flowers and seeds of the latter, but the pod is neither 4-angled nor winged.
It is constricted around the seeds, but the exocarp is spongy. In Sesban
the pod is hardly constricted, the calyx- teeth are more evident, the banner
usually has a callosity, and the seeds are different.

The type of the genus, D. longifolia (Cav.) Rydb., has a rather peculiar
history. It was first described by Cavanilles (loc. cit.), and independently
by Ortega1 under the same name. Cavanilles' species was transferred to
Piscidia by Willdenow.2 De Candolle,3 when he established the genus
Daubentonia, included it in that genus, but from the short description given
it is evident that he had in mind a yellow-flowered Daubentonia of the
southern United States and northern Mexico. In his Prodromus4 he re-
peated his error, and besides described on page 265 a Sesbania longifolia
based on Aeschynomene longifolia Ortega. His description fits Cavanilles'
plant. When Watson"' merged Daubentonia into Sesbania, he, influenced
by De Candolle's misconception, thought himself forced to give the yellow-
flowered Daubentonia a new name, S. Cavanillesii S. Wats., as there was
already a 5. longifolia (Ort.) DC. Unfortunately, however, he based this
name on Aeschynomene longifolia Cav., and technically he gave a new name

xOrt. Dec. 70. 1797-1800.
2Sp. PI. 3: 920. 1803.
» Mem. Leg. 286. 1823.
4 Prodr. 2: 267. 1825.
6 Bibl. Ind. 258. 1878.

498 AMERICAN JOURNAL OF BOTANY [Vol. 10,

to the species of Daubentoniopsis instead. Pollard,6 following De Candolle's
and Watson's interpretation of Cavanilles' plant, was of the opinion that
this, being the first one described, should retain the specific name longifolia,
and therefore proposed the name S. mexicana Poll, for Ortega's plant. As
the two are the same, he merely added a new synonym to our species of
Daubentoniopsis. The yellow-flowered Daubentonia is still nameless.

ii. Daubentonia DC. Mem. Leg. 285. 1823

Shrubs or trees, having abruptly pinnate leaves, with many leaflets and
deciduous stipules. Flower in axillary racemes, resembling closely those of
Sesban, but the calyx-lobes are still smaller, the calyx-tube being merely
undulate on the margins, slightly ciliate at the teeth. The banner is without
callosities, and the wings and keel-petals are without basal auricles. The
pod is more or less 4-angled, but somewhat compressed. The endocarp is
membranous and constricted around the seeds, the exocarp more or less
spongy, the sutures are thick, each produced into two sharp ridges or wings.
The seeds are reniform.

Illustration: Plate XXXV K. Daubentonia punicea (Cav.) DC,
X 1/2; 1. calyx, X 2; 2. stamens, 3. pistil, 4. banner, 5. wing, 6. keel-petal,
7. pod, 8. cross section of the same, X 1.

The type of the genus is D. punicea (Cav.) DC, based on Piscidia
punicea Cav.

The genus consists of about half a dozen species, of which 3 are natives
of South America, one of the southern United States and northern Mexico,
Daubentonia Drummondii Rydb. {Daubentonia longifolia? T. & G. Fl. N.
Am. 1 : 293. 1838), and I or 2 of Mexico. I have based it on D. longifolia?
T. & G., rather than on D. longifolia DC, in part as to description, for that
name really belongs to Daubentoniopsis longifolia, as stated before. One of
the South American species, D. punicea, has also been found introduced in
Florida and Mississippi.

12. Glottidium Desv. Jour. Bot. 1: 119. 1813

Annual herbs, having abruptly pinnate leaves, with many leaflets and
deciduous stipules. The flowers are in axillary racemes or panicles. The
calyx and corolla are almost exactly like those of Sesban, the banner with
callosities, the wings and keel-petals of the same shape as those of that
genus. The pod, however, is different, broad, stipitate, compressed, 2-
valved, and 2-seeded, the valves at last separating in two layers, the ehdo-
carp very thin and papery, the exocarp firmer, somewhat inflated, but not
bladdery as in Diphysa.

Illustration: Plate XXXV L. Glottidium vesicarium (J acq.) Harper,
X 2/3; I. calyx, 2. stamens, 3. pistil, 4. banner, 5. wing, 6. keel-petal, X 2;
7. pod, X 2/3; 8. pod in cross section, X 1 ; 9. seed, X 2/3.

The genus is monotypic and was based on Aeschynomene platycarpa
Michx., which is the same as Glottidium vesicarium (J acq.) Harper.

New York Botanical Garden

e Bull. Torrey Bot. Club 24: 154. 1897.

American Journal of Botany.

Volume X, Plate XXXIII.

_ HT„m„rr A»»TTT>T<~A1Vr F A T» A PRAE

American Journal of Botany.

Volume X, Plate XXXIV.

Rydberg: North American Fabaceae

American Journal of Botany

Volume X, Plate XXXV

Rydberg: North American Fabaceae

iiiiiiiiTiiffiiiSSiiSiiif1 Garden Libra7

3 5185 00293 3339