ALS axt

\ oO Sy pic =
Veg Fink —*

Reprinted from Ruopora, Vol. 32, July to December, 1930 Wo 2 wa nl \ eo

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS

By Joun M. Foag, Jr.

Dates or Issuge

tee eee a 28 June, 1930
ee eae et 28 July, 1930
eee ee ee 4 September, 1930
re a. 14 October, 1930
A ce 3 November, 1930
ee ec 18 December, 1930

Missoun: BOTANicAaS
BARDEN Liprany

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY—NO. XCI

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS

JOHN M. Foaa, Jr.
Part I. Generat Discussion
LOCATION

Tue Elizabeth Islands are formed by a partly submerged ridge of
morainal hills which extends WSW from Woods Hole, Massachusetts,
for a distance of about 16 miles. This ridge has been separated from
the mainland, as well as divided into islands, in comparatively recent
geological time. To the northwest this chain of islands is washed by
the waters of Buzzards Bay, while along their southeastern and
southern shores runs Vineyard Sound, a channel 4 miles wide which
flows between the Elizabeth Islands and Martha’s Vineyard. To-
gether, the Elizabeths and Martha’s Vineyard constitute Dukes
County, the former being included in Gosnold Township.

Seven main islands and twelve smaller ones make up the Elizabeth
Islands. Starting at their eastern end the seven principal divisions
are, in order: Nonamesset, Uncatena, Naushon, Pasque, Nasha-
wena, Cuttyhunk and Penikese. All of these lie in a more or less
direct line with the exception of Penikese, which is separated from the
main axis of the chain, being just one mile due north of Cuttyhunk.
Between Nonamesset and Naushon lie Monohansett and Buck
Islands, separated by narrow channels or “ gutters” which have been
bridged. To the north, and lying in Hadley Harbor, are Captain’s
and Ram’s Head Islands, the latter appearing on some maps as

[JULY

“st NDS

3aSayINId

Rhodora

J SLLASNHIVSSVYW
sar ALNNOD S32ING

ae ‘SGNV1S| HLESVZING

120

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 121

Puritan Island. To the south of the gutters lie East Buck and West
Buck Islands, although here again confusion exists, as West Buck oc-
casionally appears on maps as “ Monohansett.” The three Weepecket
Islands extend northward from the eastern end of Naushon and now
constitute a bird sanctuary. With the mention of Gull Island, a
small strip of sand lying east of Penikese, and Pine Island, imme-
diately to the northeast of Nonamesset, the subject of the minor
islands may be dismissed, for the remaining islets are too small to
have received formal names.

HISTORICAL INTEREST OF THE ELIZABETH ISLANDS

A unique historical interest attaches to the Elizabeth Islands
through the fact that upon the outermost of the chain was made the
first attempt to establish an English settlement in North America.
On the 4th of June (Old Style, May 25th), 1602, Captain Bartholomew
Gosnold, after having named Cape Cod and Dover Cliff (now Gay
Head), fixed upon the island of Cuttyhunk as the site of a future
settlement and, in honor of his sovereign queen, called it Elizabeth’s
Isle, which name has since been applied to the entire group. Here,
upon a tiny islet in a large pond at the west end, the crew of Gosnold’s
ship, the “ Concord,” constructed a rude fort, and here they lived for a
period of three weeks. This settlement, short-lived though it was,
thus antedates the founding of Jamestown by five years and that of
Plymouth by eighteen years, a fact which was commemorated by the
erection and dedication of a monument to Gosnold on Cuttyhunk
upon the occasion of the tercentenary of the original landfall.!

Cuttyhunk is the only member of the Elizabeth Islands which has
been able to boast a- permanent population. The little town of
Gosnold, named after its illustrious founder, has long existed as a fish-
ing village at the east end of the island and today has about one hun-
dred inhabitants. During the whaling days schooners bound for
New Bedford were accustomed to stop at Cuttyhunk to pick up their
pilots.

Another claim to fame on the part of one of the Elizabeth Islands
may be made for Penikese, the smallest and most desolate member of
the chain. Here, in the summer of 1873, Louis Agassiz founded his
School which, through the generosity of the New York merchant who

‘For further details concerning the history of Gosnold on Cuttyhunk and the exer-
Cises which

marked the dedication of the monument in 1903, see the Old Dartmouth
Historical Sketches, nos. 1 and 4. New Bedford, Mass. (1903).

122 Rhodora [JULY

donated the island and funds for the construction of a laboratory,
became known as the Anderson School of Natural History. To this
summer school, the first of its kind in the country, came students from
all over the United States and the roll included names which later
became known as belonging to some of the foremost figures in Ameri-
can biology. Following Agassiz’s death in December, 1873, the school
was continued for one summer by his son Alexander, but thereafter
was abandoned and the island reverted to the State of Massachusetts,
later to be used as a leper colony, from which function it was released
only in 1921.

The island of Naushon has for several generations been the property
of the Forbes family, various members of which have summer homes
at the east end, near Hadley Harbor. To the Forbeses also belong
Nonamesset, Uncatena and Nashawena and it is only through the
generosity and hospitality of the owners that it has been possible to
carry on the botanical exploration which forms the basis for the
present survey.

PREVIOUS BOTANICAL WORK ON THE ELIZABETH ISLANDS

Mention has been made above of Gosnold’s visit to the Elizabeth
Islands in 1602. With Gosnold on that expedition were Gilbert
Archer and John Brereton, “gentlemen and historians.” The former
has left us a very readable and illuminating account of the voyage.

One of the chief objects which Gosnold had in visiting the New
World was to collect and carry home to England a cargo of native
Sassafras which was then much in demand because of its supposed
medicinal value. According to Archer, Elizabeth’s Isle (Cuttyhunk)
was in 1602 overgrown with wood, a fact not without interest in view
of the present treeless nature of this and of several other islands of the
chain. Not only did Gosnold’s party find there the Sassafras which
they sought, but mention is also made of cedar, oak, beech and ash.
The very islet upon which the fort was built is spoken of as cedar-
covered. Hills Hope (Penikese) was likewise overgrown with cedar,
and Naushon which was also visited is referred to as being forested, a
character which this island, almost alone of the Elizabeths, has re-
tained in large measure down to the present day.

References to the plant life of the Elizabeth Islands seem to be
lacking for a period of more than 250 years, but we may well suppose
that during that interval extensive deforestation was carried on and

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 123

the islands divested of their original tree growth. Certain it is that
within the memory of no living inhabitant have there been trees on
Cuttyhunk or Penikese, except the few which have been planted by the
hand of man. It is probable that this statement also holds for Pasque
and the greater part of Nonamesset and Uncatena. Nashawena
still has considerable areas which are more or less wooded, and Nau-
shon, as already mentioned, has apparently retained much of its
original forest.

The first published report, known to the writer, on the flora of any
of the Elizabeth Islands dates from the year 1874 and concerns the
island of Penikese. Among the students attracted to Agassiz’s
laboratory in the summer of 1873 was David Starr Jordan, who began
his scientific career with a botanical publication. The task assigned
to Jordan by Agassiz was an enumeration of the plants growing
upon the island and in the waters surrounding it. As the result
of this study Jordan published a list! of the flora in which he included
not only flowering plants but cryptogams as well. So far as the
lower forms were concerned, since no attempt was made at micro-
scopic examination, only the most obvious species were included.
The list enumerates 83 species of algae, 2 mosses, 1 fern and 113 species
of flowering plants. Although the marine algae were preserved
and the original set is still in existence, the writer has it on the author-
ity of Dr. Jordan that no collection of specimens of the higher plants
was made, and we have therefore only the published names as records.
In 1923, on the occasion of the fiftieth anniversary of the founding
of the Anderson School, a botanical survey of Penikese was made by
the staff and students of the Marine Biological Laboratory at Woods
Hole, Massachusetts. The results of this survey were published in
Ruopora for 1924, and, insofar as they indicate the possible direction
of change i in the elements of the flora of one of the Elizabeth Islands,
will be referred to later.

Since 1873 nearly a score of botanists have visited the Elizabeth
Islands and brought back specimens which are in one or more of our
eastern herbaria. Since these records have been incorporated into
the catalog which forms the second part of this study, a brief chrono-
logical account of these collectors is here presented.

Walter and C. E. Faxon, in 1873 and 1875 respectively, collected a
few specimens on Nast ; these sheets are in the Gray Herbarium.

1Jordan, D.S. “The Flora of Penikese."’ Am. Nat. viii. 193 (1874).

124 Rhodora [JoLy

In 1890 a Miss Weir collected on Naushon a few sheets which are
now in the herbarium of the New England Botanical Club.

In August, 1898, Dr. Arthur Hollick made a trip to the islands with
a view to studying the geological formations there presented. In
the published account of this survey! the author makes mention of
some of the plants which he found growing on the various members of
the chain. The few specimens which he collected are in the New
York Botanical Garden herbarium.

In the herbarium of the Marine Biological Laboratory at Woods
Hole are several plants collected on Naushon in 1901. Some of these
bear the name of S. B. Sipe while the others are merely signed “E. A.
S.” Inquiry has shown that these initials refer to Miss Elizabeth A.
Simons (now Mrs. Eldred Jungerich).

Also from 1901 there dates the largest collection made on any of the
Elizabeths, previous to that which forms the basis of the present re-
port. In the Library of the Gray Herbarium there is a manuscript
list compiled by Mrs. Alice R. Northrop of the plants of Nasha-
wena. This list contains the names of 335 species of flowering plants
and ferns and therefore constitutes a substantial contribution to our
knowledge of the flora of this island. Mrs. Northrop spent the sum-
mer of 1901 and a part of that of 1903 on Nashawena and was thus
admirably situated for making careful botanical observations. Her
list includes many surprises: species which are either absent or very
locally known from southeastern Massachusetts, some of them con-
stituting, indeed, notable extensions in range. To this class of rari-
ties belong such plants as Cyperus erythrorhizos, Uvularia perfoliata,
Habenaria bracteata, Arethusa bulbosa, Rumex verticillatus, Coptis
groenlandica, Ranunculus reptans and Hydrocotyle Canbyi. Un-
fortunately not a single one of these specialties appears to be corrobo-
rated by herbarium material. However, Mrs. Northrop did collect
some specimens for permanent record, about one-fifth of the names
on her list being represented by sheets in the collections of the New
York Botanical Garden. And the fact that a few of her most interest-
_ ing records, such as Habenaria blephariglottis, Liparis (Leptorchis)
Loeselii, Tipularia discolor and Asclepias verticillata are substantiated
by specimens and that others, such as Arisaema triphyllum, Medeola
virginiana, Ranunculus delphinifolius, M. yriophyllum pinnatum,

1 Hollick, A.

A Reconnaissance of the Elizabeth Islands. Cont. Geol. Dept.
Columbia Univ. xi. no. 72 (1901).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 125

Hydrocotyle Canbyi, Cornus florida and Trientalis borealis, have been
duplicated by the writer either for Nashawena or other islands of the
group, makes it necessary to give definite weight to the plants men-
tioned on Mrs. Northrop’s list.

In the herbarium of the New England Botanical Club is a sheet of
Habenaria orbiculata collected on Naushon by Lillian MacRae in
July, 1904. This is the only specimen bearing the name of this
collector seen by the writer.

A. H. Moore also visited the Elizabeth Islands in 1904 and several
sheets of his collecting from Naushon and Penikese are in the collec-
tions of the New England Botanical Club.

In 1906, J. A. Cushman paid two visits to the islands, collecting on
Nonamesset on July 27, and on Naushon, in company with Max
Morse, on August 25. Specimens from these trips are in the herbar-
ium of the Boston Society of Natural History.

Naushon and Nashawena were visited by E. F. Williams on July
10, 1911. The few plants collected on this occasion are in the New
England Botanical Club herbarium.

In 1911, also, F. W. Pennell made several collecting trips to the
Elizabeths, touching chiefly Nonamesset, Naushon, Nashawena and
Cuttyhunk. Dr. Pennell’s specimens, numbering about fifty, were
distributed to the Marine Biological Laboratory and the University
of Pennsylvania.

Thirty sheets from Cuttyhunk, collected by S. N. F. Sanford: in
1917, are now in the New England Botanical Club.

Scattered collections were made on various islands by W. R. Taylor
from 1917 to 1921. These specimens, with the exception of a sheet of
Liparis Loeselii which is now in the New England Club, are either at
the Marine Biological Laboratory or at the University of Pennsylvania.

Dr. H. K. Svenson visited Pasque on September 8, 1926 and collect-
ed a few specimens which are now in the herbarium of the New Eng-
land Botanical Club.

On August 10, 1927, Professor M. L. Fernald and the writer visited
Uncatena and N aushon, and the material collected upon that occasion
has been distributed to the New England Botanical Club and the
University of Pennsylvania.

few specimens were collected by E. W Hervey on Cuttyhunk.
ese bear no date and are now in the New England Botanical Club
herbarium.

126 Rhodora [JuLy

COLLECTIONS MADE DURING COURSE OF PRESENT STUDY

By far the largest number of records accumulated to form the basis
for the present account of the flora of the Elizabeth Islands represent
collections made by the writer over a period of six years. From 1923

to 1928 inclusive, each island was visited many times and hundreds
of specimens were collected. This material has been worked over and
specimens have been distributed to the following institutions: Gray
Herbarium, New England Botanical Club, University of Pennsylvania,
Missouri Botanical Garden, Cornell University and Marine Biological
Laboratory, Woods Hole, Massachusetts.

In the Fiftieth Anniversary Survey of Penikese already referred to,’
the writer contributed the list of Spermatophytes collected in 1923.
Subsequent visits have resulted in several additions to that list and
made possible a more careful analysis of conditions on the island, and,
as stated earlier, such evidences of vegetational changes as are thus
afforded will be dealt with in a later section.

TOPOGRAPHY OF THE ELIZABETH ISLANDS

The general topography of the Elizabeth Islands is that of gently
undulating morainal hills with a maximum elevation of about 170
feet. All the features characteristic of typical morainal regions are
here presented, from the rounded hills and depressions, the latter
often occupied by ponds or peaty bogs, to the boulders, some the size
of a small house, which are scattered everywhere. Nowhere except
along the beaches is there any considerable stretch of flat land: a walk
across any of the islands necessitates repeated ascent and descent of
the rolling hills.

Along the south sides of the islands, facing Vineyard Sound, the
shore tends to be steep and precipitous, often presenting sheer sand
and gravel faces nearly a hundred feet high, rising abruptly from a
narrow cobble beach. In general the highest land is along this south
shore, and the ground slopes away gradually to the opposite side of the
islands where there are frequent coves and low brackish swamps or
fresh ponds.
The ponds which occur in the hollows in the open, rounded hills are,
in some cases, merely small pools which may form desiccated bog-holes
or disappear entirely in very dry seasons, or, in other cases, are

‘The Flora of Penikese, Fifty Years After. Edited by I. F. Lewis. R ORA,
xxvi. 181-195, 211-219, 222-229 (1924). . ae

1930] Fogg,—Flora of the Elizabeth Islands, Massachsuetts 127

sufficiently large to have been designated as lakes. The largest of
these latter, West End Pond on Naushon, is more than a quarter of a
mile in diameter.

The character of the beaches varies from those of the rugged
boulder type, liberally bestrewn with huge rounded stones, to those
which offer a smooth sandy shelf. The characteristic type lies be-
tween these extremes, and wefind for the most part a shingle or cobble
beach with here and there piles of boulders and flat sandy patches.
Dunes of shifting sand are rare and occur extensively only at the east
end of Nashawena, and along the north shore of Naushon, west of
Kettle Cove.

In those islands which have been deprived of their trees the open
barren hills are covered with grasses, or other low growth, while the
dry hollows or protected lee slopes harbor dense patches of scrub
vegetation, made up mostly of Myrica caroliniensis and species of
Gaylussacia or Vaccinium. Occasionally an extensive boggy hollow
will be densely wooded, the tops of the trees (Nyssa sylvatica, Quercus
velutina, Acer rubrum, she conforming to the height and contour of
the surrounding slopes

The most conspicuous vegetational feature of the islands, aside from
the open grassy downs, is the dense growth of rather low beech
woods which clothes the greater part of Naushon and smaller areas on
some of the other islands. From a distance these woods are seen to
fit in closely with the general topography, due, doubtless, to the high
wind velocity which would tend to level forest growth to the existing
lines of the hills and ridges.

CHARACTERISTIC FEATURES OF THE SEPARATE ISLANDS

As the various members of the Elizabeth Islands exhibit some
diversity as regards general topographic and vegetational features,
and also in the influence which man has exerted upon the latter, a
brief description of each of the seven main divisions is here given.

Nonamesset. This island, the easternmost of the chain, is roughly
oblong with a length of 114 miles and a greatest width of 14 mile. Its
longitudinal axis lies east and west. The eastern a ee ty eto of
Nonamesset is essentially treeless, save for
sion, while the western quarter is heavily clothed with mixed beech
and oak woods. The rather sharp line between these two areas, a line
which follows, for the most part, an old stone wall, indicates that the

128 Rhodora [JuLy

treeless nature of the eastern portion is due to artificial denudation,
it probably having been cleared for purposes of agriculture or grazing.
Altogether there are about 15 ponds of more or less permanent char-
acter on Nonamesset and all but one of these lie in the exposed eastern
portion. Munsod Pond, as it appears on the charts, has now become
merely an arm of Lackey’s Bay and the narrow bar which formerly
protected it has been submerged. Nonamesset is connected with
Naushon by three bridges which cross the narrow gates or “gutters”
separating Monohansett and Buck Islands. The East and West
Gutters are open and the rapid current flows through them as through
a mill race, but the Middle Gutter is “blind,” being closed by the stone
wall that forms the bridge.

Uncarena is a triangular-shaped island about 34 of a mile long and
14 a mile wide. It lies to the west of Nonamesset across Hadley
Harbor and forms a wedge that juts northward into Buzzards Bay.
Uncatena is practically treeless, except for a small natural area in the
extreme southern corner, and here again the assumption is that active
deforestation has been carried on, for it was from the originally
densely wooded nature of Nonamesset and Uncatena that Woods Hole
is said to have derived its name. There are about a dozen small
fresh water ponds on Uncatena during the course of a moderately
rainy summer and a large inland brackish pond drains eastward into
Hadley Harbor by a sluice-way which makes of the northeastern part
of the island a peninsula. Dry, exposed, undulating grassland
characterizes the vegetation of nine-tenths of this island, with an in-
crease of scrubby thicket to the westward. The summer home of
Mr. Malcolm Forbes is located on the east side of Uncatena and a
large central area is fenced off for grazing. This island connects by a
bridge across the Northwest Gutter with Naushon.

Navusuon is the largest of the Elizabeth Islands. It is 614 miles
long and averages 1 mile wide, with a width of about 114 miles at its
widest point. It extends from Hadley Harbor almost due southwest
to Robinson’s Hole which separates it from Pasque. By far the
greater part (perhaps 34) of the surface of Naushon is covered with a
dense growth of trees. In some regions, like the area near French
Watering Place, these woods present an almost pure stand of beech,
in others there is considerable admixture of oak, hickory, hop
hornbeam, maple and black gum. Almost the only portions of Nau-
shon which are not wooded are those right along the shore or some of

1930]. Fogg,—Flora of the Elizabeth Islands, Massachusetts 129

the higher exposed ridges in the central part of the island. Along
the north shore, on one of these treeless stretches, Scotch Broom
(Cytisus scoparius) was introduced some years ago and has taken hold
so vigorously that it now solidly occupies an area of several acres. In
similar spots along this same shore, various conifers (notably larch,
Scotch pine and several spruces) have been set out as a windbreak
and these appear to be no more than holding their own. Naushon has
many ponds, nearly all of them fresh. West End Pond, Mary’s
Lake and French Watering Place are the three largest, in the order
named, and there are about a score of smaller ones, depending upon
the degree of rainfall. The island is indented by two large coves,
Tarpaulin Cove and Kettle Cove, which are nearly across from each
other on opposite shores. It has been suggested that it was by the
approximation of similar indentations that Nashawena was separated
from Pasque in comparatively recent times and that the same process
may be going on here in Naushon. The gravel cliffs which in many
places form the south shore of the island are high and very precipitous.
East of Tarpaulin Cove they are usually barren on their crests, while
westward they are frequently wooded right to the very edge. Nau-
shon presents several areas of low brackish marshland, the most
extensive being near Job’s Neck and at the west end near West End
Pond. At the east end of the island are the several residences of
members of the Forbes family. Here, also are tracts under cultivation
and scattered farmhouses and outhouses.

Pasque. Lying across Robinsons Hole from Naushon is the island
of Pasque, or Peskinese, as it was formerly called. Pasque is roughly
oval in form, with a long axis, lying east and west, of 114 miles, and a
short one about 1 mile. It is almost entirely destitute of trees, save
for a few protected hollows. The extreme eastern end is low and
marshy, and is drained by a sinuous tidal stream. Numerous fresh
water ponds are scattered around the rim of the island, for the central
part is high and arid. As on the other islands, the depressions among
the hills near the shore are often peaty and boggy and several ex-
tensive areas of this sort are to be found here. Pasque was formerly
the property of a fishing club. As evidences of this are a large frame
club-house, a landing wharf, a truck patch and outhouses, including
an ice-house on the edge of a small pond; all at the eastern end of the
island. The ownership of Pasque has just recently passed into private
hands. Between Pasque and Nashawena lies Quicks Hole.

130 Rhodora [JULY

NasHAwENa. ‘The second largest island of the group is Nashawena
(formerly Nashuina), or “Little Naushon,” with a length of 3 miles
and an average width of about a mile. Nashawena contains large
wooded areas, chiefly toward the east end, although in the troughs
between the long ridges of hills that traverse the island longitudinally
trees may be found, except in the western quarter. The shores and
the extreme western portion are barren and open, as are likewise the
higher hills throughout. Behind the line of dune hills, which forms
most of the eastern border of the island, lie two large ponds of fresh
water separated by a low swampy area: these are known as “the
dune ponds.”’ The easternmost of these ponds is separated from the
sound by a low barrier beach, and evidence indicates that inundations
of salt water probably occur occasionally during the severe storms of
winter. It seems also likely that these two ponds have from time to
time been connected as a single body of water. Here again, as on
Naushon, the highest land is along the southern side of the island
and the slope is to the north. Many of the hollows, especially around
the margin of the island, harbor small ponds, and in addition to the
dune ponds there are several good sized bodies of water, the largest
of which is Choptauk Lake, two-thirds of the way to the west end.
There are numerous swampy and boggy areas on Nashawena, most
of them occurring in the central parts of the island in the wide de-
pressions between the lines of hills. In many cases these swampy
hollows are overgrown by tangles and thickets so dense as to be well
nigh impenetrable. It is supposed that, at one time, much of the
open land on Nashawena was under cultivation. The old stone farm
house, said to have been built in 1725, still stands, although a modern
wing has been added to it in recent years. An obscure gravestone
bears the date 1736. Today the island belongs to the Forbes family
and is used chiefly for the raising of sheep, of which there are thought
to be about 700. The caretaker and his wife, Captain and Mrs.
Mark Jamison, occupy the new wing of the old farmhouse, and to
their kind hospitality the writer owes a very enjoyable and botanically
profitable visit to the island in July, 1928.

CurryHunx. West of Nashawena, and separated from it by a very
narrow passage, Canapitsit Channel, lies the island which Gosnold in
1602 named Elizabeth’s Isle but which has reverted, with some
modification, to the old Indian name of Cuttyhunk. Roughly oblong
in shape, the main body of Cuttyhunk is some 2 miles long, with a

1930] Fogg.—Flora of the Elizabeth Islands, Massachusetts 131

greatest width of about 34 of a mile, and lies along a NE-SW axis.
From its southeastern corner a narrow sand-spit runs due east for 34
of a mile toward Nashawena. On this strip is located U. S. Coast
Guard Station No. 50. The northeastern quarter of the island is
occupied by a large body of salt water known as Cuttyhunk Pond.
To the east this is dredged to Cuttyhunk Harbor and thus offers a
land-locked refuge to small vessels. West of Cuttyhunk Pond is the
little village of Gosnold, spread out upon the sheltered east-facing
slope which leads down toward the wharf. All of the central and
western portions of the island are open grassy downs, exposed to the
full blast of winds from the Atlantic. From the lookout station on
the highest hill, near the center of the island, the land slopes gently
away to the south and southwest. The low-lying western part of the
island is occupied by two bodies of water. One of these is fresh and
furnishes the supply of ice for the inhabitants; it has long been known
to visiting botanists as Ice-house Pond or Sheep Pond. The other is
the pond made known to fame by Gosnold and is variously termed
Gosnold Pond or West End Pond. Although originally mentioned
as a fresh-water pond, and still referred to as such, this body of water,
which is separated from the open sea by a very narrow cobble barrier
beach, is certainly frequently inundated. In 1927, Potamogeton
bupleuroides and Ruppia maritima, var. longipes, both reliable indica-
tors of a brackish medium, were found growing in it and in 1928
various species of marine algae were collected along its northern
shores. Toward the western end of this pond is the tiny island
upon which stands the monument to Gosnold already mention-
ed, an unpretentious structure of rough native stone. Not far from
this, but on the extreme western edge of the main island, stands the
Cuttyhunk lighthouse. Several smaller ponds and numerous boggy
hollows, some of them rather extensive, are scattered over the western
and southern portion of the island, but these tend to disappear late in
summer when the rainfall is less abundant.

PeNnIKESE. The smallest of the main divisions of the Elizabeth
Islands lies a mile to the north of Cuttyhunk. Known also to Gos-
nold as a cedar covered isle, and called by him Hills Hope, this island
likewise goes today by its Indian name and is, if possible, even more
barren and treeless than Cuttyhunk. Penikese is about 24 of a mile
long and 14 a mile wide, with a broadly spatulate peninsula running
eastward for 14 mile from its northern end. Its contour resembles

132 Rhodora [JuLy

that of the other islands, the highest point, however, being only about
70 feet. Its few trees are mostly the result of an attempt at a planting
made around the building which, in the days of the leper colony,
served as the home of the resident physician, although several scrub
willows grow in one of the hollows around the margin of a small pond
on the east side. Of the former luxuriant forest growth Dr. Jordan,
writing in 1874, says, “there is now no trace left save the rotten roots
of a solitary beech stump and a few branches of red cedar and red
maple (?) found buried in the muck of a small swamp.” The status
of the ponds on Penikese appears to be even more precarious than of
those on the other islands. In favorable seasons six small ponds, two
of them brackish, may be found; during a dry summer the number has
been known to be reduced to half. Penikese, then, is dominated by
open, grassy downs with the exception of the narrow cobbly strip
which connects the two portions of the island. That part of the open
hillsides forming the northern slope of the island has been taken over
by the terns (common and roseate), thousands of which here find their
breeding ground. In fact, now that the island has reverted to its
wilder state, these birds show a tendency to usurp it altogether. It is
extremely difficult to walk across any of the grassland areas during
the nesting season without stepping upon the eggs or the young birds.
The handsome stone residence building, on the east side near the land-
ing, has been partly demolished, leaving only a portion of the structure
to house the caretaker who is still stationed there. The frame cottages
on the west side of the island, formerly occupied by the unfortunate
lepers, were destroyed in 1927, a single concrete structure being all
that remains. This, and the tiny graveyard at the extreme north end
of the island, bear mute testimony to the use to which Penikese was
put from 1905 to 1921. The remains of an old wooden reservoir cap
the highest hill on the island, while, set in a large boulder near by, is
a bronze tablet placed there in 1923 to commemorate the fiftieth
anniversary of the founding of the Anderson School of Natural History
by Jean Louis Rodolphe Agassiz in 1873.

(To be continued)

*

Ps

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts ~ 147 ‘

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS

JoHn M. Foae, Jr.
(Continued from page 132)
HABITATS

Despite their almost uniformly bleak and arid nature, the Elizabeth
Islands offer a considerable diversity of habitats and the chief of these,
together with a few of the most characteristic plants of each, are here
described.

BracuEs. These may be of boulders, cobbles or pure white sand.
In the first two cases, few plants may be sought for, although on the
shingle or cobble beaches Ammophila breviligulata sometimes manages
to get a foothold in the loose stones. The sandy beaches, however,
offer a habitat for certain characteristic species, of which the following
may be noted:

Ammophila breviligulata Lathyrus maritimus

Triplasis rg ie Euphorbia polygonifolia

Polygon Convolvulus pert, var. pubescens
Atriplex mile var. hastata Solanum

Salsola K Solidago seats tierivetis

Arenaria peploides, var. robusta Xanthium echinatum

Cakile edentula Sonchus oleraceus

On some of the beaches dead Eel Grass, Zostera marina, has been
piled up by the waves, forming dense mats often two or three feet
thick. ,

Satt Marsu. ‘There are no extensive salt marshes on the islands,
but a few restricted areas of this nature do occur, such as those along
the southwestern shore of Nonamesset, the eastern side of Uncatena,
the northeastern end of Naushon facing Lackeys Bay, the regions on
the north shore at the west end of the same island, the southeastern
corner of Pasque and the extreme west end of Nashawena. Other
similar patches, scattered throughout, are too limited or local to
merit enumeration. As typical of these low, brackish marshlands,
or the slightly elevated peaty areas bordering them, may be cited:

Typha angustifolia Cyperus elgg

Andropogon glomeratus C. strigosu

Echinochloa Walteri Peden pervala

s —_ ina oe var. = E. uniglumis, var. halophila
S.p E. rostellata

Distichlis spicata Scirpus Olneyi

148 Rhodora [AueustT

8S. validus Salicornia europaea

. campestris, var. paludosus Suaeda linearis
Carex hormathodes Spergularia leiosperma
Juncus bufonius Hibiscus Moscheutos
J. Gerardi tilimnium capillaceum
J. articulatus, var. obtusatus Pluchea camphorata

Rumex maritimus, var. fueginus

On the brackish mud flats, like those bordering Cuttyhunk Pond,
occur such characteristic things as:
Puccinellia paupercula, var. Suaeda linearis

alaskana . Spergularia leiosperma
Salicornia europaea Plantago oliganthos
S. ambigua

BrackisH Ponps. Either in the salt marsh areas themselves, or
near the shore and separated from the sea only by narrow shingle
barriers, occur several brackish ponds, in which may be found such
plants as:

Potamogeton bupleuroides Ruppia maritima, var. subcapitata
P. pectinatus R. maritima, var. longipes

FresH WaTeR Ponds. There are at least 65 fresh water ponds of a
more or less permanent nature on the Elizabeth Islands. Of: these,
some are mere pools 20 or 30 yards across in low hollows or kettle-holes,
while others, such as West End Pond on N aushon, may have a width
of nearly one-third of a mile. With this wide difference in size there
goes a corresponding diversity of pond bottoms and the accompanying
floras. Some of the ponds have grassy bottoms and merely represent
hollows which have become filled with rain water; others have a sandy
or cobbly bottom; while the bottoms of still others are formed of a
thick grayish clay. As representative of the plants, floating or sub-
merged, of these fresh water ponds may be mentioned:

Potamogeton Oakesianus Elatine minima
4 Samoan i puvizia palustris
. diversifolius iophyllu
Vallisneria americana rg isars pangs
Glyceria acutiflora M. tenellum
Lemna minor ; Proserpinaca palustris
Nymphozanthos variegatus Hydrocotyle umbellata
ymphaea odorata H. Canbyi
Ranunculus delphinifolius H. verticillata
Callitriche heterophylla Nymphoides lacunosum
C. palustris icularia geminiscapa

Ponp Marains. The margins of the fresh water ponds support, in

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 149

most cases, a rather characteristic flora, the components of which de-
pend usually upon the nature of the shore.
Around those ponds with a pure sandy border the following plants

may be listed as fairly typical:

Cyperus dentatus

Juncus pelocarpus
aris
J. marginatus

Ranunculus Cymbalaria
otentilla pacifica
a mph Aig

KH TD key ee
gay

asu
I lysanthes ontose
Coreopsis rosea

Surrounding those ponds, however, which offer a peaty or boggy
border, a somewhat different series may be cited as representative:

Thelypteris palustris, var. Iris versicolor
pubescens Habenaria lacera
Sagittaria latifolia Drosera rotundifolia
Sparganium americanum D. interme
Glyceria obtusa Spiraea tomentose
allida olygala eruciat
Eleocharis acicularis ippichin Sates
Scirpus cype um
Rynchospora alba V ila. i anoeaata
R. capitellata Rhe
Carex lurida Sevtclbicin epilobifalin
deep septangulare Lycopus
roliniana Gratiola aurea
Bidens connata

“yris ¢
Juncus pibehcmorn
J. acuminatus

It is not suggested that the two classes of plants above presented
be taken as mutually exclusive. In general, however, these species
exhibit a marked preference for the habitat under which they are
listed.

Swamps. Swampy areas occur near some of the larger ponds,
notably west of French Watering Place on Naushon and aroun
portions of the dune ponds on Nashawena. In addition, most of the
islands boast one or more swampy hollows and Naushon and Nasha-
wena each has several rather extensive swamps in low-lying depres-
sions near the shore. A few of the characteristic plants of this type
of habitat may be noted:

Typha latifolia Juncus effusus, var. costulatus
ina Michauxiana Spiraea tomentosa
gmites communis Rosa pal

Dulichium arundinaceum Impatiens biflo

Scirpus validus Decodon verticillatus

8. cyperinus Sium sua

150 Rhodora : [Auaeust

Asclepias incarnata, var. pulchra Cephalanthus occidentalis
Lysimachia terrestris Eupatorium verticillatum

Bogs. In addition to the restricted boggy areas in the moist
hollows and those forming pond margins, there are several rather
extensive bogs of a permanent character. Chief among these may be
mentioned one chain of bogs at the east end of Pasque and another
series at the west end of Cuttyhunk. Many of the plants which
occur around the peaty borders of the small ponds grow also in these
open bogs, but certain other species reach their fullest development
only in the more extensive areas. The following is but a partial list
of some of the more conspicuous of these bog plants:

Woodwardia areolata Juncus effusus, var. costulatus
Thelypteris palustris, var. Habenaria clavellata

pu H. lacera
Osmunda regalis, var. Pogonia ophioglossoides

spectabilis Calopogon pulchellus
Lycopodium inundatum, var. Drosera rotundifolia

igelovii rmedia
anicum longifolium Rubus hispidus

Glyceria obtusa Polygala cruciata
Eriopho ginicum Viola lanceolata
Rynchospora fusca Epilobium palustre, var. monticola
Carex cephalantha Clethra alnifolia
C. Howei Rhododendron viscosum
C. canescens, var. disjuncta Chamaedaphne calyculata
C. virescens Vaccinium macrocarpon
© fee inte ee
Eriocaulon septangulare Menyanthes trifoliata, var. minor
Xyris caroliniana

- As already indicated, the larger part of the surface
of the Elizabeth Islands is dominated by open, undulating grassland.
The following list, incomplete though it is, will convey more adequate-

y than could any description an impression of the character of these

bleak grassy downs:
Dennstaedtia punctilobula Aristida purpurascens
Andropogon scoparius, var. Phleum pratense
frequens Holcus lanatus
-aspalum pubescens Poa pratensis
Panicum virgatum, var. spissum Festuca rub
P. depauperatum 3 Agropyron repens
¥, Lindheimeri, var. fasciculatum Cyperus filiculmis, var,
] . meridionale macilen
~ Carex albolutescens
: rocarpon ilicea
Anthaiietaaen odoratum

C. Muhlenbergii

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 151

C. Swanii

Juncus tenuis

J. Greenei

Sisyrinchium angustifolium
Spiranthes gracilis

Rumex Acetosella

Hypericum perforatum
Helianthemum canadense
H. dumosum

H, Bicknellii

Hudsonia tomentosa

Lechea maritima

Viola fimbriatula

Daucus Carot
Trichostema dichotomum
Linaria canadensi

Chrysopsis falcata
Solidago suaveolens

‘ .

MMMM
E
Fe:

°
=r
fe)

ap
xnaphalium obtusifolium
Rudbeckia hirta
\chillea Millefolium
Chrysanthemum Leucanthemum,
var. pinnatifidum

roiOp 4S 7P bp
=]
5
fare
n
$9
a
5
co
tS)
o
ia")
©

gia virginica
Leontodon autumnalis
Hieracium Gronovii

Scrusp Growtu. Under the protected lee of the hills, in dry
sheltered hollows or bordering the woods, where they form a transition
zone between the grassland and the woodland, occur open patches or
dense, scrubby thickets of low shrubs, of which the following may be
designated as characteristic:

Myrica Gale Tlex verticillata
M. caroliniensis Clethra alnifolia
Betula pop a Rhododendron viscosum
Pyrus arbu Leucothoe racemosa
Amelanchier oblongifolia Lyonia ligustrina
Rubus Andrewsianus Gaylussacia frondosa

$a virginiana G. baccata
Prunus serotina Vaccinium corymbosum
P. maritima V. atrococcum
Rhus copallina Viburnum dentatum

Wooptanp. In certain areas, especially near the center of Nau-
shon, the native woods are made up of almost pure stands of beech,
Fagus grandifolia. These trees grow nowhere very tall, averaging,
perhaps, 30-40 feet, and their low, flat, leafy crowns meet overhead,
forming a thick roof through which a subdued light filters. This
climax beech forest may also be seen on a somewhat reduced scale in
portions of Nonamesset and Nashawena. Usually, however, the
wooded areas, although they may be dominated by beech, contain a

152 Rhodora [Avaust

liberal sprinkling of certain other species, most prevalent among which
are:

Ostrya virginiana Sassafras officinale
Quercus alba Acer rubrum

. velutina Nyssa sylvatica
Hamamelis virginiana

In addition to these important constituents of the densely forested
portions, a few other trees dccur scattered here and there, seldom
entering conspicuously into the formation of the heavy woods. As
such may be named:

Pinus rigida Prunus serotina
Chamaecyparis thyoides Rhus Vernix
Juniperus virginiana lex opaca

alba Cornus florida

Mention has already been made of the efforts which were carried on
to introduce certain trees either as a windbreak or for ornamental
purposes. Some of these, such as white poplar, ailanthus and catalpa,
have taken hold and are spreading, while others apparently just
manage to survive. A partial list of these introductions follows:

Pinus sylvestris Betula pubescens

Larix decidua B. pendula

Picea Abies Gleditsia triacanthos
P. glauca Robinia Pseudo-Acacia
P. pungens Ailanthus glandulosa

f a Catalpa bignonioides
Populus alba

The herbaceous flora which enjoys the protection of the native
woodland of the islands is for the most part a rather meagre one. The
dry, leaf-covered floor of the pure beech woods is almost uniformly
sterile, so far as vascular plants are concerned, although such an
habitat presents a rich and varied mycological flora, especially follow-
ing a heavy rain. In the more open mixed woods,
characteristic species inhabit the shaded knolls.
cited:

Pteridium aquilinum, var.

however, several
As examples may be

c Sanicula canadensis

latiusculum Monotropa uniflora
Thelypteris noveboracensis Epifagus virginiana
Panicum otomum Galium pilosum
Carex cephalophora

Certain of the low depressions or hollows in the woodland areas are
Swampy and, in addition to high-bush blueberries (Vaccinium corym-
bosum and V, atrococcum), may harbor such plants as:

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 153

Sparganium eurycarpum Juncus effusus, var. solutus
. americanum Decodon verticillatus

Sagittaria latifolia Siam suave

Glyceria striata Lycopus uniflorus

Carex lupulina Erechtites hieracifolia

Other similar depressions are moss-covered and boggy and in such
situations may be found:

Carex Howei Oakesia sessilifolia

C. canescens, var. disjuncta Maianthemum canadense

C. brunnescens, var. sphaero- Medeola virginiana
Trientalis borealis

Arisaema triphyllum

CHANGES IN THE FLORA OF THE ELIZABETH ISLANDS

Here it is proposed to indicate the possible direction and nature of
the changes in the flora of these islands. It has seemed advisable to
put on record certain facts which illustrate what has already taken
place in this respect and to point out others which may be of interest
to the future student of the islands in interpreting further changes.

The original wooded nature of all of the Elizabeth Islands has
previously been alluded to, as has also the fact that the present treeless
condition of some members of the chain is apparently the direct
result of cutting by man. Right here the question may very well be
asked, “Why have the islands thus denuded never regained their
forest growth?” In attempting to solve this problem two chief
factors must be taken into account and their relative importance
weighed.

In the first place, sheep have been raised more or less extensively on
the islands from time to time and the effects of these browsing animals
in cropping off the young vegetation must not be lost sight of. Despite
the numbers and activities of the sheep, however, they have not suc-
ceeded in keeping down completely the herbaceous growth in those
areas which they inhabit. Even on Nashawena, where their numbers
are greatest, the open grassy downs where they graze boast a large
number of species of grasses as well as other plants and one has no
difficulty in collecting perfect and unmutilated specimens of any plant
which he desires. While evidences of grazing are certainly not absent,
the region in general does not present the devastated appearance which
so often results where sheep have been allowed to run wild; and the
fact that so many herbaceous and shrubby plants are able to make a
showing, especially in the protected hollows, would seem to indicate

154 Rhodora [Aucusr

that it can scarcely have been the sheep alone which kept back the
developing growth so severely that the trees were unable to regain
their foothold.

That is, not only would the Elizabeth Islands, together with Long
Island, Block Island, Martha’s Vineyard and N. antucket, have formed
a portion of a continuous land surface, but they would have been some
miles inland from the actual coast line. The nearest approach of the
20-fathom line to Cuttyhunk is at a point almost due south, where it is
now about 20 miles (32 kilometers) out from the shore. To the south-
east this distance increases to about 90 miles (146 kilometers) as the
submerged contour swings out to sea to conform roughly to the out-

The point requiring present emphasis is this: if all the islands were
heavily wooded at a time when perhaps their inland location afforded
€m some protection, it seems highly probable that later, when they

llick, A. Plant Distribution as a Factor in the Interpretation of Geological
Phenomena, with Special Reference to Long Island and Vicinity. Trans. New York
Acad - xii. 189-202 (1893), :

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 155

than obtain at present, while those less fortunate have suffered
through their comparatively recent exposure to the unmitigated
forces of the Atlantic. That excessively high wind velocity is an
effective factor in retarding tree growth is nowhere better shown than
on the eastern side of Nantucket with its extensive scrub oak barrens.
This is further borne out on the Elizabeths by the fact that in the
open, unprotected areas scrub growth forms only in the more or less
sheltered hollows and the occasional isolated sapling which does get a
start elsewhere remains dwarfed and stunted.

It is true that on Martha’s Vineyard the woods along certain sec-
tions of the north shore have been cut for their timber more than once,
and that new growth has been quickly made. But this slope enjoys
the protection of the high line of morainal hills, averaging 200-300
feet, which shelter it from the winds of the open sea. No such protec-
tion exists on the Elizabeth Islands and the lack of it, rather than the
presence of sheep, appears to be the determining factor in the failure
of natural reforestation. In the light of these facts, it would seem
futile to hope that the devastated areas can ever regain their former
wooded luxuriance. .

Another phase of vegetational change which it seems worth while
putting on record is the behavior of certain introduced species on the
Elizabeth Islands.

Reference has already been made to the planting of Scotch Broom,

ytisus scoparius, on Naushon. It is interesting to note that, in the
account of his reconnaissance made between August 10 and 16, 1898,
Dr. Hollick says of this species that it was “planted over extensive
areas” on Naushon but that it “did not appear to be in a very thriving
condition, ”’! Today Cytisus occupies solidly an area of many acres
along the north shore of the island, near Kettle Cove. On the 10th
of August, 1927, Professor Fernald and the writer visited this locality
and had the unique experience of wandering through this exotic
Plantation. The plants grow very close together, and are often 6-8
feet tall, and the tendency in attempting to traverse the area is to
lose completely one’s sense of direction. Unless checked in some way,
Cytisus bids fair to encroach even further upon the surrounding region
and to usurp in time a much larger area than that which it now domi-
nates. Although introduced also on Pasque and Nashawena, Scotch
Broom has nowhere else made the showing that it has upon Naushon.

‘Hollick, A. Cont. Geol. Dept. Columbia Univ. xi. no. 72. 391 (1901).

156 Rhodora [AuausT

Another leguminous plant that has been successful in establishing
itself is the Woad-waxen, Genista tinctoria. This species was intro-
duced at the extreme east end of Naushon, near Hadley Harbor. It
now occupies almost solidly a large field in this vicinity and occurs
scattered elsewhere over the open hillsides here as well as on Uncatena.

At several places the Tree of Heaven, Ailanthus glandulosa, has
become thoroughly naturalized and appears to be spreading rapidly.
This is especially true on Naushon, north of Tarpaulin Cove, where, in
at least one protected hollow, this tree has formed an extensive and
almost impenetrable thicket.

There remains to be considered in this connection such light as is
thrown on the nature of vegetational changes by an examination of
David Starr Jordan’s account of the flora of Penikese as he found it
in 1873.!_ The author states it as his hope that his list may have an
interest for future botanists, especially “as showing which plants
survive a prolonged struggle for existence against grass and sheep.”
And as this is the first published list of the flora of any of the Elizabeth
Islands, it forms our chief basis for a study of those changes which may
have occurred over a considerable period of years.

In the paper entitled “The Flora of Penikese, Fifty Years After,’’
which has already been mentioned, Dr. I. F. Lewis summarizes the
numerical differences berween Jordan’s list and that compiled as the
result of the survey made in 1923.2 It is not intended to duplicate
that summary here, but subsequent exploration by the present writer
has yielded so many additional records, and a closer scrutiny of the

terminology employed in the earlier list has resulted in a so much -

better understanding of the discrepancies involved, that it seems well
to consider, as briefly as possible, just how much significance, if any,
attaches to the marked difference in the superficial aspects of the two
lists.

The list for 1873, compiled by Dr. Jordan, contains 1 fern and 113
flowering plants, whereas the present list includes 3 ferns, 1 gymno-
sperm (introduced) and 178 flowering plants. Of the 114 species of
vascular plants tabulated by Jordan for Penikese, including Gull
Island, a considerable number (about 25) have not been found as the
result of recent investigations. On the other hand, of the 182 species
of vascular plants on the present list an even larger number (probably
100) were not enumerated in the earlier report.

1Jordan. l.c. p. 193.

* Lewis, I. F. Ruoponra, xxvi. 188 (1924).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 157

In comparing these two lists the necessary allowances must be made
for the very natural changes in nomenclature which have come about
during the intervening 55 years. Dr. Jordan states, in a letter to the
writer, that the names he employed were those found in the edition of
Gray’s Manual then most recent (Ed. 5). With this fact in mind, it
then becomes possible to reconcile a few of the disparities in the two
lists. In general, these discrepancies fall roughly into three categories
which may be briefly described as follows:

In the first place, there are those cases in which a difference of
names involves direct synonymy. Thus, the plant listed by Jordan as
Dicksonia punctilobula Kunze is surely the same as that which we are
today calling Dennstaedtia punctilobula (Michx.) Moore. Similarly,
his Panicum Crusgalli L. corresponds to our Echinochloa Crusgalli
(L.) Beauv., his Triticum repens L. to our Agropyron repens (L.)
Beauv., his Scirpus pungens Vahl to our S. americanus Pers., his
Maruta Cotula DC. to our Anthemis Cotula L., and so on.

The second class of discrepancies includes cases involving mistaken
identity or in which an older species has become recognized as con-
sisting of two or more separate and distinct entities. For example,
Calamagrostis arenaria Roth of Jordan’s list is certainly the plant
known today as Ammophila breviligulata Fernald. This is not at all
a case of direct synonymy, but merely an instance where an American
plant, as beautifully pointed out by Fernald, has proved upon study to
be entirely distinct from its Old World ally. Again, Jordan’s Spartina
stricta Roth is doubtless our S. alterniflora Loisel. var. pilosa (Merrill)
Fernald, his Scirpus maritimus L. our S. campestris Britton, var.
paludosus (A. Nels.) Fernald, his Sisyrinchium Bermudiana L. our
S. angustifolium Mill. Spergularia salina Presl, of Jordan’s list,
appears not to grow on Penikese at the present time, but S. lecosperma
(Kindb.) F. Schmidt is fairly common and we are presumably warrant-
ed in applying a modern interpretation to the older name. Compar-
able to this are Cerastium viscosum L. for which we find only C.
vulgatum L., Viola sagittata Ait., which is represented only by V.
fimbriatula Sm., Scutellaria galericulata L. which is replaced by S.
eprlobiifolia Hasnilion, and a host of similar cases. In all such
instances, then, we are probably justified in assuming that a plant
bearing an unallowable name on the early list is represented today by
the name of a recently recognized segregate or a closely related species,
rather that that it has actually disappeared from the flora.

cmmnenctena Ste
IBRARY |
; oF |
SS AQUINGTON |

a

158 Rhodora [AuausT

Finally, there are several plants on the Jordan list concerning the
identity of which, in the complete absence of herbarium material, it is
futile even to hazard a guess. Panicum dichotomum L., for example,
which so far has not been collected, may ultimately be found still
growing on the island, or, since that name was applied in a very broad
sense in 1873, Dr. Jordan may really have had reference to P. oricola,
P. meridionale, P. Lindheimeri, var. fasciculatum or to something
still different not yet reported from the island. Likewise, Carex
straminea Schkuhr, a name used loosely before this group had received
critical study, may be equivalent to either C. Longii, C. hormathodes or
C. silicea or to all three. And again Polygonum Hydropiper L., not
known from Penikese today, may equal P. punctatum EIll., which is
ubiquitous, or some other species not yet found. Through the unfor-
tunate lack of preserved vouchers, therefore, all such ambiguous
references, when uncorroborated by subsequent collections, must be
discredited.

After having made all due allowances, however, for inequalities
arising from synonymy, modern revisions and ambiguous records, there
still remain slightly more than a score of plants of the 1873 list which
recent searches have failed to reveal. Most of these are species which
occur on the other islands and their absence from Penikese may be
merely an apparent one, to be remedied by further scrutiny. As

a
arenaria, Salsola Kali, Euphorbia polygonifolia, E. maculata, Hyper-
tcum mutilum, and Asclepias incarnata, var. pulchra. With the
exception of the last two species named, which may have disappeared
as the result of gradual drying up of the ponds, there is every reason

A few others, such as Ruppia maritima, Salicornia europaea (S.
herbacea) and Suaeda maritima, are plants of brackish situations and

lose their brackish character. Even during the six years from 1923
to 1928, South Pond has become distinctly less brackish, both as to
the nature of its margin and its algal flora. The same can apparently
be said for the ponds on the peninsula, and while these may be merely
local and transitory phenomena, they suggest the gradual diminution
of brackish areas formerly more extensive, a change worth noting.
Three of Jordan’s plants which were listed from Gull Island only,

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 159

namely, Rhus Toxicodendron, Coelopleurum lucidum (Archangelica
Gmelini) and Limonium carolinianum (Statice Limonium), are not
only still missing from Penikese, but have completely disappeared
from Gull Island as well. The total absence of Poison Ivy from
Penikese, in the face of vigorous and repeated search for it, is one of
the queer and not altogether unpleasant surprises of this island.
Finally, Jordan lists three plants which are not only unknown from
Penikese but, so far, have been collected on none of the other Elizabeth
Islands: these are Puccinellia (Glyceria) maritima, Salix discolor, and
Iva oraria (I. frutescens). The first of these may well refer to P.
paupercula, var. alaskana, known only from Cuttyhunk, the next
might easily have been an introduction which had died out, and Iva
probably is another of the diminishing salt marsh tribe, a species
which it would be interesting to add to the list of Elizabeth Islands
plants.

Turning now to the modern list of the flora of the island, we find,
after again making the necessary correction for synonymy, errors, etc.,
that of the 182 species of vascular plants which it includes, at least
96 (53)% can not in any way be identified with anything on the earlier
list. This large number of species not seen or listed by Dr. Jordan can
conveniently be divided into three groups, as follows: (1) Garden
escapes, about 20 species; (2) Cosmopolitan adventives, about 20
species; (3) Native plants, over 50 species.

Garden escapes. This includes a few ornamentals which may well
have been planted during the days of the leper occupation, some of
which have spread, while others have just barely managed to persist; a
few have escaped from the cultivated area near the site of the old
laboratory building which was destroyed by fire in 1896. A partial
list includes:

Lilium tigrinum Oenothera grandiflora
Asparagus officinalis Ligustrum vulgare
Gysophila paniculata Digitalis purpu
Dianthus barbatus Lonicera japonica
Rubus laciniatus Helianthus annuus
Rosa rugosa Coreopsis lanceolata

Here also, since this is a class of plants the introduction of which
would appear to have been premeditated, should be mentioned a few
trees which were set out around the dwelling of the resident physician
namely:

160 Rhodora [AuGusT
Pinus sylvestris Populus alba
Salix pentandra Acer Pseudo-Platanus
A. platanoides

Cosmopolitan adventives. These are the ever present European
introductions which occur more or less commonly in dry sterile soils
and cleared ground, especially near the haunts of man. It is rather
surprising that so many plants of this class should be lacking from
Jordan’s list, but they have probably been brought in with fodder,
building materials and other supplies. A few of these may be cited:

Avena sativa Trifolium pratense
Dactylis glomerata Hypericum perforatum
Bromus secalinus Daucus Caro

B. hordeaceus Convolvulus arvensis
Carex contigua Linaria vulgaris
Polygonum Convolvulus Tanacetum vulgare
Stellaria graminea Sonchus arvensis
Sisymbrium altissimum 5. oleraceus
Ranunculus acris S. asper

Native plants. As noted above, more than one-half (96 species) of
the plants on the present list of the flora of Penikese appear to have
reached the island since 1873. Of these 96 species, about 40 have
received consideration in the two classes just dealt with; their appear-
ance on the island since 1873 may be accounted for in the light of their
being introductions, either accidental or intentional. Permitting of
no such simple explanation, however, is the occurrence today on the
island of more than 50 species of native plants which were not recorded
as present in 1873 by Dr. Jordan. A few of these, such as Ranunculus
delphinifolius, Potentilla pumila, Callitriche heterophylla and I lysanthes
inaequalis, are rather inconspicuous forms and might conceivably
have been overlooked in the preparation of the original report. Others
are late-blooming members of the Compositae and, as Jordan lists
only one Golden-rod, Solidago sempervirens, and not a single Aster, it
would seem that he had not remained on the island long enough to
obtain a fair sample of the flora of late summer and might thus have
completely missed Solidago juncea, S. rugosa, S. nemoralis, S. canaden-
sis, S. tenuifolia, Aster undulatus, A. multiflorus and A. vimineus, all
of which appear on the present list. In this group also, might be
placed such things as Gnaphaliwm obtusifolium, Rudbeckia hirta, and
Bidens connata, although these plants are generally recognizable by the
first of August, as, indeed, are most of the Goldenrods and Asters
listed above. Incapable, however, of any such interpretations as

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 161

those just offered, is the present existence on Penikese of the following
plants, most of them conspicuous, some of them dominant elements
of the vegetation:

Athyrium angustum eoabe rotundifolia
Thelypteris palustris, var. Sisyrinchium graminoides

pubescens sme ie reo rg
Typha latifolia Rum us, var. fueginus
Panicum virgatum, var. spissum Amelanchier remo lia
Danthonia spicata Rubus pergratus
preabtiage — Rosa sh iss tb
Sci va. Prunus serotina
Carex harnatinacics Rhus typhina

. Silic eerie horse a

J Juncus be mae Oenothera bienn
J. Gre Ligusticum ner evel
d. F edfaace, var. costulatus Asclepias syriaca
J. acuminatus Galium Claytoni
J. articulatus Sambucus canadensis

It is difficult to believe that all these plants could have escaped the
attention of the compiler of the original list and we are rather forced to
the conclusion that they have made their advent to the island since
1873. By just what means they may have made their way to Penikese
and just how much significance may be attached to their occurrence
there today, are matters for conjecture. Certainly they are not
species preéminently adapted for wind dispersal, although Typha and
Asclepias constitute exceptions to this statement, and spores of the
two ferns may possibly have been transported by that agency. A
few of them, such as Scirpus validus, Juncus acuminatus, J. articulatus
and the Rumex, which grow in or around ponds, may have been
brought in by birds, while a few others have fleshy seeds or fruits and
may also have been introduced in this way. On the other hand, the
presence of this large block of recent arrivals may be merely illustra-
tive of what takes place on these islands when the practice of raising
sheep is discontinued, although it is difficult at first sight to see just
why these particular species should have been kept down until recent
times when so many others were not only present in 1873 but have sur-
vived the “ prolonged struggle for existence against grass and sheep.”

(To be continued)

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS

JoHn M. Foae, Jr.
(Continued from page 161)
THE ORIGIN OF THE FLORA

Any attempt to unravel the geographic origin of the chief elements
comprising the flora of the Elizabeth Islands must not only concern
itself with a close scrutiny of the vegetation of the immediately
adjacent regions, but must also take into account supplementary
evidence from two main sources, namely, the history and nature of
the coastal plain and its flora, on the one hand, and the study of the
morainal deposits of which most of southeastern Massachusetts is
composed, on the other. In fact, so far as concerns the Elizabeth
Islands, these two problems are rather inextricably linked and one can
scarcely be considered independently of the other.

HE Coastau Pian. Perhaps no geographic province in North
America has received greater botanical attention over a long period of
time than the Atlantic coastal plain. Occupying a narrow strip east
or southeast of the Piedmont Plateau, from which it is more or less
sharply marked off by the fall line, the coastal plain has been thought
of as extending from the Gulf States and Florida northeastward
through the Southern and Middle Atlantic States and reaching its
northern limit in northeastern New Jersey, near the Hackensack
Marshes, with a representation eastward on Staten Island, Long
Island, and the immediate coast district of southeastern Massachu-
setts.

The surface of the coastal plain presents in general a very gentle
slope to the southeast, which, in southern New Jersey for instance,

168 Rhodora [SEPTEMBER

averages 5 to 6 feet per mile and is seldom over 10 to 15 feet. East-
ward, beneath the waters of the Atlantic, the coastal plain continues
with the same gentle slope to the 100-fathom mark, where, about 100
miles from shore, it suddenly drops off to abysmal depths. In the
southern states the elevated portion of the coastal plain widens to
about 150 miles, while the submarine portion dwindles, finally, off the
east coast of Florida, to disappear almost entirely. Northward the
submerged portion increases in width, reaching 500 miles off the coast
of Newfoundland, while the subaerial portion diminishes, becoming a
mere fringe of islands and the peninsula of Cape Cod, and finally
disappears altogether. Throughout, the soils of the coastal plain are
of a recent nature, being largely Tertiary and Quaternary, and it
appears likely that the present fall line represents roughly the shore
line at the end of Cretaceous time.

The flora of the coastal plain has long been recognized as distinctly
southern in character, due partly to the nature of its constituent
soils (for the most part, sands, clays, gravels, etc.) which have per-
mitted a northeasterly extension of a Carolinian flora, and partly to
its climate, for the temperature is appreciably milder than that of the
only slightly more elevated continental mass to the westward.

In the state of New Jersey, three-fifths of which lies inside the
coastal plain province, Dr. Witmer Stone recorded in 1910! the
presence of 479 species of austro-riparian affinities, plants ranging
from Florida, Georgia or the Carolinas north to southern New Jersey,
some of them reaching Long Island, Massachusetts or, aS we now
know, even farther northward. The recorded number of such species
has been materially increased by recent study.

In 1911 there appeared in Ruopora a very significant paper by
Professor Fernald describing a botanical expedition to Newfoundland
and southern Labrador.? In part II of this paper, the author discusses
the Geographic Origin of the Flora of Newfoundland. Analysis of the

constituent floral elements there represented reveals that 274 indigen-
ous plants (35% of the total flora) are southwestern types and that,
of these, 60 species (over 7% of the total flora) are Carolinian types,
being known from southern New J ersey (or even farther southward),
Long Island, southeastern Massachusetts, Nova Scotia and Newfound-

Flora of the Pin - Rept. N. J. Mus. (1910).
?Fernald, M.L. A Botanical Expedition to Newf, 3}

+Stone, W. The Plants of Southern New Jersey with Especial Reference to the
e Barrens, Ann
‘ 1 and Soutt Labrador.
Ruopora, xiii, 109-162 (1911).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 169

land, but not found inland or in continental eastern Canada. As typi-
cal of this Carolinian subclass are cited: Schizaea pusilla, Ammophila
breviligulata, Carex hormathodes, C. silicea, Corema Conradii, Hudsonia
ertcoides, Myriophyllum tenellum, Utricularia geminiscapa and others.
n an endeavor to account for the presence in Newfoundland of this
coastal plain or Carolinian element, Professor Fernald, after consider-
ing the part which may have been played by birds, ocean currents,
floating ice and logs, and winds, and concluding that they are all
inadequate in explaining this distribution, turns to the question of a
post-glacial land bridge. Hollick’s paper on this subject has already
been referred to, and Professor Fernald, quick to see the phytogeo-
graphic significance of these conclusions, quotes at length from Hol-
lick’s statements and appends corroborative data from other sources.

Evidence derived from a study of the conditions attending the last,
or Wisconsin, glaciation indicates that the amount of water then
withdrawn from the ocean may have been sufficient, in conjunction
with the tendency to uplift already noted, to leave uncovered a con-
siderable portion of the now submerged continental shelf from the
South Atlantic states to Nova Scotia and Newfoundland. Of course,
as Professor R. A. Daly has recently pointed out, the apparent
upward trend of the coastal shelf, resulting in part from the removal of
this vast volume of water from the sea, would have been counteracted
by the lowering pressure exerted by the tremendous weight of the ice
on the continental mass. Nevertheless, it appears that during the
Wisconsin advance, and for some time following it, a very considerable
portion of the coastal bench must have been above sea level, forming,
with the exception of shallow channels such as that draining the Gulf
of Maine or Cabot Strait, a nearly continuous platform for the migra-
tion of plants, and animals as well, northeastward from the southern
states to Nova Scotia and Newfoundland. This strip of silicious
soils probably offered a nearly uniform habitat for the advance of
species of austro-riparian affinities and their extension northward may
even have occurred at a time when the ice had not fully retreated from
the mainland of the continent.

Later, with the melting of the glacial ice and the liberation of vast
quantities of water to the ocean, and perhaps through the operation
of other factors as well, this continental shelf underwent a period of
submergence which resulted in the drowning of this coastal plain flora
except in those areas, higher than the rest and often widely separated,

170 Rhodora [SEPTEMBER

which suffered no such submergence. Hence we have today in the Pine
Barrens of New Jersey, on Cape Cod and the adjacent islands, in south-
western Nova Scotia and in certain parts of Newfoundland the relics
of this formerly continuous flora. That these plants are nearly all
species of sterile or silicious soils indicates that they were admirably
adapted for migration northeastward along this post-glacial land
bridge and probably explains why they have never subsequently ex-
tended their ranges to the better, richer soils immediately inland but
continue to exhibit the disrupted distribution so well typified by
Schizaea pusilla and Corema Conradii.s

In treating this subject elsewhere, Professor Fernald says, “Of
greater interest are the coastal plain species, because they represent
in New England, eastern Canada and Newfoundland a relic of the
extensive flora which during the late Tertiary migrated northward
along the then highly elevated continental shelf and at the drowning
of the shelf were left as relics at isolated points. This isolated remnant
of the flora derived from the southern coastal plain is represented by
about 200 species north of New Jersey, and nearly every excursion to
southwestern Rhode Island, Cape Cod, Plymouth County (Massa-
chusetts), Nantucket, southern Nova Scotia, Cape Breton, eastern
New Brunswick, Prince Edward Island, the Magdalen Islands or
southeastern Newfoundland, adds to the number of thus isolated
species known to us or extends our knowledge of those already recog-
nized. ’”?

And in 1921, in discussing the results of botanical exploration in
Nova Scotia, the same author states that “if there were need of further
evidence that, since the Pleistocene glaciation the continental shelf
of eastern North America has been high in the air, affording an essen-
tially continuous line of migration across the mouth of the Gulf of
Maine to Nova Scotia, thence to Newfoundland, that evidence is now
abundantly at hand. A striking feature of this migration northward
of the southern coastal plain flora is the fact that several distinctive
species or genera, Schizaca pusilla, Lophiola, Habenaria flava, and
perhaps Ceratiola, reached Nova Scotia without establishing colonies
on Long Island, Cape Cod or Nantucket. This would seem to
indicate that the uplifted shelf was a region of some complexity or
else some subtle qualities in the habitats of these plants. ””

bp wees c. Plate 90, opp. p. 140.
oe han Pv ot BrP” 2m Ys Paaet on Eacad

*Fernald, M. L. The Gray Herbarium Expedition to Nova tia. 1 20. Ruo-
pora, xxiii. 168 (1921), ibaa aes

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 171

Enough has probably been said to indicate the overwhelming
amount of botanical evidence in favor of a post-Pleistocene land
connection permitting the northeastward extension of a southern
coastal plain flora. It now becomes imperative to inquire what
part this connection played in the migration of plants to the area
under immediate consideration. However, before discussing the
direct bearing of these findings upon the problem as presented by the
Elizabeth Islands, it will be found helpful to consider what has already
been learned concerning the adjacent areas, especially Nantucket,
Martha’s Vineyard and Cape Cod.

Nantucket. Nantucket, from its isolated position to the south-
east, might be expected to have caught more of these coastal migrants
than the areas to the west and northwest and is therefore considered
first. For our modern knowledge of the flora of this island we are in-
debted to the discerning and painstaking researches of the late Eugene
P. Bicknell, whose account of “The Ferns and Flowering Plants of
Nantucket” appeared in the Bulletin of the Torrey Botanical Club
from 1908 to 1919. The final section of this paper is devoted to a
consideration of the origin of Nantucket’s flora.! Omitting the
hybrids, Bicknell finds upon the island 1103 [1108] species of plants
of which 362 [31%] are listed as introduced and 746 as native. Of
the 746 indigenous species, “over one-half [373 +]... may fairly
be accounted as prevailingly more southern in their general distribu-
tion,” while “something over 150 species . . . are at least pre-
vailingly more northern in their general distribution. ”

Turning first to the plants of southern affinities, we find that over a
hundred of them reach their northeastern limit of range in south-
eastern Massachusetts, others reach Vermont, New Hampshire or
Maine, others occur in the Maritime Provinces, while a small group
is found in Newfoundland. The author then gives a list of 38 plants
which appear not to have been found at any more northern or eastern
point than Nantucket. It is of interest to note, in passing, that only 8
of these are known from the Elizabeth Islands. More than 190 of the
Prevailingly southern plants occur in the Pine Barrens of New Jersey,
over 300 are plants of the coastal plain elsewhere in that state, while
all of Nantucket’s southern-ranging maritime plants, about 40 species,
also occur in New Jersey. Thus we have over 530 species [consider-
ably more than 50%] in the Nantucket flora which display this south-

Bicknell, E. P. Bull. Torrey Club, xlvi. 423 (1919.).

172 Rhodora [SEPTEMBER

ern relationship. To account for this large percentage of austral
forms Bicknell resorts to Fernald’s views on the submerged coastal
shelf and sees isolated on Nantucket the remnants of an extensive flora
of southern derivation belonging to the New England seaboard of
Tertiary time, “a flora lost to our later day with these broad coastal
tracts which now lie beneath the sea. Yet not wholly lost. We find
it still, much of it, we may believe, in the less disturbed flora of our
more southern coastal plain, and we find its remnants persisting as the
merest fringe along the withdrawn more northern coast-lines of the
present day. And isolated on Nantucket it has been preserved to us
in that assemblage of southward ranging plants, now a primary
element in the general composition of the flora. ”

Turning now to the more northern element in the Nantucket flora
we find a group of over 150 species of plants. Of this number about
15 are found nowhere at a more southern point, while about 45 are
near the southern limit of their coastwise range; others reach south to
Long Island and a large number find their southern limit in New
Jersey. In this connection it is interesting to note that of the list of 59
species given by Dr. Stone as reaching from the Maritime Provinces
south to New Jersey, 18 are unknown from Nantucket. Stone’s list,
as it appears in his Flora of Southern New Jersey,? is here given. It
should be observed that the terminology has been revised so as to
correspond to that employed in the present Catalog of Elizabeth
Islands plants. The letter “N” following the name of the plant indi-
cates it is known fron Nantucket, “M” from Martha’s Vineyard, and
“E” from the Elizabeth Islands,

Isoetes Braunii N Eriophorum tenellum NME
Lycopodium inundatum E cile N
a pusilla Carex lanuginosa NE

Potamogeton Oakesianus NME C. trichoea:
Scheuchzeria palustris, var. C. exilis

americana C. livida

riglochin maritima NME C. canescens, var. disjuncta NME
Hierochloe odorata N C. rostrata, var. utriculata N
Spart chauxiana NME C.limosa E

dinacea C. silicea NME

lyceria canadensis NME Eriocaulon septangulare NME
G. obtusa Juncus articulatus NME

.grandis N see J. pelocarpus
Scirpus subterminalis Sisyrinchium angustifolium NE
8. campestris, var. paludosus NME Populus tremuloides N

1 Bicknell. 1. ¢. p.
*Stone, W. l.c. p.

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 173

i? AES ares NM Fivperie P boreale NME
Spe Bebbia: N H. ope
ose maritima N Myasphllien tenelum NME
Chenopodium rubrum N Arctostaphylos uva-ursi, var.
Arenaria lateriflora NME coacti NM
Nymphozanthus variegatus NME Vaccinium pennsylvanicum NM
ctaearubra N Glaux maritima N
Ranunculus Cymbalaria NME Menyanthes trifoliata, var.
Polanisia grave minor NE
a singing: “NME Lirsoestia a inci NME

Dalibarda repens Utricularia interm

Geum strictum Plantago oiganthos NME
Lathyrus maritimus NME Solidago u N
Geranium Robertianum Aster nemoralis

Corema Conradii Xanthium commune

As will be noted, of this list of essentially northern forms Nantucket
has 41 representatives, Martha’s Vineyard 27 and the Elizabeth
Islands only 25

Bicknell gives a list of northern plants which occur on Nantucket
but are unknown from the coastal plain of New Jersey, only a few of
them passing on to Long Island. The significant feature of this list
in the present connection is that, although it contains 44 plants, not
more than a half dozen of these are found on the Elizabeth Islands.

Mr. Bicknell endeavors to account for the presence of these northern
forms, especially those of a maritime character, on Nantucket and
elsewhere to the southwest by supposing that the same marginal land
connection which allowed the plants of the southern coastal plain
to reach Newfoundland would have permitted a counter extension of
northern species to the southwest, perhaps at a later date. The
author also points out that no farther away than Cape Cod there are
established others of these northern species which have not reached
Nantucket and likewise that Cape Cod possesses an extensive coastal
plain flora which is not represented on this seaward island only a few
miles to the southeast. These last two facts are significant because
they apply, even in more marked degree, to the Elizabeth Islands. Not
only, as indicated in speaking of the lists of northern species found on
Nantucket, are many of these boreal forms lacking from the Elizabeths,
but we also fail to find there that large element of southern coastal
plain types which is conspicuous on Nantucket and almost dominant
in certain regions on the Cape.

Marrtua’s VINEYARD. Unfortunately, far too little is known con-
cerning the flora of Martha’s Vineyard to permit of drawing any

174 Rhodora [SmpTEMBER

conclusions as to the origin of the elements there represented. It is
rather surprising that this relatively large and very attractive island
should have escaped careful botanical treatment, but such is the case.
This is not to say that no botanist has ever visited the Vineyard for the
purpose of collecting specimens, for there have been over a score of
independent collections made. A little more than a century ago,
1829, William Oakes visited the island and recorded some interest-

ing finds. One of the most important collections made was that of —

Sydney Harris, who from 1891 to 1904, and again later in 1911 and

1914, collected many sheets, mostly from around Chilmark. The |

island was visited by C. A. Weatherby in 1900, by Professor Fernald
in 1901, by A. H. Moore in 1904, by J. A. Cushman in 1906 and 1911,
by E. P. Bicknell in 1909 (and again in 1912 and 1913), by F. W.
Pennell in 1911, by Miss Magaret Heatley, (now Mrs. C. E. Moss),
beginning in 1916; and all of these brought back material which has
been distributed to one or more of the large herbaria of the eastern
United States. Perhaps the largest collections made were those of
Frank C. Seymour in 1916 and 1917; Seymour’s specimens have been
sent out by the Gray Herbarium. But, so far, no one has published
any coherent account of the flora of the island, and New England
botanists in general know less about its vegetation than about that of
many a more isolated area.

The writer has undertaken to draw together in a single list all the
records based upon specimens available from Martha’s Vineyard. In
the course of this task a systematic census was taken of the material in
the New England Botanical Club and the Boston Society of Natural
History. This resulted in the compilation of a list which includes
about 700 plants. In an effort to supplement this knowledge two
field trips were made to the island, one in August, 1927, and the other,
in company with Professor Fernald, in August, 1928. From the
information thus derived only one conclusion can be safely drawn,
namely that from our present insufficient knowledge of the island, the
surface only of which seems to have been touched, it is impossible to

speak intelligently of the origin of its flora. One or two general |
statements, however, can probably be made with a fair degree of

assurance.

In the first place, it seems evident that the long line of high hills |
which flanks the north shore from Menemsha to West Chop supports —
a flora of a northern or continental nature. Several plants were

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 175

found here, in the richer soils of the wooded slope, which are either
absent from or far from common in southeastern Massachusetts, and
it seems very likely that careful search will disclose many more things
of this nature.

A second point which seems entitled to emphasis is, that, in general,
the flora of Martha’s Vineyard is far from being closely related to that
of the southern coastal plain. Further study may serve to disprove
this observation, but Professor Fernald and the writer, while exploring
the region around Gay Head, were unable to escape the conviction
that the flora dealt with was continental rather than coastal in char-
acter. Time and again the impression was borne home that the
countless southern plants which form the primary element in the flora
of the middle part of Cape Cod were conspicuously absent. An
exception to this general statement may, perhaps, be constituted by
the flora of the eastern part of the island, the region around Edgar-
town, where, apparently, there is a larger representation of austro-
riparian types than may be met with elsewhere on the Vineyard. If
these observations be justified, they will be found to fit in rather well
with the interpretation of the effects of glacial activities upon the
origin of the flora of southeastern Massachusetts.

Care Cop. In speaking of the flora of Cape Cod it is necessary first
to have very clearly in mind the fact that botanically, as well as geo-
logically, this region is far from being a unit, but, rather, may be
divided, more or less distinctly, into three separate provinces, which,
for the sake of oe may be designated the “Upper,” “Mid-
dle” and “Lower” C

“Upper” Cape. This inipliniley roughly Sandwich, Bourne, Mashpee,
Falmouth and the western half of Barnstable. The line of morainal
hills which traverses this province from north to south may be traced
southwest from Woods Hole, as it is of this same ridge that the
Elizabeth Islands are formed. On the mainland these hills are rather
heavily wooded and the superficial aspect of this part of the Cape is
that generally associated with an Alleghanian flora with a slight tinge
of the Canadian. This impression is borne out by a study of the
plants which occur here, many of which are either entirely lacking or
only very locally known elsewhere on the Cape. There are well over
150 such plants, constituting a list too lengthy for eaweing here; the
following few species may, however, be cited as typi

176 Rhodora [SEPTEMBER
Polypodium virginian B. papyrifera ;
Polystichum pereutgt idee Alnus noveboracensis
Thelypteris Phegopteris Ranunculus recurvatus
Osmunda Claytonia ants Thalictrum revolutum
Botrychium virginia Actaea rubra
Lycopodium Saeldiagd Aquilegia canadensis

I Chrysosplenium americanum
Potamogeton Robbinsii Potentilla tridentata
Panicum subvillosum Rubus allegheniensis

P. latifolium Nemopanthus mucronata
Oryzopsis pungens Celastrus scande

Cinna arundinacea Cirecaea latifolia

Glyceria grandis Rhododendron oan pea
G. acutiflora Fraxinus american

Scirpus debilis Scrophularia tandeotath
Carex tribuloides Pedicularis canadensis

C. scabrata q iosiethy perfoliatum
Juncus secundus Viburnum acerifolium
Smilax herbacea ia inflata

Trillium cernuum Solidago ulmifolia
Habenaria dilatata = nemoralis

Malaxis unifolia A. acuminatus

Betula lutea

Few, if any, of these plants, as will be seen, may be looked upon as
sion characteristic of a southern coastal plain flora.

“Middle” Cape. This province embraces the eastern part of
Barnstable, all of Yarmouth, Dennis, Brewster, Harwich, and, perhaps,
Chatham and Orleans. A ridge of morainal hills extending east and
west along the north shore from Sandwich to Dennis, forms the “ back-
bone” of this part of the Cape and is flanked to the south by a broad
outwash plain. This area is not without its trees, but in certain

parts scrub oaks predominate and the appearance of the vegetation —

differs strikingly from that of the inner Cape. The most salient
botanical feature of this province resides in the plants of the numer-

ous ponds and pond margins, many of which are sandy or peaty and —

offer an ideal habitat for an extensive flora of an austro-riparian

nature. Altogether there are over 200 species which are either |
peculiar to this part of the Cape or which here find their greatest _
development, being represented only casually in the other two pro- .

vinces. A partial list of these plants follows:

Meas teris sim
teridium aquilinum, var. pseudo-
caudatum ae

Aairopogo bide
aspa um paammoph um
cai een tus Panicum vi

taria Engelmanniana P. Bicknellii

.

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 177

P. microcarpon

P. annul

P. mattamuskeetense

Fr. mattamuskeetense, var.
P. sp =
P. Wrightianum
“oo pmeumcag
auburn

YP er rata:

. polyanthes

P, Ashei

Se ieal auciflor

8 partina eynosurvides
Glyceri a lax
G. Ferna

: >uceinellia fasciculata

N us
8. campestris, var. novae-angliae
8. Eriophorum
Fuirena ee
Hemicarpha micrant
Rynchospora Torn eyana
R. pte Big! discutiens
arex str,

ps ieaee

campestris, var. echinata
ao perbum
Lachnanthes tinctoria
Salix se

Myrioa a sendentioaie
Fagus grandifolia, var.
pe a

Q

Q. Mi

Q. Prinoides, var. rufescens |
- umbellata

- Setaceum
Chenopodium leptophyllum

Acnida cannabina

Rubus tardatus
Enslenii

R. multispinus

Lespedeza Brittonii

Amphicarpa Pitcheri

Linum str riatum
L. floridanum, var. intercursum
alustris

Ceanothus americanus, var. inter-
medius
ET Se vitacea

icum dissimulatum
Helianthemum dumosum
H. Bicknellii
H. propinquum
Viola emarginata
V. incognita, var. Forbesii
L

Sabatia Kennedyana
campanulata

rPRndat
: Py SS
oO

eo
8 &
ae
Boss
G
-&
B

Siachys hy hyssopifola

iyco

Tice « su lake

agp own Lsgueracye var.
caes

U tour bi pi

U. resu

U. peri

xalium tinctorium

fupatorium hyssopifolium

Shrysopsis falcata

solidago erecta

. pu

. Elliottii
Sericocarpus linifolius
Antennaria petaloidea

=

TRHrHTATHe

178 Rhodora [SEPTEMBER

A. fallax Bidens coronata
Gnaphalium obtusifolium, var. Prenanthes serpentaria

micradenium

As will be seen at a glance, the overwhelming majority of species
in this list are plants of prevailingly southern affinities. Many of
them are common in the Pine Barrens of New Jersey and a consider-
able number were not known to occur north of there, or perhaps
Long Island, when the last revision was made of Gray’s Manual in
1908, but have been added to the Massachusetts flora only as the
result of recent investigation. The foregoing list is not in any sense
an exhaustive one, nor could it hope to be, for nearly every summer’s
exploration adds to the already large number of southern coastal
plain species which are known to occur on the central part of Cape Cod.

“Lower” Cape. At Orleans the Cape makes a right-angle turn
and continues almost due north through Eastham, Wellfleet and
Truro to Provincetown, which is at the extreme tip. This “fore-
arm” or outer portion of the Cape is characterized by rather high,
undulating hills, the axes of which for the most part run east an
west. It is thought that the troughs between these hills may have
formed the delta of a glacial river which drained into Lake Agassiz,
a large body of water impounded by the ice and today represented
by Cape Cod Bay. From Truro northward this part of the Cape
consists of a wave-built sand spit. A few of the plants peculiar to
the Outer Cape, such as Andropogon scoparius, var. polyclados,

Muhlenbergia mexicana, Cyperus filicinus, var. microdontus, Orontium |

aquaticum, Opuntia vulgaris, Aureolaria pedicularia and Baccharis
halimifolia, are far-ranging southern species, but the significant
feature of the flora of this province is that most of its specialties are
forms which display a northern or at least a continental affinity.
The following are a few of the plants belonging to this category:

Potamogeton natans Allium canadense

Elymus arenarius, var. villosus Liparis Loeselii

Scirpus atrocinctus lix lucida

Eriophorum spissum Ranunculus sceleratus

ag oo Muhlenbergii, var. enervis Cardamine parviflora, var.
5 arenicola

C.0 cligosperma Pyrus melanocarpa

C. lasio Potentilla tridentata

C. osc ae Rubus idaeus, var. strigosus
C. bullata R. orarius

Juncus effusus, var. Pylaei R. amnicola

J. articulatus, var. obtusatus R. recurvicaulis

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 179

R. arcuans Arctostaphylos Uva-ursi, var.
Prunus virginiana coacti
- et Sida Vaccinium pennsylvanicum, var.
Corema radii P< ice sat
Tlex puctinillate, var. fastigiata V. Oxycoe
Hudsonia tomentosa, var. inter- Galium trifidum, var. halophilum

media Linnaea borealis, var. americana
Circaea alpin: Bidens cernua
Cornus hecanere Lactuca Morssii

Hieracium marianum

Most of the species listed above are of a prevailingly northern
distribution; their affinities are with the widely dispersed Canadian
flora to the west and northwest. A few of them are here at, or near,
the southern limit of their ranges.

Thus, it will be seen, each of the three natural divisions of Cape
Cod possesses a rather distinctive flora: that of the Inner Cape is
essentially of a continental Alleghanian-Canadian character; that of
the Middle Cape is colored by the presence of a considerable number
of Carolinian or even Louisianian types; while that of the Outer
Cape is rendered striking by the occurrence of so many species of
Canadian, or in some cases even Hudsonian, affinities. An attempt
to explain the underlying reasons which account for this differentia-
tion will be made in summing up the evidence for the origin of the
flora of the Elizabeth Islands themselves.

EuizaBetu Istanps. Turning now to the region under immediate
consideration, we find that the total number of species, varieties
and forms of vascular plants known to occur upon the Elizabeth
Islands is 686. Of these, 128 (1814%) are introduced, while 558
(8114%) may safely be classed as indigenous.

Introductions. 'The subject of those species of plants introduced on
the Elizabeth Islands was dealt with at such length in the section
on Changes in the Flora that it seems scarcely necessary to develop
it further here. It need merely be pointed out that by far the larger
part of these introductions is comprised of those ubiquitous European
and Asiatic adventives which everywhere throughout eastern North
America have taken possession of recently cleared ground or dis-
turbed sandy areas, often completely dominating our native flora.
The rest of the exotics are either garden escapes, as exemplified in
the discussion of the foreign elements in the flora of Penikese, or
species, such as the Cytisus or the various Spruces, which have been
deliberately planted by man for a special purpose. A few plants

180 Rhodora [SEPTEMBER

native to North America are certainly not of indigenous occurrence
on the islands, but have been introduced either accidentally or inten-
tionally. As examples of such may be mentioned Glauciwm flavum
and Solanum triflorum, which grow on the beach near Tarpaulin
Cove, Naushon, and Juniperus communis, var. depressa, which
appears to have been planted along the north shore at the west end of
the same island.
(To be continued)

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS

JoHN M. Foae, Jr.
(Continued from page 180)

Native plants. The 558 species of indigenous plants on the Eliza-
beth Islands fall rather clearly into three fairly well differentiated
categories. In the first place, there is the southern element—plants
of the southern coastal plain which range north from Florida or the
Gulf States to achieve their northern limit in southeastern Massa-
chusetts, some of them passing on to Nova Scotia or Newfoundland;
then there is a group of species of northern affinities, many of which
have already been mentioned in speaking of Nantucket and Cape
Cod, which range south or southwest to Massachusetts or, at most,
New Jersey. And, finally, there is a large and very important
block of plants which fall into neither of the two preceding classes,
but belong rather to a continental upland flora than to that which
characterizes the lowlying reaches of most of Cape Cod and the ad-
jacent islands. It will be well to analyze briefly the constituents of
these three groupings before proceeding further.

The Southern Element. In rather striking contrast to the situa-
tion found on Nantucket, where, it will be remembered, over 50%
of the indigenous flora is prevailingly more southern in its hue, as well
as on the middle part of Cape Cod, where, as has been seen, the na-
tive flora is preéminently that of the southern coastal plain, is the

1930) Fogg,—Flora of the Elizabeth Islands, Massachusetts 209

fact that on the Elizabeth Islands this southern element finds ex-
pression in something less than 20% of the total indigenous flora.

It is true that there are a few species of plants of the southern
coastal plain which, on the Elizabeth Islands are near the very north-
eastern limit of their distribution. Paspalum setaceum, for example,
is known in Massachusetts only from the Elizabeths and Nantucket.’
Panicum longifolium is found nowhere east of Pasque Island, at which
place it is abundant in the peaty bog hollows,’ although it is repre-
sented in Nova Scotia by var. tusketense.2 Tipularia discolor is near
the northeastern limit of its range on one of the Elizabeth Islands
(Nashawena) and on Martha’s Vineyard. Rumezx verticillatus, long
known from Block Island, but otherwise rare in New England, has
recently been collected on the Elizabeths. Hydrocotyle Canbyz and
H. verticillata, both known on the basis of old records from Woods
Hole (for years their only known station in New England) have,
during the course of the present survey, been discovered on the
Elizabeth Islands as well. Solidago minor ranges from Alabama
and Florida to Virginia, then “jumps” to Nantucket, where it was
reported by Bicknell, and is now known to occur on Naushon, the
largest of the Elizabeths. Thus it will be seen, that these islands,
as is true of nearly every other locality along the coast from New
Jersey northward, are not totally lacking in records which represent
interesting, or even spectacular, northern extensions of plants which
are essentially southern in their affinities.

In general, however, the flora of the Elizabeth Islands far from sug-
gests that of the coastal plain. The following enumeration, which in-
cludes the seven species just mentioned, constitutes a nearly complete
list of the plants known from these islands which are also characteristic
species of the coastal strip, ranging from the Gulf States, Florida or
Georgia northeastward. Many of them, of course, continue farther
north and east, being known from Nova Scotia, New Brunswick or
even Newfoundland, but they are, for the most part, plants of a
Pronounced austro-riparian origin. It would be difficult to make
such a list comprehensive, for, in the absence of adequate data
concerning the complete ranges of every species, it is not always easy
to state categorically whether a plant belongs exclusively to the
Southern coastal plain, or whether it enjoys an Alleghanian-Carolin-
‘Bee Weatherby, Ruopora, xxx. 133 (1928).
Ogg, Ruopora, xxxi. 39 (1929

ernald, Ruopora, xxili. 192 (1921).

210

ian distribution.

Rhodora

[OcToBER

All of the species listed below, with the exception

of those marked with an asterisk, are known also from Cape Cod.

Woodwardia virginica
teris simulata
Lycopodium inundatum, var.
igelovii
Chamaecyparis thyoides
Potamogeton Oakesianus
P. pulcher
Andropogon scoparius, var.
polyclados
vinginioas
Paspalum setaceum
ul

p ns
anicum meridionale

albe: arlense

id hac Bpclaclaclach.al-4
QDs . ° . .
ieee,

Spartina alterniflora, var. pilosa
*Diplachne maritima
ed sale erythrorhizos

Eleocharis rostellata
Fimbristylis autumnalis
us Olneyi

ABE

ustus
ynchospora capitellata
Longii

olololorer:.
©

rs
r%

var.
Luzula cam i
Smilax meant lia
] ris prismatica

i
pee
g
2
s

Boehmeria cylindrica, var.
D ondiana

*Rumex verticillatus
Polygonum glaucum
P, tum
Mes intermedia

D. marilandicum
Lespedeza —
Polygala cru

Euphorbia solpranitolia

“sa
Hibiscus eepobaiee

Hyperic ginicu
Helianthemum pains

Fig hye
cr
2 oO
He
o
&

Yecodon a
exia Vir.
{yriophyllum seabratum

ier ee ee |
wee Pee © ee Ga

Hydrocotyle umbellata —
anbyi
[. verticillata
ium capillaceum
Clethra alnifo
Rhododendron viscosum, var.

ehantantal
oo

ucum

Leucothoe racemosa

amolus floribundus
vee virginica

Be

pani ta
v Leen str et lacunosum
\sclepias ve
eomtate ense, var.

2) hb be ES ES
ee Ue

ito:
Tysanthes inaequalis
Gratiola aurea

is atti
Utricularia gibba
*Plantago virginica

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 211

Eupatorium ny erostolam Pluchea camphorata

a pro ethos Gnaphalium ee)
*S. min Coreopsis rose

8. ten ifolia Krigia Vieatiie

Aster > ple Lactuca hirsuta

A. vimineus Hieracium Gronovii

A point requiring particular emphasis is, that many of these species
are by no means common on the islands. Indeed, a few of them,
such as Thelypteris simulata, Paspalum setaceum, Panicum longi-
folium, P. Commonsianum, Cenchrus pauciflorus, Diplachne mari-
tima, Eleocharis rostellata, Carex straminea, C. alata, C. Mitchelliana,
LIuzula campestris, var. echinata, Tipularia discolor, Rumex verticil-
latus, Prunus maritima, Ilex glabra, Sanicula canadensis, Hydro-
cotyle Canbyi, Rhododendron viscosum, var. glaucum, Triostewm per-
foliatum, Solidago minor and Coreopsis rosea, are known only from a
single locality, while certain others, though less restricted, are never-
theless rare and local. And seldom, if ever, are these coastal plain
plants present in sufficient abundance to create the impression,
inescapable on Cape Cod, of a southern flora transplanted almost
en masse.

Pursuing this last idea further, it will be found interesting to
contrast with the list just given a list of some of the southern coastal
plain plants which are known to occur on Cape Cod (most of them
from the Middle Cape), but which have not yet been found on the
Elizabeth Islands:

Pteridium aquilinum, var. Aristida dichotoma
pseudocaudatum . gracili :
Sagittaria Hugenaansann Spartina cynosuroides
8. i Tridens flavus
S. teres Cyperus filicinus, var.
+ aspalum psammophilum cas

anicum verrucosum C. Gra

- Bicknellii aan Robbinsii

- Microcarpon E. melanocarpa

- annulum Palscheys packs

- mattamuskeetense Scirpus atrovirens, var

- Spretum georgianus

- Wrightianum §. Eriophorum

- auburne Fuirena squarrosa

. tsugetorum Hemi ha micrantha

columbianum Rynchospora macrostachya

¥. polyanthes R. = i vm
P. Ashei R. Torre
P. scoparium R. apitellata, var. discutiens

212

Scleria reticularis
arto annectens
ntimesrens
C. bullata, var. Greenei
aema . Stewar sonii

Rhodora

Acer rubrum, var. tridens
Ceanothus americanus, var.
intermedi

Vitis cordifolia

Hypericum adpressum
Hudsonia ericoides
:
1

[OcToBER

Pelta
Orontium aquaticum Viola _emarginata
Xyris F, ulifoli:
Juncus cnr Opuntia vulgaris
J. aristulat xia mariana
Lilium su pee Oenothera linearis
Aletris farinosa longipedi t
Lachnanthes tinctoria Proserpinaca pectinata
i plenifolia Pr edia
Quercus stellata Lilaeopsis chinensis
3 a Sabatia corapatarnts
ilicif a
ees umbellata Onosmodium ' virginianum
Polygonum setaceum Stachys hyssopifolia
Polygonella articulata Lycopus sessilifolius
Acnida cannabina Agalinis purpurea
osera filiformis Aureolaria pedicularia, var.
Cassia Chamaecrista caesariensis
Crotalaria sagittalis Utricularia inflata
Lupinus U. subulata
Tephrosia virginian Viburn
Desmodium sstunidivolitis Eupatorium Merbannetoltuna
D. marilandicum Mikania scande
ag ary procumbens : solidago erecta
a wo.
Ls aneantihs lia Aster spectabilis
Sirophostyles helvola A. subulatus
mic var. A. tenuifolius
Baccharis halimifolia
Polya N Nuttalli Bidens coronata
Corema Conradi Lactuca floridana

Thus it will be seen that, while there occur on the Elizabeth Islands
something like 100 species belonging to a wideranging southwestern
flora, Cape Cod not only has practically every one of these same plants,

but boasts in addition at least an equal number of species of the same
class which, so far as is known, are totally lacking from the islands.

It may be worth while to note, in passing, that, while an overwhelm-
ingly large proportion of the more than 200 prevailingly southern
species which occur on Cape Cod occur likewise on Nantucket (and a
considerably smaller proportion on Martha’s Vineyard), that island
has caught a number of these southern migrants which appear not to
have succeeded in reaching the Cape. Several of these may be listed:

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 213

Eleocharis tricostata Ascyron hypericoides

Scleria triglomerata Lespedeza virginic

asin’ Walteriana Lechea Leggettii

Ha ria ciliaris Ludvigia alte

Quercus pagodaefolia Pyenanthemum ° cartiedibnbeais
Polygonum robustius Schwalbea americana
Amaranthus pumilus Aster concolor

ausailae: fosimtlia

Sufficient evidence has probably been adduced to bear out the
contention that the relations of the flora of the Elizabeth Islands to
that of the southern coastal plain are anything but prominently
marked, and that this southern, or southwestern, element is much
more strongly represented in the closely adjacent regions, especially
Cape Cod and Nantucket. An attempt to determine the causes
which account for this break in distribution will shortly be made.

The Northern Element. Although lacking many of the northern
types which distinguish the floras of parts of Nantucket and the
“Lower” Cape, the Elizabeth Islands are not entirely without their
representation of plants whose affinities are prevailingly boreal. In

, about 50 such species, constituting nearly 9% of the total native
flora, may be considered as belonging to this class. It is significant to
contrast this number with the 150 northern plants (over 20%) listed
by Bicknell for Nantucket.

In general, these northern species which occur on the Elizabeth
Islands are plants which range from Labrador and Newfoundland
south to Massachusetts and New Jersey or, in a few cases, Delaware
or Maryland. Many of them range south of New England along
the mountains but reach the coastwise southern limit of their distri-
bution in Massachusetts, Long Island or New Jersey. In the list
which follows those species marked with an asterisk are to be looked
upon as essentially maritime.

"Ruppia maritima, var. subeapitata carer hormathodes

oc mariti .s

ce:
e \grostis <loretien ee compacta . canescens, var. disjuncta
hau . limosa

>. lanuginosa
Sriocaulon septangulare

: uncus pelocarpus
militaris

C

KR

Glyceria o ony

G. canadensis

Puccinellia paupercula, var.
alaskana

]
Wisheharts u iglw umis J arti iculatus
Scirpus campestris, var. paludosus Bieysnahitm angustifolium
nellum Liparis Loeselii
]

<i ene
ceils populifolia

214 Rhodora [OcTOBER

*Rumex maritimus, var. fueginus Hypericum boreale
Arenaria lateriflora Epilobium palustre, var. monticola
Sagina procumbens Myriophyllum tenellum
*Ranunculus an *Ligusticum scothicum
‘Drosera rotundifo *Coelopleurum lucidum
Fragaria a var. Menyanthes trifoliata, var. minor
terrae-novae Chamaedaphne calyculata
*Potentilla pacifica Vaccinium macrocarpon
*Lathyrus maritimus Limosella subulata
Rhus ra, var. borealis *Plantago juncoides, var. decipiens
Ilex verticillata, var. fastigiata Anaphalis margaritacea

The Elizabeth Islands, then, appear to have received their share
of those far-ranging northern types which probably owe their existence
in coastal New England, and southwestward, to the former presence
of the broad continental shelf, already referred to, which permitted
of their extension to the southwest and then, following its submerg-
ence, left them stranded at isolated localities along the coast. This
would also explain why these islands possess fewer such plants than
Martha’s Vineyard (see p. 173) and still fewer than either Nantucket
or the outer portion of Cape Cod. For, if these boreal species reached
southeastern New England from off the elevated coastal bench to the
eastward, then it seems logical to assume that a greater number of
them would have found a refuge on Nantucket and the “Lower”
Cape and that a smaller proportion would have succeeded in finding
their way to the areas inland to the west, especially if, as may well
have been the case, the retreat of the glacial ice from the latter region
lagged appreciably behind its retreat from Nantucket, Martha’s
Vineyard and Cape Cod.

The Continental Element. It is only when we come to consider
the continental element as it appears on the Elizabeth Islands that
we find ourselves dealing with the type of vegetation which lends a
dominating color to their flora. Probably more than 400 plants
(about 70% of the total indigenous flora) from these islands are
neither prevailingly southern nor northern in their distributional
affinities but belong, rather, to a widespread continental flora which
might be characterized, somewhat arbitrarily, as Canadian-Alleghan-
ian in nature. This, it will be recognized immediately, is an attribute
which the Elizabeths share in common with the upper or inner part
of Cape Cod, and, indeed, there is every reason to suppose that the
flora of these islands may, until comparatively recent geologic times,
have been continuous with that of the line of hills which runs from

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 215

Falmouth to Bourne, and even north nearly to Plymouth. This
entire ridge represents a terminal moraine which, beyond Cuttyhunk,
dips below sea level reappearing, as some geologists believe, to form
Block Island and, farther west, a portion of Long Island. The
separation of the Elizabeth Islands from the mainland and their
division into the seven present members of the chain are, as has
already been indicated, relatively modern events. So it is entirely
in keeping with the past history of this region that so many of the
types which are abundant throughout the Falmouth area should
likewise be common on the islands.

The relation of the flora of the Elizabeth Islands to that of the
mainland comes out most clearly upon an examination of the forest
types. The trees have already been listed (p. 151), but it seems
entirely justifiable to repeat in this place that the native woods of
the islands are made up not only of beech, Fagus grandifolia (surely
not a coastal plain type), but also contain Carya alba, Ostrya virgin-~
iana, Quercus alba, Q. velutina, Sassafras officinale, Hamamelis virgin-
tana, Acer rubrum, Cornus florida and Nyssa sylvatica. Now these
trees, while they may and do occur on the coastal plain, are neverthe-
less more common and more clearly at home on the richer soils of the
Piedmont and the areas inland, often reaching their finest develop-
ment on wet wooded slopes and the alluvia of river valleys.
® Under the trees listed above, on the wetter parts of the forest
floor, occur such plants as Carex lupulina, Arisaema triphyllum,
Oakesia sessilifolia, Maianthemum canadense, Medeola virginiana
and Trientalis borealis. These again are types more commonly
associated with an Alleghanian woodland flora. The beech drops,
Epifagus virginiana, a rare plant in southeastern Massachusetts,
occurs everywhere in the wooded parts of Naushon and Nashawena,
and many other cases of this sort might be cited.

It would be superfluous to list here all of the species of Canadian-
Alleghanian affinities which occur on the Elizabeth Islands. An
enumeration of them would include most of the names of native
plants in the Catalog that follows which have not been listed above in
dealing either with the Southern or Northern Elements. A few of
the most typical, however, not including the few trees mentioned
above, may be given for the sake of comparison:

0. um virginianum Athyrium angustum
Asplenium platyneuron Osmunda cinnamomea

216 Rhodora [OcTroBER
Ophioglossum = Amelanchier oe
Isoetes Engelman: Geum canaden
Sparganium apiedinot Geranium maculatum
Sagittaria latifolia regen ‘ogee
Andropogon furcatus A.
Glyceria striata Callitriche r Retaedliyila
G: Rhus typhina

ilymus virginicus R. Vernix
Cyperus diandrus R. Toxicodendron
C. rivularis Tlex verticillata
Scirpus ioeid S Impatiens biflora
Eriophorum virginicum Vitis labrusca
Carex rosea, var. an a V. aestivalis
C. cep ophora Viola papilionacea
C. crinita V. pallens
C. virescens Ludvigia palustris
C. communis i a
C. pennsylvanica, var. separans Sium suave
C. digitalis eracleum lanatum
C. debilis, ~~ Rudgei Monotropa uniflora
C. lup V pope
Arisaema tri iphyllum Epigaea
Symplocarpus f oetidus G 2 ee co
Acorus Calamus pereonnoca quadrifolia
Juncus effusus, var. solutus L. terrestris
Oakesia sessilifolia 7 ental borealis
Lilium ‘hila delphicum Apocynum androsaemifolium
Mai ——— canadense Verl erbena hasta

via beactenta Pyenanthemum muticum

H. eevee P. osum —
H. orbiculata
Arethusa bulbosa it een Be occidentalis
Boehmeria cylindrica Triosteum perfolia:
Polygonum scandens Sambucus canadensis
Phytolacca americana Lobelia cardinalis
Ranunculus delphinifolius Solidago juncea
! ne vir 8. canade
Coptis groenlandica Aster divaricatus

piraea tomentosa Cirsium discolor

The great bulk of these plants are species primarily of the interior;
they attain their fullest development in the Piedmont and the Up-
lands and their occurrence on the coastal plain may be regarded, in
most cases, as rather casual.

In summing up, it need merely be pointed out that the Elizabeth
Islands, while serving, as does every other locality along the Atlantic
coast, as a meeting ground for both northern and southern species
of plants, exhibit both qualitatively and quantitatively a very strong

1930] | Fogg,—Flora of the Elizabeth Islands, Massachusetts 217

relationship with a widely dispersed flora of a continental nature, a
fact which seems readily explicable upon the basis of the close connec-
tion existing between these islands and the inner, hilly part of Cape
Cod, both as regards geologic history and general topography.

Finally, there remains to be considered, as briefly as may be, the
subject of the last glacial advance over this region and its possible
effects in influencing the present-day distribution of the flora.

Glacial History. Probably the greatest student of the geology
of southeastern Massachusetts since the days of N. S. Shaler, was
the late J. B. Woodworth of Harvard University. Professor Wood-
worth had prepared, shortly before his death, an exhaustive treatment
of the glacial history of the Cape Cod region. This manuscript,
unfortunately, still awaits publication and the details which it em-
bodies are not yet available. Happily, however, Woodworth had
related his broader conclusions to A. P. Brigham, geologist to Colgate
University, and the main arguments are set forth by Brighain in his
popular book entitled “ Cape Cod and The Old Colony.”

According to Woodworth, the advance of the last or Wisconsin ice
over southeastern Massachusetts took place not as a solid sheet,
but in the form of three tongues or lobes. One of these, the “ Buzz-
ards Bay Lobe,”’ came down over the region now occupied by Buzz-
ards Bay and deposited as a frontal moraine much of the material
which now forms the line of high hills along the northwest shore of
Martha’s Vineyard, from Menemsha to West Chop. Then, following
an interval which represented a retreat and a second advance of the
ice, this lobe laid down, as a secondary moraine, the ridge which made
the Elizabeth Islands and the “Upper Cape.” The line of the
islands, as may be seen from a map, almost exactly parallels the line
of the morainal hills on the northwest shore of the Vineyard.

The second lobe, which lay to the east of the “ Buzzards Bay Lobe,”
advanced southward over what is now the middle section of Cape Cod
and laid down, as a terminal moraine, the sand, gravel and boulders
which form the northeast shore of Martha’s Vineyard and the higher,
crescent-shaped portion of Nantucket. This Woodworth terms the
“Cape Cod Lobe.” This lobe then retreated, as did the Buzzards
Bay Lobe and, as its secondary moraine, deposited the till which com-
poses the “backbone” of Cape Cod from Sandwich to Brewster and
Orleans. Thus, Martha’s Vineyard was built by the combined
action of two lobes and the central and southern parts of the island

218 Rhodora

represent an outwash or apron plain derived from two separate
moraines. The southern and southeastern parts of Nantucket and
almost the entire south shore of the Cape likewise represent outwash
plains, both formed from the materials deposited by the Cape Cod

be.

Still further to the eastward, the third, or “South Channel Lobe”
advanced over the area now submerged and known as Georges Banks.
With the deposits of this lobe we are not so much concerned, as they
now lie mostly beneath the sea, save for such materials as may have
contributed to the building of the outer or lower part of Cape Cod.
How far this lobe may have extended eastward over the then elevated
continental shelf is apparently not definitely known.

t now becomes pertinent to inquire into the relative ages of these
deposits and as to whether any evidence is forthcoming to indicate at
what time and in what manner the various lobes retreated. Probably
Woodworth’s report, when it becomes available, will throw much
light on this question. However, the writer has it on the authority
of Dr. Wigglesworth of the Boston Society of Natural History, who
is conversant with Woodworth’s views, and who is himself a student
of the geology of Martha’s Vineyard, that in all probability the middle
or Cape Cod lobe was the first one to retreat. If this was the case,
it then means that Nantucket, the eastern part of Martha’s Vineyard
and the central part of Cape Cod were free of ice at a time when the
regions to the east and to the west were still covered by the South
Channel and Buzzards Bay lobes respectively. Remembering that
the coastal shelf was probably considerably higher at that time than
it is today, and that the Vineyard, Nantucket and the Cape may well
have been continuous dry land, it at once becomes apparent that
there was thus opened up an area which soon became available as a
refuge for that migration of southern coastal plain species of plants
which probably began as soon as the ice commenced to retreat.

t is necessary to point out here that there is a lack of complete
agreement as to the exact period of subsidence of the continental
shelf and as to whether this migration might have occurred previous
to the advent of the Wisconsin ice or whether it could not possibly
have taken place until after the glacier had receded.

Douglas Johnson, the eminent student of coastal phenomena, in
discussing the date of submergence of the Banks cuesta (i. e., the
New England-Acadian portion of the outer coastal shelf) states that

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 219

“we should expect the subsidence to be at least post-Miocene and
more probably post-Pliocene.”! And further, “It seems probable
that the date of submergence of the drowned topography must be
post-Tertiary.”? Johnson, then, is inclined to view the elevation
of the continental shelf as pre-glacial rather than post-glacial, a
condition which would have necessitated the plant migration having
antedated the advent of the Wisconsin ice. And, indeed, Fernald
sees no reason why these plants should not have moved northeast-
ward along the exposed shelf before the coming of the Wisconsin
glaciation and “have persisted outside the subsequently glaciated
area, finally taking possession of their present isolated habitats on
the receding of the ice.’’

In connection with the present study, however, it matters little
whether these species of the southern coastal plain reached the New
England area before or after the last glaciation. In either case they
must have moved inland from off the broad shelf to the eastward to
take the places left vacant for them by the recession of the ice, and
if we are justified in assuming that it was the Cape Cod Lobe of the
glacier which receded first, then we are in a position to understand
why so many of these species are to be found upon Nantucket and
the “Middle” Cape, and, to a lesser degree, upon the eastern portion
of Martha’s Vineyard and are so generally lacking from the Elizabeth
Islands and inner Cape Cod. Even if the western half of the Vineyard,
the Elizabeth Islands and the “Upper” Cape were free from ice at

e time when this migration was operative, it seems likely that they
offered a type of habitat which was less attractive to these coastal
Plain plants than the low-lying silicious areas of Nantucket and the
‘Middle” Cape which they must have reached first and where they
today abound, seldom exhibiting a tendency to widen their ranges
into the neighboring regions. And although, as Fernald suggests,

e plants may have persisted upon the outer shelf while the ice
still covered the area inland, it is nevertheless probable that the
Nantucket-“ Middle” Cape region would have been the first to wit-
hess their return.

Conclusion: In summing up, it may be said that, considered from
the viewpoint of broad, geographic origins, the native flora of the
‘Johnson, D. The New England-Acadian Shoreline. New York. 302 (1925).

“Sa “% rae f Studies on the Northeast-
Ward ie ag le cme Marg seye oo ee. Sci. 4th Ser. xl. 18 (1915).

220 Rhodora [OcroBER

Elizabeth Islands is seen to consist of three distinct elements. In
the first place, there are those species (less than 20%) which exhibit a
relationship with the flora of the southern coastal plain. Their
presence upon these islands is to be explained upon the basis of a
former land connection with New Jersey and southward which took
the form of an elevation of the outer coastal bench, now submerged,
and which, either prior to or following the Wisconsin glaciation,
permitted of the migration of plants from the southwest to the New
England area and even farther north and east. That so many more
of these southern plants occur upon Cape Cod and Nantucket than
upon the Elizabeth Islands is probably to be explained by the be-
havior of the glacial lobes which covered this area and which, by their
differential recession, seem to have rendered the former areas access-
ible to occupation by plants at an earlier date. Secondly, there is a
small percentage (less than 9%) of plants displaying a boreal affinity,
the occurrence of which may be attributed to a counter extension
southward along this same uplifted shelf. And the fact that the Eliza-
beth Islands have received a smaller number of these northern repre-
sentatives than either Cape Cod or Nantucket is probably to be
accounted for on the basis of their inland position and the character
of the habitat which they offer, which is, in general, less favorable
for these northern plants than the situations which they occupy on
the “Lower” Cape. And, finally, there is the overwhelming majority
(over 70%) of plants occurring on the Elizabeth Islands which show
an essential relationship with the flora of the mainland and which
give to the islands the dominating character of a Canadian-Alleghan-
ian region. The prevalence of this continental element is doubtless
due to the close geologic and physiographic similarity existing between
these islands and the “Upper” Cape. Thus it will be seen that the
evidence derived from a study of the geographic origin of the flora of
the Elizabeth Islands fits rather well into what is already known
concerning the history of the neighboring territory and that these
islands take their place botanically as an extension of the adjacent
mainland rather than as a link in that chain of outposts of a formerly
continuous but now highly disrupted coastal plain flora extending
from the South Atlantic States to Newfoundland.

Acknowledgements. The writer desires to express his profound
and grateful appreciation of the stimulating genius of Professor M. L.
Fernald under whose guidance this study has been conducted and
without whose inspiration it could never have been completed.

1930] Fernald,—Gentiana procera 221

Mr. C. A. Weatherby and Mr. Bayard Long have both rendered
invaluable assistance in their willingness to aid in the determination
of critical material. To them the writer’s deepest thanks are due.

He is also indebted to Professor L. H. Bailey, who has kindly
examined several specimens of Rubus, to Professor K. M. Wiegand,
who has looked over some of the Amelanchier material, and to Mrs.
Agnes Chase who has given her opinion on a few sheets of Panicum.

(To be continued)

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS

JoHN M. Foaa, Jr.
(Continued from page 221)
Part II. AnnoraTep List oF THE VASCULAR FLORA

In the Catalog of Vascular Plants of the Elizabeth Islands which
follows several conventions and abbreviations have, for the sake of
convenience, been adopted; these are here explained:

NOMENCLATURE. The International Rules of Botanical Nomen-
clature have been followed.

ELEMENTS IN THE FLORA. The various elements constituting the
flora have been differentiated thus:

Indigenous species appear in gs letters.

Introduced species are given

Discredited records are coeladen i : cone

CITATIONS AND SYNONYMY. Synonyms are given in italics and,
in general, are included only when they represent names which have
been superseded since the last edition of Gray’s Manual (Ed. 7, 1908).
Usually, in such cases, a reference is given to the place where the
new name was published or its status discussed.

LOCALITIES. The seven main islands are indicated by taking the
first three letters of the name of each; thus:

ote a including Pine Island.

NAU: Naushon including Captain’s Island, Ram’s Head, Mono-
uck, East Buck, West Buck and the Weepeckets.

PAS:
NAS: Passe
CUT: hunk,

PEN: Penikese, including Gull Island.

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 227

COLLECTORS. Those who have collected specimens from the
islands, or who have reported plants from there, are designated by
the use of the surname; this is followed by the serial collection number,
where the collector has employed a numbering system, otherwise by
the date. A list of the chief collectors is here given in alphabetical
order:

Cushman, J. A. (1906) Northrop, Alice R. (1901, 1903)
Duggar, B. M. (1911) Pennell, F. W. Ate

Faxon, C. E. (18 Sanford, 8. N. F. (1917)
Faxon, Walter (1873) Simons, Elizabeth A. (1901)
Fernald, M. L. (1927) Sipe, 8.

Hervey, E. W. (no date) Svenson, H. K. (1926)

Hollick, Arthur (1898) Taylor W.1 R. (1919-1921)
Jordan, David S. (1873) Weir, “

MacRae, Lillian (1904) Willams, Er. F. (G91)

Moore, A.H. (1904)

Further information concerning these collectors and the distribu-
tion of their material may be found in the section on Previous Botan-
ical Work on the Islands (Pages 123-125).

N. B. Serial numbers unpreceded by a name are to be taken as
representing material collected by the present writer.

HERBARIA. The following system of initials has been employed
to designate the herbaria in which specimens have been seen or to
which they are known to have been distributed:

(A) Academy of Natural Sciences of Philadelphia

(B) Boston Society of Natural History

(C) Cornell University

(G) Gray Herbarium, Harvard University

(J) Arnold Arboretum, Jamaica Plain, Mass.

(M) i Bo arde
(N) New England Botanical Club
(P) University of P ania

Muse
Marine Biological Laboratory, - Woods Hole, Mass.
(Y) N ae York eset Garden
(0) no specim

POLYPODIACEAE
PotyPpopium Mein ranuM L, P. vulgare of eastern American
ce ‘gondii mald, Ruopora, xxiv. 125 (1922). Rare; seen
‘anon a large boulder in the woods on Naushon. ‘NAU:
Hollick (, 2489 (N); NAS: Northrop (0).
Prermium aquitinum (L.) Kuhn, var. sere (Desv.)
Underwood. Pteris aquilina of the Manual. Apparently not common
Sandy woods. NON: 2672 (N); NAU:3329 (P); NAS: Northrop (0).

228 Rhodora [NovEMBER

Woopwarpia vireinica (L.) Sm. Locally abundant; forms
considerable growth around the dune ponds on Nashawena. NAU:
Williams, July 10, 1911 (N); NAS: Northrop, July, 1901 (Y), 1772
N,P).

W. areotata (L.) Moore. Locally abundant; with ie
palustris, var. pubescens occupying boggy hollows in open hills.
re a neg 8, 1926 (N); NAS: Northrop, Aug. 1901 (Y), 3492 wn. >,
CUT: 2324 (G N,P M).

hae PLATYNEURON (L.) Oakes. Reported by Mrs. North-
rop from Nashawena; no specimen seen

ATHYRIUM ANGUsTUM (Willd.) Presl. ——— ge ako of
American authors in part. See Butters, Ruopora, xix. 190 (1917).
Boggy woods and open hillsides. NAS: 3515 (P); CUT: 586 (N).

A. ANGUSTUM, var. ELATIUS (Link) Butters. Ruopora, xix. 1
ti Dry, exposed hillside. PEN: 459 (N,P,W).

LYPTERIS PALUSTRIS (Salisb.) Schott, var. PUBESCENS (Lawson)
Fernald Aspidium Thelypteris of Manual. See Fernald, RHopora,
xxxi. 34 (1929). oe n in low, boggy ground and ated borders
of as NON: 2238 (N); UNC: 2993 (N); NAU e, Aug. 1901
(W); NAS: Norticcs (0), 2349 (N,P); CUT: okey yee 15, 1917
O, 2323 (N); PEN: 460 ms P,W,M).

nepal

SIMULATA (Dav.) N ieuwl. Aspidium simulatum Dav. See
Weathaby: RHODORA, Xxi. 1 74, 178 eg hee rare on the
islands. NAU: Sipe, July, 1901 (W );P enson, Sept. 8, 1926 (N).

T. NOVEBORACENSIS (L.) Nieuw ate ense

y iter ium J
Sw. Moist woodland. NON: pas (G,N,P,M,C); NAS: Northrop
(0), 3514 (P).

T. sprnutosa (O. F. Miiller) Nieuwl. Aspidium spinulosum (O. F.
Miiller) Sw. Reported by Mrs. Northrop from Nashawena; no
o— seen.

NSTAEDTIA PUNCTILOBULA (Michx.) Moore. Dicksonia pune-
tilobula rh oar Gray. Rocky, open hillsides. NAS: Northrop (0),
3509 (P); CUT: 2534 N, P); PEN: 461 (N,P,W W,M).

es SENSIBILIS L. Apparently not abundant. NAS: North-
rop (o); CUT: Sanford, Aug. 15, 1917 (N).

OSMUNDACEAE
OsmunDA REGALIs L., SPECTABILIS (Willd.) Gray. Occasional
in open bogs. NAS: Novtheop (0), 1777 (P).
O. cinnamomea L. Bogs and wet hollows. NON: 2881 (N);
UNC: 3029 (N); NAS: ican (o), 3513 (P); CUT: 3450 (P).

ee

OPHIOGLOssUM VULGATUM L. d only in a sandy field at the
east end of Cuttyhunk, near rhonictac! Pond. NAS: Northrop (0);
CUT: 3582 (N,P).

1930] Fogg.—Flora of the Elizabeth Islands, Massachusetts 229

[BoTRYCHIUM MATRICARIAE (Schrank) Spreng. is reported by Mrs.
Northrop for Nashawena. This may refer to B. dissectum Spreng.
or one of its allies, but in the absence of herbarium material it seems
best to disregard the record entirely. ]

EQUISETACEAE

EQUISETUM ARVENSE L. Seen only in a boggy clearing in the woods
at the east end of Naushon. NAU:701 (P,W)

LYCOPODIACEAE

LYCoPoDIUM INUNDATUM L., var. BiaeLovit Tuckerm. Peaty mar-
gins of ponds. NON: 3381 (N); NAU: 3865 (P).

ISOETACEAE

Isonrres ENGELMANNI A. Br. A single sheet of this species papa
7 “oe ahi by C. E. Faxon, but bearing no date, is in the

er

PINACEAE
InUs Strosus L. Reported by Mrs. Northrop from Nashawena,
sion probably introduced; no specimen seen
.RIGIDA Mill. A small group of these trees grows at the extreme
east end of Naushon near the West ear Seen nowhere else,
although perhaps more common. NAU: 3425 (N,P).

P. sylvestris L. Scotch Pine has been ‘apa duced on several of the
islands but seems nowhere to be spreading. NAU: 3871 (N); NAS:
Northrop (0), 3482 (P); PEN: 462 (W).

Larix decidua Mil anted at several localities along the north

e
NAU: 3602 (J.P).
glauca Voss. P. canadensis B. S. P. NAU: 3690 (J,P), 3691
(J, P),” 3869 (J,N
- pungens Engelm. NAU: 3868 (J,P).

Cuamarcyparis THyowes (L.) B.S. P. Of infrequent occurrence
on the islands, although plentiful pit a large pond near Tarpaulin
Cove on Naushon. NAU: Williams, July 10, 1911 (N), 3884 (P,M,C).

UNIPERUS COMMUNIS L., var. DEPRESSA Pursh. Collected only on
the north shore of Naushon, where it may possibly have been intro-
duced. NAU: 3872 (N,P).

J. vinctntana L. Plentiful in the woods near the East Gutter on |

Nonamesset; not seen elsewhere. NON: 2668 (N, P); NAS: Northrop

(0).

TYPHACEAE
yy AXPHA Lact AL. Swamps and ns borders. NON: 2675 (N);
AS: Northrop (o); PEN: 463 (N.W).

230 ‘Rhodora [NOVEMBER |

T. ancustirotia L. Mostly in brackish areas near the shore.
NON: 3165 (N); NAU: 3887 (P); NAS: Northrop (0), 1766 (P).

SPARGANIACEAE

ene ge EURYCARPUM Engelm. Found at two stations; both
mpy hollows in woods at east end of Naushon. NAU:
ies (N, P),: 3678 (P, re

S. ANDROCLADUM (Engelm.) eae S. lucidum Fernald & Eames.
See Fernald, Ruopora, end 27 (1922). Growing " Bye 2 area
at east ie of Naushon near Hadley Harbor. NAU: 2469 (N,P).

S. AMERICANUM Nutt. eating var. androcladum ‘a sci Manual.
Sw fait oods and pond holes er common. NON: 2904 (N,P);
UNC: 3015 (N, P), 3113 (P); NAU: 2382 (P,C), 2383 (N,P,M); NAS:
3527 (N,P,M).

NAJADACEAE

PoTaMoGETON OaKEsIANus Robbins. In a small pool bordered by
Decodon verticillatus along the north shore of Naushon, near the east
end. y Nes 3867 (N,P,M).

b AD ER Tuckerm. The commonest Pondweed on the islands;
met with feotaan tly} in the small pond holes, where it often continues
to grow upon the mud after ponds become dry in late summer.
This species produces fruit rather abundantly. UNC: 2997 (G,N,P),
3013 (P,M); ey : 711 (P); PAS: 710 (P,M); NAS: Fazon, C. E.,
no oe eee 5 ON),

P. BUPLEUROIDES Fernald. Found growing abundantly in Gosnold

+ Wes t End Pond on Cuttyhunk on July 15, 1925. This pond,
foribaetd fresh, had evidently been inundated by the the preceding
winter; it mot remained brackish ever since. CUT: 1001 (P,

[P. pustLius L. is reported by Mrs. Northrop for N sahatwenik but
no specimen mg tes been seen. Riasanech as P. pusillus has been @
source of considerable confusion and there is no way of deciding the
correct application of this name without material this record can no
be allowed

P. piversiroiius Raf. P. hybridus of the Manual, in part. Col-
lected only from the sandy borders of the dune ponds on Nashawena,
but spatasng more widespread. NAS: 1083 (P, W), 1775 (N,P).

ae us L. Abundant in a somewhat brackish pond along
the “alee pee of Nineineact should be found elsewhere. NON:
2875 (N,P).

Ruppia MARITIMA L., var. LONGIPES Hagstrém. See Fernald &
Wiegand, aay RA, xvi. 125 (1914). ‘NON: 2246, (N,P,M,C), 2876
(N); NAU: Pennell 3151 (W); CUT: 2333 (G,N,P,M,C).

All of the material of R. maritima with long podogynes (1-6 cm.)
seems to have, at least in maturity, peduncles which are well over 3
em. in length and which are often much longer and usually spiralled.

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 231

If var. rostrata Agardh, with mature peduncles 0.5-3 cm. long, occurs
on these islands it has still to be collected.

eg MARITIMA L., var. SUBCAPITATA Fernald & Wiegand. R#opora,

126 (1914). NON: 2251 (N,P); NAU: Duggar, July, 1911 (G);
Pennell 3152 (P,W); PAS: 712 (N,P,W).

This variety with very short podogynes (2-6 mm.) was originally
described from material collected on Naushon. Its range has since
been extended from Quebec and Prince Edward Island to Block
Island, Rhode Island.

ZOSTERA MARINA L. Frequent in the shallow coves and small
tidal streams. NON: 2285 (G, N,P); NAS: Northrop (0); PEN: 464
(W), 1762 (N,P

Material collected in Sheep Pen Harbor on Nonamesset and from
near the landing on Penikese is remarkable for the copiousness with
which it produces flowers and fruits; this fact has been observed
without fail for several seasons.

JUNCAGINACEAE
TRIGLocHIN MaRITIMA L. Apparently rare on the Elizabeth
Islands, although common in brackish marshes on the adjacent main-
land. NAS: Northrop, Aug. 1901 (Y).

ALISMACEAE
SAGITTARIA LATIFOLIA Willd. Not common; swampy hollow in
woods near Tarpaulin Cove. NAU: 2384 (N).
S. Latiron1a Willd., forma HastaTa (Pursh) Robinson. Form
with very narrow, acute leaves. UNC: 3003 (N).

HYDROCHARITACEAE

VALLISNERIA AMERICANA pes V. spiralis of Manual; see
Fernald, Ruopora, xx. 108 (1918). Abundant in “ ee Pond”

near west end of Cuttyhunk. CUT: 2532 (G,N,P,

GRAMINEAE

POGON scopaRius Michx., var. FREQUENS Hubbard. See
. xix. 103 (1917). Abundant everywhere o UNG: hills and
sandy lowlands. NON: 2585 & C), 2911 ee C: 3109 (N),
NAU: 3924 (N,P,M); PAS: Sept. 8, 1926 (ND, 3753 (P);

seaetis var. littoralis Non Hitche. See Ruop
: xix. 103 (igi7), Collected on the white sand beach bordering West
nd Pond on Naushon. NAU: 2940 (G,N,P).

232 Rhodora [NoVEMBER

A. GLomEeRATUs (Walt.) B. S. P. Seen only in a low brackish
marsh tes be east end of Naushon; probably more abundant. NA
3333 (N,P

A, Mico L. Open grassland and pond borders: material
collected on Nonamesset has Pg roan hairy lower leaves and

eaths. NON: 3915 (N,P,M,C); UNC: 3118 (N).

A. Furcatus Muhl. Sandy fields ee clearings; not common.
bey 3011 (N,P); NAU: 2603 (N,P).

Digitaria sanguinalis (L.) Scop. Peete sandy areas. UNC:
3107 (P,M), 3053 (N,P); PEN: Jordan
ASPALUM SETACEUM Michx. Found growing abundantly on the
exposed grassy slopes at the extreme east end of Nashawena. NAS:
3549 (N,P).

This grass, rare in New England, is known elsewhere in Massa-
chusetts only from Nantucket. See Weatherby, Ruopora, xxx. 133
(1928).

P. pusescens Muhl. Including P. Muhlenbergit Nash. See

Weatherby, Ruopora, xxx. 134 (1928). Abundant everywhere on
open grassy hillsides. UNC: 2967 (N,P), 3642 (P); NAU: 2599 (N),

3885 (P,C); CUT: Sanford, ~ 15, 1917 (N), 2538 (N,P,M).

PANICUM CAPILLARE Seen growing only ace border of salt
marsh at west end of Nonamesset. NON: 2899 (N,P,M

¥. ay ere tala Michx. Occasional in moist grassy hollows.
NON: P)

FP. barat um L., var. sPissuM Linder. Ruopora, xxiv. 14 (1922).
This pera with ‘densely tufted culms seems entirely to replace the
species on the Elizabeth pga Rg onry: in open grassland, wet
hollows and sandy clearings. 32 (P); UNC: 3030 (N,P. aM)
NAU: 3908 (N,P); PAS: ie <P, c, 3786 (N,P,M); NAS:
(N,P); PEN: 1093 (P,W).

P. toneirotium Torr. This species, known elsewhere in Massa-
chusetts only from Marion, Plymouth County, was collected in the
peaty and boggy hollows at the east end of Pasque on a 28, 1928.
See Fogs, RuHoporRA, xxxi. 39 (1929). PAS: 3787 (N,P,W

DEPAUPERATUM .. Var. PSILOPHYLLUM Fernald. “RHODORA,
xxiii. 193 (1921). This | species is represented on the islands only by
the smooth-leaved variety. Grassy and sandy slopes. NON: 2579
(N); NAU: sos sett P), 3319 ue
picuoromuM L. Mostly in open sandy woods. NON: 2650
), 3423 ®, MD). 'N AU: 2478 iN, W), 2483 (P), 3906 (N,P,C).
INDHEIMERI Nash, var. FascicuLatum (Torr.) Fernald.
P. ay ee re Ashe. P. huachucae, var. silvicola Hitche. & Chase.
tennesseense Ashe. See Fernald, Ruopora, xxiii. 223 ant
NON: 2295 (N,P,W), 2301 (W,C), 2589 (P), 2633 (P,M), 2648 (P,W),
2660 (N), a 76 (P), sake ,U), 3375 (P), 3376 (P,U), 3420 (P,U);
NAU: 2931 (N,P), 3903 (P,U); CUT: 2266 (N,P,M), 3433 (P,U,C),
3456 Pt UD): PEN: 478 (N,P,W), 3405 (P).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 233

This is the most ubiquitous and the most variable Panicum on the
islands. Of the specimens cited Mrs. Agnes Chase, who has kindly
examined some of the material from the Elizabeths, has designated
no. 3374 as P. implicatum Scribn., no. 3420 as typical P. huachucae
Ashe, and nos. 3376, 3433, 3456 and 3903 as P. huachucae, var. silvi-
cola Hitche. & Chase. In speaking of no. 3420 Mrs. Chase says,
“The pubescence on upper surface of blades varies from very scant
or almost wanting to fairly copious.” Thus, degree of pubescence
varies, not only as between separate plants, but in different parts of
the same plant and, in view of the many intermediates which make it
impossible here to draw sharp specific or even varietal lines, it seems
best to follow the more conservative treatment of this species pro-
posed by Fernald, l. c. p. 226. Viewed in this light, the Elizabeth
Islands material with spikelets 1.5-2.1 mm. long and leaf-blades
short-pilose to glabrous above, falls into the variety fasciculatum.

P. MERIDIONALE Ashe. oa ieee on a _— slopes.
Often very slender with leaves not 1.5 m NON: 2582
(P); NAU: Taylor, July 24, 1917 (P): NAS: D452. iN), 3550 (N,P);
CUT: 2537 (N,P); PEN: 3404 (P).

- ALBEMARLENSE Ashe. This southern species, found abundantly
recent years on the “ Middle” Cape, is known from two secre =
the E Elizabeth Islands, both of them exposed sandy areas. UNC:3
(N,P); CUT: 2264 (N,P,W).
ue NoMa ie & Chase. Sandy woods, hillsides and pond
718 (W), 2492. (P), 27 i (N), 3681 (P); NAS: 2359
a) 3466 ®). “DEN. 479 (P,W).

The lines between this and P. meridionale are not always sharp;
in fact, the two behave more like varieties of one polymorphic species.
Differences in size of spikelets are trivial, pubescence cannot be relied
upon, and habit varies with the situation in which these plants grow.
To be sure, there is a well marked meridionale extreme and an equally
characteristic oricola one, but between them an almost complete
Series of integradations.

P. Commonstanum Ashe. Met with only on an open sandy oe
Na the copies shore of Naushon, west of Kettle Cove. NAU:3

SPHAEROCARPON Ell. Occasional on open downs. NON:
2647 (N,W), sal (Ps NAS: 3556 (N). ee
- CLANDES' andy woods and clearings; not common.
UNC: 3041 (N); WAU.3148 W, P).

Echinochloa Crusgalli (L.) Be: A weed near cultivated areas.

UNC: 3063 (N); NAS: Northrop Gs PEN: 1092 (P, W).

234 Rhodora [NovEMBER

E. Watrert (Pursh) Nash. Moist, usually brackish, situations.
NON: 3188 (N); UNC: 2992 (N); NAU: 3156 (C), 3888 (P,M,C).

SETARIA GENICULATA (Lam.) Beauv. See Hitche. Contr. U. S.
Nat. Herb. xxii. 168 (1920). Open grassland and around borders of
brackish areas. NON: 2695 (N,P); UNC: 3033 (N,P), 3103 (N,P).

This species is in urgent need of further study, for even a casual
examination of a large series of specimens is sufficient to indicate
that our perennial Foxtail of the salt marshes of New England is
distinct from the far-ranging southern plant which occurs in tropical
and subtropical America.

S. glauca (L.) Beauv. See Ruopora, xxxi. 109, 110 (1929). Dis-
turbed sandy area near beach. UNC: 3098 (N).

S. viridis (L.) Beauv. Reported by Jordan from Penikese; no
specimen.

CENCHRUS PAUCIFLORUS Benth. C. carolinianus of Manual. See
Chase, Contrib. U. S. Nat. Herb., xxii. 67 (1920). Seen only in sandy
field near east end of Pasque. PAS:3766 (N,P,M,C).

[C. rrrBULOWEs L., reported by Mrs. Northrop from Nashawena,
probably refers to the preceding. C. tribuloides is unknown east of
Long Island.]

LEERSIA ORYZOIDES (L.) Sw., forma INCLUSA (Wiesb.) Fogg. See
Ruopora, xxx. 81 (1928). The only material of the Rice Cut-grass
seen on the islands grew on the sandy margin of Mary’s Lake and had
the panicles enclosed within the upper leaf-sheaths. NAU: 3135
(G,N,P,M).

Phalaris canariensis L. Sandy area near landing at east end of
Pasque. PAS: 1794 (P).

‘Anthoxanthum odoratum L. Everywhere in open grassland. NON:
2250 (N,P,M); NAU: 3315 (P,C); PEN: 468 (N,P,W,M).

SrrpA AVENACEA L. Apparently rare and local. NAS: Northrop

ARISTIDA PURPURASCENS Poir. Abundant on exposed hillsides,
especially on Naushon. NAU: 1010 (P,W), 3140 (N), 3683 (P).

Phleum pratense L. Seemingly only on outer islands, where abund-
ant on open slopes. CUT: 2280 (N); PEN: 480 (W).

AGROSTIS STOLONIFERA L. A. alba of most American authors, not

. See Malte, Canada Dept. Mines, Bull. no. 50. (1926). Grassland
and sandy fields. NON: 2653 (N), 3418 (P); PAS: 3772 (P); NAS:

(Lam.) Mey. See Malte, 1. c. aches and brackish situations.
NON: 2721 (P), 3417 (N); NAU: 2713 (N); PEN: 3397 (N,P,M,C)
A. TENuIS Sibth. A. vulgaris With. A. alba, var. vulgaris (Wi
Thurb. Doubtfully A. capillaris L. See Malte, 1. c. Dry hillsides
and hollows. NON: 3352 (P); PEN: 466 (P,W), 3391 (P)

.

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 235

A. nyematis (Walt.) B. S. P. Sandy woods; occasional. NAU:
Pennell 3124 (W,P), 3918 (N).

A. PERENNANS (Walt.) Tuckerm. Sandy woods and clearings
more frequent than the preceding. NAU: 2679 (P), 3881 (N, p),
3904 (N).

CALAMAGROSTIS oe a Beauv. Wet hollows and
borders of ponds. NON: Ds NAS: 1007 (P,W), Williams,
July 10, 1911 (N); CUT: Setar: §

C. CINNOIDES (Muhl.) Barton. oe only in a boggy thicket near
Job’s Neck. NAU: 2604 (N,P).

AMMOPHILA _iygtber ag Fernald. A. arenaria of Manual.
See Fernald, Ruopora, xxii. 70 (1920). pee i? ~sigred beaches.
NON: 2698 (N, P,M,C); UN c: 3058 (N); PEN

clearings. NON: 2291 (N,P); NAU 2709 (N,P); NAS: Northrop
(0); CUT: Taylor 2408 (P), 3455 @, 0) - PEN: 477 (N,P,W,M).

cc geet Sere (L.) Trin. Dry woods near Tarpaulin
Cove. NAU: 1009 (N,P,W).
Avena sativa L. Sandy sig near Laer ree area. PEN: 469 (W).

veabnatans spicaTA (L.) B ant on dry knolls and
open clearings. NON: 2249 \N, P) : oe pa . PEN: 473 (P,W).

Several trends are evident in material of this species from these
islands and elsewhere in eastern North America. Specimens collected
from the beach near French Watering Place, Naushon, for example,
have lemmas with the aristate teeth of D. sericea, although lacking
the pubescence of that species. This material can be satisfactorily
referred to nothing included in our existing treatments of the genus.
It is hoped that the intensive study of our American material already
embarked upon by the writer may lead to a better understanding of
these puzzling forms. A large series of collections is being held for
critical examination and comparison with types, but the preceding
may be referred to as fairly typical of D. spicata as it is commonly
understood.

SpartiNnA MICHAUXIANA ag Boggy hollows and pon
margins, often near shore. NON: 2908 (N); NAU: 2380 (N); NAS:
2341 (N,P); CUT: 3440 (P).

S. ALTERNIFLORA Loisel., var. PILOSA (Merrill) Fernald. S. glabra,
var. pilosa Merrill. See Fernald, Ruopora, xviii. 179 (1916). Brack-
ish marshes and upper borders of beaches. NON: 2896 (P), 2909
emt (N,P,M); UNC: 3095 (N, ®): NAU: 3685 (N); PAS: Suenson,
2

or Aine ) Muhl. Brackish ae ice fairly common. NON:
M,C); NAS: Northrop (0); PEN: 483
8. estas (Ait.) Muhl., var. suncea (Michx.) Hitche. Seen grow-

236 Rhodora [NovEMBER

ing only on sandy upper beach on Uncatena. NON: Cushman 126
(B); UNC: 3049 (N).

S. parens (Ait.) Muhl., var. cazsprrosa (A. A. Eaton) Hitche.
This variety, characterized by its densely tufted habit, was collec ted
in a brackish marsh near Tarpaulin Cove, where it formed a dense
turf. NAU: 2377 (N,P).

DIPLAcHNE MaRITIMA Bickn. Bull. Torr. Bot. Cl. xxxv. 195 (1908.)
sao (NP. only este a brackish ditch near Tarpaulin Cove. NAU:
3889 (N

PHRAGMITES COMMUNIS Trin. Occasional in swampy situations.
NAU: 1803 OP, gts 2608 (P); NAS: W. Faxon, Aug. 1873 (G),
N sang Aug. 1

TRIPLASIS Savane (Walt.) Chapm. Abundant on beaches and
low sa sahligiatae tches. NON: 2871 (N,P,W,M,C); UNC: 3009 (N,P,M);
NAU: 2510 (N), 2732 (P); PAS: 3754 (P).

Pm. rae: ACEA ae Steud. Arora every where

open grassland and in sandy clearings, except on the outer islands.
NON: 2580 (P.V W) 3167 (N); ONC: 3651 (P): NAU: 3919 (N,P,C);
PAS: 3765 (P,M); NAS: Northrop, Aug. 1901 (Y).

An intensive search has failed completely to reveal var. spectabilis
Gray with smooth sheaths, which is abundant on Martha’s Mes:
yard and Nantucket.

Pte beeen raion ie (L.) Greene. With Spartina patens in brackish

arshes 901 (P), — ay) UNC: 2694 (P), 3097 (N);
PAS: 3737 (PM; PEN: 1091 (P,W

Dactylis glomerata L. Sandy soil near cultivated area on Penikese;
NE Wil va seen, although probably more general. PEN: 472
Poa annua L. Not com Disturbed sandy areas on two of
the carpe pe 3311 (P); PEN: 3389 (P).

[P. s a Ehrh. Reported by Jordan from Penikese. This
may pire to oes palustris L. (P. triflora ee but, in the absence of

— ne — best to ene the record.|

Fair on on gti downs. NON: 3249
ww P), 3351 ‘®); PEN: ist 4s PW), 3388 (N,P,M).

Gtycerta opTusA (Muhl.) Trin. Pond borders and moist hollows.
NAU: 2934 (N,P); NAS: Northrop, Aug. 1901 (Y), 2348 (N,P,C).

G. ne (Michx.) Trin. Border of Decodon swamp near

arpa
G. pinkie (Lam.) Hitche. a snereata (Willd.) Trin. Boggy
woods; oS pain oC a 7 abit 79 (N).
G.P A (Torr.) T woods and wet hollows. NON:
2575 (NP), ‘3358 ‘®), 3371 ey “NAU: 2470 @; NAS: 3525 (N,P,M,
ae Total pends ‘dae ere almost apn gta t wee in
haw

eral sm n Nonamesset and one NON:
2243 (GN): 3844 ‘es 3380 (Px C); NAS: 3524 (N, P,M).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 237

[PUCCINELLIA MARITIMA (Huds.) Parl. Reported by Jordan from
Penikese but probably refers to the following which, though unknown
from Rites. grows on the neighboring island of Cuttyhun

P. paupERcULA (Holm) Fernald & Weatherby, var. ALASKANA
(Scribn. & Merr.) Fernald & Weatherby. RuHopora, xviii. 18 (1916).
NP.Wa on mud flats bordering Cuttyhunk Pond. CUT: 3572

»P,W,M,C).

stuca it yuros L. Edge of sandy road leading to Tarpaulin Cove.
NAU. 2509 (N,P).
F. octortora Walt. Apparently not general. NON: Pennell
2847 (W).

F. rupra L. ta of Ag commonest grasses on the islands. Gravel
banks, borders of beac open downs, sandy clearings, etc.

2845 (W,P), 3349 (P), per (N,P), 3370 (P), 3416 (P); NAU: o712
(N,P,M), 3877 (N,P,C); CUT: 2279 (N,P); PEN: 476 (N,

Extremely variable as to habit, texture of hae and degree of

—

AL. Reported by Mrs. Northrop from Nashawena and by
Tacan. rian Penikese. No material seen, but this record is allowed
to stand for the present.
. elatior L. Found only on Penikese, where abundant in certain
areas, PEN: 4 475 (P,W).

Bromus secalinus L. Growing in dry sandy ground along east
shore of Penikese. PEN: 470 (W).

B. hordeac . Sige ae sandy areas on Nonamesset and along
path near landing on Penikese. NON: Pennell 2839 (W), 3304
(N,P), 3347 (P.M C): PEN: 3412 (N,P).

Agropyron repens L. A pparently not abundant except on pod
islands, where it takes s possession of large sandy areas. CUT: 2
(N,P,M); PEN: 465 (W), 3387 (N,P).

LYMUS virGINicus L. Collected only on Gull Island, Penikese,
Where the plant is unusually robust. Reported by Jordan from
Penikese proper, but not seen there recently. P°PEN: 474 (N,P,W).

virGinicus L., var. HALOPHILUS (Bickn.) Wiegand. RHopoRA
xx. 83 (1918). Dry sandy banks facing the sea at west end of Un-
catena. UNC: 3119 (N), 3648 (P).

CYPERACEAE

CyPERUs pIANDRUS Torr. Wet depressions and pond borders.
NON: 3198 8 (N); UNC: 3021 (N); watt ae (N,P.W):

C. rivutarts Kunth. Seen only with the preceding on
grassy bank bordering French Watering ge Mis NAU: 2739 (N):;
NAS. Northro

- FILICINUS Vahl. C. — Eddy. Borders of fresh ponds and
low brackish areas. NON: 2700 (N.P), 2889 (P); UNC: 3007 (N,P),
3048 (P); NAU: Cushman 313 (B), 2725 (N), 3864 (P).

238 Rhodora [Nova

C. pentatus Torr. Not common; sandy aiynes a several void
on Naushon. NAU: 2467 (C), 2920 (P, M), 2935 (N,P

C,. ERYTHRORHIZOS Muhl. Reported by Mrs. Nsiat aiak for Nasha-
wena; no eae seen.

. FE eee Peete of brackish marshes and low swales; not
common. “NON 897 (N); NAU: 2369 (G,P,C), 3155 (N, P,W,M).

C. STRIGOSUS L Of “ai frequent occurrence in pond-holes
borders of salt marshes, upper beaches, and moist clearings. NO
2665 (P); oe 2982 (P), 091 (N,P), 3057 (W): "PAS: 3781 (P).

C. strigosus L., var. RopusTiIoR Kunth. Moist depression in
open hills. UNC: 3083 (N,P,W)

C. stricosus L., var. composirus Britton. Small pond-hole at
west end of island. UNC: 3020 (N,P,W

C. rruicutmis Vahl, var. MACILENTUS Fernald. Dry open ground

and sandy clearings everywhere. NON: 2587 (N); UNC: 3061 OM);
NAU: 742 (P,W); PAS: 3769 (N); NAS: Northrop, as “C. filiculm
(0), on (P); CUT: Sanford, Aug. 15, 1917 (N), 2265 (P,M).

HIUM See um (L.) Britton. Swampy _ hollows.
NAU: 745 (W); PAS: 744 (W); NAS: Northrop, Aug. 1901 (Y), 2348
(N,P); CUT: Sanford, ea 15, 1917 (N

ELEOCHARIS PARVULA (R. & S.) Link. Sei cirpus nanus Spreng.
See Svenson, Ruopora, xxxi. 168 (1929). Brackish ponds and salt
marshes. NON: 2287 (P,M), 2708 (N,P); UNC: 3050 (P); NAU:
1804 (N); NAS: Pennell 2924 (W

E. ostusa (Willd.) Schultes. “Moist hollows and pond margins.
NON: 2574 (P,M), 3204 (N,P); NAS: Northrop, Aug. 1901, labelled

“E. ovata” (Y); CUT: a sept Aug. 15, 1917 (N); PEN: 485 (W).

E. actcutaris (L.) R. & S. Peaty a nd sandy pond _ borders:
re mea ch not tr ge ‘NON: 2913 (N, Pp); NAU: 3507 (N,P,M);

ordan
E. SMAtuit eition. ey Fernald & Brackett, Ruopora, xxxi.
57-77 (1929). bing Ge d boggy pond m and low swampy

argins
areas. ea 002 NP) NAS 3498 P,W);
in 340 (PW). (N,P); (NPM); CUT: 746 (P,W)

. Bosca (Link) Schultes, var. ie ae Fernald & Brace
y or peaty margins of brackish ponds; rather frequen

NON: 2703 (N,P,M,C), 2912 (P,M): NAU: 1028 (N.P,W), 3316 (P):

NAS: Williams, July 10, 1911 (N), 1029 (P,W).

ea — Let ies et infrequent. pe
marsh a tte

cok eee Rey a north shore of Naushon sant Ke

aaron AUTUMNALIS (L.) R F. Frankii Steud.

Blake, 25 ( & S. fe |
HODORA, xx. 25 (1918). G 31
(NSP); NAU. 2138 WN. tes rassy pond borders. NON:

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 239

ScIRPUS AMERICANUS Pers. Ev verywhere bordering ponds, both
fresh and saline. NON: 2239 (P), 2888 (M); UNC: 2998 (P), 3025

S, Otneyt aod Seen growing at edge of salt marsh along south
shore of Nonamesset; probably more general. NON: 2898 (N,P);
NAS: Northrop, Aug. 1901 (Y).

S. VALIDUS Vahl. sneer in wet depressions and swa rate.
borders of ponds. NON: 44 (N,P); UNC: 2999 (N,P); N
1033 (P,W); ee rer ang as “S. lacustris” (0); CUT: 3573 (P);
PEN: 488 (N,P,W,M).

S. ROBUSTUS Pursh. Notcommon. NAS: Northrop, Aug. 1901 (Y )

- CAMPESTRIS Britton, var. paLuposus (A. Ne son) Fernald
Occasional in brackish marshes. NON: 2288 (P,M), 3209 (N, P);
PEN: 487 (N,P,W,M M,C).

. CAMPESTRIS Britton, var. NOVAE-ANGLIAE (Britton) Fernald.
Collected ripe: from swampy clearing in woods near Job’s Neck.
NAU: 3154 (N,P).

S. cyPerinus (L.) Kunth. Everywhere in wet depressions and
around ponds. NON: 2566 (N, P), 3169 (P,M); UNC: 2990 (N,P);
NAU: 2922 2 (N,P,M), 3149 (P); PAS: Svenson, Sept. 8, 1926 (N),
3791 (C); NAS Tee ©), 2340 (N,P); CUT: 2308 (N, P).

S. pee (L.) Kunth, var. pettus Fernald. Found only along
the edge of a swamp tle a eben Cove. NAU: 2395 (N).

ERIOPHORUM TENELLUM Nutt. Occasional in boggy situations.
NAU: Pennell 2905 (P); NAS: Northrop, Aug. 1901, labelled “E.

30 set

E. vireintcum L. Rather frequent in open poe and i pee
hollows. NON: 2879 (N,P.M), 3180 (P); PAS: 3792 (N,P), 8
Sept. 8, 1926 N), NAS: N netoe (o), 3497 (P); CUT: Sanford, A oe
17 (P,C).

Ryncuos?ora FUSCA a 5 Ait. f. Apparently oy seen only in a
low, boggy hollow near Tarpaulin Cove. NAU: 2512 (P,W).

» ALBA 1. Open bogs and peaty Rs seemingly
hot abundant, although plentiful in an extensive bog at the a end of
Cuttyhunk. NAS: Northrop, Aug. 1901 (Y); CUT: 2319 (N,

pVAPITELLATA (Michx.) Vahl. R. glomerata of Manual § in “part.
See Blake, ee xx. 25 (1918). Much ei eae than the

on pond-! oles and pond bo NON: 2644 (N);
UNC: 3014 is PD, 3116 (P,M,C); NAU: ba (N); PAS: 1782 (N'P),
NAS: 3471 (P),

T SCUS MARISCOIDES rape Kuntze. Cladium mariscoides

bei ‘ba acy xxv. 49 (1923). Mostly on sandy beaches

29 NAU: 2924 (N); PAS: 3776 (P); NAS: Pennell
17 (W), 2339 (N); CUT: 10 40 (P,W).

240 Rhodora [NovEMBER

CAREX SCOPARIA Schkuhr. Not common on the Elizabeth Islands.
Dryish border of small pond. NON: 3384 (N,P,M,C); PEN: Jordan
(0).

C. scoparia Schkuhr, var. SUBTURBINATA Fernald & Wiegand.
Ruopora xiv. 115 (1912). er around margin of Decodon swamp
near Tarpaulin Cove. NAU: 2399 (N).

C. Lonemt Mackenzie. C. See of Manual. See Macken-
~ Bull. Torrey Bot. Club, xlix. 372 (1922). Frequent in open

assland. UNC: 3078 (N), og (P): tae Ae: (P); CUT: Sanford,
y OP 15, 1917 (N), 1013 (N,P,W,M); PEN: 848 (N,P,W).

C. sTRAMINEA Willd. This is the plant with obovate perigynia
selected by Mackenzie as true C. straminea; see Bull. Torrey Bot.
Club, xlii. "603 (1915). It has been found growing only on the edge
of a bog at the west end of Cuttyhunk. CUT: 2327 (N,P).

C. a iy Fernald. Moist, ouieiaaally brackish, situa-
tions as well as slopes near ahiorea: NON: 2297 (N); U.
3644 (P); NAS: Williams, July 10, 1911 (N); CUT: 726 (P,W); PEN:
pies WN, hg

Dry, sandy banks and hillsides. PAS:
ws, P); ‘CUT: hers ‘N), 2267 (P,M,C); PEN: 1497 (P), 1498 ON PW).
. ALATA Torr. Plentiful around ee an extensive boggy area
a the center of Nashawena. NAS: 2 ;

C, FESTUCACEA erays This is te iiall Eebed sedge retained
by Mackenzie as genuine C. festucacea. See Bull. Torrey Bot. Cre:
a a ” 15). Adoanentiy rare and local on the islands. NA

eo Hower Mackenzie. C. poides, var. capillacea o-
Fernald See oe apa A setae Torey ate Club, ee 245 (1910).
Abu po ers. NO 4

NAS: Pennell 2922 (W), 3542 (N.P): C
C. incomPERTA Bicknell. Bull. To: sls nice Club, xxxv. 494 (1908).

C. srorsa E. C. Howe. Seen growing only in a boggy wooded
nova on the south shore of Nonamesset. NON: 3379 tN).
ant.

EP
(Baily) Fern. See Bull. Torrey Bot. Cl b, 1. 346 (1923). B woods
at east end of Nashawena. NAS: 3 4 ae
. CANESCENS L., var. DISJUNCTA sha ald. Boggy woods and wet

Bey
of ponds. NON: 3378 (N), 3398 (P); NAU: Williams, July _

borders
10, 1911 (N), 1019 (P,W
- BRUNNESCENS Poir., var. SPHAEROSTACHYA (Tuckerm.) Kiiken
Fernald, RHopDORA, xxviii. 162 (1926). Rare. Boggy hollow is
woods a ~ end of nag NAU: 2480 (P,W). sail
RADIATA Dewey. Apparently |
Coleted in in dry saute wen at east end of Nevshor NAU: 2475

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 241

C. contigua Hoppe. C. muricata of most American authors. See
Mackenzie, Bull. Torrey Bot. Club, 1. 236 (1923). Dry, open,
ae ground: NAU: 734 (P,W); CUT: 2283 (N,P), Taylor 2413 (P);
PEN: 484 (P,W

C. Munuensercn Schkuhr. Rather frequent on exposed downs
and in sandy oe: NON: Pennell 2838 (W), 2245 (N), 3361
(P,M); NAU: 733 (W).

- CEPHALOPHORA Muhl. Sandy oak and beech woods; local.
NON: 2594 (N,P,W); NAU: 2463 (P).

i: ee a Michx. Wet depressions. NAU: 2394 (N,P,M,
C); NAS: 3485 (N).

C. os epalllgtl (Kiikent.) Mackenzie. See Britton & Brown,
Ill. Fl. ed. 2, ce 371. (1913). Low, wet hollows; local. NAS: 3473
(P); CUT: 2281 (N).

C. cRIN te am. ampy and boggy depressions. NAU: 723
(W); 2398 ®), 2936 wy NAS: 3475 (P).

C. Mircuetiiana Curtis. See Weat therby, Raopora, xxv. 17
(1923). Rare on the Elizabeths. NAS: Pennell 2921, labelled “C.
crinita” (W).

A species of the southern coastal plain ranging north to Cape Cod
and Plymouth County.

C. virescens Muhl. Boggy woods; rare and local. NAU: 1020
(P,W), 3141 (N,P,M).
. SWANII (Fernald) Mackenzie. Bull. Torrey Bot. Club, oe
246 (isi0 0). C. virescens, var. Swanii Fernald. Dry, open ground
and sandy woods. NON: 2588 (N,P), 3354 FO NAU: Pennell
3132 Ww), 3680 (N); CUT: 1012 (N,P,W,M), 3452 (P).

- COMMUNIS Bailey. Dry, sandy bee Gate ke rare. NAU:
3324 (N,P),

- VARIA Muhl. Sandy clearings, exposed knolls and _ hillsides;
widespread. nome ti the material ie the scarcely separable
forma colorata (Bailey) Kiikent. NON: 3284 (N,P,C), 3300 (NP),
3301 (P), 3285 (P): AU. 3322 (N,P); NAS. 2361 (N,P,W,M).

PENNSYLVANICA Lam. Sandy banks and open grassland. NON:
8283 (N,P), 3353 (P,C); NAU: 5321 (NP PM.
C. PENNSYLVANICA Lam., var. SEPA in E. C. Howe, N. Y.
State Mus. 48 Ann. Rep. 174 (1896). a g atte (Willd.) Fernald.
Sen oly C). dry beech woods at east end of Naushon. NAU: 2476

LIMOSA L. A far northern sedge which reaches the southern
limit of its coastwise distribution in a bog at the west end of Cuttyhunk.
CUT: 2311 (N,P,W).

C. pigrratis Willd. Apparently not common; sandy beech woods
Rete eer Harbor. NAU: 2461 (P), 3323 (N,P).
h Lis Michx., var. RupGet Bailey. Collected only in a moist
ollow o = "the hillsides of Cutty hunk. CUT: 1011 (P).

242 Rhodora [NovEMBER

C. LANUGINOSA Michx. Peaty margin of a aenay brackish pond
at the aloe west end of Nashawena. bir 00 (N,P).

C.c A Boott. Swampy clearings a capt she locally
an NON: 2905 (N,P); CUT: os coe ea 15, 1917 (N), 733
Ww

urIDA Wahlenb. Pond borders and moist hollows. UNC:
Hs (N, bas a“ (P.C):N ee 2385 (N,P,M); NAS: 3474 (N,P).

C. tuputina Muhl. Boggy clearings and low moist situations.
NAU: 2486 (N, ry, 3882 (P, M C): PAS: 519 (P,W); CUT: 731 (P,W).
ARACEAE

ARISAEMA TRIPHYLLUM (L.) Schott. This seems to be the only
Arisaema on the islands. ie is abundant in “in bogg By and swampy

woods at Pres east end of Naushon, and in similar situations on Nasha-
wena. : 752 .(N), 3133 (W), 3317 (P,M,C), 3331 (P); NAS
Ni ae rn (o 3528 (N,P).

SYMPLOCARPUS FOETIDUS (L.) Nutt. Seen only in a low swampy
ney at the extreme west end of Nashawena. NAS: Northrop (0), 3496

Acorus CaLamus L. Wet grassy hollows and swampy pond bor-
nS locally eusank NAS: Northrop (0), 3565 (P); CUT: 2282
LEMNACEAE

Lemna Minor L. Occasional in small ponds in clearings. NAU:
2376 (N,P).

ERIOCAULACEAE
ERIOCAULON SEPTANGULARE Ue E. articulatum (Huds.) Morong.
Sandy and peaty pond margins. NAU: Pennell 3118 (P), 2508
(P,M), 2942 (N); NAS: Neithews. oan 1901 (Y).
XYRIDACEAE

CAROLINIANA Walt. a gach am not sommon. Sand

X. TorTa Smith. YX, ftom Muhl. Reported by Mrs. Northrop
from Nashawena; no specimen seen :
PONTEDERIACEAE
PonTEDERIA corpaTa L. Collected only in swampy border of
or ponds on Nashawena. NAS: Noveiep (o), 2368 (N,P), 3547

JUNCACEAE

UNCUS BUFONIUS L. Moist sandy or peaty soil. NAS:

JUN
Aug. 1901 W
ice se SNE illiams 1279 (N), Pennell 2918 (W), es (P,M PM);

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 243

J. Gerarvi Loisel. Brackish areas; general. NON: 2877 (N);
PAS: 1779 (P); NAS: Northrop, Aug. 1901 (Y); CUT: 2276 (N,P,M,
C); PEN: 1094 (P,W), 1753 (N,P).

Willd.

shores, sandy paths and open fields. NON: 3191 (N); UNC: 3645
(P); NAU: 3136 (N); NAS: Northrop, Aug. 1901 (Y), 3468 (P); CUT:
3441 (N,P,M); PEN: 492 (P,W), 3400 (P).

Much of the material which has been collected can not satisfactorily
be referred here, but is being held pending a critical comparison of
our American J. tenuis with Old World collections, a comparison
which may necessitate a careful revision of the entire species. The
preceding few numbers are, however, cited as fairly typical.

J. tenuis Willd., var. Wriutamsu Fernald. This phase, with
strongly divergent floriferous branchlets, has been met with occasion-
ally. NAU: 2400 (N), 2710 (P,M).

J. pichoromus Ell. NAU: 1800 (P,W), 3162 (P); PEN: 1756
(N,P,W).

Under this heading is grouped a large series of variations, of which
the variety platyphyllus, perhaps specifically distinct, represents
only one trend. This matter is under consideration at present, and,
until the species can be completely studied and revised, a few numbers
only, out of an extensive series of collections, are cited.

J. DICHOTOMUS Ell., var. pLarypHyLLus Wiegand. PAS: 1781 (P);
CUT: 1039 (P).

- GREENEI Oakes & Tuckerm. This is one of the commonest

]
NON: 2247 (N »P,M,C), UNC: 3130 (P); NAU: 766 (P,W), 1802 (P);
: (65 (W); NAS: Northrop, Aug. 1901 (Y), 236

(1991) sus L., var. cosru.atus Fernald. Ruopora, xxiii. 239

(Py ). Abundant in wet hollows and around ponds. NON: 2303
M,C), 2567 (N,P); NAU: 2397 (P); PAS: Svenson, Sept. 8, 1926

m3 AS: orthrop, Aug. 1901 (Y), Williams, July 10, 1911 (N),
J (N,P), 2347 (P); CUT: 2320 (N,P,M); PEN: 852 (P,W).

(1910) Frusus L., var. sotutus Fernald & Wi

‘. ) _ Collected by the writer only in a boggy hollow ear

on sh ulin Cove. NAU: 2883 (N,P,M,C); NAS: Northrop, Aug. 1901,

Set with var. costulatus (Y). :

‘CANADENSIS J. Gay. Abundant in moist depressions and around
3778 ( C: 2994 (P); 3114 (N,P,C); NAU: 2918 (N,P,M); PAS:
1917 St NAS: Northrop, Aug. 1901 (Y); CUT: Sanford, Aug. 15,

J. PELocaRPUS M. ,
borders ey. Sandy shores and sandy or peaty pon
- NON: 2638 (BP), 3003 (N); NAU: 1042 (W), 2379 (P),

244 Rhodora | NovEMBER

2724 a ee - (N,P); PAS: 3780 (P); NAS: Northrop, Aug. 1901
me dan (0).

J. MI sale im Bigel. Sandy and swampy pond margins. NAU:
1038 (P), 2917 (N,P); NAS: Northrop, Aug. 1901 (Y), Williams, July
10, 1911 (N), hers 2915 ¢, W), 1037 (P,W).

. ACUM wets He undant in wet to pease — ty: boggy

peaty borders i ee onds in open NON: 2242 (P,C),
3704 (N), 3163 (NPM), UNG: 2995 (N >) NAU: 1041 . W), 2391
(N,P); PAS: 757 (W); NAS: Northrop, Aug. 1901 Y), Pennell 2909
w); CUT: 3442 (P); PEN: 489 (N,P,W,M), 1757 (P,W

. ARTICULATUS L. Occasional in ‘boggy depressions. PEN: 490!
ey 1499 (Ww).
J. ARTI tus L., var. oprusatus Engelm. Peaty, usually
brackish, ey N AU: 3300 WN: CUT: 3574 (N,P).

MARGINATUS Rostk. ddy borders of ponds. UNC:3117 (N);
NAU: 2390 wget me al “P): PAS: 3779 (P); NAS: Northeae
Aug. 1901 (Y), 3

LuzuLa ee (L.) D a war. ieee (Ehrh.) Celak.
Sandy bank at east end of N: Nonamsaet. NON: 3369 (N).

L. campestris (L.) DC., var. ECHINATA (Small) Fernald & Wiegand.
Ruopora, xv. 38 (1913). ‘Collected only in sandy beech woods on
Naushon. NAU: 3325 (N,P).

LILIACEAE

Uvunaria PeRFouIATA L. Reported by Mrs. Northrop for Nasha-

wena; no specimen seen
AKESIA SESSILIFOLIA (L.) Wats. Seen in a boggy, wooded

hollow along the south shore of tiene NON: 3295 (N,P);
NAS: Northrop (0

Litium PHILADELPHICUM L. Reported by Mrs. Northrop from —
Nashawena and said by wife of caretaker there to ow abundantly at
certain Isiewaiacy on the island today; no specimens

L. tigrinum ., Found persisting around Le cottages on
ae ae PEN: 494 (P,W). |

sparagus officin ait ka Th or soil near areas under cultivation.

NAS: Northrop (0); PEN: 493 (N,P, :

MAIANTHEMUM CANADENSE Teck Abe ndant in boggy ooded
jane a south shore of Nonamesset. NON: 3296 (N); ; NAS:

0

EDEOLA VIRGINIANA L. Boggy woodland on a ee and

edge of brackish marsh at east end of Naushon. NON: 3337 (P);
NAU: 3332 (N); NAS: Northrop (o).

SMILAx athe) L. Rather abundant in thickets and bigs
overgrown hollows. NON: 2296 (N,P); UNC: 3044 (N); :
Northrop (0); PEN: 495 (W).

‘This is the materia] erroniously reported from Penikese as J. debilis Gray if
Ropora, xxvi. 223 (1924),

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 245

S. HERBACEA L. Reported by Mrs. Northrop from Nashawena;
no specimen seen.

S. gLauca Walt., var. LEUROPHYLLA Blake. See Ruopora, xx.
78 (1918). Known only on basis of Mrs. Northrop’s collection from
Nashawena. NAS: Northrop, “July-Aug.” 1901 (Y).

AMARYLLIDACEAE

Hypoxis uirsuta (L.) Coville. Abundant on open, sandy slopes
and knolls at west end of Cuttyhunk. NAS: Northrop (0); CUT:
Sanford, Aug. 15, 1917 (N); 2310 (P).

IRIDACEAE

IRIS VERSICOLOR L. Wet depressions and pond borders; abundant.
NON: 2678 (N); UNC: 3031 (P), 3121 (N); NAS: Northrop (0),
2357 (N); PEN: 496 (P,W).

I. prismatica Pursh. Apparently not common; reported and
collected by Mrs. Northrop from Nashawena. NAS: Northrop,
“July-Aug.” 1901 (Y).

SISYRINCHIUM ANGUSTIFOLIUM Mill. Seen growing only in dry
sand along east shore of Penikese, where fairly plentiful. NAS:
Northrop (0); PEN: 497 (P,W).

S. GRaMINoIDEs Bicknell. S. gramineum Curtis, not Lam. Grassy
slopes and pond borders. NON: 2300 (P), 3360 (N,P); CUT: 3449
(N,P,C); PEN: 3408 (N,P,M), 3415 (P).

. ATLANTICUM Bickn. Wet hollows, open slopes and bogs. UNC:
@y. (P); CUT: 771 (W), 2318 (N,P,M,C), Hervey (N); PEN: 3401

ORCHIDACEAE

HABENARIA BRACTEATA (Willd.) R. Br. Reported by Mrs. North-
Top as “common in meadow” on Nashawena; no specimen seen.
NAS: Northrop (o).
This is one of the surprises of the Northrop list. It is hard to be-
lieve that the compiler could have confused this species of the damp
upland woods with H. clavellata, H. blephariglottis or H. lacera, all of
which she likewise reported and the first two of which are represented
by specimens at the New York Botanical Garden. It is greatly to
be regretted that she did not also collect H. bracteata. As it is,
gs er, there seems no course but to let this record stand.

- CLAVELLATA (Michx.) Spreng. Open bogs, boggy pond borders
and wet, peaty depressions. NAS: Northrop, July 1901 (Y), 3540
(N,P); CUT: Sanford, Aug. 15, 1917 (N), 2527 (P.)

H. onBIcuLaTa Torr. NAU: MacRae, July 25, 1904 (N).
Collected by Lillian J. MacRae on Naushon, July 25, 1904.
Specimen originally in the local herbarium at Woods Hole but since

246 Rhodora [NovEMBER

transferred to the collections of the New England Botanical Club.
This, like H. bracteata, constitutes another remarkable addition to
the flora of southeastern Massachusetts, for this plant has not pre-
viously been reported from nearer than Norfolk County.

a algerie (Willd.) Torr. Apparently rare; known only

om Sy throp’s collection on Nashawena. N AS: Northrop,
89 190 rw).

H. tacera (Michx.) R. Br. Swampy and boggy hollows. NAU:
iN (N): PAS: 1787 (P); NAS: Northrop (0); CUT: 2315 (P), aun

PoGoNIA OPHIOGLOSSOIDES (L.) Ker. Locally abundant in open
peaty and boggy depressions. PAS: 1789 (P); NAS: Northrop (0),
Williams, July 10, 1911 (N); CUT: 524 (W).

ALOPOGON PULCHELLUS (Sw.) R.Br. Not faba oer with

the preceding. PAS: 1788 (N,P); NAS: Northrop (0); C 23 (W).

RETHUSA BULBOSA L. Mrs. N orthrop reports this pes Nasha-

wena with the comment “quite common in some places.” Unfortu-

nately, no specimen is extant, but there seems to be no good reason for
questioning this record. N AS: Northrop (o).

SprrantHes Becxn Lindl. Rare; 4 only on dry hillside west
of French Watering Place. NAU: 2 946 (N).

- GRACILIS (Bigel.) Beck. Dry, grassy, open slopes; fairly abund-
ant. NON: 2622 (N); NAU: 2601 (N); NAS: Northrop, Aug. 1901

(L.) Richard. Local; wet, peaty hollows. UNC:
piss WN, P); “NAU: Cushman & Morse, “Aug. 25, 1906 (B).

_ Liparis Lorsexm (L.) Richard. Local; Mrs. Northrop’s comment
is “moist hollow in face southern cliffs.” Taylor’s sheet is from
border of Sheep Pond. NAS: Northrop, Aug. 1901 (Y); CUT: Taylor,
July 15, 1919 (N).

TYPULARIA DISCOLOR (Pursh) Nutt. This rare orchid, hither
known on the islands of Massachusetts! only from an old collection fe
Martha’ s Vineyard, is listed by Mrs . Northrop from Nashawena as

“quite common in 3 or 4 Toealitees, & he part island.” Fortunate-
ly, i in this instance a specimen was preserved and is now in the herba-
y of the New York B Botanical Garden. NAS: Northrop, Aug. 1901

SALICACEAE
[Saurx aMyGpaLomes Andersons. Appears on Mrs. Northrop’s
Nashawena list. This record probably refers to S. alba or some other
— and, in the absence of material, can scarcely be admitted.]
S. pentandra L, Introduced on Naushon and Penikese; apparently
thriving on latter island. NAU: 3873 (P); CUT: 590 (N,P,W,M,C).
-alba L. X? This hybrid, probably S. alba X fragilis, is abund-
1 See Fogg, Ruopora, xxxii. 114 (1930).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 247
ant around a small freshwater pond on Penikese. PEN: 591 (N,P,W,
3 DISCOLOR Muhl. Reported by Jordan from Penikese; no speci-

ce alba L. In boggy woods near farmhouse on Nashawen
and around dwellings on Penikese. NAS: Northrop (0), 3511 );
PEN: 499 (N,P,W).

. TACAMAHACcA Mill. P. balsamifera of most American authors.

See Sargent, Jour. Arn. Arb. i. 61 (1919). Reported by Mrs. North-
rop for Nashawena; no specimen seen

P. BALSAMIFERA L. P. deltoides Marsh. See Sargent, Jour. Arn.
Arb., i. 62 (1919). Small clump near landing on Penikese. PEN:
500 (P,W).

MYRICACEAE
Myrica Gate L. Only locally abundant, as around dune ponds o
Nashawena. NAU: Williams, July 10, 1911 (N); NAS: Northrop oy,
1773 (N,P,W,M).

- CAROLINIENSIS Mill.'. A common shrub on the Elizabeths;
thickets, pond margins, and sheltered hollows in open hills. NON:
2237 (N, P); UNC: 2951 (N); PAS: 1783 (N,P,W); NAS: Northrop (0),
2342 (N); CUT: Sanford, Aug. 15, 1917 (N); PEN: 498 (N,P,W).

JUGLANDACEAE
mig tena L. On Mrs. Northrop’s list for Nashawena; no
specimen
ARYA ALBA a (L.) K och. Occasional in mixed woods. UNC: 2454
(N,P); NAU: 777 (P, W; NAS: Northrop (o).

BETULACEAE
Corytus cornuta Marsh. C. rostrata Ait. Reported by Mrs.
Northrop for Nashawena; no specimen seen
TRYA VIRGINIANA (Mill. ) - Koch. An important constituent of
the native woods , especially on Nonamesset and Naushon. NON:
2593 (N); NAU: 779 (P,W); NAS: North rop (0).
ETULA LuTEA Michx. Mrs. Northrop reports two specimens near
Choptauk Lak ke, Nashawena; no material seen
- POPULIFOLIA Marsh. Bordering woods and in protected sept
where it often forms den kar NON: 2635 (N,P,M); UN
pout WAS: Northros (0); PEN: Jordan (0).
N la Roth. Planted and established near eastern end of
aushon, “NAU: 2471 (N,P).
B. pubescens Ehrh. B. alba, var. pubescens Spach. With the last.
NAU: 2472 (N,P,M).

oon was the original spelling employed by Miller, Gard. Dict. ed. 8 (1768) and not
linensis, ag usually given.

248 Rhodora [NOVEMBER

FAGACEAE

FaGus GRANDIFOLIA Ehrh. The most important tree on the islands.
Forms a pure stand over large areas on Naushon and Nonamesset, and
occurs mixed with other trees on these two islands as well as parts of
Uncatena and Nashawena. When growing in the heavier woods this
tree apparently does not form fruit, but an occasional isolated speci-
men in the open will be found laden with nuts. NON: 2592 (N);
NAU: 3147 (N); NAS: Northrop (0), 3532 (P,M).

F. GRANDIFOLIA Ehrh., forma PUBESCENS Fernald & Rehder. See
Rehder, Ruopora, ix. 111 (1907). Occasional with the species.
UNC: 3125 (N); NAU: 3187 (P), 3693 (N,P).

UERCUS ALBA L, Abundant; dry sandy woods. NON: 2657 (P);
UNC: 3127 (N); NAS: Northrop (0), 3531 (P).

-RUBRAL. A single tree near landing on Penikese, where possibly
introduced. PEN: 592 (N,P,W).

Q. veLuTiIva Lam. Next to the Beech the most important constit-
uent of the wooded areas. NON: 2656 (N), 3309 (P,M), 3916 (P);
UNC: 3126 (P); NAU: 780 (P,W), 3880 (P); NAS: Northrop (0),
3530 (N,P). :

[Q. MariLanpica Moench is reported from Nashawena by Mrs.
Northrop but is not represented by a specimen. The Black Jack Oak

own northeast of Long Island and it seems probable that this
record refers to some other species, possibly a sapling of Q. velutina.]

URTICACEAE
Utmus americana L. Reported by Mrs. Northrop from Nasha-
wena, no sp seen
a urens L. Listed by Mrs. Northrop as growing “about
houses” on Nashawena; no specimen seen.

POLYGONACEAE

RumMex Britannica L. Reported by Mrs. Northrop from Nashaw-
Be no tae —
_ A. crispus L. aste places and pond margins; often’ appearing
indigenous. NAS: Northrop (0); PEN: 598 (N,P,W,M), 3399 (P).
Pin et olius L. eae the Nashawena collection was —

Swampy margin of thick woods. NAS: North ,

(N); PEN: Jordan (0). gas cia uiceaphit aig .

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 249

R. VERTICILLATUS L. Swampy margin of oul at east end of
Nashawena with the last. NAS: Northrop (0), 3529 (N).

This and Block Island are the only stations for the plant in south-
eastern New England.

R. maritimus L., var. ruEcinus (Phil.) Dusén. R. persicaroides,
in part, aes Manual. See St. John, Ruopora, xvii. 73 in “9

R. Acetosella L. Expose d knolls, slopes and sandy clearings. NAS:
ie ah (0); CUT: 2262 (N); de 597 (N, P, W).

OLYGONUM GLAUCUM Nutt. P. maritimum of Manual. See Fer-
oe Ruopora, xv. 69 (1913). ye mas obble beaches. NON:
68 (P); UNC: 3051 (N); NAU: Cushman & Morse, Aug. wen 1906

B) Williams, July 10, 1911 (N); NAS: eae Aug. 1901
P. avicutarE L. Sandy patches; not common. NAS: Northrop
(0); PEN: 594 (P,W).
- PENSYLVANICUM L. Not common; borders of small ponds. UNC:
2087 (N).
PENSYLVANICUM L., var. LAEVIGATUM Fernald. RHopora, xix.
70° (1917). Moist situations; local. PAS: Svenson, Sept. 8, 1926 (N).
YDROPIPER L., var. PROJECTUM Stanford. RHopoRA, xxix. 86
(1927). Seen aye from margin of small pond near east end of Nona-
. es 02 (N).
NCTA ae ll. P. acre HBK. See Stanford, RHopora, xxix
77 ‘Gan. where, in a facta, seer is bmg for attributing priority
- punctatum over P. acr ders, ee abba and
wet ane eee ally. NON: ae N, P ‘M), 26 99 (P,M,C);
2986 (N,P), 3018 (P); NAU: 2371 (P), ap (P,C), 3150 (NPM):
CUT: “ eae 15, 1917 (N); PEN: 5 93 (N,P,W).
P. Persic riaL. Pond m margins and moist depressions. UNC: 3017
iv PAS: Seeson Sept. 8, 1926 (N); NAS: Northrop (0); PEN: 596
P. nypRopiprromes Michx. Apparently not common. PAS:
Stenson, Sept. 8, 1926 (N).
P. opELousanum Riddell, var. ADENOCALYX Stanford. See Ruo-
(NP. XXVili. 28 (1926). Pond borders. NON: 3207 (P); PAS: 3747

- Convolvulus L. cao over thicket and low growth surrounding
pond. Setoak 595 (N,P,W,M,C). ne oak
cae Coverin scrub growth in swampy hollows.
3172 (NE P): UN 038 (NP). ie
© eile L. ” This is species, with smaller fruit than the pre-

ceding, was found only on Pine Island, Nonamesset, where it formed
we tangle over the other vegetation on the beach. NON: 3912

250 Rhodora [NovEMBER

CHENOPODIACEAE
henopodium album L. Beaches and exposed sandy areas. UNC:
a (N); NAS: sheen (0); CUT: Pennell 2993 (P); PEN: 1401
(P,W).

ATRIPLEX ARENARIA Nutt. Sea beaches; not abundant. UNC:
3106 (N); PAS: 1793 (N,P), 3756 (P,M); NAS: Northrop (0); PEN:
Jordan

A. nh hey L., var. HasTaTa (L.) Gray. One of the commonest
plants of the sea beaches and disturbed pa areas near the shore
UNC: 2978 (N,P); PAS: 1791 (N,P); N Northrop (0), ew (P);
CUT: Pennell 2996 (P), 2272 (N); DEN. “aD (P,W), 3386 (M,C).

SALICORNIA MUCRONATA Bigel. Salt marsh areas; less Ais roe
se ye or ds NAU: Pennell 3146 (W); PAS: Svenson, Sept. 8

26 (N

ae EUROPAEA L, Salt marshes and brackish mud flats. NON: 2892
(N); UNC: 3094 (N,P); PAS: Swenson, Sept. 8, 1926 (N); 3742 (P);
NAS: Northrop (0); CUT: 3570 (N,P); PEN: Jordan

. AMBIGUA Mic Same situations as last. NAU: Pennell 3112
®) a 3145 (W); PAS: Svenson, Sept. 8, 1926 (N), 3741 (N,P);
T: 3571 (N,P,M,C).

perce MARITIMA (L.) Dumort. Reported by Jordan from Peni-
kese; no specimen.

S. LINEARIS (Ell.) Mog. Salt marshes and brackish muds. UNC:
2963 (P,W), 3076 (N); PAS: 3739 (N,P); CUT: 3567 (P,M »C).

gt LI eaches vase Persone marshes. NON: 2286
(N,P); NAU: Cushman & Morse, 5, 1906 (B), 3886 (P); ‘NAS:
Northrop (o), — (N); en si ©; PEN: Jordan (o).

S. Kaur L., var. carotrntana (Walt.) Nutt. With the last; less

abundant, UNC: 2961 (P), 3075 (N,P); PAS: 1792 (N,P).

AMARANTHAC
Amaranthus retroflecus L. Dry sandy soil near cultivated area.
PEN: 1402 (N,P,W,M).

PHYTOLACCACEAE
PuyToLacca AMERICANA L. P. decandra L. See Ruopora, xvi

180 (1915). Apparently rare; seen growing only along edge of boggy
aig at east end of Pasque. NAU: Pennell 3174 (W); PAS:

oe
Mollugo verticillata L. » grassy slopes sandy patches.
NON: 2891 (N); UNC: a ‘N, P); NAS: et i (0), 3470 (P).
CARYOPHYLLACEAE

any ne RUBRA nee ‘Deby, C. Presl. Dry exposed slopes an
NAS p (0), Williams, July 10, 1911 &
rrr MD CUT: sare py” "PEN: 1410 (P,W).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 251

S. satina J. & C. Presl. Sand and cobble beach along Northwest
Gutter. Rare; represented on these islands usually by the next species.
NAU: 3132 (N, W).

S. LEIOSPERMA (Kindb.) F. Schmidt. S. marina of Manual, in

cobbly beaches, mud flats and brackish marshes. NON: 2290
UNC: 2962 (N,P,M); NAU: 2455 (N); PAS: 3738 (N,P,C); CUT:
3569 (N,P); PEN: 1409 (P,W).

Spergula arvensis L. Not common. NAS: Northrop (0); PEN:
Jordan (0).

SAGINA PROCUMBENS L. Upper beaches and dry, gravelly banks.
NON: 2692 (N,P); NAU: 2728 (N); NAS: Northrop (0), 3464 (P);
PEN: 1408 (W).

ARENARIA LATERIFLORA L. Seen growing only in sandy thicket
along north shore of Naushon near the east end. NAU: 3318 (N);
NAS: Northrop, “July—Aug.” 1901 (Y).

A. PePLomEs L., var. RoBuUsTA Fernald. See Ruopora, xi. 109
(1909). Occasional on sandy or cobbly beaches. CUT: 2274 (N,P);
PEN: 1403 (W)

A. serpyllifolia L. Sandy and gravelly areas; occasional. NON:
Pennell 2840 (W); NAU: 3424 (N); NAS: Northrop (0).
cae graminea L. Grassland on Penikese. PEN: 1411 (P,W);

.

1760

(0); CUT: 3438 (P); PEN: 1404 (N,P,W).

Agrostemma Githago L. Reported by Mrs. Northrop from Nasha-
wena; no specimen.

Lychnis alba Mill. Occasional in grassland on Penikese. PEN:
1407 (W).

, Sypsophila paniculata L. Persistent around leper cottages on Pen-
ikese. PEN: 1406 (P,W).
Dianthus barbatus L. Same situation as the preceding. PEN: 1405

D; Hineria: 1... Seen only on edge of wet hollow at west end of
Cuttyhunk, CUT: 3445 (N,P).

PORTULACACEAE

af ortulaca oleracea L. Sandy ground, usually near cultivated areas.
AS: Northrop (0); PEN: 1412 (W).

NYMPHAEACEAE
Nymprozantuos VARIEGATUS (Engelm.) Fernald. Nymphaea ad
vena, var. variegata Fernald. See Fernald, Ruopora, xxi. 183 (1919).

252 Rhodora [NOVEMBER

Dune ponds on Nashawena and Sheep Pond, Cuttyhunk; possibly
more general. NAS: Northrop (0), 2366 (N,P); CUT: 786 (W), 2531

NympHaEa oporata Ait. Castalia odorata Woodv. & Wood. See
RwHopoRA, xviii. 161 (1916). Rather general in smaller ponds as well
as Sheep Pond, Cuttyhunk. UNC: 3004 (P); NAU: 2734 (N); NAS:
Northrop (0), 2367 (N,P); CUT: 2530 (N).

RANUNCULACEAE
RanuncuLtus CympaLarta Pursh. Sandy and gravelly shores;
rather frequent. NON: 2302 (N,P); NAU: 528 (W,P); CUT: Hervey,
no date, (N); PEN: Northrop (0), 1414 (W).
. DELPHINIFOLIUS Torr. Small,
hills. NAS: Northrop (0), 3522 (P); PEN: 451 (P,W), 1759 (N,P,M).
[R. reEprans L. appears on Mrs. Northrop’s list. It is hard to say

east end of Nonamesset; possibly more abundant. NON: 3282 (N)
. acris L. Rather common in open grassland, especially on outer
islands. NAS: Northrop (0); CUT: 2325 (N); PEN: 1413 (N,P,W).

THALICTRUM REVOLUTUM DC. Reported by Mrs. Northrop as
T’. purpurascens, but had probably best be referred here. NAS:
Northrop (0).

ANEMONE VIRGINIANA L. Collected in dry, sandy woods on “ Cap-
tain’s Island,” Naushon. NAU: 3907 (P).

A. QUINQUEFOLIA L. Reported by Mrs. Northrop for Nashawena;
no specimen seen.

CopTis GROENLANDICA (Oeder) Fernald. C. trifolia of the Manual.
See Fernald, Ruopora, xxxi. 136 (1929). Included in Mrs. Northrop’s
list with the comment “in one filled up pond under fern lvs. Very
small, no fruit.” NAS: Northrop (o).

This is an interesting addition to the flora of the islands. Coptis
is unknown from Barnstable County, although represented in both
Plymouth and Bristol Counties. Moreover, it is reported by Bick-
nell for Nantucket, where, as on N ashawena, it may be a remnant of a
formerly more dominant continental flora. There seems to be no
real reason, even in the absence of herbarium material, to discredit
this record.

BERBERIDACEAE

Berberis Thunbergii DC. Occurs scattered over the open hills at

the east end of Uncatena. UNC: 2975 (N)

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 253

LAURACEAE

SASSAFRAS OFFICINALE Nees & Eberm. S. variifolium (Salisb.)
Ktze. See Blake, Ruopora, xx. 98 (1918). Scattered ied
throughout all of the wooded a reas. NAS: Northrop (0), 3533 (P).

S. OFFICINALE Nees & Eberm. var. aLBipum (Nutt.) Blake. Ruo-
DORA, xx. 98 (1918). Occasional with the species. UNC: 3123 (N).

ENZOIN AESTIVALE .) Nees. Reported by Mrs. Northrop from
Nashawena; no specimen

PAPAVERACEAE

Glaucium flavum Crantz. Established on beach near Tarpaulin
yee. NAU: 789 (P,W), 1798 (N).

CRUCIFERAE

sitet meana (L.) DC. Dry sandy soil near landing at east end
of Pasque. Hn didboh Ee M).

Bisco vIRGINICUM L. Open grassland and low sandy sie
UNC: 3056 (N,P); "PAS: 3762 (P); NAS: Northrop (0); CUT: 2259

»P); PEN: 1417 (N »P,W,M,C).

apsella a Bursa~pastoris a Medic. Occasional in disturbed situ-
ations. NAS: Northrop (0); PEN: 854

AKILE EDENTULA (Bigel.) Hook. Sand y i and cobbly sea beaches,
especially aon islands. NAS: Northrop (0); CUT: 2271 (N); 3431
(P); PEN: 1416 (P,W).

aioe, SEI um L. Disturbed sandy soil near cultivated
areas. NAS: Northrop (o); CUT: 2268 (N,P,M 0); PEN: 1418 (W).

t sativus L. Same situations as last. PEN: 1419 (N, PW).

ren se (L.) Cosson. On peninsula, east end of Penikese.

. ‘ : (N,P. 3 ria
wy altissimum L. Same situations as the preceding. PEN: 1420

Rorippa N asturtium-aquaticum (L.) Schinz & Thel. Radicula Nas-
turtium-aquaticum Britt. & Rendle. Seen growing along a moist rill
draining into Sheep Pond, Cuttyhunk. NAS: Northrop (0); CUT:
10 054 (P ,W).

sprang
ge ERA ROTUNDIFOLIA L. Boggy situations and borders of ponds.
eb a - (N), 3362 ri PAS: 1786 (N,P); NAS: Northrop (0),

Dz. “acd a the D. longifolia of Manual. Apparently not
N. rders of dune ponds on Nashawena. NAS: 1778 (N,P),

orthrop o. Williams, July. 10, 1911 (N).

254 Rhodora [NovEMBER

SAXIFRAGACEAE

RIBEs HIRTELLUM Michx., var. catcicoLa Fernald. See Fernald,
RHopora, xiii. 76 (1911). Seemingly rare. NON: 2844 (P).

HAMAMELIDACEAE
HAMAMELIS VIRGINIANA L. Dry to moist woodland and thickets.
Some of this material approaches the variety parvifolia Nutt. NON:
2591 (N,P); NAU: 3152 (N,P); NAS: Noviheep (o), 3512 (P).

ROSACEAE

SPIRAEA TOMENTOSA L. Pond borders and moist hollows; abundant.
NON: 2564 (N); PAS: 1784 oe P); NAS: Northrop (0), 2345 (N,P).

YRUS ARBUTIFOLIA (L.) L Thickets and wet depressions; of
frequent occurrence. NON: a (N,P); UNC: 3046 (N,P); NAS:
1774 (N,P), 3495 (P).

P. arsutiro.ta (L.) L. f., var. ATROPURPUREA (Britton) Robinson.
Same situations as the last but less common. NON: 3175 (N,P);
UNC: a (N,P); NAS: Northrop, Aug. 1901 (Y).

MELANCHIER OBLONGIFOLIA (T. & G.) Roem. Dense thickets and
dubltsen hollows; abundant. NON: 2630 (P), 2883 (N,P), 3205 (P),
3202 (P,C), 3308 (P.M, C); PEN: 1421 (P,W).

Cratarcus prurnosa (Wendl.) C. Koch. Single tree on dry hill-
side near Tarpaulin Cove. NAU: 1801 (N, 1 gt

Cratarcus spp. This genus is to be met with scattered through-
out the dry, sandy beech woods, mostly on Naushon. Several
collections have been made, but it has not been found possible to
match this material with existing specimens or to reconcile it with
available descriptions. Each individual plant appears to possess its
own characteristics which differ from those of its neighbor, and,
rather than give to these specimens names which assume a definite-
ness we sige ne not own, the seh sia is here left in abeyance.

FRa A Duch Dry ye slopes and hollows.
NON: 3987 (N):; ‘NAS: NV io og (0): PEN: 1422 (W).

. VIRGINIANA Duchesne, var. po ONE (Rydb.) Fernald &
Wiegand. RwopoRA, xiii. 106 ae Seen only in a sandy meadow
Pond. CUT: & N,P).

OTENTILLA NORVEGICA L., var. HIRSUTA

iensis (Michx.) Lehm. P. mon-
speliensis L. See Bibliot. Bot. 16, Heft Ixxi. 404 (1908). were open
gas i (N,P,W

entea Dr d
PEN: 1423 (N,P,W WM). y and gravelly soil. NAS: Northrop (0);

P. recta L. Sandy cl AU:
3923 (N,P). y clearing in woods near West Gutter. N

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 255

P. pacirica Howell. P. Anserina of Manual, in part. See Fernald,
Ruopora, xi. 1 (1909). ar ae beach along south shore of Nonames-
set. NON: 2874 (N,P,M); NAS: Northrop, as “ P. Anserina” (0).

P. pumita Poir. Dry tolls ste exposed sandy and ee hill-
sides. NON: 2677 (N,P), 3288 (P,M sc); PEN: 1425 (N,P,W

nsis L., var. simPLEx (Michx.) T. & G The Elizabeth
Islands material wee all has narrow, mre rt: leaves and short,
more or less appres ee meres Dry slopes and sandy clearings.
NON: 3355 (N,P ™):C 536 (N).

GEUM CANADENSE J tes he ently rare; boggy woods at east end
of Nashawena. NAS: Northrop (0), 3544

Rusus occientaus L. Edge of woods at west end of Nonamesset.
NON: 2676 (N).

R. PERGRATUsS Blanchard. ee and margins of boggy hollows.
NAS: 3487 (P); PEN: 1428 (N,P,W).

laciniatus Bing Well sctablighed on peninsula of Penikese.
PEN: 1429 (N,

. AN ieee tadartc Blanchard. This is the commonest Blackberry
on the Islands, tilde in wet overgrown hollows and in thickets on
open slopes. NON: 2577 (P). 2626 (N,P); UNC: 3034 (P); NAU:
Pennell 3135 (W), 2496 (P), 2497 (P).

R. uispipus L. Abundant; apparently as much at home on the
dry exposed hillsides as in the boggy depressions. NAS: Northrop (0),
3539 (N): CUT: 2312 (P), 2321 (N,P,M).

R. FLaGELLARIS Willd. R. procumbens of many authors. R. vil-
losus _ the manual. See Bailey, Gentes Herb. i. 234 fern PEN.
over dry open sandy and gravelly ground. NAS: Northrop (o
1430 (N, PW), bg sr y er 14

Rosa rugosa Thunb. Established and thriving, usually along beach-
€s, on —e of the islands. NAU: Taylor 3077 (P); CUT: 3439 (N);
PEN: 1 427 (P,W).

R. Saisie i Reported by Mrs. Northrop from Nashawena;
NO specimen seen

R. patustris Marsh. R. carolina “ gore and mos t American
authors. See Fernald, RHODORA, xx. (1918). Wet margins of
Ponds. NAS: Northrop, July, 1901 (Y), 2364 (N,W); PEN: 1426

R. carouna L. R. humilis Marsh See Fernald, 1.c. Wet hollows
cored Ae NON: 2631 (N,P); NAS: Northrop (0), 3477 (P);

R. NANELLA Rydb. N. Am. FI. xxii (pt. 6), 497 (1918). Material
collected from cine thicket on Uncatena has dwarf stature and very
itTow leaflets and corresponds in all essentials to Rydberg’s descrip-
tion of this species. Wer 3112 (N).
aio VIRGINIANA Mill. ist hollows and swampy borders of ponds.

ON: Pennell 2843 (P), an ‘N, P), 3346 (P,M); PAS: 3746 (N,P).

Pr ghee Apeb Ehrh. Sheltered hollows, thickets and sandy
cared 71 (N,P); UNC: 3108 (P); NAU: 2501 (N,P); NAS:

orthrop (0), 3519 (P); PEN: 1431 (W).

256 Rhodora [NOVEMBER

P. maritima Wang. Apparently not very common; forms a den
thicket at east end of Nonamesset. NON: 3206 (N), 3303 (P, M):
NAS: Northrop (0).

LEGUMINOSAE
Gleditsia triacanthos L. Sandy woods near 05 (8). at east end
U: 3925 (P

Genista tinctoria L. Well established and spreading in fields at east
end of Naushon and on Uncatena. UNC: 2976 (N); NAU: 796 (N,
,W

ye
Cytisus scoparius (L.) Link. gg and spreading on parts of
rare (see p. 155); local elsewhere. NAU: Miss Weir, June 22,
90 (N), Moore, 1903 (N); PAS: Hollick, Aug. 1898 (Y); NAS: North-
).

ays (0

Ulex Pett ede L. A single large clump along road near Mary’s
Lake. NAU: 797 Sey

rifolium arvens pen ground and sandy areas. NON:
2669 (P); UNC: 3012 a, NAS. Northrop (0): CUT: 2257 (N); PEN:
1434 (P,W).

T. pratense L. Grassland; occasional. NAS: Northrop (0); PEN:
1485 (P,W).

. repens L. Open fields and grassy levels. NAS: Northrop (0);
cur: 2261 (N,P); PEN: 3402 (N,P).

I’. hybridum L. Open grassland; occasional. NAS: N. orthrop (0);
PEN: 1436 (P,W).

Y’. agraritum L. Disturbed sandy soil near areas under cultivation.
PEN: 1433 (N,P,W).

Coronilla varia L. Abundant on grassy hillside toward west end
of Pasque. PAS: 795 (P,W).

Robinia Pseudo-Acacia L. At several localities along the north
shore of Naushon, where saat ign planted but now spreading. NAU:
3874 (N,P); NAS: Northrop (o).

DESMODIUM RIGIDUM (E iL) DC. Sandy woods and hills; not com-
mon aes 2977 OEE NAU: 2600

M (Muhl.) DC. Sandy woods and clearings. NAU:
es (P. M, 0), 3928 (N,P).

LESPEDEZA PROCUMBENS Michx. Not abundant; seen only in sandy
woods on “ Captain’s Island.” NAU: 3909 (N,P .M).

L. viotacea (L.) — Reported from Nashawena by Mrs. North-
Top; No specimens s

L. uiera (L.) Saal On Mrs. Northrop’s Nashawena list; n0
specimen seen

L. caprrata Michx. sf hae a abundant; open, sandy bank
on Uncatena, west end. UNC

Vicia tetrasperma (L.) each: Diewes sandy area near landing.
PEN: 1 “hotel a

V.C ,0 und. CUT: Taylor 3076 (P); PEN:
“a ae PW. e rag pen gro UT: Tay (P)

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 257

V. villosa Roth. Dry sandy strip near Cuttyhunk Pond. CUT:
3579 (N,P).

LaTHyrus Maritimus (L.) Bigel. Sea beaches; commoner on outer
islands. NAS: Northrop (0); CUT: 2263 (N,P); PEN: 1432 (N,P,W).

L. patustris L. Known only from single collection on Nashawena.
This on approaches var. pilosus. NAS: Northrop, “July—Aug.”
1901 (Y).

APIOS TUBEROSA Moench. Occasional in wet hollows and around
ponds. UNC: 3089 (N); NAS: Northrop (0); CUT: 2328 (N,P).

AMPHICARPA Monorca (L.) Ell. Sandy woods near Hadley Harbor.
NAU: 3917 (N,P); NAS: Northrop (o).

LINACEAE
Linum striatum Walt. Dry hillsides and peaty pond margins.
CUT: 2330 (P), 2538 (N,P).
L. virentanum L. Dry sandy woods. NAU: Williams & Collins,
July 10, 1911 (N), 2381 (N,P).

OXALIDACEAE
OXALIs srricra L. Open sandy hillsides and sandy clearings.
UNC: 3096 (P); NAU: Pennell 3142 (W), 2498 (P), 3682 (P); NAS:
Northrop (0); PEN: 1439 (N »P,W).

GERANIACEAE

GERANIUM MacuLaTuM L. Seen only on an exposed gravel bank
at upper margin of beach along north shore of Nonamesset. NON:
3364 (N,P).

G. caRoLintanum L. Burnt-over ground near hospital building on
Penikese. PEN: 1440 (W).

SIMARUBACEAE

Ailanthus glandulosa Desf. Naturalized and spreading at several
localities on Naushon and Nashawena. NAS: Northrop (0), 3481 (P).

POLYGALACEAE
Potyeata POLYGAMA Walt. Open sandy and grassy hillsides.
NAU: 2477 (N); NAS: NV orthrop, “ July-Aug.” 1901 (Y).
P. sancuinra L. Low boggy ground. Not common. NAS:
Northrop (0), 3563 (N).
-, CRuctaTa L. Peaty and boggy pond borders; apparently not
widespread. CUT: Sanford, Aug. 15, 1917 (N), 2514 (N,P). |

EUPHORBIACEAE
ACALYPHA virGINIcA L. Known only from a sheet of Dr. Pennell’s
collecting on Naushon. NAU: Pennell 3140 (W,P).

258 Rhodora [NovEMBER

A. DIGYNEIA Raf. See Weatherby, Ruopora, xxix. 193 (1927).
Rare. Boggy depression in open hills near east end of Uncatena.
ai C: popes a‘ 2

ns Gray. Dry hillsides and moist depressions; local.
UNC: 2974 ‘P), 3087 (N).
UPHORBIA POLYGONIFOLIA L. Sea beaches and low sandy stretches.
NON: 2870 (N); UNC: 3008 (N); PAS: 3755 (P,M); NAS: Northrop,
Baa ssa ” 1901 %); PEN: Jordan (o).

E. macutata L. Sandy beaches and dry exposed slopes. NON:

2869 (N): ‘NAU: un (P), PEN: Jordan (0).

CALLITRICHACEAE

CALLITRICHE PaLUsTRIS L. Apparently rare and local. NAS:
Rites. (0), 3521 (P).
C.

ETEROPHYLLA Pursh. Small ponds in hills; more frequent than
last. NON: 2235 (N); PEN: 456 (W), 1095 (P, W), 1758 (N,P).

(To be continued)

THE FLORA OF THE ELIZABETH ISLANDS,
MASSACHUSETTS
Joun M. Foaa, Jr.
(Continued from page 258)
ANACARDIACEAE
Occasional in sheltered hollows on hillsides.

Ruvs
NAS: Northrop (0), he (N); PEN: 1441 (P,W), 3409 (N,P).
» GLABRA (L.) Gray. No material seen; reported by Mrs. Northrop
eilaly a

me me GLABRA (L.) Gray, var. BOREALIS aa. See Britton’s Man.
No. St. ih 601 (1901). Forming a small grove along the North-
west one te Ws me Hog: of inca kana: “UN C: 2453 (P,M).
n slopes and moist depressions. UNC: 3039
Os : PAS: 4 3793 WN, P), NAS: Northrop, Aug. 13, 1901
PAS. Vernix L. Swam y woods and thickets. NAU: "2485 (N,P);
3794 (N); NAS: Nevthaut (0).
Ex OXICODENDRON L. Thickets, copses and overgrown hollows.
Extremely variable as to habit, leaf-shape and fruit. NON: 3345
380’ M); UNC: 3040 (N,P); PAS: 1785 (N,P); NAS: Northrop (0),
(N); CUT: 2314 (N); PEN: Jordan (0).

AQUIFOLIACEAE
LEX OPacA Ait. Occasional i & ecco wooded areas. NAU: 2488
OM °C); NAS: Northrop (0), 3535 (N).
(an Ait., forma suBINTEGRA Weatherby. Ruopora, xxiii. 118
). Sandy woods with the last; rare. NAU: 2484 (N,P,W,M,C).

264 Rhodora. [DECEMBER

I. verticrLuata (L.) Gray. Moist overgrown ey oy and wet
borders of ponds. NON: ye See P,M); NAS: 3490 (P).

I. verticituaTa (L.) G var. FASTIGIATA (Bickn.) Fernald.
RwoporRA, xxiii. 274 (921). i similar situations as last; occasional.
NON: 2884 (N, P, M); NAS; Northrop, “July—Aug.” 1901 (Y), 3494
N

I. ragvicata (Pursh) Gray. Reported by Mrs. Northrop for Nash-
ste Tees specimen seen
A (L.) Gray. The Inkberry is evidently very rare on the
Elizabeth Tats, Although sought for repeatedly, it is known only
from a single specimen collected on Nashawena by Mrs. Northrop.
NAS: Northrop, Aug. 1901 (Y)

ACERACEAE
ACER RUBRUM L. Sandy and boggy woods and thickets; var. tridens
was nowhere seen. NON: 2655 (N,P); NAS: Northrop (0), 3534 (N).
A. platanoides L. Planted in small grove near hospital building on
Penikese. PEN: 1443 (W).
pe te eudo-Platanus L. A few trees with the preceding. PEN: 3410
M

SAPINDACEAE
Aesculus Hippocastanum L. Included in Mrs. N orthrop’s list with
the comment, “Single good sized tree in midst of native trees about
Choptauk Lake. Introduced.” No specimen seen.

BALSAMINACEAE

IMPATIENS BIFLORA Walt. Wet borders of ponds. NON: 2887 (N);
NAU: 2735 (N,P); NAS: Northrop (o).

VITACEAE os
a QUINQUEFOLIA (L.) Planch. Psedera quingquefolia
Greene. Occasional in thickets. NON: 2294 (N,P); NAS: Northrop —
(0); PEN: 856 (W).
tricuspidata (Sieb. & Zucc.) Planch. ues over stone wall
fk hospital building on Penikese. PEN: 341 :
Vitis Lasrusca L. Thickets oo se halline’ UNC: 3028 (N);
NAS: Northrop, July, 1901 (Y), 3493 (P).
V. AESTIVALIS Michx. Trailing over wie trees, edge of sandy woods. —
NAU: 2490 (P).
MALVACEAE o
Malva rotundifolia L. Dry sand soil, waste places. NAU: Pennell —
3115 (W); NAS: Northrop (0): "REN: 14 42 (PW). :
Althaea officinalis L. : Neos in listing this species for
diaries says, “2 or 3 plants on different parts of shore.” No :
$ :

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 265

iscus Moscueutos L. Borders of salt marshes. NON: 2673
), ‘UNC. 2985 (N,P); NAU: 2378 (N).

HYPERICACEAE

H. BOREALE (@riton Bickn. Pond bor ders and wet hollows;
abundant. NON: 2642 (P,W), “pee ay P); UNC: 2983 (N); NAU:
2726 (N,P); PAS: oe (P); NAS: 3523 (P).

H. mutitum L. Same habitats as a Thet: less frequent. NON: 2569
); n Pennell 3125 (W), 2389 (P); NAS: Northrop (0); PEN:
ordan (0).

H. canapense L. Peaty pond margins and boggy hollows. NON:
2645 (N); UNC: 3023 (N,P), 3085 (N,P); NAS: Northrop (o); CUT:
Sanford, Aug. 15, 191 7 (N).

H. Gentranores (L.) BSP. Dry sandy hillsides and clearings.
NON: 2694 (P); UNG: 3086 ON) NAS: Northrop (0).

’ ie yn UM a ei eae dy and peaty pond borders, often partly
immersed. NON: 2640 (P), 3171 (N,P,C); UNC: 3022 (N); NAU:
eel ts wre Fete (o), 2351 (P,M); CUT: Sanford, Aug. 15,

ELATINACEAE

Evatine minima (Nutt.) F. & M. See Fernald, Ruovora, xix. 10
(1917). Sandy bottoms and borders of ponds. NON: 2706 (N,P);
NAU: Pennell 3202 (W), 2727 (P), 2943 (N); PAS: 1061 (P), 1064 (P);
NAS: Northrop (0); CUT: 2533 (N).

CISTACEAE

HELIANTHEMUM CANADENSE ae Mies: Open grassland and dry
sandy reiaeych abu ndant NON 625 ne P), 2578 ae ee 2903

rhe ea Fernald. H. we of Manua r Sg es DORA, XX1.

2494 (P,W), 2944 Sua ON ); PAS. Svenson, Sept. 8, 1926 (N); NAS:
Northrop, Aug. 1901 (Y), 3557 (P,M).

I im ONIA TOMENTOSA Nutt. ein ary areas. NAU: W: iliams,
‘illy 10 , 1911 (N); PAS: 800 (N,P,M); NAS: Northrop, “July-Aug.”
901 1), ilar, July 10, 1911 (N), Pennell 2916 (W), 2365 (P,M).

N A EIl. Apparently not common; sandy woods on
Naushon,, ‘NAU: "2462 2 (P).

266 Rhodora [DECEMBER

L. MarITIMA Leggett. Abundant on open downs and sandy stretch-
es. NON: 2584 (N), 2624 (P); NAU: 2473 (P), 2481 (P), 2597 (N);
PAS: 3770 (P).

L. Leccerri Britton & Hollick. Reported by Mrs. Northrop
from Nashawena; no specimen seen.

VIOLACEAE

VIOLA PAPILIONACEA Pursh. Not common; seen only in sandy
beech woods on Naushon. NAU: 3314 (P), 3326 (N,P).

V. FIMBRIATULA Sm. Abundant on open downs and dry, grassy
slopes. NON: 2581 (N), 3289 (P); NAU: Pennell 3129 (W), 3126 (P);
CUT: Sanford, Aug. 15, 1917 (N), 2457 (P,M); PEN: 1444 (N,P,W).

V. sacittata Ait. Reported by Jordan for Penikese and by Mrs.
Northrop for Nashawena but probably refers to the preceding. ]

- LANCEOLATA L. Open bogs and peaty borders of ponds. NON:
2568 (P,M), 2894 (N), 3307 (P); UNC: 3088 (N); NAU: Pennell 3122
(W), 2511 (N); NAS: Northrop (0); CUT: Sanford, Aug. 15, 1917 (N).

V. PALLENS (Banks) Brainerd. Peaty and boggy hollow at west end
of Nashawena. NAS: 3499 (N,P).

[V. BLanpDa Willd. Appears on Mrs. Northrop’s Nashawena list.
In absence of specimens it seems most likely that this refers to the
preceding. |

LYTHRACEAE

Decopon verticitiatus (L.) Ell. Abundant in swamps, borders
of ponds an wet thickets. NON: 2563 (N,P), 3170 (N,P);
NAU: 2387 (N); PAS: 3795 (P); NAS: Northrop (0), 2338 (N,P,M,C);
CUT: Sanford, Aug. 15, 1917 (N).

MELASTOMACEAE

Ruexra virerntca L. Wet, peaty depressions and borders of small
ponds. NON: 2639 (N,P,M); NAS: Northrop, “July—-Aug.” 1901
(Y), Northrop, Sept. 1901 (Y), 3548 (P); CUT: 2522 (P).

ONAGRACEAE
Lupvicia patusrris (L.) Ell. Floating or growing on moist mud;
2468 (NP); PAS. She Pr (G) 2571 (PM), 3164 (NP); NAUS
1); PAS: 3784 (P); NAS: North 26 (N,P); CUT:
at (P); PEN: 457 (P,W). orthrop (0), 3526 (N,P)
P

Place; perhaps more general. NAU: 2718

E. cotoratum Muhl. Seen only on the shingle at French Watering — :
(N) ae

|

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 267

E. GLANDULOSUM Lehm., var. ADENOCAULON (Haussk.) Fernald.
E. adenocaulon Haussk. See Ruopora, xx. 34 (1918). Reported from
Nashawena by Mrs. Northrop; no specimen seen.

ENOTHERA MURICATA L. Dry sandy soil, east end of Pasque;
probably ey common. PAS: 3761 (P).

O. BrenNIs L. Dry sterile ground, occasionally in grassland. NAS:
Northrop (0); PEN: 1446 (N,P,W).

_ 0. grandiflora Ait. On Penikese, where probably introduced. PEN:
1447 (W).
HALORAGIDACEAE

MYRIOPHYLLUM SCABRATUM boar Rather frequent; small ponds
and wet pond holes. NON: 1 (G, y ren ee UNC: 2996
(N,P,M,C), 3647 (P); NAU: sor (N); N axon, Sept. 1875
(G), Northrop (0); PEN: 453 (N,P,W), ee Oy.

M. numILe (Raf.) Morong. Abundant and variable; forma capil-
lacewm and forma natans have both been met with but hatte to be

2705

nothing more ene mere ecological variations. N (N,P,
M,C); NAU: 2503 N, NED PAS: 845 (N,P); NAS: Northrop (0),
3520 (N, P); CUT: BG ).

. TENELLUM Bige ag ddy borders and bottoms of several small

ponds. NAU: 2401 (N), Dis (N,P); PAS: 1068 (W); NAS: Northrop
(0); CUT: Hervey, no date (N).

PROSERPINACA PALUSTRIS L. Ponds and pond borders, often sub-
merged. NON: 2570 (P), 2643 (N,P), 3382 (P,M); UNC: 3026 (N);
PAS: 554 (P,W); NAS: Northrop, July, 1901 (Y); CUT: 803 (W),
Sanford, Aug. 15, 1917 (N).

UMBELLIFERAE

SANICULA CANADENSIS L. Dry sandy woods; not common. NAU:
2459 (P,M), 3927 (N,P,C).

op ETDROCOT TEE pennies L. Shallow water of ponds and muddy

argins. UNC: 3001 (N); NAU: Williams et al., July
10,191 1911 (N), 1805 &) 2502 (N,P), 2729 (M), 2947 (N,P,C); NAS:
hrop, “July— 1 (Y), Pennell 2923 (W); CUT: Sanford,

Aug. 15, 1917 Oy

H. Canny Coult. & Rose. NAU: 1069 (G), 2948 (N,P).
In the Gray Herbarium there is a specimen of H. Canbyi labelled
A. umbellata, var. ambigua which was collected at “Woods Holl,
Sept. 10, 1876, on shores of Nobska Pond.” This sheet is further

labelled “ Herb. W. G. Farlow,” and, though no collector’s name
‘Appears, may have been collected by Farlow himself. It constituted

€ only known record for this plant north of southern New Jersey
until 1925 when, on July 14, Dr. H. K. Svenson and the writer found
it growing around the shores of French Watering Place on Naushon.

268 Rhodora [DECEMBER

On Aug. 31, 1927, the writer again visited this locality and found the
plant, this time along another portion of the shore of the same pond.
H. Canbyi is thus definitely established as a member of the flora of
New England.

H. verTIcILLaATA Thunb. NON: 2895 (G,N,P); UNC: 2981 (G,
N,P,W), 3000 (P).

The reported occurrence of this species in Massachusetts has
apparently been based upon a sheet collected by Thomas Morong at
Woods Hole on Aug. 10, 1877. This sheet, which is in the Gray
Herbarium, bears the determination “ H. interrupta Muhl.” and is
perfectly typical H. verticillata as understood today. Also, there is a
sheet of this species in the herbarium of the Boston Society of Natural
History collected at New Bedford and further labelled merely “Oakes
Coll.” Southwest of Massachusetts the plant has been known only
from Block Island before “jumping” to southern New Jersey (Cape
May County).

On August 30, 1927, while exploring the west end of Nonamesset,
the writer came upon a small pond in the open hills which was filled
with a species of Hydrocotyle. This, upon examination, proved to be
H. verticillata. Three days later, September 2, this same species Was
found growing in profusion in two small grassy ponds along the north
shore of Uncatena toward the east end. Thus, a second member of
this genus is restored to our list for this region after a lapse of half a
century.

PTILIMNIUM CaPILLACEUM (Michx.) Raf. Borders of brackish ponds
and salt marshes. NON: 2595 (N), 2664 (P); UNC: 3019 (N,P);
NAU: 2370 (P,M,C), 2715 (N,P); NAS: NV orthrop (0), 3501 (P).

icuTA MACULATA L. On Mrs. Northrop’s list for Nashawena;

Stu suave Walt. S. cicutaefolium Schrank. See Blake, RHODORAs
xvil. 131 (1915). Swampy hollows in woods and borders of ponds.
UNC: 2988 (P); NAU: 2386 (N,P); NAS: Northrop (0), 2350 (N,P);
CUT: 3443 (P). N
IGUSTICUM scoTHICcUM L. Sandy and gravelly beaches. NON:
3210 (N); UNC: 2979 (N,P); PAS: 357 (PW) ; PEN. 1449 (P,W).
+ COELOPLEURUM LUCIDUM (L.) Fernald. C. actaeifolium (Michx.)
; See Fernald, Ruopora, xxi, 144 (1919). Sand and cobble
beaches; not very common. CUT: 3435 (N,P); PEN: Jordan (0).
Pastinaca sativa L. Reported by Mrs. Northrop from Nashawena;
ho specimen seen. bank
ERACLEUM LANATUM Michx. Seen growing only on gravelly
along East Gutter. NON: 3310 (N); NAS: Rortiron am

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 269

ucus Carota L. Dry grassland; eae S not common. NAS:
Thon (0); CUT: 2326 (N,P); PEN: 1448 (P,W).

CORNACEAE
Cornus Fioripa L. Fairly abundant in boggy woods in hollows
along south shore of Nona main OOLe s elsew. ere. Seen in full

hollows and in the moist portions < ro beech-oak forests. NON:
2636 (P), 2667 (W, My 2886 (N,P); NAU: Simons, July 23, 1901 (W),
2932 (N); NAS: Northrop (o), Wakiome July 10, 1911 *(N), 3517 (P).

ERICACEAE
CLETHRA ALNIFoLIA L. Thickets and pond .borders; abundant.
NON: 2293 (P,C), 2627 (N,P), oo ~ M); UNC: 3649 (NP);
NAS: Northrop (0), 1771 (P); CUT: 2 Rh
CHIMAPHILA MACULATA (L.) Pursh. ug monica by Mrs. Northrop
from Nashawena; no specimen seen.

YROLA ROTUNDIFO ouia L., var. AMERICANA (Sweet) Fernald. P.
americana Sweet. See Ru HODORA, xxii. 122 (1920). Mrs. Northrop,
who reports this from N ashawena, says, “ Few plants in one station.”

© specimen seen.

Monorropa untFtora L. Dry sandy beech and oak woods.
Williams, el 10, 1911 (N), 806 (P,W); NAS: Northrop (0), 3537 re),
- HyPpoPitHys L. Seen only in ay woods at east end of Nash-
awena:; eaithacn more ponpen NAS: (N).

- HyPopituys L., var. RUBRA (Torr, Farwell. See Amer. Midl.
Nat., x. 39 (1926). The ¢ crimson variety was collected in dense beech
ar yo Witches’ Glen on Naushon; seen nowhere else. NAU:

“S VISCOSUM L) ai var. GLAUCUM (Michx.) fdas. Known
only from a single pea collected by Williams on Maire:
ea more abunda NAS: Williams, (ast 10, 1911 (N).

EUCOTHOE RACEMOSA (L. Occasional around sandy mar-
gins of ponds. NAU: 2482 (P, My 2938 (N,
Lyonta LiGusTRINA (L.) DC. Dense thickets and wet borders of
oe a 2628 (N,P); UNC: 3650 (N,P,M); NAS: Northrop
, PGP).
a DAPHNE CALycuLATA (L.) Moench. Seen growing only

d borders of oe nds on Nakawesd NAS: Northrop, Aug.

1901 (Y), 1770 0, .
PIGAEA REPENS vies Reporte d by Mrs. Northrop from Nashawena
With the eke: “On bank near Mirror Lake.” No specimens seen.

270 Rhodora [DECEMBER

GAULTHERIA PROCUMBENS . Apparently not common. Reported
from Nashawena and found there growing sparingly in boggy woods
at east end. NAS: Northrop ma ene (N).

GayLussacia Dumosa (Andr.) T. & G., va Scrapers Fernald.
See Ruopora, xiii. 95 (1911). Also Ruo Hooal. xxiii. 248 (1921).
Decidedly not common; so far known only fro P Nahiveue

Northrop, July, 1901 (Y), Norton. § Bily—Auie. * 1901 CY), Taylor,
July 6, 1920 (P).

. FRonpDosA (L.) T. & G. Occasional i . hoagie and about ponds.
NON: 2236 (N,P); NAS: Northrop, Aug. 1 01 (Y).

G. BACCATA (Wang .) C. Koch. This tthe commonest Huckleberry
on the islands; it is everywhere in the open areas, often forming an
impenetrable growth i in uN sna ws or under the lee of protecting
hills. NON: 2576 (N,P), 2 Ope sa 2504 (P,C), 2941 (N,P,
M), 3689 (P); NAS: Natbiag. (0), 3559 (P).

VACCINIUM CORYMBOSUM L, Ate ndant around ponds,
nee in boggy woods. NON: 2637 (P); N AS: Nore
0

V. aTRococcum (Gray) Heller. Same situations as last but appar-
aes less common; flowers and fruit earlier. NAS: 17 ;

MACROCARPON Ait. Open bogs and peaty hollows. NAS: North-
rop vey CUT: 2313 (P), 2519 (N,P,M).

PLUMBAGINACEAE

Limonrum Nasutt Small, var. rrIcHoGonuM Blake. L. carolinianum
of the Manual. See Ruopora, xxv. 58 (1923). The following num-
bers, out of a large vg being held for study, may be cited. UNC:
3122 (P); NAU: 2456 (P).

In his recent revision of the Sea Lavenders Blake characterizes the
New England material as having the calyx densely pubescent, at
least at the base. Nevertheless, specimens collected from the Eliza-
beth Islands show the calyx nearly smooth or with a few scattered
hairs and, according to Blake’s treatment, would come closer to L.
carolinianum, which name he restricts to the southern material with
glabrous calyx. It is obvious that the last word has not yet been
Pee concerning this genus as it is represented in the eastern United

ta
PRIMULACEAE
SaMOLUs FLoRIBUNDUS HBK. Sandy and not
: y and peaty pond margins;
permis NON: 2873 (N); NAU: 1078 (W), 2722 (N). NAU
IMACHIA QUADRIFOLIA L. Sandy woods; infrequent. ;
3327 (P); NAS: Northrop (0). ALG

- TERRESTRIS (L.) BSP. Open bogs and pond borders. NA‘:
Northrop (0), 2344 (N); CUT: 2316 e. re :

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 271

TRIENTALIS BOREALIS Raf. 7. americana (Pers.) Pursh. See
Blanchard, RHopora, xi. 236 (1909). Boggy woods. NON: 3297
(N,P); NAU: 3328 (N, FP); art Northrop (o).

Anagallis arvensis L. Cobble beaches and sandy waste places.
UNC: 2957 (N); NAS: Northrop (0); CUT: 3434 (P); PEN: 1450
(N,P,W).

OLEACEAE

FRAXINUS AMERICANA L, Certainly not common; seen only on
edge of swampy hollow along north shore of N ashawena. NAS:
Northrop (0), 3480 (N).

NSYLVANICA areas Reported by Mrs. Northrop from Nasha-
wena; no specimen
es The vulgare a ‘haaied from cultivation on Penikese. PEN:
pares
Bartonta virerinica (L.) B Open bogs and peaty hollows or
pond bo el eg 3782 Ips NAS: Northrop, Aug. 1901 (Y), 3508
(N); ee PF).
ee “atichs,) Robinson. A single specimen of t
reer a was collected in the woods on Naushon by Professor Fernald
and the writer on Aug. 10, 1927, but the material was subsequently
maa. The plant is surely uncommon on the Elizabeth Islands.
NYANTHES TRIFOLIATA L., var. MINOR Michx. See RHopoRA,
XXX1. 198 (1929). Seen growing ‘only i in a small boggy area bordering
ree Pond. NAS: Northrop (0); CUT: Hervey, no date (N), 810

NYMPHOIDES LACUNOSUM an Fernald. Collected in a small
pond on Nonamesset. NON: 2707 (P).

APOCYNACEAE

APOCYNUM ANDROSAEMIFOLIUM L. Exposed sandy bank bordering
beach on Uncatena; possibly more general. UNC: 3639 (N,P).

ASCLEPIADACEAE
ASCLEPIAsS TUBEROSA L. Sandy hillsides and clearings; rare and
local. NAU: Simons, July 16, 1901 (W), 2602 (N); PAS: 589 (W).
Pers

Milkweed on the islands. Swamps and wet pond margins everywhere.
NON: 2674 (N,P); UNC: 3050 (P): NAT: 2736 (N), NAS: Northrop
(0), son ); PEN: Jordan
ca L. Onenawaalri in ‘dry sandy ground. NAS: Northrop
(0); PEN: 1452 (W).
A. PHYTOLACcomDES Pursh. Reported by Mrs. Northrop as ie exal-
tata with the comment “Not uncommon.” No specimen see
A. verticittata L. Known only from Mrs. Northrop’s ~cilectiun
from Nashawena. NAS: adr vg Aug. 1901 (Y).

272 Rhodora [DECEMBER

CONVOLVULACEAE
ConvoLvuLus sEpIuM L. Occasional in a and overgrown
depressions. NON: 2890 (P); NAS: Northro
1uM L., var. PUBESCENS (Gray) Perhatil On sea beaches;
apparently not abundant. 1454 (N,P,
C. arvensis L. Cobble beach of serine on Penikese. PEN: 1453

Cuscura Gronovi Willd. Swamps, boggy eta 2 bs thickets,
ete., NAU: Pennell 3127 (W), 2495 (P), 2605 (N,P,M), 2914 (P,C),
2933 (P), 3139 NEM); 3890 (P,C); NAS: Northrop (o), " viliioma
July 10, 1911 (N), 2353 (P).

This is apparently the only species of Dodder on the islands. A
large and variable series has been ‘collected and has been studied in
the light of T. G. Yuncker’s recent monograph of the genus (1921),
but it has not been found possible to recognize more than this single
polymorphic species.

VERBENACEAE

VERBENA HASTATA L. Collected on gravelly upper beach on Unca-

- on in swampy clearing on Naushon. UNC: 2980 (N); NAU:
LABIATAE

‘TEUCRIUM CANADENSE L.., var. LITTORALE (Bickn. ) avg Abund-
ant near era borders of brackish areas, wet thickets,

Cushman, July 27, 1906 (B), 2586 (N); UNC: 2960 (P), 3047 (N,P ,C);
NAU: my (P), 2500 (P,M); NAS: linet: (0); PEN: 1460 (N,P W).

TRICHOSTEMA DIcHoToMuM L, Fre open downs and sandy

on “NON: 2623 (N); UNC: 2973 (Py, 3120 (N); NAS: Northrop

SCUTELLARIA LATERIFLORA L. te as “common” on Nashawena
by Mrs. Northrop; no specimen

S. EPILOBUFOLIA Hamilton. ce “dilerionlata of the Manual. See

UNC: 3045 (P); NAU: Cushman & Morse, “Au ug. 2 1906 (B), Pen
3116 (P), 2373 (P,M), 2607 (N); NAS: No rtp | ©), 3476 (P); cur:
Sanford, Aug. 15, 1917 (N), 815 (W,P); PEN: 1459 (N,P,W), 3406 (P)-
rrubium oulgare L. Sandy soil, waste ae NAU: Moore 1
) tl 011 (W) Northrop (o), Williams, July 10, 1911 (N);

peta Cataria L. Dry, sterile soil. UNC: 3054 (N); NAS: North-
ia oh )s BEN: 1458 (N,P,W). AS:
ru vulgaris L. A N
Nort (o) Reo g eee pparently not common; dry ground.
VvuLGaris L., var. LaNCEOLATA (Barton) Fernald. RHODORA, XV:
179 (1913). "The indigenous variety pee : met with on dry, sandy
slopes near center of Uncatena. UNC: 3032 (P).

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 273

Leonurus Cardiaca L. Fairly abundant on Penikese; not collected
elsewhere. NAS: Northrop (0); PEN: 1455 (W).

PycNANTHEMUM FLEXuUosuUM (Walt.) BSP. Known only from spec-
imen in Woods Hole herbarium collected on Naushon. NAU: Simons,
July 23, 1901 (W).

P. muticum (Michx.) Pers. Reported and collected by Mrs.
Northrop from Nashawena. NAS: Northrop, Aug. 1901 (Y).

Lycopus vireinicus L. Rare and local on the islands; seen growing
only on the sandy beach See West End Pond. NAU: Cushman
& Morse, Aug. 25, 1906 (B), 2926 (N).

L. UNIFLORUS Michx. The A abundant pha of the gen
here; pond margins, wet hollows, etc. NON: 2565 (N,P,W), 3663
(P); UNC: 2969 (N); NAU: Williams, we 10, ait (N), 2487 (P);
NAS: hoot ); vee 2525 (N); PEN: 1 7 (W).

L. americanus Muhl. Frequent; eee situations. UNC: 2970
Ne Nas: N iether re 3749 WN, P); CUT: 2524 (N,P), PEN: 1456

C).

Mentha spicata L. Occasional in wet hollows. NAS: Northrop (0),
3491 (P,M).

M. Cardiaca Gerarde. Reported by Mrs. Northrop from Nasha-
wena; no specimen seen

rispa L. Collected i in a swampy hollow at the east end of Cutty-
hier CUT: 2284 (N).

SOLANACEAE
Solanum Dulcamara L. Seen only bee! over a dense thicket
bordering pond on Uncatena. UNC: N).

- NIGRUM L. Very abundant, as a areal plant, often fringing the
sandy or cobbly beaches with a solid growth. NON: Cushman, July
ye 2(N. 5 ee (N,P); NAS: Northrop (0); CUT: 3446 (P); PEN:

S. rostratum Dunal. Sandy waste areas; infrequent. NAU: Taylor,
July 24, 1921 (P); PEN: Hollick (o).
8. triflorum Nutt. This western species was found growing on
beach near Tarpaulin Cove; possibly introduced in the days when
e Cove served as a refuge for sailing vessels. NAU: 1806 (N).
atura Stramonium L. Sandy and cobbly beaches; of casual occur
ree PNG: 2958 (N); NAU: Cushman & Morse, Aug. 25, 1906 (B);
61 (W).
Fp Table L. On Mrs. Northrop’s Nashawena list; may refer to

€ preceding; no specimen seen.

ae
erbascum sus L. Open, ground and waste places.
NAU: 2720 N, Py. NAS: Northrop (0); Sen: 1465 (N,P,W
maria vulgaris Hill. ogeeyone g not common on the islands; seen
only on Penikese. PEN: 857 (W). -

274 Rhodora [DECEMBER

L. CANADENSIS ol Dumont. Grassy slopes and open sandy
stretches. NON: 3 (P); NAS: Northrop (0), 3483 (N); CUT:
2331 (N); PEN: hap ‘P, W), 3403 (P

LIMOSELLA SUBULATA Ives. L. a aquatica, var. tenuifolia of Manual.
See Ruopora, xx. 160 (1918). Sandy and muddy margins of ponds
and Nashawena. PAS: 3745 (N,P); NAS: Northrop,

2

g
Fo

: f
ILysaANTHES pUBIA (L.) Barnhart. ss aid psd Mrs. Northrop
from Nashawena as I. gratioloides; no specim
I, rvaEquatis (Walt.) Pennell. J. ies (Michx,) Robins.
See Pennell, Torreya, xix. 149 (1919). Sandy and peaty pond borders.
PN. 2702 (N); NAS: Williams, July 10, 1911 (N); PEN: 454 (N,

ATIOLA AUREA Muhl. Peaty margins of small ponds. NON:
rd (P), 3194 (N,P), 3383 (N,P); NAU: Taylor 2347 (P), 1079 (P, W).

a ee L. A few plants persistent around leper cottages
on Penike PEN: 1463 (W).

Vendincs SERPYLLIFOLIA L. Reported from Nashawena by Mrs.
Northrop; no specimen seen.

V. PEREGRINA L. Occasional in sandy soil. The Naushon material
is very sparingly pubescent, approaching var. xalapensis (HBK.) Pen-
nell. NAU: 3330 (P); NAS: Northrop (0); CUT: Hervey, no date (N).

_ arvensis L. Seen only in a sandy field near center of Nonamesset.
NON: 3306 (N,P).
GALINIS PURPUREA (L.) Pennell. Gerardia purpurea L. See
Bull. Torrey Bot. Club, xl. 126 (1913). Listed by Mrs. Northrou for
ashawena; no specimen seen.

A. maritma Raf. Gerardia maritima Raf. Seen only on border of
brackish area on Uncatena; probably more general. UNC: 3005 (N,P).

MELAMPYRUM LINEARE Lam. Mrs. Northrop reports this for Nash-
awena; no specimen seen.

PEpIcULARIS CANADENSIS L, Dry, sandy beech woods near Hadley
Harbor. NAU: 3313 (N,P).

LENTIBULARIACEAE

UrricuLaRIA GEMINISCAPA Benj. U. clandestina Nutt. Ina small
pond on Nashawena; probably crea NAS: 3504 (P).

U. crpsa L. Known only from sheet in Woods Hole herbarium
apy ou Naushon. NAU: Simons, July 23, 1901 (W).

U. srrtora Lam. Reported from Nashawena by Mrs. Northrop:
no specimen s en. :

cornuTa Michx. Reported and collected by Mrs. Northrop:

NAS: Northrop, Aug. 1901 (Y). voir geal ad

_OROBANCHACEAE
Epiracus virerntana (L.) Bart. Fairly abundant in — woods

on Naushon; should be found on other islands as well. NAU: Wr
Ric. July 10, 1911 (N), 2937 (P), 3138 (N). :

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 275

BIGNONIACEAE

Catalpa bignonioides Walt. Well naturalized in certain areas on
Nashawena. NAS: 3546 (P).

PLANTAGINACEAE

PLANTAGO MAJOR L. Cobbly beaches, grassy slopes, etc.
2697 (P), 2900 (N); NAS: Werke (0); PEN: 1467 AW), oT

(N,P)

or L., ir INTERMEDIA (Gilibert) Dene. Material so labelled
was ‘collected by A. H. Moore on Penikese in 1904. Here also may be
referred beatae vely sabres collected by Hollick on Nadshion and
by Mrs. Northrop (labelled P. halophila Bickn.) on Nashawena. This
is done with the realization that the maritime phase of P. major is in
need of very careful study in order that its status may be satisfactorily
determined. NAU: Hollick, Aug. 11, 1898 (Y); NAS: Northrop,
Aug. 1901 (Y); PEN: Moore 1916 (G, N).

P. suncomwEs Lam., var. DECIPIENS (Barneoud) Fernald. P. mari-
tima of American authors in part, not L. See Fernald, Ruopora,
xxvii. 100 (1925). Collected on exposed ethers bie along north
shore of Naushon west of Kettle Cove. NAU: 3876 (N,P).

r. “acer peo R. & S. P. maritima of Feith authors in part.
See Fernald, |. c. 93-104. Salt marshes and brackish muds. PAS:
3740 (N, P.M oO. ‘CUT: 3568 (P),

P. lanceolata L. Grassland; not common. NAS: Northrop (0);
PEN: 1466 (P,W), 3394 (N,P).

P. lanceolata L., var. sphaerostachya Mert. & Koch. See Fernald,
Ryopora, xxiv. 203 (1922). Sandy and cobbly beach on Uncatena.

P. lanceolata L., var. sphaerostachya Mert. & Koch., forma eriophora
(Hoffmans. & L Link) Beck von Man. See Fernald, 1.c. Sandy soil at
east end of Cuttyhunk. CUT: 2269 (N).

P. aristata Michx. Occasional in sandy fields. UNC: 3652 (N,P);

U: Simons, July 23, 1901 (W); CUT: ehh: ).

P. vrrerntca L. Mrs. Northrop reports this from Nashawena,

saying, “Common on downs, about’ 2 in. high.” No specimen seen.

RUBIACEAE
atium Apartne L. Listed by Mrs. Northrop for Nashawena;
no specimen seen.
. PILosuM Ait. Dry sandy woods and clearings. NON: 2562 (P);
NAU: Pennell 3150 0(W), 2030 A aeons (P); NAS: Northrop (0).

- CIRCAEZANS M. ected by Pennell on Naushon. NAU:
Pennell 3110

TRIFIDUM 1 Borders of brackish areas; not common. NAU:
2513 (N), 3684 (P,M); NAS: Northrop to) 3541 (P).

Erroneously reported from Penikese; see Ryopora, xxvi. 228
(1924). The Penikese material proves to be G. Claytoni.

276 Rhodora [DECEMBER

G. Cuaytont Michx. Wet thickets, pond margins and gras
slope: very abundant. NON: 2572 (C), 2646 (N), 2666 (P), 2880 ®),
6 (P); UNC: 3027 (N); NAU: 2374 (P), 2716 (N,P,M); PAS
3748 (P); CUT: 2516 (N,P); PEN: 1468 (N,P,W).
. TRIFLORUM Michx. Collected only at aA a beach near east
end of N aushon; ee more general. NAU: 2 ;
TCHELLA REPENS L. Reported by Mrs. pon from Nasha-
wena; no area seen.
CEPHA ALANTHUS OCCIDENTALIS L. Not common; collected in boggy
pasture on Nashawena. NAS: Northrop (o), 3564 (Fy,

CAPRIFOLIACEAE

Lonicera japonica Thunb. Occasional as an escape. UNC: 3037
(N); PEN: 1469 (N,P,W).

RIOSTEUM sponte we hg L. Seen only on dry open hillside on
Pasque. PAS: 567 (N,P).

VIBURNUM ACERIFOLIUM L. Reported by Mrs. Northrop for Nash-
awena; no specimen seen.

V. PUBESCENS (Ait.) Pursh. V. venosum Britton. See Blake,
Ruopora, xx. 11 (1918). This is the only Viburnum seen growing
upon the islands. It is of frequent occurrence in moist hollows and
in thickets bordering ponds. rips 3010 (N,P); NAU: 2458 (P,M).
3320 (P,C); NAS: 3489 (N), 3518 (P).

VIBURNUM DENTATUM L. is reported by Mrs. Northrop from Nash-
awena. It is possible that this ies to the preceding, but in the ab-
sence — specimens this record is permitted to stand.

UCUS CANADENSIS L. Occasional; thickets and moist depres-
pew “NAS: Northrop (0), 3488 (N); PEN: 1470 (N,P,W).

CUCURBITACEAE
Cucurbita maxima Duchesne. Sandy soil near cultivated area on
Penikees PEN: 1471 (W).
CAMPANULACEAE
SPECULARIA PERFOLIATA (L.) A. DC. Reported from Nashawena
by Mrs. Northrop; no specimen seen <e
LOBELIACEAE
Lose s L. Found onl cl at the
east eno of reaaage N AU: 2466 (P,M i heir renin
: COMPOSITAE
EvPaToriuM VERTICILLATUM Lam Northrop reports E. mac-

a Mrs.
atum from Nashawena, but it is probable that this refers to the plant
treated by Wiegand as E. verticillatum (See Ruopora, xxii. 57 (1920),

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 277

a coastal plain species the general absence of which from the Eliza-
beths is indeed surprisi ae ie Northrop (0).

E. # Pex sca L. Rather frequent on open grassy and sandy
slopes. NON: 2590 (P); UNC: 2953 (N); NAU: Oakes? (Herb.
Siw) (B); PAS. 1790 (P,W).

E. VERBENAEFOLIUM Michx. Reported by Mrs. Northrop from
Nashawena; no specimen seen

E. PERFOLIATUM L. Wet hollows; uncommon. UNC: 2966 (N);
NAS: Northrop (0).

CHRYSOPSIS FALCATA (Pursh) Ell. Open downs and sandy clearings.
“a 2583 (P); NAU: 2596 (N,P); PAS: 3788 (N); NAS: Northrop

SoLipaco BicoLor L. Dry, sandy woods; not abundant. NAU:
3161 (N), 3905 (P).

S. uLiginosa Nutt. is reported from N ashawena by Mrs. Northrop;
this doubtless refers to some other species, but since, in the absence of
ec pee it : cat baggy to say which one, this record is wore ]

EMPERVIRENS L. Sandy and cobbly beaches. NON: 3211 (P);

UNC: 3110 (N) 3 a 3744 (P); NAS: Northrop (0); CUT: 3070 (N);
PEN: 1492 (N,P,W).

S. suNcEA Ait. Known only from bare gravelly slopes on Penikese.

PEN: 1489 (P,W).

. SUAVEOLENS Schoepf. SS. odora Ait. See Standley, Ruopora,

xxi. 69 (1919). Collected only in dryish peaty border of bog on Cutty-

hunk; perhaps more abundant. NAS: Northrop (0); CUT: 2520 (N,P,

S. Evuiorm T. & G. Dry thickets on Uncatena. UNC: 3105 (P),
3111 (N,P,M).
S. Rucosa Mill. Dry, boulder-covered slopes on Penikese. PEN:
1491 (N,P,W).
S. RUGOSA ae var. ASPERA Guns baer S. aspera Ait.
HODORA, xvii. 7 (1 915). or y downs and pond borders.
ee nae 3186 (N,P,M); UNC: 309 3 (P); CUT: 2521 (N,P).
s Ait. xt meer UNC: 2971 (N, P); PAS: 3797
(P); NAS as: : Northrop (0); P. 90 (W).
» CA pepe L. Callected -_ on dry southwest slopes of Pen-
ikese, ‘PEN: | 488 (W).
S. ALTISSIMA an spec hhaear only in thickets on dry hillsides of Un-
any UNC:3
S. stent y cmatiel “Tee rywhere on open downs and borders of
Gaylussacia thickets. NON: 2561 (P); 3168 (N,P); UNC: 3084 (N,
P, ine NAU: Cushman, Aug. 25, 1906 (B), re (M,C), 2719 (N, P)
nee sn op (0); CUT: Sanford, Au
PEN. ipa rop (0); C ft g
S. GRAMIN ees “ays lisb. Same situations as the preceding;
ends 50 ts abundant. NON: 3179 (N,P,M,C); UNC: 2952 (N, P);
¥)-

278 Rhodora [DEcEMBER

S. minor (Michx.) Fernald. This southern species, ranging from
Mississippi and Florida to Virginia and reported by Bicknell from
Nantucket, was collected by Professor Fernald and the writer on open
ay el hillsides near the central part of Naushon. NAU: 2506 (N,P).

TER -aacinge L. Occasional in dry open woods. NON: 319
Si i 160 (N,P
NS ats Gra ssland oa sandy thickets. NON: 3190 (P);
Nau: 3142 LN); ae Nort p (0).
is common. NON: 3183 yy P), 3187
wr C): UNC: 3080 tN Pao. NAU: 3158 (N,P,M); PEN: 1474 (W ).

A. rrtcorpEs L. Reported by Hollick from Pasque; no specimen
seen, nor has this Aster, so abundant on the adjacent mainland, been
foune rowing on the Elizabeth Islands, despite diligent search for it.

LTIFLORUs Ait. Fairly abundant on grassy slopes. NON:
i (N), 3192 (N,P); PAS: 3751 (P, C); CUT: 3577 (P,M); PEN:
ae C, ve
“ae h the last. NON: 3178 (P), 3189 (N,P,W);
eee @) UNC: ee MN P), 3091 (P); NAU: 3144 (P), 3159 (P);
PAS: 3764 (N,P).

A. misbpeit L., var. cormprroiius (Michx.) T. & G. Clearing in
dry beech monde near Mary’s Lake. NAU: 3145 (N, P).

A. puMosus X ? This plant, probably representing A. dumosus
x simi, was collected in sandy woods on Naushon. NAU: 3148

A. v1 ae Lam. Abundant in penne woods and on open downs.

NON: 3196 (P,M,C), 3200 (N,P,W,M); UNC: 3100 (N,P); NAU:
3134 (N,P,C), 3157 (P); PEN: 1475 (P).

A. LaTeRIFLorus (L.) Britton. Apparently not common; dry,
shady beech woods on Naushon. NAU: 3146 (N).

A. Novi-BELcit L, Abundant in dry, open ground, especially along
yee of thickets; extremely variable. NON: 3181 (W,C), 3182
QM), 31 Lee 92 & PC), ).g201 (N,P); UNC: 3090 (N, P), 3101 (P); PAS:

A. cae Mill Reported by Mrs. Northrop from Nasha-
_ no specimen —

. LINARIIFOLIUS L open hillsides; not frequent. UNC:
os (N); PAS: 3790 oe NAS: Northrop (o). NC: 3643
RIGERON RAMOSUS (Walt.) BSP. n sandy slopes. U
(P); NAS: Northrop (0. ee Open sandy slop
. CANADENSIS L. Disturbed, sandy soil; rare. PEN; 1485 85 (W).

E pusiLtus Nutt. See Robinson, Ruopora, xv. 205 (1913). Oped
sandy slopes. NON: 2902 (N,P); PAS: 3768 (5 PEN: 1486 86 (P, he

ee wns.

specimen seen :
Pideuis ¢: CAMPHORATA (L.) DC. Margins of brackish ponds and

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 279

borders of salt marshes. NON: 2289 (N), 2662 (P); UNC: 3036 (N); .
NAU: Cushman & Morse, Aug. 25, 1906 (B), 1799 (N,P,W).

ANTENNARIA PLANTAGINIFOLIA (L.) Richards. Grassy knolls and
exposed hillsides. NON: 2670 a 3280 (N,P); UNC: 3640 (N,P);
NAU: “ye ar) abe ribet 0).

A. NEGLECTA Greene. cairiatbin as last. NON: 3281 (P);
NAS: 3506 (P), 3554 (N); “OUT: 345. = (N,P).
ANAPHALIS MARGARITACEA (L.) B. & H. Grassland and open

ground everywhere. NON: 2560 (P): UNC: 3035 (N); PAS: 3796
@); ae Northrop (0); pope 3454 (N,P).

UM sebered tiom L. G. pes Aes gece Michx.
ne oie ite slopes wer clearings. UNC: 2954 (N), wen (P).
NAU: 2730 (N):; PAS: 3 3760 (P); NAS: Mines: (o); CUT: Sanford,
Aug. 15, 1917 (N); PEN: 1482 (N,P,W,M,C).

G. uLiernosum L. ve menor: and low sandy ground, occasion-
ally spake a salt marshes. 3006 AU: Cushman &
ise Sg Sk , 1906 (B), eae TN. P): NAS: "Northrop (o); PEN:

"6 PURPUREUM L, Dry, sandy soil; rare and local. NAU: 834 (W);
NAS: 3507 (N); CUT: Hervey, no dat

Iva orarta Bartlett. Reported by J ordan from Penikese; no spec-
imen

AMBROSIA ARTEMISIFOLIA L. Not common except on Penikese
where abundant on grassy slopes of aa NAS: Northrop (0);
PEN: 1473 (P,W).

XANTHIUM CANADENSE Mill. Reported by Mrs. Northrop from
Nashawena; no eae seen.

ECHINATUM Occasional on beaches. NON: 3177 (N),
3914 (P); PAS: ite rs PEN: 458
udbeckia hirta L. dpe UN C: 3641 (P); NAS: North-
rop (0); CUT: 2329 «N) PPEN: 1 P,W

Helianthus annuus “ seeds au near area under cultivation
on Penikese. PEN: 1483 o W).

Coreopsis rosea Nutt. Apparently rare; seen only on sandy beach

rdering West End Pond. NIU: 2916 (N,P,M,C).

C. lanceolata L. Sasigsiogy: around leper cottages on Penikese,
where planted. PEN:

IDENS CONNATA ae Border of small pond in hills at north end
of Penikese. PEN: 1476 (N,P,W
gala Muhl., var. PETIOLATA (Nutt. ) Farwell. See Ruopora,
7 (1908). Wet "borders of ponds; occasional. UNC: 3016 (P),
3115 (N,P); NAU: 3133 (N,P,W,M,C).
CHILLEA drain! Me nesoree eveON, 31 eA eo ee

hillsides and in sandy and cle NON: 3166 (N); UNC:

PENN: Met Novheeg ah 3555 (P): ys CUT: 2260 , M), 2322 (N);
N,P,W).

iia Conk t. Disturbed sandy areas. UNC: 2955 (N);

NAS: Northrop (0); CUT: 3580 D); PEN: 858 (N,P,W).

280 Rhodora [DEcEMBER

nitro 2 Leucanthemum L., var. pinnatifidum Lecoq. & La-
motte. Certainly not common; occasional on open grasslands. NAS:
Notieop (0); PEN: 1477 (N,P
Ace onsig vulgare L. cL. soil near habitations. PAS: 3767 (N);
PEN: W).
ries Stelleriana Bess. Occasional on sandy and cobbly beaches.
NON: Cushman, July 27, 1906 (B); CUT: 2275 (N,P,M).
ERECHTITES HIERACIFOLIA Cw) Raf. Sandy clearings; moist pond
margins and depressions. NON: 2696 (N); NAU: 2928 (N); NAS:

. HIERACIFOLIA (L.) Raf., var. PREALTA (Raf.) Fernald. See
gr pane xix. 27 (1917). Moist hollow. NON: 2661 (N,P).

E. m OCARPA Fernald. See Ruopora, a 24 (1917). Moist

> i asa local. UNC: 2965 (P,C) ; 2984 (NP y
retium minus Bernh. Reported by Mrs. Sata for Nashawena;
no specimen see

Cirsium cae ee (L.) Hill. Sandy and peea! banks and hill-
sides. NON: 2693 (N,P), — July 28 6 (B); NAU: 2717
(N,P); NAS: Northrop (0); CUT: 827 (P,W); PEN: 1479 (P,W).

C. HorriDuULUM Michx. C. spinosissimum of the Man oF oe
Robinson, Rxopora, xiii. 239 (1911). Occasional in ope
situations. NON: 2299 (N); NAU: Miss Weir, June 2 22. 1890 Pas
NAS: Northrop (o).

C. piscotor (Muhl.) Spreng. Occasional in sandy clearings and
borders of thickets. UNC: 3104 vd PAS: 3752 ee

entaurea arenaria Beh. “Re ported from eshenet by Hollick
as the first record for North America;! no specimen se

Cichorium Intybus L. Listed by Mrs. N orthrop tas Nashawena;

n seen
reap vinctvtca (L.) Willd. Sandy fields. NON: 3305 (NP);
NAU: 3879 (N,P).

sea (Walt.) Blake. K. amplexicaulis Nutt. See RHODORA,
xvii. 135 (1915). On Mrs. Northrop’s Nashawena list; no specimet
seen.

Leontodon peatuemenales L. Oceasional i in aE or on dry banks. —

L. ananake La r. pratensis Sock: In tall pee on slopes at
west end of Penikese, "PEN: 1 1484 We
araxacum officinale Weber. mmon; collected only neat

—- on Pens, NAS: NAS: Norv ae PEN: 1496 (N,P,W). Ze
ional on beach d sandy areas. :
Northrop ( (0); PEN: ‘a08 N, PW W.M, a. es and sandy ar

1 Hollick, A. Cont. Geol. Dept. Columbia Univ xi. no. 72. (1901)

1930] Fogg,—Flora of the Elizabeth Islands, Massachusetts 281

S. oleraceus L. Seen mostly growing in the litter of upper sea
beaches. CUT: 3432 (P); PEN: 1495 (N,P,W), 1755 (P).
per (L.) Hill. With the last; less common. CUT: Pennell 2992
: 1494 (W).

ACTUCA CANADENSIS L. Reported by Mrs. Northrop from Nash-
awena; no specimen seen.

. CANADENSIS L., var. oldies Mpeoe: Gray, forma ANGUS-
TATA Wiegand. See RuHoporA, xxii. 10 (1920). Collected in dry,
sandy soil at east end of Pasque. PAS: as (P).

L. nirsuta Muhl. Seen growing only in sandy woods on Naushon;
probably more common. NAU: 2464 (N); NAS: Northrop (0).

PRENANTHES morn Felon Mrs. Northrop reports this from
Nashawena; no specim

ig oe aa) eke Sandy woods; apparently rare.
NAU: 3921 (P).
Hir ee PANICULATUM L. Dry, sandy woods; not common.
NAU: 3920 (N,

H. MARIANUM Willd. Reported from Nashawena by Mrs. Northrop;
no specimen see

H. scapruM Miss. On Mrs. Northrop’s list for Nashawena; no
specimen seen.

<meted L. This is the ee member of the genus on
these islands; dry woods, sandy cl gs, etc. NON: 2634 (P), 3185
(P); NAU: = (N); PAS: 572 (Ww), ‘3775 (N); NAS: 3536 (P).

. CANADENSE Michx. Apparently not abundant; collected only

in dry, sandy soil at east end of Pasque. PAS: 3774 (N); NAS:
Northrop (o).

UnIversITy oF PENNSYLVANIA.

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY.

No. XCII.

ISSUED CCT 20 1930

a 2QLet | MARY
DN iv Ay),
STUDIES IN THE BoraGINnacEAE,—VIII. Un ~ 7 FONT
ne eid

iL<
I. Observations on the Species of Cordia and T , neler at |
from Brazil, Paraguay, Uruguay and Argentina. ,
1. Prefatory Discussion

hg Reva bi cA A 3
me ameetinent of Condit. 35.50. on hee wee 5
3. Treatment of Tournefortia ge en eee nes 66
I. Taxonomic notes concerning various Borages......----.-+-- 89

By

Ivan M. JOHNSTON.

LISHED BY
THE GRAY HERBARIUM OF ANAS? UNIVERSITY
CAMBRIDGE, MASS.,
1930.

I. OBSERVATIONS ON THE SPECIES OF CORDIA AND
TOURNEFORTIA KNOWN FROM BRAZIL, PARAGUAY,
URUGUAY AND ARGENTINA.

1. Preratory Discussion.

ALMostT simultaneously about two years ago I received two large
collections of Cordia and Tournefortia from Brazil and Argentina.
In an attempt to name this material it quickly became evident that
the existing treatments of these genera for the species of Brazil and
the countries to the south, indeed for those of the whole of South
America, were antiquated and completely inadequate and that even
reasonably precise identifications for the specimens would necessitate
a rather extensive survey of all the species of eastern and southern
South America. Upon attempting this survey I soon found so much
that needed careful attention as to arouse particular interest and I
decided to give the groups more detailed study and to prepare a revision
of them which was so obviously needed. These genera, accordingly,
have been a special interest for the past year, during which I have seen
many hundred specimens including practically all those of historic or
critical interest found either in this country or abroad. As a result I
am able to present in this paper critical synopses of the species of
Cordia and Tournefortia known from Brazil, Paraguay, Uruguay and
Argentina. Although by no means a final treatment of these complex
groups in the region mentioned I hope it will at least be a workable one
and that it may make identifications of materials from the region not
only less onerous but far more precise than in the past.

For ordinary considerations the work on the species of Cordia and

ournefortia in our region may be said to begin with Chamisso who
published, in 1829 and 1833, a large number of the most common
species. This work is of a model sort, Chamisso having carefully
compared his plants with authentic material of the older species and
described them at length with a keen appreciation of detail. It is
marred only by the very unfortunate vagueness as to the precise
locality at which the types were obtained, an obscurity which in the
last analysis must be attributed to Sellow who collected the specimens.
If we except the sketchy hack compilation by Giirke, E. & P. Nat.
Pflanzenf. iv. Abt. 3a, 81-85 and 91-92 (1893), the only notable general
4ecounts of our species are those by De Candolle in the Prodromus,
ak 468-501 and 513-531 (1845), and by Fresenius in the Flora Brasil-
‘ensis, Vili, pt. 1, 3-28 and 48-58 (1857), which were published respec-

tively 85 and 73 years ago. De Candolle published an account of the

4 JOHNSTON

genera for the world. The studies were made in his usual scholarly
manner, his treatment still standing as the last and best general ac-
count of these two large and difficult genera. So far as it relates to the
species here under consideration it is necessarily brief and contains
many errors, but it is nevertheless remarkable in view of the fact that
its author had from the region less than fifty specimens of Cordia and
much less than half that number of specimens of Tournefortia. The
work on the Brazilian species of Cordia and Tournefortia by Fresenius
is quite uninspired and although impressive in setting falls much
below the scholarly standard generally maintained in the Flora
Brasiliensis, being scarcely more than an indiscriminate compilation
deriving much of its merit and part of its faults from the work of De
Candolle upon which it was largely based. The author showed no
particular feeling for relationships and seems actually to have had
scruples against the reduction of any species proposed by his prede-
cessors, an attitude of mind which I have been unable to evoke in
saving from the depths of synonymy the majority of the species he
proposed.

The subjoined treatment of the Cordias and Tournefortias from
Brazil, Paraguay, Uruguay and Argentina is the first one ever bas
upon a large series of specimens. For its preparation I visited Europé
in 1929 and studied the important collections of the Royal Botanic
Gardens at Kew (K), the British Museum of Natural History at Sou
Kensington (BM), the Botanical Museum at Berlin-Dahlem (BD),
and the Delessert Herbarium (DL) and Barbey-Boissier Herbarium
(BB) at Geneva. Also studied at these herbaria were various smal
but very important collections that have been kept as units. The most
significant of these for my work were the Willdenow Herbarium (Willd)
at Berlin and the Prodromus Herbarium (DC) and Hassler Herbarium
(Hass) at the Delessert Herbarium in Geneva. In the United State
I have visited and studied the collections at the New York Botanical
Garden (NY) and at the National Herbarium at Washington (US).
For detailed study at the Gray Herbarium I have also receive on
loan a large number of selected specimens from various of the instl-
tituins I have mentioned. My most important loan, however, was the
very extensive and extremely interesting one received from the
tanical Section of the Riks Museum at Stockholm (RS). Supple-
menting the large amount of authentic and diversified mater
studied either abroad or here in the United States are three lars
collections of Cordia and Tournefortia which I have had the very good
fortune to receive from South American botanists. The first of thes?
was from Sr. F. C. Hoehne of Sio Paulo and consisted of the ve

STUDIES IN THE BORAGINACEAE 5

interesting Brazilian collections from the herbarium of the Botanical
Section of the Museu Paulista (SP), a section that has subsequently
become a part of the Instituto Biologico de Defesa Agricola e Animal.
From Prof. M. Lillo of the University of Tucumén I received another
valuable collection, largely’ of species from northwestern Argentina.
Beautifully supplementing this, and thus giving me a very repre-
sentative collection of the Argentine species for study, is a collection
received more recently, through the interest of Prof. José Molfino,
from the herbarium of the Instituto de Botanica y Farmacologia of
Buenos Aires. Without access to the wealth of material that I have
been able to examine from all the above named sources I could scarcely
have prepared this paper. I was welcomed and treated most kindly
by those in charge of the institutions which I was fortunate enough to
visit and have met with continued courtesy and helpfulness from those
of my other friends that have assisted me from a distance by sending
specimens and notes. To all these, both to those whom I have met
and to those whom I know only through correspondence, I wish to
express my keenest appreciation and sincerest thanks for their valu-
able help and sympathetic interest.

Practically all the material examined during the preparation of
this paper has been cited under the species concerned. The location
of the particular specimens seen is indicated by initials (characteristic
of the names of the various herbaria) that are enclosed in parentheses
and directly follow the collection-number. In citing collections I
have, with the investment of no small effort, done my best to arrange
them geographically, grouping them first under countries and then
under the major political division of the country. This procedure is
very time-consuming and risks possible error, but in any case has
permitted me to give the collections a logical arrangement and one
best calculated to bring out most clearly the salient features of geo-
graphical distribution. It may be well to explain that I have purposely
ignored one boundary that has a place on the map of Brazil, namely
that between the State of Rio de Janeiro and the small Federal Dis-
trict which includes and extends somewhat to the west of the city of
Rio Janeiro. I have arbitrarily treated the Federal District as though
it were a part of the State of Rio Janeiro. In defense of this action I
can plead, not ignorance, but merely reasons of convenience.

2. TREATMENT OF CORDIA.

J .
Among the species of Cordia in our area only those ranging to the
south and west of the southern portions of the state of Minas Geraes

6 JOHNSTON

have been collected sufficiently to be regarded as reasonably well
understood, but even these, except about the important cities near
the coast, must be collected very much more before the details of
their distribution are known and the final word on their classification
be said. North of southern Minas Geraes most of the collecting of
Cordia has been done in eastern Bahia and about the city of Paré.
Not only do we have very imperfect knowledge as to just what species
occur to the north of southern Minas Geraes but we have only frag-
mentary information as to the distribution there of even the most
common species. It seems certain that it will be in this extensive
region that most of the as yet undescribed species will be discovered.
The region contains a number of endemics, most of them apparently
local, some having not been re-collected since they were describ
fifty to seventy-five years ago. It will be after a study of extensive
collections from this eastern corner of Brazil that the most important
additions and corrections to this present treatment will be made. The
species of the Amazon Basin are also in much need of further repre-
sentation. The existing collections of them are most inadequate and
must be greatly augmented before we can have even a reasonably
workable provisional treatment of them. The one here presented is,
from lack of sufficient material, very unsatisfactory.

Not only must the final statement on the classification of the
Brazilian species of Cordia await the time when more representative
and numerous collections are to be had for study but it must also await
the day when more progress has been made in the classifying of the
species of the genus from adjacent countries. Our spicate species of
the section Varronia can be said to have had only preliminary treat
ment until the species of northwestern South America, where that pat-
ticular group centers, have been worked out in detail and our plants
correlated with them. Our Amazonian species of the section Pih-
cordia can not be definitely settled before the related ones of norther®
South America, particularly those of the Guianas, are understood.
Not only will these necessary studies of extralimital forms undoubtedly
give reasons for some readjustments in specific lines but, I fear,
almost certainly show the necessity of certain changes in accepted
specific names also. There are a half dozen species of Cordia from
French Guiana that were described over a hundred years ago whi
have yet to be identified or correlated with any of the more recent
species. The identification of such obscure or confused species must
have its effects on the taxonomy of our plants.

Although this paper is primarily a regional study of Cordia I have
taken every opportunity to study our plants in relation to those from

STUDIES IN THE BORAGINACEAE 7

other parts of America. As a result of the general familiarity with the
various American species thus obtained and with the benefit of certain
special studies I have been able to work out what seems to be a natural
grouping of the American species of Cordia. The definition and
characterization of the sections of Cordia that are presented in this
paper are accordingly new. It will be noted that I have been unable
to agree with the system of subgeneric groups, based upon anatomical
studies, that has been proposed by Mez, Bot. Jahrb. xii. 526-588
(1890). The definition of Mez’s groups (indiscriminately called by
him groups, sections and subgenera) is entirely in terms of anatomy.
The result is a series of groups too impracticable for ordinary syste-
matic work and, if we may judge of them by the representative
species cited, too open to serious challenge as to their naturalness.
Unfortified by a microtome I am quite unable to justify such groups
as those Mez has treated as Myxac, Strigosae, Pilicordia and Crassi-
foliae. They seem to be quite heterogeneous and incapable of defini-
tion by those characters of habit, leaves, flowers and fruit that are,
not only readily examined, but have been tested and are known to
have phylogenetic significance.

Key To THE SECTIONS OF CoRDIA IN AMERICA.

lla marcescent and very tardily deciduous.
as with fiber. sae fd Bene es walls, ih base of or sausage-
ped, capped inous discoid base of the style,
completely and a eae by the calyx and eles >
c AG saa es ae rite ae eee GERASCANTHUS, pg. 8.
Fruit with bonty Min Stine ir — below the middle con
t g the style, only loosely
2 pectially lothed be tie calyx, falling, away naked; a
xican RHABDOCALYX.
Corley a Se Le and soon deciduous
Ts in pte es or heads or glomerules, baie i in small open

ecovoeeeve tees

Fl y only very shallowly and broadly lo _.Varronia, pg. 18.
owers in very loose panicles or cymes; fruit larger, 6-16 mm.

Corolla large and on gt 1.5-6.5 em. long; pa be remain-

ei yank completely investing the fruit, usually
poh accresc: ha eee nov ieakees a corpiA, pg- 40.
Corolla small, 0. 6-1. 5 cm. long; cal explanate or

reflexed or falling cin scarcely if at pe scereacen nt, not

Corolla funnelform iow broader than lone: distinctly
emarginate; calyx 3-5-toothed and usually irregularly

co
which ranges from Mexico and the
West Indies to pot tens South America; Calyptracordia
Brit CALYPTRACORDIA.

ele eb AS Oo ee a ee ee 8 ee eS Oe Re ee 8 Ee Fee ee

8 : JOHNSTON

Corolla salverform; lobes commonly elongate and longer
han broad, apex rounded.
Stems with conspicuous subnodal swellings that serve as

SNe ee Es oi tien S's g Puysocwapa, pg. 46.
Stems without subnodal swellings; plant not myrmeco-
p.

Section Gerascanthus [P. Browne] Don, Gen. Syst. iv. 380 (1837);
DC. Prodr. ix. 471 (1845). Gerascanthus [P. Browne, Hist. Jamaica
170, t. 29, f. 3 (1756)] Raf. Sylva Tellur. 40 (1838); Lindley, Veg.
Kingd. 629 (1847); type-species, C. Gerascanthus L. Cordia subgenus
Gerascanthus Schlecht. & Cham. Linnaea v. 115 (1830). Cerdana R.
& P. Prodr. 37, t. 6, (1794); type-species, C. alliodora R. & P. Cordia-
Cerdanae R. & S. Syst. iv. 467 (1819). Cordiada Vell. Fl. Flum. 98
(1825) and Icones ii. t. 156 (1827); type-species, C. trichotoma Vell.

The section Gerascanthus is restricted to America and consists of
about a dozen species. Of these only one, C. alliodora, is widely
distributed, the remaining species centering in Mexico and southern
Brazil. The section is a very distinct and natural one being closely
related only with the monotypic Mexican section, Rhabdocalyx with
which it agrees in all corolla-structures but differs from in profoun
structures of fruit. The fruit of the section Gerascanthus is, indeed,
unique in the genus. It is sausage-shaped or ellipsoidal and has the
rounded or truncate apex capped by a definite discoid cartilaginous
expansion of the style-base. The carpellary walls are not thickened
and bony but thin, fibrous and chartaceous. Only a single seed is
normally matured. I have a suspicion that the flowers are only 1-2-
ovulate. In maturing the fruit becomes tightly invested by the calyx
and commonly also by the tube of the marcescent corolla, When it
is ripe the flower detaches at the base of the calyx from the short
pedicels and the whole, calyx, corolla and inclosed fruit, fall away, with
the corolla perhaps acting in some species as a parachute and thus
aiding somewhat in the dispersal of the seed. Also distinctive of the
section Gerascanthus are its marcescent corollas, a development which
it shares in the genus only with C. elaeagnoides DC., the single species
of the related section Rhabdocalyx. These corollas, white or occ
sionally yellow, when fresh, do. not drop away after anthesis but
drying buff or brownish remain expanded, almost unshriveled, i
position until they fall with the calyx and fruit or, as rarely happeDS —
are pushed out of the calyx on top of the maturing ovary. es
corolla is somewhat funnel-form with a well developed tube that
abruptly widened above to form a shallow throat and five ascending

STUDIES IN THE BORAGINACEAE 9

usually pinnately veined lobes. Usually included in the calyx the tube

protrudes conspicuously only in C. insignis. In most species the

corolla-lobes are oblong with more or less definitely parallel sides and

a truncate, rounded or abruptly obtuse apex. In C. insignis and C.

glabrata, however, they tend to be orbicular and in C. latiloba they
are ovate and evidently contracted to the apex. The calyx is cylin-

drical and except for species like C. Goeldiana is definitely 10-costate.

It is not conspicuously accrescent in age being only stretched by the

ripening fruit which it incloses.

Key To Spectres.
Leaves glabrous.
Corolla-lobes deltoid-ovate, about as long as broad, from below

e middle contracted to a broadly triangular apex; Rio

UE a iE FUEL i ENE ENE OY Fe 1. C. latiloba.
Corolla-lobes oblong, about twice as long as broad, apex obtuse :
ig PAIRS Sey a ab a analePre ge nalaren ant col has Semana Clb oad 2. C. Goeldiana,

Leaves pubescent. : :
Pubescence on stems and leaves almost exclusively simple; co-

mm SGxture, tsually dlonigate... ©. 6.5.66. ck chops ce nes . C. insignis.

Corolla moderate; calyx 10-14 mm. long; leaves somewhat

lue above, texture much less firm, outline ovate or
ess | MPRA a apn ot ASHORE ae Dag | Hl aber myip lias 4. C. glabrata.

bescence on stems and leaves almost exclusively stellate. :

Corolla small, lobes 1.5-2.5 mm. broad; Matto Grosso. .5. C. alliodora.

Corolla moderate, lobes 3-7 mm. broad: east and south of
M RON i dic ia aks divi etd codes we ss 6. C. trichotoma.
I. Cordia latiloba, sp. nov., arbor ?; foliis graciliter longi-petiolatis
ellipticis vel obovato-ellipticis integerrimis in sicco nigrescentibus,
subtus subpallidioribus opacis minute elevato-reticulatis, apice abrupte
cutis vel breviter acuminatis, basi gradatim vel rariter abrupte
Contractis, lamina petiolo duplo vel triplo longiori 2.5-7 cm. lata
4-15 cm. longa; inflorescentia terminali paniculata multiflori, partibus
juvenalibus sparse inconspicue brunneo-puberulentis minute glandu-
liferis mox glabrescentibus; pedicellis gracilibus rigidis 2-5 mm. longis;
calycibus cylindricis 10-costatis ut videtur glabris sed sub lente

inconspicue sparseque glandulifero-puberulentis mox glabrescentibus

2 mm. longis 2-3 mm. crassis fructiferis accrescentibus fructus
arcte investientibus apice inlobos 2-3 rotundos vel obtusos inaequales

3 mm. profunde fissis; corolla infundibuliformi marcescenti 25-33
mm. longa glaberrima tubo in calyce occulto, lobis 5 deltoideo-ovatis
ascendentibus 9-11 mm. longis latisque apice late acutis haud trunca-
~~ sinibus aliquantum plicatis; staminibus 5 apice tubi corollae in-
Sertis filamentis ad basem versus barbellatis; fructu botuliformi ca.
12 mm. longo ca. 3.5 mm. crasso glaberrimo.

10 JOHNSTON

BRAZIL. Rio JANeErRo: indefinite, Glaziou 1106 (Kew, Type); Rio Janeiro
Riedel (RS).

This species is closely related only to C. Goeldiana, a plant known
only from a single station in eastern Parad. It differs in having the
corolla-lobes deltoid-ovate, about as broad as long and _ broadly
acute, rather than having them oblong, twice as long as broad and
obtuse or even retuse. In addition the new species has the calyces and
the branchlets of the inflorescence much less copiously puberulent and
glandular than its northern relative. The corolla-lobes are very
broadly acute, the margins, which start to converge below the middle
of the lobe, meet at the apex forming a definite angle of 65°-90°. The
species is a distinctive one in the section Gerascanthus and is notable
even among the representatives of the entire genus in South America
because of its glabrous herbage, very elongate petioles and deltoid
corolla-lobes. The only material I have seen of it consists of two col-
lections with incomplete data indicating that the home of the species
must be in the general region centering about Rio Janeiro. That
such a well marked species can have lurked in this region and be so
poorly represented in herbaria is only to be explained by assuming it
to be very rare or local.

2. C, Goeldiana Huber, Bol. Mus. Goeldi vi. 89 and 201 (1910).

Known only from the eastern part of the state of Pard.

BRAZIL. Pari: Pei i, rai ;
1907, Goel 8510 (BM, BD DL, HS, worvens)s Beweber Get, 3, 1007,
Siqueira Rodrigues 8788 (BM).

This very distinct species is known only from the type-locality.
Its nearest relative is C. latiloba, a species from the region about Rio
Janeiro. Huber described the fruit of C. Goeldiana as being globose
and 2 cm. in diameter. I have a very strong suspicion that this is
quite incorrect and that perhaps Huber’s description is based upo?
extraneous material belonging to some wholly different species, Pe?
haps of even a different family. Cordia Goeldiana is so clearly a mem-
ber of the section Gerascanthus and so obviously related to C. latiloba
that it is unbelievable that it should have fruit so very markedly
different from that species and the other members of the section. !
feel certain that the fruit of Huber’s species will be found to be e>
sentially the same as that produced by C. latiloba.

3. C. insignis Cham. Linnaea viii. 122 (1833); Fresen. in Mart.
Fil. Bras. vu. pt. 1,3, t. 1 (1857). Lithocardium insigne Kuntze, Rev-
Gen. ii. 977 (1891). Gerascanthus aspera Mart. (Herb. Fl. Bras- P&
166, sub. no. 269) Flora Regenb. xxi. Bd. 2, Beibl. 86 (1838). C.

STUDIES IN THE BORAGINACEAE ll

Martu DC. Prodr. ix. 471 (1845). C. Nettoana Taub. Bot. Jahrb. xv.
Beibl. 38, 11 (Feb. 1893). C. jucunda Moore, Trans. Linn. Soe. ser. 2,
iv. 401 (Dec. 1893).

From Minas Geraes and Rio Janeiro westward through Matto
Grosso to easternmost Bolivia.

BRAZIL. Mrnas Geraes: indefinite, Lhotzky (BD, Trpx of C. insignis).
jou 112 :

Aug. 4, a Cc
Hoehne 1612 (BD); Morro de Ernesto, Cuyaba, Martius Herb. Bras 269 (NY,

, UC, ISOTYPES OF G. aspera); s 3 m. tall, corolla creamy white, rocky
soil, Cuyabdé, May 1927, Smith 125 (K); arborescent shrub 2 m. tall, branches
few and somewhat exuose, bark smooth, corolla yellowish green oming
white, Cuyabd, April 2, 1894, Malme 1504 (RS); shrub ca. 1.7 m. tall, Cuyaba,
Jan. 10, 1902, Malme 1680 (RS); large shrub, Serra do Itapirapuan, April 28,
1894, Lindman A3355 (RS); Matto Grosso, 1886, Leeson (BM, Tyre of C.
jJucunda); shrubby plant of the campo, 1-2 m. tall

BOLIVIA. Santa Cruz: corolla yellowish, Velasco, 200 m., July 1892,
Kuntze (NY, BD).

Var. glabrifolia (Kuntze), comb. nov.—Leaves glabrescent beneath
and conspicuously less reticulate than in the typical form.—Lithocar-
dium insigne, var. glabrifolium Kuntze, Rev. Gen. iii. pt. 2, 206 (1898).

Known only from Matto Grosso.

BRAZIL. Marto Grosso: corolla white, tree 7-8 m. tall, Corumbé, July
20, 1908, Hoehne 100 (BD); shrub 1 m. tall, corolla white, Sao Luiz de Caceres,
Jan. 1909, Hoehne 1120 (BD); indefinite, July 1892, Kuntze (NY, trPx; BD,
ISOTYPE),

This species is quickly distinguished from all other members of its
section by its very large white or yellowish corollas and by its com-
monly rather large leaves. It appears to be a loosely branched bush
or small tree 1-3 m. tall. A small proportion of the specimens ex-
amined have the twigs tunneled apparently by ants. The common
form of the species has leaves of heavy texture that are prominently
veined and hence conspicuously reticulate beneath. The lower surface
's usually velvety. Under the name, var. glabrifolia, I have followed
Kuntze in recognizing a peculiar form of the species from Matto
Grosso. This plant has leaves that are usually smaller than is common
in the usual form of the species and are glabrescent and very much
less evidently reticulate beneath.

4. C. glabrata (Mart.) A. DC. Prodr. ix. 473 (1845); Fresen. in
Mart. Fl. Bras. viii. pt. 1, 6 (1857). Gerascanthus glabrata (Herb. F|
Bras. pg. 167, sub no. 269) Flora Regenb. xxi. Bd. 2, Beibl. 87 (1838).
Lithocardium glabratum Kuntze, Rev. Gen. ii. 977 (1891). C. longi-
beda Mez, Bot. Jahrb. xii. 550 (1890). L. longipedum Kuntze, |. e.

12 JOHNSTON

C. glabrata, var. longipeda Chodat, Bull. Soc. Bot. Genéve, ser. 2, xil.
212 (1921). L. longipedum, var. glabrifolium Kuntze, Rev. Gen. ii.
pt. 2, 206 (1898). L. longipedum, var. eriophyllum Kuntze, |. ¢. 206.
C. glabrata, var. amambayénsis Chodat, 1. ec. C. glabrata, var. orbicu-
laris Chodat & Vischer in Chodat, I. ec. fig. 305. C. longituba Chodat
& Vischer in Chodat, |. ¢. 213, fig. 310 and 313.

In Brazil from Piauhy and central Minas Geraes southwestward
through Goyaz and Matto Grosso to southern Paraguay.

ZIL. Pravuny: tree 12 m. tall, corolla white, common on banks of the
Gurgoa, Aug. 1839, Gardner 2686 (NY, K, , DC, DL, BB). Govyaz: betw.
Meiaponte and Caisara, Oct. 23, 1827, Burchell 6290 (K); small tree 6-9 m.
tall, corolla white, throat darkish, base of lobes with dark lines, near Chapel of

anto Amaro, Goyaz, July 8, 1828, Burchell 7323 (G, K); tree 6-9 m. tal,
Goyaz, June 15, 1828, Burchell 7183 (K); tree 4-6 m. tall, Goyaz, June 1828,
Burchell 7186 (1K)
Glaziou 21782 (N

GUAY: a tree wi it of Populus alba, solitary or in groups on
campo, Colonia Risso near Rio Apa, Oct. 20, 1893, Malme, 1092b (RS);
Sierra de Sr cigefe Sept., Rojas in Hassler 10600 (BB, TYPE of v

no, a
1915, Hass i . 1914, Char
ae iy role: eer Paraguarf, eg ig1d, Chedal 368 (5B); indefinite,
A well marked species, which is restricted to our area, where it #8
characterized in its section by its simple trichomes and its ovate
leaves that are glabrous and shiny above. It appears to be prevailingly
deciduous and is said to develop, at the beginning of the growing
season, its inflorescence before its leaves. There is a striking variation
in the amount of pubescence developed on the leaves. In some forms
the mature foliage is glabrous or only obscurely villulose below, ™
others it is pale and satiny with very abundant closely appressed silky
airs. The type of L. longipedum var. eriophyllum, C. glabrata var
orbicularis and C. longituba all represent this form with the leaves

STUDIES IN THE BORAGINACEAE 13

silky-pubescent beneath. As this is a very striking phase of the
species it may be convenient to have a name for it; it may appropri-
ately be called C. glabrata, forma eriophylla (Kuntze), comb. nov.
This form appears to be restricted to the southern parts of the range
of the species where it appears to grow with the typical glabrate
phase. Malme has noted on one of his labels that the trees of this
form along the Rio Apa recall those of Populus alba. Cordia longituba
seems to have been based upon a very mature specimen of the typical
form in which the maturing fruit has pushed the marcescent corolla out
of the calyx and hence exposed the tube. Cordia glabrata shows
tendencies towards myrmecophily, stems of specimens frequently
having been tunneled and inhabited by ants. As in C. insignis,
however, these stems show only a slight increase over their normal
diameter and are never distorted by conspicuous irregular swellings
as is common in C. alliodora.

5. C. alliodora (R. & P.) Cham. ex DC. Prodr. ix. 472 (1845); Urban,
Symb. Ant. viii. 574 (1921). Cerdana alliodora R. &. P. Fl. Peru. ii.
47, t. 184 (1799). Lithocardium alliodorum Kuntze, Rev. Gen. ii. 976
(1891). L. Gerascanthus, var. alliodorum Kuntze, |. ¢. iii. pt. 2, 206
(1898). Cordia Cerdana R. & S. Syst. iv. 467 (1819). Cordia cujaben-
sis Manso & Lhotzky ex Cham. Linnaea viii. 121 (1833); DC. Prodr.
ix. 473 (1845). Cerdana cujabensis Manso ex DC. |. c. 473, in synony-
my. Cordia velutina Mart. (Herb. Fl. Bras. pg. 165, no. 268) Flora
Regenb. xxi. Bd. 2, Beibl. 85 (1838). Gerascanthus velutina Mart. ex
Fresen. in Mart. Fl. Bras. viii. pt. 1, 5 (1857). ? @. vulgaris Mart.
(Herb. Fl. Bras. pg. 166, sub no. 268) Flora Regenb. |. c. 86, in dis-
cussion. Cordia Gerascanthus, var. ? subcanescens DC. 1. ¢. 472. Cordia
Gerascanthus, var. domingensis Cham. 1. c. 121; DC. 1. ¢. 472.
Gerascanthus, var. domingense Kuntze, Rev. Gen. iii. pt. 2, 206 (1898).
Cordia alliodora, var. glabra A. DC. 1. ¢. 472. ? Cordia Hartwissiana
Regel, Gartenflora vi. 342 (1857). ? L. Hartwigsiana Kuntze, Rev.
Gen. ii. 977 (1891). Cordia Rusbyi Chodat, Bull. Soc. Bot. Genéve,
Ser. 2, xii. 187 and 198 (1921), not Britt. (1896). Cordia Gerascanthus,
'. martinicensis Chodat, |. c. 210. Cordia Gerascanthus, f. micrantha
Chodat, l. ce. 210. Cordia consanguinea Klotzsch ex Chodat, 1. ¢. 211,
incidental mention. Cordia alliodora, var. boliviana Chodat & Vischer
in Chodat, |. c.211. Cordia andina Chodat, |. ¢. 211. Cordia Chamis-
soniana, var. complicata [R. & P.] Chodat, l. c. 215. Cerdania com-
plicata R. & P. ex Chodat, |. c. 215, in synonymy. Cordia Goudoti
Chodat, 1. c. 215. Cordia macrantha Chodat, |. c. 215. ? Cerdania
eraltata R. & P. ex Chodat, |. c. 216, incidental mention. Cordia Ger-
ascanthus of Jacq. Sel. Stirp. Amer. 43, t. 175, f. 16 (1763) and many

14 JOHNSTON

subsequent authors, notably Griseb. Fl. W. Indies 478 (1861) and Cho-
dat, |. c. 209; not Cordia Gerascanthus L. Syst. ed. 10, 936 (1759) nor
Gerascanthus P. Browne, Hist. Jamaica 170, t. 29, f. 3 (1756)!! ef.
Johnston, Contr. Gray Herb. Ixxiii. 77-78 (1924).

From Mexico, Central America and the West Indies south into
Venezuela and Colombia and then southward along the Andes to
Amazonian Bolivia from whence it extends eastward into south-
western Brazil (Matto Grosso).

RAZIL. Marro Grosso: Cuyaba, June 1832, Manso & Lhotzky 13 (BD,
TYPE of C. cuj is; DC, 1soryPE); Morro do Ernesto near Cuyaba, Martius
Herb. Bras. 268 (NY, K, DC, DL); Cuyabé, June 8, 1902, Malme 1651 (RS);
Santa Anna da Chapada, July 27, 1902, Malme 2110 (RS).

This widely distributed plant just enters our area, being known there
only from Matto Grosso where it replaces the closely related C.
trichotoma. Although C. alliodora and C. trichotoma are very closely
related, it is to be noted that the two differ rather strikingly in their
relation with ants. Cordia alliodora is a well-known myrmecophyte,
its stems, usually just below the inflorescence, developing conspicuous
irregular globose or turbinate gall-like swellings that become ant-
domatia. Material of C. trichotoma only very rarely shows any sug-
gestion of such developments. In the very rare cases where the stems
are occupied by ants they are merely tunneled and although slightly
increasing in thickness over the normal they are never conspicuously
deformed as in C. alliodora. The complete and rather staggering list
of synonyms of C. alliodora which I have given above might give the
impression that the species, as here defined, is an uncommonly variable
one. There is some variation in the size of the flowers and some it
the amount of pubescence on the leaves, but less than in the related
C. trichotoma. Chodat, |. ¢., the only person who has tried to split up
the species, has failed most evidently in defining clearly, not only his
numerous varieties, but even his proposed specific segregates. I am
of the opinion that several varieties or forms would account for the
principal phases of this species. Chodat, l. c., having insisted that the
name Cordia Gerascanthus is the correct name for this species, it may
be noted in substantiation of my arguments made several years 28%
i C., that I have recently examined the specimens in the Linnaean Her-
barium which are the basis of Cordia Gerascanthus L. These speck
mens were collected by Browne in Jamaica and represent the plant
treated by him as Gerascanthus. There can be no doubt whatsoeve!
that they represent the plant described as C. gerascanthoides HBE!
and hence one quite distinct from the one I am here calling C. alliodor@.

C. trichotoma (Vell.) Arrab. ex Steud. Nom. ed. 2, 419 (1840).

STUDIES IN THE BORAGINACEAE 15

Cordiada trichotoma Vell. Fl. Flum. 98 (1825) and Icones ii. t. 156
(1827). Cordia frondosa Schott ex Spreng. Syst. iv. pt. 2, 403 (1827).
Cordia tomentosa Cham. Linnaea iv. 472 (1829), not R. & S. (1819).
Gerascanthus excelsa Mart. (Herb. Fl. Bras. pg. 166, sub no. 269)
Flora Regenb. xxi. Bd. 2, Beibl. 86 (1838). Cordia excelsa A. DC.
Prodr. ix. 473. (1845). Lithocardium excelsum Kuntze, Rev. Gen. ii.
977 (1891). Cordia Chamissoniana Steud. |. c. 417, not Don (1837).
L. Chamissonianum Kuntze, |. c. 976. Cordia hypoleuca DC. 1. e. 472.
L. hypoleucum Kuntze, |. ce. 977. Cordia alliodora, var. tomentosa
A. DC. 1. c. 472; Fresen. in Mart. FI. Bras. viii. pt. 1, 4 (1857). Cordia
asterophora Mart. ex. Fresen. |. c. 5. L. asterophorum Kuntze, |. e.
976. L. Gerascanthus, var. puberulum Kuntze, Rev. Gen. iii. 206
(1898). Cordia ei cenasitiuet: f. puberula Kuntze ex Fries, Ark. Bot.
vi. no. 11, 20 (1 906). Cordia hypoleuca, f. minor Chodat, Bull. Herb.
Boiss. ser. 2, ii. 815 (1902). Cordia Hassleriana Chodat, 1. ¢. 815.
Cordia oometnanessieniet var. Gardner, var. Blancheti, var. Marti, var.
nemorensis and var. Aemilii Chodat, Bull. Soc. Bot. Genéve, ser. 2,
xii. 214 (1921). Cordia cujabensis, var. glabrescens Hass. ex Chodat,

. ¢. 214, in synonymy.

Eastern Brazil from Cearé to Rio Grande do Sul and extending
West through Paraguay to northwestern Argentina and adjacent
Bolivia. Doubtfully occurring also in northernmost Brazil near the
Guianan border.

BRAZIL. Amazonas: tree 5-12 m. tall, Mniam, Surum o Branco,
Nov. 1909, Ule 8290 (K, BD, DL). nan! Crato, Sept. 1838, Gaviner 1781
ej Crato, 1838, Gardner 1780 (BB, TYPE of C. Cham v. Ga rdneri; G, lel
j , . Rro Granve po Norte: bank of Potengy River n tal,
1914, Dawe 13 (K). EEENaupuco: Pertimatine, July 1925, Pickel 1068
(BD). Banta: j woods, Catingas, Martius (BD, 1sotyPx of C. asterophora);
Sara Jacobina, Blanchet 2580. (BB, tyPx of C. Cham. v. Blanchett; NY, , BM,
504) DC, DL). Minas Gerass: tree 9 m. tall, Rio Claro, June 1840, Gardner
7.40 (NY, K, DL, RS); tree 7.5 m. tall, betw. Rio Claro and Sao
sa 1840, Gardner 5040 (BM); Oar Preto, Casaretto 2915 ( Soe

arandahy and Queluz, Glaziou’ 11282 ( Kk, Bb), 14143 (K, BM Bb, DL)
an ete OR K BD, DL); ‘aacons, ‘Glaziou 14145 (NY, K, BD) and
i eg 7

E);
wior Tee Wise K, BM Ss Fon is Campos, Herb. Mendonga 1 188
(BD); (San an Jae do i 234 Apri Ns a Hoehne & Gehrt, Mus. Paul. i 17642
21,

Guest, 1874, Mosén 2006 (Rs); Campinas, June 20, 1875, Mosén 3905 (RS);

16 JOHNSTON

1 )
and 11743 (RS). Rio Granpe po Sux: Sao Leopoldo, Feb. 17, 1902, Malme
1374 (RS); Ijuhy, April 4, 1893, Lindman (RS); Victoria on Rio St. Leopold,
Tweedie 28 (K); near Porto Alegre, For 426 (K, BM). Inperinite: Brazil,
ex herb. Spreng (BD, authentic C. frondosa); Bras. aequin. Sellow 1571 and 1578
(BD, type of C. tomentosa); Brazil, Sellow (K, BM, BD, DC, DL); Brazil,
Pohl, Herb. Mus. Vind. 1738 (K, BD); Brazil, Martius Herb. Bras. 1067 (BB,
TYPE of C. Cham. v. Martii; G, NY, K. , BD, , , ISOTYPES) and 486
(NY, K, BM, BD, DC, DL, tsorypes of G. excelsa); Brazil, Glaziou 12087 (K).
PARAGUAY: Alto Parana, Fieberg 6164 (G, K, BM, BD); Sierra de Amam-
bay, May, Hassler 11225 (BB, Typr of C. Cham. v. Aemilii; D,
11235 (G, K, BM, BD, DL, Hass, BB); Itacurubi to Union, June 1898,
; 3

. hypol. £. minor and C. .Y.
nemorensis; K, isoTyPES); tree 10-15 m., Villa Rica, Feb. 1829, Jérgensen 3469
(G); tree 10-20 m. Villa Rica, July 1929, Jérgensen 3737 (G).

ARGENTINA. Muisionus: near Iguazi, May 5, 1902, Thays 11 (K);

erto Aguirre, July 1914, Curran 672 (G, US, BM); Puerto Aguirre, Jan.
1922, Molfino (G); Puerto Leon, July 1914, Curran 701 (G, US, BM); San
Ana, May 1910, Rodriguez 10318 (G); Misiones, April 1888, Niederlein A61
(BD). Saxra: Tabacal near Oran, July 1873, } our
TYPE of L. Gerasc. v. um; DL, RS . Jusuy: Ledesma,
July 4, 1911, Lillo, Stuckert Herb. 22345 (DL); Ledesma, July 5, 1911, Lallo

3 ; tree ca. 1 i ae Laguna de la Brea, June 13, 1901,

Cordia trichotoma is very closely related to C. alliodora, differing
only in its larger flowers, usually more copious and conspicuous pUu-
bescence and more eastern and southern distribution. Although these
differences are not strong ones and even though the two species do
show a slight tendency to intergrade I am inclined to believe that
their acceptance as species is perhaps desirable since the division
is a natural one and seems to be in the vast majority of instances 4
practicable one also.

Cordia trichotoma exhibits some variation in the size of the flowers
but its most conspicuous variations are those involving the characte?,
abundance, distribution and to some extent the color of the stellate-
pubescence particularly on the leaves. The extremes of this variation
are very striking and quite different at first sight. They have,
accordingly, attracted attention and have been largely the cause of
the imposing array of synonyms that I have listed above. These
extreme variations of pubescence, however, are thoroughly connected
by an abundance of intermediate forms and, what I consider very

STUDIES IN THE BORAGINACEAE 17

significant, show very little if any tendency to be geographically cor-
related. Since these extreme forms have attracted and no doubt will
continue to attract more attention than they merit I have defined
below, for convenience and in the hope that they may forestall further
multiplication of synonyms, a group of four forms into which all the
material of this species may be placed by those who feel the need of a
detailed classification of such fluctuating forms.

Forma typica. Lower surface of the leaves felty with a very dense
close usually yellowish indument of abundant minute stellate hairs;
upper leaf-surface glabrate or with minute sparse inconspicuous
stellate hairs.

Forma Blancheti (Chodat), comb. nov. Lower surface of the leaves
velvety with a loose evident grayish or grayish-yellow tomentum of
abundant stellate hairs! upper surface of leaves sparsely stellate or
glabrate—C. Chamissoniana, var. Blancheti Chodat, l. ¢.

Forma puberula (Kuntze), comb. nov. Lower surface of the leaves
very sparsely clothed with stellate hairs, frequently almost glabrescent;
upper surface usually glabrescent.—L. Gerascanthus, var. puberulum
Kuntze, 1. c.

Forma tomentosa (Cham.) comb. nov. Upper as well as lower
surface of the leaves clothed with a loose very abundant fulvescent
tomentum of evident stellate hairs.—C. tomentosa Cham., |. c.

he type of C. trichotoma came from Brazil in “silvis maritimis
Regii Praedii Sanctae Crusis” at the western end of the Federal
District. Unfortunately no mention is made in the original descrip-
tion of the pubescence and so consequently it is necessary to place
the species entirely upon what can be derived from our knowledge
of its type-locality and from a study of the plate given by Vellozo.
The form of our species that has been collected by others in the Federal
District is almost exclusively that which I have indicated as the forma
‘ypica. The plate given by Vellozo has no pubescence indicated, but
uch as the artist in his various illustrations has indicated only

the most conspicuous types of pubescence I believe that we may as-
sume that the plate of the present species of Cordia was probably
ased upon the form with very fine dense felty yellowish indument,
that is to say upon the predominating form in the Federal District,
the forma typica. I have seen this form from Paraguay, northeastern
ntina and from the Brazilian states south of Rio Janeiro. Cordia

; is a clear synonym. Cordia excelsa is typical except that it
wea in having the calyx only weakly rather than strongly

The forma Blancheti is probably the most widely distributed form

18 JOHNSTON

of the species occurring throughout most of its range. It differs from
typica only in its velvety, much looser.indument of much coarser
trichomes. The lower leaf-surface is obviously hairy and not as in the
f. typica so densely and finely pubescent that extreme forms might,
with a hasty glance, be thought to be glabrous with a waxy indument.
To this common form I would refer C. Chamissoniana vars. Gardner,
Blancheti, nemorensis, Aemilii and C. cujabensis var. glabrescens.
The forma Blancheti is slightly less distinct than the f. typica, being
connected by very numerous intergrades with both f. tomentosa and
f. puberula.

The forma puberula is simply the extreme phase of f. Blancheti in
which the leaves become glabrescent or are only sparsely pubescent
beneath. It occurs sporadically in various parts of the range of the
species and seems to be less Common than Blancheti. It has synonyms
in C. asterophora and C. Hassleriana.

The forma tomentosa is another extreme phase of f. Blancheti, in
this case one in which the very copious velvety tomentum is form
on both surfaces of the leaves rather than on merely the lower one.
It appears to be uncommon and seems to occur only in the coastal
states of Brazil north of Rio Janeiro. It has synonyms in C. tomentosa,
C. Chamissoniana and C. Chamissoniana var. Martii.

Section Varronia [P. Browne] Don, Gen. Syst. iv. 382 (1837).
Varronia [P. Browne, Hist. Jamaica 172, t. 13, f. 2 (1756)] L. Syst.
ed. 10, 916 (1754) and Gen. ed. 6, 102 (1764), type-species ? C. globost
HBK. Cordia-Varroniae HBK. Nov. Gen. et Sp. iii. 71 (1818);
& S. Syst. iv. 458 (1819). Cordia subgenus Varronia Schlecht. &
Cham. Linnaea y. 115 (1830). Cordia § Sebestena subsect. Varromu
Endl. Gen. 644 (1838). Not Varronia DC. nor Cordia § Varrona
Giirke. Cordia-Dasycephalae HBK. |. c. 76 (1818), type-species,
C. globosa HBK. Cordia § Varronia subsect. Dasycephalae Don, lL.
383 (1837). Cordia § Myza subsect. Dasycephalae Giirke in E. & P.
Nat. Pflanzenf. iv. Abt. 3a, 84 (1893). Cordiopsis Desv. ex Hamilton

Prodr. Pl. Ind. Oce. 23 (1825); Urban, Symbol. Ant. viii. 580 (1921),

type-species, C. mirabilloides Ham. Cordia § Cordiopsis A. DC. Prodt-
1x. 498 (1845). Catonia Raf. Fl. Tellur. ii. 36 (1836), type-speci®

C. lantanoides Raf. Piloisia Raf. Sylva Tellur. 43 (1838), type-specie
P. globosa (Jacq.) Raf. Topiaris Raf. Sylva Tellur. 43 (1838), tyP® |

species, 7’. geniculata Raf

The section Varronia is confined to America, where its approximate —

ly 60 species are widely dispersed in the tropics. The greatest diversitY
of form, however, is found in Mexico and in Brazil. It contains

shrubs or suffrutescent plants. The group is polymorphous bt"

STUDIES IN THE BORAGINACEAE 19

evidently a natural one and because of the large number of species,
their variability and their indefiniteness probably the most perplexing
and poorly understood in the entire genus. The most characteristic
feature of the species of this section is the development of dense,
distinctly spicate, globose or glomerate flower-clusters. Although
the vast majority of the species have the numerous flowers densely
congested, a few show not unexpected departures from this habit.
The widely distributed C. corymbosa commonly has the rather numer-
ous flowers in definite glomerules but frequently these loosen up to
form flat cymes several centimeters in diameter. Another extreme
is found in the few species related to C. Greggi Torr. of Mexico. Here
the flowers are very few and form small loose cymes that occasionally
are reduced to a single flower. Morphologically the characteristic
spikes and heads of the section are very interesting being what Gray
has termed “reversed or determinate.”’ The flowering in them begins
at the apex and proceeds downward towards the base thus showing
their cymose origin. The calyx of Varronia is cupulate, crateriform
or somewhat globose and is always clearly accrescent at maturity and
More or less sheaths the fruit. It is never explanate. In some forms
it becomes a deep cup closely investing the lower half or two thirds of
the mature fruit, in others it becomes slightly inflated, completely
Surrounds the fruit and eventifally seems to fall away with it en-

ut with occasional exception is characterized by its subentire or
very shallowly lobed limb. The corolla-lobes are not particularly
Conspicuous, usually being short, extremely broad and rounded or
truncate. The fruit is either globose or ovoid and appears to be
always small, becoming only 3-6 mm. in length. It is generally 1-2-
— although it is not uncommon for plants to develop all 4 of its
ules.

Key To THE SPECIEs.

ge oree, aie cm. long, limb many times as broad as the tube,
and thro t : . “ :
Tips of caly Bettas evidently differentiated ee
mm. lo .

dA
: ot 5

20 JOHNSTON

Leaves with distinct slender petioles 1-2 cm. long; corolla very
large 4-5.5 cm. long; loosely branched shrub becoming 25

all; of eer PABRM Cel Ok ee een ba tis 7. C. grandiflora.

Leaves sessile or the cuneate base rarely drawn out into a
winged petiole 1-5 mm. long; corolla smaller, 2—-2.5 em.
long; undershrub with several ine or sparsely branched

stems 1-5 dm. tall; dry campos............ 8. C. paucidentata.

oN of pices ne seeenersieced or ‘with pi rae inconspicuous
es ca. 1 m
Calyx and peduncles labret. conspicuously studded with
mrmbe Sines Blinds 6s. oss ik ok C. longifolia.
Calyx ee peduncles evidently strigose or hirsute, sparingly
an
eae lea surface oe or. _ pin a few scattered
nspicuo ale disks of miner-
alived mi leaves lanceolate, re ena aliens crenate or
serrate; corolla-lobes ovate...............-.. 10. C. poliophylla.
Upper leaf-surfaces a hsdlanthy strigose, conspicuous disks
of pale mineralized cells absent; leaves ovate to lanceo-
-_ Bey 0 trae serrate; sero shee at broad
on, Seatfeiroulat. 6 eS i leucocephala.
Corolla at gpaen 1 em. long or less, limb only 2-4 times as
veo d as the abe throat and tube seldom slacks differenti.

Inflorescence clearly spicate, the spikes 3-20 times as long as

Leaves muriculate or verrucose above, not hairy, ane re
i ee ae . verbenaced.

es hairy abov
Leaves 5 em "broad, usually lanceolate or lance-ovate;

Paciancles and ‘leafy stems with appressed, usually curled a
Gs ca 4 os Bl etis yuck a ui yer: 13. C. chacoensts.

eee i ee ee Bee ae ee oes ee le be we Wee ee

ently becomi ming scandent.............. ib. vi. cali
Taflorepsiines cae or glomerate or y becoming som
t atin never more than twice as long as br
sapos oon or rarely and forming . _
Si ate Ue i tuk ae Sov asces Jee . corymbosa.

Upper ‘net en A feaoeot: Loan the oe rves)
abrous, ughen nly the tubercles res eaibe
mene pina undeveloped rs; calyx-lobes ithe
monly tip ear a d ; plants of the
Amaz Be teaite & ster sa eee ; ass 7 C. buddleyoides-
Upper leaf-surfaces hai alyx-lo

bes ti vay
short stout ap rendages i . all; plant tipped by (except Ate
in Goyaz) east a and south of the Amazon Basin. 18. C. azillans.
Peduncles exclusively. pect or if not then opposite the
leaves or internodal, eu
Freely coger. g shrubs 1— oa, tall 2 pene yeh —

posi mo
leaves with slender evident petioles 5-10 mm. long.
Lower leaf-surface conspicuously floccose, white .19. - Leucomalla

STUDIES IN THE BORAGINACEAE 21

Lower re not at all floccose, green.
peduncles very sparingly hairy, covered
with hy ais ely numerous sessile or stalked
PO sis hrcecoig hye heh agin vin tamale era C. corchorifolia.
Gave ane Sci aeantie conspicuously hairy, glands
sparse or absen

Leaves very 29 0 somewhat velvety, at

with stalked glands; flower-heads 1.5-2

Ss

Jaga sessile if present; flower-h
n diameter; calyx-lobes with much
hia 3 conspiewous appendages only 0. co

ntly with simple or once-branched leafy stems
spri Seine rom a woody caudex; peduncles termin:
eres ‘sieeve or with i ianeniances petioles up to 5
Calysclobes Miah very conspicuous apical linear ap-
5 mm. long; leaves vig rs or ovate-
oblong, broadest below the middle, lower ea

to
Stems sca hispid with slender spreading hairs;
lea SB eink heoed. 3. hese 24. C. —
Stems tae, clothed eg apa or incurved hairs
. bro

sp
vein-areoles; flower-heads very dense an
MNUULY TULUM. fo56 os ion Vig Sees ee 25. C. truncata,
Leaves elongate, narrowly oblanceolate, the sec-
ondary veining very obscurely elo

ni
flower-heads less dense, fewer-flow ie
RVG? TAOS. os os ac hs cc eae 6. C. guaranitica.

7. C. grandiflora (Desv.) HBK. Nov. Gen. et Sp. iii. 77 (1818);
Fresen. in Mart. Fl. Bras. viii. pt. i, 21, t. 3, f. 2 (1857). Varronia
grandiflora Desy. Jour. de Bot. i. 273 (1808). Lithocardium grandi-
florum Kuntze, Rev. Gen. ii. 977 (1891). V. lantanoides Willd. ex
Cham. Linn inaea iv, 492 (1829), in synonymy. C. rufa Klotzsch ex
ae, Fauna u. Fl. Brit. Guiana 960 (1848), nomen.

wn only from central Venezuela and eastward along the Orinoco

to British Guiana and from the following single collection on the lower
Amazon j in Brazil

22 JOHNSTON

BRAZIL. Par: shrub ca. 2.5 m. tall, growing into the water, corolla
white, Rio Trombetas, vicinity of Obidos, Dec. 1849, Spruce 515 (G, K, BM,
BD, DL, BB).

A very distinct species and the one having the largest flowers in its
section. The plant must be a rare one for despite its conspicuous
flowers it has been collected in only a few scattered localities, in all
cases apparently in low wet ground near rivers. The corollas, sugges-
tive of those of Ipomoea, have a tube 6-9 mm. long and 2-2.5 mm.
thick, and a very well developed throat 2-2.5 em. long and 11-14 mm.
thick at the summit; the limb is 3-4 cm. in diameter and has broad
shallow ascending lobes. In our Brazilian material the throat is
narrowly obconic and is gradually expanded upward from the tube.
In the specimens from British Guiana and Venezuela, however, the
throat is campanulate and very abruptly expanded at the base. In
other characters our plant is remarkably like these from farther north.
The species, in fact, is very stable and shows practically no other
noteworthy variations. The type of C. grandiflora was obtained in
central Venezuela.

8. C. paucidentata Fresen. in Mart. Fl. Bras. viii. pt. 1, 25 (1857).
C. sessilifolia, var. macrantha Cham. Linnaea viii. 130 (1833). Lith-
ocardium macranthum Kuntze, Rev. Gen. ii. 976 (1891). C. uruguaya
Arech. Anal. Mus. Nac. Montevideo, ser. 1, iv. 16, t. 7 (1902) and L.¢.
vii (Fl. Urug.) 170 (1911). C. paucidentata, var. subulata Chodat &
Hass. Bull. Herb. Boiss. ser. 2, v. 481 (1905). C. paucidentata, vat.
valenzuelensis Chodat, Bull. Soc. Bot. Gendve ser. 2, xii. 217 (1921).

rom northeastern Paraguay and from Parana in Brazil southward
into northern Uruguay.
GUAY: suffrutex 4-5 dm. tall, Punta Pardé, Dec., Rojas in Hassler

PARA
9903 (BM, BD, Hass); suffrutex 3-5 dm. tall, Yerbales, Sierra de Maracay4,
: D SOTYPES);

pucay, Dec., Hassler 1620 (Hass); on ant-nests in camp®
nta Maria, Feb. 2, 1896, Anisits 2692 (RS); Cerro del Gobierno, Dee.

ARGENTINA. Mistonzs: San I cio, Oct. 23, 1913, Quiroga (G); betw-
Santa Ana and Loreto, Feb. 14, 1883, Niederletn (BD); oon Feb. 1922,
Molfino (G); Posadas, Feb. 1922, Molfino (G); Posadas, Jan. 1922, Molfino (G);
Posadas, Dec. 1, 1909, Rodriguez 10054 (G).

ZIL. Paran&: shrubby campo, Rio Capinzal, Feb. 29, 1916, Dusé”
17885 (RS). Rio Granpe po Sut: grassy campo, Cruz Alto, Jan. 20, 1903,
Malme 1163 (RS). Inperinrrn: Bras. merid. Sellow 8096 and di728. (BD:
pov neg material of C. paucidentata and C. sessilifolia v. macrantha); Bras.

URUGUAY. Anrricas: Catalan, Nov. 27, 1927, Herter 999 (G, RS).

STUDIES IN THE BORAGINACEAE 25

A well marked, very distinct species which appears to frequent
open places, particularly on dry campos. Among the South American
species which have this habit C. paucidentata is quickly distinguished
by its conspicuous corollas. The plant may be a shrubby perennial
with a tuft of subsimple stems 1.5-3 dm. tall or may become a shrub
with only a few very sparsely and strictly branched stems 5 or rarely
as much as 8 dm. tall. The corollas are funnelform, 2-2.5 em. long
and nearly as broad. The broad limb is only very shallowly lobed.
Thave seen a photograph of the type of C. uruguaya, this was obtained
for me several years ago through the kindness of Dr. Herter of Monte-
video. The species is clearly conspecific with C. paucidentata, the
type in fact being remarkably like the type-collection of C. sessilifolia,
var. macrantha Cham.; the nomenclatorial basis of the name C
paucidentata Fresen. I am unable to distinguish the two varieties -
proposed by Chodat.

9. C. longifolia A. DC. Prodr. ix. 495 (1845). Lithocardium longi-
folium Kuntze, Rev. Gen. ii. 977 (1891).

hown only from the eastern part of the state of Bahia.

BRAZIL. Banta: Muritiba, 1847, Blanchet (DC); Feira Santa Anna,
Blanchet (DC); Bahia, Blanchet 268 (DL); corolla white, forests, Bahia, 1834,
nchet 1739 , TYPE; BM BB, isotypes); ‘‘Ilhios, Moricand” 1739
Mes hn sion: y. Bahia, 1840, Blanchet 3190A (DL); Bahia, 1840,

: (DL); Muritiba, Blanchet 3986 (K, DL, BB); corolla white,
Bahia, Blanchet 23 (BM); corolla white, shrub in shady woods, Feira Santa

This extremely well marked species is known only from Bahia
Where it appears to have been collected only by Blanchet. The
flowers are borne in a short stout not very densely flowered spike
terminating a very slender elongate distinctly supra-axillary peduncle.
Although the calyces and peduncle are devoid of any hairs they have
4 pruinose covering of abundant minute sessile glands. The calyces are
funnel form, the basal part being distinctly contracted into a definite
tube almost 1 mm. thick and nearly 2 mm. long. Above this tube the
calyx expands rapidly forming an open throat. The calyx-lobes are
triangular, 1-1.5 mm. long, and have unappendaged tips. After the
fall of the corolla the calyx-lobes are more or less inflexed and conse-
quently enclose the developing fruit. The fruit, which is ovoid-tur-

» almost 5 mm. long and 3 mm. thick and conspicuously lacunose,
develops in the tube of the calyx and consequently distorts and appar-
ae with frequency even ruptures it as it increases in size in maturing.
it developed corolla, unknown to De Candolle, is white and fun-
th orm having a very slender tube, 4-5 mm. long and 1 mm. thick,

“t much surpasses the calyx. The wide corolla-throat is abruptly

24 JOHNSTON

expanded from the tube. The limb of the corolla is nearly 2 em.
broad and, by sinus 4-5 mm. deep, is cut into five broad ascending
emarginate lobes. The leaves are lanceolate, somewhat firm in tex-
ture, with the lower surface noticeably pale. As with the inflorescence
it is pruinose with minute sessile glands. The upper surface is oc-
casionally provided with pustules formed by the thickening and
mineralization of disks of epidermal cells. Although similar to the
hair-bases formed on the leaves of other species, these disks do not
produce hairs in C. longifolia.

10. C. poliophylla Fresen. in Mart. Fl. Bras. viii. pt. 1, 26 (1857).
Lithocardium poliophylla Kuntze, Rev. Gen. ii. 977 (1891).

Known only from southeastern Bahia or adjacent Espirito Santo.

BRAZIL. Esprrriro Santo or Banta: betw. Vittoria and Bahia, Sellow

- (BD, coryrr). InpEFrnire: no locality given, Blanchet (DL).

I have not seen the type of this species which was collected by
Martius at an undesignated locality “in silvarum Oceano conter-
minarum margine,” but I have studied a cotype at Berlin that was
collected by Sellow somewhere on the coast in southern Bahia or
northern Espirito Santo. My study of the original description and
of the cotype, however, makes me feel confident in referring to
species a collection made in Bahia by Blanchet. This came from the
Moricand Herbarium, now a part of the Delessert Herbarium at
Geneva, and is labeled as having been received under and perhaps bed
a part of Blanchet’s no. 3/90, a number which represents C. longifolia.
Cordia poliophylla is most closely related to C. longifolia but seems t0
be quickly distinguished by its abundant appressed pubescence and
particularly by its pubescent globose and stouter less elongate calyces-
Blanchet’s specimen has corollas with the tube ca. 8 mm. long, almost
2 mm. thick and much surpassing the calyx. The throat is abruptly
expanded, broad and open. The lobes are ovate, obtuse and 5-6 mm.
long. Although the corolla is glabrous, outside it is sprinkled with
short-stipitate yellowish glands. .

Al. C. leucocephala Moricand, Pl. Nouv. d’Amér. 148, t. 88 (1846):
Tithocardium leucocephalum Kuntze, Rev. Gen. ii. 977 (1891). C.
leucocalyx Fresen. in Mart. Fl. Bras. viii. pt. 1, 22, t. 7 (1857). L.
leucocalye Kuntze, |. c. 977. C. affinis Fresen. 1. c. 22, t. 9, f. 10:
L. affine Kuntze, |. ¢. 976. C. striata Fresen. |. c. 23. L. striatum
Kuntze, l.¢. 977. C. nivea Fresen. |. c. 26. L. niveum Kuntze, |. ¢. 97:

Ranging from northwestern Cearé southward through easter
Piauhy and central and east-central Bahia to northern Minas Geraes:

BRAZIL. Cxark: Jatob4, March 30, 1910. Loe 2 (RS), Pravae:
shrub 18 dm. tall, corolla white, near Boa Beperanee’ Feb’ 1839, Gardner 226°

STUDIES IN THE BORAGINACEAE 25

(K, BM). Bauza: corolla white, Serra do Assurud towards the Rio San Fran-
cisco, 1838, Blanchet 2854 (N Y, K, BM, BD, DL , IsOTYPIC of C. nivea ?);

corolla yellow, Pougo d’Areia near Jacobina Blanchet 3880 (DL, Type of C.

ocephala; BM, BB, tsoryrEs); Pouco d’Areia, Jan. 1844, Blanchet 3879
(DL); near Cachoeira, Feb. 1819. Martius (BD, tsoryrs of C. affinis). Minas
ERAES: near Formigas, July 1818, Martius (BD, coryrE of C. affinis).
In founding his species Moricand cited two collections of Blanchet,
nos. 2854 and 3880, giving both of them as from “Serra Acurua.”
Blanchet 2854 is in bud only and doubtlessly came from the Serra
Assurué. The other collection, no. 3880, however, is in ower. There
can be no doubt that this latter collection is not only the basis of
Moricand’s illustration but also of the important parts of his descrip-
tion. Consequently I have accepted Blanchet 3880 as the type-collec-
tion, although in doing so I question strongly that the collection came
from the locality given by Moricand. The locality, Serra Acurua,
written on the label of the type in a script different from those which
are familiar on Blanchet’s labels, does not agree with that given on
the labels of this collection in the De Candolle or Boissier herbaria or
at the British Museum. In these three herbaria the specimens of
Blanchet 3880 are provided with the data which I have used above
In citing collections examined. Cordia nivea is a clear synonym of
C. cocephala having no doubt been based upon another specimen
of Blanchet 3880. I have seen the two basic collections of C. affinis and
80 Teel confident in referring that species also to the synonymy of C.
leucocephala, have seen no authentic material of either C. striata or

its identity. Cordia striata comes from east-central Bahia, the region

in which Blanchet has collected good C. leuwcocephala. As a species,

C. leucocephala is well marked by the elongate throat of its large corollas,

by the pronounced crowding of the ovoid calyces in the head and by

its leaves which are evidently strigose above. It may be mentioned

that Blanchet, in notes on his labels, has given the corolla as either
te or yellow.

12. C. verbenacea DC. Prodr. ix. 491 (1845). Lithocardium ver-
benaceum Kuntze, Rev. Gen. ii. 977 (1891). C. salicina DC. 1. e. 492.
L. salicinum Kuntze, |. c. 977. L. Fresenii Kuntze, |. c. 976. C.
cylindristachya and curassavica auct. bras.

In eastern Brazil, generally at no great distance inland, from
Cearé to Rio Grande do Sul; in the south apparently ranging westward

ugh Misiones and southern Paraguay to northwestern Argentina.

BRAZIL. . : : . Per-
NAMBUCO: TR cig Sh 1887, Schenck 4184 (BD). Tlial Ontingh

26 JOHNSTON

near Calderao, Oct. 1906, Ule 7263 (K, BD); Prov. Jacobina, 1842, Blanchet
3608 (K, BM, BB); Bahia, Blanchet 211 (DL); vicinity of Toca de Onca, June
1915, Rose & Russell 20125 (NY). Mrnas Geraszs: betw. Carandahy and
Ouro the ee 13036 (K

a y,
Itaborahy, March 1888, Mendonca 946 (BD); shrub becoming 18 dm. tall,
i i . 18, 1921, Holway 1318 (NY); Ilha do Governa-
20, Patschke 180 (BD); shrub, corolla white, SAo Christo-

o Janeiro, Aug. 16, 1825, Burchell 777 (K); Rio Janeiro, Gaudic
(DC, TYPE of C. verbenacea; fi , ISOTYPES), Vauthier 204 (DC, TrPx of
salicina; G, DL, tsoryprs), Gaudichaud 103 ( DL), Gardner 84 (NY, BM,
672 }

: ). Sanra CaTHarina: shru

Capoeira, road to Brusque, Nov. 12, 1886, Schenck 4184 (BD) ; Barra do Itaja-
hy, Mueller 378 (K); Itajahy, Oct. 1885, Ule 479 (BD); dunes, Laguna, June
24, 1 8431 (RS). Rio Granp :

sa Emigration Station, Tristeza, Reineck & Czermak 125 (DC, ;
Grande do Sul, shore of L. da Mangueira, Oct. 29, 1901, Malme 60 (RS); 8.

ure ring

ARGENTINA. Murstonus: San José, Feb. 1922, Molfino (G); Posadas,
Jan. 1922, Molfino (G); Posadas near “La Granja,”’ Nov. 14, 1907, Ekman
1784 (RS); Santa Ana, Sept. 1909, Rodrigues 10019 (G). Sara: Cerro de
Bernardino, March 8, 1905, Spegazzini (G); Reprisa to Campo Gasano,
Tartagal, 500 m., Feb. 6, 1925, Schreiter 3817 (G). pica

: Encarnacién, Sept. 1915, H. ; Rio Pileoma:

Maxpod: 1406-01, ee Ge). 48 ae; ee

This species, along with C. multispicata, C. chacoensis and C. intonsd,
belongs to the group of spicate-flowered shrubs, exemplified by C.
cylindristachya (R. & P.) R. & S. and C. curassavica (Jacq.) R. & 5»
which has its center of distribution beyond our limits in northwestern
South America. The group is one of exceptional difficulty and is very
imperfectly understood. Its species are extremely variable and ill
defined; their taxonomy is complicated and indefinite through an over
excess of names, many of which have been almost traditionally mis-
applied. It is a group in crying need of detailed attention but 18

tely adopted a provincial view-point and have
for our plants the names which have been based upon plants from the

STUDIES IN THE BORAGINACEAE 27

area and hence indubitably applicable to them. While I am rather
certain that some of our plants of this group are either identical with,
or at most no more than varietally distinct from, certain plants of
northern and western South America I feel it preferable to maintain
them as species provisionally distinct than to guess at their relation-
ships in this complex very poorly understood group and thereby,
perhaps, add another bit of confusion by a misidentification and a

the several “species”’ found in our area and that, although I have been
by no means blind to it, I have offered very scanty notes on the obvious
relation which our plants show with the extralimital ones.

Cordia verbenacea seems to be closely related and perhaps conspecific
with the plants of the West Indies and northern Venezuela which appear
to be properly referable to C. curassavica (Jaeq.) R. & S. and C.
macrostachya (Jacq.) R. & S. In our area the species is reasonably
well defined and occurs in its characteristic typical form, generally
near the coast, from Cearé to Sao Paulo. In this form the elongate,
frequently lanceolate leaves are glabrous above and usually roughened
by small and commonly abundant minute tubercles or murications that
are morphologically the bases of undeveloped trichomes. The spikes
are usually slender and elongate and tend to be interrupted in age.
This typical form of the species has about the same range of distribu-
tion as (. multispicata, but except for a very few collections that may
Possibly represent hybrids, the two species are quickly separated on size
and shape of leaves and on the presence or absence of hairs on the upper
leaf-face, The difficulties in defining C. verbenacea appear in southern-
Most Brazil, in northern Argentina and southern Paraguay where it
meets and completely intergrades with C. chacoensis. This latter
Species is one of the interior, ranging in northern Argentina and in
P. araguay. In its extreme forms it differs from C. verbenacea in having
Usually smaller proportionately broader leaves that are strigose or

ute above, as well as usually hairy calyces and denser, shorter and
somewhat thicker spikes: In sorting the specimens preparatory for
citing them in this paper I have been finally forced to distinguish C.
verbenacea and C. chacoensis on the presence or absence of hairs on the
pe leaf-surface. This has brought most of the coastal plants under
i @ and most of those from the interior under C. chacoensis,
es though certain plants from the general range of the latter have had
ses Placed in (. verbenacea. These latter plants as well as those from
© two most southern states of Brazil have the broadish leaves and the

28 JOHNSTON

stout spikes of C. chacoensis. The shape of the leaves and spikes,
however, do not furnish characters capable of precise definition and
hence I feel it best to place my emphasis on the more practicable
criterion of leaf-pubescence and accept these aberrant southern and
western plants as forms verging off towards C. chacoensis. As will no
doubt have been inferred I am inclined to believe that C. chacoensis
and C. verbenacea are incapable of even reasonably sharp separation
and so consequently am of the opinion that we are probably concerned
here with no more than a species and a geographic variety. Being un-
certain as to the status of C. verbenacea I have, however, kept them
both as provisionally distinct.

13. C. chacoensis Chodat, Bull. Soc. Bot. Geneve ser. 2, xii. 218
(1921).

From southern Matto Grosso southward through Paraguay to
Corrientes and thence westward to Salta and La Rioja.

BRAZIL. Marto Grosso: Rio Paraguay near Pao d’Assucar, Feb. 1892:
Moore 1086 (BM, BD). ;

PARAGUAY: along the upper Rio Apa, Feb., Hassler 8442 (BM, BD, DL,
Hass, BB); Puerto Casado, El Chaco, Oct. 23, 1893, Lindman A2231 (RS);

ar Co ién, Aug., Hassler 7248 (BB,

» BD, Hass, DL); Concepcién, Sept. 1892, Kuntze (NY, BD); Est

Santa Maria, Jan. 14, 1897, Anisits 2607 (RS); Villa Encarnacion, Oct. 16,
1886, Niederlein 1826 (BD).

ARGENTINA. Misiones: Posadas, Feb. 9, 1883, Niederlein (BD); road

San Javier, Feb. 1922, Molfino (G). CorrreNnTEs: Empedrada, Au » 23,

1892, Niederlein 304 (BD). Sata: Rio Blanco betw. Bado Hondo and Santa

n ent. * °

po. . Burruyacu, 800 m.
1100 m., Dee. 10, 1906, Lillo 5333 (G).
1600 m., Dec. 24, 1928, Venturi 7848 (G)

I assemble here a group of plants, which are probably not specifically
distinct from C. verbenacea, but which differ in being definitely plants
of the interior, rather than of the coast, and in having hairy uP
leaf-surfaces, frequently shorter and broader leaves, shorter and some
what thicker spikes and perhaps more pubescent calyces. For further
discussion see my notes under (. verbenacea. ws

14. C. intonsa, nom. nov. (. hirsuta Fresen. in Mart. Fl. Bras. 1
pt. 1, 19 (1857), not Willd. (1798). Lithocardium hirsutum Kunts
Rev. Gen. ii. 977 (1891).

Known only from northern Minas Geraes.

STUDIES IN THE BORAGINACEAE 29

BRAZIL. Munas Gerazs: betw. Salgado and Vao do Parandn, Sept. 1818,
Martius (BD, 1soryre).

I know this species only from the type-collection which was made
by Martius in northwestern Minas Geraes. It is related to C. ver-
benacea but differs in having the younger parts of the stems hirsute,
the spikes very dense and the leaves as distinctly hairy above as in C.
chacoensis.

15. C. multispicata Cham. Linnaea iv. 490 (1829); Fresen. in Mart.
Fl. Bras. viii. pt. 1, 17 (1857). Lithocardium multispicatum Kuntze,
Rev. Gen. ii. 977 (1891). C. bahiensis DC. Prodr. ix. 489 (1845).
Varronia spicata Salam. ex DC. 1. c. in synonymy. C. glandulosa
Fresen. 1. c. 19. LL. glandulosum Kuntze, |. ¢. 977. C. campestris
Warm. Kjoeb. Vidensk. Meddel. 1867: 12, f. 2 (1868).

From the mouth of the Amazon southward through eastern Goyaz
and the coastal states of Brazil to Rio Janeiro.

BRAZIL. Pars: shrub 1
1929, Killip & Smith 30403 (G, US); compact shrub 9 dm. tall, corolla white,
Tapané near Pard, Oct. 25, 1929, Killip & Smith 30287 (US); banks of Ura
Sie ard, Aug. 15, 1849, Spruce ); stems elongate and scandent,

ite 6 395 es

truncate and white, Par, Aug. 20, 1829, Burchell 9558 (K); a shrub 1.5 m.
tall in aie but becoming scandent in forest, corolla white, Paré, Dec. 30,
529, 9987 (G, K); near Para, Feb. 4, 1907, Brenning 929 (BD);
vicinity of Pard, 1849, Spruce (K, ; Vicinity of Pardé, Dec. 16, 1907,
41 (BM, BD, DL, RS) and 282 (DL); corolla purple, beach at Prainha,
Noy. 26, 1873, Traill 661 (K). Marannao: Maranhao near Rio Grande,
Feb. 24, 1904, Brenning 970 (BD). Crark: Quixada, Feb. 27, 1910, Loefgren

rs . PERNAMBUCO i rd

y Gardner 1081 ; near Pernambuco, Nov. 1837,
ego shay M corolla white, assu, Oct. 21, ag OF ve Ys me
amage corolla white, Dois Irmaos, July 15, 188 ea
Ramage (BM); Pernambuco, April 18, 1918, Curran 61 (G), ALAGOAs: shrub

ca. 18 dm. tall, banks of Rio San Francisco near Piassabasst, March 1838,
Garaner 1365 (K, BM). Goyaz: shrub, corolla white, Formosa Dec. 24, 1894,
20 21781 (K, BD, DL, RS). Banta: Ilha Itaparica near , Casar
(BAI) OL); Bahia, Blanchet 142 (DC), 142-633 (BM), 633 (NY, DC), 295
M), 735 (BM, DC), 877 (BM, DL), and 1235 (BM, DL); Bahia, Salz-
ann (K, DL) 376 (DG, trex of (. bahiensis); Bahia, Lockhart (BM); Bahia,
Sony? 1920, Silva, Mus. Paul. 4181 (G); near Villa Ilhéos, Martius (BD,
ndulosa). Minas Grraks: in campo near Cabezeiras da
indefinj ch 8, 1864, Warming (BD, tsoryrE of C. campestris) ;
1830. M2 Claussen (K, DL) 221 (G, RS). Rio Janerro: Rio Jan Nov.
» Meyen (BD). Ixpurivrre: Brazil, Sellow (K, BM, BD, DL) 46 (BD,
BD). ©: ™ultispicata), Claussen 417 in pt. (BB), Herb. Mus. Vind. 1745 (K,

This is a large bushy or more or less clambering shrub probably most
closely related to C. Aubletii DC. and such related forms of northern
eek America and the West Indies. It is distinguished in our area

¥ its large broad ovate or oblong leaves which are evidently hirsute

30 JOHNSTON

on the upper surface. The types of C. multispicata, C. bahiensis and
C. campestris are very much alike and represent the common form of the
species, a form which does not have a very abundant velvety indument
on the lower leaf-face. In the type of C. glandulosa, as in some col-
lections made by Blanchet, no. 877, the lower surfaces of the leaves
have a much denser and much paler indument that is almost felty
rather than velvety: This pale form is probably the most striking
variation in the species but I doubt if it is striking enough to merit
special nomenclatorial recognition.

16. C bosa (L.) Don, Gen. Syst. iv. 383 (1837); Urban,
Symb. Ant. iv. 519 (1910). Lantana corymbosa L. Sp. Pl. 628 (1753).
C. ulmifolia Juss. in, Dum.-Cours. Bot. Cult. ed. 1, ii. 148 (1802).
Varronia guasumaefolia Desv. Jour. de Bot. i. 276 (1808). C. guazu-
maefolia R. & S. Syst. iv. 463 (1819). L. guazumifolium Kuntze, Rev.
Gen. iii. pt. 2, 206 (1898). C. discolor Cham. Linnaea iv. 482 (1829).
Lithocardium discolor Kuntze, Rev. Gen. ii. 977 (1891). C. urticifolia
Cham. Linnaea iy. 483 (1829). C. hermanniaefolia Cham. |. ec. 484.
L. hermanniaefolium Kuntze, Rev. Gen. ii. 977 (1891). C. hermanniae-
folia, var. calycina Cham. |. c. 486. C. Salzmanni DC. Prodr. iv. 494
(1845). C. Salzmanni, var. lanceolata Fresen. in Mart. Fl. Bras-
vill. pt. 1, 20 (1857). C. lapensis Warm. Kjoeb. Vidensk. Meddel.
1867: 9, f. 4 (1868). L. corymbosum, forma glabriusculum Kuntze,
Rev. Gen. iii. pt. 2, 206 (1898). C. paraguariensis Chodat & Hass.
Bull Herb. Boiss. ser. 2, v. 305 (1905); Chodat, Bull Soc. Bot. Genéve
ser. 2, xii. 217 (1921).

Widely distributed in the American tropics and extending south to
the mouth of the Rio La Plata where in the vicinity of Buenos Aires It
sets the southern limit for the genus Cordia in America.

BRAZIL. Amazonas: shrub 1-2 m. tall, Jurué Miry, Rio Jurué, July 1901,
5687 (K, BD, DL); Rio Negro, Jan. 11, 1887, Moura 568 (BD); shrub ca
Cero

eh 1850, Spruce ak
PERNAMBUCO: Tapera, Feb. 1927, Pickel 549 (BD); Inaranham, hej 7, 1887,
Ridley a M). Bana: ar ; Bahia,
Lockhart (BM); Bahia, Blanchet 1171 (BM, DL}; Bahia, Salemann (K, DL);
Bahi 37 sane : é .

> tu ‘0 F pt. Di

Luzia do Rio das Velhas, Jan. 31, 1895, Schwacke 11459 (BD); Sitio, Mareh 19,
1887, Schenck (BD); Caldas, Mosén 1506 (RS), Mosén 361 (is), Li ii
(RS), Regnell IIT 908 (BD, RS); Alberto, Caldas, Feb. 17, 1847, Regnell

—

STUDIES IN THE BORAGINACEAE 31

8 (RS); Serra de Caldas, 1848, Regnell IIT 908 (RS); indefinite, Claussen
tk) , (BB) and 224 (G, NY , RS). Rio JANEIRO: Mau, Aug ads 7, Schenck
: ‘a

, Mus. Paul. ers joan
ningham 229 (BM); Morumbi, March 7, 1827, Burchell 4406 (K ); See Recat:
Jan. 18, 1827, Burchell 3895 (K); Sao Paulo near Parca da Alegria, March 24

4681-5 : i

182%, Burchell ; Campinas, Mus. Paul. 22464 ; Campinas,
Nov. 8, 1873, Severin 176 ogy Marim, Marc 874, Mosén 1508
RS); Ribeirdo to Preto, Oct. 25, 1889, Loefgren 1447, Mus. Paul. 11311 (G);

ei 93, Mus '
de Alferes Rodrigues, Oct. 1899, Edwall, Mus. Paul. 11316 (G); Jundiacanga,
Loefgren 249, Mus. Paul. 11314 (G); Itapira, May 17, 1927,

H ); Therezina, _ 22, 1911, SANTA

Catarina: Santa Catharina, Mueller 3 06, (K); near Itajahy. Feb. 1886,

Ule 512 (BD); Santa Catharina, audi ich aud 161 (NY, DL). Rio GranvE
: 1 57

A 6 (DL) and Ihering 43 (B Marto Grosso: moa) da Chapada,
Shop (Rey. — 129 (BM); Santa Anna da Chapada, May 24, 1903, Malme
: il, Herb
42

Lund (DC), Martiue Herb. Fl. Bras. 1080 (G, NY, K, — DL, BBY

7 jog Sellow (K, BM, BD, BB, authentic C. herma iaefolia) ; Breen merid.,
ellow D, authentic var. calycina) ; Sellow 138? | (BD, var. calyci
more UAY. Rio Negro: Isla del Vizcaino, Jan, 1908, Berr 04424 and 4456
G yes Cotontas: Isla San Gabriel , Jan. 1902, Berro 1757 (BB): Tela San
pak Feb. 1920 Motfino. 96 (08 GS. San, Jie "Salvas de San José, Pavon,
P 1895, Arechaoa eta (BD).
AY: near Igatimi, Nov., Hassler 5552 (BM, BD); Loma Clavel,
n Ov. “ gee ex Hassler 2618 (G, BM, BD, DL, Hass BB):
Fiebrntrts Oct., Hassler 1335 (BM, Hass, BB); Cord.’ de Altos, Nov. 1902,
BB); 458 (K, BD, Hass, DL); Cond. dealee , Dec., Hassler 3600 (BM, Hass
me Or ag. Ypacaray, 1913, Hassler 12131, 12609 and 12609a (G, BM, BD,
Recaray, 1¢ 0, Sept. 1896, Endlich 329 (BD); near Valenzuela,

32 JOHNSTON

valley of Rio Y-aca, Feb. 1900, Hassler 7065 (BM, Hass, BB); upper Rio Y-acé,
Feb. 1900, Hassler 7069 (BD, BB); near Paraguari, Dec. 1900, Hassler 6575
(BB, Type of C. paraguariensis; , Hass, DL, 1tsoTypEs); betw. Paragua

and Ibitimi, Feb. 11, 1876, Balansa 2654 (DL); Ibitimi, Sept. 1874, Balansa
2043 (K, DL, BB); Villa Rica, 1928, Jérgensen 3735 (G); Villa Rica, Dec. 1874,
Balansa 2655 (DL); Villa Occidental, May 1874, Balansa 2044 (K); Encar-
nacion, Oct. 15, 1886, er ge 2220 (BD); Rio Pileomayo, Morong 987 (G,
‘ ! ate 4

ARGENTINA. Misiones: Falls of Iguazt, King Albert Archipel., March
17, 1883, Niederlein (BD); Campo Eré, Jan. 1887, Niederlein 1818 (BD);
east of San Daniel, ee 12, 1883, Niederlein (BD); San Ignacio, Dee. 16,

reto, March 2
1790 and 1791 (RS); Posadas near ‘‘La Granja,”’ Dec. 24, 1907, Ekman 1785
and 1786 (RS); San José, Feb. 1922, Molfino (G); Apostoles, Jan. 29, 1926,
Parodi 7024 and Clos 2077 (G). Corrientes: Monte de Tareiry, Dept. San
Tomé, Oct. 11, 1886, Niederlein 1834 (BD). Entre Rio: Arroyo Cupalen,

ru BuENOS Ss:
Chafia, Delta of the Parana, Jan. 1916, Hawman 26 /90 (G); a ot
L 86. . For-

Jan. 1 int : : ier :
1874, Cae £ ee (BD); Verte. Macacee=t oe ilo. 12054 (G);
Naranjal, 500 m., Schreiter 1596 (G).

One of the most widely distributed of American Cordias and at least
the most frequently collected if not the most common species in our
area. Over most of its range it is quickly distinguished by its glomerate
or small cymose inflorescences. Only along the Andes is it apt to
confused, for in that region are found its closest relatives, such habit-
ally similar species as C. scaberrima HBK., C. lanceolata HBK., ¢:
lantanoides Spreng., etc. Ranging widely and being rather variable
C. corymbosa has a large synonymy much of which is founded upon
material from north of our region. In the present treatment I have
given only the basic names and those synonyms established upo?
plants from the area in which we are interested.

There are several distinctive variations of C. corymbosa developed
in our region. Having spent considerable time sorting and studying
the specimens I have finally become convinced that these variations
are only vaguely correlated with geography, that from much inter
gradation they are very indefinite and that in all probability they
Tepresent chiefly ecological forms. The most important of thest
variations are in pubescence. Since this affects the gross aspect °
the plant and gave the reason for most of the synonyms it may a
desirable to have names for its major manifestations. The extrem®

STUDIES IN THE BORAGINACEAE 33

of these principal forms of the species may be, accordingly, distin-
guished by the following key.

Lower leaf-surface decidedly tomentulose with an abundant fine
c scence.

Coarse hairs sparse or absent, the fine pubescence frequently
BON Saeco ei ee orma typica,
freer term Rtiidane. 2250) a ae forma intonsa,
Lower leaf-surface without a fine pubescence or this very sparse.
oarse hairs absent or sparse and weakly developed, appressed.
forma detonsa.
Coarse hairs abundant, usually more or less spreading...... forma hirsuta.
The type of C. corymbosa comes from the West Indies where the
prevailing form is that which I have indicated as the forma typica.

boliviana Gandoger. The typical form occurs over most of the range
of the species. In our area it has been most frequently collected about
Rio Janeiro where it appears to be locally the predominating form.

brown. Only
Mass of fine tomentum. The types of C. hermanniaefolia and its
variety calycina represent a form with rather numerous scattered
Coarse hairs which tends towards the forma intonsa. Cordia lapensis
Warm., to judge from description, must be a phase of the forma
typica in which the fine tomentum is pallid rather than distinetly
ferruginose.

The form

a intonsa is a frequent one in our area and is characterized
by having

numerous erect or frequently even appressed hairs arising

ma fine tomentum similar to, although usually not quite so abun-
dant as that found on the lower leaf-surfaces in forma typica. The
limits of the form are indefinite for it intergrades completely into both
forma typica and forma hirsuta.

The forma detonsa appears to be most frequent from Ecuador to
Bolivia, where it has been called Varronia dichotoma R. & P., and C.
bifurcataR.& S. Tt

araguay and northern Argentina. It is also sporadic in the coastal
il. The lower surfaces of the leaves in this form are

pubescence found in forma typica or this pubescence

'S short Coarsened and very sparse. The lower surface of the leaves is

Frequently glandular and commonly develops a few scattering coarse

ly with mits. As might be expected the form intergrades complete-
forma hirsuta and to a less extent with forma typica.

34 JOHNSTON

The forma hirsuta is a frequent one in southern Brazil and north-
eastern Argentina. It has been described as C. urticifolia Cham. and
C. paraguariensis Chodat & Hass. and possibly also as C. guasumae-
folia Desv. In this form the lower leaf-surface may be more or less
glandular but it is practically devoid of the fine tomentum found
among the coarse hairs on the leaves in perhaps unseparable forma
intonsa.

Besides the extensive variation in pubescence, C. corymbosa also
exhibits variation not only in the size of leaf-blade but also in the
conspicuousness of the marginal toothing. More significant, however,
is the variation in the development of the inflorescence. In some forms
this is glomerate at maturity being quite congested and suggesting
the truly capitate flower-clusters of those species which are typical
of the section Varronia. In other forms the inflorescence loosens up
in maturity frequently to form open cymes several centimeters broad
that are suggestive, on a smaller scale, of the inflorescence in the
section Pilicordia. Cordia corymbosa, indeed, seems to be a primitive
member of its section, and in the variable development of its inflores-
cence probably gives us a good idea of the manner in which the
characteristic capitate and spicate flower-clusters of the section Var-
ronia must have evolved.

17. C. buddleoides Rusby, Mem. Torr. Bot. Cl. vi. $3 (18%):
Lithocardium guazumifolium, var. santacruzense Kuntze, Rey. Ge?
ili. 206 (1898).

Amazonian headwaters from central Bolivia to northeastern Pert.

PERU. Lorero: :

San Gaban, Aug. 18541 Oar tae (kK, RS Ts a _

BOLIVIA. Coxontas: corolla pale, shrub 1-3 m. tall, Cobija, Rio Act®
Dec. 1911, Ule 9714 (K, BD). La Paz: Mapiri, July-Aug. 1892, Bang 15
(NY, TyPE of C. buddleoides; G, isotype). Santa Cruz: Sierra de Santa
May 1892, Kuntze (NY), TyPE of var. santacruzense).

While this species has not been actually collected in Brazil it ®
most certainly to be expected within our western limits inasmuch 4s
Ule has obtained it at Cobija in northwestern Bolivia just across TH
Rio Acre from Brazilian territory. The types of C. buddleoides and 2
guazumifolium var. santacruzense are very much alike and clearly
represent the same species. Lechler’s collection from San G®
matches these types closely, but Ule’s material has the flowers wit
cylindrical spike rather than in a globose head. Although striking .
first sight this difference is probably due to the state of maturity,
material collected by Rusby, Kuntze and Lechler having bee? ‘

tained at a very early stage in flowering and that of Ule in early fruiting : |

the |

STUDIES IN THE BORAGINACEAE 35

condition after the fall of the corollas. Poeppig’s collection from
northern Peru is the most different among the specimens cited. In
this collection the calyx-lobes lack the definite apical appendages so
clearly developed in the other specimens from Peru and Bolivia.

18. C, axillaris, sp. nov. C. patens, var. monocephala Cham.
Linnaea iv. 486 (1829). ?C. patens, var. polycephala Cham. |.c. C.
patens, var. angustifolia Warm. Kjoeb. Vidensk. Meddel. 1867: 11
(1868). C. patens of Fresen. in Mart. FI. Bras. viii. pt. 1, 21 (1857).

Known in Brazil from Goyaz, western and southern Minas Geraes,
northern and western Rio Janeiro and eastern Sado Paulo, and in
Argentina from Misiones.

BRAZIL. Goyaz: shrub ca. 12 dm. tall, corolla white, road betw. Tambau
and Cercado, Aug. 23, 1827, Burchell 5326 (K); shrub 3 m. tall, along stream
at Farinha Podre, Sept. 11, 1827, Burchell 5695 (K). Minas Gerazs: shrub
nearly 2 m. tall, Uberava, Sept. 2, 1848, Regnell III 909 (K, BD, RS); Gran-
dahy, Sept. 17, 1818, Sellow B1565-c658 (BD). Rio Janerro: shrub, corolla

sy righ ;

4
E
e
a
a"
iz
)
Sg
wo
[o>
~
s
&
£
(a
e
<q
bo
<=)
%
es)
Ca
Q
7
i=)
&
&
co

NDEFIN .
yar: monocephala; K, BM); Bras. trop. Sellow (BD, authentic var. polycephala;
K); Brasil, Sellow 1564 (BD, authentic var. monocephala); Brasil, Pohl. Herb.
Mus Vind. rd (K, in pt., BD

ARGENTINA. Misiones: subscandent, margin of forest, Salto de Iguazi,
Sept. 23, 1910, Rodriguez 10460 (G).

Var. gymnoc a, var. nov. A varietate genuina differt capitulis
paullo minoribus, calycibus maturitate vix accrescentibus fructu
evidenter brevioribus.

Southern Minas Geraes.
is aE. Minas Grrass: suffruticose, ocean, whites Out reek, a cg
1 (Ken. oe (BD); indefinite, St. Hilaire B', 285 (K); inde

, TYPE).
_ The present species, based upon C. patens var. monocephala Cham.,
includes those plants from Brazil treated as C. patens by Fresenius
and subsequent authors. The type of C. patens HBK., Nov. Gen. et
Sp. ili. 75 (1818), was collected by Humboldt and Bonpland in north-
rastern Venezuela about 110 km. southeasterly from Cumana. I
have hot seen any authentic material of this species. Mr. Killip has
Written me that during his studies of Cordia in the European herbaria
Be also failed to find any authentic collections of the species, even at
P avis where one would expect to find the type. As no plants closely
ager ting C. avillaris or C. buddleoides, the species that have been
identified with C. patens HBK., have been found in Venezuela I am

36 JOHNSTON

inclined to agree with Mr. Killip in believing that true C. patens HBK.
must be some form of the variable and widely distributed C. corymbosa
(L.) Don. In any case as it seems very improbable that our Brazilian
plant will prove referable to this obscure species of northern South
America its treatment as a distinct species appears both justifiable
and desirable.

The exact source of the type of C. patens var. monocephala Cham.,
and hence that of C. avillaris, is uncertain. It was originally given as
merely from “ Brasilia tropica.’’ The authentic collections at Berlin
named by Chamisso are evidently part of the same collections as the
specimens from Sellow’s private herbarium which bear the number
1564, but no exact geographic data. Perhaps significant is the fact
that this number is the next smaller than the first number in the com-
plex designation, “ B1565, c658,”’ found on another sheet from Sellow’s
private collection. This latter complex number occurs with material
that is very evidently part of the same collection as that upon which
C. patens var. polycephala Cham. was founded, and is accompanied
by definite geographical data giving its source as Grandahy, a locality
in southern Minas Geraes between Barbacena and Ouro Preto.

After some indecision I am now of the opinion that the two varieties
of C. patens described by Chamisso represent merely the fruiting
(var. monocephala) and the early flowering (var. polycephala) states of
the same species. It is even possible that both may have been 0
tained in the same region. In the authentic material of both varieties
the leaves, stems and calyces are covered with a darkly brown pubes-
cence, the coarse stems being conspicuously shaggy with spreading
brown hairs. These are the striking characters of the plants which
I have cited from Rio Janeiro, Sao Paulo and Misiones. The material
collected by Burchell in Goyaz and by Regnell in Minas Geraes,
however, differs slightly in having much less conspicuous, shorter
pubescence and a decidedly lighter cast to the plant as a whole. 48
this northern material is not mature I consequently hesitate to #1V°
it the varietal recognition which I suspect the study of comp
material will show it to deserve. Z

The calyces of C. azillaris are crowded in the head and become qui’
congested as the fruit matures and the calyces enlarge. The calyx 8
accrescent becoming inflated and completely, although loosely, invests
the fruit. Certain material from Minas Geraes, however, does not
have accrescent calyces, the smaller heads being less dense a0
fruit at maturity protruding from the calyx. ‘This is the form that!
have recognized above as constituting the var. gymnocarpa. It:
to be a distinctly more slender plant than the typical C. axillaris and

sealing). remeron

STUDIES IN THE BORAGINACEAE 37

it is quite possible that it may eventually be thought distinct enough to
merit specific rank.

19. C. leucomalla Taub. Bot. Jahrb. xv. Beibl. 38, 14 (1893).

State of Rio Janeiro.

BRAZIL. Rio Janerro: shrub, Ilha do Cabo Frio, Oct. 17, 1899, Ule 4775
(BD); shrub in woods, Laranjeiras, Glaziow 4146 (BD, TPE; K, IsorypPE);
Morro Flamengo, Miers (BM); shrub nearly 2 m. tall, Praia de Copacabana,
Feb. 1841, Gardner 5563 (K, BM); Tedreiva, Miers 3931 (K).

This well marked plant is known only from the vicinity of Rio
Janeiro and from the coast to the eastward. It appears to be a loosely
branched shrub of thickets and to be a very distinct species. Its
Most conspicuous features are its very contrastedly bicolored leaves,
these being green and somewhat velvety above and white with a floc-
cose tomentum beneath. The dense heads of small flowers are usually
ovoid.

20. C. corchorifolia A. DC. Prodr. ix. 496 (1845). Inthocardium
corchorifolium Kuntze, Rev. Gen. ii. 976 (1891).

own only from southeastern Alagéas and from Bahia, growing
probably at no great distance from the coast.

BRAZIL. Atacéas: shrub 1.5 m. tall, banks of the Rio San Francisco near
Penedo, March 1838, Gardner 1364 (K, BM, DL). Banta: Bahia, 1833,
Blanchet 1073 (DL); Bahia, Blanchet 2073 (DL); small shrub, in moist shady
Places, corolla white, Dec., Blanchet (DL). INDEFINITE: Brazil, Herb. Dowville
(DL, Type).

Very closely related to C. globosa and perhaps only a well marked
local form of it differing in its much sparser pubescence and in the
Presence of very abundant commonly stipitate glands on the herbage
and calyces,
ai, ©, caput-medusae Taub. Bot. Jahrb. xv. Beibl. 38, 15 (1893).
Known only from a single collection made in southern Minas
eraes,

G

PRAZIL. Minas Gerazs: shrub, Gandarela, Glaziou 152873 (BD, TyPE;
ISOTYPE),

Obviously related to C. globosa but differing in its large heads, its
very long appendages on the calyx-lobes, its copious glandularity and
Its Conspicuously hairy almost velutinous leaves. Until more collec-
tons of the plant are available for study, however, the status of the
Species must remain a little doubtful.

P 22. C. globosa (Jacq.) HBK. Nov. Gen. et Sp. iii. 76 (1818); Urban,
Symb. Ant. iv. 520 (1910). Varronia globosa Jacq. Enum, Syst. Pl.
* (1760) and Sel. Stirp. Amer. 41 (1763).

38 JOHNSTON

Widely distributed in the West Indies and in northern South Amer-
ica; in our area known only from the island of Fernando do Noronha
and on the Brazilian mainland in Alagéas and in southeastern Minas
Geraes.

BRAZIL. Perrnampuco: shrub ca. 18 dm. tall, Sapate, Fernando do
Noronha, Aug. 20, 1887, Ridley, Lea & Ramage 89 (K). AxaGo6as: shrub ca.
18 dm. tall, very common on banks of the Rio San Francisco, Feb. 1838,
Gardner 1367 (K); at Propia, Feb. 1838, Gardner 1367 (BM). Munas Grrazs:
shrub, Serra da Mantiqueira, Jodo Gomez, Glaziou 13037 (K, BD); shrub, Joao
Gomez, Glaziou 14141 (K).

I am unable to distinguish the above cited collections from C.
globosa, a species widely distributed north of our area. In Brazil it
is partially replaced by the closely related C. corchorifolia and C. caput-
medusae, plants which may eventually be found unworthy of specific
separation. I have not attempted to give complete synonymy of
this species since all of its names are based upon material from out-
side our area.

23. C. calocephala Cham. Linnaea iv. 488 (1829) and l. e. viii. 129
(1833); Fresen. in Mart. Fl. Bras. viii. pt. 1, 24, t. 8 (1857). Latho-
cardium calocephalum Kuntze, Rev. Gen. ii. 976 (1891).

Southern Goyaz, western Minas Geraes and adjacent Sao Paulo.

BRAZIL. G 3-10 dm. tall,
at Canja, Nov. 5, 1828, Burchell 8300 (K); ca. 6 dm. tall, corolla white, near

Caldas, Dec. 27, 1862, hone IT 213 (BD, RS); indefinite, 1845, Widgren (RS);
e, 1840, Claussen 222 (G, K, DL, RS). Sao Pa R

pee — Hu 213 (RS); Mursa Jundiahy, Feb. 4, 1915, bra
Fey Ripe At an. DEFINITE: Bras. trop., Sellow 5611 (BD, TY

_ A very well marked species, which probably has its closest relative
in the evidently distinct C. sessilifolia Cham. The exceptionally
well developed and very conspicuously elongate apical appendages
on the calyx-lobes are very characteristic. The brown flower-heads,
which are usually globose at first, tend to become 2-3 times as long
broad in maturity. The plant is apparently a small shrub with a very
ne sparsely branched coarse stems and becomes as much as 12 f

24. C. sessilifolia Cham. Linnaea iv. 488 (1829). Lithocardiwm
sessilifolium Kuntze, Rev. Gen. ii. 977 (1891). C. sessilifolia, va
macrantha Cham. |. ¢. viii. 129 (1833).

Minas and Goyaz; reported from Bahia by Fresenius.

sseetan

STUDIES IN THE BORAGINACEAE 39

ZIL. Goyaz: calyx dark, corolla white campanulate, betw. Ribeirao
dos Macacos and Olhos da Agoa, Sept. 2, 1829, Burchell 7506 (K). Munas
Grraks: Quartel de S. Antonio, 1818, Sellow B1562-c655 (BD, Type). INpEF-
INITE: Brasil, Sellow (K), 1562 (BD), Claussen 410 (BB).

Var. tomentosa (Chodat & Hass.), comb. nov. C. villicaulis, var.
tomentosa Chodat & Hass. Bull. Herb. Boiss. ser. 2, v. 481 (1905).
C. villicaulis Fresen. in Mart. Fl. Bras. vill. pt. 1, 24 (1857). Litho-
cardium villicaule Kuntze, Rev. Gen. ii. 977 (1891). C. caaguazuensis
Chodat, Bull. Soc. Bot. Genave ser. 2, xil. 216 (1921).

Ranging from Sao Paulo and Parana, in Brazil, to northeastern
Paraguay; Fresenius reports it from Goyaz.

. SAo Pauto: Canna Verde, April 1848, Regnell III 910 (RS);
Sarapuhy, Oct. 30, 1887, Loefgren 289, Mus. Paul. 11272 (G). Paran: betw.
eg and Itararé, Dec. 11, 1910, Dusén 10936 (RS); Itararé, Nov. 7, 1910

n 10399 (RS). Inperunrre: Brazil, Pohl, Herb. Mus, Vind. 1740 (K, BD):
Brazil, Sellow 5414 (BD)
PARAGUA

campo near Caaguazi, March 1908, Hassler 9108 (phe aw of C. caagua-
; Hass, DL, tsorypns); indefinite, Hassler 6355 (Hass.).

Cordia sessilifolia and its variety agree with C. calocephala in having
brownish hairs on the calyx and stems. In the present plants, however,
these are more distinctly spreading and more conspicuous than in
C. calocephala. The calyx also is very different, the tips of the lobes
having inconspicuous appendages only 1-2 mm. long; the calyx proper,

Pas truncata Fresen. in Mart. Fl. Bras. viii. pt. 1, 25. (1857).
thocardium truncatum Kuntze, Rev. Gen. ii. 977 (1891).
yaz"and Minas Geraes.
BRAZIL. Goyaz: small shrub, be ’ ;
Glazioy : , betw. Chico Lobo and the head of the Rio
re: 21779 (K, BD, DL). Minis GeraEs: indefinite, Claussen 220
Vind. 73 K, DL, RS). INDEFINITE: “Serra dos Cristaes,’’ Pohl, Herb. Mus.
- 1741 (K, tsorypr ?),
* Although probably most closely related to C. sessilifolia, from which
1S Separated by its very different appressed much less brownish

40 JOHNSTON

pubescence, this species is most difficult to separate by readily stat-
able characters from the very obviously distinct, C. guaranitica.

. C, guaranitica Chodat & Hass. Bull. Herb. Boiss. ser. 2, v. 305
(1905); Chodat, Bull. Soc. Bot. Genéve ser. 2, xii. 217, f. 312 (1921).
C. guaranitica, var. pedunculosa Chodat & Hass. l. c. C. guaranitica,
var. foliosa Chodat & Hass. 1. ec.

Northeastern Bolivia, Paraguay and northern Argentina.
BOLIVIA. Santa Cruz: Velasco, 200 m., July 1892, Kuntze (NY, as L.
nniaefolium).
PARAGUAY: shrubby 4-8 dm. tall, in campo, San Rafael, Rio Apa, Oct.
BB ‘unculosa; BM ;

Hassler 7655 (BB, Type of var. F , Hass, ISOTYPES);

shrubby 5-10 dm. tall, in campo near ec., Hassler 5619 (G, BM,
, DL, Hass, ); shrubby 8-10 dm. tall in sandy pl near

Tobaty, Sept. 1900, Hassler 6359 (BM, BD, D BB); shrubby 3-10 dm.

, BM, BD, DL Altos, Nov. 1902,
Fiebrig 444 (BD); old fields, Cord. de Altos, Sept. 1902, Fiebrig 110 (BD);
campos de Cord. de Altos, Jan., Hassler 1022 (Hass.); shrubby 5-10 dm.

campo near Valenzuela, upper Rio Y-acd, Jan. 1900, Hassler 6962 (BB, TYPE

L ‘ . Santa
Anisits 2667 (RS); “‘Camp Dr. Abente,” Dec. 1898, Anisits 2748 (RS). we
ARGENTINA. Sarra: Ordn, Rio Santa Marfa, Feb. 8, 1906, Spegazzin

A very distinct species probably closest to C. truncata, but readily
separated from its relative by its paler cast, the antrorse pubescence
on the more elongate leaves, its more slender habit, more conspicuously
appendaged calyx-tips, fewer-flowered less dense heads and more
southern range. It appears to be a few-stemmed strictly branched
shrub 5-10 dm. tall frequenting dry open ground. The two varieties
of this species that have been described are based upon characters
too variable and trivial to warrant recognition.

Section Eucordia Johnston. Cordia L. Gen. 87 (1754) and Sp. Pl
190 (1753), type-species, C’. Sebestena L. Sebesten Adans. Fam. Pl. “a
177 (1763); Britt. & Wils. Bot. Porto Rico ii. 124 (1925), type-species
U. SebestenaL. Cordia-Sebestenae HBK. Nov. Gen. et Sp. iii. 68 (1818)
type-species, C. Sebestena L. Cordia § Sebestenae Don, Gen. Syst
lv. 374 (1837). Not Sebestena Gaertn. Fruct. et Sem. PI. i. 364, t- 76 :
f. 1 (1788). Pavonia Dombey ex Lam. Tab. Encye. i. 421 (1791), ™
Synonymy; type-species, C’. lutea Lam. _ Coilanthera Raf. Sylva Tellur.
38 (1838), type-species, C. rotundifolia R. & P. Ectemis Raf. 1. ¢ 39
(1838), type-species, C. lutea Lam. Cordia § Sebestenoides DC. Prodr.
ix. 476 (1845), type-species, C. speciosa Willd. ?Macria Teno?
Mem. Math. Fis., Soc. Ital. Sci. Modena xxiv. 366 (1847); type-specl®
M. callipticantha Tenore, which see in my list of doubtful speci® |

STUDIES IN THE BORAGINACEAE 41

Paradigma Miers, Trans. Linn. Soc. Bot. ser. 2, 1. 31, t. 8a (1875),
type-species, C. Galeottiana A. Rich. Plethostephia Miers, 1. e. a2; t.
8b (1875), type-species, C. angiocarpa A. Rich. Lithocardium [L.]
Kuntze, Rev. Gen. ii. 438 (1891), type-species, C. Sebestena L.

Under this section I have assembled a very diverse group of coarse
shrubs and trees characterized by their large conspicuous corollas
and large fruit that is more or less completely sheathed or even fully
enclosed by the calyx. There are about a dozen species in America
and certainly one, and possibly several more, in southern Asia. In
the section there are some very interesting and rather extreme develop-
ments in the structure and number of floral parts. Cordia Sebestena
L. of America and (. subcordata Lam. of the Old World have the calyx
tightly and completely enclosing and more or less definitely adnate to
the fruit. In C. angiocarpa A. Rich. and C. Galeottiana A. Rich. the
stamens and corolla-lobes are increased in number, the corolla-lobes
in the former species becoming as many as 15. In C. decandra H. & A.
the stamens are uniformly 10. In another species, C. rotundifolia R.
& P., the stamens and corolla-lobes vary independently between 5
and 9 in number. Among the Brazilian species, such as C. superba,
there appears to be a slight tendency towards zygomorphy. Although
this great diversity in corolla-structures suggests heterogeneity I am
of the opinion that the group is really a natural one.

The sectional Synonyms given above are those applying to the

4 species that seems to be very closely related to the American C.
Sebestena L,

Since Britton & Wilson, Bot. Porto Rico ii. 125 (1925), have re-
cently selected Cordia glabrata L. as the type of Cordia and have

Plant Linnaeus called C. Sebestena. Turning finally to the description
in the Genera Plantarum, |. c., we find that Linnaeus according to his
dre mcf. Pennell, Proc. Acad. Nat. Sci. Philadelphia lxxxii. 10 (1930),

W it to cover one species in particular, in this case C. Sebestena.

42 JOHNSTON

Certainly the phrase “drupa. . . calyce tecta’’ is not applicable
to C. Myza or C. glabra the other two species mentioned in the com-
panion volume, the Species Plantarum for 1753. Since the original
generic description of the genus Cordia was thus based upon C, Sebes-
tena, since the cited synonym, Sebestena Dill., applies to that species
and since even the name Cordia itself was historically first applied to
it, I am entirely unable to see why that species should be passed over
and C. glabra selected as the type of the genus. Cordia glabra is even
the last in order of the three species of the genus given in the first
edition of the Species Plantarum.

Key TO THE SPECIEs.

apparently axillary on slender pedicels.
Leaves ovate, under surface glabrous except for dense tufts of
ir in the axils of the primary nerves; flowers apparently
pringing directly from the leaf-axils............_. 27. C. mucronata.
Leaves lanceolate, generally pubescent beneath; flowers ae J
mes . C.

Leaves more or less abundantly pubescent, opaque, margins fre-
quently serrate.
Leaves with abundant spreading brownish hairs, the pubes-
cence evident, more or less velvety; buds brownish
ee ea eu Oi vilece lod used )cck oy 30. C. rufescens.
Leaves strigose or short-hirtellous beneath, usually somewhat
sca strigose.
Corolla 30 mm. long or more
Corolla 13-15 mm. long; rare plant

vet;
brid; buds

27. C. mucronata Fresen. in Mart. Fl. Bras. viii. pt. 1, 9 (1857)
Lithocardium mucronatum Kuntze, Rev. Gen. ii. 977 (1891).
States of Rio Janeiro and Espirito Santo; rare.
. BRAZIL. Rio Janerro: Cabo Frio, May 5, 1887, Schenck 3843 (BD);
vicinity of Rio Janeiro, Schott,” Herb. Mus. Vind. | 747 (K); vicinity of
Janeiro, Glaziou 7778 (K, BD). InpeFintre: Brazil, Sellow (BM) 254 (BD)
and 472 (BD, TYPE).
: A very distinct species readily recognized by its very reduce
inflorescence and hence apparently axillary flowers, and by its sma”
ovate leaves which are glabrous except for tufts of hairs that occu!
im the axils of the veins on the under surface. Fresenius reports
species, the type, from the coast between Vittoria and Bahia and henc? .
from the state of Espirito Santo or from southern Bahia. Otherwisé
I know it only from the stations I have cited above.

STUDIES IN THE BORAGINACEAE 43

28. C. candida Vell. Fl. Flum. 98 (1825) and Icones ii. t. 155 (1827).
Known only from the region about Rio Janeiro; rare.

BRAZIL. Rio JANerro: shrub, corolla white, betw. Jerecino and Realengo,
oh 12088 (K, BD). Inverinrre: no locality given, Bowie & Cunningham

A very distinct species. Its inflorescence is unique, consisting of
one or more slender-stemmed cymes which spring directly from the
upper leaf-axils. Except for C. mucronata, the other species of this
relationship have the cymes borne on an elongate terminal peduncle.
In C. candida the peduncle is not developed and what would appear as
branches in the pedunculate panicle of the other species here appear
as axillary cymes. These cymes have a slender axis which apparently
becomes more or less spreading or even reflexed in age. The fruit is
drupaceous. The type of C. candida is said to have come from “ Silvis
Parochiae dictae Campo grande,” evidently the present town of
Campo Grande which lies about 40 km. west of Rio Janeiro.

29. C. taguahyensis Vell. Fl. Flum. 98 (1825) and Icones ii. t. 154
(1827). -C. glabra Cham. Linnaea viii. 124 (1833), not L. (1753);
Fresen. in Mart. Fl. Bras. viii. pt. 1, 8 (1857). C. glomerata Lem. Jard.
Fleur. iv. t. 326 (1853). Sebestena glomerata Lem. 1. c. in synonymy.

Coastal states of Brazil from Maranhao to Parana.

BRAZIL. Maranuio: Vianna, July 24, 1922, Ozimo de Carvalho 85, Mus.
Paul. 8112 (G). Cwari: shrub 3-10 mm. tall, corolla white, Serra Baturité,
Sept. 1910, Ule 9898 (K, BD). Prrnamsuco: shrub 2 m. tall, Tapera, Dec.
1926, Pickel 1241 (BD). Banta: Bahia, Blanchet 320 (BM, DL), 239 (DL),
am (BM, DL), 1386 and 1915 (DL); Ilhéos, Blanchet (DL);.Muritiba, 1842,
Blanchet 3612 (BM, DL, BB); indefinite, Blanchet 606 (NY) 905 (BM); Rio

uty Basin, 1915, Curran 182 (US). Rio JANzrRo: corolla white, rugose
Frechal an

* Be)
Burchell 1460 (K); small shrub by side of stream in Laran :
BR Gardner Be a); Pedra de Gavea, Martius po Pras. 462 (NY, K,
> :

BD, DC, D
Gio; Bowie & Cunni ; Sete P Baretto, Sept. 26, 1872,
Irma, 6051 (K, BD); Serra da Bica, Feb. 1897, 4297 (BD); Pedra Dois
(K) 0s, March 23, 1929, Smith 2131 (G); vicinity of Rio Janeiro, Glaziou 5951
i ; Rio Janeiro, Vauthier 74 (DC, DL), Miers 3020 (K,
14, Seeidichaud (BD), Graham (K} and Widgren (RS). Rio Jane

29 raw 339 (BD, authentic C. glabra). Sio PauLo: Sao Se fio, Mar
Pi 892, Edwall 1733, Mus. Paul. 18191 (SP). Parand: Porto de Cima, Serra
Mus. P2007, Jan. 24, 1914, Dusén 14395 (RS). Inpertnrre: Brazil, Herb.

“8: Vind. 1748 (K), Sellow 95 (K), Sellow (K).

Among the species with large corollas this one is quickly distin-
Or by its glabrous leaves that are green above and pale and some-
What shiny beneath. Its nearest relative is C. superba, a species

44 JOHNSTON

having more or less pubescent leaves in which the lower surface is dull
and the margins are frequently toothed. The flowers of this relative,
also, are larger and develop a somewhat larger and more evidently
pubescent calyx that is frequently calypterate. Cordia superba ap-
pears to range more commonly in the north and somewhat more in the
interior than does the present species. The type of C. taguahyensis
is given as from the Rio Taguahy, evidently a stream in the western
end of the Federal District and possibly the one called Rio Gandi
Mirim on recent maps.

30. C. rufescens A. DC. Prodr. ix. 476 (1845). Lithocardium
rufescens Kuntze, Rev. Gen. ii. 977 (1891). C. piauhiensis Fresen. in
Mart. FI. Bras. viii. pt. 1, 9 (1857). L. piauhiense Kuntze, |. ¢. 977.

Eastern Brazil from Ceara to Parand, usually well back from the
coast.

BRAZIL. Ceari: a shrub, ca. 12 dm. tall, corolla white, Serra do Araripe,
Oct. 1838, Gardner 1782 (G, NY, K, BM, DL); Parnahyba, Serra do Rey
Aug. 1921, Luetzelburg 12504 and 12519 (BD}; Ceard, 1910, Loefgren (RS):
Prauny: shrub, Serra Branca, Jan. 1907, Ule #191 (K, BD); upper Piauby,
May, Martius (BD, authentic C. piauhiensis). Banta: Serra Assurud, 1838,
Blanchet 2821 DL, Typr of C. rufescens; NY, K, BM, BD, B ;
a oo 3607 (K, BM, DL, BB). Mrnas Gerass: indefinite,

istov: 284 Bk : Itapet
» 1887, Loefgren 415, Mus. Paul. 11271 (G). Parand: Pateisimonio,
March 14, 1914, Dusén (RS); Jaguariahyva, 740 m., Jan. 1, 1915, Dusén 16248
(G, RS); Jaguariahyva, May 20, 1914, Jonsson 397a (RS).

This species is very closely and very evidently related to C. superba,
particularly to the broad-leaved form that has been described as the
var. elliptica. As it differs from (. superba only in having a more
abundant spreading pubescence on the lower leaf-surfaces and in the
inflorescence, it may eventually be considered no more than varietally
distinct from that species. It might be noted here that Blanchet’
collection no. 2821, cited by De Candolle, is a mixture of two forms
which differ in having the upper leaf-surface glabrous in one form
decidedly scabrous in the other.

31. C. superba Cham. Linnaea iv. 474 (1829) and I. e. viii. 123
(1833); Hook. Bot. Mag. Ixxxi. t. 4888 (1855); Fresen. in Mart. Fi.
Bras. viii. pt. 1, 6, t. 3, f. 1 (1857). Lithocardium superbum Kunta
Rev. Gen. ii. 977 (1891). C. superba, var. cuncata Cham. 1. c. WV.
474. C. superba, var. elliptica Cham. 1. . iv. 474. C. grandiflot
Lindl. Bot. Reg. xviii. t. 1491 (1832), not HBK (1818). €. Blanchett
DC. Prodr. ix. 477 (1845). L. Blanchetii Kuntze, 1. c. 976 ©
Schottiana Fresen. 1. c. 7. ?C. intermedia Fresen. |. c. 8. ? inter
medium Kuntze, |. ¢. 977. ©. ; pomoeaeflora Hook. Bot. Mag. x=

STUDIES IN THE BORAGINACEAE 45

t. 5027 (1858). C. atrofusca Taub. Bot. Jahrb. xv. Beibl. 38, 12 (1893).
? C. glabra of Hook. Bot. Mag. xev. t. 5774 (1869).
Eastern Brazil, from Rio Grande do Norte to Sao Paulo.

BRAZIL. Rio Granpe vo Norte: near Natal, July 1914, Dawe (K).
NAMBUCO: small tree ca. 3.5 m. tall, Island of Itamarica, Dec. 1837,
cuore 1082 (NY, K, BM, DL); banks of Rio San cage at Villa Nova,
arch 1 co

Nova on banks of Rio San F rancisco about opposite Penedo, March 1838
Gardner 1366 (K); Olinda, Jan. 25, 1925, Pickel 619, Mus. Paul. 18128 (G).

OYAZ: ca. 18 dm. tall, corolla white, near Boa Esperanca, May 13-14,
1829, Burchell 9049 (K). Banta: Bahia, Luschnath 61 (BD); Cachoeira,
Blanchet 283 (DC); Jacobina, Blanchet 2618 (DC, tyre of C. Blanchetii; D ):
5 ;

indefinite, Blanchet 3954 (K); near Toca de Onca, June 1915, Rose & Russe
20096 (US). INAS GERAES: shru m. tall, corolla white, Cocaes, Aug.
1840, Gardner 5089 (K, BM); corolla white, Rio Novo, Sept. 1894, Schwacke
10939 (BD); Pauso Alegre, April 27, 1927, Hoehne, Mus. Paul. 19212 (G);

ca tall, Machado, Caldas, Nov. 18, 1854, Lind (RS); Ma-
chado towards Faz ista, Caldas, Dec. 20, 1864, I (RS);
inde , Dec. 17, 1868, Regnell III 906 (BD); indefinite, Claussen 408 (BD)
Rio JANErRo: corol s , Cabo Frio, Glaziou 11 , BD); vicinity
of Rio Janeiro, G 2 (K) and 13035 (K, BD, DL); cul ‘
Pontas, Bare Gla. (BD, TYPE o TO,

tto, 1 tC. ; AULO:
luza Queiroz, 1919, Gehrt 1632 (BM); Campinas, Dec. 1894, Novaes 2893,
Mus. Paul. 11274 (G); shrub, Sao Joas de Boa Vista, Dec. 15, 1875, Mosén
. at. ignot., ex hort. Kew (K, rype of C.. ipomoeaeflora);
Bras. trop., Sellow (K, BD, , DL, isoryric of C. superba v. cuneata); rye
Sellow DL, tsoryric of C. superba v. elliptica); Bras. trop., Sellow 1574 an
9305 (BD, authentic C. superba v. elliptica),
The present species is a variable one. The leaves vary in shape from
oblong-oblanceolate to elliptical-obovate, the two extremes having
n named by Chamisso respectively var. cuncata and var. elliptica.
fe leaf-margins are entire or are often very evidently serrate. The
foliage of C. superba is commonly scabrous with strigose or weakly
spreading hairs, Although the pubescence is usually not conspicuous
there are some forms which have a comparatively copious pubescence
and vary off thus toward C. rufescens, a species that differs from C.
Superba only in the conspicuous dense spreading brownish pubescence
that clothes and gives a velvety surface to the lower face of its leaves.
Fresenius attempted to limit C. superba to those forms in which the
u
In these forms the calyx-lobes remain joined together and the tip of
calyx, as a calyptera, is pushed off as the corolla expands. This

's Very impressive in its extreme development. A study of much

46 JOHNSTON

tention to the interesting but very uncertain tendency for the calyx to
open by a more or less definite apical circumscission. I have not
seen authentic material of C. Schottiana or C. intermedia and am so
forced to judge of them entirely upon their published descriptions.
Gardner’s collection no. 1/082 is cited by Nees, cf. DC. Prodr. xi. 142
(1847), under Dipteracanthus viscidulus, a habitally somewhat similar
genus of the Acanthaceae.

32. C. anabaptista Cham. Linnaea viii. 512 (1833). Lithocardium
anabaptistum Kuntze, Rev. Gen. ii. 976 (1891). C. ambigua Cham.
l. ce. 125, not Cham. (1830).

BRAZIL. InpeEFrinrrE: no locality given, Sellow (BD, TYPE).

This peculiar plant is known only from the type which was collected
by Sellow at some unknown locality in Brazil. It is obviously related
to C. taguahyensis and C. superba, but is quickly distinguished by its
very much smaller corollas.

Section Physoclada DC. Prodr. ix. 475 (1845), type-species, 0.
nodosa Lam. Physoclada (DC) Lindley, Veg. Kingd. ed. 2, 629 (1847).

A monotypic section characterized by the unique and highly com-
plex subnodal structure that is manifest in conspicuous swellings of
the stem which eventually serve as ant-domatia. The calyx is also
peculiar. It does not open to form more or less regular teeth or lobes
but is irregularly ruptured by the expanding corolla. It is not at all
accrescent nor explanate in age but if not broken apart and thrown
off by the developing fruit simply disintegrates and thus reveals its
fibrous structure. is

33. C. nodosa Lam. Tab. Encyc. i. 422 (1791); Poir. Encye. vil
43 (1806). Lithocardium nodosum Kuntze, Rev. Gen. ii. 977 (1891).
C. hirsuta Willd. Sp. Pl. i. 1076 (1798). Firensia hirsuta Raf. Sylva
Tellur. 40 (1838). C. formicarum Hoffm. ex R. & S. Syst. iv. 800
(1819). C. miranda DC. Prodr. ix. 475 (1845). L. mirandum Kunte,
lic. C. hispidissima DC. |. c. L. hispidissimum Kuntze, |. ¢. ©
nodosa, var. glabrior Fresen. in Mart. F). Bras. viii. pt. 1, 16 (1857).

. nodosa, var. hispidissima Fresen. |. c. 17. C. nodosa, var.
gustifolia Fresen. |. c. 17. C. umbrosa Spruce ex Rusby, Bull. Tort.
Bot. Cl. xxvi. 147 (1899). C. volubilis Pittier (Explor. Bot. Cuenca
de Maracaibo 41) Bol. Comer. e Indust. iv. pg. ? (1923) and Jour.
Wash. Acad. Sci. xix. 184 (1929). :

Occurring in the state of Pard and adjacent Maranhao, in outlymns
stations in Bahia and in western Brazil along the principal tributaries
of the Amazon up which it extends into northern Bolivia, east
Peru and southern Venezuela. The remainder of its range 18 ™

STUDIES IN THE BORAGINACEAE 47

northern South America where it is known only from French, Dutch
and British Guiana and from northwestern Venezuela.

; ; a, Jurud, May
, Rio Puris, Sept. 29, 1874, Traill 560

. Dard near Montalegre, Nov. 24, 1873, Trail 561 (K); Prainha, Dec. 17,
1873, Traill 562 (K); corolla white or greenish, Santarém, Aug. 1850, Spruce 775
(K); sm about 3.5 m. tall, drupes vermillion, Santarém, Mare :
pruce775 (K); ‘pecan Oct. 29, 1929, Killip & Smith 30355 (G, US) i shrub 4
.O-5.5 m. c es

€,
June 7, 1829, Burchell 9357 (K). Maranwio: Tury-asst, Oct. 13, 1923,
Snethlage 268 (BD). Banta: Cruz de Casma, Jan. 28, 1835, Luschnath 237
(BD); near Lanarro, 1832, Blanchet 995 (DC, Typx of C. hispidissima; NY,

TPE of C. miranda; K, isoryrr). Rio JANEIRO: cultivated in Quinta do Boa
Vista, Sao Christovao, Glaziou 9980 (K). InpEFtNrTE: Brazil, Sellow (BD),
Wallis (BD) and Gardner 984 (DL).

This is an extremely distinct species. It varies somewhat in the
abundance and distribution of pubescence and in the size and shape of
its leaves. This variation seems to be quite erratic and shows no
tendency at all to be geographically correlated. Various forms of the
Species are to be found in any large series of specimens from a single
region. Consequently I do not believe that these merit nomenclatorial
recognition. The plant is well known because of the peculiar sub-
nodal swellings which are a classical example of ant-domatia. The
Most exhaustive and interesting discussion of this case of myrmeco-
phy usm and of the very strange morphology of these curious subnodal
thickenings of the stem in Cordia nodosa is that given by I. S. Bailey,
Bot. Gaz. Ixxvii. 32-49 t. 6-7 (1924).

Section Pilicordia A. DC. Prodr. ix. 474 (1845), type-species C.
diversifolia Pav. Pilicordia Lindley, Veg. Kingdom ed. 2, 629 (1847).
Borellia Necker, Elem. i. 275 (1790), type-species, C. tetrandra Aubl.
Toquera Raf. Sylva Tellur. 40 (1838), type-species, C’. togueva Aubl.
Colococea Raf. Sylva Tellur. 40 (1838), type-species, C. macrophylla L.

enadena Raf. Sylva Tellur. 41 (1838), type-species, C. elliptica Sw.
Hymenesthes Miers, Trans. Linn. Soc. Bot. i. 26, t. 6B (1875), type-
Species, H. nitida Miers.

nder this section I have placed a well marked group of trees and
shrubs from the American tropics. The sectional name was originally
*pplied to a few of the species which have conspicuously sulcate

48 JOHNSTON

calyces, but ignoring this particular trait, I have expanded the original
concept to make it include a total of about 25 habitally similar species.
These are as a group characterized by having a well developed loose
paniculate inflorescence bearing commonly rather small flowers. The
calyces are either smooth or sulcate and in form very from campanu-
late to cylindrical. They become only very slightly if at all accrescent
in age and commonly become quite explanate or are soon disrupted
and reflexed by the maturing fruit. The calyx, consequently, is not
conspicuously associated with the fruit and never sheathes nor covers
any noticeable part of it. In a few species it is early circumscissile
at the base and unaltered is soon pushed off when the ovary begins to
grow. The corollas of this group are rather distinctive consisting of a
rather short tube, commonly included in the calyx, a short weakly
developed throat, 4-5 well developed ovate or quite oblong spreading
or reflexed petals and usually 4-5 evidently exserted stamens. In
most species the corolla is 6-8 mm. long, only in the coarse-flowered
Caribbean species related to C. reticulata Vahl do the corollas become
10-13 mm. long. Most of the species have decidedly drupaceous
fruits with a conspicuous mucilaginous pulp. The stone of these
fruits is depressed-globose to ovoid or rarely even obovoid in shape
and shows a decided and rather characteristic tendency to become
quite asymmetrical. In length it varies between 6 and 15 mm. The
leaves are generally broad and rather commonly turn quite dark im
drying. Most of the species are trees and the majority of them are
confined to South America where they are particularly numerous 12
Brazil and on the north coast. The group as a whole is most closely
related to that of the Old World known as the Section Myzxa. To de-
termine the exact relationship of ‘these two sections, however, would
necessitate the comparison, close study and many dissections of @
much more extended and representative suite of these Old W
plants than I have available. As the species of the section Myz4
seem to have a distinctive facies, different styles and general structures
more or less intermediate between the sections Pilicordia and E !
I have been willing, for the present at least, to believe them sectionally
separable from our American forms. This problem, however, needs
study.
Key To THE SpEciEs.

Calyx regularly costate, sometimes weakly so and the ribs obscured

by a copious loose pubescence but then the flowers crowded

into spike-like or head-like clusters on the stiff divaricate

Calpe var Wiley cto: ith
Rete willed y — pte wencurlged 10-ribbed, clothed with a

STUDIES IN THE BORAGINACEAE 49

Calyx at anthesis 2 mm. or less Re Agee coat ed,
10-14 cm. long, 5-7 em. broad, plane, the veins not con-
ne etether impressed; stems wicaser: aiiteseuely er.
Ct brachyphyla
Calyx at anthesis 3-4 mm. thick; leaves very large, flies
25 em. long and 12 em. bro ad, more or les Ai a ate, with

the debitiducy and tertiary nerves cons uously im-
anes above; stems coarse, pubescence sale very
MMOUMEY OOTORND Moo oti la ss yee 35. C. trichoclada.

tye with 70 regular although not very prominent ribs obscured

«gles nd puuetent. oc... 30. C. Chadien Manes.
i atthe rate, becoming 7 cm. broad, densely and con-
pic y brown-tomentose. 05... gs. scien 37. C. Gardneri.
Calyx Sahbed ¢ or if ribbed only very imperfectly and irregularl m
mostly strigose; branches of the inflorescence never wit

dntinke ee Pls ey sis apa Ae Ok. Gg eee ea de GE . C. toqueve.
Up surface of leaves agente with short appressed ir

ints or
decidedly heterom
Leaves glabrous fae rh adest at or above the middle,
ei fer below with erect or et cps antrorsely ascend-
stiffish or closely ap ao bee eee eS C. Sprucet.
Tiaves | lite above with short appre poe’ hairs, becaent
at or below the middle, fais: ne om tie the nerves

with short appresse rs pointing in various directions
but usually inward concentrically dveeaniie the middle
of the nerve-areolae

wer paiace of ere S green, scabrid with very short

8 rs; te: leavoe lanceolate......... 40. C. scabrifolia.
Lower surface of lave pallid, finely felty with ager but
“50 ous appressed “gs nder hairs; leaves sige
vate

ie ovsde of aoe renbed haire: 2. oo eS 41. C. sericicalyz.
Leaves green or rufous — the pubescence sparse

or —" or less tom :

Leaf broadest at or ahore the middle; Amazonian... .43. C. Ulei.
Leaf oe below the ah ; Extra-Ama zonian.

Leaves glabrous above; rare................ C. acutifolia.
Cal Leaves sista pabiesent above; common. .45. C. Sellowiana.
caves Prous oF very sparingly pubescent.
a owes or ieaiean
Leaves 1-3(—5) em teats buds obo .......46. C. ecalyculata.
i Lea 3 emi. becail: buds obl aponevian ai C. magnoliaefolia.
ves rather Shandnstty, pubescent on one or both faces.
Lea scabrous, pubescence of slender usually —

EON Vie ON SU STEW La chsh Cul eee 8. C. silvestris.

50 JOHNSTON

Leaves sepbrous beneath with stiffish hairs he peers

Leaves lanceolate, apex acute; buds obov. 49. C. ochnacea.
mane ate to seein tea obtuse er round
at apex, abruptly obovate.
50. C. scabrida.

34. C. trachyphylla Mart. (Herb. FI. Bras. pg. 246, no. 412) Flora
Regenb. xxiv. Bd. 2, Beibl. 6 (1841); Fresen. in Mart. FI. Bras. viii.
pt. 1, 10, t. 4 (1857). Lithocardiwm trachyphyllum Kuntze, Rev.
Gen. if 977 (1891).

Known only from southeast Bahia.

BRAZIL. Banta: woods near Ilhéos, Dec. 1818, Martius, Herb. Fl. Bras
412 (G, NY, K, faa ea DC, DL, 1s tte corolla white, Bahia, , Sa
1835-7, Luschnath BD); southern Bahia a, 1840, Blanchet’ 3189a (DC);
indefinite, 1840, Blechet 3189 (DL, BB).

In habit this plant is extremely suggestive of such species as 0.
scabrifolia, C. scabrida and some forms of C. Sellowiana. It is,
however, quickly separated from them by its very strongly ribbed
calyx, and is, I believe, much more closely related to C. trichoclada.
aed material I have cited forms a very uniform series.

. C. trichoclada DC. Prodr. ix. 474 (1845). Lithocardium trich0-
eck Kuntze, Rev. Gen. ii. 977 (1891). C. macrophylla Vell. Fl.
Flum. 97 (1825) and Icones ii. t. 152 (1827); not L. (1763). C. grandis
Cham. Linnaea iv. 473 (1829), not Roxb. (1824). C. Sellowiana Don.
Gen. Syst. iv. 381 (1837), not Cham. (1829). C. grandifolia DC. |. ¢.
475; Fresen. in Mart. Fl. Bras. viii. pt. 1, 10 (1857).

: Ranging from Bahia to Sao Paulo, although best known from about
Rio Janeiro.

clad ef D,
Joao ind Cline 14141 (NY), 1 14148 (K, BD, DL) ee 15274 WS, K, "B
Rio JANEIRO: Necmelbe, Miers 1845 (US); Morro Flame’ a Miers 3655 1 d
ca. 2 m. tall, along the aqueduct’ near Rio Janeiro, Aug. 16, 1825,
Burchell 799 (K); fruit yellow, arboresce at - m. ta i, Corcovado, Dec. 31, BS)
Burchell 1 , Corcovado, Jul 1873, Mosén 23 M }
Corco covado, Glasion 1505 (Ky 1) and 6 6050 cK, BD), "Lholeke 41 (DC); M a
da Babyl —_ Botafogo, July 1887, Mendonca 639 (BD); corolla ¥:
inha

Capoeinn Rio, Sept. 1876, Schwacke 11855 (BD); Morro da No®
Cintra, sing 1887, Ule 784 (BD); Praia da Fai Casaretto 1554 oe
ie Janeiro, Nov. 4, 1815 Bonke & Solander (BM); Rio Janeiro, W’ Tiete,
DL) and 7 Tweedie i 312 (K). § o Pauto: small tree, corolla white, Alto
aed 19, 1901, Loefgren 5843, Mus Poul. 11312 (G); tree, corolla white 7"
Biologica, Alto da Serra, Sept. 27, 1922 . Paul. 7979 (G).. ie
DEFINITE erid. Sellow (K, BM. DL. a7 K. 398 (BD, authente

grandis) ce Li74_B398 8 (BD).

A very distinct species quickly distinguished by its large bulla ee
leaves, coarse usually very shaggy stems, spreading inflorescence

STUDIES IN THE BORAGINACEAE 51

coarse strongly ribbed calyces. It is very erratic in its distribution
being known only from Bahia, south central Minas Geraes, vicinity
of Rio Janeiro and eastern Sao Paulo. F resenius, |. ¢., reports it from
the Rio Xipoté, Minas Geraes. This station is to the southwest of
Ouro Preto and from the same region as Joao Gomez where Glaziou
obtained it. The plant’ must be rather common about Rio Janeiro.
A study of the variation exhibited by the many collections from this
last locality shows conclusively that C. trichoclada, from Bahia, is com-
pletely indistinguishable from the more southern C. grandifolia. The
two species are accordingly here united under the prior name.

36. C. Chamissoniana Don, Gen. Syst. iv. 381 (1837). C. latifolia
Cham. Linnaea viii. 126 (1833), not Roxb. (1814). C. platyphylla
Steud, Nom. ed. 2, i. 418 (1840); Fresen. in Mart. Fl. Bras. viii. pt.
15 (1857). Lithocardium platyphyllum Kuntze, Rev. Gen. ii. 977
1891).

Known only from the type collected at an undesignated locality
in Brazil.

BRAZIL: locality unknown, Sellow (BD, type).

A close relative of C. trichoclada from which it is separated by having
the slightly smaller calyces with a somewhat looser pubescence and
only obscure rather than conspicuous ribs. In addition, the flowers of
C. Chamissoniana seem to be more definitely spicate and the pubes-
cence of the proportionately broader leaves scantier than in its relative.
The source of the type and only known collection of the species is not
on record,

37. C. Gardneri, sp. nov., fruticosa 2-2.5 m. alta; caulibus sulcatis
cum pilis gracilibus rigidusculis brunneis flexuosis abundantibus laxe
tomentosis; foliis alternis elliptico-ovatis vel lanceolato-oblongis
1-2 dm. longis 5-7 em. latis firmis pinnate nervatis, supra fuscescenti-
bus pilis rigidusculis antrorse adpressis brunneis plus minusve ves-
"tis, subtus pallidioribus nervis prominentibus rugosis cum pilis

rotundatis, margine integris sinuolatis, nervis primariis 9-10-jugatis
Simplicibus ascendentibus per nervos secondarios tenues subrectos

—— 15 mm longo stricto densissime brunneo-villoso; inflores-
ike “aminali nigida divaricate spa ue breviramea brunneo-

52 JOHNSTON

vel capitula densa 1.5-2 cm. diametro formantibus; calycibus cam-
panulatis 4-5 mm. longis 3.5 mm. crassis cum pilis abundantibus laxe
adpressis flexuosis gracilibus brunneo-vestitis 10-costatis (costis per
pilos obscuratis) basi rotundis, dentibus 5 late triangularibus 1 mm.
longis; corolla alba (?) 6-7 mm. longa, tubo cylindrico ca. 3 mm. longo,
lobis oblongis 5 recurvatis ca. 3 mm. longis; filamentis exsertis usque
ad 5 mm. longis ad basem versus villosis; ovario glaberrimo in stylum
subulatum sparse adpresseque villosum abrupte contracto; fructu
ignoto.

Southeast parts of Minas Geraes.

BRAZIL. Minas Gerass: Arraial das Mercés, Oct. 1840, Gardner 5126
(Kew, type); Rio Novo, 1896, Araujo, in Herb. Schwacke 11974 (BD).

This species is very distinct and, because of the spicate arrangement
of its flowers, rather suggestive of some of the Verbenaccac. Its dense
indument serves to separate it from its immediate relatives.

38. C. toqueve Aubl. Hist. Pl. Guian. Fr. i. 228, t. 90 (1774).
Lithocardium toqueve Kuntze, Rev. Gen. ii. 977 (1891). C. heterophylla
Poir. Dict. Sci. Nat. x. 409 (1818) ; Willd. ex R. & S. Syst. iv. 800 (1819);
Cham. Linnaea iv. 480 (1829). L. heterophyllum Kuntze, |. ¢ 97.
? C. scandens Poir. 1. c. 410. C. pubescens Willd. ex R. & S. 1. ¢. 800;
Cham. |. c. iv. 479 (1829) and 1. ¢. viii. 127 (1833). L. pubescens
Kuntze, |. ¢. 977. 7, oquera tomentosa Raf. Sylva Tellur. 40 (1838).
C. hebecarpa DC. Prodr. ix. 488 (1845). L. hebecarpum Kuntze, |. ¢

Occurring in eastern Brazil, northern South America and along the
Andes to Bolivia.

BRAZIL. Pern )
March 1924, Pickel 620, Mus. Paul. 18120 SP). Banta: indefinite, Blan
44 and 1834 (DC; rvpps of C. hebecarpa), 131 (BM), 191 (BM, DL), 49% (Dt
and 1839 (K, BD, BB); Injule Itaparica near Bahia, Casaretto 2023 (DL).
INDEFINITE: Brazil, Sellow 31 (BD).

I have seen the type of C. toqueve in the British Museum and the
specimen of C, heterophylla in the Willdenow Herbarium.
specimens came from French Guiana and are clearly conspecific.
type of C. pubescens came from near Caracas and like other Venezuela”
collections of the species differs from the Guianan plants merely “
having a slightly less copious pubescence on the foliage. The
zilian material, however, varies somewhat in the abundance of pubes
cence and clearly demonstrates that the very slight difference betwe?
C. to and C. pubescens is obliterated by intermediate slat
Cordia hebecarpa, from Brazil, is clearly synonymous with @. # by
: ta toqueve is distinguished among all the species of our a
its decidedly heteromorphic foliage. The foliage on a single b

STUDIES IN THE BORAGINACEAE 53

consists of suborbicular or broadly ovate leaves which alternate with
or are frequently opposite much larger more elongate elliptical, nar-
rowly ovate or even lance-oblong ones. In tropical America there are
several species with a very similar leaf-heteromorphism. Some of
these have passed as C. heterophylla. That name, however, belongs
in the synonymy of C. toqueve which in turn is properly applied only
to the South American plant which occurs in eastern Brazil, the
Guianas and Venezuela and apparently south along the Andes to
Bolivia. This plant is quickly and decisively distinguished from all
of its relatives with strongly heteromorphic foliage by having a very
hairy ovary, very pubescent fruit and denser pubescence on the leaves.
In northern South America it occurs with only one close relative,
C. ierensis Britton, which is known only from Surinam, Trinidad,
Venezuela and Colombia and which differs from our plant in its very
fine inconspicuous pubescence, perhaps less firm leaves and decidedly
glabrous ovary and fruit. The plants of the West Indies and Central
America which have been called C. heterophylla are referable to C.
suleata DC. and to C. macrophylla L. As already intimated these have
decidedly glabrous fruit and ovaries.
- Sprucei Mez. Bot. Jahrb. xii. 549 (1890). Lithocardium
Spruceit Kuntze, Rev. Gen. ij. 977 (1891).
Known only from the mouth of the Rio Negro and in northern Peru.
BRAZIL. Amazonas: Barra do Rio Negro, Dec.—March 1850-1, Spruce
1019 (BD, part of TYPE); vicinity of Barra, Dec.—March, 1850-51, Spruce sine
‘ K, BM); tree 4.5 m. tall, corolla white, Barra to Matiriho, Jan. 1851

- Lorero: tree 4.5-6 m. tall, Iquitos, ca. 100 m., Aug. 1929, Killip
& Smith snove (UB) 5-6 m. tall, Iquitos, ca , Aug

type-specimen of C. Sprucei. It is given as bearing Spruce’s number
- Although the first and best set of Spruce’s collections is at Kew
there is no collection of a Cordia there bearing number 1019. Con-
sulting Spruce’s manuscripts and his field-books, which form a part of
the Kew Library, I found that Spruce applied the number in question
to a plant collected between January and September near Santarém

at was determined as Gustavia brasiliana DC. I believe that the
type of C. Sprucei should have properly borne the number 1234!

70. C. scabrifolia A. DC. Prodr. ix. 485 (1845).

Known only from British Guiana and from scattered stations in
the Amazon Basin in Brazil and southeastern Peru.

RAZIL. Amazonas: tree 6 m. tall lla white, Rio Negro, gapé above
Cabuquena, Dec. 1851, Navas 1942 (kK. 'BM). Pan: ‘hae or small tree

54 JOHNSTON

m. tall, near Pard, Nov. 8, 1819, Burchell 9720 (K); Prainha, Nov. 18,
l . Goyaz: small tree 5.5 m. tall, near Porto Real, Dee.
17, 1828, Burchell 8452 (K). Inperinrre: Brazil, Newman (DL).

BRITISH GUIANA: “Upper Essequibo,”’ Schomburgk 911 (BB, TyPE;

, BM, BD, tsorypss).

PERU. Mapre pe Dros: tree 8-15 m. tall, corolla white, Seringal Auristella,
Rio Acre, Aug. 1911, Ule 9717 (K, BD).

This is a very well marked species known only from the collections
cited above. It is quickly distinguished by the character and distribu-
tion of its very scanty short pubescence. The leaves, which dry a
brown, have on the lower surface sparse, appressed, minute hairs that
do not point in one direction but, developing on the slightly prominent
veins and veinlets, are directed towards the center of the adjacent
vein-areoles. The nature and distribution of these minute stout hairs
on the leaves is quite constant and hence very characteristic; once
they have been studied under a lens and their peculiarities noted the
species becomes quickly distinguished by them thereafter. Another
peculiarity of the pubescence in this species is the suprabasal attach-
ment of the hairs, particularly on the calyx. The calyx is antrorsely
strigose, the hairs being abruptly bent at the point of attachment and
very closely appressed. The suprabasal attachment has apparently
resulted from the development of a callous or possibly glandula
thickening below the attachment and in line with the main body of
the hair. A suggestion of this condition is occasionally found in other
species, notably C. sericicalyz.

. 41..C. sericicalyx A. DC. Prodr. ix. 485 (1845). Lithoeardium
sericicalyx Kuntze, Rev. Gen. ii. 977 (1891). C. bicolor A. DC. Prodr.
ix. 485 (1845). L. bicolor Kuntze, l. c. 976. ? C. dichotoma Klotasch
ex Schomburgk, Fauna u. Fl. Brit. Guiana 1084 (1848), nomen. L.
Lockartii Kuntze, |. c. 438. C. Lockartii Kuntze, |. c. 438, in synonymy:
C. trichostyla Pittier, Contr. U. S. Nat. Herb. xviii. 252, f. 102 (1917):
C. carnosa Rusby, Three Hundred N. Sp. So. Am. Pl. 104 (1920). ©:
corvacea Killip, Jour. Wash. Acad. Sci. xvii. 329 (1927).

urring in Amazon headwaters of western Bolivia, in scattered
stations south and east of the Amazon Basin in Brazil, in northern
: uth America from Surinam to Colombia, in Panama and Guatemala
Hs are America and on Trinidad and St. Vincent in the West
ndles,

i Feb. 1910, Ue
.Ceari: a large shrub near Crato, Nov. 1838, Gardner 1781 (B
NEIRO: indefinite, 1844, Wi (RS). M a ae
— Anna da Chapada, Oct. 6, 1902, Malme 2455 (RS); small tree, Sarre
us: do Chapada, May 26, 1903, Malme 2456a (RS); indefinite, Smith

STUDIES IN THE BORAGINACEAE 55

Once recognized this species can be quickly distinguished from all
other American species by the character of the pubescence on its foliage.
The leaves are green above although somewhat dulled by a copious
closely appressed pubescence in which the slightly antrorse hairs, on
either side of the midrib, are directed towards the leaf-margin. The
under leaf-surface is pallid with a very dense thin felty indument.
This pubescence is formed of fine closely appressed slender hairs,
which spring laterally from the veins and are directed inward towards
the center of the vein-areoles, which are thus covered with a pallid
pubescence. The veins and veinlets themselves are very sparsely
clothed and consequently are very evident because of the contrast of
their dark color against the gray of the hair-covered areoles.

The type of C. sericicalyx came from British Guiana. The ovary
in the type collection is glabrous but otherwise the plant seems to be
quite similar to the plants that have been described as C. bicolor, C.

ckartii, C. trichostyla, C. carnosa and C. coriacea. In the species as
Thave defined it the ovary is prevailingly pubescent. The only plants
with glabrous ovaries that I have seen are those of the type-collection
of C. sericicalyx and those collected by Gardner in Cearé. The ovaries
In the collections made by Malme, which I have cited above, are very
sparingly pubescent. As the vegetative characters of the plants that
Thave referred to C. sericicalyx are so constant and unique, and since
they seem to characterize a species with a natural and very credible,
albeit disrupted, distribution I feel that in this case the presence or
absence of pubescence on the ovary should be ignored in defining the
Species.

42. C. tetrandra Aubl. Hist. Pl. Guian. Fr. i. 222, t. 87 (1775).
Lithocardium tetrandrum Kuntze, Rev. Gen. ii. 976 ( 1891). C. cordi-
folia HBK: Nov. Gen. et Sp. iii, 70 (1818); Willd. ex R. & S. Syst. iv.
800 (1819). J, cordifolium Kuntze, |. c. 976. C. muneco HBK. 1. ec.
ore 1895). L. memory Kuntze, |. c. 977. Borellia aspera Raf.
Sylva Tellur. 41 (1838). C. umbraculifera DC. Prodr. ix. 484 (1845);
Fresen. in Mart. FI. Bras. viii. pt. 1, 16 (1857). L. umbraculiferum
Kuntze, |. e. 977,

ortheast coast of Brazil, in northern South America and along the
ae to Bolivia. Occasionally in cultivation beyond its natural

BRAZIL, Pari: : . tall, mouth of the
Sg Cag a to er lg hog ge al

. inqcaver north of Campinha, Pard, Dec. 30, 1829, Burchell 9968 (K).
March 18 tenn r. Allemao 1124 (US). PERNaMBuCo: egos Ym,
tree, eo , * ickel 620, Mus. Paul. 18120 (SP). Rio JANEIRO: cultivated

olla white, Passeio Publico, Rio Janeiro, Glaziou 5949 (K, BD);

56 JOHNSTON

cultivated tree, corolla white, Quinta da Béa Vista, Rio Janeiro, Glaziou 9534
(NY, K, BD).

Among the South American species of this section C. tetrandra is
quickly distinguished by its well developed petioles. The large non-
heteromorphic leaves are usually coriaceous, oblong and conspicuously
bicolored and are frequently cordate at the base. The upper surface
of the leaves is green although provided with an inconspicuous indu-
ment of short slender appressed hairs. The lower surface is rufous
being densely and evidently pubescent and usually more or less vel-
vety. The petioles on the larger leaves are 2 cm. long or more, about
twice the length of those in the related species. The ovary is glabrous
and is commonly provided with a rather characteristic glandular or
callous ring above its middle. The stamens, almost always 5in number,
are either glabrous or villous. The inflorescence frequently becomes
extremely broad and corymbose. The species is well marked and
very distinctive. :

43. C. Ulei, sp. nov., arborescens 5-12 m. alta; ramulis inconspicue
sulcatis hirsutulis brunnescentibus; foliis firmis penninervatis 5-1
mm. longe petiolatis obovato-oblongis vel oblongo-oblanceolatis 14-22
cm. longis 4-10 cm. latis saepissime supra medium latioribus saltem
infra medium ad basem versus contractis, basi cuneatis vel raritet
immam ad basem abrupte obtusis vel rotundatis, apice saepe rotun-
datis vel obtusis et abrupte (5-10 mm. longe) acuminatis, margine
integerrimis plus minusve inconspicueque revolutis, supra glabris a
tantum in costa et in nervis primariis sparse inconspicue pilifers,
subtus pallidioribus pubescentibus, per nervos secundarios tenues
subrectos transverse conjunctos verrucosis, nervis primariis 5-6-jugatis
conspicuis; inflorescentia terminali laxe paniculatimque e cincinnls
terminalibus composita pilis rigidulis curvatis ascendentibus brunnes-
centibus ca. 0.7 mm. longis vestita breviter pedunculata foliis vi
superante; floribus sessilibus; calyce poculiformi 4-5 mm. longo ©
2 mm. crasso plus minusve strigoso; corolla alba fere ad 8 mm. long
5-mera, tubo cylindrico, filamentis gracilibus exsertis ad basem vers's
villosis; ovario glaberrimo; fructu ignoto.

BRAZIL. Acre: Seringal S. Francisco, Rio Acre, March 1911, Ule 976
(Herb. Berol. rypz; Kew, ISOTYPE).

This species is probably most closely related to C. Spruce of ee
Manfos region, but differs in having a glabrous rather than a deci
pubescent ovary and fruit. It is probably also related to C. Sellowan!
of the coastal states of southeastern Brazil from which it differs
having the leaves not only broadest above the middle but larger, with

.

STUDIES IN THE BORAGINACEAE oy

a somewhat sparser pubescence and drying commonly much darker.
The calyx in flower is scantily covered with appressed hairs whereas
in C. Sellowiana it is densely and completely clothed. In the typical
form of C. Ulei the upper surface of the leaf is glabrous except along
the midrib and the principal nerves and hence differs in this trait from
that of C. Sellowiana in which the upper surface of the foliage is
always pubescent. Material from along the Ucayale and the Amazon
rivers in eastern Peru, however, represent a variety of C. Ulei with the
leaves pubescent on both surfaces. As this plant may be expected
along the upper Amazon in western Brazil it is placed on record here
as,—

C. Ulei, var. ucayaliensis, var. nov., a varietate genuina differt foliis
utrinque pubescentibus.

ERU. Lorero: corolla white, tree 6 m. tall, branched 3 m. above the
ground, trunk 21 cm. thick, flood-free forest, Yarina Cocha, 155 m. alt., Oct.

26, 1925, Tessmann 5457 (Herb. Berol., Type); Caballo Cocha on the Amazon
River, Aug. 1929, Williams 2333 (G).

The upper surface on the leaves in this variety is evidently strigose
and consequently scabrid. A collection by Poeppig, no. 2/96, in the
herbarium at Berlin, no doubt came from the Peruvian montafia and
1S very probably also referable here.

34. C. acutifolia Fresen. in Mart. Fl. Bras. viii. pt. 1. 11 (1857).
Lithocardium acutifolium Kuntze, Rev. Gen. ii. 976 (1891).

BRAZIL: no locality given, Sellow 744 and 1203 (BD, Typr).

A peculiar plant known only from material collected by Sellow at

an unknown locality in Brazil. The species is closely related to C.

Sellowiana, particularly to the form of that species which was de-
scribed as C. obscura. It is distinguished, however, by having the
Upper leaf-surfaces glabrous and the leaf-outline rather different from
any T have seen in C. Sellowiana and much more suggestive of that of
¢. scabrifolia,

45. C. Sellowiana Cham. Linnaea iv. 478 (1829). Lithocardium
Sellowianum Kuntze, Rev. Gen. ii. 977 (1891). ©. obscura Cham.
I. e. 480, L. obscurum Kuntze, |. ¢. 977. C. obscura, var. glabrior
Cham. 1. ¢. viii, 128 (1833). C. obscura, var. tomentosa Cham. I. c. viii.
128 (1833). C. brachypoda DC. Prodr. ix. 487 (1845). L. brachypodum
Kintze, l. ¢. 976. C. obscura, var. magnifolia Fresen. in Mart. FI.

fas. Vili. pt. 1, 15 (1857).

nging in the coastal states of Brazil from Bahia to Parand, in the
south extending inland to northeastern Paraguay.

58 JOHNSTON

burgo and Conego, J K

(K); 2863 (K) and 5950 (K, BD); Organ Mts., Wilkes Exped. (U8); ee

4 m., road betw. Frechal and Magé, Burchell 2689 (K); Rio Magé tow ;

Faz. da Lagoa, Burchell 2751 (K); shrub ca. 2.5 m. tall, near Magé, Mare
5. © .

near church of N. 8. da Gloria, April 1827, Burchell 4685 (K); convent of be
ento, May 4, 1815, Bowie Cunningham (B ar Sado Bernardo, an.
1827, Burchell 3971 (K); Butantan, J
c. 2, 1918, Hoehne, Mus. Paul. 2
tert 12.

RS); Paranagud, March 7, 1914, Dusén 1a (G, RS). INDEFINITE: Bras.

merid. Sellow (K, BM DL); Bras. trop., Sellow 399 (BD, authentic C. oboe

ras., Se (B

1666 (BD, pea Re Sellowiana); Bras. merid., Sellow 1566-c659 (BD);
06

: Sierra de Amambay, Sept Rojas in Hassler 10595 (BM, BD);

Sierra de Amambay, Oct., Rojas in Hassler 11176 (Hass)

This appears to be especially common, in any case has been pee
frequently collected, in southern Minas Geraes, Rio Janeiro
eastern Sao Paulo. The species is variable in pubescence. The
common forms are those which are copiously pubescent with s
ing hairs and so more or less definitely tomentose on the lower surface
of the leaves. This form has been described under the names ©
Sellowiana, C. obscura var. tomentosa and C’. brachypoda. The ere
extreme in pubescence is the comparatively uncommon form at”
has been described as C. obscura and C. obscura var. glabrior. oe eo
a pubescence that is sparse and closely appressed, so decidedly re
that its leaves might at a glance be taken for glabrous. This gre
leaved form with its sparse appressed pubescence is completely 0”
nected by intermediates with the more common one that is tomentost
with abundant spreading hairs. The two extreme forms occur 12

STUDIES IN THE BORAGINACEAE 59

same region and I am inclined to believe them merely phases of a
single species.

Among the species in our area (. Sellowiana is most apt to be con-
fused with C. toqueve since both species grow in Bahia. Cordia toqueve,
however, is quickly distinguished by its pubescent rather than glab-
rous ovary and fruit and by its very decidedly heteromorphic leaves.
The leaves of C. Scllowiana, as in other species of the section, are often
slightly heteromorphic but never so pronouncedly so as is commonly
encountered in C. toqueve. Although ranging entirely outside and
southeast of the Amazon Basin C. Sellowiana in appearance approaches

- Ulei from near the Peruvian-Brazilian frontier and C. Sprucei from
the lower Rio N egro. Both of these Amazonian species have leaves
which are broadest at or above the middle. The leaves of C. Sellowiana
are ovate to oblong-lanceolate and almost universally broadest below
the middle. Of all the many collections of the present species seen by
me only two have the leaves clearly broadest above the middle and
these have the dense pubescence of the typical form which is very
different from the scanty indument of the two Amazonian species
mentioned.

46. C. ecalyculata Vell. Fl. Flum. 96 (1825) and Icones ii. t. 149
(1827). ? C. digynia Vell. 1. c. 97 and 1. c. t. 153. C. salicifolia Cham.
Linnaea vi. 481 (1829) and 1. c. viii. 129 (1833). Lithocardium salici-
folium Kuntze, Rev. Gen. ii. 977 (1891). C. leptocaula Fresen. in
Mart. Fl. Bras. vill. pt. 1, 14 (1857). L. leptocaulon Kuntze, |. c. 977.
¢. coffecides Warm. Kjoeb. Vidensk. Meddel. 1867: 4 and 44, fig. 3
(1868). Patagonula Glaziovii Mez, Bot. Jahrb. xii. Beibl. 27, pg. 17
(1890). C. Glaziovii Taub. Bot. Jahrb. xv. Beibl. 38, pg. 13 (1893).

Ranging in the coastal states of Brazil from southern Minas Geraes
to Rio Grande do Sul and westward through northeastern Argentina to
Paraguay,

Corl ); Rio Novo, Oct. 29, 1889, raujo 14827 (US); Parahybunda,
Fo 1840, Gardner 5038 (K, BM); indefinite, Claussen (K).’ R10 JANEIRO:

). Rr
M e Novo Friburgo, Glaziou 18384 (K, BD, DL); Serra do Picu,
Fat: 1, 1886, Schenck 1481° (BD). Sio Pacvo; Sov0
5, 1875, Mosén 3445 (RS); Serra de Caracol, Dec. 10, 1875, Mosén 4307

): Biol ose” 3445 (RS); Serr
(G): O, Ological Station, Alto do Serra, Feb. 3, 1922, Hoehne, Mus. Paul. 57.
ampinas, Nov. 15, 1896, Novaes 939, Mus. Paul. 11283 (G); near Pico
8)

a ards Sio Paulo, Dec. 22, 1826, Burchell 3741 and
Bee (K); near Cubatio at Ri nd Pair , Dec: 10, 1826, Burchell K);
tan, Jan. 1921, Gehrt, Mus. Paul. 5304 (G); Loreto, Vecchi, Mus. Paul

tan an
18137 (G): St Em _ : 30. og deny
? St. Ernestina, Sept. 24, 1920, Gehrt, Mus. Paul. 4520 (G); Piraci:

Dee ultemans, Mus. Paul 11278 (SP). Parani: Therezina, Jan. 21, 1911,
cin ta SANTA

1910, Dusén 11309; (RS).
CarmaRiNa: Herval, June 7, 1911 1888 (RS). R1o GRANDE Do Suz:

60 JOHNSTON

Tristeza near Porto Alegre, Dec. 11, 1901, Malme 817 (RS); Porto Alegre to
Menino Deus, Feb. 24, 1902, Malme 817a (RS); Excol. Santo Angelo, Jan. 19,
1893, Lindman (RS); Rio Grande do Sul, Ihering 323, Mus. Paul. 18122 (G).

DEFINITE: Brazil, Sellow (BD, Type of C. salicifolia; K, DL, tsorypss), 272,
1567, 5056, 25II + 5116 and 5788-241 (BD); Glaziou 13476 (BD, rrpx of P.
Glaziovii; K, DL, tsorypEs); Claussen 406 (BB); Pohl, Herb. Mus. Vind. 1924

. P - fs
PARAGUAY: region of the Alta Parandé, Fiebrig 5606 (BD, Hass) and
5865 (K, BM, Hass); near Igatimi, Sierra Maracayti, Nov., Hassler 5545 (BD,
DL, Hass, BB); Cord. de Altos, Nov. 29, 1902, Fiebrig 518 (K, BD, DL, Hass);
; 08 ;

Ss, ; Paraguay Central, 1307, Hassler 1505 (BM, Hass, BB); Villa Rie
Jan. 1929, Jérgensen 3470 (G); Villa Encarnacion, Bettfreund 62 (BD).

Cordia ecalyculata is closely related to C. magnoliaefolia, differing
in having smaller narrower and usually less coriaceous leaves as well
as less elongate obovoid, rather than eylindrical-obovoid buds.
Typical C. ecalyculata comes from the region about Rio Janeiro and
appears to represent that phase of the species, as here defined, which
has rather broad firm leaves. This typical form with its broadish
coriaceous leaves I know only from the northern part of the species
range where the names C.. Glaziovii and C. digyna have been applied to
it. The much more frequently collected and much more widely dis
tributed form with less firm, lanceolate leaves is that which has bee?
called C. salicifolia and C. coffeoides. Cordia leptocaula seems to bea
form of C. salicifolia distinguished by its conspicuously villous anther
As I have accepted C. ecalyculata it appears to be a shrub or tree be-
coming at least 6 m. in height. The flowers are white or ochroleucols
and have the filaments very variable as to the amount of pubescen?
upon them. The fruit is ovoid or obovoid and becomes fleshy and red.
The pulp is said to become mucilaginous. The names “ Cha de
Bugre” and “Cha de Frade” have been applied to the Brazilian plants
and “Gomita” to the Argentine. : :

47. C. magnoliaefolia Cham. Linnaea iv. 476 (1829). Lithocardiu
magnoliaefolium Kuntze, Rev. Gen. ii. 977 (1891). C. obliqua V
Fl. Flum. 97 (1825), not Willd. (1798). C. obligua Vell. Ieones 1 :
150 (1827), lapsu calami. C. diospyrifolia Cham. |. c. 477. L. diosP"
rifolium Kuntze, |. ¢.

Ranging from southern Minas Geraes to Paranda.

BRAZIL. M BD); ben
Rio Paro and Cero, Caldas 15 181 fag 1908 3); Coo

STUDIES IN THE BORAGINACEAE 61

Caldas, Nov. 30 1873, Mosén 959 (RS); Caconde, Regnell II 202 (RS); Caldas,
Dee. 1862, Regnell II 202 (RS); indefinite, 1845, Widgren (RS). Rrio JANErRo:
an Mts., Feb. 1838, Miers (BM); Valley of Catumbf, Jan. 31, 1826, Bur-

me i8 “ou 5952 Me

C 96 : i; ;

1896, Ule 3845 (BD); indefinite, Glaziou 1057 (K) and 7780 (K, BD). Sio
: . 14, 1913, Brade 7467 (G, BD); Sao Paulo, 1815, Bowie

& Cunningham (BM). Paran&: near Morretes, Jan. 4, 1914, Dusén 1 4287 (G);

. Marumb n. 4, 1914, Dusén 1 Fac.
1912, Dusén 14058 (RS); Valle Grande, Sept. 16, 1915, Dusén (RS). InpEFI-
: il, Pohl, Herb. Mus. Vind. 1746 (K, BD); Brazil, Sellow 724, 1569,
1570 and 5358-489 (BD, type suite of C. magnoliaefolia); Sellow, (K, DL).

>
=
S
ot
o
&
es

This species is characterized by its large, coriaceous, glabrous
leaves which are oblanceolate or elliptical-oblanceolate in outline.
There appear to be two phases of the species, one with rather strictly
branched and evidently ped late inflorescences and rather narrowly
oblanceolate leaves, and the other with more laxly branched very
shortly pedunculate inflorescences and much broader leaves. The
second form comes chiefly from Minas Geraes and has been described
as C. diospyrifolia. Although rather distinct in their extremes the
two forms are well connected by intermediates.

48. C. silvestris Fresen. in Mart. Fl. Bras. viii. pt. 1, 12 (1857).
Lithocardium silvestre Kuntze, Rev. Gen. ii. 977 (1891).

: An apparently rare plant which occurs from Bahia to Santa Catha-
es “a reappears in a doubtful form along the Amazon in western

Taz.

IL. Amazonas: Marary Jurud, Sept. 1900, Ule 5191 (BD); Barra,
Banta: F. Santa Anna, 1850, Blanchet (BB);

, . 16. ers (BM);
} : Glaziou 41 K, BD);
is; Theresopolis, Dec. 1896, (le 4298 isp) Parana: Rio Canetti Dec. 28
1, Dusén 13649 (RS). Santa CATHARINA: Itajahy, Mueller 444 (K).
The leaves in C’. silvestris are broadly oblanceolate to obovate or
el 'ptical and have the under surface with slender appressed hairs, not
ngid enough to make the surface noticeably scabrid. It is closely
elated to C. ecalyculata from some forms of which it differs in scarcely
more than the pubescence on the foliage. Usually, however, it is
af Separated from C. ecalyculata by its proportionately broader
I have referred to C. silvestris with much hesitation and doubt
ee plants collected in the upper Amazon Basin. These agree with
*Pical C. silvestris in general habit and leaf-outline but tend to have
es stiffish hairs on the leaves and too abundant hairs on the calyx.
; 18 Amazonian plant is involved in a series of problems concerning
uch other Amazon plants as C. scabrida, C. Sprucei and C. Ulei and

62 JOHNSTON

its final disposition, consequently, must await the time when more
numerous collections from the region are available for study. It is
possible that it may merit recognition as a distinct species.

49..C. ochnacea DC. Prodr. ix. 485 (1845). Lithocardiwm ochnaceum
Kuntze, Rev. Gen. ii. 977 (1891). ? C. brachytricha Fresen. in Mart.
Fl. Bras. viii. pt. 1, 13 (1857).

A rare species known only from Pernambuco and Rio Janeiro.

Martius (BD, ISOTYPE of C. brachytricha). Rio JANEIRO: near Santa —_
° . . D, DL)

Beyrich (BD);

18583 (K, BD); near Rio Janeiro, Feb. 23, 1815, Bowie & Cuntingian (BM);
Rio Janeiro, Beyric . INbEFiniTE: Brazil, Lhotsky (DC, Trex of C.
ochnacea), Luschnath (BD) and Sellow (BM, DL, RS).

A species related to C. silvestris from which it differs in its somewhat
firmer usually decidedly lanceolate leaves, which are decidedly seab-
rous with stiffish closely appressed hairs below. In C. silvestris the
hairs on the leaves are more slender and by no means so stiff. Glaziou’s
collections, above cited, are quite like the type of C. ochnacea, which
no doubt also came from the vicinity of Rio Janeiro.

50. €. scabrida Mart. ex Fresen. in Mart. Fl. Bras. viii. pt. 1, IJ
(1857). Lithocardium scabridum Kuntze, Rev. Gen. ii. 439 (1891):

Known with certainty only from the lower Amazon, principally
in the vicinity of Pard.

Part: near Santarém, Martius (K BD, isotypes); shrub 3 m.

s ay, 1929,

b 2-3 m. tall, corolla white, Ilha do Mosqueir

near Pard, Nov. 1929, Killip & Smith 30434 (G, US); shrub 2-2.5 m. tall

fruit red, Ilha do Mosqueiro, Nov. 1929, Killip & Smith 30425 (G, US);
Braganga Railroad east of Pard, June 1920, Goeldi 1103 (G, NY, US, DL).

This species is known definitely only from the collections cited
above. It is, however, related to certain forms of French Guiana ane
consequently may be eventually identified with some of the very
obscure and poorly understood species of that country. The very
sparsely strigose calyces and the scabrid foliage set off the plant as
we now know it rather well among the related Amazonian specie
As with the other species from that area, however, material from
more stations is needed before its limits can be defined with any
certainty. I suspect C. scabrida will be found to intergrade W!
some of the other species known from higher up the Amazon.

EXxcLupep OR UNIDENTIFIED SPECIES.

Cordia amplifolia Mez, Bot. Jahrb. xii. 538 (1890), not A.
(1845). Lithocardium Mezianum Kuntze, Rev. Gen. ii. 976 (1891).

STUDIES IN THE BORAGINACEAE 63

C. Meziana Giirke in E. & P. Nat. Pflanzenf. iv. Abt. 3a, 84 (1893).—
Described from material lacking both flowers and fruit. The type was
collected by Martius in woods near Pedra Branca, a locality about
300 km. northwest from the city of Bahia. It is described as having
subglabrous obovate or oblong membranaceous entire-margined leaves
about 36 cm. long and 11.5 cm. wide. Mez suggests a relationship
with C. magnoliaefolia. I have a suspicion, however, that the plant
may not be a Cordia.

ordia asperrima Spreng. Syst. i. 649 (1825); Cham. Linnaea iv.
492 (1829).—Said to be a Labiate, Hyptis membranacea Benth.

been collected in the Serra da Estrella, state of Rio Janeiro. The ty
number, Glaziou 1073, is given as being represented in the herbaria at
Kew, Berlin and Geneva. As I did not discover this collection among
the sheets of Cordia at these herbaria I suspect that it may not
represent a species of Cordia and that its true affinities having been
recognized the specimens have been removed from the genus-covers.
: ia fragrans Fresen. in Mart. Fl. Bras. viii. pt. 1, 28 (1857),
in discussion.— According to Fresenius this name was found in the
herbarium at Vienna where it is said to occur on material referable
to Aegiphila obducta Vell.
Cordia flavescens Aubl. Hist. Pl. Guian. Fr. i. 226, t. 89 (1775).
Lithocardium flavescens Kuntze, Rev. Gen. ii. 977 (1891). C. sarmen-
a Lam. Tab. Encyc. i. 422 (1791). C. echitoides Lam. ex Dietr.
Synop. i. 612 (1839), in synonymy. Firensia Scop. Introd. 157 (1777).
Bab, have seen the authentic material of this species preserved at the
British Museum and at the Riks Museum in Stockholm. The species
'S not a Cordia but apparently a member of some genus of the Laura-
ceae. Although I went through the entire collection of American
Lauraceae at the British Museum I was quite unable to find a match
~ +. Dr. Gleason has also examined the plant and is likewise in-
clined to believe it might be one of the Lawraceae but was also unable
0 find a place for it in that family. Although clearly to be excluded
sesh the genus Cordia the identity of C. flavescens unfortunately
remains doubtful.
Vag i® Macrocephala (Desv.) HBK. Nov. Gen. et Sp. iii. 77 (1818).
V mait macrocephala Desv. Jour. de Bot. i. 273-4 (1808).—When
* Macrocephala was originally published it was reported as having
B n collected by Dombey in Brazil. As Dombey did not collect in
Tazil and as the species:is known in South America only from the

64 JOHNSTON

west coast where Dombey did collect, it is evident that the species
was reported from Brazil only through an error, probably clerical.

Cordia Neowediana DC. Prodr. ix. 498 (1845); Fresen. in Mart.
Fl. Bras. viii. pt. i, 23 (1857). Varronia macrocephala of Nees &
Martius, Nov. Acta, Acad. Caes. Leop.-Carol. Nat. Cur. xi. 78 (1823).
Lithocardium Neowiedianum Kuntze, Rev. Gen. ii. 977 (1891).—This
species is based upon a description of a plant given by Nees & Martius
in their account of the botanical collections made in Brazil by Prince
Maximilian of Wied-Neuwied. The plant described was determined
by them as V. macrocephala and was said to have been collected in
southeastern Bahia near the Ilhéos River. It apparently belongs to
the section Varronia, its flowers being described as being in dense
globose heads the size of a walnut. The calyx is given as being ob-
conic and very hispid with darkly ferrugineous hairs. The deltoid
lobes are subulate at the apex. The leaves are ovate-oblong and hispid.
These characters suggest that the plant might be near if not actually
C. calocephala. Nees & Martius, however, speak of the corolla as
“magna, infundibuliformis, alba; limbo inaequali quadrifido, lacinia
supra majore” which would Suggest that it has large conspicuous
flowers. Hence the species may represent a plant, not seen by me,
which would key out with C. paucidentata.

Cordia oncocalyx Allem. Trab. Comm. Sci. Explor. Bot. i. 11, cum
tab. (1862).—This is a species of Auxemma, A. oncocalyzx (Allem.)
Baillon.

Cordia Patagonula Aiton, Hort. Kew. i. 259 (1789).—Patagonulla
americana L,

Cordia tetraphylla Aubl. Hist. Pl. Guian. Fr. i. 224, t. 88 (1777);
Fresen. in Mart. Fl. Bras. viii. pt. 1, 13 (1857). Lithocardiwm
tetraphyllum Kuntze, Rev. Gen. ii, 977 (1891). Firensia lutea Raf.
Sylva Tellur. 40 (1838).—Fresenius reports a collection of this speci¢s
from the Para District, Dutch Guiana, and intimates that the species
may be expected in the Amazon Region. I have not seen any materia
of any Cordia from any region that is possibly referable to this
obscure species. I do not believe that the fruit and fruiting branch
which forms the bulk of Aublet’s illustration represents a genuine
species of Cordia. The flowers shown in the illustration appear to be
those of a Cordia, but in any case I do not believe that they wel
produced by the plant with the leaves and fruit shown.

; Vand. Fl. Lusit. et Bras. 14 (1788); Roemer, Script: Pl.
Hispan. Lusitan. Bras. 88 (1796); DC. Prodr. x. 177 (1846). M.
arborea Steud. Nom. ed. 2, ii. 87 (1841). ?M. racemifera Vand. &
Jackson, Index Kew. iii. 140 (1894).—This genus was based up?

STUDIES IN THE BORAGINACEAE 65

Brazilian material supplied Vandelli by Vellozo. T hough its brief
description is entirely inadequate for its complete recognition most
students have been content to follow De Candolle, |. ¢., and place it
as a doubtful synonym of Cordia. I can not claim to be any more
successful in identifying the genus than my predecessors and can
only add that if the plant described is a Cordia it is most probably
C. Sellowiana Cham. since of the species of this genus most apt to

" have been collected by Vellozo it agrees best with the short description

given by Vandelli. The significant details given by him are as follows,
“foliis inferioribus ovato-oblongis, obliquis, acuminatis, supra scabris,
subtus subtomentosis; floribus cymosis; caule arboreo.” A copy,
rearranged as to form but complete as to substance, has been given
by De Candolle, l. ec. De Candolle, however, has included in his copy
the binomial “Cyrilla racemifera” which comes directly after the
description of Macielia in the original text. Vandelli did not, I believe,
intend to have this binomial directly associated with his genus Macielia.
A study of his work will show that he did not give species for any of
's new genera. It should also be noted that the specific name se-
lected for the Cyrilla, i. e. racemifera, is one scarcely consistent with
the phrase, “ floribus cymosis’’ of the description of Macielia. As
I do not believe that the name Cyrilla racemifera is properly to be
associated with Macielia I accordingly doubt the validity of the com-
bination, Maciclia racemifera, which was coined in the Index Kewensis.
Tenore, Mem. Mat. Fis., Soc. Ital. Sci. Modena xxiv. 366
(1847), M. callipticantha Tenore, |. c. 367; Caesati, Atti R. Accad.
Sci. Fis. e Mat. Napoli vi. no. 11, 5, t. 3 (1875). M. calliptiantha
Tenore ex Pasquale, Cat. R. Orto Bot. Napoli 63 (1867). M. brasili-
ensis Tenore ex Pasquale, |. c. M. calliptychantha Caesati, |. c. Cordia
rufescens, var. glabrata Caesati, 1. c.—The genus Macria was based by
Tenore upon a very young (4-5-year old) plant, grown from seed
said to have originated in Brazil, which flowered at Modena in 1846.
flowers produced were either very faultily described by Tenore or,
“ seems likely, were precocious ones abnormal in development. In
aby case the original description and Tenore’s plate (published sub-
‘quently by Caesati, 1. c.) present a most suspicious combination in
the number of floral parts. The style is given as consisting of 3 bifid
and the stamens, corolla- and calyx-lobes as being 5-8. No
Was developed. The leaves are described as being shiny and quite
I eget myrtle-green above and yellowish or even golden beneath.
= inclined to believe that we are concerned here with a plant of C.
an yensis, or possibly C. superba, that was modified by cultivation
abnormal in the number of its floral parts.

fruit

66 JOHNSTON

2. TREATMENT OF TOURNEFORTIA.

This study of Tournefortia is the second of a series of projected
papers of the classification of the Subfamily Heliotropioideae. Al-
though this study is directly concerned with clarifying the classifica-
tion of the species from the eastern half of South America the treat-
ment of sections presented here is the result of a careful examination
of all the American species and most of those of the Old World as well. .
The two sections recognized are very distinct. The discussion of the
phylogenetic significance of these profoundly different groups is
reserved for a future paper in this series.

Key To THE SEctions.
Fruit obscurely if at all lobed, evidently bicarpellate, breaking up
into 1- or 2-seeded irregular nutlets ; embryo straight.
1. EurourNnerortia, pg. 66.
Fruit evidently lobed, obscurely bicarpellate, consisting of 4 con-
imilar 1-seeded nutlets; embryo curved...... . . 2. CypHocyEMa, pg. 72.
1. Section Eutournefortia Johnston; type-species, 7’. hirsutissima
L. Tournefortia-Pittoniae HBK. Nov. Gen. et Sp. iii. 80 (1818); type
species 7’. hirsutissima L. Tournefortia § Pittonia Don, Gen. Syst. iv.
366 (1837); type-species 7’. hirsutissima L, Pittonia Plumier ex Adans.
Fam. PI. ii. 177 (1863); type-species 7. hirsutissima L. Oskampia Ral.
Sylva Tellur. 123 (1838); type species 0. scandens and O. hirsuta Ral.
see note below. Tournefortia § Mallotonia Griseb. Fl. West. Ind. 483
(1861); type-species, 7’. gnaphalodes R.Br. Mallotonia Britton, Ann.
Mo. Bot. Gard. ii. 47 (1915); type-species, 7. gnaphalodes R. Br.

Giirke, and the fifth is a Heliotropium, H. ternatum Vahl. The thir
species treated, 7. volubilis L., has fruit not in accord with the gene
description in the Genera Plantarum, 68 (1754). Accordingly ther?
remain four species from among which the type-species of the gels

STUDIES IN THE BORAGINACEAE 67

should be selected, i. e., no. 2 T. hirsutissima, no. 4 T. foetidissima,
no. 6 7. glabra and no. 7 T. siberica. Upon consulting the Genera
Plantarum (1754) it is found that Linnaeus cites Pittonia Plum. as a
generic synonym and refers to the third plate of Plumier’s Nova

lantarum Americanarum Genera (1703). In fact Linnaeus accepted
Plumier’s generic concept almost in its entirety. His rejection of the
name Pittonia, in favor of the newly coined, Tournefortia, was merely
on the ground that the latter rather than the first part of the family
name of Joseph Pitton de Tournefort should be used in making the
generic name, cf. L. Hort. Cliff. 49 (1737). The crude figures on
Plumier’s plate, t. 3, which were cited by Linnaeus probably represent
the floral parts of 7. hirsutissima, for certainly among the six species
which appear under phrase names in Plumier’s text, l. c. 5, that species
is the only one that could have inspired the figures mentioned. With
the exception of a few details, which he apparently obtained from
Plumier’s generic description, Linnaeus appears to have based his
generic description upon the figures given by Plumier and so, thus,

This selection keeps the name associated with the largest part of the
genus as currently accepted and is in accord with an earlier selection
of a genotype made by Britton & Millspaugh, Bahama Fl. 361 (1920).

Trefer to 7, ournefortia, as a generic synonym, Oskampia Raf. Sylva
Tellur. 123 (1838). The original description of this genus is as follows,—

770. OsKampra Raf. (bot.) cal. Spart. cor. infund. 5fida, stam. 5.
antheris subsess, styl. clavatus, stig. capit. bacca scabra 2loc. 4sp.
Pruct. scandens, fol. alt. fl. subrac—Types O. scandens and hirsuta Raf.
Tournefortia do. L. auct. very dist. G. of Vines, berries not biporose.”
Just What species of Tournefortia are referred to by Rafinesque is

K

Stems shaggy with abundant slender spreading brownish hairs

unm. long; corolla 10-12 mm. long; calyx-lobes conspicu-
ous'y subulate; doubtfully from our area..............-- 1. T. obscura.

or with sh irs.

8 » Wie stigma apparently sessile in the mature fruit.
sae and leaves conspicuously and abundantly short-hirsute;
Ste oubtfully from our area 2. 7

ms and leaves glabrous or rarel arsely and inconspicu-
ualy ap ‘distributed in our iit

Ce ee er ee ee |

T. bicolor.

68 JOHNSTON

Style well developed and obvious even on the mature fruit.
rolla-lobes elongate, less than 1 mm. broad; leaves alternate;

Masonian head-waters... 0)... 6.506.668 eS 4, T. Ulei.
Corolla-lobes broad, over 1 mm. broad ; leaves very frequently oe
opposite; northern Argentina and adjacent Bolivia. ...5. 7. Lilloi,

1. Tournefortia obscura A. DC. Prodr. ix. 517 (1845); Fresen. in
Mart. FI. Bras. viii. pt. 1, 49, adnot. (1857).

Occurring in Central America and in the countries of northwestern
South America from British Guiana to Bolivia. Only doubttfully

accredited to Brazil.
BRAZIL: indefinite, Glocker 227 (BM).

Fresenius, |. ¢., reports this species from Guiana and Maynas (east-
ern Peru) and states that he has not seen it from Brazil. Although it
does not appear to be common, it has been collected in British Guiana,
Venezuela, Colombia, Panama, Peru and Bolivia in tropical condi-
tions at low altitudes. While I suspect that the species will eventually
be found in the Amazon Basin in western Brazil I am inclined to believe
Glocker’s collection, despite its label, did not come from Brazil
This collection formed a part of the Shuttleworth Herbarium now in
the British Museum. It was probably made by the Fernand Glocker
who is reported by Nees, DC. Prodr. xi. 728 (1847), as having collected
extensively in Bahia. From my own observations I can believe that
some of the collections by Glocker in the British Museum may well
have come from Bahia, although I should guess that most of them
came from near Rio Janeiro. Whether Glocker’s collection no. 22/
really came from Bahia or Rio Janeiro or whether it is merely evidence
of some confusion of material or labels can not be decided now. Per
sonally I do not believe the specimen is of Brazilian origin.

2. T. hirsu L. Sp. Pl. 140 (1753); Fresen. in Mart. Fl. Bras.
Vili. pt. 1. 49 (1857). Messerschmidtia hirsutissima R. & S. Syst. IV.
541 (1819); Walp. Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur. 1%
suppl. 1, 371 (1843). ;

Central America, the West Indies and northern South America;
doubtfully reported from Brazil.

BRAZIL. Rio Janemo: Rio Janeiro, Nov. 1830, Meyen 3426 (BD).

STUDIES IN THE BORAGINACEAE 69

ica at Rio Janeiro and from there sailed directly to Chile and Peru.
After collecting in southern Peru and about Lima he sailed for Hawaii
and the Orient. Consequently he visited no locality from which
l’. hirsutissima is with certainty known. It seems very improbable
that this large shrub, if present, could remain undiscovered so many
years in a region so frequently visited and repeatedly botanized as
that about Rio Janeiro. I am therefore forced to the belief that the
specimen at Berlin was not collected by Meyen, but by some other
person, and that Meyen’s label, through some confusion of data and
material, has been incorrectly associated with it. In any case, until
7. Misectizsinia has been rediscovered at Rio Janeiro or elsewhere in
our area, the authenticity of Meyen’s specimen, and consequently the
record based upon it, should be questioned.

3. T. bicolor Sw. Prodr. 40 (1788) and FI. Ind. Oce. i. 344 (1797).
r. laevigata Lam. Tab. Encye. i. 416 (1791); Poir. Encye. v. 356 (1804 Ds
Fresen. in Mart. Fl. Bras. viii. pt. 1, 49 (1857). 7. latifolia Willd. in
R. & S. Syst. iv. 540 (1819). 7. niléde HBK. Nov. Gen. et Sp. iii. 84
(1819). 7. glaberrima Salam. ex DC. Prodr. ix. 519 (1845); Fresn.
Le.49, T. laevigata, var. latifolia DC. 1. ec. 519. 7. bicolor, subvar.

latifolia Gomez, Anal. Hist. Nat. Madrid xix. 256 (1890). 7. bicolor
var. laevigata ecke: l. ec. ? Heliotropium citrifolium Lehm. Gétting-
ische Gelehrte Anzeiger 1817: 1515 (Sept. 1817). Stenostomum?
dichotomum DC. Prodr. iv. 461 (1830); Moc. & Sesse, Calq. Fl. Mex.
t. 482 (1874); Hemsley, Biol. Cent. Am. ii. 372 (1882). Antirrhoea ?
dichotoma Hemsley, |. c. 42 (1881).

Widely distributed in tropical America.

1p PAZIL. Amazonas: Rio Negro near confluence with Rio Solimées, May
, Spruce 1491 (G, K, BM); corolla white, shrub 1-4 m., Jurud Miry, Rio

Jurud, June i901, Gle wae (K, BD). Pari: tree 4-6 m. tall, i929,
aunt & Smith 30290 (G, US)? corolla white, Para, Oct. 25 25, 1829, urchell
laee (G, K). Pernamsuco: corolla whit te, Iguarassu, Nov. 3, 1887, Ridley,

BD); Kamage (BM); Beberibe near P. emnambico, Jun June 4, 188%, Se henck oO
:; Olinda, N36, » Picket a. Mus, raul 2 3 (G). Bax mS tad

AHIA:
eee tt ma), si Pog car Laken
oe (BM, D DO), v1 )tT18 B (BM), 1858 BM, 1690.0 0 (DC) ay 2h ABM)
anto, ovo,
io ee 10038 8 (BD); betw. Rio oes. sre noes Beni 1894, Sakiieke
(BD); Parahybuna, 1901, Sebastido de Nasconcellos, Mus . Paul. 18129
(SP); P Paraiso lis ; ,

2
ee
Tee
Se:
Dy
EBS
=
re8.
Tas
225
as
7
x2
ae
a
oe
BB.
S38

(K, BM). Rito JANEIRO: Valle the Rio Comprido, A

Canine 83 83 (NY, K K, BM); corolla white a Ses, Cateté and lotatona. ein 7,
Burchell 2989’ (K); Port D); betw.
BM, and Campo Fri rio, Sellow 200 and 1577 (BD); Rio Janeiro, Miers
pe aad (K), Gaudichaud 531 (BD, DC) and Tweedie 1261 (K);
(K) Toe, Ontos, Lhotsky 104 (DC); Or gan Mts., June 1838, Miers (BM) 4203
Apr. 12 aye, Dec. 1892, iniee (NY); vi vicinity of Monte Serrat, Mt. Itatiaya,
* "*, 1929, Smith 2808 (G); indefinite, Glaziow 4891 (BM, RS), 4991 (K,

70 JOHNSTON

BD), 7779 (K, BD) and 11294 (RS). Sio Pavto: Campinas, Dec. 1894,
Novaes, Mus. Paul. 11261 (G); San José do R. Pardo, Sept. 17, 1889, Loefgren
1380, Mus. Paul. 11253 (G); edge of the matto, Guarujé, Feb. 28, 1929,
Smith 2023 (G); corolla white, S. Vicinte near Santos, Feb. 10, 1875, Mosén
3446 (RS). Parand: Morretes, Nov. 6, 1908, Dusén 7005 (RS); Morretes,
Sept. 22, 1914, Dusén 957a (G, RS); Serra do Mar, Volta Grande, Dusén 3622
(RS); Cubataéo, Dec. 26, 1911, Dusén 13669 (RS). Santa CATHARINA:

ear Blumenau, Sept. 1888, Ule 950 (US, BD); Blumenau, 1884, Muiiller (BD);
Itajahy, Muller 443 (K). Inpmrrnrre: Brazil, Sellow 49, 783 and 5840 (BD),
Claussen 18 (BD), Scouler (K) and Boog j

PARAGUAY. corolla white, Yerbales, Sierra Maracayt, Hassler 5185 (K,

M, BD, Hass, BB).

The Brazilian material of this species is indistinguishable from tha
found on the Caribbean islands and the a jacent mainland and is
obviously conspecific. The most striking variation presented by our
plants is that found in the material from the Amazon basin. Spruce’s
collections from Amazonas have lanceolate or subulate calyx-lobes
which become 4 mm. long and in addition have the veins of the lower
leaf-surfaces evidently strigose. This Amazonian form is similar to
the plant of Central America described as 1. bicolor var. calycosa,
Donn. Sm. Bot. Gaz. xiv. 27 (1889), and is also very much like the
British Guianan plant described as 7. Schomburgkii, DC. Prodr. 1x.
517 (1845). The common form of 7’. bicolor in our region has ap-
parently glabrous leaves and has very short calyx-lobes which
only about 1 mm. long. I have seen none of our plants with the leaves
sparsely though evidently strigose such as are frequently found ”
Central America. I might add here that this phase of the species
was the basis of the very vaguely described and consequently poorly
understood 7. Billbergiana Beurl. of which I have recently had the
opportunity to examine the t pe.

4. T. Ulei Vaupel, Notizbl. vi. 186 (1914).

Known only from the headwaters of the Amazon; rare.

BRAZIL. Amazonas: Falls of the Madeira, Oct. 1886, Rusby 2049 (G, NY:
é Fie ie, Porvenir, Rio Tahuamanu, ca. 30 km. south of

obija, Jan. 9711 (BD, TrprE: K ISOTYPE).

VENEZUELA. Amazonas: near San Carlos, April 1853, Spruce 2985 (K)

This species is closely related to 7. ovalifolia Rusby (= r. chin-
chensis Killip) which ranges at higher altitudes along the Cordilleras
in Bolivia and Peru. It differs, however, in having the flowers sessile
and the larger, proportionately broader, ovate or elliptical leav®
more abruptly contracted at the base. The specimens examined
none of the leaves opposite or subopposite. The branches of f
ovalifolia have, on the other hand, some of the lower leaves coat
opposite or subopposite. The specimens from Venezuela are glabra

STUDIES IN THE BORAGINACEAE 71

but otherwise are quite similar to the plants of the Brazil-Bolivian
frontier.

5. T. Lilloi, sp. nov., fruticosa 1-3 m. alta pilis minutis incurvatis
rigidusculis numerosis vestita mox glabrata; ramis teretibus gracilibus;
foliis lanceolatis vel non rariter elliptico-ovatis 1.5-6.5 em. latis 6-15
em. longis herbaceis nervosis integerrimis plerumque oppositis vel
suboppositis vel rarissime ternatis subtus pallidioribus apice acumi-
natis basi in petiolum 6-11 mm. longum gradatim contractis; inflores-
centia terminali foliis plus minusve superante laxe dichotomo-ramosa
1-7 em. longe pedunculata; spicis saepe ca. 8 ebracteatis 2.5-5.5 cm.
longis; floribus in spicis unilateralibus sessilibus vel subsessilibus vel
infimis rariter usque ad 2 mm. longe pedicellatis 1-3 mm. distantibus;
calycibus ca. 2 mm. longis ad basem versus in lobos erectos subulatos
vel lanceolatos incisis pubescentibus; corollis albis hypocrateriformi-
bus extus strigosis intus glaberrimis, tubo cylindrico 5-8 mm. longo

.-2 mm. crasso quam calyce 3-4-plo longiori, limbo paullo concavo
ca. 5 mm. diametro, lobis ovatis vel semiorbicularibus latioribus quam
longis ca. 1.5 mm. longis post anthesin plus minusve recurvatis;
antheris distinctis 1.4—1.7 mm. longis glabris supra medium tubo

Ventraliter et verticaliter compresso apice truncato paulo supra

plus minusve suleatis; stylo subpersistenti maturitate 1.5-3 mm.
longo quam stigmate 2-3-plo longiori; ovario glabro.—T. undulata of
Griseb. [Symb. Argent.] Abh. K. Ges. Wiss, Gattingen xxiv. 270 (1879);
0, [Resefia Fitogeogr. Tucumfn] Prim. Reunién Nat. Soc. Argent.,
Bot, 221 (1919). 7. laevigata in pt. of Kuntze, Rev. Gen. iii. pt. 2,
206 (1898), 7. polystachya of Fries, Ark. Bot. vi. no. 11, 21 (1906).
l. sp. Vaupel, Mededeel. Rijks Herb. no. 46, 12 (1922).
g in northwestern Argentina from southern Tucumdan north-
hry through Salta and southeastern Jujuy to adjacent southernmost
Ivia,

ZO, ; a ?
Oct. 1390120, 1911, Rodriguez R181 and 10983 (G); “El Cedrél,” Orén,
1990, 13, 1913,

72 JOHNSTON

Oct. 11, 1929, Venturi 9637 (G) ; Jujuy, Oct. 1892, Kuntze (NY, BD). Tuev-
ie] 22, 1907, Lillo 7099 (Typr, Gray Herb.) ; Yerba

m., Oct. 12, 1917, Schreiter 109 (G); Yerba Buena, 600 m., March 19, 1909,
Venturi 316 (G); vicinity of Tucumdn, Oct. 1902, Baer 23 (BD, RS); Rio Chico
to La Samy gt m., Nov. 13, 1913, Lillo 15154 (G); near Santa Ana, Nov.

The exact relationship of this very distinct previously unnamed
species is uncertain. It belongs with the relatively small group of
species of the section Eutournefortia which commonly develop opposite
or subopposite leaves. Its relations are to be expected in Bolivia or
Peru. The species of these countries, however, which most suggest
T. Lilloi have fruit that is rounded or broadly conic and is drawn up
at the apex into a slender elongate persistent style. In the present
species the fruit is deeply sunken in at the apex and has a very much
shorter and less conspicuous style. The species is named in honor of
Prof. Miguel Lillo of the University of Tucumén to whom I am in-
debted for many of the specimens I have cited in this paper. I find
it a great pleasure to be able to associate Dr. Lillo’s name with another
distinctive species of the flora which he has studied for so many years.

II. Section Cyphocyema Johnston: type-species 7. membranacet
(Gardn.) DC. cf. Miers, Contr. Bot. ii. tab. 53b (1869). Verrucaria
Medic, Malv. 103 (1789); type-species 7. volubilis L.; not Verrucaria
Wegg (1780), a genus of lichens. Messerschmidia R. & S. Syst. iv.
pg. li and 541 (1919); not Messersmidia L. Mant. 42 (1767) not
Messerschmidia Murr. Syst. ed. 13, pg. 161 (1774); regarding these
names see discussion below. Tournefortia-Messerschmidtiae HBK.
Nov. Gen. et Sp. iii. 78 (1818); as to plants but not nomenclatorial

of the section Cyphocyema are characterized by the elongate very
narrow corolla-lobes which have involute margins and are subvalvate
in the bud. In many ways the section seems more closely related

STUDIES IN THE BORAGINACEAE 73

Heliotropium § Orthostachys than to T ournefortia § Eutournefortia.
The most critical discussion of the characters of this section have been
given by Miers, Contr. Bot. ii. 192 and 202-4 (1869). Its distinctive
characters were indicated much earlier, however, by Robert Brown,
Prodr. 496 (1810).

The name Messerschmidia has been applied to this group of Tourne-
fortia by Roemer & Schultes, Syst. iv. pg. li. and 541 (1819), Don,
Gen. Syst. iv. 370 (1837), Miers, Ann. & Mag. Nat. Hist. ser. 4, ii.
191-204 (1868) and Contr. Bot. ii. 202-215 (1869), and a few others.
This name, which has been subjected to several variations of spelling,
was first applied by Linnaeus, Mant. 42 (1767), to the Asiatic 7.
siberica L., a plant which has the general fruit-structures of Eutourne-
fortia. Linnaeus spelt the name, Messersmidia, and maintained this
spelling in all his writings. The first orthographic variation occurred
in the 13th edition of the Systema, 161 (1774), by J. A. Murray who
Properly applied the generic name to the Asiatic plant but who spelled
it Messerschmidia. This variant spelling generally prevailed until
Kunth, HBK. Nov. Gen. et Sp. iii. 78 (1818), introduced the final
variation, Messerschmidtia. Although Kunth attributed the name to
Linnaeus, he applied it incorrectly to the present group of American
Tourncfortiae. With Roemer & Schultes, |. c., the use of the name

recognition, arbitrarily applying to the American plants the name
Messerschmidia. The Messersmidia of Linnaeus was subordinated as
4 group of Tournefortia under the name “ Messerschmidiae antiquae,”
9. To prevent confusion in their own work (R. & S. 1. c. pg. li)

they referred to our group of American plants as “ Messerschmidiae
vero.” This arbitrary new application of Messerschmidia was followed
by Don, 1. ¢. 269 and 270, Endlicher, Gen. 646 (1838) and Miers, 1. e.
De Candolle, Prodr. ix. 98 (1845), however, was able to introduce
i ‘r complications, by giving the name still another application.
18 Tournefortia § Messerschmidia, apparently formed around M.
ef, Suppl. 132 (1781), consists of four species from Africa
M the Canary Islands. The original Linnean species of Messersmidia,
cc Se L. (= T’.. siberica L.), by De Candolle, l. c. 514, was placed
os . rguzia in the monotypic section Tournefortia § Arguzia. In his
udy of the sroup, Giirke, E. & P. Nat. Pflanzenf. iv. Abt. 3a, 94
Mem’ recognized that the species treated under Tournefortia §
Holi ne by De Candolle, |. c., were more properly placed under
“oropium. Tn rectifying De Candolle’s treatment, however, he

74 JOHNSTON

disregarded the nomenclatorial basis of the sectional name used by
De Candolle and, adding one more blunder to the sad account of
Messersmidia, erected Heliotropium § Messerschmidia to include the
African and Canary Island plants mentioned. As if Messersmidia
had not already been complicated sufficiently by wide variations in
spelling and applications, Dalla Torre & Harms, Gen. Siphogam. 425
(1900), approaching the matter from a purely bibliographic interest
and lacking any acquaintance with the plants concerned, have treated
the name as a section of Heliotropium and cite under it as synony-
mous a group of references and uses of the name which apply not only
to the African and Canary Island plants but to those of Asia and of
America as well. It becomes more than obvious that the unfortunate
name, Messersmedia, has been both misspelled and recklessly mi i
As the original species of Messersmidia, i. e. T. siberica L., is very
distinct and may eventually be thought worthy of generic recognition
_ it seems best that the name Messersmidia (= Arguzia Raf.) should be
maintained for the Asiatic plant. The application of the name
Messerschmidia to the African and Canary Island species of Helvtro-
pium can not be justified. Its use for the American species of Tourneé-
fortia was most arbitrary and irregular and can not be sanctioned.
In any case the confusion that has attended the history of the name
is quite sufficient to warrant its rejection as a nomen confusum, at
least as far as our American plants are concerned. Consequently I
have provided a new sectional name for those of our plants w
have the lobed fruit and the curved embryo. The name alludes, of
course, to the shape of the embryo.

Although the section Cyphocyema as a group is evidently well
marked, the species which compose it are variable in the extreme and
are notorious for the difficulties they present to classification.
group is only weakly developed in the Andean Region, having tW°
principal centers of distribution, one in the Caribbean Region and the
other in Brazil. Before the section can be said to have even a reason
ably satisfactory treatment it must be studied as a whole. Hence
publishing the following treatment of the species of eastern South
America it is fully realized that it presents no final classification 2h
must be subjected to future revision when the precise relation a
species with the extralimital ones, particularly those of the Caribbea
Region, becomes more clear. My purpose in this paper is to giv¢®
careful, although not detailed, survey of the principal forms of the
section found in Brazil, Paraguay and Argentina, Efforts, hence; bave
been directed at determining just what parts of the variable assem)™
found there constitute reasonably definable and natural entitle

iilisappe™

STUDIES IN THE BORAGINACEAE 75

worthy of major recognition. Consequently although I have been
much concerned with the interrelations of the species of our area I
have not made a particular effort to discover their precise relation to
the extralimital ones. I have treated as specifically distinct all our

treated as species will no doubt eventually be found to be no more
than geographic variations of species occurring in northwestern
South America or the West Indies. In order that there be no mis-
understandings I have, however, freely stated in the discussion of
each species my ideas regarding their relations and the degree of the
relationship. I have thought this procedure more desirable than the
making of a large number of perhaps unnecessary new names.

Although some of the species I have accepted are extremely variable
Thave made only a single new subspecific name. It will be noted, how-
ever, that I have discussed at length under each species its more

elaborate classification of tropical South American plants are not
needed and can not help but be very premature at this time I prefer
to leave botanical bibliography unburdened by a host of ill timed
Polynomials and the opportunity for such detailed classification open
to the resident field-botanist of the Tropics who will be more able to
Judge of the needs and of the extent of the subspecific classification
really usefyl and desirable.

Key To Specigs.
Tnflorescence axillary, frequently also terminal.
t fulvous, conspicuously strigose-tomentose; leaves ovate or oy.
Plan borbicular ; Inflorescence cylindrical, stiffish..... . 6. T. mapirensis.
' green, only obscurely if at all strigose; leaves lanceolate to
narrowly elliptical ; inflorescence slender and widely branched.

Inflo : 7. T. breviflora,
Conc terminal, not distinctly axillary.
rova-tube decidedly elongate, 4-10 mm. long, usually much
short lobes

ee. ee.
Mature leaves more or less strigose or tomentose, the youngest ;
Corolla) 2, Ways densely strigose or tomentose’... 9. T. paniculata,
Tolla-tube legs strikingly te.
I with the lo ore t
hflorescence ra
Secondary branches short and stiff; flowers congested.
Teaves glabrate or merely strigose beneath with straight
®ppressed not very abundant hairs.

76 JOHNSTON

¥YT. - Yet
pal

rigose, the fay ry (ase
sparse) not restricted to “tee ene: blade ut not
leathery, lanceolate or rarely ovate, 1.5-3 a. broad.

10. 7’. Gardneri.
Peper surface of leaves glabrous or hairy only along the

pressed veins; leaf-blade leathery, ovate, 3-6.5 cm

te oh re Sg Se UE ee eer cae Ne ee 11. T. subsessilis.

Leaves tomentose beneath with an sig rast indument of
slender somewhat curved or ou y

Upper surface of leaves glabrous................ T. subsessilis.
Upper — e of leaves father abundantly icc or
velutinous.

Under surface of Jeaves and the corolla-tube with an
bescence; fruit pubescent.12. T. candidula.

ade ha aiek GEL EO UNG cee oaks Leek bes oe . T. villosa.
fnticeemmuniaey vei loosely orn throughout, the coisa
ranches slender and elongate; flowers loose .
Leaves evidently although Pa very abundantly strigose or
oti glabrous, the hairs on the upper r leaf-surface usu-
rovided od ce ells 14. sf rubicunda.
; peri cheat tomentose, the hai t oid base
aves sein tat fulvous beneath; sgn a larger
n the follow te mii es . T. Salzmanni.

freoty bra NON a a ee 16. T. m embranacet.
ee hich Ghic baud hud sie T. salicifolia.
6. T. mapirensis Lingelsheim in Fedde, er vii, 244 (1909).
1’. subrotunda Rusby, Bull. N. Y. Bot. Garden viii. 115 (1912).
Amazonian head-waters from northeastern Peru to east cen
gar Via.
CRE: scandent, corolla greenish yellow, Seringal 5. Francisco,

July 19 1 lear Ule
y 1911, Ule 9713 K, BD); frui Frene
re = BD). ( ); t bright yellow, Seringal Trancisco,

E — twining ig with pale ereceaat — scented flowers
urim , Rio ay 1855, Spruce 3889 (G, K,
BOLIVIA. Bent Rurrenabaque, 300 t. 15, 1921, Rusby 1290 “
Ha Pat: C near Mapiri, , Nov. 1907, :

oe very distinct species asia ee cn its tawny pubescene®
and well developed definitely axillary rather than terminal inflores:
cences. The flowering branches are leafless or bear only a very
much reduced leaves.

- T. breviflora DC. Prodr. ix. 520 (1845); Fresen. in Mart- 4
Bras. viii. pt. 1, 50 (1857). 7. Vauthieri DC. 1. c. 526; Fresen. ct

STUDIES IN THE BORAGINACEAE 77

Messerschmidtia V authieri Miers, Ann. & Mag. Nat. Hist. ser. 4, ii.
196 (1868) and Contr. Bot. ii. 207 (1869). 7. macroloba DC. 1. ¢. 527;
Fresen. 1. ¢. 55. M. macroloba Miers, I. c. 196 and 1. e. 207. 7. lance-
olata Fresen. |. ¢. 55, t. 12. M. organensis Miers, |. c. 194 and 1. ¢. 205.
M. ramiflora Miers, 1. c. 197 and 1. ¢. 208. 7. catharinensis Vaupel,
Notizbl. vi. 181 (1914). 7. gracillima Vaupel, |. c. 183. ? T. aapuryen-
sis Vaupel, |. c. 186.

Ranging in the coastal states of Brazil from southernmost Minas
Geraes to Santa Catharina and from thence into Paraguay; apparently
reappearing in Acre Terr.

ner TYPE of ramifl
Caldas, June 18, 1861, Regnell III 915 (RS). Rio JANEtRO: near Petropolis,
ca. 750 m., July 1882, Ball (K); Imbuy Lane, Organ Mts., Dec. 1837, Miers
- or, an Mts., Miers 4204 (K); Serra dos Orgaos,

Nov. 25, 1898, Loefgren 4181, Mus. Paul. 11260 (G); Campinas, Dec.
1894, Novaes 147, Mus. Paul. 11952 (SP.). and: Ponta Grossa, 19,
1911, Dusén 13216 (RS); Ponta Grossa, June 15079 (G, RS);
Ponta Grossa, Aug. 8, 1909, Dusén 8041 (RS); lacareh¥, Sept. 6, 1915, Dusén
17135 (RS). Santa Catarina: Blumenau d Velha, July 1888,
Ule 811 (BD, TYPE of T’. catherensis). Acre: scandent, fl. greenish yellow,
Seringal § Francisco, Alto Xapury, Oct. 1911, Ule 9710 (BD, tyre of T.

;4Puryensis; K, isorypE). INDEFIN,TE: Brazil, 1834, Lund (DC, Type of T.
a

PARAGUAY. Sierra de Amambay, Hassler 11290 (G, K, BM, BD, Hass);
i 179 Sing hae 1893, Lindman A1837 (RS); Picado— uneal, June Hasaer
Bernardino, Aug 48, 1 aray; Aug. 1913, Hassler 11791 (G, BM, BD);

» Aug. 18, 1916, Osten 8953 (G).

Although its rather large number of synonyms would not suggest it,
this is one of the most readily recognized of the South American species
Of Tournefortia. It is one of the two known species of the genus with
tecidedly axillary inflorescences. The bractless flower clusters are

ly and widely branched and are developed not only at the apex
rie stems but usually also from many of the leaf-axils. The lanceo-

* leaves are glabrous or inconspicuously and very sparsely strigose
watt: The lower surfaces of the leaves are paler than the upper and
Po be glabrous or have a very fine sparse strigosity somewhat more

‘int than on the upper surface. The mature leaves of this
Species, however, always appear to be glabrous at first glance. The
a are well developed being slender and elongate, commonly
onl fing the corolla-tube. Of all the synonyms listed above

¥ A: tapuryensis seems to merit particular note. The type of this

78 JOHNSTON

species was collected in the head-waters of the Amazon in southwestern
Brazil. It can be separated by no important characters from the
frequently collected plant of the coastal states of Brazil and adjacent
Paraguay although from distributional considerations this is rather
unexpected. Perhaps further collections of this form from Acre may
reveal characters which will eventually justify its recognition as
distinct.

8. T. peruviana Poir. Encyc. Suppl. iv. 425 (1816); Urban, Symb.
Ant. iv. 524 (1910). 7. volubilis of R. & P. Fl. Peruv. ii. 24, t. 148b
(1799). 7. scandens Willd. Enum. i. 188 (1809), not Mill. (1768).

From Central America and the West Indies ranging south to
eastern Peru and the Amazon River; rare in our area.

BRAZIL. Amazonas: corolla green, Parané-mir4 dos Ramos, Nov. 1850,
aera 1129 (K); betw. Santarém and Barra do Rio N egro, Oct. 1853, Spruce
(G, K, BM). Par: climbing on low trees and shrubs, fruit yellow, Santarém,
April 1850, Spruce 804 (K).

This species is a common one in northwestern South America but
enters our area only in the Amazon Basin. I have seen materia
collected by Ruiz & Pavon and determined by them as 7. volubilis
and am consequently able fully to substantiate Urban in his applica
tion of the name 7. peruviana!

9. T. paniculata Cham. Linnaea iv. 468 (1829); Fresen. in Mart.
Fl. Bras. viii. pt. 1, 50 (1857). Messerschmidtia paniculata Don,
Gen. Syst. iv. 370 (1837). Heliotropium scandens Vell. Fl. Flum.
(1825) and Icones ii. t. 41 (1827), not T. scandens Mill. (1768) not
Willd. (1809). 7. elegans Cham. 1. c. 469; Fresen. 1. c. 50, t. 9, fig: 5
M. elegans Don, |. ¢. 371. T. brachiata A. DC. Prodr. ix. 525 (1845);
Fresen. 1. ¢. 51, t. 9, fig. 2. 7. intermedia Fresen. 1. c: 50, t. 9, fig. 6.
T. grandifolia Fresen. 1. ¢. 56, t. 13. 7. brachiata, f. grandifolia Chodat
& Hass. Bull. Herb. Boiss. ser. 2, v. 482 (1005). 2 T, graciliflors
Rusby, Bull. Torr. Bot. Cl. xxvi. 148 (1899).

Widely distributed in our area where it extends from the lowe
Amazon southward in Brazil through Goyaz and Matto Grosso ”
Santa Cruz (Bolivia), Paraguay, Misiones (Argentina), Rio Grande
do Sul and southernmost Minas Geraes. It is absent in easternmost
Brazil. Aberrant forms of the species occur in northwestern Argentin’
(var. austrina), eastern Peru, Colombia and British Guiana (1
spigelliflora A. DC.).

BRAZIL. Pari: Prai ; :a cli
shrub infores proce Nati dede oe Eee oe TH nel
brachiatn; K, GM, wotrres); betw. Natividade and Porto F al, Burch ot
Dec. 4, 1921, Holway 136/ (US); f’ prot Sais oe nas Hehoay 1404 (NY);

STUDIES IN THE BORAGINACEAE 79

near Rio Verde, Caldas, Dec. 15, 1873, Mosén 1505 (RS); Caldas, 1847, Regnell

I 309 (K, BD); Caldas, 1862, Kegnell I 309 (RS); Caldas, 1854, Lindberg 161

1 i tdgren (RS); indefinite, Claussen (K), Claussen 225

(G, NY, RS), Claussen 407 (BB, coryrs of 7. brachiata). Rio JANErRo: Serra
LOU 0

1918, Novaes, Mus. Paul.
894, Novaes 324, Mus. Paul. 11957 (SP); Santa Anna, 1908,
Usteri, Mus. Paul. 11255 (G); Sapucahy, Jan. 1893, Loefgren & Edwall 2141,
1 i ;

NITE: Bras. Meri é

Merid. Sellow (K, BD, DC, ? isorypss of T. ¢ gans); Bras. Merid. S

P1575e669 (BD, trpxr of 7. a aad Brazil, Herb. Mus. Vind. 1584 (K);
rade (BD

I
and 3294 (BD); Alto del Vas betw. San Pedro and Campifias de Américo, Dec.
6, 1886, Niederlein 1741 (BD); Yerbal, Campo Grande, Nov. 26, 1884, N
- 60 (BD); San Pedro, Nov. 7, 1886, N
— 1922, Molfino (G); corolla greenish yellow, Santa Ana, Nov. 10, 1913,
Rg 12468 (G); Posada near “La Granja,” Nov. 7, 1907, Ekman 1775

PARAGUAY. Upper Rio Apa, Hassler 7848 (BM, BD, Hass, BB), 7848a
ee Hass, BB) and 7699 (BM, ‘Hass, BB); Centurién, betw. Rio Apa and
ae Aquidaban, Nov., Fiebrig 4240 (K, BM, BD, Hass); Sierra Maraca "4,
1900 Hassler 5180 (K, BM, BD, Hass, BB); Cerros de Tobaty, Cord. Central,
1900, Hassler 6209 (BM, BD, Hass, BB); Asuncién, Oct. 1875, Balansa (K
Villa’ Rien’ ind); Cord. de Villa Rica, 1905, Hassler 8594 (BM, BD, Hass);

? 3736 (G),
3129 (BD): ¢ ANTA Cruz: Bosques de la Barreras, Prov. Sara, Ciabacs
Steinbach ‘5 105 (Ne Pampa Monte, Prov. Sara, 500 m., Nov. 16, :

As IT have defined 7. paniculata it consists of a group of forms
Separated from 7’. peruviana by having the leaves, at least the imma-
ture ones, more or less evidently and abundantly pubescent. These
Pubescent Plants occur almost exclusively in the area to the south and
“ast of the range of 7’ peruviana. Although reasonably satisfactory

ee differentiating character, the presence or absence of pubescence

ae merely a matter of degree. I am, hence, by no means
eae than future students will all agree as to the advisibility of
ee T. paniculata as specifically distinct from 7. perunana.
in th : Plants T have assembled under 7. paniculata are very variable

© outline of the leaves and in the amount of pubescence. The most

co
"mon leaf-form is more or less broadly ovate and acute. However,

80 JOHNSTON

plants with lanceolate leaves are also common, particularly in the
southern states of Brazil. As there are numerous collections ex-
hibiting leaves more or less intermediate between these two extremes
I have considered the difference in leaf-outline unworthy of nomen-
clatorial recognition.

The bulk of the material of 7. paniculata has leaves that are evi-
dently though not abundantly strigose. It is only in some forms from
the Amazon Basin, where the species approaches closely the main
range of 7’. peruviana, that the very sparsely strigose or nearly glab-
rous forms occur. Occurring sporadically in the same area as the
common evidently strigose forms of 7. paniculata are plants which
have leaves that are rufous or less commonly cinereous and are col
spicuously villose-strigose. These conspicuously pubescent plants are
mere extremes. Any extended suite of specimens will show numerous
intermediate forms connecting them by transitions to the common
rather green and merely strigose forms. There is absolutely no correla-
tion in the variation of leaf-outline and that of the amount of pubes
cence.

The inflorescence of 7’. paniculata also shows considerable variation.
It may be cylindrical or pyramidal with an evident and only sligh
flexuous axis or it may be very diffuse and have the axis strongly and
irregularly flexuous. It may have only a few or it may have Very
numerous flowers. The branches may be short or very elongate:
This great variability seems to be quite erratic and is not concomitant
with the variation of either leaf-outline or pubescence. ;

The corollas exhibit variation in size and, in accord with the foliag®
variations in the amount of pubescence. Some plants have almost
tubular corollas. The prevailing form, however, is trumpet-shapet-
There is also noticeable variation in the size of the corollas and im the
proportional length of lobes and tube.

The types of 7. paniculata and T., elegans are so similar that they
might well be respectively the fruiting and the flowering phases vi
plant. They are evidently strigose and have lanceolate leaves. The
types of 7. brachiata, T. intermedia and T. grandifolia are forms 4
acute ovate leaves; they are merely strigose. The type collection
T. graciliflora is a very puzzling plant. Its large thin oblong-ov8®
leaves are glabrate. Only the small immature leaves (seen only 0 ]
the specimen at Kew) are densely strigose. I am forced to the —
clusion that 7. graciliflora is an extreme shade-form of 7. pam .

In the synonymy, citation of specimens and discussion give? wie!
I have been concerned only with the plants coming from the =
south of Par and southeast of the Rio Madeira. Coming from

STUDIES IN THE BORAGINACEAE SL

side of this area, however, there are a few collections which belong to
1. paniculata in the broad sense and which should be considered in any
study of the species. Some of the material comes from the region to
the west and north of our area. There are two collections from Tara-
poto in eastern Peru made by Spruce (no. 4916) and by Ule (no.
6568) which in their evidently strigose herbage and acute ovate leaves
are like the bulk of the Brazilian material. These two collections, the
only material possibly referable to the species which I have seen from
Peru, are notable only because of their very slender elongate corollas.
There is also a collection from Valparaiso, Dept. Antioquia, Colombia,
made by Pennell (no. 10799). This is darkly tawny with an evident
short appressed pubescence. In the shape of its leaves, size and form
of the corolla, ete., the plant is otherwise quite like the glabrous or
glabrate 7. peruviana. The collection is of especial interest since, as
it must be referred to 7. paniculata, it consequently sets the northwest
limit of the species in a region in which otherwise only 7. peruviana
has been noted. Coming from British Guiana there are also a few
specimens which must be associated with 7. paniculata. These were
collected by Schomburgk (nos. 749, 427 and 427-6698) apparently
at P irara, at the head of the Rio Branco near the Brazil-British
Guiana border, cf. Schomb. Faun. & Fl. Brit. Guian. 1151 (1848).
They subsequently became the type material of 7. spigeliaeflora
A, DC, Prodr. ix. 525 (1845). The material is notable in 7. paniculata
for its soft spreading conspicuous pubescence, its slender elongate
corollas and rather well developed corolla-lobes. Among the three
extralimital forms I have mentioned I would consider it alone as
worthy of possible varietal recognition.

The collections I have mentioned in the last paragraph came from
Peru, Colombia and British Guj d outside our area. There
Temains for discussion, however, a group of forms which range in north-
Western Argentine and which accordingly should be disposed of in this

. mere aspect, a study of them in conjunction with material from

araguay and Brazil reveals the fact that the plant of northwestern
: hg is not at all readily separable from 7. paniculata taken as
: ‘ “i alizing this fact I very early in this study decided that
Cac: mn Argentina were unworthy of nomenclatorial recognition.
ha ing back to this problem at various times I have been repeatedly

Pressed, however, by the rather characteristic aspect of the north-

82 JOHNSTON

western plant and particularly by its restriction to a very natural area
in Argentina. These facts have finally decided me to give this ill
defined form nomenclatorial recognition as follows,—

Var. austrina, var. nov., planta argentinensis et austro-boliviensis a
var. genuina differt paniculis longirameis diffusis, lobis corollae con-
spicuis subulatis ad basem versus saepe expansis et plus minusve
evidenter coalitis.— 7. elegens of Griseb. [P]. Lorentz. 184] Abh.
Ges. Wiss. pena xix. 232 (1874) and [Symb. Argent.] I. e. xxiv.
270 (1879). T. Salzmannii of Kuntze, Rev. Gen. iii. pt. 2, 206 (1898).
T. grandifolia of Lillo [Resefia Fitogeogr. Tucumén] Prim. Reunién
Nat. Soc. Argentina, Bot. 219 (1919).

In Argentina from southeastern Catamarca, through Tucuméan,
eastern Jujuy, and western Salta to Tarija and Chaco in extreme
southern Bolivia.

BOLIVIA. Tarisa: Bermejo, Nov. 18, 1903, Fiebrig 2136 (G, K, BD).

HACO: noo Oct. 2a 1910, ‘Steinbach 1637 (

ARGEN or Saura: “El Cedrél,” Orén, Nov. Lt 1918, Rodrigue te
Venturi 8109 Fund 8176 (G). Jusuy: edilitak 1100 m., Jai 2, 1912, Lillo 11818

ucUMAN: — yellowish, Cerro del Cam mpo , 800 m., Nov. :

Venturi 7472 (G); La Ramada, Dec. 1914, Lillo 3032M (rves, Gray Herb)

Cerro Duraznillo, 750 m., Jan. 1, 1914, Lillo 15998 (G); Higiies Sola, ©.
; ., Oct. 17,

2, Lorentz 199 and 274 (BD); Yoru Buena, 650 m., Jan. 1900, Lillo 13,
(G); La Cocha, 440 m., F oe 1919, Lillo 283B Q; Le Calera, 500 m., Nov

1913, Lillo 15153 (G);'Tas Cu chillas, 1100 m., Dec. 14, (OR
e, 300 m., Oct. 1913, Lillo 15152 (G); 3 Barranea - la Tom (G).
Nov. 1, 1921, Sekeotie vets (G); Las Pavas, Dec: 1916, Torgensen “sl

CATAMA La
Sica Hae Come ah Son (BO aa
nate (G).

As I have intimated above, this variety is not at all sharply set of
from typical 7. paniculata. It is significant only in so far as it pies
to be a practicable geographical variety. If its distinct gross b
proves illusionary or if it is found to be too closely ap BRS y

many forms in Paraguay and Brazil it will have little to justify 1
recognition. Its establishment does, however, provide a cony
means whereby the botanists of Argentina can distinguish the | a
of Tucumén and Salta from the rather differently appearing plan
of Misiones and adjacent Brazil. Fl.

10. T. Gardneri A. DC. Prodr. ix. 526 (1845); Fresen. in Mart 4
Bras. viii. pt. 1, 54 (1857). Messerschmidia subulata Gardn. in H 4
London Jour. Bot. i. 582 (1842); Miers, Ann. & Mag. Nat. Hist. se

STUDIES IN THE BORAGINACEAE 83

ii. 193 (1868) and Contr. Bot. ii. 204 (1869); not 7. subulata Hochst.
ex DC. (1845). 7. restingicola Vaupel, Notizbl. vi. 184 (1914).

Brazil; at no great distance from the coast in the states of Rio
Janeiro and south to Rio Grande do Sul.

lo Janeiro, 1844, gren . SAo Pavto: vicinity of Sao Paulo, Rio
Tiete, July 26, 1827, Burchell 4791 (K); Pinheiros, Sio Paulo, July 23, 1897,
us. Paul. 11248 (G); Sao Paulo, Oct. 1913, Brade 7067, Mus.

GE ANTA

105 (BD); Hamman, Oct. 13, 1886, Schenck 602 (BD). Rio GRANDE Do SUL:
scrambling in hedges and edges of forest, Puerto Alegre, T'weedie (K). InpEFI-
Nite: Brazil, Mart. Herb. Brasil 1223 (K, BM, BD, DC); Brasil, Herb. Mus.
Vind. 1695 (K); Brasil, Blanchet 107 (BM).

Usually a slender climbing plant with lanceolate herbaceous and
strigose leaves. In the type, which seems to be a form from shaded
situations, the hairs on the foliage are very slender, quite scattered
and inconspicuous. In some forms of the species, however, the
Strigose pubescence on the leaves is coarse, abundant and quite
evident. The type of 7. restingicola is one of the latter forms and ob-
viously belongs with the plants I have referred to 7. Gardneri, although
Its leaves are oblong-ovate rather than of the characteristic lanceolate
outline of the species. Blanchet’s collection (no. 107) is also aberrant
m leaf-outline. The fruit of 7. Gardneri is finely strigose.

N. 'T. subsessilis Cham. Linnaea viii. 119 (1833); DC. Prodr. ix.
821 (1845); Fresen in Mart. Fl. Bras. viii. pt. 1, 53 (1857). Messer-
‘chmidtia subsessilis Don, Gen. Syst. iv. 370 (1837); Miers, Ann. &
Mag. Nat. Hist. ser. 4, if. 203 (1868) and Contr. Bot. ii, 214 (1869).
7’. maritima Salam. ex DC. 1. c. 526; Fresen. L. c. 54. 7. pyrrhotrichia
Fresen. le. 5

Eastern Brazil in Bahia and perhaps also in Minas Geraes.

BRAZIL Banra: : . sy DC

. : sandy plain near ocean, Bahia, 1830, Salzmann 371 (DC,

Ree Of T. maritima). erveurenen Vittoria—Bahia, Sellow (BD); Brazil,

ever BD, type of 7’. sessilis); Brazil, Sellow (BD, TYPE of T. pyrrhotrichia);
» Glocker 406-16E (BM).

ao Species appears to be rare. Although it is probably most
Yas ty related to T. villosa, of which it may be only a glabrous deriva-
vd 't is more difficult to separate in the key from 7. Gardner.

S characters are chiefly found in the leaves and in the lack of pu-

84 JOHNSTON

bescence. The leaves are leathery in texture and ovate to lance-ovate
in outline, being about twice as long as broad. The upper surface is
somewhat shiny and is glabrous or has only a few hairs along the im-
pressed veins. The lower surface may be sparsely tomentose but is
more commonly provided with only a few scattered appressed slender

-hairs. I know 7. pyrrhotrichia from material at Berlin determined by
Fresenius. This may be type material. The plant agrees very closely
in details with and is almost certainly part of the same collection as
the sheet (also at Berlin) from Sellow which is labeled as from “ Vit-
toria—Bahia.” The under surface of the leaves is sparsely and loosely
tomentose. The types of 7. subsessilis and T. maritima are extremely
alike. The under surface of the leaves has a very sparse inconspicuous
indument of scattered slender appressed hairs. The fruit in the
present species is glabrous or, at least, glabrate.

12. T, candidula (Miers), comb. nov. Messerschmidtia candidula
Ann. & Mag. Nat. Hist. ser. 4, ii. 202 (1868) and Contr. Bot. i. 213
(1869). 7. lanuginosa Vaupel, Notizbl. vi. 183 (1914). 7. sericea of
DC. Prodr. ix. 524 (1845) and Fresen. in Mart. Fl. Bras. viii. pt-],
54 (1857) as to Brazilian plant, not Vahl.

In eastern Brazil from Maranhao to Rio Janeiro.

Restinga de Benifica near Fortaleza, Sept. 1910, Ule 9097 (BD, TYPE of I.
nuginosa; K, 1soTyPE); Ceard, 1926, Bolland (K). PERNAMBUCO: CO
dull yellow, Iguarassu, Dee. 2, , Ridley, Lea & Ramage (BM); Itap tee
1887, Ridley, & | e (BM); Olinda, Feb. 5, 1925, Pickel 758, #W
Paul. 18130 (G); fruticose, ca. 1.5 m. tall, in dry bushy places, Pernamb

ov. 1837, Gardner 1078 (BM, Type of M. candidula; , K, IsOTTPAG
Banta: St. Thomé betw. Jacobina and Villanova, Blanchet 3789/37 (BM, oo
BB); Bahia, Blanchet 3350 (K, BM, BD, DC, BB). Rro Janetro: “Sao J

a Barra near Campos,” Glaziou 11293 (K, BD).

A very distinct species readily recognized by the characters
white pubescence which covers the under surface of the leaves ®
clothes the corolla-tube. While the lower leaf-surface is white tome
tose, the upper is green with an ashy cast resulting from scattered
pale appressed hairs. The fruit is ashy with a fine pubescence.

1. %. villosa Salzm. ex DC. Prodr. ix. 524 (1845); F resen. 10 eat
Fl. Bras. viii. pt. 1, 52 (1857). Messerschmidtia villosa Miers,
& Mag. Nat. Hist. ser. 4, ii. 202 (1868) and Contr. Bot. ii. 213 (IS!
a: Pohlit Fresen. 1. c. 52. M. Pohlii Miers, |. c. 203 and |. ig :
T. villosa, f. rubriflora Wawra, Bot. Reise Maximilian nach Bras.
68 (1866). 7. Chamissoniana Vaupel, Notizbl. vi. 182 (1914).
speciosa Vaupel, |. c. 185.

In southeastern Brazil from Bahia to Rio Grande do Sul.

STUDIES IN THE BORAGINACEAE 85

BRAZIL. Banta: Bahia, Blanchet 215 (DC) and 2202 (BM, BD); indefinite,
corolla pale yellow, Salzmann 369 (DC, TypE of 7 villosa); indefinite, Salz-
ma k

); woody vine, corolla yellow, Therezopolis, Oct. 2, 1921, Holway
); a climber, Imbuhy, Organ Mts., 1837, Gardner 546 (K, BM);
Organ Mts., Wilkes Exped. (US); climbing in woods on the Corcovado, Sept.
1897, Ule 4465 (BD, typr of T. speciosa); Rio Janeiro, Jan. 4, 1887, Schenck
1853 (BD); Rio Janeiro, 1844, Widgren (RS). Sio Pavno: Campinas, Novaes
ini v. 17, 1887, Loefgren 387, Mus. Paul. 11259 (G).
THARINA: Santa Catharina, June 1868, Mueller 147 (K). Rio Granpe
bo Sut: Campos to V. Vittoria, Sellow 303 (BD). INDEFINITE: Brazil, Claussen
412 and 417 (BB), Sellow 226 (BD), Sellow (BD, ryrx of T. Chamissoniana),
Bowie & Cunningham (BM), Pohl. in Herb. Mus. Vind. 1597 (K, BD, 1so-
TrPES of 7’. Pohlii).

446 (US
1183 (US

Among its relatives, those with a stiff inflorescence and crowded
flowers, this species is quickly distinguished by its leaves which are
velvety above and tawny with a conspicuous silky tomentum beneath.
The type of 7. villosa is more conspicuously pubescent than is common
m the species. Its stems are somewhat shaggy and the under leaf-
surface is quite silky. The type of 7. Chamissoniana is quite similar
as to pubescence although the primary branches of its inflorescence,
like those in all the other specimens I have referred to the species, are
hot so elongate as they are in the type of 7. villosa, The types of
I. Pohlii and T. speciosa are quite alike and represent the common

l4. T. rubicunda Salzm. ex DC. Prodr. ix. 526 (1845); Fresen. in
Mart. Fl. Bras. viii. pt. 1, 54 (1857). 7. psilostachya, var. rubicunda
Chodat, Bull. Herb. Boiss. ser. 2, ii. 816 (1902). 7. Martti Fresen.

Hist. ser. 4,194 (1868) and Contr. Bot. ii. 205 (1869). M. vicina Miers,
l. ¢. 197 and |. c. 208. M. valga Miers, |. c. 198 and 1. c. 209. T.
Herzogii Vaupel, Mededeel. Rijks Herb. xlvi. 10 (1922).

Along the east coast of Brazil (Pernambuco to Rio Grande do Sul)
and westward through Matto Grosso, Paraguay and Corrientes to
Tucumén, Jujuy and southern Bolivia.

BRAZIL. > OW ickel 935, Mus. Paul.
id (G). P Spor ag oak an sheet banda of is S. Franci near
tall, 0, Met 1888, ie 1362 (BM, TYPE of M. vicina); bush ca. ca m.

}; Near Bahia, Blanchet 468 (BEIy ion DG, rrPe

: ’ 462 (BM), Salzmann (K), Salzmann 272 ( J, TY
a ‘cunda), Lhotsky (DC) ae Malme (RS); near Ilhéos, Martius (K,
ean Rio Janzrro: Rio Janeiro, 1847, Forssell (RS). Rio GRANDE vo Sut:
*nder climber in hedges near Porto Alegre, Tweedie (K); corolla greenish

86 JOHNSTON

prc Pa brick-red, pra Rident near Cu Bote Ded alme 12068
( ; flood plain of. river near Cuyabd, Oct. 28, 10, Malme 2538 (RS).
PA sa Pose rién, Nov. 2, oneda 4140 (G, K ‘ 5 ;

Fiebrig $23 and 529 (G, K, BD, Hass).
RGENT Misronzs: Posadas near ‘“ La ie Nov. 1907,

1788 (RS). Corrrentes: Isla Apipé ea March 2, 1883, Niederlein “BD;
forest of Riachuelo near Corrientes, Jan. 19, 1883, N: iederlein (BD). Formosa:
Pozo del Tigre, _ 1928, Parodi 8508 (G); indefinite, Jorgensen 2247 (G).
Satta: San José, Feb. 1873, Lorentz & Hieronymus 241 (NY, BD, RS); Oran,
Feb. 1916, i ad 26/84 (G); Monte Alto, Ordn, Nov. 19, 1913, Rodrigues
1131 (GQ); corolla dark purple, Rio Piedras, Ordn, Nov. 30, 1911, Rodriguez 130

>
oF
e

and 10982 (G). Jusuy: Ledesma, Jan. 22, 1906, Spegazzini (G). Tu

pot sich aly El Cadillal, 650. m., Nov. 24, 1907, PVillo ? 206 (G).
BOLIVIA. Santa Cruz: Samai vata, prov. llegrande }

BD). Cuaco: Villa Montes, Rio Pileomayo, Nov. 1910, Herzog 1136 (BD,
PE of T. Herzogii); Villa Montes, Pflanz 2086 (BD); Tatarenda, Apr. 4

1902, Fries 1565 (RS). Tarisa: Bermejo, Nov. 17, 1904, Fiebrig 2122 an

(
2136 (G, BD). La Paz: race Nor Yungas, Dec. 1917, Buchtien 641
(G); ? Sorata, Mandon 389 (G, K , RS).

has the leaves practically glabrous and entirely lacks the disks
thickened cells. The specimens from Centurién, Paraguay (F webrig
41 40) and Sorata, Bolivia (Mandon 389) are aberrant in the vely
copious strigosity on the leaves. The remainder of the specimens
cited are rather uniform. Among the species of our area 7. r

is probably closest to 7. breviflora from which it is separated by its
much less freely climbing or almost erect habit of growth, its lack of
axillary inflorescences, its usually conspicuous strigose pubescence

its peculiar disks of thickened cells on the upper leaf-surface-

real relation of the species is, however, with 7. psilostachya HBK.,.a
extremely variable and very poorly understood species of northwester?
South America. Although 7. rubicunda may be only a geogra
variety of 7. psilostachya, characterized by its very detached rang®
its coarser and usually less copious pubescence, and the strong devel :
opment of the hair-bases, I am inclined to believe that it had best be |

STUDIES IN THE BORAGINACEAE 87

treated as distinct until the variations and intricacies of the relation-
ship of 7. psilostachya are very much better understood.

15. T. Salzmanni DC. Prodr. ix. 524 (1845); Fresen. in Mart. FI.
Bras. viii. pt. 1, 51 (1857). Messerschmidtia Salzemanni Miers, Ann.
& Mag. Nat. Hist. ser. 4, ii. 201 (1868) and Contr. Bot. ii. 212 (1869).
lr. vulgaris Salam. ex DC. 1. c. in synonymy. 7. Salzmanni, var.
paraguariensis Chodat, Bull. Herb. Boiss. ser. 1, vii, append. 1, 78
(1899). 7. Salzmannii, var. Blanchetiana Chodat & Hass. Bull. Herb.
Boiss. ser. 2, v. 482 (1905).

From Ceara to Rio Janeiro in eastern Brazil and westward into
Paraguay, southern Bolivia and the northern states of Argentina;
apparently also in extreme northern Brazil.

RAZI 8
892 (K, . ods r Crato, Oct.
1838, Gardner 1785 (K, BM). PrrNamBuco: Tapera, 1926, Pickel 1137 (BD).
Banta: St. Thomé betw. Jacobina and Villanova, _ 3787 (BM, a
2 of v:

(K); corolla pale yellow, , 1830, Salz YPE
manni); Bahia, Lhotsky (DC), Lockhart (BM), Blanchet 125 (DC) and Blanchet
215 (BM) ANEIRO: near Rio Janeiro, Glaziou 112 ;

PA Corolla yellowish green, Colonia Risso near Rio Apa, Sept
25, 1892, Malme 986 D, RS); betw. Rio Apa Rio Aquidabin, Fiebrig
prov (G, K, , ass); corolla yellowish gr ntral, Dec
ripe Hassler 6532 (BM, Hass, BB); west bank of Rio Pileomayo, G
at. 23° 20-30’, O ojas in Hassler 2398 (BM, BD, Hass); corolla

. ’ ct. 190 BM
{ellowish green, thickets near Concepcién, 1901-2, Hassler 7430 (BM, BD,
BB); Asuncién, Oct., Morong 800 (G, K, BM); corolla yellowish green,
Asuncié6n nsa 20 5

corolla yellowish green, woods near Altos, Aug., Hassler 891 (BB, Type of var.
Loc , K, ,B ass, ISOTYPES); corolla olivaceous, w:
ardino, Nov. 1897, Hassler 3460 (G, K, BM, BD, Hass, BB);
Yann sPucay, Sept. 1913, Hassler 12152 (G, BM, BD, Hass); near Lake
ARGS arch 1913, Hassler 12142 (G, BM, BD, Hass).
F GENTINA. Formosa: San Francisco de Laish{, 1923, Herb. M
H ia Beni

. o oT ar a. , *
Schr corolla greenish, Tembaréacién, Rio Bermejo, Oran, 500 m., Nov. 25, 1925,
reer 5044 (G). . Susur: Ledesma, Jan. 22, 1906, Spegazzini (G); corolla
'G); Sez sierra de Calilegua, Ledesma, 800 m., Oct. 12, 1927, Venturi 5392
(G)’ terra de Santa Barbara, San Pedro, 750 m., Oct. 17, 1929, Venturi 9638
BORIVIA. La Pas: Coripati Bang 2076 (G
: : pati, Nor Yungas, March 1, 1894, Bang G,
pi BM, BD); Milluguaya, Nor Yungas, 1300 m., : , Buchtien
(G). Tania: Villa Montes, Pflanz #181 (BD). Gran Cuaco: Tata-
April 5, 1902, Fries 1532 (RS).
Pte Species is very closely related to 7. floribunda HBK. of northern
It a America but differs in having corollas about twice as large.
also has a detached range. It is also closely related to 7. membrana-

88 JOHNSTON

cea, of the region about Rio Janeiro, from which it differs in its larger
more acute leaves, more darkly tawny pubescence of the foliage,
evidently larger corollas and more elongate corolla-lobes.

16. T, membranacea (Gardn.) DC. Prodr. ix. 530 (1845). Messer-
schmidia membranacea Gardn. in Hook. London Jour. Bot. i. 181
(1842); Miers, Ann. & Mag. Nat. Hist. ser. 4, ii. 200 (1868) and Contr.
Bot. ii. 211 (1869). 7. laxiflora DC., 1. c. 525. 7. floribunda of Fresen.
in Mart. FI. Bras. viii. pt. 1, 51 (1857), in pt.

Known only from near the coast in the states of Rio Janeiro and
Sao Paulo, Brazil.

rae ZIL. Rio Janerro: Restinga de Copacabana, Dec. 17, 1886, Schenck
1575 (BD); corolla greenish, subscandent, Ilha da Governador, Nov. 5, 1825,
Bavhot 1105-112 (K); flower yellowish’ green, hills west of Catumbi, Dec.
29 ad cheb 1312 (K) ; scandent, 15-18 dm. tall, betw. Sao vio

m i
SOTYPEs); aque 1 Dus With ; or,
Gaudichaud 1 830 (DC, TYPE of T. laxiflora; BD, 1soryPn), Vavth ier 236 (DC,
flora), ug rin 531 bis (DC), T wee Sone, (K), Claussen
61 (i (BM), Sellen L938 (BD), Ule 827 (BD), Meyen (BD), Rudio (BD), Wi
( Anderss ), Gay (N } d. (UI 3 oP.

(G). Inperirre: Brazil, Bowie & Cunningham (BM), Banks & Solander
(BM) and Herb. Mus. Vind. 1594 (G, K, BD).

While it is possible that this may be only a local form of the widely
distributed 7. Salzmanni, differing in its much less tawny pubescenee,
smaller flowers and much less elongate corolla-lobes, it seems to be
closest to 7. floribunda HBK., a species to which it was referred by
many past authors and from hel in fact, it ee to differ only
in its detached range and its more obtuse leav

17. T. salicifolia (Gardn.) DC. Prodr. ix. 530 (1845); Fresen. in
Mart. Fl. Bras. viii. pt. 1, 56 (1857). Messerschmidia salict
Gardn. in Hook. London Jour. Bot. i. 181 (1842); Miers, Ann. & Mag:
Nat. Hist. ser. 4, ii. 195 (1868) and Contr. Bot. ii. 206 (1869). :
volubilis, var. hirsuta Fresen. 1. ¢

Known only from the meh Rio Janeiro, Brazil.

MERAZIL. Rio JaNnerro: Botafogo Bay, Sept. 1836, Gardner 61 (Bi)
Morro cern, Sept. 1836, Gaviner 8. 81 (BM oe of M. salicifolia; G, ”
IsoTYPES); M go, Miers 3596 (K); Morro Flamengo,

GSD); eames BD); north pe of Pao de Assucar, March 14, 1887, 2, Ak PS
; of Pedra Dois Irmios, March, 23 1920 nee 2137 oar
vanes 1837, Tweedie 1235 (K); Rio Janeiro, pet Pohl i”
: Brazil, 1814-17, Bowie Cunningham 91- ne (BM); Brazil,
Bet Men Vind. 1596 (K, BD).

STUDIES IN THE BORAGINACEAE 89

This is a reasonably well defined local species, which is known only
from the vicinity of the city of Rio Janeiro. It is probably most
closely related to 7. membranacea, which also occurs in the same
region, but differs in its lanceolate leaves, more congested inflores-
cence, slightly larger flowers, more elongate calyx-lobes and spreading
pubescence. —

UNRECOGNIZED SPECIES.

Heliotropium erectum Vell. Fl. Flum. 69 (1825) & Icones ii. t. 42
(1827).—Perhaps a species of 7. ournefortia, but if so, one not known to
me. The type came from the region about Rio Janeiro.

Tournefortia Blanchetii DC. Prodr. ix. 524 (1845).—The type of
this species was collected by Blanchet (no. 1914) in Bahia. Besides
the type I have seen an authentic specimen (lacking the inflorescences)
in the herbarium of the British Museum. I am uncertain as to its
correct disposition. The flowers and inflorescence suggests 7’. ruln-
_ cunda and it may be that the plant is only a very strigose, broad-leaved
form of that species. The only other species particularly suggested
by T. Blanchetii is T. Gardneri. That latter, however, has somewhat

gapotamica Spreng. Syst. iv. pt. 2, 66 (1827); DC.
Prodr. 1x. 530 (1845).—I know this species only from the short descrip-
tion, Which is as follows, “T. foliis oblongis obtusis glabris medio
dentatis subtus flavicantibus, spicis axillaribus elongatis, calycibus
verticillatis hirsutis. Rio grande. Sello.”” The identity of the species
‘$a complete mystery to me.

Il. TAXONOMIC NOTES CONCERNING VARIOUS BORAGES.

HEuioTRoPIuM PROCUMBENS Mill. Dict. ed. 8, no. 10 (1768). H.
americanum Mill. Dict. ed. 8, no. 11 (1768). H. brasilianum Roth,
Nov. Pl. Sp. 103 (1821). H. rigidulum DC. Prodr. ix. 540 (1845).—

© at the British Museum of Natural History last year I made
Pitas search for Miller’s type of H. americanum and eventually
ood, It. It is specifically identical with H. procumbens as is also H.
Botannn yy the type of which I found in Roth’s herbarium at the
of Ppa Museum in Berlin. I have also examined the type-material
cae nigidulum at Geneva and can now definitely place this obscure
Maan the synonymy of H. procumbens Mill. also.

ELIOTROPIUM LEIOCARPUM Morong, Ann. N. Y. Acad. Sci. vii.

90 JOHNSTON

168 (1892). H. minarum Glaz. Bull. Soc. Bot. France lvii. Mem. 3e,
479 (1910).—I have examined the authentic material of H. minarwn
found in the Delessert Herbarium at Geneva. This species, scarcely
more than a nomen nudum, falls into the synonymy of H. leiocarpum.
It is rather slender, loosely branched and has rather thinnish leaves
but seems to be clearly referable to H. leiocarpum. The plant is most
interesting since it represents a conspicuous range-extension. No

Pp (§ ys) utum, sp. nov., annuum prostra-
tum multicaule foliosum pilis rigidiusculis ca. 0.8 mm. longis ad-
pressis pallidis sparse vestitum; caulibus gracilibus laxe ramosis 5-20
em. longis ad 1 mm. crassis, internodiis usque ad 2 cm. longis; foliis
alternis vel inferioribus rariter suboppositis ad apicem versus caulis
paullo reductis; lamina ovata vel oblongo-lanceolata 3-8 mm. longa
1.5-4 mm. lata concolori firmiter herbacea costata sed enervata
integerrima apice acuta basi in petiolum gracilem 0.5-1 mm. longum
plus minusve abrupte contracta; floribus distantibus inter euphylla
secus ramulos extra-axillaribus; pedicello maturitate ascendenti ad
apicem versus plus minusve incrassato 2-2.5 mm. longo imam ad
basem tarde deciduis; corolla 5-6 mm. longitudine et diametro infu
dibuliformi extus faucibus et medio parte loborum sparse adpresse
villosa; tubo ca. 1.5 mm. longo ad 1 mm. crasso intus supra insertionem
staminum puberulento calyce conspicue breviori; faucibus gradatim
ampliatis intus glabris tubo duplo longioribus; lobis ovatis ascendenti-
bus 1.5 mm. longis distantibus, dentibus inter lobos late deltoide's
plicatis ca. 0.3 mm. longis; staminibus tubo corollae inclusis m A
versus tubi insertis; filamentis brevissimis; antheris ca. 1.2 mm. ongis
ovato-oblongis supra stigma fornicatis, apice truncatis cohaerentibus

glanduloso-puberulentis; stylo ad 0.5 mm. longo gracili, disco sug

matoso angustissimo ad 1 mm. diametro, columna stigmatis gra“
stylo subaequilonga; ovario villosissimo; fructibus conspicue rostratls
et ad basem versus quadri-cornutis quadrangulatis maturitate
mm. crassis 3-4 mm. longis in nucula 4 uniovulatas disruptis lobis
calycis paullo brevioribus; nuculis dorso leviter concavis adpres?
villosis, ventre cum faciebus duabus planis arcte contiguis.—)

Ixpra: in a rice-field near Mangalor, Canara, March, PI. India Or.
Ed. Hohenacker, 1851, no. 82 (Type, Brit. Mus.).—This is @ V
interesting and very distinct new member of the Subsection Agillant,
Johnston, Contr. Gray Herb. Ixxxi. 48 (1928), a group hee
thought to be endemic to America. The species is characte

by its very villous fruit which is not only drawn up into a very

2-2.5 .

|
,

:

STUDIES IN THE BORAGINACEAE 91

developed beak but is also provided with evident blunt cornute pro-
cesses that develop above the base on each of its four angles. These
processes are duplex in structure being composed of prolongations
of adjacent portions of the margins of the closely contiguous nutlets.
I know of no other species in the entire genus that has similar out-
growths on the fruit. The type has been determined as H. marifolium
Retz, a species which it somewhat simulates in habit but from which
it differs widely in inflorescence and fruit. Heliotropium cornutum
must be very rare or local for otherwise it would be very difficult to
understand how such a very striking species could go so long un-
recognized.

It should be noted that H. cornutum is not the only representative
of the Subsection Azillaria in the Old World. This very character-
istic group also has throughly typical members in Africa and Mada-
gascar. As with the Indian species these have also been confounded
with H. marifolium. I take this opportunity, therefore, to give their
‘synonymy and to cite such specimens of them as I have seen. They
are as follows,—

Heliotropium madagascariense (Vatke), comb. nov. volvulus
madagascariensis Vatke, Linnaea xliii. 522 (1882); Hallier, Bot. Jahrb.
Xvi. 536 (1893) and Bull. Herb. Boiss. vi. 723 (1898).—MapDaGascar:
flowers white, in locis sterilibus, Mojanga, lat. 15° 43’, June 1879,
Hildebrandt 3035 (Berlin) —Of the three species of the Azillaria in
the Old World this one seems to be closest to the American species,

eo to H. antillanum Urban. From that species, however, it is
Teadily

—e A. Baclei differing in its looser more evident pubescence, more
a habit and much smaller fruit.

Seka’ it Pt—Arnica: Quoja, West Africa, Bacle (DC, TYPE);
ara 1928, Hayes 543 (Kew); Sierra Leone, 1915, Thomas 8836
(Kew); dry sandy places, Makunde, Limba, Sierra Leone, April 1892,
Scott-Elliot 5717 (Kew, Brit. Mus.).

‘Topium Baclei DC., var. rostratum, nom. nov. H. katan-
aie ake in De Wild. Etudes Fl. Katanga iii. 223 (1903); Baker,
Bot Pn Afric a iv. pt. 2, 43 (1905). H. nigerinum A. Chev. Explor.
big: Oceid. Fr. i. 450 (1920), nomen. H. marifolium of Baker,
Dis og pt.—AFrica: near water, flowers bright yellow, Sesheke

3 Rhodesia, Gairdner 186 (Kew); wet sands near river, flowers

92 JOHNSTON

yellow, Victoria Falls, Zambesi, Feb. 1906, Allen 289 (Kew); flowers
bright yellow, Elephant Marsh, North Nyasaland, Dec. 1887, Scott
(Kew); flowers yellow, dry sandbank of the Ugala River, Tanganyika
Terr., March 29, 1882, Béhm 123 (Berlin); Lukafu, Katanga, Verdick
174 (Berlin); Mokele, Belg. Congo, Sept. 1914, Vanderyst 4591 (Brit.
Mus.); Kasai, Belg. Congo, Sept. 1914, Vanderyst 4629 (Brit. Mus.);
dunes along streams in the marigots of Sudan, 1880, Lécard 42 (Berlin;
Delessert); dunes along the Niger near Sebi, Sudan, July 11, 1899,
Chevalier 1168 (Kew).—The African material of the Avillaria seems to
represent a single species. In its typical form it is confined to west
Africa and is characterized by having its fruit broadly conical and
not long-beaked. The forms of the species in central Africa, however,
have the fruit drawn up into a very long and conspicuous beak.
Because of the geographical correlation of this fruit variation I have
thought it worthy of varietal recognition. The forms with beaked
fruit have been published upon as H. katangense and H. nigerinuil.
In treating them for the first time as a variety, however, I have taken
the opportunity of applying to them a really descriptive name. The
varietal name may be considered as having the same type as 4:
katangense Giirke.
icostoma calcarea (Vatke), comb. nov. Heliotropium calearewm
Vatke, Linnaea xliii. 318 (1882). S. albidum Franchet in Révolh
Fauna et Fl. Pays Comalis (Sert. Somal.) 46 (1882). S. verrucosult
Beck. in Paulits. Harar 457, fig. 3-6 (1888). H. albo-hispidum Baket,
Kew Bull. 220 (1895). 8. strigosa Deflers, Bull. Soc. Bot. France xl
120 (1896). H. Vatkei Baker in Thiselton-Dyer, Fl. Trop. Africa)’:
pt. 2, 39 (1905).—This species is known only from Somaliland and
southwestern Arabia. It is very closely related to S. Kotschy!
(Boiss. & Hohen.) B. & -H. ex Jackson, Index Kew, iv. 885 (1895);
from the head of the Persian Gulf and may be no more than @ variety
of that species distinguished by its coarsely tuberculate rather than
smoothish nutlets. Sericostoma calcarea varies in the size of its leav®
and somewhat also in the size of the slightly irregular corollas, :
by no means sufficiently to justify the recognition of the species tha
T have listed above in synonymy. The synonyms that were dé
as species of Heliotropium have been universally accepted mn
latter genus, despite the fact that even a most hasty dissection

have revealed their true generic affinities. It perhaps should a

mentioned here that although H. calcareum Vatke and 5. “able
Franch. were both published in 1882 the former name has indubt

priority since Franchet, |. c. 8, cites Vatke’s paper in his bibliogr™P*’
a arenaria(Vatke), comb.nov. Heliotropium aren” —

;

STUDIES IN THE BORAGINACEAE 93

Vatke, Linnaea xliii. 319 (1882).—This remarkable species is known
only from the type collected on the coastal hills near Baraua, Somali-
land. It is characterized by its congested inflorescence, shallowly
lobed calyx, oblanceolate leaves and its firm oddly shaped corollas.
The corolla is coarse and herbaceous in texture and is pustulate out-
side. The tube is cylindrical but the faux is globular.

MACROMERIA LONGIFLORA Don, Edinb. New Philos. Jour. xiii. 239
(1832) Onosmodium longiflorum Macbr. Contr. Gray Herb. xlix. 21
(1917). M. discolor Benth. Pl. Hartw. 49 (1840). O. discolor Macbr.
l. e. 20.—I have examined the type of Don’s M. longiflora, which is
how preserved in the British Museum. It is unquestionably the same
species as that which was subsequently described as M. discolor Benth.
Don’s type, which was a part of the Ruiz Herbarium formerly owned
by Lambert, bears the manuscript name, Lithospermum longiflorum.
Another old specimen of this very same species, also bearing the name
L. longiflorum, is to be found among the Pavon collections in the
Boissier Herbarium. Number 905 of Sesse & Mocifio’s plates in the
library of the Delessert Herbarium apparently also represents the
Species. Although the old specimens from the collections of Ruiz and
of Pavon bear the name Lithospermum longifolium I rather doubt
that they are the same as the plant described under that name in

€ & Mociiio’s posthumous work, the Flora Mexicana 29 (1894).
It seems searcely likely that M. discolor, i. e. M. longiflora, which is
characterized by its closely appressed, almost strigose, pubescence
would be described as having hirsute stems and leaves.

In my treatment of Macromeria, Contr. Gray Herb. Ixx. 14 (1924),
identified M. viridiflora DC. with M. Thurberi (Gray) Mack. This
disposition I am now inclined to doubt. Macromeria viridiflora is
root entirely upon number 904 of the De Candolle set of copies from

this plate,

+ eret to Say that this opportunity did not assist me in definitely
Pacing M. viridiflora. 1 can only suggest that plate no. 904 may

thi tenes hence unrecognizable. With more data at hand regarding
tine i Pution of M. Thurber and with more information as to the
could of Sesse and Mocifio I am inclined to believe now that they
Co; sane have collected the northern species, M. Thurberi.
co) ey until authentic material of M. viridiflora is seen or until
904 a made in central Mexico are found which agree with plate
0 ve that the name M. viridiflora had best be treated as too
“ure and doubtful for recognition.

94 JOHNSTON

tantha mendocina, sp. nov., annua ca. 15cm. alta ascendenter
ramosa hispida; foliis linearibus vel lineari-spathulatis 1.5-3 em. longis
1-3 mm. latis ad apicem versus latioribus abundanter minu
pustulatis apice obtusis vel subrotundis superioribus paullo reductis;
spicis solitariis vel rariter geminatis breviter pedunculatis 34 em.
longis evidenter foliaceo-bracteatis, bracteis calycibus plus minusve
duplo longioribus; corolla inconspicua 2-2.5 mm. longa alba, lobis
ca. 0.5 mm. longis ascendentibus; calycibus fructiferis laxe dispositis
ascendentibus 5-6 mm. longis ad basem versus 2-2.5 mm. erassis
paullo asymmetricis tarde vel vix deciduis basi rotundis; pedicellis
evidentibus rigidis 1.5-2 mm. longis, lobis subulatis vel subulato-
linearibus conniventibus sed ad apicem versus recurvatis imam ad
basem connatis marginibus breviter hispido-villosis, costa lobi valde
prominenti crassa indurata setis 1-1.5 mm. longis rigidis pallidis hor-
rida; ovulis 4; nuculis 4 evidenter heteromorphis; nucula axillam
maxima persistenti pallida 2-2.3 mm. longa ca. 1.1 mm. lata lanceolata
ovata vel anguste ovata abundanter minuteque tuberculato-granulata
et evidenter tuberculata papillatave, basi obtusa, lateribus rotundis,
apice acuta, ventre ca. 34 longitudinis ad gynobasem quadrangular-
columnarem 1-—1.2 mm. longam affixa, sulco paullo asymmetrico ad
apicem versus nuculae clauso ad medium versus in areolam latam vis
excavatam abrupte expanso; nuculis 3 homomorphis deciduis crassis
1.8-1.9 mm. longis paullo lucentibus ovato-oblongis inconspicué
minuteque tuberculato-granulatis et evidenter papillatis, dorso ™
tundis, marginibus rotundis, lateribus subplanis, ventre ca. 34 longr
tudinis ad gynobasem affixis, sulcis asymmetricis apice nuculae clausis
paullo infra apicem in areolam cuneatam vel lanceolatam evidenter
excavatam ampliatis; stylo ca. 0.8 mm. longo nuculis 3 homomorphis
superanti nuculae axillari aequanti vel rariter paullo superantl.—
ARGENTINA: dry place, Potrerillos, Prov. Mendoza, 1400 m. alt., Oct
20, 1907, Corn. Osten 5129 (tyPE, Herb. Boissier; fragments, Gray
Herb.).—This species, while very distinct from any of the

South American members of the genus, is very closely related to and

perhaps no more than varietally distinct from the North Americ!
C. minima Rydb., a plant which ranges almost exclusively east ofthe
Rocky Mountains and extends from northern Texas northward ee!
Canada. The Argentine plant differs from its northern relative .
being a more slender plant, having slightly smaller less consp!

thickened fruiting calyces and. in having evidently tubereulate 1

papillate odd-nutlets. In C. minima the odd nutlet, i. e. the ®
one, 1s minutely and finely tuberculate- or papillate-granulate ®%
C. mendocina but lacks the large coarse tuberculations present ™,

o

i ciate

STUDIES IN THE BORAGINACEAE 95

southern species. Among the species of South America C. mendocina
may be quickly distinguished by the very evident heteromorphy of
its nutlets, the axial one of each fruit differing from the remaining
ones not only in persistence but in size, shape, color and roughenings.
It is of particular interest being clearly a member of the Texanae,
Johnston, Contr. Gray Herb. Ixxiv. 54-60 (1925), a group heretofore
thought to be confined to North America.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

XCIII.

SPECIFIC SEGREGATIONS AND IDENTITIES IN SOME
FLORAS OF EASTERN NORTH AMERICA
AND THE OLD WORLD.

By M. L. Ferna.p.

Reprinted from Ruopora, Vol. 33,
February, 1931.

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY.—NO. XCIII

SPECIFIC SEGREGATIONS AND IDENTITIES IN SOME
FLORAS OF EASTERN NORTH AMERICA
AND THE OLD WORLD!

M. L. FERNALD
(Plate 204)

I HAVE assumed that I am expected to base my discussion primar-
ily on some features of phytogeography which I have personally ex-
amined, rather than to attempt extensive generalizations concerning
floras with which I am absolutely unfamiliar. Granted this assump-
tion, I must start with a consideration of problems centering on &
limited area of eastern North America, the region of the Atlantic
slope extending from the Labrador Peninsula to the Atlantic States
and the Mississippi Valley (map 1). This district, however, is suffi-
ciently extensive clearly to exhibit large areas which, since the dawn
. re, have had quite different physiographic and floristic

8

Viewed from the standpoint of availability for occupation by flower-
ing plants, the oldest large section of the region is the southern half
oe Appalachian Upland, extending from central New York to
northern Georgia and northern Alabama, and west of the Mississippi
tepresented by the Ozark Plateau. Never, since it was first occupied

pBtiovan title for the joint Discussion on Geographical
‘na srtton aa 3 its J Retation to the Concept of Species of seers E (Phytogeography
Siete T (Taxonomy and Nomenclature) but, at the invita tion of the
sass Pcemageiy xpanded and presented orp an pel nectar yo in the

[FEpruary

Rhodora

26

| a
ae

of |

Nunatak areas
vd exposed during

since the close of

Cretaceous.

dy ed since the close OF
ene Paleozoic

the Tertiary.

pO Raa 3
Pat

Exposed
[it i the S
Exposed since the clase

: Sa
r at
f

=

BOPrZ

i WK

vailability for Plant-Occupation of Eastern North 4? | 3

Map 1, Periods of A
since the Paleozoic,

1931} Fernald,—Specific Segregations and Identities 27

by Angiosperms, has the Appalachian Upland of the United States
(and Canada) been invaded by seas; and, except for its northern ex-
tension, it lies wholly south of the limits of the Pleistocene glaciation.
During the Cretaceous, while this southern half of the Appalachian
region was covered by land-vegetation, the lower marginal country,
east, south and far to the west and northwest, was submerged under
the Cretaceous seas. In the Tertiary, likewise, much of the low-lying
Coastal Plain was again covered by shallow seas; and, furthermore,
the outer margin of the Coastal Plain is often of very modern or
Quaternary origin.

North of the southern Appalachians and extending north to north-
ern Labrador and the subarctic and arctic continental coast is the
area which was covered by Pleistocene ice, the last advance of which,
the Wisconsin, is estimated to have begun its frontal decay only
25,000-30,000 years ago, while the northern limit of this advance in
eastern America is represented by the living glacial field of eastern Baf-
‘fin Island and the al t continental ice-cap of Greenland Although the
Wisconsin or most recent continental ice-sheet in the median latitudes
of North America and its confluent valley-glaciers together denuded
or modified a vast area, there are several nunatak areas known, where
the last glaciers did not develop or where their work was so restricted
43 to leave essentially undisturbed some regions of mountains, high
Plateaus and precipitous headlands and, in the famous “ driftless
area” of Wisconsin, Illinois, Iowa and Minnesota, even of prairie. In
the area we are specially considering the most notable of these nuna-
taks! or regions unglaciated or only slightly denuded by the last con-
tnental ice-sheet are the Torngat Mountains, just south of the eastern
entrance to Hudson Strait, large areas of Newfoundland, especially
centering on the Long Range near the West Coast, the eastern half
of the Gaspé Peninsula of Quebec, and some other smaller and iso-
lated spots about the Gulf of St. Lawrence.

Temperate eastern North America has, then, an extensive area (the
Southern Appalachian Upland) in which land-plants have had an
°pportunity to spread since the advent of Angiosperms; others (limit-

Portions of the inner Coastal Plain and much of the central plain
of the United States) where living floras could have taken possession
only after the withdrawal of the Cretaceous seas; others (much of the

"For enumeration and discussion of these areas see Fernald, Persistence of

P ,
sg in Unglaciated Areas of Boreal America, Mem. Amer. Acad. xv. No. III. (Mem.
Y Herb. ii.) 295-317 (1925).

28 Rhodora [FEBRUARY

Atlantic and Gulf Coastal Plains) unavailable to land-plants until
the tipping off of the Eocene and the Miocene waters; others (the
nunatak areas) which apparently retained their present distinctive
floras through at least the last glaciation; still another, the largest of
all, the vast region of Canada and the Northern States which has
become available for wholesale occupation by plants only since the
decay of the Wisconsin ice, within the last few thousand years; al-
most as young or in some regions even younger, scattered areas at the
outer margin of the Coastal Plain, of very recent or Quatenary origin;
and, youngest of all, the million or more square miles cleared and
fundamentally altered by the white man during the last three cen-
turies.

These strongly contrasted periods of availability to plant-popula-
tions of different eastern American areas are, of course, closely ee
ed in other regions: for instance, in Europe, with the western third of
the Iberian tableland uninvaded by the Cretaceous and the Tertiary
seas which drowned much of the continent; others, like much of the
Mediterranean basin, largely unavailable until the withdrawal of the
Tertiary seas; and still others, such as southern Scandinavia, “a a
few thousand years removed from submergence under the Grea
tic Glacier. The distinctive floras which characterize the differe
physiographic areas of temperate eastern North America are “
definite, and it is my purpose, in so far as time will allow, to direct
attention to the different degrees of specific segregation shown, esp
cially by the amphigean or the world-wide genera in these different
floras.

In the ancient Appalachian Upland of the United States the =
standing phytogeographic feature is, of course, the great mesoP'
forest of Mesozoic or early Cenozoic genera, many of them formerly
found likewise in Europe, western America and the Arctic but now
restricted to one or more areas of eastern North America, easte™
central or southwestern Asia! or the geologically ancient peninsulas 0

1
ws Rr ne occ Stn ssn Sa

ted out (

see L. Amoen. Acad. ii. 336 (1751). This relationship bas subsequently ot
requently emphasized and many typical examples are well known. Consed! other

sis righ urn. 29
Antiquity and Dispersal of Vascular Plants, Quart. Rev. Biol. i. 222, ee sf
reg Relationships of the Floras of the Northern e Boao Proc. Intern

1931] Fernald,—Specifie Segregations and Identities 29

- —
NP OSS aS,
A - vs
ide Die ley\ \.
Te AN \
7 at rr “~) \ \ \ Ne
at ie ee Th E- ¥
pT A tos _\ -\_ Fothergi la
\ eres
} ines be) /
ae i as é

——
\

if / ui
ae “Buc kleya es fae s
.disty chophylla x / en

sentative, Nevivsta; 4, of Buckieya; 5, of DipHy.veta.

southeastern Europe: Fothergilla (map 2)! with a few excessively rare
species in the southern Appalachians and one on the wooded slopes
of Kashmir and adjacent Afghanistan; Hamamelis, with wide-ranging
eastern American species and localized Asiatic representatives ; Mag-
nolia, highly developed in both southeastern North America and south-
eastern Asia; Neviusia (MAP 3), like an apetalous representative of
the local Chinese Kerria? and known only from a very restricted
station in the foothills of Alabama; Pachystima, Stewartia, Symplo-
carpus, Panax, Shortia, Phryma, Triosteum and scores and scores of
others. It was this vast assemblage in the southern Appalachian
H ' The world-maps used to show ranges of plants in this paper are made on Goode's

. phic Projection, published and copyrighted by the University of Chicago.
Kerria japonica DC., in spite of its specific name, is endemic in China.

30 Rhodora [Fepruary

region as well as the remarkable relics on the Coastal Plain of the
United States which stimulated Guppy, congratulating American
students upon the superabundance of these ancient types still living
and picturing the emotions of a student appreciative of his oppor-
tunity, to write:

“In the woods around him were growing the Liquidambar, the
Sassafras, and other shrubs and trees that had flourished in the Meso-
zoic ages in the spot where he was standing. Their remains crowded
the Cretaceous deposits exhibited in the cliffs near by. Specimens
of the past and of the present were in his hands. Though the differ-
ence in kind was very slight . . . the difference in time, measured in hu-
man lives, amounted to eternity. Itisastory of perpetuity rather than
of change .. . He begins with the cosmopolitanism of such types
in the Cretaceous age and he ends with their more restricted distri-
bution and somewhat greater specialization now. ... Those old
genera become the genera of today; and the genera of today, though
the genera of a thousand ages, are ‘but as yesterday’ in the history of
flowering plants.”

Not only, as Guppy so vividly stated, are these Appalachian gener
of today the genera of a thousand ages; their species are also ancient
and usually sharply differentiated. No one with good material would
fail to distinguish the two or three local Japanese and Chinese species
of Buckleya? from the famously rare American B. distichophylla Tort:
(maP 4), the latter American shrub parasitic on the roots of T suga,
the American Menispermum canadense L. from the Asiatic M. daur-
cum DC., the extremely local American Cladrastis lutea (Michx.) K.
Koch from the Japanese C. platycarpa Mak. and the other Asiatic spe
cies; nor such herbs as the highly localized Appalachian Diphyllew
cymosa Michx. (map 5) from the less localized Asiatic D. Gray" B
Schmidt, or the American Podophyllum peltatum L. from the Asiatie P :
emodi Wall. Such illustrations of this eastern Asiatic-eastern Amel
can generic similarity but specific differentiation could be increas
by hundreds but these must now suffice. a

Turning to genera, subgenera or sections shared by Appalachian
America and eastern or southeastern continental Europe, baie! aga
get sharp specific differentiations: the Old World Platanus
‘ Bi sged 3 ton ted in the Western World: an Appreciation, Am. Journ. ac

n in-
* For memoranda on the Asiatic ranges of Buckleya and some other genera see
debted to the kindness of Dr. Handel-Mazzetti.

an : . TAS
wt —\—\—-\f FR Mee RN Wn ak ‘aon CEL

Map 6, Geographic Range of Piatranvs; 7, of Cercis: 8, of ComaNpRa.

(1g61

salytuep] pues suonesaiseg oytoadg—‘pyeuseg

32 Rhodora [FEBRUARY

L. (map 6) and de American P. occidentalis L. (with some related
trees farther west); the Appalachian Anemone lancifolia Pursh, which
has generally been confused with the European A. trifolia L. but which
differs' in constant details of the flowers; the Eurasian Ceres Siliqua-
strum L. (Map 7) and the American C. canadensis L., with other species
in western North America and in central China; the common American
genus Comandra (Map 8) with a single Eurasian representative, the
comparatively localized C. elegans (Rochel) Reichenb.; and so on

ough dozens of groups, the southeastern European and the eastern
American species usually having western North American represent-
atives as well. In the cases just cited the Appalachian and the Old
World species are sharply distinguished, but there are a few (perhaps
a dozen) cases of Angiosperms (and still more in the Pteridophytes)
where the differentiation of species is not so complete, the eastern
American and the Asiatic plants being considered by some systema-
tists as recognizable species, by others as merely geographic varieties:
such cases as Tovara virginiana (L.) Raf., Polygonum sagittatum L,
Liriodendron Tulipifera L.,2 Phryma Leptostachya L. and Aruneus
sylvester Kosteletz.; while in a few cases, such as Symplocarpus foetidus
(L.) Nutt. and Monotropa uniflora L., search has failed to reveal even
significant varietal differences. Admitting these very few exceptions
(which will become significant when we have examined the histori
relation of the mesophytic Alleghenian flora to the prevailingly xe
phytic and hydrophytic Coastal Plain groups), however, when the
hundreds of species of Appalachian angiospermous genera are com-
pared with their Old World representatives the general conclusion is

1 See Fernald, reget. ans 184 (192

n from
these simple, emarginate ‘leaves’ has convinced me that they are the leafle
—* leaf, a conclusion which had previously been r eached by some other cretar

North American and Europea? can pale
ceous a recently published book on sd Ancestors,’ the Ameri
loam ta ia ~ W. Berry, speaks of the occurrence of Liriodendron in G Greenland bt
I venture to think that there is no satisfactory evidence in support of this

Though it is not possible from the available material to refer the leaflets pies a
recent gen us with complete confidence, the > probability is that they ye Pteroca!

allied L/G40CF

pus, genera with a wide ics and both -
geographical distribution in the tropics at oak preset, —

a similar position in the Leguminosae. "*—Seward, Arctic Vegetation P
Journ. Royal Hort. Soc. |. pt. 1: 15 (1925).

1931] Fernald,—Specific Segregations and Identities 33

apparent: that in nearly all groups the species of the Western Hemi-
sphere are completely segregated from those of the Eastern; that we
here have stable or essentially stable specific entities.

Descending to the broad outer Coastal Plain of Atlantic North
America, the region of silicious and acid peaty soils of Tertiary or
later origin, we come to a change in the flora so striking as immedi-
ately to challenge attention; for, whereas the distinctive non-endemic
genera of the southern Appalachian region are those shared with tem-
perate eastern and central Asia or with southern or eastern Europe,
the Old World elements in the indigenous Coastal Plain flora of At-
lantic North America are prevailingly the groups with wide tropical
and subtropical range, the families, tribes, genera and sections shared
by the Americas with tropical and subtropical Africa, tropical and
subtropical Asia, and Australia but for the most part unknown in
temperate Europe and largely unrepresented in temperate Asia. For
instance, if I may be permitted to leave for a moment the Angio-
sperms and to draw an illustration from the Pteridophytes, the an-
cient genus Schizaca (MAP 9), with 20-25 species confined to the South-
ern Hemisphere and the Tropics, in the Old World unknown north
of Madagascar, the Seychelles, India and the Philippines; in the New
World wanting north of tropical Mexico and subtropical Florida,
but with the single famous exception, Schizaea pusilla Pursh. This
solitary northern species was long supposed to be restricted to a lim-
ited area of the northern Coastal Plain, in New Jersey, but we now
know that the regions of its wide-spread abundance are the peats of
more northern Nova Scotia and still more northern Newfoundland;
S. pusilla on the latter island being a typical species both of the high-
est unglaciated tablelands and of the lower regions recently covered

y Wisconsin ice. Very similar tropical or austral relationships are
shown by the majority of non-endemic genera and tribes of Angio-
opens (and especially of Monocotyledons) on the Atlantic Coastal

m: the Haemadoraceae (map 10) restricted to Australia, South
Africa, tropical America and the Atlantic Coastal Plain north to
Massachusetts; the tribe Conostylideae (map 11) of the Amaryllidaceae
with 50 species in southwestern Australia, 1 in South Africa and 2 in
eastern North America, the American species both excessively local
but one of them extending north to Nova Scotia; the Xyridaceae
(MAP 12) confined to the Southern Hemisphere and the Tropics ex-
pt for an extension northward along the eastern margin of North

34 Rhodora [FEBRUARY

Bie 2 :
ae |)

if

Haemadorace

4
% g fa / 1
Map 9, Geographic Range of Scurzama, S. PUSILEA in ellipse; 10, of the H

1931] Fernald,—Specific Segregations and Identities 35

12, of the Xyrrpacean.,

, Tribe ConosryiipRaAB?:

<

a)

=

-

&

2 °

‘ 5

1S } =

ee i =
ee = a
ol o
vs] z

mn « We oe eee _

‘= eb =
oT ks —y
ee aot S
Bees tee a Bh 7s

C

he
poe

Map 11,

36 Rhodora [FEBRUARY

America quite to latitude 50° in Newfoundland, with a slight isolation
inland in the north about the Great Lakes and an area of disrupted
colonies (of X. torta Sm.) on the 4000-foot silicious tablelands of the
southern Appalachians; and so on with prevailingly tropical or aus-
tral groups like Lophotocarpus, Eriocaulon, Hypoxis, Drosera (MaP 28),
Galactia, Lilaeopsis, the Pontederiaceae, Burmanniaceae, Podostemaceae
and hundreds of other groups.

Although the eastern American representatives of the tropical and
austral groups are there now concentrated on the soils of Tertiary,
Quaternary or even Pleistocene origin, they are by no means restrict-
ed to the comparatively youthful areas. Here and there along the
ancient tableland-crests of the Appalachian system, wherever the
primitive, open xerophytic or hydrophytie conditions prevail and
where the denuding action of the Pleistocene ice was not too severe, We
are learning to expect rare and highly localized members of the so-
called Coastal Plain flora, like Xyris torta (map 31) just referred to.
Through Mesozoic time the Appalachian Mountains of the present
day did not exist as mountains but, as Berkey succinctly puts it,
“The continent stood much lower than now. Portions that are now
mountain tops and the crests of ridges were then constituent parts of
the rock floor of the peneplain not much above sea level. . . . Such
conditions prevailed over a very large region—certainly all of the
eastern portion of the United States.”! The following, from among
the very numerous corroborative statements, are representative of
the consensus of geological opinion. “Jurassic time throughout the
greater part of North America was one of erosion. . . . This ef
sion cycle brought about the final transformation from the old top
graphic expression of high Appalachian . . . mountains to a nearly
base-leveled land.”? “Once more there was prolonged quiet. The
fault-block mountains were persistently attacked by the erosive agents
and were reduced [in early Cretaceous], like their Appalachian. pt
decessors, to an almost plane surface, close to sea-level.”? “ It 8
known as the ‘Cretaceous Peneplain,’ because of its best development
during the Cretaceous period. This vast plain extended over the areas
of the Appalachian Mountains, Piedmont Plateau, all of New Yc
state, the Berkshire Hills, and the Green and [the] White Mountain
- +... The Cretaceous period was closed in eastern North America

1 Berkey, N. Y. State Mus. Bull. no. 146: 67 (1911).
2 Schuchert, Historical Geol. 846 (1915).
* Daly, Our Mobile Earth, 297 (1926).

1981] Fernald,—Specific Segregations and Identities 37

by a disturbance which produced an upwarp of this vast Cretaceous
peneplain with maximum uplift of from 2000 to 3000 feet . . . to pro-
duce a broad dome sloping eastward and westward, and northward to
the Gulf of St. Lawrence and southward to the Gulf of Mexico.”!

And studies by the geologists farther north clearly indicate that the
peneplained or baseleveled condition of the ancient mountain-axes ex-
tended at least to Newfoundland and apparently to northern Labrador,
if not beyond. For example, Alcock, reporting on the Shickshock
Mountains of Gaspé, famous, like the tablelands of western New-
foundland, for having stood undisturbed above the local ice-sheets
of the Pleistocene,” says: “Gaspé peninsula . . . is a region of Appa-
lachian structure, ... The interior of the peninsula is a plateau
dissected by deep valleys ... The most striking feature is, prob-
ably, the mature character of the upland topography . . . the sky-
line is comparatively even in all directions. . . . Gaspé peninsula,
therefore, is a region which must have been base-leveled and, later,
uplifted and dissected... . A youthful topography, represented by
these steep-walled valleys, is superimposed upon the old age topog-
taphy of the upland surface.”? Similarly, Twenhofel, whose inti-
mate knowledge of the physiography of the region is unequaled, says:
“To one approaching Newfoundland from Sidney [Cape Breton] . . . ,
the Most impressive feature is the high flat-topped upland, here ris-
ing almost vertically from the sea. ... If. Newfoundland be ob-
served from the Labrador side, one feature will attract and maintain
the attention: the flat-topped upland, standing boldly and prom-
mnently in view.... The sky line of the Long Range is strik-
ingly horizontal and the appearance of an equal height in all its parts
'S hot a fiction resulting from a distant view, for it remains the same
hear as well as far... . The accordance of the summit levels of
the highlands, . . . the presence of well preserved flat-topped moun-
tains at many localities with the projected plane of their summits

ne oe le no evidence that glaciers worked on the highest parts of the Shicksbock
bec, ©; —Coleman, Physiography and Glacial Geology of Gaspé Peninsula, Que-
' an. Dept. Mines, Geol. Surv. Bull. 34 (1922); ‘‘The Shickshock Mountains,

°ccupied by local glaciers during the Pleistocene, esca he erosion e
continental i ts.'’—Alcock, Across Gaspé, xiv. 208 (1924); ‘‘the
lek ce Part of the Long Range . seems to have remained as an area free from

Which pre-glacial plants could survive.”—Coleman, The_Pleistxene of New-

aco - Geol. xxxiv. 220 (1926).

a i Geology of Lemieuz- Township, Gaspé County, Quebec, Geol. Surv. Can.
m- Rep. 1921, Pt. D, 76, 77 (1922).

38 Rhodora [FEBRUARY

truncating all kinds of structure and rock, . . : these are considered
evidence of the present dissection, but one time perfection of a péne-
plain, a plain of erosion of remarkable perfection extending over the
whole of Newfoundland. ... . On this ancient plain the rivers
were free to wander where they would, . . . . They probably crossed
the site of the present mountain ranges and, when the land arose,
each stream struggled to maintain its position. . . . In the eastern
United States throughout the Appalachians, the existence of an
extensive peneplain, completed before the end of Cretaceous time, is
now universally admitted and with this base level that of Newfound-
land is tentatively correlated, and the period of development and
close of the cycle assumed to be the same.”! Beginning with
close of the Cretaceous and continuing, with minor interruptions,
through Tertiary time this floor of ancient Appalachian rock, only
slightly above sea-level, was gradually uplifted and deeply weath-
ered and sculptured to produce the Appalachian Upland of today;
and as the upwarping of the older land progressed the marginal
region, which had been overlapped by the shallow Tertiary seas, was
also elevated, tipping off the Tertiary waters and extending the
emerging continent a full 100 miles (160 kilometers) eastward to the
margin of the now submerged continental shelf. ‘The American geol-
ogist, Miller, thus puts the matter: “The uplift of the great Creta-
ceous peneplain was an event of prime importance for the easter
United States, because it literally furnishes us with the beginning of
the history of most of the existing relief features of the Appalachian
district as well as New York and much of New England. Hence we
assert, with emphasis, that all the principal topographic features of
this region as we see them today date from the uplift of the Cretaceous
peneplain, because they have been produced by the dissection of that
upraised surface. This dissection was largely the work of erosion.
.... All the valleys, great and small, such as the Champlaim,
Connecticut, Mohawk, Hudson, the Great Lakes valleys, and the
valleys of the Appalachians, have been produced since the uplift of
the peneplain. ”2

In view of this well attested physiographic and ecological change
of the Appalachian Upland, from a baseleveled or coastal-plain status,
with sluggish drainage and innumerable shallow pools and boggy de-

‘ Twenhofel, Physiography of Newfoundland, Am, Journ. Sci. ser. 4, xxxill. 7 1%
and 19 (1912).

* Miller, Hist, Geol. ed. 3: 326-329 (1928). .

|
,
|

1931] Fernald,—Specific Segregations and Identities 39

pressions through much of the Cretaceous, to its present elevated and
deeply dissected condition, it is certainly significant that in favor-
able habitats at different points on the now uplifted ancient peneplain
we find relic-colonies of such tropical or austral genera as Schizaea,
Lygodium, Stenophyllus, Eriocaulon, Xyris, Lobelia and that peculiar
terrestrial group of Utricularia, sometimes treated as a genus, Stomois-
ia, with 50 species in tropical Asia, Africa, Australia, eastern South
America and eastern North America; and mingled with the typical
lowland plants of tropical or austral affinity on the high Appalachian
crests and tablelands we sometimes find similar relic-colonies of
characteristic endemic American genera of the Atlantic Coastal Plain:
Orontium, Xerophyllum, Helonias, Calopogon, Cleistes, Sarracenia,
udsonia, Rhexia, Leiophyllum, Bartonia and numerous others.
Schizaea pusilla of the Pine Barrens of New Jersey and of the peaty
barrens of Nova Scotia and Newfoundland occurs, likewise, on the
highest unglaciated tablelands (the uplifted Cretaceous peneplain so
much enphasized above) of Newfoundland. My first experience with
't was on the diorite tableland of the Blomidon Range, my published
hote reading: “Among the tufts of Scirpus caespitosus . . . was
the famous little fern of the New Jersey Pine Barrens, Schizaea pusilla.
- +.» Here at 2000 feet altitude it abounded over many acres, nestling
in the bases of the Scirpus tussocks.”! Fifteen years later the great
student of Pleistocene geology, Coleman, visiting Blomidon, wrote:
From this [1,560 feet] to the highest point reached (1,700 feet) . . -
€re was no evidence of glaciation.”? Lygodiwm palmatum (Bernh.)
Sw., the only living member of this tropical genus now found on the
Appalachian tablelands, has upland stations scattered from western
New England to Tennessee; Eriocaulon septangulare With., Xyris
@ Ries, Lobelia Dortmanna L. and Utricularia (Stomoisia) cor-
muta Michx. abound in pools on the tablelands of Newfoundland,
there often associated with Schizaca pusilla; the Eriocaulon, Xyris,
Utricularia and Lobelia also occurring, along with the Coastal Plain
Lygodium palmatum, Orontium aquaticum L., Amianthium muscae-
(Walt.) Gray, Aletris farinosa L. and Rhezxia virginica L. on
Witblelands of Pennsylvania.’ Xerophyllum asphodeloides (L.)
and Zigadenus leimanthoides Gray, ordinarily Coastal Plain
Plants, have long been known from the high tablelands of North
Fernald, Ruopora, xiii, 133 1

J

» JOurn, Geol. xxxiv. 210 (1926).
2 -
Porter, Flora of the Pocono Plateau, Ruopora, i. 183-185 (1899).

40 Rhodora | FEBRUARY

Carolina; Leiophyllum prostratum Loudon and L. Hugeri (Small) K.
Sch. are the Appalachian crest representatives of the Coastal Plain L.
buzifolium (Berg.) Ell. Helonias bullata L., for nearly two centuries
supposed to be endemic on the Coastal plain, was found! in 1909 on
the mountains of western North Carolina; and Cleistes divaricata (L.)
Ames (Pogonia divaricata R. Br.), also considered a typical species of
the Coastal Plain, is well known from high tablelands of the southern
Appalachians.? In fact, this relationship of the Coastal Plain and the
Appalachian tablelands has long been known. ‘Thus, in 1879, Red-
field, describing the Carolina mountain trip with Canby, Gray, Hyams
and Sargent, wrote: “on the 12th was made the ascent of Table Rock,
a most remarkable summit, belonging to the Blue Ridge system, pre-
senting in one direction the tabular profile which gives it its name,
and from another a sharp conical outline not unlike that of the Matter-
horn. The botanists returned from it laden with plants, and it was
curious to see among them so many of the species which are associated
with the sandy barrens and swamps of southern New Jersey.”®

In some cases plants of Australian, Malayan and African affinity
are now found in America almost wholly in the North, in the region
which was largely under the Pleistocene ice. This most unlooked-for
disruption of range is well illustrated by Potamogeton, sub § Javaniet
(map 13), the group of dimorphic plants which Hagstroém considered
the most primitive of pondweeds with floating leaves The mem
of this primitive subsection occur locally in Australia, more generally
from Java to Japan and Burma, on Madagascar and across southern
and central Africa; but outside these tropical and subtropical areas
of the Old World the Javanici are known only as two extremely rare
species, so rare that each new station is heralded as a notable range
extension, Potamogeton Vaseyi Robbins and P. lateralis Morong, ©™
fined to a limited region of the northeastern United States and ad-
jacent Canada. A very similar disruption is displayed by Myno-
phyllum, § Tessaronia, the plants of Eumyriophyllum with four, i
of eight, stamens, plants of tropical and subtropical India, Madagas@?
and eastern North America, but with a slight representation also
Pacific North America.

* House, Muhlenbergia, vi. 73 (1910).

* Gattinger, Tenn. Fl. 84 (1887); Small & Heller, Mem. Torr. Bot. Cl. iil. 20- ie
= ' Kearney, Plant World, i. 35 (1897); Wherry, Journ. Wash. Acad. Sei. xvill

* Redfield, Bull. Torr. Bot. Cl. vi. 335 (1879).

1931] Fernald,—Specific Segregations and Identities 4]

Map 13, Geographic Range of Poramoarron, sub § JavVANIcI; 14, of Drospyros.

geton,
vanic!

Tetaine
sub § Ja

42 Rhodora [FEBRUARY

Whether the wide-ranging herbaceous angiospermous groups of
the Atlantic region of North America were Mesozoic or early Cenozoic
genera of the whole Northern Hemisphere, which have subsequently
restricted their ranges principally to the Tropics, we cannot now
absolutely demonstrate. As herbs they have left few if any fossils;
but Carboniferous, Jurassic and Cretaceous members of the Schizae-
aceae occurred in Europe and northern Asia, even in arctic Spitzbergen;!
and the ligneous and consequently easily fossilized genus Diospyros
(map 14), which today has an essentially tropical distribution quite
comparable with the present ranges of Schizaea, Lygodium, Potamo-
geton, sub § Javanici, Lophotocarpus, Xyris, Eriocaulon, the Burman-
miaceae (MAP 18), the Droseraceae (map 28), the Halorrhagidaceae,
tribe Halorrhageae (Map 15) and hundreds of other groups (except
that, unlike them, Diospyros is considered indigenous also from
northern India to the Caucasus and the Black Sea), in mid-Creta-
ceous covered practically the whole Northern Hemisphere, quite to the
Arctic. Berry’s vivid account is to the point: “In that grand dis-
play of dicotyledonous genera which during the mid-Cretaceous re
placed the old Mesozoic flora of ferns, cyeads, and conifers and which
appeared with such apparent suddenness at a number of points in the
Northern Hemisphere, we find unmistakable evidence of the abund-
ance and wide distribution of species of Diospyros. No less than 17
different forms have been described from the rocks of this age, and
the localities where they have been found are scattered from Australia
to Bohemia, Greenland, and Vancouver Island. . . . they seem to
have been especially at home along the Cretaceous coast of the At-
lantic and along the border of the Mediterranean Sea which extended
northwestward from the Gulf of Mexico over much of our present
Great Plains area.”? In other words, their greatest display of Cre-
taceous fossils is in the margin of the area covered by the Cretaceous
seas, which then bordered the baseleveled land which in the Tertiary
became uplifted and dissected to form the present Appalachian Up
land with its characteristic climax mesophytic forest. Very similarly,
ligneous Nyssa (mar 16), today with a disrupted range somewhat
similar to that of Myriophyllum, § Tessaronia, formerly had a b
range across Eurasia and North America, reaching Spitzberge? and

A Syn Potonié in Engl. & Prantl, Naturl. Pflanzenfam, 1.4 372 (1900): also Sewer!
1 Study in Contrasts: The Present and Past Distribution of certain Ferns, Journ.
xlvi, 233, 234 (1922),
* Berry, Tree Ancestors, 257 (1923),

1931]

Fernald,—Specific Segregations and Identities 4

Ves Saks 23
eae fe Ce

See" ad

| g!
Tribe: Halorrha

3

ORRHAGEAE; 16, of Nyssa.

, Tribe Hau

IDACEAE

Map 15, Geographic Range of the HaLtorr#ac

44 Rhodora [FEBRUARY

Greenland. Again drawing upon Berry’s Tree Ancestors: “A large
number of fossil forms of gums have been described . . . found in
Upper Cretaceous deposits in Wyoming, Nebraska, Kansas, and Ala-
bama. During the early Tertiary gums are extraordinarily abund-
ant. ... Itis worthy of comment that all of the Upper Cretaceous
and Eocene gums are North American or Arctic,. . . In the far north
they occurred at that time in Alaska, Greenland and Spitzbergen.”!
This present restriction primarily to the Tropics of groups formerly
wide-spread in the Arctic has been summarized by our President in a
single sentence: “A study of ancient floras reveals the capacity of
plants as travelers; many flowering plants, conifers, and ferns that are
now confined, or almost confined, to the tropics of the southern hemi-
sphere were in the Cretaceous and Tertiary periods represented by
species widely spread over the northern hemisphere, passing far with-
in the Arctic Circle.”? And all of us will call to mind groups, pre-
vailingly tropical today, which are also characteristic of temperate
es North America and which have a fewnorthern representatives
n Europe as well: Eriocaulon, Echinodorus, Fuirena, Fimbristylis,
M ariscus (Cladium), Smilax, the Dioscoreaceae, Hibiscus and Ammanit.
A few species in eastern North America have today ranges quite
comparable, except for minor details, with that of Schizaea pusilla,
admittedly a relic from the boreal Cretaceous or earlier dispersal of
its progenitors. This parallelism of range (on some of the Appala-
chian plateaus, uplifted from their sea-level position of the Cretaceous,
and on the Atlantic Coastal Plain or its margin, with few if any inter
mediate stations) is well shown by one of the most distinct and rarest
of pondweeds, Potamogeton confervoides Reichenb. (PLATE 204; also
MAP 33). This remarkable species has the free ligules and continuous
spike of the comparatively advanced subgenus Eupotamogeton; it
constitutes Hagstrém’s subsection Monticoli of Eupotamogeton, ei
as he has Btated, it exhibits such primitive anatomical characters
to form a “ transition to the [primitive] Coleogeton species” and in in ms
delicate foliage, its branching, its creeping rhizome with subte
tubers, and its long, terminal peduncle it so strongly simulates P.
pectinatus L. (PLATE 204), P. filiformis Pers. and other species of
most primitive subgenus Coleogeton that only by the closest examil”

Fp ii Journ. Royal Hort. Soc. 1. pt. 1: 17 (1925). tenskaP®
3 Hagstrém, Critical Researches on the Potamogetons, Kungl, Svenska ‘2 e
akad. Handl. lv. No. 5: 86 (19.

45

Specific Segregations and Identities

Fernald,

1931]

pulp

peti
'

46 Rhodora [FEBRUARY

ation can it be distinguished from them. It certainly is not a mere
accident that today the obvious relic-species, Schizaea pusilla, on the
uplifted Cretaceous peneplain of Newfoundland (and apparently also
of Cape Breton) should there border the shallow pools in which the
rare and demonstrably primitive Potamogeton confervoides is found;
nor that these two, along with the equally rare and ancient Lyco-
podium carolinianum L., should be close neighbors in that haven of
relic-species, the Pine Barrens of New Jersey, for, surely, the
amazingly disrupted austral occurrence (MAP 17)! of restricted
colonies of L. carolinianum and its immediate allies? is good evi-
dence, if evidence were needed, that that species (in the broad
sense) is of great antiquity. It is also not a mere accident that, ex-
cept for rare stations in the silicious Coastal area, P. confervoides
should be known in Pennsylvania, New York and New England only
in pools on or among the highest mountains (“Alleghany Mountains,”
sent by Schweinitz to Reichenbach; Pocono Plateau, Pennsylvania,
where associated with the relic colonies of Coastal Plain types enumer-
ated on p. 39; Adirondack Mts., New York; Taconic and Green Mts.
Massachusetts and Vermont; White Mts., New Hampshire; Mt. Ka-
tahdin, Maine).

The present-day ranges of the tropical groups which are represented
in temperate eastern North America closely coincide with the exist-
ing continental and insular remnants of hypothetical Gondwana,
the great equatorial land which, in the Permian, is supposed by some
to have stretched from Australia and peninsular India to Africa and
South America. But those who have reconstructed Gondwana Land
(map 19) tell us that, in Lower Cretaceous time Gondwana had become
much disrupted and that by mid-Cretaceous (MAP 20) it had broken
into remote Australia, Lemuria, Ethiopia, and Amazonia. Further-
more, during the supposed existence of Gondwana that equatorial
land has been pictured as having had no direct connection with the
Atlantic United States, and subsequently (maps 22-27)* the gaP ise
tween the southeastern United States and Brazil is said to have been

1 Doubtless a few tropical and austral stations have been overlooked, since “_
“Floras’’ omit the Pteridophyta.

se and L. mere

Rg Including L. drepanoides Blume, L. Drummondii Spring, L goyozen
or yagi & Lloyd, L. paradorum Spring, L. sarcocaulon Welw. and
r

* Map 19 is copied, with permission, from Schuchert, Hist. Geol. fig. 434 (New bg
Jobn Wiley & Sons, Inc.); map 20 from Schuchert's fig. 48

Maps 22-27 are from Schuchert, Geological History of the Antillean Region, =.

Geol. Soc. Am. xl. 337-360 (1929).

L. tuber

1931]

Fernald,—Specifie Segregations and Identities

Positive

ig WB cracistion =
bea Uncertain
at glaciation

i ok ae ar sa er aaa oe 20°

sre 5 oe
a a ey 120° 140 160° 180°
Map 19, the Wortp in Permian time (after Schuchert).
0" 2 ° 2° ao x» We laa i” we se"
h & -
= == a + | less seas
AAS Arata Cir, SR (OR Sled SPs SO 7. 4
i pA. | ONS R
: PA)
ane 4
yes 5 Ys
as = a : a
= 4 =
== NA ee E oa e
2 /
‘ 4 x 3
= = Ni
> = = =F
—= = f F
S<] MIDDLE CRETACEOUS
tr)
i ee 7 ee

Map 20, the Wortp in Mrp-Cretacrovs time (after Schuchert).

48 Rhodora [FEBRUARY
2 1 Essie
f S J
2 Fi Mies 7 WE, eat
: Q ? a ie + a
% : | es Si ee 3 ok
ee eet — pasa e
pe x 0) = Sia eee aa
q \e
” \ \ _ =
Catt ri aes Temtapher®

Map 21, the Permian GuactaTion (after Coleman)!

27 (upper

Upper
UPPER foce CENE

), in the
per right), in the Miocene (all after Schuchert

p 22 (lower left), Geographic Connections of the Fingers: in het LoweR —
vs R CRETACEOUS; 24 (uD eft), in
; 26 (middle pera

ert).

hay OLIGOCENE:

1931] Fernald,—Specific Segregations and Identities 49

continuously maintained and often broadened. Consequently, Gon-
dwana, as usually defined and whether a hypothetical or a demon-
strated Permian land, need not necessarily concern our immediate
problem. If, however, with Alphonse De Candolle and many other
thoughtful students, we conceive the angiospermous groups of today
as having lived in Jurassic time! or if, with Velenovsky, Mrs. Arber?
and numerous others, we push the ancestral Angiosperms (the Pro-
tangiosperms of Engler*) back beyond the Cretaceous, and also
greatly extend the area of Gondwana in the Western Hemisphere
to reach the southeastern United States, then that hypothetical
equatorial land will immediately come into the picture. In this con-
nection it isworth noting that my associate, Dr. Hugh M. Raup, calls
attention to the very striking fact, that hundreds of tropical and
austral groups of today have restricted ranges almost coincident with
the known areas of Permian or Permo-Carboniferous glaciation (see
MaPs 21 and 19; also Maps 9-18). If, indeed, Angiosperms existed at
that time and if youthful and aggressive groups then spread into newly
available territory as do virile and dominating groups today, it would
seem only natural that, with the decay of the Permian glaciers, the most
aggressive plants of the adjacent territory should quickly have taken
possession and that their descendents should still linger about their
ancestral centers. By some the herbaceous habit is associated with
cold climates, and many herbs (as well as shrubs and trees) undoubt-
edly originated during the Pleistocene, while many others seem to
be even more youthful; but, surely, many of the herbaceous species of
today are as ancient as their living ligneous associates. And it is
certainly significant that the greatest glaciation in the history of the
world was the Permo-Carboniferous (Map 21).4 If it be true that the
development of herbs has been largely a response to refrigerating
climates, it must be apparent that, given the progenitor-stocks, herbs
would,had a phenomenal epoch of development during the Permo-
arboniferous glaciation.

oe nos jou cies aoe Cycadtes @ oo aia he
icc et l’Inde. Existaient-elles déja dans I’hémisph?re austral, au

exique. ete., quand les terrains jurassiques se formaient en Europe? C’est
tae mement probable; matheureusement on connait peu de fossiles de ape pe7* loin a

s +

ere d’années & des distances aussi grandes.” —A. DC. Mon. Phan. 1. 35 (1878).

Habit in of literature and pertinent discussion see Arber, Agnes, The Tree
in Angiosperms: its Origin and Meaning, New Phyt. xxvii. 71-84 (1929).

enf. ed 2, xiv®, 138-145 (1926).

‘Map 21 reproduced, with permission, from p. 91 of Coleman, Ice Ages Recent

Ancient (New York. Macmillan).

50 Rhodora [FEBRUARY

Gondwana was, after all, only a hypothetical land. Originally re-
constructed and subsequently greatly extended to account for the
disrupted localities of now fossil species, it furnishes a sort of explan-
ation of early tropical dispersal. But, at this very day, hundreds of
living groups of Angiosperms (as well as many groups of Pterido-
phytes) have a “Gondwana” or relic pantropical range. Neverthe-
less, there is now no circum-tropical connection, such as is generally
assumed for Permian time, between these areas of living plants; conse-
quently, the hypothetical status of Gondwana becomes strikingly
accentuated.

aving these purely theoretical realms and returning to better
demonstrated and, consequently, firmer geological and _paleonto-
logical grounds, we are at least justified in looking upon the present
Atlantic North American representatives of prevailingly tropical
groups as lineal descendents of plants which reached eastern North
America out of the North at a time, during the Mesozoic and perhaps
early Cenozoic, when the ancient rock-floor which now constitutes the
Appalachian Upland was peneplained or near sea-level. With the
Cretaceous seas (see Map 20) occupying the north-and-south center
of the present continent and its southern and southeastern margin
and the Tertiary seas later invading much of the southern and eastern
portion of the same area, the only section of present eastern contin-
ental North America then continuously accessible to these plants was
the ancient baseleveled Appalachian region with the Laurentian
shield to the north and northwest.! Then, as the later uplifting of
the once peneplained Appalachian region progressed, making halts at
intervals through Tertiary time, with the consequent deep dissection
of the extensive plateau and its inevitable conversion from a low
Cretaceous plain with retarded drainage into a vast upland mesophy-
tic area, ready for occupation by the abundant mesophytic types
the climax forest which could now freely mingle between Asia,

western America, eastern North America and Europe, the members

of the tropical and subtropical groups (and with them many endemle
groups) represented in eastern North America were largely forced (by
drainage of the area and by invasion by the hoard of actively colont-
zing mesophytic types) to abandon their once congenial but now Leoal

+In this connection, the absence from Pacific North America of the a —

tropical and oe groups which linger on the Atlantic Coastal Plain pletely

cant; during the Cretaceous eastern and western North America were com
isolated from each other (see Map 20).

;

1931] Fernald,—Specific Segregations and Identities 51

congenial haunts on the Appalachian area and to move out to the
newly available xerophytic and hydrophytic habitats, chiefly on the
Coastal Plain (though in some instances west and northwest of the
Appalachian axis), where the acid savannahs, bogs, shallow pools and
dry sands supply the ecological conditions in which these descendents
of Cretaceous and early Tertiary hydrophytes and xerophytes can
still survive.

In 1897, Dr. T. H. Kearney, directing attention to “ The Pine-bar-
ren Flora in the East Tennessee Mountains,’’! said: “One would hardly
expect to find a large proportion of this flat country flora hidden
among the high ridges of the Alleghanies; yet this is unmistakably

e case. Especially along the picturesque French Broad river, . . .

there is a notable incursion of plants usually considered typical of

the coastal plain,” and, after enumerating several such species,
Kearney continued: “What are we to infer from the presence of
these austro-riparian plants among the flora of northern origin that
chiefly covers these mountains? Possibly they are the advance-
guard of an invading army. Much more probably, however, they
are the lingering survivals of a more southern flora, once widely dis-
tributed over the southern Appalachian region.” Three years later,
developing the subject more fully and interpreting the migration to
the Coastal Plain as a response to Pleistocene cold, Kearney” suggest-
ed that “it may be conceived that while some individuals of each hy-
pothetical Pliocene ancestral species maintained themselves in well-
sheltered situations and were not forced [by Pleistocene cold] to a
change of abode,’ others escaped the changing environment by a grad-
ual retreat into the warmer lowlands. The individuals which remained
m the mountains were the direct ancestors of the present Appalachian
Species; while those which migrated and later accustomed themselves
oe Plant World, i. 33-35 (1897).

farney, The Lower Austral Element in the Flora of the Southern Appalachian

. 830-842 (1900), especially p. 839.
not eo 2 ct UPON vegetation of ice-sheets hundreds of miles away was probably
fate Sreat as was formerly supposed. Witness the occurrence today of splendid
on the glaciers of Alaska; and the occurrence at the ice-free margin of other-
Greenland of such plants of sea-level in the eastern United States
codsia ilvensis, Cystopteris fragilis, Equisetum sylvaticum, E. hiemale, Deschampsia

=
g
<
z3
3

$s canadensis,

Oxycoccus, Menyanthes trifoliata, Utricularia

N and Linnaea borealis, var. americana; or the occurrence on tablelands

With typical or Gaspé of members of chiefly tropical groups almost side-by-side
Plants of Ellesmereland and Melville Island.

52 Rhodora [FEBRUARY

in the Coastal Plain . . . gave rise to the Austro-riparian species
that attract our attention today because of their close resemblance to
Appalachian forms.” In 1904 the late Professor Harshberger! sug-
gested a similar derivation of some of the endemic North American
plants of the Coastal Plain from the crests of the Alleghenies. Some-
what later, Dr. Witmer Stone stated that “The flora at the top of
Meadow Mountain, the summit of the Alleghenies [in Garrett County,
Maryland], . . . was very much as in the pine barrens of New Jersey,
and is quite likely a remnant of an early primitive flora such as we
have there.’”? In 1924 Dr. John K. Small expressed the same view.
The relic-genus Narthecium (or Abama) has one wide-ranging species
in Europe, one or two highly localized species in the eastern Medi-
terranean region, one in Japan, one in California and Oregon; and it
has long been of great local interest in eastern America through N.
americanum Ker., a famously rare plant of the very heart of the New
Jersey Pine Barrens. But in 1919 a species closely related to N. amer-
icanum was discovered at a single station in the mountains of North
Carolina. In describing the mountain plant as Abama montana,
Small said: “It is not surprising that a bog-asphodel should come
to light in the mountains of North Carolina, as several kinds of plants
otherwise known only in the pine-barrens of the middle Atlantic Coast-
al Plain also grow in the Appalachians. . . . It is scarce, eviden y
rare, and may be on the verge of extinction. It may be that in this
species we have one of the progenitors of the Abama of the Coastal
Plain, for the high mountain region was the reservoir whence many of
our Coastal Plain plants were derived.’ :

The derivation of much of the flora of the youthful Coastal Plain
of the United States from the ancient Appalachian Upland, which m
the Cretaceous itself had a coastal-plain status, is strongly supported
by the very striking condition in the Antillean and Mexican regions:
As shown by Schuchert# (maps 22-27), portions of the Greater Antilles
(Cuba, Haiti, Jamaica, Porto Rico, etc.) have been in continuous ©
istence at least since the Lower Cretaceous and much of the time
connected with the Central American and Mexican tablelands; but
the Bahama Islands did not rise above the sea until the Miocen®

' Harshberger, Proc. Acad. Nat Sci. Phila. 1904: 607

, Stone. as reported in Bartonia, No. 5: 16 (1912).

Small, Torreya, xxiv. 86 (1924).

i-
‘Schuchert, Geological History of the Antillean Region, Bull. Geol. Soc. Am. x1. 33
360 (1929), from which my maps 22-27 are photographed.

1931] Fernald,—Specific Segregations and Identities 53

while all but the base of the Peninsula of Yucatan is even younger.
It is, therefore, significant, when we check the living floras of the
Bahamas! and of Yucatan,’ to find that on these very youthful areas,
which are far isolated from more ancient upland centers, there are no
species of Schizaea (Map 9), the Haemadoraceae (MaP 10), the Xyrid-
aceae (MAP 12) and Lycopodiaceae (see MAP 17); but that on the older
Greater Antilles and the old Mexican or Central American Plateaus
or their margins these groups of the young Coastal Plain of the
United States are all represented. Ancient Diospyros (MAP 14), also,
is on the Greater Antilles but, although locally present in Yucatan,
it is absent, as a native, from the Bahamas. Many other tropical or
austral groups which today have a northern remnant persisting on
the young Atlantic Coastal Plain of the United States and on the
ancient Greater Antilles and the equally ancient Mexican or Cent-
ral American Plateaus or at their margins, are absent, likewise, from
the young Bahamas and from still younger Yucatan (except occa-
sionally as cultivated species or as relics of cultivation): such groups
as the Eriocaulaceae, in Mexico up to 3355 m. (11,000 ft.), the Marant-
aceae, up to 915 m. (3000 ft.), the Burmanniaceae (MaP 18), up to 1675
m. (5500 ft.), Boehmeria, up to 1585 m. (5200 ft.), Clitora, up to 2135
m. (7000 ft.), Rotala, up to 1525 m. (5000 ft.), Myriophyllum § Tess-
aronia, up to 1740 m. (5700 ft.) and Dyschoriste, up to 2400 m. (8000
ft.); while Psilocarya occurs on the Greater Antilles but not in Mexico
and Litsea in Mexico ascends to 3050 m. (10,000 ft.). Similarly,
on the Mexican Plateau we find representatives of Xyris at 1830 m.
(6000 ft.) and of Schizaca at 1200 m. (3925 ft.).

The always interesting genus Drosera (MAP 28),' with three circum-
boreal species and another (D. linearis Goldie) confined to boreal
America, has most of its living species concentrated in the Southern
Hemisphere and in southern Asia, with a few related species on the
Greater Antilles, on the Coastal Plain of the United States and in

entral America; but the genus is unknown from either Yucatan or
the Bahamas. The unique D. filiformis Raf., however, differs from
all other American species in its hardened bulbiform woolly base, its
linear-filiform and very prolonged (up to 6 dm.) leaves without elon-

"See Bri
«i Standley, Flora of Yucsen, Field Mus. Publ. ‘275-—Bot. Ser. ill. no. 3 (1930),

' The as - sor while this paper is paging gentle Pflanzenr. iv ™, supple-
mented by mate ete big from that of = nisiek a lank ska
to accord with the enect e Gray Herbarium. : sdb

rea
alta.

|
}——|———+

] pt |
eek dei
/ i
f \
\ \
é a eee
\
x
Anak
Ne Nes /

/
oe

ww Zz24

~

BLOpouy

SW

ie SS —\
Map 28, Geographic Range of Drosrra, the boreal Species indicated by Stippling, and of Drosopny.iuvum (in ellipse).

Auvouda 7]

Drosera filiformis

ins ane oo es at oF
sie since the ae of
the Cretaceous.

S Exposed since the close of
ary.

o the Tertia
oo of VE, — areas wholly or partly
pace Vea d during Ple ane
Pens cated after disappearance
dre: wigs [ed or Pleistocene Ce.
Pa LY a 2 ge ocastal plain areas ocgee
during the Quaten

Woodway dia. Existing glacial ‘ae

( Areolata

Map 29, Geographi
and of he Ap a (us ae Pare of nase FILIFORMIS (upper sa acti black a
Station sho saan le black area); of WoopwaRDIA a a (an aie
rded, at the sakes end of Lake

56 Rhodora [FEBRUARY

gate petiole, its large purple petals and its seeds; it is far more closely
related to some of the Indian, Malayan and Australian species than
to any of the broad-leaved, long-petioled and white-flowered species
(of § Rossolis) with which it grows. It is restricted to two wholly
disrupted areas of the Coastal Plain (see MaP 29) and its present dis-
tribution is best explained as a result of migration in two directions
off the now uplifted Appalachian Upland. Outside the now essen-
tially austral and tropical genus Drosera the family has two monotypes:
Drosophyllum lusitanicum L. (see MAP 28) of the ancient Iberian
Tableland and adjacent Morocco and Dionaea muscipula Ellis (see
Map 29) of the Cretaceous (hardly Tertiary or Quaternary) Coastal
Plain of the Carolinas.! The antiquity of the insectivorous Drose-
raceae, therefore, can scarcely be doubted.

Not only, as we have seen, do the tropical groups represented oD
the Atlantic Coastal Plain of the United States have representatives
on the high tablelands of the Appalachian Mountains, on the older
of the Antilles and on the high plateaus of Mexico or of Central Ameri-
ca: they are also well represented at low levels west and northwest
of the Appalachian axis, particularly in the upper Mississippi Valley,
and about the Great Lakes where the sands and peats of Pleistocene
or later origin offer them peculiarly attractive habitats. Their mi-
gration out from the now elevated Appalachian area to the oastal
Plain on one side, to the region of the Great Lakes on the other, 1s
clearly indicated by Woodwardia areolata (L.) Moore (MAP 30), @
characteristic fern of the Coastal Plain but also found on the table-
lands of the southern Appalachians at 3000 ft. (915 m.), and other-
wise known only in northwestern Ohio and in southwestern Michigan.

This movement out from the Appalachian center is even better
displayed by the very distinct, bulbous Xyris torta Sm. (MAP 31), 8
common species of the Coastal Plain sands, also at 2000-4000 ft. (610-
1220 m.) on the high silicious tablelands from North Carolina and
Tennessee southward, and with a rather extensive area about
Great Lakes. In these cases (and scores of others which might
cited) migration from the ancient Appalachian region is @ reasonable
explanation of the present segregations. In some other cases We
not now (or yet) know any existing stations high on the Appalachians
but it is fair to assume that during the baseleveled Cretaceous stage
of this area they must have there existed; and it is reasonable to
1 For all known stations see Coker, The Distribution of Venus's Flv Trap (Dionaee
muscipuia), Journ. Elisha Mitchell Sci. Soc. xliii. 221 (1928).

e
O Exposed since the close of
the Paleozoic.

Exposed since the close of
the Cretaceous.

Exposed since the close of

- the Tertiary.

Nunatak areas wholly or partly
ZI exposed during Pleistocene,

ea Exposed after disappearance

of Pleistocene ice.
Coastal plain areas exposed
<3 during the Quaternary

Existing glacial fields,

—

5

“Potam oge ton

confer voides

° a
Se 31, Geographic Range of Xrris Torta; 32, of Frrrena SQuaRROsA; 33, of
MOGRTON CONFERVOIDES. —aHARY
TIBRARY

oF

AW ASHINGTC

58 Rhodora [FEBRUARY

lieve that, to the already surprisingly large number of relic-colonies
of tropical types known on the mountains of the Carolinas and Ten-
nessee, stations of members of these other groups may soon be added.
Potamogeton, sub § Javanici (MAP 13), not yet known either on the
Appalachian Upland or on the true Coastal Plain, is decidedly an
exception. Much more typical is Fuirena squarrosa Michx. (map 32),
characteristic of Coastal Plain sands, with a slight overlap on the
ancient Piedmont area, but unknown from the high tablelands, and
isolated on the sands of southern Michigan and northern Indiana.
The Burmanniaceae (Map 18) yield an even more striking instance.
The family is widespread in tropical and subtropical regions but,
like many other groups, it has a slight extension northward in eastern
Asia and another around the Gulf and Atlantic Coastal Plains of the
United States. North of Virginia the family is unknown, save for
the extraordinary little species of prairies at Chicago, Thismia ameri-
cana N. E. Pfeiffer. This, as noteworthy a plant as is known in the
Great Lake area, is the more remarkable since it belongs to a small
subfamily otherwise confined to the Tropics, chiefly of the Old World.
The cases I have presented are merely a few illustrations but, with-
out further enumeration or further analysis, enough have been pre
sented to make it clear that, to a large extent the tropical groups
which today have a representation on the Atlantic Coastal Plain of
the United States or about the Great Lakes are also found persisting
(often as rare relics) on the high tablelands of the Mexican Plateau
as well as on the tablelands of the Appalachian Mountains or at other
spots back of and older than the outer Coastal Plain. I am aware
that, in looking upon the members of these and other primitive groups
which now occur upon or among the mountains of north-¢
Pennsylvania, northern New York and western and northern New
England as preglacial relics there, I am not in full accord with ort
dox Pleistocene geology, which asks us to believe that the late 1¢&
sheets so vigorously overrode these mountains as to denude them
of all life. The parallelism is so definite, however, between
Coastal Plain species having representatives on the elevated tablelands
of the more southern and quite unglaciated Appalachians, aS well as
on the Central American, Mexican and Antillean tablelands, and their
occurrence on the supposedly denuded mountains farther north, that
the botanist, at least, is forced to believe that the Pleistocene 4

' Pfeiffer, Morphology of Thismia americana, Bot. Gaz. vii. 122-134 (1914).

1931] Fernald,—Specific Segregations and Identities 59

struction of plants on the high mountains of New England, New
York and northeastern Pennsylvania has been vastly overestimated.
It is far more logical to look upon the very primitive Potamogeton con-
fervoides (MaP 33), an aquatic with plump fruits not adapted for fly-
ing, as a preglacial relic on the high mountains of northern and west-
ern New England, on the Adirondacks and on the Pocono Plateau
(as well as on the unglaciated or but slightly glaciated tablelands of
Newfoundland and of Cape Breton), than to picture it as making a
post-Wisconsin leap from the coastal pools to the highest mountain-
masses far to the northwest and north and successfully finding wel-
coming pools on them all, but not landing in the thousands of inter-
mediate lowland ponds. P. confervoides, one of the most primitive
members of the worldwide subgenus Eupotamogeton, is so rare that
only a few botanists have ever seen it growing. If any lowland pond-
weed were in post-Wisconsin time to do the improbable and invade the
elevated mountain-pools from sea-level, it would less unreasonably
be a common or, in the region, essentially ubiquitous and less prim-
itive species (also of peaty depressions and pools), such as P. pusillus
L., P. Spirillus Tuckerm., P. epihydrus Raf.! and P. gramineus L.;
but these almost weed-like species of all pools at low levels, north,
south, east and west, have not invaded the mountain-ponds of New
England, New York and Pennsylvania, where the primitive and ex-
cessively rare Coastal Plain P. confervoides and its Coastal Plain
congener, P. Oakesianus Robbins, alone represent the genus. Near
¢ White Mountain stations of P. confervoides occur two endemic
alpine xerophytes: Potentilla Robbinsiana Oakes, nearest related
‘o the Siberian P. elegans C. & S.; and Gewm Peckii Pursh, the isolated
representative of G. calthifolium Mertens of the North Pacific and
& Sea region. Obviously these two endemic neighbors of the
Coastal Plain Potamogeton confervoides on the White Mountains have
not originated there in post-Wisconsin time.
Tf this general reasoning is correct, that the Coastal Plain plants
of the eastern United States before occupying the youthful Coastal
Plain had had direct progenitors on the Cretaceous peneplain which,

Pop lthough not found in the high subalpine ponds of New England and New York,
beta epihudrus is, surely, to be sought there. On the great tableland of the
Pp

sio :
Muriophytium Farweltii Morong, the most northern representative of the tropical
stal Plain Myriophyllum § Tessaronia.

60 Rhodora [FEBRUARY

tains, it follows that these Coastal Plain groups are of as ancient an-
cestry as any angiosperms now living in eastern North America. It
is, therefore, particularly noteworthy that in world-wide groups their
representatives are invariably regarded by monographers as wholly
differentiated specifically from their relatives now living in Africa,
tropical India and Australia.

When we turn northward to the physiographically very youthful
soils (or lack of soils) which have resulted from the Pleistocene glaci-
ation, much of the vast area of eastern Canada and the northernmost
States, a region in which the plant-population is, in some places, just
moving in to occupy the newly available territory, we find a great
assemblage of dominantly boreal (or extreme austral) genera: Spar-
ganium, Festuca, Glyceria, Agropyron, Calamagrostis, Eriophorum,
Carex, Juncus, Salix, Alnus, Betula, Sorbus, Rubus, Fragaria, Po-
tentilla, Rosa, Lathyrus, Epilobium, Pyrola and hundreds of others.
Although many stable species in this flora, like Sparganiwm minimum
Fries, Potamogeton filiformis Pers., P. obtusifolius Mert. & Koch, P.
praelongus Wulfen, Triglochin palustris L. and maritima L., Cinna
latifolia (Trev.) Griseb., Eriophorum gracile Koch, Caltha palustris
L., Pyrola minor L. and Lysimachiathyrsiflora L., are regularly regarded
as identical with plants of Eurasia, others, like Scheuchzeria palustris
L.! and Menyanthes trifoliata L.,? in eastern America are sufficiently
different from the Eurasian plants in size of flower or fruit as to 1
ceive varietal designations; but no one has seriously proposed id
treat them as different species. And when we get into heteromorphic
tangles like Festuca rubra, the Poa pratensis jumble, Agropyron repen®
Calamagrostis canadensis (including Langsdorfi), the Carex vesieart
morass, Juncus effusus, the Salix glauca complex, Alnus incana, the
Betula alba forest, Cerastium arvense, Sorbus, Rubus idaeus, and &
hundred other always perplexing mazes, common experience on both
sides of the Atlantic is sufficient to indicate the utter lack in these
groups of the clear specific differentiations which every systematist
looks for but in these plants fails to find. F

The key-note in this youthful flora of the vast Canadian regio™
mostly available for plant-occupation only during the last few thou-
sand years, is, then, the essential lack of fundamental s ific segre”
gation from the European representatives of these plants. Thus ¥
in marked contrast with the condition shown in the am 8

1See Fernald, Ruopora, xxv. 178 (1923).
? See Fernald, Raopora, xxxi. 195 (1929).

1931] Fernald,—Specific Segregations and Identities 61

genera now characteristic of the ancient Appalachian Upland and
with the sharp specific differentiation of the Coastal Plain species
from their allies in Africa, tropical Asia and Australia. Whether we
derive the disrupted colonies of the tropical groups out of the wide-
spread boreal flora of the Cretaceous and early Tertiary, which seems
to me the logical deduction, or whether we prefer to picture them as
having spread across a hypothetical trans-Atlantic land, it is clear
that long after the tropical floras had become geographically separated
the northern lands still maintained a broad connection. Many geolo-
gists have postulated a late Tertiary and Pleistocene uplift of the
Arctic lands. Thus, Mecking,! following Suess, says of the Arctic:
“At the end of the Tertiary began an uplift which continued into the
Glacial Period. At the climax of the latter there followed a general
sinking, which, by filling the valleys and troughs, created the sounds
and then gave way to a newer uplift which still continues”; and Sim-
mons, summarizing the various studies of former continuity of Arctic
lands, concludes: “That the northern parts [of North America] . . .
were lifted considerably above their previous level, cannot be doubted
. most authors are inclined to refer the beginning of the rise .
to the last part of the tertiary era.”? The amount of this late Ter-
tiary and early Pleistocene uplift has generally been estimated at
2000-3000 feet (600-900 meters); and study of Dr. Nansen’s bathy-
metric map? of the Arctic Basin shows that an elevation of 2000 feet
(600 meters) would today unite northwestern America and Asia and
would connect northeastern America by two routes with Greenland,
thence via Iceland with Scotland and Scandinavia. It is, then.
quite clear that in late Tertiary, and probably in early Pleistocene time,
e northern plants had essentially free routes for interchange be-
tween Eurasia and North America. Consequently, sharp specific
differentiation in the comparatively youthful plants of amphigean
relationship which are chiefly restricted to recently glaciated northern
urope and aa glaciated northern America is the exception
rather than the ru
Time does not seanit my now specially discussing the outstanding
Ry Mecking, The Polar nz iptted a Regional Geography, Geogr. Polar Reg., Am. Geogr.
Sp. Pub. no. 8: 220 (19
“ae ve eg A Survey of a qetncerew of the or American Archipelago,
ne, v. Arsskrift. n. f. Afd. 2, Bd. 9, Nr. 19: 152 (19 he Problems of
ymetric Map of the Arctic Basin in Nansen, sig Ocxanooraphie Probl

= Still eres Arctic Regions, in Problems of Polar
bl. no. 7, t. 1 (1928).

62 Rhodora [FEBRUARY

features of certain other floras in eastern America, like the high degree
of relic-endemism on the nunatak areas, a topic already discussed by
me in other papers! (190 rare and localized endemics already known,
and 160 others with relic-colonies isolated by 1500-4500 miles from
other colonies of the species), or the multiplicity of intergradient
forms? of Panicum, Salix, Amelanchier, Crataegus, Rubus, Vaccinium
(§ Cyanococcus) and Aster and the dominance of more than 1000 ag-
gressive species introduced from Europe during the last three centuries
on the clearings made by the white man, the youngest spots ofall. From
what I have been able to present, however, certain deductions, by
way of summary, seem justified:

1. The characteristic flora of the present southern Appalachian
Upland, made up largely of groups which were cosmopolitan in the
Mesozoic or early Cenozoic but now reduced to geographically segre-
gated remnants in the Northern Hemisphere, is obviously ancient
and its species are usually (but not always) clearly differentiated
from those of the same genera in eastern and central Asia or in south-
eastern Europe.

2. The Coastal Plain flora of Atlantic North America is distin-
guished by the abundance of tropical groups represented. Although
these plants now chiefly occupy open silicious, peaty and aquatic
habitats in comparatively youthful regions of eastern North America,
it is probable that they or their progenitors formerly existed on the
area of the ancient Appalachian Upland, especially in the Cretaceous,
when that primitive region of the continent was baseleveled and
reduced essentially to sea-level and at the time when the trop!
groups of today were widespread in the North. Then, with the
_ Tertiary uplift of the Appalachian region and its final conversion into
a vast well-drained mesophytic area available to the groups
now constitute the climax forests of the Appalachian Upland, the
Cretaceous xerophytes and hydrophytes which had previously 0%
cupied the ground gradually moved out to the newly available and
for them more congenial Coastal Plain and similar habitats to the
west and northwest. These species of the Coastal Plain, whose d-
from thelr

FE Pe
Journ. Bot. xi. 564 (1924) ; Persistence of Plants in Unglaciated Areas of
Mem. Am. Acad. xv. 242-244 (1925); Some Relationships of the Floras die:
Hemisphere, Proc. Internat. Congr Pl. Sci. ii. 1493, 1494 (1929). Quart.
2 For see Fernald, The Antiquity and Dispersal of Vascular P' -_
Rev. Biol. i. 236-238 (1926).

1931] Fernald,—Specific Segregations and Identities 63

African, Indian and Australian allies during the Cretaceous or the
early Tertiary, are thoroughly differentiated from the Old World
species.

3. The nunatak areas, the scattered unglaciated or only slightly
glaciated spots within the region of more widespread glaciation, are
distinguished by a high degree of relic-endemism.

4. The dominant groups of the region of northeastern America
only lately freed from Pleistocene ice are largely those of recently
glaciated northern Europe. These northern lands have been so re-
cently segregated geographically that until geologically recent time
there has been free interchange in the floras; and in the ecologically
youthful Canadian and northern European regions identity of species
in the amphigean genera is so general as to become the rule.

5. The recently cleared lands, like the recently glaciated areas
and the Quaternary sands, are also characterized by a multiplicity of
variable and often intricately intergrading and hybridizing forms, as
well as by the aggressive dominance of unchanged species introduced
within three centuries from the youthful flora of Europe.

Some of the points I have discussed are possibly new, though many
of them are familiar and very old-fashioned doctrines; but since famil-
larity so often breeds contempt and the latest doctrines are so fre-
quently supposed to be the best, it may not be amiss to call attention
anew to the overwhelming evidence of the soundness of some of the
older deductions,

Rhodora Plate 204

FLORA OF NEWFOUNDLAND
REGION OF CAPE Ray

Qf 285 — : ey
6 Pe as ion conferees Nels
Picok peant edidiok
Chaldewe potls In peaty barvene omesy the gee

Puee

Photo, by H. xg. Raup.

I

> a ay ond Ww BS
OTAMOGETON CONFERVOIDES (left) and P. pectinaTus (right) X 19.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

XCIV.

THE VASCULAR FLORA OF ST. PAUL ISLAND,
NOVA SCOTIA.

By Lity M. Perry.

Reprinted from Ruovora, Vol. 33,
ay, 1931

ee ee ee

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY—NO. XCIV.

THE VASCULAR FLORA OF ST. PAUL ISLAND, NOVA SCOTIA
Lity M. Perry

St. Pau, one of the smaller islands in the Gulf of St. Lawrence,
lies about twelve miles northeast of Cape North, Nova Scotia. It is
well-known that at least portions of the Magdalens, Anticosti and the
Mingan Islands in the Gulf at some distance from the mainland
escaped the Pleistocene denudation and consequently have been of
particular interest phytogeographically. Presumably St. Paul also
was far enough from the mainland to continue unglaciated, except
perhaps locally; if so, its flora too might offer some interesting relics.
This possibility and the smallness of the island induced Dr. Muriel V.

and me to spend a month’s vacation there during the summer
of 1929 collecting botanical specimens.

It was exceedingly difficult to obtain in advance much information
Concerning St. Paul. Apart from wireless and the government boats
which carry supplies, equipment and employees, there is no regular
means of communication with the mainland; no geological survey of
the island has been made and consequently no topographic maps are
available. Finally, however, Miss Roscoe discovered Miss Campbell,
a former resident, who gave us many concrete facts about the place
and supplied us with a map which, although drawn with the details of
shipwrecks and currents especially for mariners, was invaluable to us
M our work.

The island in outline is elliptical-oval, at best not more than a
mile wide and approximately four miles long. About a quarter of a
mile from the northeast end the sea runs through a narrow channel

106 Rhodora [May

North Last light

e E. Channel

Martin Powers Cove

Petrie's Pe.
Loohout Fb.
Trimty love

At/antie Cove

fog Whistie

Paul /s/and, WV. S.
South West Light

MAP OF ST. PAUL ISLAND

1931] Perry,—Vascular Flora of St. Paul Island, Nova Scotia 107

so that, in fact, there are really two islands. The part cut off

is perhaps a fifth of a mile wide and lacks trees or shrubs of any kind.
On it are located the northeast light and the foghorn. A sliding cable
with chairlike attachment serves as a means of transportation across
the “Tittle” and here one receives as much of a thrill as if riding over
the Niagara whirlpool. The island rises rather abruptly from the sea,
although the surface is only moderately rolling with two slight ele-
vations, Norwegian and Coggin mountains. On the summit of
Coggin Mountain is a small bronze tablet set in concrete and stamped
“for information write the Geodetic Survey of Canada.”’ From Mr.
Ogilvie, the Director of the Survey, I learned that this tablet is a
triangulation station mark of the Geodetic Survey of Canada and its
position is as follows: latitude 47° 11’ 40.2” N., longitude 60° 09’ 06.2”
W., elevation 550 feet above the sea level. This particular station was
used not only for the purpose of accurately locating the island and the
lighthouses at either end, but also to aid in connecting points in
Newfoundland to those in Nova Scotia. /

rom the economic point of view St. Paul Island is of little value},

since for the most part it is covered with a somewhat stunted foregt of «| -
balsam fir and white spruce. It is really a government station man t.

aged by the Department of Marine and Fisheries; and no person, ia <g

unless shipwrecked, is permitted to land without the consent of the
Canadian government. Last summer eight government employees
with their families were living there. Although the history of the
island will not be considered here, one thing should not be overlooked.
Hitherto St. Paul has been a menace to seafaring people on the Gulf
or St. Lawrence. The coast is rugged, rocky, irregular and without
harbors. There are three or four coves where a stranger might expect
to find shelter, but experience has proved them to be more dangerous

ing a storm than the open sea and the lee side of the island affords
the only possible haven. Not only have storms played havoc with
Vessels but also bewildering fogs and unusually treacherous currents.
Practically every little headland has received its name from some
ship destroyed thereon. So numerous were the wrecks that, until
the year 1924, a life-saving station was maintained on the island under
the direction of a governor; and a crew of eight men patrolled the coast
every day on the lookout for wrecks or ships in distress. Since that
time, however, the Canadian government has established a radio
direction-finding station and the operation of this practically elimin-

108 Rhodora [May

ates the probability of any large wreck. The operator of any boat or
vessel carrying a wireless outfit may ask for and receive his bearing
at any time of the day or night.

As early as 1883 we find the island-flora already studied by Dr. A.
H. MacKay and his manuscript-list incorporated in Macoun’s “ Cata-
logue of Canadian Plants.”’ Nevertheless, so great has been the prog-
ress in the field of systematic botany that earlier determinations are
often obscure and many times characters have been misinterpreted;
for this reason a restudy of the specimens in the light of present knowl-
edge is necessary to establish their identities. For instance, Professor
Fernald, while revising the genus Oxytropis in eastern North America,
found that it had been reported from St. Paul island. Through in-
quiry, he learned that Dr. MacKay’s record had been made on tenta-
tive field identification, hence the specimens were unavailable for
further study. Since “no Ozxytropis is known about the Gulf of St.
Lawrence except in the Gaspé region and in western Newfoundland
(four species),” Professor Fernald emphasized the fact that it was
highly desirable to obtain specimens from St. Paul. He also sug-
gested that the flora there would probably reveal significant geographic
relationships. Now the scope of the island is too limited to attract
outstanding field-botanists; but Miss Roscoe and I believed that here
was a somewhat isolated unit whereby we might contribute something
to the knowledge of the flora of Nova Scotia and at the same time
profit by some valuable field-experience.

t seemed a relatively simple problem to get a boat to transfer us
from the mainland to the island; nevertheless, after seeing people
there wait ten days for favoring wind, tide and weather in order to 0
safely ashore, we realized how fortunate we had been to spend only one
day at Bay of St. Lawrence. The weather was ideal for tenting and
collecting specimens; yet, in spite of almost no fog nor rain, so grea
was the humidity that it was impossible to dry either specimens oT
driers out-doors.

In the vicinity of the Radio Station at Atlantic Cove were miniature
gardens; and the few fields, at least in part, had been cultivated
although not many genera have been introduced. The vegetation
of the fields was various. The predominant grasses, Anthoxanthum,
Agrostis and Phleum, were associated with a profusion of Rumet
Acetosella, Stellaria graminea, Ranunculus repens, R. acris, Euphrasa
canadensis, E. americana, Rhinanthus Crista-galli var. fallax, Chrysa"-

|

1931) Perry,—Vascular Flora of St. Paul Island, Nova Scotia 109
thenum Leucanthemum var., Leontodon autumnalis var. pratensis
Cirsium arvense and occasional patches of Trifolium, Prunella,
Galeopsis Tetrahit var., Plantago major and Centaurea nigra. In the
moister or more meadowy portions Viola cucullata and Arenaria lateri-
flora were the showy members of the plant-association. Near the
base of one hill Potentilla canadensis var. simplex was growing with
scattering bits of sphagnum. Invariably the dry rocky places were
covered with Sisyrinchium angustifolium, Potentilla tridentata, Vac-
cinium Vitis-Idaea var. minus and clumps of Iris versicolor. Instead
of Spiraea on the open slopes, Sanguisorba canadensis var. latifolia
grew in abundance. This Alaskan variety has been collected in
Quebec at Natashquan and Pontchartrain, and on the island of
Anticosti, but apparently has not been known south of Anticosti.

rdering the swamp of this region was a somewhat solid turf of
Glyceria canadensis, Scirpus rubrotinctus, S. atrocinctus, Eriophorum
urginicum, Carex scoparia, Juncus filiformis and Luzula campestris
var. multiflora. Beyond this the ground was spongy. Usually Carex
stellulata var. cephalantha and C. paupercula filled the margins of
little “bog-holes.” The larger part of the swamp was covered with
tussocks of Scirpus cespitosus var. callosus supplemented in open
spots by Rynchospora alba and Epilobium palustre var. monticola.
The whole area was interspersed with Eriophorum angustifolium,
Habenaria dilatata, H. fimbriata, Sarracenia purpurea and afew clumps
of Tris versicolor. Just beyond in a typical sphagnum bog the incon-
spicuous Malaxis unifolia was growing with an abundance of Haben-
ara clavellata and Equisetum arvense, and nearby a scrubby fir con-
cealed a Meagre specimen of Kalmia polifolia, a species not noted
elsewhere on the island.

Owing to zones of dense undergrowth of brushwood, travelling in the
forest Was exceedingly difficult and in many places well-nigh impos-
sible, Sometimes the scrub was sufficiently compact to carry our
Weight, but oftener it swallowed us and we had to hunt for a way
throu, - The flora of the woods varied scarcely at all in the different
regions. Athryium angustum, var. rubellum, Thelypteris P. hegopteris,

: §pinulosa vars., Clintonia, Maianthemum, Streptopus, Habenaria
obtusata, Coptis, Ribes glandulosum, Aralia nudicaulis, Cornus cana-

*verywhere, Empetrum nigrum or Chiogenes hispidula covered the

110 Rhodora [May

exposed knolls and Carex trisperma carpeted the little swales and
glades. Southeast of Lena Lake, Polypodium virginianum, Taxus
canadensis, Lycopodium lucidulum, Mitella nuda, Oxalis montana and
Viola incognita were frequent. The forest flora, however, was not
entirely uninteresting. Carex crinita var. simulans collected in a
swale on the east slope of Coggin Mountain, although known from
Newfoundland to Massachusetts chiefly in the mountains, is reported
for the first time from Nova Scotia; also Solidago uniligulata var.
neglecta not definitely reported east of Yarmouth county was fairly
abundant here. The prevailing tree was Abies balsamea although
Picea canadensis was generally distributed throughout. Possibly an
occasional “bog” spruce grew there, but we failed to find any in our
cruising. The chief deciduous elements were Alnus, Betula, Acer,
Pyrus, Amelanchier, Lonicera, Viburnum, and Sambucus. These were
not plentiful nor were they much larger than undershrubs. Lonicera
canadensis reaches its northeastern range-limit on St. Paul, and Amel-
anchier Fernaldii, a native of the region about the Gulf of St. Law-
rence, is an addition to the Nova Scotian flora.

On account of the high winds and salt spray the border of the forest
was greatly dwarfed. Often the firs were simply flattened dense mats
which grew to the very edge of the sea-cliffs; but, between the woods
and the cliffs were also stretches of barren which yielded a variety of
interesting things. The barrens were nearly always on headlands and
similar in character, although usually each was marked by some
species lacking in the others. Here and there in slightly protected
places stunted trees weathered the gales, but Empetrum migrum
flourished regardless of shelter. Occasionally sprawled with it were
Juniperus horizontalis and J. communis var. megistocarpa similar
in habit to var. montana yet distinguished by larger berries wit
larger seeds. This variety was described from Sable Island, but it also
occurs on the Magdalen Islands and in western Newfoundland. ‘
the broad barren, Agrostis, Ammophila, Deschampsia flexuosa, Smila-
cina stellata, Iris setosa var. canadensis, Euphrasia, Solidago bicolor and
Aster novi-belgii were the general vegetative elements. Equisetum,
Agropyron, Carex hormathodes, C. silicea, C. maritima or Juncus bal-
ticus var. littoralis filled the damp or sphagnous depressions. At
southwest end of the island some two hundred feet above the s@
level, Vaccinium uliginosum var. alpinum, much cropped by the g0# -
maintained a struggle for existence. Near West Point the bore

1931} Perry,—Vascular Flora of St. Paul Island, Nova Scotia 111

species Cornus suecica grew in abundance and close by a colony of
Anaphalis margaritacea f. anochlora prospered. The barren at Trinity
Cove was by far the largest, yet its vegetation contained only three
particularly interesting things: Deschampsia flexuosa var. montana,
Rhinanthus groenlandicus and Campanula rotundifolia var. alaskana.
These natives of Newfoundland and Gaspé are here noted for the
first time from Nova Scotia. To this enumeration of boreal species
should be added Silene acaulis var. exscapa (abundant also at the
southwest of the island), Salix Uva-ursi, S. cordifolia var. callicarpaea
and Solidago multiradiata, all collected on a little slope south of the
“Tittle” (North East Channel), the two latter the first from south
of Newfoundland, Anticosti and Gaspé.

Near the sea cliffs, Plantago juncoides var. decipiens replaced the
Gramineae. Along the upper edges of the bluffs were mats of Ceras-
tium arvense or Lathyrus maritimus and clumps of Sedum roseum.
At the northeast Oxytropis johannensis grew in profusion. This, the
only member of the genus known from Nova Scotia, has not been
found farther south in Newfoundland than Cape St. George, although
it occurs in Quebec, Maine and east along the upper St. John river in
New Brunswick. The chief crevice plants were Puccinellia paupercula
var. alaskana, Festuca rubra var., Sagina procumbens, Plantago juncoi-
des var. decipiens and Solidago sempervirens.

St. Paul rises so abruptly from the sea that it is unsafe to try to
reach the water from more than three or four points on the island.
Logically enough there is a characteristic lack of beaches. Neverthe-
ess, on the west coast a small sand beach has developed, and in other
Places below the cliffs sandy margins are appearing. The beach
below West Landing maintained a few beach-types: Mertensia mari-
fima growing in mats from 6 inches to 2 feet in diameter, Atriplex
maritima var. hastata, Cakile edentula, Polygonum Rait and Arenaria
Peploides var. robusta (a single specimen of each of the last two).

While scouting to obtain the lay of the land, we discovered ap-
Parently a low open stretch in the forest northeast of Petrie’s Pond.
We were eager to explore it, but, since the vegetation of the barren
Was More mature owing to its exposed location, we bent our efforts in
that direction. We started at West Landing and, after following the
Trinity Cove barren almost to Petrie’s Point, we came to a slight ridge
marking a definite change in the plant-association. On the other
Side of the elevation, a little stream was trickling from a series of

112 Rhodora [May

shallow pools and sunken spots of ooze or slimy mud. The locality
was too fascinating to resist. From the pools we took Equisetum
arvense f. decumbens and Triglochin palustre. We were puzzled by a
network of delicate strands on the ooze; however, thinking they were
probably indeterminable, we passed them by for species manifestly
not yet represented in our collection: Selaginella selaginoides nestling
under sedge tussocks and tiny banks, Panicum boreale, Danthonia
spicata, Carex gynocrates (rare on the island and new to Nova Scotia),
C: Oedert var. pumila, Thalictrum polygamum, Potentilla fruticosa,
Primula mistassinica and Pinguicula vulgaris. It was impossible to
pick up everything that day so we planned to return the following
Monday. A heavy rain on Sunday swelled the brooklet and filled
the oozy hollows with water. Imagine our surprise to find some pools
almost choked with submersed masses of Utricularia minor.
specimens were coarser than those of the typical form, the bladders
larger and the leaves flattened rather than capillary. There is a form
platyloba which occurs wherever the plant creeps from the water into
a muddy habitat, and I am inclined to believe that our slightly
abnormal specimens may be the result of ecological conditions under
which they live.

The stream drained a boggy meadow in which grasses and sedges
were predominant. Here we gathered Agrostis h yemalis, Calamagrostis
canadensis, Scirpus cespitosus var. callosus, S. hudsonianus, Eri-
phorum viridi-carinatum, Carex pauciflora, C. rostrata and var. utricu-
lata, Epilobium palustre var. monticola, Andromeda glaucophylla and,
curiously enough, Gaultheria procumbens. Bordering the expanse was
a thicket of Betula pumila, Pyrus dumosa, Rubus recurvicaulis, Rosa
nitida, Acer rubrum, Lonicera villosa var. Solonis and other common
types. The swampy and boggy area in one direction was lying adja
cent to Petrie’s Pond and proved to be the section we had wished to
explore. At the opposite end, it opened into a larger bog dotted with
a number of small Sons often etait soporte as s to floral specialty, usually
one of the foll Spa , Potamogeton epihydrus,
Lycopodium Selago, i indisbuns 0 or Juncus canadensis. Here Cares

owei, a coastal plain species, reaches its present northern limit.
Just beyond the immediate source of White Spring (the stream
draining the bog) a beautiful patch of Pogonia ophioglossoides flour-
ished and Scattering specimens of Solidago uniligulata var. neglecta
were beginning to blossom.

1931] Perry,—Vascular Flora of St. Paul Island, Nova Scotia 113

Apart from the brooklet already mentioned, the streams and their
banks were disappointing botanically. This was probably owing to
their smallness and the rapidity of the flow of water as well as to the
wetness of the land in general. The lakes, Ethel and Lena, were more
attractive. In Ethel Lake the wading was good and we added to our
list Isoetes macrospora, Eriocaulon septangulare, Nymphozanthus
variegatus, Lysimachia terrestris and Lobelia Dortmanna. The latter
is shghtly aberrant in that the lip of the corolla is somewhat more
coarsely and profusely pubescent than that of the typical form and
the protrusion of the pubescence causes the sinuses between the lobes
to appear ciliate. Lena Lake was a different proposition. We had
been warned to keep away from the lower end if we valued our lives.
The muddy bottom, although seemingly solid, had several times
given way suddenly, so that it was impossible to ascertain the real
depth of water. But alas, here if anywhere ought to be good collecting.
The vegetation grew from the margin nearly to the middle of the
lake. We visited the place soon after landing and noted that the
Plants were too immature to collect ; incidentally, we discovered by
stepping into the water, that good-sized leeches were there in abund-
ance. Just before leaving the island, we went back again to pick up
Whatever we could without venturing too far from the bank. In
three weeks’ interval a huge colony of Scirpus acutus had grown from
# harrow strip along the margin almost to the centre of the lake.
We obtained Potamogeton Oakesianus, Scirpus subterminalis, Carex
rostrata var. utriculata, Juncus militaris, Utricularia geminiscapa, U.
ochroleuca, Aster nemoralis and var. major. U. ochroleuca has been
Teported in North America only from Greenland; and while this may
*ppear to be a long range-extension, both Professor Fernald and I
believe that the specimens (which unfortunately are sterile) compare
favorably with authentic material of this species of Hartman in the

Tay Herbarium. Petrie’s Pond was very small and had a quaking
Margin from which grew a profusion of Carex limosa.

It was our aim to make a complete collection of the vascular flora
of St. Paul. Unfortunately, we did not reach a small part of the island,
served for a last venture since we had been assured that it was almost
inaccessible. Four days in advance, we knew that a boat was coming
from the mainland for a government official and that we might return
with him. By working into the wee small hours of the morning, we
had hoped to clear our presses and yet to have left an entire day in

114 Rhodora [May

which to pick up at least one specimen of each of the desirable plants
of Mica Head, the unexplored region. We had not one day but four
so filled with heavy fog that this last field-trip faded into an impos-
sibility. In spite of the lacuna, our collection is fairly representative
of the flora of St. Paul. It includes about 2360 sheets of specimens, in
all 418 numbers embracing 160 genera and approximately 300 species
and varieties. In large part, these are common to the flora of Nova
Scotia; some are found only in Cape Breton and St. Paul, while others,
20 in all, were collected in the province for the first time last summer
and most of them illustrate definite range-extensions from the north.
It is significant that these northern plants are characteristic of slightly
or not at all glaciated regions in Newfoundland, Quebec, Anticosti
or the Magdalen Islands; hence, if I may draw a parallel conclusion,
it seems not unreasonable to infer that St. Paul also escaped denuda-
tion by the Wisconsin ice-sheet.

I am under obligation to many people who have aided me through-
out my work on the flora of St. Paul Island. First of all, I wish to
express my appreciation to Professor Fernald both for suggesting the
trip and for the expert assistance and criticism which he has kindly and
promptly given from the beginning of our preparation. To Dr. B. L.
Robinson, I am deeply grateful for the generous support which made
this trip possible, as well as for the personal interest he has shown
in the work. I am greatly indebted to all the Canadian government
officials and employees who contributed to the success and pleasure
of our little expedition. I wish particularly to thank Mr. E. W.
Hawken, Assistant Deputy Minister of the Marine Department, and
Mr. C. H. Harvey, Agent of Marine and Fisheries at Halifax, for
permission to visit the island; Major J. J. McLean, Superintendent
of Lighthouses, for shipping our equipment on the “Lady Laurier
and also for many other courtesies; and Mr. H. M. Sutherland,
Division Superintendent of Radio, for the special favor of permitting
us to board at the Wireless Station. I take much pleasure in thanking
Miss Campbell for a map of the island which greatly facilitated our
work; and, I am highly appreciative of the helpful and courteous
assistance of the government employees during our sojourn 0? the
island. Their spirit of friendliness added much to the enjoyment of
our work there.

i Ti eee Sk a le

1931] Perry,—Vascular Flora of St. Paul Island, Nova Scotia 115

VASCULAR PLANTS COLLECTED ON St. Pau Istanp, 1929

In the following list, the names of the species marked * are recorded
for the first time, apparently, from Nova Scotia. The introduced
species are in italics.

sedi VIRGINIANUM L. Between Atlantic Cove and Lena

airly common on the southern _ of ome —

Shei LATIUSCULUM (Desv.) Hier

ATHYRIUM ANGUSTUM (Willd.) pe var. onan oe (Gilbert)
Butters. F en nt throughout the woods.

THELYPTERIS Puecopteris (L.) Slosson. Abundant shee egaa

ia NOVEHORACENST (L.) Nieuwl. General in the ope

T. sprnu ae F. Muell.) Nieuwl., var. INTERMEDIA (Muhl.)
Nieuwl. ‘e

i¥ Se iGioss 0. F. Muell.) Nieuwl., var. americana (Fisch.)
Weathe tby. Common

Almost as Slack as the two varieties just mentioned was a
form transitional between the two.

OsmuNDA REGALIS L., var. SPECTABILIS (Willd.) Gray. Margin of
Petrie’s Pond and Lena ta ke.

0. CLayrontana L. Occasional on the southeast side of the island.
‘al CINNAMOMEA L. Frequent in the swampy region of Atlantic

Equiserum ARVENSE L. Com

E. arvense L., f. DECUMBENS (Meo) Klinge. Muddy bottom of
stream between Petrie’s Pond and White Spring.
var. Henin: Sy Milde. Woods near the

Prey iy Cane in Nova Scotia on the Barrasois River, Cape

Breton

L. uucwutum Michx. General southeast of Lena
ties pe aro L. Muddy margin of water-hole,
ite

oo See L. Dry hillock near the path to the North East

Lake.
bog at head of

: Neither this nor the following species were seen elsewhere on the
island.
Ligh opscuruM L. Dry hillock near the path to the North East
~
SELAGINELLA SELAGINoIDES (L.) Link. Borders of tussocks; over-
Sorin ging aay of sluggish stream; between Petrie’s Pond and White
ar

cere MacRospora Dur. Abundant in Ethel Lake.

116 Rhodora [May

Not frequent in the province.

Taxus CANADENSIS Willd. General on the southern part of the
island.

rset oe (Mill.) BSP. Commonly distributed throughout
the for

erat BALSAMEA (L.) Mill., var. pHaNEROLEPIS Fernald. Every-
where. This rather than the typical form is the predominant tree on
the on nd.

JUNIPERUS ComMuNIS L., var. MEGISTOCARPA Fernald & St. John.
fone of forest, Trinity Cov.

Known also on Sable island, N ova Scotia.

J. communis L., var. MONTANA Ait. Trinity Cove barren. Not
soe tiful.

Bie ETON OaKEsIANus Robbins. Shallow water, Lena Lake,
P. eprHyprus Raf. Pond in bog above Petrie’s Pond.

ZOSTERA MARINA L. In wash of tide.

+ wou maRiTiMA L. Stagnant pool in rocks, North East

T PALUSTRIS L. Sluggish stream between Petrie’s Pond and
White Spring.

PANICUM BOREALE Nash. Banks of stream just mentioned.

Abilieaitoss sean L. Abundant in all the open places.
Pace EFFUSUM L. Woodland, southeast slope of Coggin “Moun-

ae nicky ASPERIFOLIA Michx. Wooded hillsides near Trinity

Phieum pratense L. yc el in — athe

AGROSTIS STOLONIFERA Ee pacta Har Barren slopes.

A. _ RYEMALIS (Walt.) BSP. y eg at head of White Spring. The
awn

Ps A. —— X A. sTOLONIFERA. [Det. Dr. M. 0. Malte]. Atlantic
ov

CaLaictont Pickertnci Gray. Bog above Petrie’s Pond.
This grass is rare on the islan

C. CANADENSIS oo Nutt. Abundant in exposed places. The
predominant grass on St. .

MOPHILA besviactars Fernald. Wind-swept barren, Trimty

Cove. Not noted elsew her :

CINNA LATIFOLIA (Trev Griseb. Woodland between Whistle
Point and South Wes .

DescHampsta emia a (L.) Trin. General on the barren and im
the woods on the west side of the island.

|
|

er NN Sa een eat aL ire

1931] Perry,—Vascular Flora of St. Paul Island, Nova Seotia 117

*D. rLexuosa (L.) Trin., var. MonTANA (L.) Ledeb. Barren,
Trinity Cove.

Var. montana is essentially a plant of the northern regions. It is
known from Greenland, Labrador, Newfoundland and Quebec
(Gaspé and Matane Counties). This is the first collection south of

ec.
Avena sativa L. Waste — Atlantic Cove.

Se eonne spicaTA (L.) Beauv. Occasional in rocky places

Spartina Micuauxtana Hitchcock. Near edge of aes slope
southwest of N. E. Channel.

We were very much surprised to find this plant of river banks and
other wet habitats growing with Carex silicea in a sheltered spot at
least fifty feet above the sea level, but evidently it derived sufficient
moisture from the drainage of the land above it to thrive.

Poa pratensis L. Occasional, Atlantic te e.

P. trivialis L. Waste places, Atlantic Cov
< toa CANADENSIS (Michx.) Trin. Baniee of swamp, Atlantic

errs PAUPERCULA (Holm) Fernald & Weatherby, var.
ALASKANA (Scribn. & Merr.) Fernald & Weatherby. Crevices of

_PStuca RuBRA L. Rocky slopes, North East Light.
Festuca rupra L., var. ARENARIA (Osbeck) Fries. With F. rubra.
Var. arenaria is another native of northern lands; it has been col-
lected previously in Greenland, Newfoundland and Quebee (Matane
and Gaspé Counties).
F. rupra L., var. yuncea (Hack.) Richter. Crevices of cliffs,
ve.

Ton re

*A. ee NUM (L) Beat, eae Bites co (Koch) Pease & Moore.
Barren, Trinity Cov

ar. eae ae been reported as far south as the alpine
tegion of New Hanpshire, but I have not found anything to indicate
that it has been collected previously in Nova Scotia.

ELEOCHARIS PALUSTRIS (L.) R. & S., var. Masor Sonder. Near the
margin of Lena Lake. Scarce.

bogs.
: S. HUDSoNTANUS (Michx.) Fernald. Peat bog between Petrie’s
ond and White Spring. Only a few patches seen.

. RUBROTINCTUS Fernald. General in swampy land.
ATROCINCTUS Fernald. Swamp, Atlantic Cove.

oe. 4 Rhodora [Mary

ERtoPHORUM ANGUSTIFOLIUM Roth. General in bogs and swamps.

E. “Lait Praca a dare (En, gelm.) Fernald. Occasional in bogs.

E. vireinicum L. Abundant i 7 swamp, Atlantic Cove.

Re coe. ALBA (L.) Vahl. Bogs

CAREX HORMATHODES Paid: Wet hollow i in barren, Trinity Cove.

C. Hower Mackenzie. Peat bog at head of White Sprin ring.

This is a southern species which as yet has not been found farther
nort

C. scoparta Schkuhr. Common.

C. stuicEa Olney. Barren near the edge of the sea cliffs.

*C. eynocrates Wormsk. Bog between Petrie’s Pond and White
Spring. Rare.

C. Exitis Dewey. Sphagnous margin of Petrie’s Pond.

C. STELLULATA Good., var. CEPHALANTHA (Bailey) Fernald. Com-
mon.

C. INTERIOR Bailey. Peat bog at head of White Spring.

C. BRUNNESCENS Poir., var. SPHAEROSTACHYA (Tuckerm.) Kiikenth.
Frequent in the open woods.

. TRISPERMA Dewey. Wet mossy fir woods near Lena Lake.

C. strpata Muhl. Occasional in swales.

C. maritima O. F. Muell. Swales on barren, also in muddy pockets
in rocks.

C. crinira Lam., var. GyNanpRa (Schwein.) Schwein. & Torr.
Swamp i in woods near the Radio Station.

ay, 3 es Lam., var. stmuLans Fernald. Woodland, Coggin
Mounta
ames Now Gay. Profuse everywhere
: artless Light. ats above Petrie’ S Pond.

C. ARCTATA Boott. Occurring casually in the open woodland.
C. DEBILIS Michx., var. RuDGEI t Bailey. Swampy woods near the
n.

: ibe er Rudge. By streamlet from the dam near the
Radio Sta
C. = i Stokes. Swamp near Petrie’s Pon
C. ROSTRATA aes var. UTRICULATA Boot) “Bailey. Shallow

RIOCAULON sce With. Ethel Lake.
JuNcus BuFontus L. Abundant sreeyaner? on open spots.
J. TENUIs Willd. Common along
J. Bauticus Willd., gs LITTORALIS Teca Windswept barren,
Trinity Cove southwes

1931] Perry,—Vascular Flora of St. Paul Island, Nova Scotia 119

J. ritirorMis L. Swale, Atlantic Cove
a EFFusUS L., var. soLuTUS Fernald & Wiegand. Margin of Lena

a EFFUSUS L., var. compactus Lej. & Court. Freque

J. BREVICAUDATUS (Engelm.) Fernald. Abundant between Lena
Lake and ae s Pond.

J. canapEnsis J. Gay. Peat bog at head of White Spring. Rare.

d. a Big Shallow water, Lena La -

. ARTICULATUS L. Common in open place:

J. Peecerarcs L., var. OBTUSATUS Rigel Wet path toward the

ae fea ap Light.
LA SALTUENSIS Fernald. Norwegian Hea

L. I caurestns (L.) DC., var. MULTIFLORA ai Celak. Common

in fi

*L. campestris (L.) DC. var. comosa (Meyer) Fernald & Wiegand.
Woodland path near Lena Lake.

This variety, generally distributed from Montana west and north-
west, has been collected in eastern America, in western Newfoundland
and at one station in Gaspé County, Quebec.

CLINTONIA BOREALIS (Ait.) Raf. tee ee throughout the woods.

SMILACINA STELLATA ead Desf. Common on the ie "3

AIANTHEMUM CANA E Desf. General} in the for
TREPTOPUS jee ac (L.) DC. Frequent at the southwest
end of the island.

S. RosEus Michx. Fairly ies distributed throughout — woodland.

Tr1s Verstcotor L. In swamps and on dry rocky pla

I. serosa ame ., Var. Gui bieas Foster. Near the cliffs on head-
lands and barrens.

De ceticero ANGUSTIFOLIUM Mill. Common

HaBENARIA DILATATA (Pursh) Gray. Usual in ‘bog

te Sprin eens
a ccon Sa: ep ree. R. Br. Occasional in bogs and
swam

wan corpaTa (L.) R. Br. Damp woods near Lena Lake.
Mata AXIS UNIFOLIA Michx. Sphagnum bog, Atlantic Cove.
*Saurx Uva-urst Pursh. Wind-swept barren southwest of N. E.

Channel

Although this species ranges farther south on the alpine summits
of New England and New York, probably this locality is the southern
limit of its aaa, near the sea-level.

: Pursh, var. CALLICARPAEA (Trautv.) Fernald.
(Prostrate). ics at the north end of the main island.

120 Rhodora [May

“Var. callicarpaea is common from northern Labrador to north-
western Newfoundland and the Shickshock Mountains, Quebec”.

Myrica CAROLINIENSIS Mill. arce.

BrTULA PAPYRIFERA Marsh. Infrequent, but general.

B. pumita L. Bogs. Casual.

This species has been collected by Nichols in Cape Breton but
it is not commonly distributed through the province.

ALNUS cRISPA (Ait.) Pursh, var. MoLuis Fernald. Fairly common.

*GEOCAULON LivipuM (Richardson) Fernald. Barren near Petrie’s
nd.

Rumex crispus L. Abundant in old fields.

R. obtusifolius L. Interspersed with the last named species.

R. Acetosella L. Apparently profuse in the fields and occasional on
the barrens.

PoLyGonum avicuLarE L. Atlantic Cove. Weed.

P. Persicaria L. Atlantic Cove. Weed.

P. scabrum Moench. Atlantic Cove. Weed.

P. Ran Babington. Sand beach beyond West Landing. Only one
plant seen.

Chenopodium album L. Occasional on hillsides.

ATRIPLEX PATULA L., Var. HASTATA (L.) Gray. Sand beach beyond
West Landing. Scarce.

SPERGULARIA LEIOSPERMA (Kindberg) F. Schmidt. Grassy slope,
North East Light.

STELLARIA BOREALIS Bigel., var. FLORIBUNDA Fernald. Between
West Landing and South West Light. Rare.
i ds.

“SILENE acauLis L., var. exscapa (All.) DC. Abundant at the
southwest end of the island, also south of N. E. Channel practically
at sea level.

This is the first record near sea level southwest of Newfoundland.

NYMPHOZANTHUS VARIEGATUS (Engelm.) Fernald. Ethel are

NYMPHAEA oporata Ait., var. RosEA Pursh. Petrie’s Pond an
Lena Lake.
we Cympatarta Pursh. Grassy slope, North East

ight.

: L. Abundant in fields.
R. repens L. Vying with the last named species in profusion.
THALICTRUM POLYGAMUM Muhl. General, variable; on rolling
slopes up to 1 m. high, in bogs as small as 3 dm. in height.

1931] Perry,—Vascular Flora of St. Paul Island, Nova Scotia 121

CopTis GROENLANDICA (Oed.) Fernald. Common in the forest.

AcTAEA RUBRA (Ait.) Willd. Open woods at the foot of Coggin
Mountain. Not seen elsewhere on the islan

Thlaspi arvense L. Between West Taide and South Bb Light.

Capsella Bursa-pastoris (L.) Medic. Weed, Atlantic Cove.

CAKILE EDENTULA (Bigel.) Hook. On cliffs and jane eahed

Brassica j juncea (L.) Cosson. Weed near the Radio Station.

B. arvensis (L.) Ktze. Waste places

R HIRTE LLUM Michx. eee pretest but casua
RUS ARBUTIFOLIA (L.) L. f., var. ATROPURPUREA ‘Britt: Robin-
son. Rocky margins of the lakes
*X P. ARSENII eaoe Arséne. (P. arbutifolia var. atropurpurea
X P. dumosa). Freq

P. oe iGiecic) Fernald. Thicket bordering bog above
Petrie’s Pon

Be ccmen Frernatpu Wiegand. Margin of Ethel Lake.

The species thus far is restricted in its distribution to the Gulf of
St. Lawrence region. It has been collected on the Magdalens, Anti-
Costi, Quebec (along the lower St. Lawrence), Newfoundland and,
ae St. Paul
Ps 2 Segui (Tausch.) Roemer. Between Ethel Lake and

ntic

Fr daa se ANA Duchesne, var. TERRAE-NOVAE (Rydb.)
Fernald & Wie _. ay the dam near the nee Station. General.

Atlantic
Fr. mt Howell. Rocky slopes and near ga of cliffs.
P. TRIDENTATA Ait. Rocky places. Comm
P. FRUTI ies L. Bogs. : :
oo IDAEws L., var. CANADENSIS Richardson. Occasional in the

4 CuaMarmorus L. Fairly general i in bogs.
R PUBESCENS Raf. Abundant in the open places
Rak RECURVICAULIS Blanchard. Peat bog ot head of White Spring.
*SANGUISORBA CANADENSIS L., var. LATIFOLIA Hook. Frequent on
8tassy slopes,

An Alaskan variety heretofore known in the east only from the

122 Rhodora [May

north shore of the St. Lawrence River (two stations), Quebec and the
island of Anticosti.

Rosa nitipa Willd. Casual in boggy places.

PRUNUS PENNSYLVANICA L. f. Top of Coggin Mountain, also
between Lena Lake and Petrie’s Pond. Rare on the island.

Trifolium pratense L. Fields.

T. repens L. Near edge of cliffs, Money Rocks.

*OXYTROPIS JOHANNENSIS Fernald. Abundant at the northeast
end of St. Paul.

Although this species has not been collected in Newfoundland south
of Cape St. George, its range extends south from Quebec into Maine
and east along the upper St. John River in New Brunswick. This is
undoubtedly the plant reported in Macoun’s Catalogue (i. 115) from

t. Paul as O.uralensis, var. pumila, and later (i. 509) as O. aretica.

Vicia angustifolia (L.) Reichard, yar. segetalis (Thuillier) Koch.
Weed, Atlantic Cove.

Latuyrus Maritimus (L.) Bigel. Border of cliffs, Trinity Cove.

L. patustnis L., var. MacRANTHUS (T. G. White) Fernald. Grassy
slope, North East Light.

Although found in Nova Scotia, it is more plentiful farther north.

HYPERICUM CANADENSE L, Path, near Lena Lake. Scarce. a6
VIOLA CUCULLATA Ait., f. PRIONOSEPALA (Greene) Brainerd. Fields.
d

V. incognita Brainerd. Deep woods southeast of Lena Lake.
EpiLosrum aneustirouium L. Atlantic Cove.
. GLANDULosUM Lehm., var. ADENOCAULON (Haussk.) Fernald:

Open woods, Atlantic Cove,

E. patustre L., var. Monticota Hausskn. Common in bogs.

OENOTHERA PERENNIS L. Waste laces, Atlantic Cove. |
Sindee ALPINA L. Dripping slope of open woods, Coggin Mou

in :

ARALIA NUDICAULIS L. A characteristic woodland species.
Carum Carvi L. Weed, Atlantic Cove. liffs.
Ligusticum scornicum L. Rocky slopes usually near the ¢

HERACLEUM LANATUM Michx. Swale, Coggin Mountain.

.
i

1931} Perry,—Vascular Flora of St. Paul Island, Nova Scotia 123

CoNIOSELINUM CHINENSE (L.) BSP. F airly general on the island,
but apparently not so in the province.

*Cornus suEcicA L. Sphagnous depression in barren, Trinity
Cove southwest

This boreal iuiaes occurs in Greenland, Newfoundland, Quebec,
Anticosti, the Magdalens, St. Paul and Alaska. After collecting
several sheets of the material at Trinity Cove, Miss Roscoe and I
noticed near Martin Power’s Cove an abundance of Cornus, the
berries of which were very like this species but the leaves were much
larger. In passing we thought it C. suecica, but I now believe it was
rather C. canadensis var. intermedia Farr. However, since we took
no specimens, I cannot make this statement with surety.

C. CANADENS

MonEsks UNIFLORA (L.) Gray. Common in the woods.

tie LORANTHA Sw. Woodland near Lena Lake.

wa L.

Vi ONOTROPA UNIFLOR Freq

ite Besa axwicon Oeder. oa vell distributed but not anywhere
abun .

A ANGUsTIFOLIA L. Chiefly at Atlantic Cove.

cto Rotsror. Wang. Sphagnum bog, Atlantic Cove. Not seen
e
_nmnontens GLAUCoPHYLLA Link. Bog above Petrie’s Pond.

CHAMAEDAPHNE CaLycuLaTa (L.) Moench. Bog on the left of the
path to North East Light.

GAULTHERIA PROCUMBENS L. Bog above Petrie’s Pond.

> Sag HISPIDULA (L.) T. & G. Abundant on knolls in open
woo

GAYLUSSACIA DUMOSA (Andr.) T. & G., var. Se Fernald.
Bog on the left of the path to North East Ligh

Vaccrytum Vitis-Ipaka L., var. MINUS Lodd. Freq
Fs a Oxycoccus L., var. INTERMEDIUM Gray. Near ‘diffs, Lookout
oin

v. PENNSYLV ANICUM Lam. Woods, Atlantic Cove. Rar

V. unicrnosum L., var. aLpInuM Bigel. Upper slope of Alco,
West Point; South West Li ght.

Previously collected by Nichols on the mountains west of Ingonish,
but not generally distributed in Nova Scotia. —

MULA MISTASSINICA ie Banks of streamlet between Pet-
Nie’s toad an ite Sprin
YEMMACHTA TERRESTRIS (L.) BSP. Shallow sabe Ethel Lake.

TRreNTaLIs BOREALIS Raf. Common in the
N, UX MARITIMA L., var. OBTUSIFOLIA Fernald. Rocky slope,

orth East Light.

124 Rhodora [May

HALENIA DEFLEXA (Sm.) Griseb. Hillside, Martin Power's Cove.

This species is listed by Nichols (Vegetation of Northern Cape
Breton, 324, 1918) as common on bleak exposed headlands. I do not
find any other record of its occurrence in the province.

MENYANTHES TRIFOLIATA L., var. MINOR Michx. Stagnant pools
and bogs.

ConvoLvuLus sepium L., var. PUBESCENS (Gray) Fernald. At-
lantic Cove.

MERTENSIA MariTIMA (L.) S. F. Gray. Sand beach beyond West

nding.

PRUNELLA VULGARIS L., var. LANCEOLATA (Barton) Fernald.
Fields, Atlantic Cove.
Galeopsis Tetrahit L., var. bifida (Boenn.) Lej. & Court. Waste
C

v
_— PURPUREA Reeks, var. Fartownt (Rob.) Fernald & Wiegand.
ns

rrens.

E. AMERICANA Wettst. Field, Atlantic Cove. Frequent.

E. CANADENSIS Townsend. Field, Atlantic Cove. Abun

*RHINANTHUS GROENLANDICUS Chabert. Near ruins of an old
house, Trinity Cove. Not noted elsewhere on the barren.

This boreal species is represented in the Gray Herbarium by speci-
mens from Greenland, Labrador, western Newfoundland, Quebec
(Saguenay Co.), Anticosti, the Mingan Islands and Alaska.

R. -Cusracaits 1. var. paizax (Wimmer & Grab.) Druce.
Abundant in fields, Atlantic Cove.

TRICULARIA GEMINISCAPA Benj. Lena Lake. .

U. minor L. Pools of streamlet between Petrie’s Pond and White
Spring.

The specimens differ from those of the typical form in having
bladders somewhat larger than usual and flattened rather -
terete leaves. Possibly they correspond to f. platyloba Meister,
which is merely the result of an ecological reaction to creeping out
on the mud.

*U. ocuroteuca R. Hartm. Lena Lake. :

This dainty little plant is most closely related to U. intermedi
which ordinarily may be identified by the fact that the bladders at

Tne on separate leafless branches. U. ochroleuca, too, has seP# .

1931] Perry,—Vascular Flora of St. Paul Island, Nova Scotia 125

a. leafless branches, but it differs from its relative in
dders also on the leaves of the immersed stems; moreover,

the a along the margins of the leaves are larger than those char-
acteristic of U. intermedia. Unfortunately our specimens are sterile,
but Gliick intimates that sterility is not an unusual occurrence in
this species. U. ochroleuca is fairly common in the northern countries
of Europe. It has been reported from two localities in Greenland,
but this is the first record south of there; hence, it is not only an
addition to the flora of Nova Scotia but also to that of Canada.

PINGUICULA reamed L. Banks of streamlet between Petrie’s
Pond oy White Spr

Found also on one nett but not elsewhere in the province.

PLANTAGO JUNCOIDES Lam. toward var. LAURENTIANA Fernald.
Rocks, North East Light.

P. JUNCOIDES Teck var. gLauca (Hornem.) Fernald. Hillside,
Martin Power’s Cove.

a os Lam., var. pEcIPIENS (Barneoud) Fernald. Barrens.

rps a. L. Field, Atlantic Cove.

GaLiuM TRIFLoRUM Michx. Woods near the Radio Station.

G. CLaytont Michx. Common in swamps and wet open places,

MircHeia REPENS L. Near the outlet of Lena Lak

Lonicera vitLosa (Michx.) R. & S., var. SOLONIS (Eaton) Fernald.
Margin of peat bog at head of White Spring.

L. CANADENSIS Marsh. Woods near Ethel Lake. -

This is the northeastern limit of its range

Spycam BOREALIs L., var. AMERICANA (Forbes) Rehder. Abundant
ef

linens casstnowEs L. Casual on the southern part of the

Sampucus racemosa L. Open places in the woodland.
this lense * ROTuNviFotta L. Hillside, Martin Power’s Cove. In
locality we found one white-flowered form. The typical was
li

ROTUNDIFOLIA _ var. ALASKANA Gray. Rocky slopes

This Alaskan variety has been collected east of British Columbia
only in western Newfoundland and Gaspé (one collection), and, now,
St. Paul (Nova Scotia).

Lopetta Dortmanna L. Ethel Lake.
i The s specimens differ slightly from the typical in the more pubescent
1p of the cor olla.
SoLIDAGo Bicotor L. Rocky slope above South West Light.

126 Rhodora [Mar

S. MACROPHYLLA Pursh. Open forest. Fairly common.

S. PUBERULA Nutt. Woodland near Martin Power’s Cove.

*S. MULTIRADIATA Ait. Barren southwest of N. E. Channel.

Distributed in the west from Manitoba to Alaska. In the east,
reported from western Newfoundland and Quebec (Gaspé and Matane
counties).

S. SEMPERVIRENS L. Both near and on c

S. untLtiguLata (DC.) Porter, var. Rone (T. & G.) Fernald.
Peat bog at the head of White Sprin

. RUGOSA Mill., var. SPHAGNOPHILA Graves. Wet gulch, Nor-

wegian Mountain.

ae Ea tel (L.) Britton. Woods, Atlantic Cove.

A. A Ait. Swamp, Atlantic Cove. Part of the specimens
oe iat asia var. STRICTUS.

A. Novi-BELGH L. Casual on the barren.

A. UMBELLATUS Mill. General in open places.

ake.
EMORALIS Ait. , Var. MAJOR Peck. Margin of Lena Lake, growing
with ne typical form
ALIS MARGARITACEA (L.) B. & H. Hillside, Atlantic Cove.
A. with Meee: (L.) B. & H., f. anocutora Fernald. Barren,
Trinity Cove peek ia
Gnaphalium uliginosum L. Waste places, Atlantic Cove.
Achillea Mi illefotium I a Clearing near South West Light.
atricaria suaveolens (Pursh) Buchenau. Hillside, Atlaarte Cove.
Chrysanthemum Leucanthemum L., var. pinnatifidum Le
Lamotte. Abundant in the fields, Atlantic Cove.
» Commie muTicum Michx. Casual. Only three or four specimens

wg arvense (L.) Scop. Swale, base of Norwegian Mountain.
Centaurea nigra L. Field, Atlantic Cove. At
Leon; autumnalis L., var. pratensis (Link) Koch. Weed, 4+
lantic Cove.

Taraxacum officinale Weber. Atlantic Cove. d

PRENANTHES icenue (Cass.) Fernald. In open forest am
along borders. Com d

. TRIFOLIOLATA (Cass) Fernald, var. NANA (Bigel.) F eral

Barren, Trinity Cov

GRAY fede

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

XCV.

ISSUED AUG 24 1931

I. Taxonomic Studies in Cuphea. By Rimo BaciGaLupPt.... .
II. eee Flora of the Guano Islands of Peru. By

Oe ia Te ee Ee eR RO ae eC ene ree vay

IV. Studies | in the Bromeliaceae—II. By L. B. SMITH......--
V. The Group of Aspleniwm fragile in South America. By C.
Ie MEATHERUY. oobi ee eee
VI. New oe from Mexico and Argentina. By I.
Be WONTON | oc co ei es ns Se ae eee cet

PUBLISHED BY :
THE GRAY HERBARIUM OF HARVARD UNIVERSITY
CAMBRIDGE, MASS., U. S. A.

I. TAXONOMIC STUDIES IN CUPHEA.
By Rimo Bacicaurt.
(Plates I-V.)

Tue large genus Cuphea, which already includes some 240 species
and several score varieties, is one whose many technicalities render
it particularly fitting that it be left in the hands of a specialist. By
far the largest genus of the Lythraceae, it early attracted the attention
of Emil Koehne of Berlin who, although known more generally, in
Germany if not abroad, for his extended and distinguished dendro-
logical researches, soon became established, in international taxo-
nomic circles at least, as the world’s foremost authority on this
largely tropical American family. His contributions to the knowledge
of this group, as indicated by the dates of his many publications
among which are included two complete monographs of the Lyth-
raceae, extended from 1873 to 1913, at which latter date began the
series of ailments and disabilities which culminated in his death
five years later.

Shortly after his death, the Gray Herbarium was fortunate enough
to acquire, through purchase, Koehne’s unpublished “Atlas Lyth-
Tacearum,’’ a set of original habital and diagnostic drawings in
Which every known species and variety of the Lythraceae published
Up to the time of his fatal illness had been delineated with consummate
skill by this thorough scholar. Interest in Cuphea as a possible
subject for taxonomic investigation was initiated by the availability
of this valuable document; indeed, without such an exceptional aid,
the writer would not have been encouraged to attempt detailed
study of a genus so large and complex. Moreover, it soon became
‘pparent that much interesting and novel material had been col-
lected from Mexico, Central America and more especially from
South America since Koehne published his last novelty in Cuphea
(C. bicolor) in 1913. From these more recently collected specimens,
Many of them kindly forwarded to the Gray Herbarium for the
Writer's purposes by Professor Samuelsson of the Museum of Botany
. Stockholm, most of the novelties here presented were described.

ong these were found specimens of four of Koehne’s still un-
Published species of Cuphea, transcripts of the original diagnoses
of two of which were very obligingly transmitted by Professor Diels
“nd Dr. Melchior of the Botanical Museum at Berlin-Dahlem.
During the summer of 1930, the writer enjoyed the privilege of

4 BACIGALUPT

examining Koehne’s collection, now permanently housed at the
herbarium in Berlin-Dahlem, among which the many Cuphea types
studied helped immeasurably to clear up dubious details. Other
European herbaria visited in this connection are those of the Con-
servatoire de Botanique in Geneva, the Institut de Botanique of
the University of Geneva, and the Muséum d’Histoire Naturelle
in Paris. A loan of the Cuphea collection of the Dudley Herbarium
of Stanford University, which included the very recent collections
of Mrs. R. S. Ferris from the west coast of Mexico, yielded many
interesting specimens from which one of the sixteen novelties pre-
sented in this paper materialized.

Koehne’s lifelong studies on his chosen family of plants were
brought together in his second monograph on the Lythraceae in
“Das Pflanzenreich,” issued in 1903. Later, he twice published,
in Engler’s serial “Botanische Jahrbiicher,” supplementary accounts
embodying his subsequent work, consisting in large measure of
novelties brought to his attention since the general treatment and,
in lesser degree, of emendations to statements therein. The results

Just cited. The exceptions are Cuphea delicatula T. S. Brandegee,
Zoe v. (1905) 213, C, humayana T. S. Brandegee, loc. cit. 214, ¢.
pannoso-cortica Rusby, Bull. N. Y. Bot. Gard. iv. (1907) 361, ©
Goldmanii Rose, Contrib. U. S. Nat. Herb. xii. (1909) 287, fig. 5:
C. trichandra Brandegee, Univ. Calif. Publ, Bot. iv. (1910) 89 =

C. Purpusii Brandegee, loc. cit. (1913) 378. Several species desct!

STUDIES IN CUPHEA 5

by and attributed to Rose were published by Koehne and subse-
quently republished by their author, loc. cit. 287-290.

Since 1913, various workers have published new species of Cuphea.
These are C. chiapensis Brandegee, Univ. Calif. Publ. Bot. vi. (1915)
187, C. chiquitensis Herzog, Mededeel. Rijks Herb. xxvii (1915) 16,
C. petalosa Herzog, loc. cit. 17, Parsonsia blepharophylla Blake,
Proc. Biol. Soc. Wash. xxxii (1919) 190 [C. blepharophylla (Blake)
Riley, Kew Bull. (1924) 211], C. scelopetala Riley, loc. cit. 212,
Parsonsia salvadorensis Standl. Journ. Wash. Acad. Sci. xiv. (1924)
240 [C. salvadorensis Standl. Field Mus. Publ. Bot. iv. (1929) 238],
C. Eckmanii O. C. Schmidt, Arkiv Bot. 20A, no. 15 (1926) 82, C.
pergracilis O. C. Schmidt, Fedde Rep. Spec. Nov: xxiv. (1927) 77,
and C. riparia Eck. & O. C. Schmidt, Fedde Rep. Spec. Nov. xxvii.
(1929) 106. C. tuberosiformis Koehne ex Dusén & Neger, Beihefte
Bot. Centralbl. xxxviii, Abt. 2 (1921) 295, is virtually a nomen
nudum. It was published in a descriptive account of the woody
anatomy of “ Xylopodien,”’ Lindman’s name for the irregularly
thickened, often rounded, very hard woody enlargements of the
toot or of the basal underground portion of the stem characteristic
of the § Oidemation in Cuphea. Only the anatomical peculiarities
of the wood-structure of the xylopodium of C. tuberosiformis are
here described. The characterization of this species is therefore
included among the following diagnoses and comments.

The writer wishes here to express his appreciation of the friendly
*neouragement and scholarly advice so unstintingly given by Pro-
fessor B. L. Robinson under whose guidance the work here presented
Was carried out. Both he and Mr. C. A. Weatherby of the Gray
Herbarium staff have aided in giving some semblance of botanical
style to the Latin diagnoses characterizing the novelties. Professor
Briquet of the Conservatoire de Botanique in Geneva authorized
the transmittal of specimens which proved of invaluable aid. For
help in the matter of photographing types, I am indebted to Drs.
L. B. Smith and A. N. Steward.

§ 2. ENANTIOCUPHEA.
_Cuphea obtusifolia Koehne in MS., spec. nov. “‘Herba perennis

6 BACIGALUPt

setuloso, bracteis infimis 5-7 mm. longis, supremis minimis, omnibus
herbaceis, setis glanduliferis parce ciliatis. Pedicelli 5-9 mm. longi,
tenuiter filiformes, violaceo-setulosi. Calyx basi gibbus, 11 mm.
ongus, laxe breviter violaceo-setulosus, setulis glanduliferis, intus
supra stamina lanatus, intus fundo glabro excepto dense. villosus;
lobus dorsalis maximus productus. Petala 6, dorsalia 2 rosea nervo
medio saturate purpureo, 5 x 2.5 mm., ventralia 4 lilacina, 4.5x2
mm., omnia cuneato-oblonga. Stamina lobos aequantia, lanata.
Pistillum 10 mm. longum, ovarium apice breviter villosum, stylus
ovario ca. 14 brevior, dense villosus; discus magnus _horizontalis,
supra concavus, subtus convexus; ovula ca. 34-38. Semina rotun-
data, 1.5 mm. longa, tuberculata, pallide crenato-marginata.”—
Plate I.

BRAZIL: Paran&: on marshy ground, Piraguara, Jan. 1, 1909, P. Dusen,
no. 7782 (TyPx, in Herb. Mus. Botan. Stockholm).

This species is to be distinguished from C. longiflora Koehne and
C. racemosa var. divergens Koehne, its closest relatives, by its un-
usually produced large dorsal calyx-lobe and its obtuse glaucescent
glabrous leaves and its likewise glaucescent, absolutely glabrous
stems.

Cuphea Lindmaniana Koehne in herb., hucusque ined. Annua
ex basi ramosa magis minusve hirsutula. Caules ad apices foliati
10-15 em. longi dense hirsutuli praesertim insuper simul sparsius
crispo-puberuli. Petioli foliorum inferiorum saepe 1-3 mm. longi vel

subnulli, ei superiorum nulli vel subnulli. Folia ovato-oblonga vel.

oblongo-lanceolata basi subrotundata vel saepius subattenuata apice
acuta ambis paginis ut caules hirsutula margine scabrida 12-18 mm.
longa, 4-6 mm. lata. Pedicelli ebracteolati oppositi in axillis foliorum
superiorum ut caulis puberuli sed minus hispiduli 4-6, demum 7-8
mm. longi. Calyx parum gibbosus 4-5 mm. longus purpurascens
purpureo-vel albido-hirsutulus, lobo dorsuali flavo margine pune
cum lobis obtusis juxtapostis producto, lobis ventralibus 3 angusts
acutis, intus ventre infra stamina magis minusve_ villosus costulis
infra petala dorsualia villosis sub staminibus ventralibus subae-
qualibus inclusis lanatus. Petala alba, 2 dorsualia 3 mm. lange
3 mm. lata, 4 ventralia 4 mm. longa, 1.7 mm. lata. Pistillum
villosum stylo ovarii tertiam partem aequante. Ovula 28-29. Se-
mina minuta valde alata glabra, 1x 0.8 mm. Discus angustiusculus
adscendens.

tapirusi, “in campo,” Nov. 11, 1908, Po
TyPE, Herb. Mus. Botan. Stockholm).

This species of the series 1 of the subsection Gastrodynameé §

BRAZIL: Parani, I
no. 7156 ('

STUDIES IN CUPHEA ff

Enantiocuphea) is most closely related to C. ramosissima Koehne.

t may be distinguished from that species by its hirsutulous leaves
and calyx, its villose ovary and calyx interior, its shorter petioles
and winged seeds.

Cuphea varia Koehne in herb., hucusque ined. Annua basi
interdum procumbens internodiis infimis saepe radices emittens.
Caules parce ramosi dense glanduloso-hirsutuli simul pilis crispis
parum puberuli basi glabrescentes 12-30 cm. longi. Petioli 1-5 mm.
longi vel foliorum superiorum saepe subnulli, ut caules vestiti. Folia
ovata vel ovato-oblonga basi subrotundata vel late attenuata apice
obtusa, 11-22 mm. longa, 7-13 mm. lata, nervis utrinsecus 4-5 nervo
medioque paginarum inferiorum saepe prominulis margine scabra
ambis paginis eorundem minorum ad basem caulis versus glabra vel
glabrata paginis superioribus eorundem majorum ad apicem caulis
versus plerumque setulis adpressis conspersa simul saepe paene
scabrida paginis inferioribus nervis interdum hirsutula. Pedicelli
ebracteolati quam caules conspicuius crispo-puberuli sparsiusque
glanduloso-hirsutuli 4-11 mm. longi. Calyx sub anthesi atroviolaceo

BRAZIL: Parand: Calmon, “in subpaludosis,” March 15, 1910, P. Dusén,
no. 9266 (TYPE, in Herb. Mus. Botan. Stockholm); Guarapuava, “in paludo-
SS, Jan. 8, 1911, P. Dusén, without number.

This species, together with C. Lindmaniana Koehne, likewise
described in these studies, is related to C. ramosissima Koehne. The
subsection Gastrodynamia (§ Enantiocuphea), of which all three
Species are members, is characterized by having the two dorsal
Petals, singularly enough for Cuphea, shorter than the four ventral
ones.

Cuphea varia Koehne is at once set apart by its dark violet dorsal
Petals and villose seeds. These latter are so small that only with a
lens is their villose character clearly discernible.

CUPHEA ORIGANIFOLIA Cham. & Schlecht. Among the recent
“ollections of E. P. Killip and Albert C. Smith from the eastern

8 BACIGALUPI

Cordillera of Colombia appear specimens referable to Cuphea ori-
gantfolia Cham. & Schlecht. This species had not heretofore been
noted north or west of Minas Geraes, Brazil. Hence these specimens
prove a very interesting extension of range. If they had not had the
erect or semi-erect gland at the base of the ovary characteristic of
this species, these specimens might easily have passed for forms of
C. racemosa (L. f.) Spreng. These Killip & Smith numbers of plants
from the department of Santander, Colombia are 15542, 16012,
17734, and 18276.

Cuphea flavisetula, spec. nov. Herba perennis ramosa_ basi
fruticulosa cire. 20 cm. alta. Rami puberulenti ac praesertim in-
super remote hispidi, petioli subnulli summum 0.5 mm. longi flave-
scentes. Folia lanceolato-elliptica utroque acuta utrinque glabra
remote perspicueque flavescente hispido-ciliata, 10-22 mm. longa,
3-8 mm. lata. Racemi 2.5-6 cm. longi; bracteae 1-3 mm. longae;
pedicelli oppositi ebracteolati ut caulis puberulenti insuper dense
breviterque glanduloso-hispiduli, 4-6 mm. longi. Calyx basi tantum
paullo gibbosus remote flavescenteque hispidus basi interdum glan-
duloso-hispidulus ceterum glaber intus infra stamina uniformiter
pilosus 7-8 mm. longus, lobis ventralibus productis anguste t-
angularibus leviter acuminatis quam dorsales multo angustioribus.
Petala oblanceolata basi attenuata apice rotundata duo dorsualia
retrorsa 2.5 mm. longa, 4 ventralibus dimidio breviora. Stamina a9
ventralia alterne subinaequalia quorum 5 inter lobos calycis sub-
exserta filamentis lanatis, filamenta staminum 2 dorsualium multo
breviora glabra. Stylus villosus calycem aequans. Ovula 12-18.

xe XICO: Tepic Trrrirory (now Nayarit): Acaponeta, June, 1897,
J.N. Rose, no. 1421 (rypr, in Gray Herb.).

This species falls into Koehne’s subsection Gastrodynamia. It is
related to C. decandra Ait. The differences between these two speci
may perhaps best be set down in tabular form.

C. DECANDRA.
Species of West Indies and Colombia.
Leaves broadly oblanceolate - ob-
long-oblanceolate, Aap son age
ensely so, min

i aes densely, a and st se iattalacty

— “dentely glandular-hirtell
tall hope ous,

Ventral ze ue, broadly tri-
angular

C. FLAVISETULA.

Species of Mexican West Coast. ee

rseey ne oe em ee se
ht conspeuousl hip hairs.
id (the ba-

Y Hip i
— laerampprtier sho: Pp: gland a
otherwise pa

wly i
Ventral sepals gente. narrow!
angular and somewhat row te

STUDIES IN CUPHEA 9

CurpHea Purpustt Brandegee, Univ. Calif. Publ. Bot. iv. (1913)
378. This species is based on Purpus, no. 6147, collected at Batios
de Carrizal, Rio de los Pescados, Vera Cruz, Mexico. It agrees
with C. decandra Ait. in all its characters save one, the marked re-
duction of the two dorsal petals. These are subulate-filiform, 1.5
mm. in length (Purpus, no. 7650) and early deciduous so that in
mature flowers they appear to be lacking. Of the lower four petals,
the outer two are 3.54 mm., the inner 4.5—-5.5 mm. long.

It may be recalled that C. decandra, along with several other
species, differs from many of the members of the subgenus Lythro-
cuphea in having the two dorsal petals smaller, rather than larger,
than the lower four, but in this species, the difference in size is not
such a striking one.

Purpusii was originally described as a four-petaled species,
“disco erecto, ovarii adpresso,” but an examination of isotype ma-
terial at the Gray Herbarium revealed a disc almost surrounding the
base of the ovary, interrupted ventrally and greatly thickened dorsally,
as in typical C. decandra.

Koehne (Pflanzenreich IV, Fam. 216, p. 110. 1903) gives Vera
Cruz, Mexico, Haiti, eastern Cuba and Colombia as the geographic
range of C. decandra, but all the Vera Cruz material at the Gray
Herbarium has the dorsal petals, in the rare instances when they are
still present, minute and linear-subulate, and coincides in all other
particulars with typical material of C. Purpusii. It is probable,
therefore, that all Mexican (Vera Cruz) specimens, heretofore re-
ferred to C. decandra, are more properly referable to C. Purpusii.
In addition to the two Purpus specimens mentioned above, Purpus, no.
1493 and Pringle, no. 7825 are referred here. j

It seems better, in the light of the facts presented, to consider C.
P urpusii a geographic variety of C. decandra. The following com-
bination is therefore proposed. Cuphea decandra Ait. var. Pur-
Pusii (Brandegee), comb. nov. ;

: According to Koehne’s drawings (loc. cit. and Atlas Lythracearum
in MS.), the inside of the calyx of C. decandra is never more than
merely biseriately pilose, but in specimens from Cuba (Wright, no.
334) and Colombia (Lehmann, no. 5612; Rusby & Pennell, no. 174),
as well as in all Vera Cruz material referable to C. decandra var.
Purpusii, each of the ribs on the inside of the calyx bears a series
of minute hairs from below the stamens to almost the very base, so
that the whole inside of the calyx appears uniformly pilose. In
Wright, no. 334, this pilosity is not so uniform as in the others, but still
the inside of the calyx is not nearly so glabrous as Koehne indicates.

10 BACIGALUPI

§ 5. BRACHYANDRA.

CupHEA MICRANTHA H. B. K. Pittier, no. 9011 and no. 9749,
from Venezuela, were erroneously distributed as Cuphea strigulosa
H. B. K

CUPHEA CARTHAGENENSIS (Jacq.) Macbride, Field Mus. Pub.
Bot. viii. (1930) 124 (C. balsamona Cham. & Schlecht.; C. hyssopi-
folia Griseb., in part, Fl. West Ind. Ids. (1860) 270, not H. B. K.,
the latter taken up by Hillebrand, Flora of the Hawaiian Islands
(1888) 131, and followed by many Hawaiian authors since; Parsonsia
pinto Heller, Minn. Bot. Stud. ii. 862. 1897). Among specimens of
C. carthagenensis erroneously labelled C. hyssopifolia H. B. K. are
Joseph F. Rock, no. 3137 and Heller, no. 2004.

Cuphea patula, as misapplied by Stewart, Proc. Calif. Acad. Sei.
ser. 4, 1. (1911) 116, not St. Hil. Stewart’s specimens from the
Galapagos Islands, which were named and distributed as C. patula,
proved upon examination to be C. carthagenensis (Jacq.) Macbride
(C. balsamona Cham. & Schlecht.), a species of somewhat weedlike
ability to spread but not previously reported from this island group.

Cuphea Ferrisiae, spec. nov. Herba annua erecta vel semi-
prostrata fere a basi divergente ramosa. Caules hispidi pilis arcuato-
adscendentibus simul crispe pubescentes, 1.5-2.5 dm. alti. Rami ut
caulis vestiti, 7-14 em. longi. Folia elliptica ad apicem basemque
versus attenuata utrinque margineque scabrida supra in 0 ng
medio ad basem versus ut in petiolo crispo-pilosiuscula subtus ™m
nervis interdum sparsissime adscendenter hirsuta basi et in parte
in petiolum decurrente ciliata ceterum glabra, 2.5-4 cm. longa, 14
2.2 cm. lata. Inflorescentiae terminales spiciformes _ bracteatae
rachibus biseriatim crispo-pilosis ad basem versus ut caules hispidis.
Bracteae inferiores euphylloideae lanceolatae acuminato-apiculatae
superiores multo angustiores, 2-2.5 mm. longae. Pedicelli aliquantum
verticellati ut rachis vestiti vel glabrati, 1-3.5 mm. longi, prope
apicem bracteolas in setulas albidas 3 reductas gerentes. Calyx
cylindraceus mox_ scariosus anguste calcaratus calcare interdum

partem petali aequante), 2 mm. longa, 4 ventralia brevius pcr
culata breviora pallidiora cuneato-elliptica. Stamina 11 gam
Inaequalia inaequalissime disposita 4 ex eis lobos aequantia. OV

-

STUDIES IN CUPHEA 11

glabrum fructiferum calycem 2% aequans, stylo lobos ventrales
calycis aequante, stigmate late capitato. Discus tenuis valde re-
trorsus apice subglobosus. Semina 3 exalata leviter retusa minute
scrobiculata, 1.5 mm. longa.

MEXICO: Nayarit: in dense forest, vicinity of San Blas, R. S. Ferris,
no. 5363 (Type, in Dudley Herb., Stanford University).

This species shows relationship to the Cuphea Wrighti group of

the subsection Lophostomum (§ Heterodon) on the one hand, and to
Cuphea Seleri of the somewhat anomalous subsection Lophostomopsis
(§ Brachyandra) on the other. Koehne (Pflanzenreich IV, Fam. 216,
p. 121) admits the relationship of this latter group to C. Wright
and C. secundiflora. Indeed, in general habit this species is not
unlike C. secundiflora.
_ If, because of its slightly produced dorsal lobe, it be, in spite of
its short calyx, included in the section Heterodon, its slender deflexed
disc indicates its relationship to C. Wrightii, a species the calyx of
which is often not more than 5 mm. in length, and to its close rel-
atives, C. lutea Rose and C. viscosa Rose. In addition to striking
habital differences, it may be separated from these by its scarious
calyx, its strict spiciform inflorescence and in having the dorsal
calyx lobe not so strikingly produced as in these.

In addition to the gland character, this species differs from C.
secundiflora in its scarious and merely hispid calyx, in the very
shortly produced dorsal calyx lobe, in having only three ovules and
absolutely glabrous filaments and style.

If, on the other hand, this species be included, because of its short
calyx, within the section ‘Brachyandra, it falls, as suggested above,
into the subsection Lophostomopsis. Here, it may be distinguished
from C. Seleri, its closest relative in this grouping, by its shorter
calyx (7-8 mm. long in C. Seleri), by its longer, more slender and
often slightly incurved spur, by its longer ovary and correspondingly
shorter style, this latter equaling or slightly exceeding the short
lower calyx lobes, in its more slender and decidedly deflexed disc,
subglobose at the tip, and in having the calyx within and the stamens
absolutely glabrous. As in C. Seleri, however, the dorsal calyx lobe
Is longer, even if not markedly, and larger than the others, and is
deflexed finally, helping with the other lobes to close the mouth of
the calyx, as in many species of the section Heterodon.

Owing to the short calyx, it may be considered more expedient, for
keying Purposes at least, to place this species in the subsection Lo-
Phostomo sis

psis.
This species is dedicated to Mrs. Roxana S. Ferris, assistant curator

12 BACIGALUPI

f the Dudley Herbarium of Stanford University, whose recent
collections on the west coast of Mexico have disclosed more than a
few interesting novelties.

§ 6. EUANDRA.

Cuphea laeviuscula, spec. nov. Herba maximam partem glabra
basi suffrutescens. Caules plures ramosi adscendentes saepe 2-3 ex
basi emanantes hinc tantum mox decorticati insuper flavescentes
interdum rubescentes ad apicem versus magis magisque pilis albidis
brevissimisque tantum uniseriatim minutimque crispo-pubescentes
ceterum glaberrimi, 25-38 cm. alti. Petioli glabri foliorum inferiorum
ad 1.5 mm. longi ei superiorum subnulli. Folia decussata lineari-

prominulis basi subattenuata vel subrotundata interdum obliqua
apice leviter attenuata vel acuta vel obtusissima, 20-25 mm. longa,
5-6 et rarius ad 8 mm. lata. Pedicelli 1-3 mm. longi glabri ad apicem
versus bracteolas minutas rotundato-ovatas setulis stipularibus
instructas gerentes. Calyx glaberrimus apud florem juniorem graci-
liter rotundato-calcaratus dorso lateribusque rubescens ad os versus
purpurascens, 7-8 mm. longus, apicibus loborum subaequalium
leviter recurvatis hinc minutim albido-pubescentibus appendicibus
obtusis minutissimis intus sat conferte biseriatim villosus ore sub
staminibus ventralibus leviter floccoso-lanatus ceterum gla .
Petala calyce vix 14 breviora 2 dorsalia intense purpurea ovali-
elliptica 4 ventralia lilacina spathulato-oblonga. Filamenta 2 brevia
dorsualia pilosissima ex 9 ventralibus paria duo exterioral anata 5
interiora longiora glabrescentia sinus sed non lobos calycis superanti
arium glabrum stylo villosiusculo apud florem juniorem ovarium
longitudine aequante et incluso. Ovula 16-18. Semina ala angustits-
a cincta. Discus latus haud crassus dorso convexus.
ARGENTINA: Misiones: Loreto, in the —- about Posadas, moist situa

tions along the Rio Yabebiry, Jan. 14, 1908, E. L. Ekman, no. 1929 (t¥P#, !? -
Herb. Mus. Bot. Stockholm). ee

phoenix, C. acinifolia, C. concinna and C. thymoides. Cup ne
uscula may be distinguished from any of these by its @

glabrous calyx (outside, at least) and by the minute uniseriate pubes-

STUDIES IN CUPHEA 13

cence of the younger stems, which, unlike any of the other species
of its group, are otherwise quite glabrous.

§ Euandra, subsection PiLatyprerus, series 2. Since the last
general treatment of the Lythraceae by Koehne in Das Pflanzenreich
(1903), five new species, all members of this series, were subsequently
described by the monographer. These are C. phoenix, C. concinna,
C. grata, C. corisperma and C. hexasperma. In addition, C. laevius-
cula, described in this paper, has been added by the present writer.
These may be more readily distinguished from each other and from
the other species relegated to this series by the following key, adapted
from the original Latin key in Koehne’s Atlas Lythracearum (ined.).

A. Spur obtuse, not incurved B :
B. Branches glabrous save for a minute recta white ;
bescence towards the tips; calyx glabrous........ C. laeviuscula
BB. Branches not glabrous, variously paneecet below as
well as above
C. Branches usua lly more 4 Ae hispid above with reddish-
rown hairs; calyx 7—12(-13) mm. long, glabrous or
ften sparingly his ae Oviulee Beko cen C. ingrata.
CC. Branches not reddish-brown-hispid; aresbay more or less
id; calyx never glabrous, 6-9 mm. long; leaves
small, 4-17 (or rarely —24) mm. Hecftoaeoal vules
D. Stems, branches and the surfaces of at least the
ounger leaves aniforint and densely glandular-
tellous and pubescent; leaves commonly attenu- ‘
ate at the meee uy Gt pak SER aE RLS ek OR eM Se C. glutinosa.

‘hirte
pubescant on their
E. Leaves finely but ioe aa
6-17 mm. long, 2-8 mm. wide.
F. Stems and bra hide minutely puberulent;
eaves 6-13 x 2-5 mm., abruptly narrowed or

somewhat rounded at the base, glab ; calyx
9 long, covered with short gland-tipped
, glabrous below the stamens the

stamens within; the two deen petals hardly 4
wider than the others; seeds 2 mm. long af ae
slightly wider; — Be ess C. acinifolia.
EE. Leaves l-nerved or the lateral v eins quite incon-
spicuous, at least the lower leaves oo obeae at the
base G.
G. Leaves commonly crowded, — heegy the
in tragic 5-15 x 2-7 rome the two dorsal
petals more than 4 wi er than the others, i
rotaned-ovete? ovules 6-8 ee eee i

14 BACIGALUPI

GG. Leaves more remote, the longest 14 mm. long,

oblong-ovate to linear; the two dorsal petals

hardly, if at all, wider than the others, widely
cuneate-oblong; ovules 4-12.................. C. thymoides.

AA. Spur quite slender, more or less incurved; calyx 10-13 mm.

H. Ovules 21-32; seeds orbicular or suborbicular; spur de-
cidedly incurved; leaves 10-35 mm. long; plants not at all
or only ve ightly viscid J.
J. Branches pilose with light hairs 2-3 mm. long, especially
eaves commonly lanceolate, bright green,
acutish to obtusish at the base; calyx 13 mm. long;
Pp EROS Gs C. grata.
JJ. Branches not pilose with long hairs; leaves commonly
oblanceolate, more or less hoary, cuneate at the base;
calyx 10-12 mm. long; petals 6-7 mm. long (perhaps
OER a eises cs ove see a C. campylocenira.
HH. Ovules 10-12; seeds quite widely ellipsoid; spur onl
slightly incurved; leaves cuneate at the base, the lower

1 MO a . corispermd,
HHH. Ovules 6; leaves rounded at the base, commonly 8—
eo, i ee eee C. hexasperma.

Cuphea obliqua, spec. nov. Suffrutex verisimiliter alt. ignota.
Caules irregulariter ramosi inferne decorticati superne pilis atro-
purpureis glanduloso-hispidi simul sat dense pilis articulatis pubes-
centes apud specimen visum 50 cm. longi. Petioli 0.5-2 mm. longi,
ut caules vestiti. Folia obliqua lanceolata vel oblongo-lanceolata

breviores setulis 1-2 munitae. Petala lilacina, 2 dorsualia oblongo-
oblanceolata, 3.5 mm. longa, 4 ventralia cuneato-oblanceolata, ®
mm. longa. Ex filamentis 9 ventralibus exteriora duo lobos calyes
superantia, interiora 7 quam lobi breviora, 3 brevissima, A
alternantia exteriorum 24 aequantia. Ovarium glabrum sty lo

ARGENTINA: Mistonzs: regi bout Posad, long the Alto-Paran
River, near the ranch known as "La Granja,” Nov. 18, 1907, B. . Hkm™
no. 1914 (type, in Herb. Mus. Botan. Stockholm).

.

STUDIES IN CUPHEA 15

This species is one of the subsection Hyssopocuphea (§ Kuandra).
It differs from C’. rubescens Koehne, its closest relative, in its glandular-
hispidulous and somewhat longer calyx, its slightly ‘illose style,
in having the leaves distinctly petioled by as much as 2 mm. in the
larger leaves, in the articulate pubescence of the younger stems and
in its much wider leaves strigulose beneath, particularly on the
nerves. The leaves of C. rubescens range from lanceolate to sub-
linear (those of C. obliqua, from lanceolate to oblong-lanceolate) and
are often so oblique at the base as to make the leaf appear falcate.

€ specific name has reference, of course, to these oblique leaves.

Cuphea paranensis, spec. nov. Suffrutex (an frutex?) viscidus
insuper adscendenter denseque ramosus alt. ignota. Rami inferne
decorticati dense glanduloso-hispiduli insuper simul sparse crispo-
puberuli folia pauca sessilia rotundato-ovata uninervia (interdum 3
nervis ex basi) ut ea ramulorum vestita 12 x 8-9 mm. in quorum
axillis ramuli gerentes, 30-40 cm. longi. Ramuli virgati ut rami
juniores vestiti 5-13 cm. longi. Folia conferta sessilia decussata
subcoriacea aut cordato-ovata internodiis paullo breviora vel ea
aequantia, 5 mm. longa, 4 mm. lata, aut elliptica-oblonga inter-
nodiis longiora, 3-4 mm. longa, 1.5-2 mm. lata, omnia adscendentia
caulibus adpressa eos obtectentia uninervia conspicue glanduloso-
ciliata utrinque atroviridia sparsius pilis adpressis glanduloso-hispidula.
Inflorescentiae compositae foliatae confertae 3-10 cm. longae. Pedi-
celli interpetiolares vel interdum paullo infrapetiolares ut caules
Vestiti prope apicem bracteolas ovato-cordatas basi late sessiles
serentes. Calyx subcoriaceus stramineus dorso lateribusque purpureo
suffultus dense uniformiterque glanduloso-hirsutulus 9-10 mm.
engus, calcare rotundato obtuso, fauce valde adscendente, lobis
brevibus latisque, intus infra stamina ventralia biseriatim villosius-
culus ceterum hine glaber, fauce intus atropurpureus sub staminibus
ventralibus ac infra petala 2 dorsualia lanatus. Vesiculae infrasta-
minales nullae. Petala rosea vena media purpurea, 2 dorsualia
cuneato-oblanceolata 4.5 mm. longa, 4 ventralia lineari-oblanceolata
5 mm. longa. Filamenta ventralia longiora lobos calycis paullo
“Perantia. Ovarium sub anthesi breve, glabrum, 2 mm. longum,
stylo villosiusculo 3 mm. longo calycem demum 2 mm. superante.

«85-10. Semina atropurpurea nigrescentia ecarunculata margine
shiny Discus latus horizontalis.—Plate III.
RAZIL: . “ss i,” Deo: 7,
1915, P. Dunne LisB aed vt, Hen has Botan, Stockholm)

The obtuse margins of the seeds of this species are those char-
acteristic of the subsection Hilariella (§ Euandra). The rounded

tuse spur of the calyx places it in series / of this subsection.

16 BACIGALUPI

This novelty seems to be not very closely related to any hitherto
described species in this subsection, though it is closer to C. acinos
St. Hil. of Minas Geraes and S. Paulo than to any of the others.
It may easily be distinguished from any species of the section Ht-
lariella, series 1, by its larger number of ovules (no other species in
this series has more than 6), and by its crowded, glandular-ciliate,
sessile, dimorphic (larger cordate-ovate and smaller elliptic-oblong)
leaves thickly clothing the branchlets.

Cuphea tuberosiformis Koehne in MS., spec. nov. “Suffrutex
vel herba perennis” rhizomate suborbiculari crasso tuberoso 2-3 em.
diam. (Ekman, no. 1928); ‘“‘caules 14-32 cm. longi, tenues, simplices,
internodia foliis pro parte breviora pro parte longiora, pubescentia
ac breviter violaceo-hirtella. Folia petiolis 1-4 mm. longis insidentia,
e basi obtusa vel late acuta ad oblonga, vel infima orbicularia, 12-50
mm. longa, et 8-25 mm. lata, acuta, supra albido-punctulata vel
strigosa setulisque saepe conspersa, subtus scabra vel scabro-hirtella,
nervis utrinsecus 6 supra interdum impressis subtus prominentibus.
Racemi distincti, bracteis infimis tantum ad 20-25 mm. longis
ceteris a 5 ad 1.5 mm. longis, decurrentibus, -+ compositi ramulis
saepe pedicellum axillarem juxta pedicellum interpetiolarem insertum,
ad 17 mm. longum, foliis 1-3 minimis, 1-1.5 mm. longis obsitum
imitantibus; pedicelli 2-12 mm. longi ut caules vestiti prope apicem

racteolas minimas virides rotundatas gerentes. Calyx calcare brevi
obtuso adjecto” (10-) “13-14 mm. longus, puberulus ac breviter
violaceo-setulosus viscosus, intus supra stamina ventre lanatus
infra eadem villosus; lobi aequilongi. Petala 6, purpureo-lilacina,
7 mm. longa, dorsalia 2 cuneato-obovata 5 mm., ventralia cuneato-
oblonga 3 mm. lata. Stamina episepala lobos aequantia vel vix
superantia, 2 lanata, 3 glabra, episepala” [epipetala] “tubum V¥
aequantia lanata, 2 dorsalibus brevibus glabris exceptis; vesiculae
infrastaminales 0. Pistillum ca. 9-10 mm. longum, ovarium glabrum,
stylus ovario aequilongus parce pilosus vel subglaber, demum brev-
iter exsertus; discus deflexus, supra concavus, subtus + trilobus.
Ovula 8 vel 9. Semina orbicularia, ca. 2.2 mm. longa, margine
obtusa.”—Plate IV

: IL: Parand: Tamandud, “in campo,” Feb. 24, 1910, Dusét
bane meng (type, in Herb. Mus. Bot, Stockholm); Capéo bes in
, OV. 41, yf. én, no. 7281; Turma N. 23, “in campo, ¢
1910, P. Dusén, no. 9142; same loc., “in campo graminoso,” Oct. 24, 1914 A
onsson, no. 1201a; Ponta Grossa, “in campo,” Jan. 10, 1915, P. Duse™
PARAGUAY: “in viciniis C \ ele
i aaguazt,” Feb. 1905, Hassler, no
Siena Misiones: Bonpland, vicinity of Posadas, Dee. 2% 190%,

The striking tuberous underground portions of this species place

STUDIES IN CUPHEA 17

it at once with the others of the subsection Oidemation (§ Euandra).
Its dorsal stamens have normal pollen-bearing anthers, a character
confining the species to series 1 of this subsection. It may be dis-
tinguished from its closest relatives, C. tuberosa Cham. & Schlecht.
and C. confertiflora St. Hil. by its larger ovate-oblong to orbicular-
ovate leaves and by having its flowers borne in racemosely branched
cymes. Its petioles are much shorter than those of C. tuberosa,
while its calyx differs from that of C. confertiflora in not having
longer, though sparser, hispidulous hairs in addition to its uniform
glandular puberulence.

Cuphea myrtifolia, spec. nov. Suffrutex verisimiliter. Rami:
basi lignescentes ex rhizomate cir. 1 cm. diam. vere lignoso inferne
pilis rufis hirsuti hine insuperque canescenti-strigulosi ac pruinoso-
glandulosi_ fusco-ferruginei insuper pallidiores apud exempla visa
simplices, 15-22 cm. alti. Petioli 0-1 mm. longi ut caulis vestiti.
Folia adscendentia anguste oblongo-elliptica ad basin versus acute
vel obtuse angustata sursum angustata vel acuminata adscendentia

sIKiA ; bri

*. * . . Fa | a
sat conferta internodiis longiora utrinqu strigulosa 1

Praesertim subtus aeneo-viridia nervo medio tantum hinc prominente
ae cinnamomeo margine plerumque sparse ciliata nervis utrinsecus
Paginae superioris 6-7 adscendentibus ad marginem versus arcuatis
confluentibusque medio sulcato omnibus singulari modo incrassato-
&xstantibus, 12-35 mm. longa, 6-14 mm. lata. Pedicelli interpetio-
ares ut caulis dense albido-strigulosi apice bracteolas sat magnas
ut pedicellus vestitas gerentes, 1-3 mm. longi. Inflorescentia vere
foliata calycibus partim foliis obtectis. Calyx fuscus ad os versus
et interdum dorso atropurpureus subcoriaceus prominente obtuseque
“ostatus ut caulis uniformiter albido-strigulosus interdum dorso
lateribusque pilis bruneis vel purpureis sparsis 1.5-2 mm. longis
obsitus calcare rotundato appendicibus quam lobi calycis dimidio
longioribus setis 2 duplo longioribus auctis, 6-9 mm. longus, Intus

cuneato-oblonga 3 mm. longa. Vesiculae infrastaminales 0.
Ovarium glabrum stylo demum calycem 0.5-1 mm. superante. Ovula
f Semina turgida exalata. Discus horizontalis latus super in-
‘heque convexus.

FRAZIL: Marto Gros May 12, 1903, Gustav
. so: Santa Anna da Chapada, May 12, .
Mae (Tvpx, in Herb. Mus. Bot. Stockholm); same locality, Oct. 13, 1902,

This novelty belongs to the subsection Oidemation (§ Euandra).

18 BACIGALUPI

It is most closely related to C. ferruginea Koehne, having among

other characters in common with that species its dark reddish pubes-

cence. It differs in having strictly decussate, sparsely long-ciliate
and not hirsute leaves, much shorter pedicels, in its much more
sparsely hirsute calyx, in having the two dorsal filament-bearing
ribs of the inside of the calyx very densely instead of only slightly
pilose, and more especially in the peculiar character of the leaf vena-
tion. On the lower surface, the lateral veins are hardly visible,
while on the upper, they are very conspicuously raised, none ending
freely at the edge of the leaf but running into the similarly raised
marginal nerve. The wider dorsal midrib appears to be a double
one, owing to the longitudinal groove which runs its entire length.

In Das Pflanzenreich IV, Fam. 216, only one species (C. Urbaniana)
appears in series 3 of the subsection Platypterus (§ Euandra). Since
that publication, Koehne described three additional species having
the carunculate seeds characteristic of this series. ‘The distinctions
between these and C. Urbaniana are set down in the following key,
largely adapted from Koehne’s manuscript.

Caruncle not appearing as an apical protuberance (merely as a
raised area on the seed surface). Ovules 4. Stem not
hirsute. Leaves sparingly setose-ciliate. Uniformly stri-
gulose calyx sparsely long-setose, biseriately villose below ,

me ee Wi ee eee Fiebrigt
Caruncle appearing as an apical protuberance. Ovules 8-12.
Calyx strigose, often with a few setose hairs, or even quite
dense rsute, as well.

Calyces and stems often violet-hirsute. Seeds suborbicular.

Calyx within glabrous below the stamens..........-- 7. Urbanvana

Plant not all hirsute. Seeds obovate-ellipsoid. Calyx .

within biseriately villose below the stamens.......-.--- C, Dusen.

Calyx very densely viscose-hirtellous and, in addition, hirsute
with hairs onger than those forming the more or less

uniform covering. s obovate-orbicular. Calyx
within biseriately villose below the stamens.........- C. carunculata.
§ 7. TrisPERMUM.

Cuphea callosa, spec. nov. Herba annua depressa basi ee
tulum lignea. Caulis simplex vel ramosus, incano-strigosus Si?

STUDIES IN CUPHEA 19

costis duabus dorsualibus infra petala dorsualia pallide purpureis
dilatatis claviformibus appendicibus inconspicuis breviter hirsutulis.
Vesiculae infrastaminales nullae. Petala subaequalia ovato-oblonga
caeruleo-purpurea calyci dimidio (cire.) aequilonga. Filamenta
ventralia longiora tubum fere aequantia omnia pilis crispis pubes-
centia duo dorsualia multo breviora inferius inserta glabra. Ovarium
glabrum. Stylus ovarium aequans glaber calice multo brevior.
Diseus. valde incurvato-deflexus. Ovula 3-4.

COLOMBIA: Depr. Sanranper: marshy soil on the northern slope of
the Mesa de los Santos, alt. 1000— meters, eastern Cordillera, Dec. 11-15,
1926, E. P. Killip & Albert C. Smith, no. 15032 (rTypr, in Gray Herb.). One
of the more robust specimens collected under this number appeared to have
been badly cropped.

This species belongs to the section Trispermum. Its closest relative
is C. impexa Koehne from Piauhy, Brazil. It may be distinguished
from this species by the presence of a petiole, however short, by its
less densely and more shortly hirsute calyx, by having the external
appendages which alternate with calyx-lobes only inconspicuously
and shortly setulose, by the two purplish dorsal ribs of the calyx
which are widened and thickened, ending as free callosities just
below the insertion of the dorsal petals (whence the name), by its
glabrous style, and by it less densely, though still conspicuously
hirsute, subrevolute and strigose-ciliolate leaves.

§ 8. PsEuDOCIRCAEA.

In his key to the species of the § Pseudocircaea (Pflanzenreich IV,
Fam. 216, p. 95. 1903), Koehne divided the species into two main
sroups, basing this separation on the presence or absence of hairs
on the thickened midrib of the lower side of the persistent petals.
Later, he found this condition far more general in this section than
he had at first supposed. From Fed. Rep. viii. (1910) 199, Koehne
may be quoted as follows: “Die Blumenblitter scheinen stets, was
“onst bei Cuphea nicht vorkommt, unterseits an der Mittelrippe
Zottig zu sein, eine Erscheinung, die ich anfangs bei einer Anzah
von Arten, von denen mir nur geringfiigige Proben vorlagen, iiberge-
sehen habe.” He thereupon revised his key to the section, and
included therein two species described since the general revision, C.

wy wrensis and C. Rojasi, both of Koehne. This revised key was
still M Manuscript at the time of Koehne’s death in 1918, nor has it
Sinee been published. It is here reproduced verbatim from Koehne’s
Manuscript,

20 BACIGALUPI

Sect. 8. PseupocrrcaBA.

A. Petioli 1-3 mm. longi. Folia nulla cordata, 5-33 mm. longa.

ere semierectus, itpalag sursum adpressus. Folia 12-32

m. longa, flor vis quovis pari maxime inaequalia.

Petala angusta v Pages

a. Folia basi att aiitiate rOhisegoianceolate: Ovula 5. .117. C. Chodatiana.

8. Folia basi Fler iat vel vix acuta rotundato-ovata vel
to-oblonga. Ovula 5-10............ Be C. ovalifolia.

b. Discus Seteontane vel subhorizontalis. Folia 5-33
Li cl Petala dorsalia plerumque veniestiis rianifeate

latior.
a. Ovula 5-13. Calyx 8-15 mm. lon
I. Calyx 8-10 m mm. ees glanduloso-hir-

TT. Ovula 9-13. Calyx 10-15 mm. longus, ities
hirsutus. Folia basi attenuata late lanceolata a
a “igen 18-33 mm. longa, floralia valde inae-

Gee ee gk bs ok ve ok does 120. C. persistens.
B. Ovula 3 3-5. Calyx 7-7.5 mm. pet pel Sictonds Se
loa basi rotundata, ovata vel o ato-oblonga, 5-25

m. longa, floralia TD idacnehe. Pedicelli Vix 1 mm

a. Annua. Ovula 3. Petioli 10-20 mm. longi; folia ovata
Mes ovato-oblonga, 10-70 mm. longa. Calyx 5-7.5
m. lon 122. C. prunellifolia.

; . Ca :
1. Ovula 5. Petioli 8-12 ating longi; folia only. :
vaya 25-45 m sag me sy inaequalia rok
~7 breviore nec lineari nec longe ciliato. Calyx
WA ORR cscs ‘i os ek 122b. C. talaverensis.
2. Ovula 6-8. * Petioli 3-9 mm. longi; folia oblongo-
epee age mm. longa, floralia eer IEY
u altero 3 mm. longo lineari longe ci Reo f
C ahve 8mm. i aera po eas - 123 ay inp

"vel la ne eolata-linearia, 20-60 mm. lon mii oral ae
ualia. Calyx 6-7 rine 1M a ~~ parietarioides

a. Folia i rotundata oblonga vel oblongo-lanceolata

apice obtusiuscula, 25-55 mm. lon, mn Calyx Aes Bae videscens
ongus. Ovula 3-8. Di io Hathcabs Oa
B. Folia multa basi subcordata vi 1 cordata Sypnees acu
T. Ovula 3 Fe ovato-oblonga

vel ovata, 20-43 m: en a nervis oS meat bcps se 6-12;
o erio £: 10 mm Rojas.
ate gba) ec iene ue : her ide Sat 125b. C.
1. Ovula 5(-7). Discus horizontalis. Folia rugosa late
— vel 2 ie 20-45 mm. lon, —s
trinsecus 11-14: Ml costata.
aa. petioli 2-5 mm. longi. alyx 126. C. :

. see
Te ee OE Se Se A Nik wee eee eR ee eee eee

STUDIES IN CUPHEA 21

§ 9. HeTERopon.

CupHea Wricuti Gray. This species has a geographical range
extending from Costa Rica northward to the mountains of .south-
eastern Arizona. In the extreme northern portion of its range,
growing with the typical form, is found one with uniformly much
reduced linear petals less than 1 mm. in length. Owing to the pres-
ence of very short slightly crisped hairs, the petals, though actually
pale violet, appear grayish in hue. This plant may be designated as
Cuphea Wrightit var. nematopetala. The distinctions may be thus
set forth :—

Var. a. typica, var. nov. Petala violacea glabra 2 dorsalia obovata
vel fere rotundata, 2-3 mm. longa, 4 ventralia 1.5-2 mm. longa,
cuneato-obovata.

Var. 8. nematopetala, var. nov. Petala pallide violacea sed
propter pilos brevissimos primo aspectu cineracea omnia linearia,
0.5-0.8 mm. longa.

ARIZONA: Bowie, Sept. 21, 1884, Marcus E. Jones (Typx, in Gray Herb.);
Bowie, Sept. 10 (18?), 1884, M. E. Jones, no. 35; Ramsey Cafion, Huachuca
Mountains, Sept. 28, 1929, M. E. J ones, no. 24989, at least in part, distributed
as C. Palmeri Wats.

Related to Cuphea lophostoma Koehne are C. cuernavacana Rose
and C, Lozani Rose, both described since the general treatment in
Das Pflanzenreich. The distinctions among these three species were
outlined in key form by Koehne in manuscript. A translation from
this Latin key follows :—

Pedicels bearing normal bractlets; dorsal calyx-lobe glabrous

within, the edge not turned back like a hood (not eeaieiaist-&:

Calyx densely long-hi ons aca:
S-NIYSULS 6 es se ee es ee

payalyx uch more shortly viscous-hirtellous.......-.----- C. lophostoma

edicels without normal bractlets, these replaced by a row of

albido-pubescens simul conspicue longe purpureo-hirsutus ad basem
ae retrorso-hispidus, 25-60 cm. altus. Petioli subnulli-6 mm.
longi. F olia anguste lanceolata basi saepe cuneata in petiolos de-

entia utrinque scabra strigosaque, 30-70 mm. longa, 9-15 mm.
lata, floralia linearia setis purpureis ciliata. Inflorescentiae confertae

um aperte paniculatae. Pedicelli 1-2 mm. longi prope apicem
bracteolati. Bracteolae ovatae longe ciliatae pilis brevibus albidis

22 BACIGALUPI

basi bulbosis in axillis instructae. Calyx basi stramineus dorso
purpureus ad os versus viridis dense breviterque albido-strigosus
simul dense longeque purpureo-hirsutus, lobo dorsuali producto,
intus sub staminibus ventralibus crispo-pilosus infra stamina duo
dorsualia biseriatim villosus, 10-12 mm. longus, calli duo infra
petala dorsualia conspicui; appendices calycis minutae setosae, duae
dorsuales multo majores setas 3 vel 4 quam lobus dorsualis longiores
gerentes. Petala duo dorsualia cyanea lamina suborbiculari basi
cuneata, 5-6 mm. longa, unguiculo tenuissimo 3-4 mm. longo, 4
ventralia lutea oblongo-oblanceolata breviter lateque unguiculata,
3-4 mm. longa. Filamenta ventralia 9 subexserta inaequalia, eX-
timis duobus barbatis, aliorum 4 leviter crispo-pilosis, 3 cum els
alternantibus omnino glabris. Vesiculae infrastaminales 8. Stylus
glaber apice bifidus demum os calycis 2.5 mm. superans. Discus
brevis crassus deflexus. Ovula 5.—Plate V.

CO: Guerrero: hills near Iguala, alt. 3500 feet, July 30, 1907,

MEXI
Pringle, no. 13942 (rypx, in Gray Herb.); same loc., Aug., 1905, J. N. Rose,
nos. 9304 and 9304a.

This species falls into Koehne’s section Heterodon (subsection
Lophostomum, series 1). Its closest relative is C. viscosa Rose, from

with C. lutea Rose, a species considered by Koehne as closely
to and intermediate between (. Wrightti and C. viscosa.
manner, ©. tenwipes may be considered intermediate between
wscosa and C. lutea.
The specific name of this novelty has reference to the long slender
1 petals

¢.

STUDIES IN CUPHEA 23

formis bracteata. Pedicelli vix graciles parte superiore bracteolati,
5-6 mm. longi. Bracteolae apice rotundatae ciliatae. Calyx crassus
fulvus insuper basique purpurascens viridi-costatus rotundato-
calearatus glanduloso-hirsutulus circ. 16 mm. longus lobo dorsuali
leviter producto intus sub staminibus ventralibus pilosiusculus infra
stamina glaber. Petala 6, duo dorsualia retrorso-erecta ovato-elliptica
5mm. longa eorundem unguiculi vix tenui 1.5 mm. longi, 4 ventralia
oblongo-spathulata sessilia 3-3.5 mm. longa. Filamenta duo dor-
sualia breviora glabra, 9 ventralia irregulariter incrassata floccoso-
lanata praecipue exteriora quorum tria subexserta. Ovarium glabrum
stylo glabro incluso. Ovula 5-7. Discus crassus apice deflexo.

MEXICO: Trpic Trerrrrory (now Nayarit): between Santa Gertrudis
and Santa Teresa, August, 1897, J. N. Rose, no. 2066 (Typx, in Gray Herb.).

This species belongs to the subsection Lophostomum. Its closest
relative is C. Karwinskii Koehne from Oaxaca, from which it may
be distinguished by its quaternate leaves and its much thicker and
glandular-hirsutulous rather than strigose-puberulent calyx.

Cuphea purpurascens, spec. nov. Herba annua radice ies

a -

simul glanduloso-hirsutulus pilis purpureis, 15-35 em. alt s. + ioli
foliorum inferiorum ut caulis vestiti, 2.5 mm. longi, ei/\foliorum

superiorum nulli vel subnulli. Folia oblanceolata vel linedri-oblanceo-

lata supra strigulosa interdum sparse adpresso-hirsutula_subtus,;item A

MEXICO: Pursia: Teh Sept., 1911, C. A. Purpus, no 5711, dis-
fated as C. Palmeri Wate. Gira, Te Herb.); at same locality, Sept.
» 1905, J. N. Rose, no. 10136.

24 BACIGALUPI

This species belongs in the subsection Lophostomum (§ Heterodon).
In Koehne’s key (Pflanzenreich IV, Fam. 216, p. 95), it would be
placed next to C. secundiflora Sesse & Moc. Related to both these
species is C. humayana Brandegee (Zoe V. 214. 1905) from Sinaloa.
The distinctions among them may be set forth in the following
amendment to Koehne’s key (loc. cit.).

B. 2 dorsal petals larger than the lower or sometimes the lower
4 lacking.

a. Dise obliquely erect.

I. Claws of the 2 dorsal petals never more than 1 the length
vules 3-6.

of the blade. O
I. Calyx 8-12 mm. long, 1-114 mm. wide, in fruit

owly amp ous, greenish, pilose, slightly
2-ribbed within, the throat slender, the mouth strik-
ingly oblique; ventr petal c

lacking; apes as long as the dorsal (longer) side of
ispidulous, especially below; leaves ovate
\y ong; petioles of
lower leaves often 10-35 mm. long). fe eee C. secundiflora.
II. Calyx 5-7 mm. long, 114-214 mm. wide, in fruit

ventral petals present, 114 mm. long; style shor
than the ventral (shorter) side of the calyx, glabrate;
leaves oblong-lanceolate to narrowly oblanceolate,
to 244 em. long; petioles mostly none to 5 mm. in
OL eae a Ee UNOS ANG Steal ea arate Mint Serge C. purpurascens.

len,

I. Claws of the 2 dorsal petals 14 as long as to equaling
the blade. Ovules 3. alyx 8-10 mm. long, in fruit
Sey iia ng Tanree lanco-oress

CUPHEA HUMAYANA Brandegee. This species was originally
described from Culiacan, Sinaloa. Subsequent collections have more
definitely established its range. It is now known to occur in Morelos
and Nayarit as well. The following specimens may be cited:—

Yuatepec, Morelos, Rose ct al.,no. 8537 (distributed under a manuscript

name by Rose); same loc., Rose & Painter, no. 6581 (distributed ‘a

C. lanceolata Ait.); foothills back of Jalisco, Nayarit, R. S. Ferns

no. 5838.

§ 10. Metviina.

oblongo-elliptica plerumque ad basem versus subcuneata apice -_

vel plerumque breviter attenuata supra atroviridia seabra parce

STUDIES IN CUPHEA 25

hispida subtus pallidiora hispidula praesertim in nervis prominentibus
et in margine scaberrima, 20-40 mm. longa, 13-26 mm. lata. In-
florescentia foliata. Pedicelli interpetiolares ut caulis vestiti, 5-8 mm.
ongi, prope apicem bracteolati. Bracteolae ovato-lanceolatae his-
pido-ciliatae. Calyx crassus cylindraceus, 20-40 mm. longus, 4-5
mm. latus, sulfureus dorso convexus hine et ad os versus coccineus
valde rotundato-calcaratus dense breviterque albido-hirtellus dorso
lateribusque coccineo-hirsutus intus sub staminibus 9 ventralibus

46 aristulas lobos paene aequantes gerentibus. Petala 6 sessilia
pallide sanguinea duo dorsalia ovata basi subcordata 4 ventralia
elliptico-oblonga omnia 2-2.5 mm. longa lobos calycis 0.5 mm.
superantia. Filamenta duo dorsalia glabra, ex ventralibus medium
glabrum aliis conspicue longius, ea utroque adjacentia et paria duo
exteriora villosa lobos calycis bene vel vix aequantia cetera duo
glabra longitudine intermedia. Ovarium glabrum, 5 mm. longum,
stylo glabro, 13 mm. longo, filamento longissimo 1 mm. breviore.
Discus latus brevis deflexus. Ovula 10 valde convexo-concava
apice late retusa.

BRAZIL: Cuars: Alta da Serra in “caapuerdio,” March 16, 1910, Albert
Léfgren, no. 256 (rypx, in herb. Mus. Botan. Stockholm); San Felix, in

carrascal,”” March 18, 1910, Albert Léfgren, no. 332.

This species fits into Koehne’s series 2 of the subsection Pachycalyx
(} Melgills). . The nine ventral stamens, as usual, are alternately
long and short, but the middle of these is conspicuously longer than
any of the others, exceeding the long exserted style by 1 mm. This
condition is very rare in Cuphea. In no other species of the section
Melvilla hitherto described is any one of the ventral stamens longer
than the others, a character at once separating C. Loefgrenti from the
other Brazilian species of series 2 of the subsection Pachycalyz.
Moreover, C. Loefgrenii differs from these (C. grandiflora, C. Gardner
and (. annulata) in the character and uniform distribution of the
Pubescence of the interior of the calyx, in having at least the two
dorsal petals slightly cordate at the base and in the number of ovules.

Cuphea Mexiae, spec. nov. Suffruticulus erectus rhizomate ]2—
16 mm. incrassato. Caulis simplex cortice chartaceo cinnamomeo
ad 1.5 m. altus praesertim insuper dense glanduloso-pilosus. Petioli
subnulli ad 1 mm. longi in axillis pilos minutos incrassatos purpureos
‘tpulares gerentes. Folia decussata oblongo-elliptica utroque sensim
‘ngustata apice leviter acuminata basi abrupte rotundata supra
Stigulosa ac sat dense hispidula subtus densissime glanduloso-pilosa

26 JOHNSTON

nervo medio venisque stramineis prominentibus nervis utrinsecus
8-11 margine strigulosa interdum revoluta, 30-55 mm. longa, 5-15
mm. lata. Folia juniora floralia pilis longis intense coccineis obsita.
Inflorescentia terminalis. Pedicelli ad 7 mm. longi prope apicem
bracteolati. Calyx coccineus sat dense pilis coccineis glanduliferis
pilosus simul sparsius pilis albidis hirsutulus dorso convexus, 26 mm.
longus, ad 4 mm. latus, intus fauce sub staminibus pilis mollibus
pilosus infra stamina glaberrimus, caleare brevissimo 1 mm. longo
haud orbiculare; lobi calycis brevissimi leviter acuminati minutim
hirsuto-ciliolati; appendices duae dorsuales lobis longiores crassius-
culae setis validis coccineis basi bulbosis margine vel interdum dorso
obsitae, ceterae minores lobis subbreviores setis similibus ciliatae.
Petala ignota an nulla (?). Stamina 11 ventralia pilosa alterne
inaequalia quorum 3 mediana ceteris longiora inclusa. Ovarium
gibbum glaberrimum styli glabri demum subexserti tertiam partem
aequans. Ovula 6. Discus dorsualis brevis crasse ovato-cordatus
leviter deflexus.

XICO: Jarisco: East of San Sebastian, Hacienda del Ototal, Arroyo
de los Hornos, Sierra Madre, alt. 1500 m., woods, on slopes, abundant, Mareh
6, 1927, Ynez Mexia, no. 1820 (typ, in Gray Herb.).

This novelty finds its closest affinity in Cuphca Watsoniana Koehne.
It may be distinguished from the latter by its copious glandular
pubescence, very short petioles, mostly narrower leaves densely
glandular-pilose beneath, strictly terminal inflorescence, uniformly
scarlet calyx (that of C. Watsoniana having a green throat), much
more striking appendages, included pilose stamens and by the dis
parity in the number of ovules (C. Watsoniana has 20-26 ovules).

Upon examination, Pringle, no. 2792, distributed as C. Laebmanne
Koehne ?, proves conspecific with C. Watsoniana Koehne. Bo
this and the type (same collector’s no. 4355) are from the same
general region.

Il. THE VASCULAR FLORA OF THE
GUANO ISLANDS OF PERU.

By Ivan M. Jounsron.
(Plates VI-V II.)
_ WERE it not for the myriads of sea-fowls that frequent them the
islands off the coast of Peru would be known only as local eit

and of concern only to fishermen and navigators. They are fod
bits of land. A few of the several score of them have a length of

FLORA OF THE GUANO ISLANDS 27

five to ten kilometers, but the great majority are very much smaller.
Practically all are found strewn close in along the Peruvian coast
ten kilometers or less off shore. The most isolated one is less than
sixty kilometers from the mainland. In themselves they are for-
bidding and merit only passing attention of Man. Under the rainless
sky which prevails over them, however, the excreta of their enormous
bird populations have accumulated, forming vast quantities of that
valued and hence much sought-after fertilizer—guano. The islands,
consequently, have achieved a fame and importance which their
small size, coastal position and general aridity and inhospitality
alone would never have given them.

To the naturalist the Peruvian islands have been known only for
their bird populations. The truly astonishing size of these and their
great economic importance early attracted attention. Concerning
the flora and the other than avian fauna of the islands, however,
searcely anything has been recorded. The most successful attempt
to remedy this surprising dearth of information regarding the strictly
land flora and fauna of the Peruvian islands has come through the
collections and studies made on them between Sept. 1919 and Feb.
1920 by Dr. Robert Cushman Murphy, then connected with the
Brooklyn Museum and now on the staff of the American Museum
of Natural History. Although his primary objects were ornitho-
logical, Dr. Murphy, taking full advantage of excellent facilities
for visiting the important Peruvian islands, collected and observed all

le land plants and animals he could find on them. His collections,
with the exception only of the plants, have been reported upon
technically ten years ago. The results of their study may be found
summarized in Dr. Murphy’s delightful book, “The Bird Islands of
Peru” which was published (Putnam, New York) in 1925.

The collection of plants made by Dr. Murphy has, through a
Variety of circumstances, remained until now unrecor in print.
It represents the first ever made of the land flora of the guano islands.

- Murphy collected upon all the important Peruvian islands except
San Lorenzo, off Callao, and visited all the islands on which any of
the higher plants might be expected to occur. His collections, hence,
indicate rather well the distribution of such plants among the islands
and the probable extent of the flora involved. As a result we can
new say with fair confidence that the higher plants are present on
only four of the Peruvian islands, namely Lobos de Tierra, San

renzo, San Gallan and Viejas. The total flora probably consists
of about two dozen species. :

The most northern of the islands mentioned, Lobos de Tierra,

28 JOHNSTON

which has the smallest flora as to number of both species and indi-
viduals, apparently owes its pitifully small flora in some part to the
occasional visits of equatorial storms. On the remaining three
islands, however, the flora is dependent upon and directly correlated
with the clouds that bathe their rather lofty crests. These fogs,
brought about by the cooling of moisture-ladened sea-winds which
have been upwardly deflected by obstructing headlands or coastal
hills, are characteristic features of the littoral of Peru and northern
Chile. Forming at between 300 and 400 meters altitude over the
islands, and somewhat higher, at about 1000 meters, over the coas
hills on the mainland, they bathe the favored peaks and slopes in
wet clouds. On the coastal hills the localized effects of the protection
they give from the sun and the moisture brought by them to the
soil results in a rather sharply defined zone of lush green vegetation
on otherwise barren seaward slopes. This winter and spring fog-
correlated vegetation, so conspicuous on the coastal hills of Peru and
northern Chile, has been called the Loma Formation. It is one of
the most distinctive plant associations in South America. Belonging
with it, both ecologically and phytogeographically, are the floras of
San Lorenzo, San Gallan and Viejas.

The Peruvian islands have a flora that is merely an impoverished
fragment of that of the adjacent mainland. The flora of Lobos de
Tierra is the only one uncorrelated with fog. ‘The most common of its
two species, Sesuvium, is a sea-dispersed halophyte widely distributed
in tropical America. Its other species, represented by a single strug-
gling tree of Prosopis, may possibly be a relic of a climatically more
favorable past. Whatever may be the case, whether its presence 1s
the result of natural dispersal or the agency of man, the colonization
of the island by Prosopis is scarcely a successful one. The remarkable
thing about Lobos de Tierra is that of all the Peruvian islands it
alone has developed a flora at low altitudes, On the islands south
of Lobos de Tierra a flora is developed only on those which reach a?
altitude of 350 meters or more and which thereby achieve fog-bathed
crests. There are only three of these; San Lorenzo, San Gallan and
Viejas. The other Peruvian islands are much lower, most of them
being less than 100 meters high and only four reaching up to betwee?
100 and 200 meters in altitude. In their alignment with topographi¢
features on the mainland the islands all clearly reveal their connections
with the continent. An elevation of the Peruvian coast by Bast
score of meters would connect most of the islands with meant
Promontories and headlands. A subsidence of similar amount —
make islands out of various headlands now part of the mainlan¢.

FLORA OF THE GUANO ISLANDS 29

Under such conditions an endemic insular flora is scarcely to be ex-

ted. I am consequently of the opinion that eventual exploration
of the headlands on the mainland adjacent to San Gallan and Viejas
will reveal there the presence of the species now found on the foggy
summits of those islands. It is known that the foggy crests of San
Lorenzo have a flora composed of species found in the Loma Forma-
tion on the adjacent mainland. While the Periloba and Polyachyrus
that are described from San Gallan seem to have affinities that are
Chilean rather than Peruvian, I believe that they represent simply
further Chilean elements in the Peruvian flora belonging with the
many others that are recognizable in the coastal regions of Peru,
particularly in the southern departments. That they have been
discovered first on the islands is simply because the headlands of
Ica and northern Arequipa have not yet been botanized.

Lobos de Tierra Island, lat. 6° 26’ S., one of the largest and most
northern of the Peruvian islands, is about 10 km. in length and lies
over 16 km. from the nearest mainland. ‘The island is rough in
topography and reaches at the south end an altitude of almost 100 m.
It is barren and quite arid, its only water coming from the equatorial
storms, that through the course of the years, occasionally reach south
to it. According to Murphy, I. ec. 319, “Lobos de Tierra is unique
among all the low islands of the Peruvian coast in that it has what
may by courtesy be termed a terrestrial flora. Minute green algae
color a few of the higher rocks along the western border; and a small,
fleshy-leaved flowering plant [Sesuwvium], which the fishermen call
lejuia, grows sparingly on the beaches of the east coast. But the
outstanding, almost startling, botanical feature of the island is a
thorny, acacia-like tree [Prosopis], which stands in a sandy sink not
far from the Caleta de Cherra and within sight of the light-house.
Deriving its water presumably from ocean seepage, this anomalous
plant—a single representative of a single species—has contrived to
exe out a lengthy existence. According to the most venerable of
the Indians, its twisted trunk and bent branches, which reach barely

feet above the ground, have had their present appearance as
far back as memory and tradition go.”

San Lorenzo Island, lat. 12° 06’ S., lies only a few kilometers to
the southwest of Callao. About 8 km. long and nearly 3 km. wide
treaches an altitude of over 370 meters. Its slopes are dry and sandy,
but its crests, reaching into the same atmospheric strata as San
Gallan and Viejas, like these islands are bathed with fog. A similar
fog-correlated vegetation is developed there. The only attempt
to list the species of plant found on San Lorenzo is that by Ramondi,

30 JOHNSTON

El Peru iv. 405 (1902), whose specific determinations are probably
mostly, if not all, incorrect. San Lorenzo not being a guano island,
and Dr. Murphy not having collected there, its flora is not treated
in detail in this paper.

San Gallan Island, lat. 13° 53’ S., lies to the southwest of the port
of Pisco and about 5.5 km. west of the Paracas Peninsula. The
island is about 4 km. long. Prevailingly low at its north end, it
slopes generally upward and culminates at the south end in cliffs and
bold headlands that reach an altitude of about 400 meters, It is
the most lofty of all the Peruvian islands. Mostly arid, it is only on
and about the foggy crests of the island that its flora is developed
(Plate vi). A quotation from Dr. Murphy’s book, l. c. 187, gives some
idea of the ecological conditions and plant distribution on the island.
“We walked through the gently sloping, sandy valley, which was
filled here and there with diving petrel burrows, and then turned up
the steep, bone-dry hillside toward the rocky central summit. At
a slight elevation we encountered the first straggling, stunted growths
of the bayonet-leaved epiphyte, Tillandsia purpurea [rather 1.
latifolia!|; and, at a height of 1100 feet the desert, with its hot stones
and resounding crust of alkaline salt, abruptly gave way to a
garden such as a landscape architect might well take as an inimitable
ideal. On this peak the crumbling granite was exposed in chaoti¢
piles, and here, within a layer of cool, moist air, were fully a dozen
species of plants with buds, flowers, and fruit all together. :
of them I had already seen on the crest of Isla Vieja, in Independencia
Bay, but the others were all new. Butterflies, grasshoppers, land
snails, and lizards dwelt among the rocks and the vegetation, while
every crevice and small patch of soil had its aggregation of petrel
tunnels. To one who had been living upon the bare low guano
islands, the fresh, verdant scene and the smell of foggy air were #
delight.” Dr. Murphy collected seventeen species of ferns and
flowering plants about the summit of San Gallan.

Viejas Island, lat. 14° 14’ S., lies close in shore in line with the
Promontories forming the north and south ends of Independenci
Bay, and separated from them by more than one and less than mine

kilometers of water. It is about 6 km. long and nearly 2 km. wide,

and reaches an altitude of about 366 meters. The crest is almost
constantly bathed in fog. On this moistened area Dr. Murphy
found three kinds of higher plants, all of them belonging to speci®
present on San Gallan.

The collection of plants made by Dr. Murphy on the guano islands;
which is the basis of the present paper, was presented to the Brooklyn

FLORA OF THE GUANO ISLANDS 31

Botanic Garden and is now preserved in the herbarium of that in-
stitution. For the privilege of studying the collection I am indebted
to Dr. C. Stuart Gager, Director of the Garden and to Dr. Henry K.
Svenson, Assistant Curator in the herbarium. Dr. Murphy has most
kindly supplied for publication some of his excellent photographs
of San Gallan and its vegetation. I am also indebted to several of
my friends for assistance in the identification of critical groups.
The ferns have been named by Dr. W. R. Maxon, the grasses by
Prof. A. S. Hitchcock and the bromeliads by Dr. Lyman B. Smith.

CATALOGUE OF THE SPECIES.

Polypodium lanceolatum L. Sp. PI. ii. 1082 (1753).
San Gallan (M. 3475). A species widely distributed in the tropics. .
Trisetum spicatum (L.) Rich. Pl. Eur. i. 59 (1890).

San Gallan (M. 3481). A form of this polymorphous species. Hitch-
cock, Contr. U. S. Nat. Herb. xxiv. 360 (1927), reports it elsewhere
in Peru only from the Andean valleys of the interior. I have collected
4 similar plant in the Loma Formation in northern Chile.

Koeleria trachyantha Ph. Fl. Atac. 55 (1860).

San Gallan (M. 3477). A plant of the coastal hills which ranges
from the Lima region south into northern Chile.

Stipa disticha Hitchc. Contr. U. S. Nat. Herb. xxiv. 288 (1925).

San Gallan (M. 3478, ryeg). A very distinct species known only
from the Dr. Murphy’s collection. Its relations seem to be southern.

Tillandsia latifolia Meyen, Reise ii. 45 (1835).

San Gallan (M. 3482). The most common and conspicuous plant
on the island, growing luxuriantly over gravelly slopes and out-crops
above 300 m. alt. and even staggling down below the cloud-zone in
occasional stunted examples to about 10 m. alt. The best devel-
opments of the plants were towards the crests on lee exposures.

€ species is widely distributed along the Peruvian coast and has
been reliably reported from San Lorenzo.

Tillandsia recurvata L. Sp. Pl. ed. 2, i. 410 (1762).

Viejas (M. 3221) and San Gallan (M. 3468). On Viejas it grows
only on crumbling rocks about the summit. On San Gallan in the
cloud belt it is not uncommon, and is usually found clinging to the
lee side of rocks, According to Dr. Smith this island plant represents
& phase of the variable and very widely distributed 7. recurvata
notable chiefly for its stout habit and long, trailing stems.

Parietaria debilis Forst. Prodr. 73 (1786).

San Gallan (M. 3471). A widely distributed herb frequent in
noist, sheltered places in western Peru.

32 JOHNSTON

Suaeda foliosa Mog. in DC. Prodr. xiii. pt. 2, 156 (1849).

San Gallan (M. 3465). A spreading succulent shrub of the coastal
hills of central and southern Peru and northern Chile.

Sesuvium Portulacastrum L. Syst. ed. 10, ii. 1058 (1759).

Lobos de Tierra (M. 3464). Growing sparingly along the beach on
the east side of the island. A widely distributed halophyte in the Ameri-
can tropics.

Prosopis chilensis (Molina) Stuntz, U. S. Bur. Pl. Indust., Invent.
Seeds xxxi. 85 (1914).

Lobos de Tierra (M. 3463). Represented on the islands by a
single, stunted, wind-swept tree slightly back and above the shore.
A photograph of this individual is to be found on the plate in Dr.

. Murphy’s book opposite page 320. The species is a widely distributed
one in well drained soils throughout most of the coastal area of Peru
and Chile.

: Geranium limae Kunth, Pflanzenr. [Heft 53] iv. Fam. 129, 74
1912),

San Gallan (M. 3473, 3474). This is the only Peruvian species of
a group that is prevailingly Chilean. The type and previously only
known specimen is given as from “Lima und §, Lorenz.” The San
Gallan plant may be somewhat more densely hairy on the stems
but otherwise appears to agree closely with the type.

Oxalis carnosa Molina Sagg. Chile ed. 2, 288 (1810).

Viejas (M. 3218) and San Gallan (M. 3469). A plant widely

distributed along the coast of western South America.

Palaua moschata Cay. Monadelph. Class. Dissertat. i. 41 (1790):

San Gallan (M. 3466). Growing chiefly in sheltered hollows
between rocks about the summit. The flowers are lavender or pale
purplish. Otherwise known only from San Lorenzo Island and the
hills near Lima. -

Daucus montanus H. & B. ex Schultes, Syst. vi. 482 (1820).

San Gallan (M. 3476). Widely distributed in western South America.

Apium ammi (Jacq.) Urban in Mart. Fl. Bras. xi. pt. 1, 341 (1879).

FLORA OF THE GUANO ISLANDS 33

stipitato-glanduliferis integerrimis crassiusculis, lamina ovata vel
obovata 10-15 mm. longa 6-8 mm. lata apice rotundata basi in
petiolo 5-10 mm. longo gradatim attenuata; basi petioli indurata
persistenti; foliis capitum caudicis rosulatis; foliis caulinis rameali-
busque 1-15 mm. distantibus superioribus vix conspicue reductis;
floribus in axillis foliorum superiorum solitariis; pedicellis gracilibus
ca. 10 mm. longis maturitate deflexis; calycibus obconicis 1 cm.
longis ultra medium in lobos duos acutos incisis, fructiferis ca. 8 mm.
diametro 6-7 mm. altis, lobis incurvatis; corolla infundibuliformi
caerulescenti 2 cm. longa extus villosula et glanduligera; tubo cy-
indrico 5 mm. longo per calycem occulto intus villoso; faucibus
gradatim ampliatis ca. 10 mm. longis ad 10 mm. diametro; filamentis
inclusis 3 et 4 mm. longis ca. 5 mm. supra basem corollae affixis ad
basem versus dilatatis, partibus inferioribus et decurrentibus dense
villosis, aliter glaberrimis; antheris ellipticis ca. 1.8 mm. longis;
stylo glaberrimo 10 mm. longo; receptaculo maturitate hemisphaerico
2-2.5 mm. diametro irregulariter profundeque lacunoso; nuculis 15-
20 percongestis distinctis evidenter 2-3-seriatis basi affixis evidenter
maequalibus 1-2.5 mm. longis plus minusve angulatis polymorphis
uigris sub lente minute delicateque alveolatis plus minusve opacis.—
Peru: fog-bathed crest of San Gallan Island, Nov. 27, 1919, R. C. Mur-
phy 3472 (tyPE, Brooklyn Bot. Garden; 1sorypr, Gray Herb.)

A very distinct species of a genus heretofore unknown in the
Peruvian flora. All the congeners of P. insularis are Chilean. None
of them, however, seems closely related to it. If it were necessary to
indicate a particular affinity I should suggest P. paradoxa (Lindl.)

t. of the coast of central Chile. The island plant can be distin-
guished from all other members of it genus by its very woody caudex
and dense, forking or even dendritic pubescence. From most species
"ls further distinguished by having the corolla-tube villous within
ry in having conspicuously unequal nutlets. The calyx is not at
all plicate and seems to be prevailingly two-lobed. In the nature of

pubescence on the herbage, in its villous corolla-tube and in the
ent of its small nutlets the plant rather suggests a species of
en The elevated gynobase, the pluriseriate arrangement
the nutlets, and the broad leaves, however, are all indicative of its
it "ai in Periloba. Its fruit, clearly and unmistakably, removes
rig Nolana, although the gross habit, though not pubescence,
y Suggest the well known species of the Peruvian coastal region,

* Prostrata L. f.

San Cam tuberiferum Dunal in DC. Prodr. xiii. pt. 1, 63 (1852).
an Gallan (M. 3470). The specimen is poor but represents,

34 JOHNSTON
almost certainly, this species. It is a plant of the coastal hills of

eru.

Solanum Murphyi Johnston, sp. nov., succulentum glaberrimum
erectum; radice crassa palariramosa ut videtur annua; caulibus 8-12
mm. crassis purpureo-tinctis breviter subdichotome ramosis 15-30
cm. altis flexuosis ad basem versus aliquanto induratis ceterum
herbaceis fistulosis; ramis 5-12 cm. longis ascendentibus saepissime
simplicibus foliosis 4-8 mm. crassis confertis vel distantibus; foliis
succosis 1,5-2.5 em. longis 5-8 mm. latis 5—7-lobatis viridis, margine
evidenter revolutis, supra minute papillatis, subtus leavibus cum
costa lata conspicua ornatis; lamina ovata vel ovato-oblonga 7-13
mm. longa in petiolum alatum abrupte contracta, lobis 2-3-jugatis
1-2.5 mm. longis et latis obtusis sinubus saepissime apertis obtusis;
inflorescentia cymosa 1-2-flora in ramis saepe terminali; pedicellis
gracilibus sulcatis 5-9 mm. longis erectis, fructiferis deflexis; calyee

4.5 mm. longo, lobis 5 erectis oblongis 3-3.5 mm. longis
1-1.5 mm. latis, apice obtusis, extus glaberrimis minute papillatis,
intus stipitato-glanduliferis; corolla purpureo-caerulea cupuliformi
vel rotato-infundibuliformi 1.5-2 em. diametro ca. 12 mm. longa
quam calyce 3-4-plo longiori extus villosula intus glabra, lobis latis
rotundis ca. 8 mm. latis ca. 1.5 mm. longis apiculatis, sinubus lob
rotundis 2 mm. latis ca. 0.9 mm. altis ornatis; filamentis aequalibus
ca. 1.5 mm. supra basem corollae affixis subulatis glaberrimis 4 mm.
longis immam ad basem ca. 0.8 mm. latis; antheris distinctis oblongis
aequalibus 2-2.5 mm. longis biporosis; stylo 5.5 mm. longo gla-
berrimo; stigmate 1 mm. longo compresso; fructu globoso 5-7 rom.
diametro.—Prru: growing about rocks on foggy summit of Viejas
Island, Nov. 17, 1919, R. C. Murphy 3219 (type, Brooklyn Bot.
Garden; 1sorypr, Gray Herb.).

An interesting member of Solanum subsection Dulcamara, chat-
acterized by its glabrous, very succulent herbage and its broadly
very shallowly lobed corollas. It seems to be related to the Chilean
S. Feuillei Dunal. The corolla of S. Murphyi has five very broad
short lobes that alternate with smaller but similarly proporti ned
lobes produced in the sinuses Along with Tillandsia and Oxalis

Same species. Except that the plants are distinctly shrubby below
the middle and seem to be persistent, they differ in no evident sd
from those of Viejas Island which have furnished the basis of my
description given above. I believe that they are conspecific.

FLORA OF THE GUANO ISLANDS 35

Polyachyrus nesites Johnston, sp. nov., suffruticosus 2.5-3 dm.
altus; caulibus vetustioribus caudicem laxe ramosus decumbens
formantibus; ramis erectis vel ascendentibus numerosis 1—2.5 dm.
longis evanescenter canescenterque arachnoideis vetustate plus
minusve glabrescentibus 2.54 mm. crassis; foliis 4-7 cm. longis 13—
26 mm. latis numerosis supra medium caulis abrupte reductis carno-
sulis pinnate lobatis, supra glabratis et sparse glandulosis, subtus

incano-arachnoideis, margine evidenter revolutis, lobis elongatis

prominente medio-costatis; alis petioli conspicuis infra basem petioli
in caulem vix decurrentibus apud folia medialia auriculas 3-4 mm.
latas amplexicaules formantibus; glomerulis capitulorum 14-18 mm.
diametro 3-12 globosis in paniculam subcorymbosam vel racemum
congestum dispositis; receptaculo lanuginoso, bracteis navicularibus
2-3 mm. longis glabratis; involucris flosculorum 1-2-floris 3.5-4 mm.
longis glaberrimis; tegulo exteriori paullo brevissime ad basem versus
callo prominenti notato ; flosculis roseis 5-6.5 mm. longis, tubo
(faucibus vix differentiatis) ca. 4 mm. longo, labia interiori in lobos 2
acutos ligulatos 2.5 mm. longos divisa, labia exteriori ovato-elliptica
apice tridentata; antheris (partibus fertilibus) ca. 1.4 mm. longis
linearibus apice appendicula oblongo-lanceolata ca. 1 mm. longa
coronatis basi cum caudis subulatis laevibus ca. 1 mm. longis ornatis;
setis pappi ca. 5 mm. longis albis—Perv: in sheltered places near
the crest of San Gallan Island, Nov. 27, 1919, R. C. Murphy 3467
(TYPE, Brooklyn Bot. Garden; photo and frag., Gray Herb.).

ong the Peruvian species probably closest to P. echinopsoides
(Hook.) DC., but differing in habit of growth, in less copious arachnoid
indument of the herbage, and in the smaller thicker leaves which are
glabrate above and have narrow elongate more separated lobes. Its
Closest relative seems to be in P. Poeppigii Kunze, of central Chile.
\ photograph of the living plant is reproduced on Plate VII.

EXPLANATION OF PLATES.

Puate VI.
gle o peatnn Island from the paprongcar vi dca the vegetation-
porting fogs that bathe the crests of the island. :
Fig. 2. San Gallan Island. A view near the summit, showing the ie owths of
ndgia latifolia that darken the lee-sides of the higher ridges.
Prate VII.
3. Solanum Murphyi sp. nov. A living plant near the summit of
Fig. 4. Polyachyrus nesites sp. nov. A plant growing near the summit
of San Gallan Island, the type station.

36 WEATHERBY

Il. NEW AMERICAN SPECIES OF TRICHOMANES.
By C. A. WEATHERBY.
(Plate VIII.)

TRICHOMANES daguense, spec. nov. Rhizoma breve erectum,
pilis tenuibus brunneis onustum. Frondes plures caespitosae 5.5-
16.5 em. longae. Stipes viridis 2-4.5 cm. longus quam lamina 4
brevior usque ad basim alatus. Lamina lanceolata basi paullo angus-
tata apice sensim attenuata 3.5(plerumque 5)-12 cm. longa 2-3.5
em. lata ad alam 1-1.5 mm. latam secundum rachin pinnatifida
nervis margineque pilis brunneis simplicibus rigidis sparse instructa.
Laciniae plus minusve remotae sinubus rotundatis interstinctae 10-
18-jugae lineari-oblongae medianae 1-2 cm. longae 3-4 mm. latae
adscendenti-patulae, fertiles apice excepto subintegrae vel leviter
undulatae truncato-obtusae apice 2-4 soros gerentes, steriles crenato-
serratae ad apicem angustum obtusum angustatae. Nervi crassills-
culi, nervilli catadromi 3-10-jugi plerumque simplices recti e costula
angulo acuto exeuntes. Cellulae plerumque oblongae c. 0.2 mu.
longae, parietibus crassis, luminibus parvis hyalinis. Indusia cylin-
drica basi angustata 1.6-1.8 mm. longa sub orificio 0.5-0.6 mm. lata
in limbum angustissimum sed distinctum suberectum vel patulum
0.2 mm. latum dilatata, in dentibus apicalibus prominentibus laci-
niarum immersa sicut alata. Receptaculum aetate longe exsertum.—
CoLomBia: forest, alt. 80-100 m., Dagua Valley, Cordoba, Dept. i
Valle, May 6-8, 1922, Killip, no. 5251, rype in U. §. Nat. Herb;
another sheet of the same number in Gray Herb.; same place and
date, Killip, no. 5098. +s

Species 7. delicato v. d. Bosch affinis, a quo differt laciniis longior!-
bus angustioribus minus dentatis, cellulis latis, indusiis longior!bUs
anguste cylindricis, venis simplicibus.

TRICHOMANES Killipii, spec. nov. Rhizoma c. 2 em. longum
erectum, apice pilis brevibus brunneis lucidis tenuibus onustun.
Frondes_plures caespitosae 8-13 cm. altae, stipitibus rachi_nervis
marginibusque pilis simplicibus rigidis brunneis sparse © oe
Stipes 24 cm. longus lamina multo brevior fuscus superne alatus
basim versus nudus. Lamina lanceolata basi paullo angustata eal
gradatim attenuata 5-9 cm. longa 1.3-2.5 cm. lata ad alam 1-1

tam secundum rachin pinnatifida, rachis plus minusve Asis
Laciniae 9-15-jugae approximatae vel subremotae sinubus rotum ies
interstinctae oblongae vel fere quadratae obtusae leviter adscenden!
integrae vel leviter undulatae vel apice paucidentatae mee

NEW SPECIES OF TRICHOMANES 37

0.7-1.3 em. longae 3-5 mm. latae, fertiles apice 1-3 soros gerentes.
Nervi tenues; nervilli catadromi 2—4-jugi 1-2-furcati imi e costula
angulo lato exeuntes deinde arcuati superiores adscendentes recti
angulo acutiore exeuntes. Cellulae laciniarum basi quadratae,
apicem versus oblongae 0.1—0.2 mm. diametro vel longae, parietibus
crassis, luminibus hyalinis majusculis. Indusia anguste cylindrica
1.8-2 mm. longa, orificio 0.4-0.5 mm. lato truncato vel in limbum
angustissimum bilabiatum erectum producto, immersa vel si indentibus
solitaria sicut exserta alata. Receptaculum aetate longe exsertum.—
CoLomsta: epiphytic in coastal thickets, alt. 0-10 m., Buenaventura,
Dept. El Valle, Oct. 5-10, 1922, Killip, no. 11727, type in Gray Herb.

Species 7’. holoptero affinis, quod differt praecipue venis fertilibus
simplicibus, indusiis in limbum conspicuum latum dilatatis. 7.
Galeottii etiam similis, quod tamen rhizomate vix erecto, lamina
latiore, stipite longiore, laciniis crebris subimbricatis patulis, venulis
crassioribus crebris differt.

T. delicatum v. d. Bosch appears to have been curiously misunder-
stood. It was based on a specimen of Cuming 21 from Quito in
“H. Berol.” and was explicitly described as having fasciculate fronds,
once-pinnatifid lamina with undulate-toothed segments, simple hairs
and very elongate, linear cells. Yet Baker (Syn. Fil. 79), followed
In the Index Filicum and even by Domin (Mem. Czech Soc. Sci. .
nM. 8. no. 2. 51 (1929)), referred it to the mainly West Indian 7.
alatum, a species in which even dwarfed fronds are strongly bipin-
natifid, with forked hairs and mostly quadrate cells. Hieronymus,
who should have had access to the type specimen, after remarking
that 7. delicatum was “‘a Hooker et Baker perverse ut synonymum
T. alati Swartz citatum” (Engler Bot. Jahr. xxxiv. 421 (1904)), him-
self disposes of it, even less convincingly, as a form of 7, Poepprgr
Presl. This last is closely related to and has been regarded by many
authors as synonymous with, 7. poly podioides; like that species, it has
a filiform, repent rootstock, scattered fronds, branched hairs and
quadrate cells. :

ere are no specimens of Cuming 21 in the Gray Herbarium;
there are, however, two sheets of a plant from Quito (Jameson 104)
remarkable for its greatly elongate cells and otherwise agreeing well
with van den Bosch’s description and with his drawings of details,
Published in Med, Rijks Herb. Leiden no. 38. 3 (1919). Here also
the citation “Jameson (Herb. Hook.)” is added to the republished

ription. In all probability the collection above ‘mentioned is
teant; I am accordingly taking the Gray specimens of it as authenti-
cally representing the species.

38 WEATHERBY

The following key attempts to define the American species of

richomanes, so far as known to me, which share the characters of
low stature, cespitose, once-pinnatifid fronds, spreading-ascending
subremote segments and unbranched hairs. 17’. macilentum and T.
ornatulum are known to me only from v. d. Bosch’s rather brief
descriptions and his posthumously published, fragmentary drawings.
They are included without entire confidence that they really belong
here, but for the purpose of indicating wherein the two species above
proposed as new differ from them, as described.

Indusium with subtruncate orifice or narrow spreading limb
not over 0.2 mm. wide.
Veins mostly simple; segments linear to rather narrowly ob-
long; cells of the leaf-tissue oblong, not more than three
m. long.
Indusium with narrow spreadin limb 0.2 mm. wide; vein-

m
oR
&
&
°
“t
5
5
r)
(aa
=)
=}

quadrate.
Segments prominently undulate-toothed ; cells of the leaf-
tissue mostly elongate, 4-6 times as long as wide, up to

_mm. long, thin-walled; indusium rather broadl

cylindric-ventricose with narrow more or less spread- B

_ ing MO ee ea T’. delicatum v. 4. B.
Segments (except sometimes at apex) entire or very shal-
lowly undulate of the leaf-tissue quadrate at
ase of segments, crowded and oblong between the

ns al ong;
cylindric with truncate orifice or narrow bilabiate erect
pe: mice i Get : Killip “i Weatherby.
ndusium with conspicuous spreading ]j .5 mm. wide; cells
thick catat 9) preading limb 0.5 m
Segments oblong or broader, coarsely 3—4-undulate-dentate
with nm narrow but rounded sinuses; veins of fertile
ee Se eee cole quadrate my phlona
as long as broad; i i i ibuliform. *
eran, ne Sond indo inunulor. F yum Ba
Segments “suboblong,” entire or 1—2-dentate, with acute
sinuses; cells 3-6 times as long as broad; indusium cyl- d.B
indric. Brazil (Rio) T. ornatulum v. @ P-

NEW SPECIES OF TRICHOMANES 39

lata basim versus gradatim angustata marginatus. Lamina deltoidea
vel late ovata vel rite ovata 2.5-6 cm. longa 2-3.5 cm. lata ad alam
12 mm. latam secundum rachin pinnatifida nervis margineque
pilis simplicibus rigidis brunneis ensiformibus prope basim articulatis
sparse obsita. Laciniae 3-6-jugae adscendentes oblongae late obtusae
basim versus integrae ad apicem grosse sparseque crenato-dentatae
subremotae sinubus rotundatis interstinctae, medianae 1-2 cm.
longae, c. 5 mm. latae. Venae crassae, venulae catadromae c. 5-
jugae simplices vel saepius 1-2-furcatae angulo acuto e costula
exeuntes. Cellulae plerumque oblongae ad nervos marginemque
irregulariter quadratae ad 0.2 mm. longae, parietibus crassis, luminibus
parvis hyalinis. Sori 1-3 in apice laciniarum superiorum immersi;
indusia subeylindrica basi angustata 2-2.4 mm. longa sub orificio

1 mm. lata in limbum angustum patulum dilatata; receptaculum
aetate longius exsertum.—Santo Domrnco: alt. 400-1000 m. Quita
Espuela, vicinity of San Francisco de Macoris, Prov. Pacificador,
April 5-17, 1922, W. L. Abbott, no. 2114a, Type in U. S. Nat. Herb.,
sheet no. 1,145,256; high ridges, c. 1100 m. alt., rare, Loma la Cam-
pana, Cordillera Central, Prov. Sto. Domingo, Feb. 2, 1929, E. L.
ry toa no. H 11507, U. S. Nat. Herb.

tamen totae diversae sunt.

TricHoMaNnes Curranii, spec. nov. Rhizoma breviter repens
‘ortuoso-ramosum trichomatibus crassis compressis articulatis pallide
brunneis onustum. Frondes subdistantes 14-18 em. altae omnibus
P artibus pilis rigidis e basi tuberculosa 2—4-ramosis deciduis plus
minusve obsitae. Stipes 5-7 cm. longus lamina brevior usque vel
fere ad basim alatus. Lamina ovata 9-11 em. longa 5-6 cm. lata
bipinnatifida basi vix angustata apice pinnatifido acuta, ala secundum
Tachin 1.5-2 mm. lata. Laciniae ec. 10-jugae lineari-oblongae medianae

cm. longae 0.8-1.2 cm. latae obtusae subremotae sinubus latis
rotundatis interstinctae profunde pinnatifidae; lacinulae oblongae
linearesve sparse crenato-denticulatae obtusae. Venae crassiusculae
ultimae in lacinulis valde adscendentes simplices vel 1-furcatae. Sori
n lacinulis singulis plerumque singuli; indusia (post maturitatem
tantum visa) late cylindrica subventricosa 1-1.4 mm. longa 0.6-0.8
mm. lata sub orificio ut videtur subtruncato vel leviter bicorne;
Teceptaculum breviter exsertum. Cellulae minimae quadratae vel
oblongae, parietibus crassis, luminibus plerumque vix ultra 0.05 mm.

ee

40 SMITH

latis.—VENEZUELA: common at 300-600 m. alt., Cerro Sta. Ana,
Paraguana Peninsula, April 13, 1917, Curran & Haman, no. 672,
TYPE in Gray Herb.

T. fimbriato affinis a quo differt textura densiore, cellulis minoribus,
lacinulis obtusis neque appendiculo capillari apicali ornatis.

The following key may serve to distinguish 7. Curranii from the
species which are apparently its nearest relatives, sharing with it
the characters of erect or short-repent rootstock with coarse, artic-
ulate, pale-brown trichomes, bipinnatifid lamina and branched hairs.
Lamina lance-elliptic, 15-30 cm. long, attenuate at base, broad-

at about the middle, segments often only shallowly
toothed, if pinnatifid the lobes produced into capillary ap-
ndages; indusium with narrow spreading limb. est
Siren Cele sae Shey Ciel aS. Boe . trigonum Desv.
Lamina ovate, mostly less than 15 em. long, broadest below
the middle, only the lowest one or two pairs of segmen

Lobes of segments acute, the sterile usually produced into a

oe sage appendage; texture thin, the rather thin-walled

cells with lumina 0.1 mm. in diameter; indusium with

narrow spreading limb. Trinidad.............. T. fimbriatum Backh.
Lobes obtuse, without capillary appendages; texture firm,

the thick-walled cells with very small lumina (0.05 mm.);__

indusium apparently subtruncate. Venezuela... 7’. Curranii Weatherby.

I am indebted to Dr. William R. Maxon of the United States
National Herbarium for the loan of indispensable material.

EXPLANATION oF Puate VIII.

Fig. 1. TricHoMANES CRASSIPILIS Weatherby, spec. nov. 4. Habit wr
6. sterile —— x 0.8; c. fertile segment x 0.8; d. indusium X 9 ©

Fig. 3. TRIcHOMANES DELICATUM van den Bosch. X 0. ; 4: b.
Fig. 4. TricHomanes Kinin Weatherby, spec. nov. Habit X 0.4; 9
sterile segment X 0.8; c. fertile segment X 0.8; d. indusium X 6.

IV. STUDIES IN THE BROMELIACEAE,—IL.

By Lyman B. Smrru.
(Plates IX—X1I.)
During the past year many points of taxonomic significance pete
come to my attention in the course of further study on the sd is
ceae, and since no immediate revision covering these ma mre”
contemplated it seems best to put them on record here. I am im aluable
to the following institutions for the opportunity to examme V

ai

STUDIES IN THE BROMELIACEAE,—II 41

specimens in their collections: the United States National Herbarium
(US), the Field Museum of Natural History (FM), the Herbarium
of the Royal Botanic Gardens at Kew (K), the Biological Institute
of Sao Paulo (SP), the National Museum of Rio de Janeiro, and the
Botanical Section of the Riks Museum at Stockholm (S). To Mr.
W. E. Broadway of Trinidad I owe much for critical observations
in the field, helpful correspondence, and frequent gifts of specimens.

The small lot of plants collected in the Rio Grongogy Basin in
eastern Brazil by Mr. H. M. Curran deserves special notice since it
includes three new species, Ronnbergia marantoides, Streptocalyx
Curranii, and S. lanatus. In two cases these species are so unusual
as to present difficulty in placing them generically. Evidently this
region in extreme southern Bahia is little known botanically and
should amply repay further investigation.

For convenience of reference genera and species are arranged in
alphabetical order.

_ Aechmea Bernoulliana Wittm. The following record is of
interest as being the first known instance of this species occurring
south of Costa Rica:

ECUADOR: Manat: imprimis in sitio El Recreo dicto lecta, 14° 8. Lat.,
Eggers 15489 (US, phot. G).

Araeococcus pectinatus, spec. nov., pergracilis, fere metralis,
stolonibus procreans: foliis eis Aracococci micranthi valde similibus,
ad 70 cm. longis, fasciculatis; vaginis anguste ovatis, tubum angustum
formantibus; laminis grosse serratis, anguste ensiformibus, acuminatis,
nterioribus canaliculatis, ad 20 mm. latis et 50 em. longis, exterioribus
valde reductis, vaginis multo brevioribus: seapo gracili, perelongato;
Scapl vaginis membranaceis, angustis, infimis internodia bene super-
antibus: inflorescentia glabra, pauciramosa, pulchre rubra, 15-22
cm. longa; bracteis primariis parvis, ovato-acuminatis, membranacels,
patulis: spicis elongatis, multifloris, 1 em. latis, erectis; thachi flexuosa,
Sulcata; bracteis florigeris ovatis, acutis vel mucronulatis, valde
Pectinatis, 7 mm. longis, flores subobscurantibus, auriculatis, reflexis:
floribus sessilibus, fructiferis solis cognitis: bacca late ovoidea, 5
mm. longa; seminis in bacca paucis, ovoideis, curvatis. Pl. XI,
: COSTA RICA: Pozo Azul de Pirris, Pacific foot-hills on a edge of
Sastal plain, alt. 95 m., 1927, C. H. Lankester 1164 (US, phot. G).

Although no good flowers were to be found on the above specimen
the open paniculate inflorescence, inferior ovary, and small few-
Seeded berries prove it to be an Araeococeus beyond any reasonable

42 SMITH

its leaves are an almost exact match (see FI. Bras. iii, pt. 3, t. 56,
fig. 1 (1891)). But it differs from all previously described species
of the genus in its relatively large and strongly pectinate floral bracts.

Glomeropitcairnia erectiflora Mez. Thanks to the efforts of
Mr. W. E. Broadway it is now possible to extend the known range
of this interesting species to Trinidad where he collected it on the
summit of Mt. Tucuche as early as 1893. Until the present, the
species has not been noted outside its type locality, the Island of
Margarita, Venezuela. At the same time the original description
may be enlarged as follows on the basis of fresh flowering material
sent recently by Mr. Broadway:

Petalis fulgide citrinis, quam sepalas paulo brevioribus, basi
biligulatis; staminibus inclusis, antheris angustis, 6 mm. longis;
stylo gracili, stamina subaequanti. Pl. X, figs. 6-7.

The presence of two scales at the base of each petal is of particular
interest, since the only other species in the genus, G. penduliflora is
characterized by a single scale on each petal (see Mez in Bull. Herb.
Boiss. ser. 2, v. 232 (1905)).

Hohenbergia Sellowiana Mez. Until now this species seems to
have been known only from the type material, a collection of Sellow
with no locality indicated except “Brasilia.” The following citations
serve to define its range more exactl y:

BRAZIL: Banta: near Bahia (Sao Salvador), 1915, Rose & Russell 19895
(US, phot. G); Bahia, in palmetis prope urbem, loco arido litori proxime,
1892, Lindman A63 (S).

Nidularium billbergioides (Schult. f.), comb. nov. Hohen-
bergia billbergioides Schult. f. in R. & S. Syst. vii. 1253 (1830). Til
landsia terminalis Vell. Fl. Flum. 137 (1825), Icones iii. t. 143 (1827);
not Nidularium terminale Ule. ;

This species is treated by Mez in De Candolle’s Monographiae
under the name, Nidularium bracteatum (Vell.) Mez. 1 illandsia
bracteata Vell. upon which the name is based is not a ibanesse
shown by the scales at the base of the petals, the free petals, am
the free second series stamens (see Fl. Flum. Icones iii. t. 125). These
characters with the serrate floral bracts (see Fl. Flum. 132) make It
Seem more probable that Tillandsia bracteata is the same as Aechmet
fasciata (Lindl.) Bak. .
_ The next name in question is Tillandsia terminalis Vell. and it
is almost certainly the same as Nidularium billbergioides. Unfort
nately, however, Ule has used the name Nidularium terminale for t
a species with serrate floral bracts so that the combination 1s 9°
usable.

doubt. Its sessile flowers would ally it with A. maicranthus, for which

STUDIES IN THE BROMELIACEAE,—II 43

The next name for this species is Hohenbergia billbergioides Schult.
f.inR. & S. and this combination for Nidulariwm is made accordingly.

Quesnelia (§ Wawraea) Hoehnei, spec. nov., epiphyta, acaulis,
stolonibus procreans, ca. 40 cm. alta: foliis paucis, anguste rosulatis,
inflorescentiam subaequantibus vel ea distincte longioribus, cellulis
sclerenchymaticis permanifeste auctis, subtus lepidibus minutis
peradpressisque dense praeditis; vagina anguste elliptica, laminam
longitudine fere aequanti, atro-violacea; lamina subligulata, apice
rotundato-apiculata vel retusa, 3-4.5 cm. lata, minute remoteque
denticulata, viridi: scapo erecto, gracili, albo-tomentoso; scapi bracteis
membranaceis, integris, pulchre roseis, anguste ellipticis, apiculatis,
infimis quam internodia brevioribus, summis sub inflorescentiam
comatim cumulatis: inflorescentia breviter spicata, subcapitata,
pauciflora, ad 8 em. longa; axi albo-tomentoso; bracteis minutis,.
membranaceis, ovario bene brevioribus: floribus erectis, sessilibus,
5 em. longis; sepalis oblongo-lanceolatis, minutissime asymmetricis,
apice anguste rotundatis, glabris, bene nervatis, ca. 20 mm. longis,
basi 2 mm. connatis; petalis atro-rubris, anguste lanceolatis, sepalis
subduplo longioribus, basi ligulis binis fimbriatis acutis, paulo zygo-
morphe dispositis; staminibus inclusis; stylo gracili; ovario sulcato,
glabro, tubo epigyno magno, ovulis excaudatis. Pl. X, figs. 3-5.
_ BRAZIL: Sio Pauo: dense forest, Estagio Biologica, Alto da Serra, alt.
1929) Mm lat. 23° 47’ S.. long. 46° 19’ W., 1929, L. B. Smith 1968 (G, type);
c Sp) oe 1887 (G); 1923, F. C. Hoehne in Inst. Biol. ee Age ni
'G, SP); Ca eats nes on Dt) Taet Biol. 8. P. 12840 (G, SP); Riveirdo
Pires, @. Edvoall tn Inst, Biol. 8. P. 12336 (SP). ee

nNS8lt. DIOl. Oo. I. . .

: This species is most nearly related to B. indecora Mez, from which
it differs in its entire membranaceous scape-bracts and rounded sepals.
tis a pleasure to name this handsome little Bromeliad after my
food friend Dr. F. C. Hoehne, whose great experience and ready
“ooperation have been an invaluable aid in studying the plants of
his country

~onnbergia marantoides, spec. nov., acaulis, ad 25 cm. alta:
foliis Tosulatis, utrinque minute lepidotis; vaginis late ellipticis;
laminis lanceolatis, 35-40 mm. latis, super vaginam valde contractis,

te acutis apiculatisque: scapo erecto, gracili, sulcato, foliis breviore;
Scapl_bracteis peranguste triangularibus, membranaceis, reflexis:
inflorescentia simplicissima, laxe spicata, glabra; bracteis florigeris
‘Bhotis, Verisimiliter deciduis: floribus sessilibus, 25 mm. longis;
Sepals nermibus, retusis, 10 mm. longis, ad 3 mm. connatis, asym-
7 etticis, bicarinatis, carinis cum eis ovarii continuis; petalis liberis,
(ex sicco) aureis, eligulatis, laminis anguste ellipticis; staminibus

44 SMITH

inclusis, serie II cum petalis alte connatis, antheris mucronatis,
granulis pollinis biporatis; ovario omnino infero, late 6-carinato, tubo
epigyno breve sed manifesto, placentis apice loculorum affixis, ovulis
longe caudatis. Pl. XI, figs. 1-3.

BRAZIL: Banta: forests of Rio Grongogy Basin, alt. 100-500 m., 1915,
H. M. Curran 142 (US, phot. G).

This plant does not fit exactly any of the genera of the Bromeliaceac
as defined by either Mez or Harms. However, it does come very
close to the Andean genus, Ronnbergia, from which it differs only in
its long-caudate ovules. Since Acchmea is made to include both
caudate and obtuse types of ovules it does not seem inconsistent
to broaden the definition of Ronnbergia in the same way.

This species resembles the two species of the subgenus Ewronnbergia
in its entire subpetiolate leaves, but differs from them very strikingly
in its broadly 6-winged ovary.

Sodiroa dissitiflora André. Here is one more instance of a
South American genus of Bromeliaceae turning up in Central America,
Greigia and Puya having been so noted by Mr. P. C. Standley, and
Araeococcus by the writer.

COSTA RICA: Cascajal, 1920, Lankester K166 (KK, phot. G); SAN José:
on tree, La Hondura, alt. 1300-1700 m., 1924, Standley 37813 (US, phot. G).

Streptocalyx Curranii, spec. nov., acaulis, ca. 4 dm. altus:
foliis rosulatis, subtus dense minuteque lepidotis, ad 6 dm. longis;
vaginis ellipticis, pallide brunneis ; laminis anguste ligulatis, late
acutis apiculatisque, 20-25 mm. latis, remote denticulatis: scap0
erecto, gracili, paulo furfuraceo; scapi bracteis lanceolatis, membra-
naceis, integris: inflorescentia pyramidata, bipinnatim paniculata,
11 em. longa, paulo furfuracea, mox glabrescenti; bracteis primarlls
anguste lanceolatis, membranaceis, quam spicae axillares bene brev-
loribus; spicis sublaxe cylindricis, quaquaverse florigeris; bractels
florigeris Ovato-acuminatis, membranaceis, ovario subaequantl ie
floribus sessilibus, suberectis; sepalis liberis, asymmetricis, apiculatis,

linis biporatis; ovario omnino infero, cylindrico, (ex sicco
sulcato, tubo epigyno manifesto, placentis prope apices loculorum
affixis, ovulis multis, obtusis. PI. XI, figs. 7-9.

BRAZIL: Banta: forests of Rio Grongogy Basin, alt. 100-500 m., 1916,
H. M. Curran 143 (US, phot. G).

This species is easily distinguished from the other members of the
subgenus Eustreptocalyx by its combination of slender scape
polystichous-flowered spikes.

;
;

STUDIES IN THE BROMELIACEAE,—II 45

Streptocalyx lanatus, spec. nov., acaulis: foliis rosulatis, subtus
minute peradpresseque lepidotis; vaginis maximis, ad 20 em. longis,
anguste ellipticis, atro-castaneis; laminis peranguste triangularibus,
basi 3 em. latis, apice involutis, grosse serratis, spinis ad 6 mm.
longis: scapo erecto, dense lanato; scapi bracteis ligulatis, apice
involutis, subtus lanatis, mox glabrescentibus, supra adpresse lepi-
dotis: inflorescentia subtripinnatim paniculata, dense cylindrica,
dense lanata; bracteis primariis eis scapalibus isomorphis; ramis
inferioribus paulo divisis; spicis -quaquaverse florigeris; bracteis
florigeris aciculosis, pungentibus, ovario subaequantibus: floribus
sessilibus, adscendente curvatis; sepalis 7 mm. longis (sine spina),
alte connatis, asymmetricis, infra apicem retusum longe spinulatis,
coriaceis; petalis liberis, eligulatis; staminibus inclusis; ovario clavato,
placentis apice loculorum affixis. PI. XI, figs. 4—5.

RAZIL: Banta: forests of Rio Grongogy Basin, alt. 100-500 m., 1915,
H. M. Curran 138 (US, phot. G)

The definition of the subgenus Eustreptocalyx as given by Mez
must be broadened to include plants with tripinnate inflorescences in
order to cover this species.

Tillandsia argentea Griseb. F ormerly known only from the
Greater Antilles.

MEXICO: Oaxaca: between San Carlos and San Bartolo, alt. 1000-1600
m., 1895, F. W. Nelson 2561 (US, phot. G).

Tillandsia Bakeri, nom. nov., Catopsis flexuosa Bak. Journ. Bot.
xxv. 175 (1887), Tillandsia flexuosa (Bak.) Mez in DC. Mon. ix.
144 (1896), not Swartz (1788).

Tillandsia crispa (Bak.) Mez. The following citation marks a
Westward extension of range for the species:
in ANAMA: Darren: Cerro de Garagar’, Sambt basin, alt. 500-974 m.,

2, H. Pittier 5658 (US, phot. G).
houlandsia Lorentziana Griseb. It is not at all surprising that
moe Species already known from Argentina, Paraguay, and Bolivia

ould appear in similar country in southern Brazil.

BRAZIL: Parani UL ‘3 : xd p M
Vo RAZIL: ne bus,” Villa Velha, 1903, P. Dusén 2810 (Museu
Nacional do Rio de Suuslro. price G) - Dusén 2756 (Mus. Nac. Rio); 1909,
Dusén 7624 (S); 1910, Dusén 9528 (S).
yrillandsia micrantha Bak. Until recently this species has been

rahe only from fruiting material, but now through the kindness of
hes W. E. Broadway, who has forwarded preserved flowering material

m Trinidad, it is possible to complete the description as follows:

46 SMITH

Petalis albis (ex cl. W. E. Broadway), quam sepalae minute longiori-
bus; staminibus profunde inclusis, stylo superantibus, antheris
ovoideis, apice minute mucronatis, filamentis validis, basi dilatatis.
Pl. IX, figs. 5-9.

The general character of petals, stamens, and pistil agrees closely
with that already known in other members of the Section Pseudo-
Catopsis, but it is interesting to note that in 7. micrantha the base
of the filament is slightly broader than the anther.

Tillandsia pruinosa Sw. Not recorded south of Colombia on
the Pacific coast heretofore.

ECUADOR: Oro: Portovelo (Gold Mine near Zaruma), alt. 600-1000 m.,
1923, Hitchcock 21259 (G).

Tosulatis, 40 cm. longis, utrinque lepidibus minutis peradpressisque
praeditis; vaginis ellipticis, 10 cm. longis, 4.5 cm. latis, atro-purpure®
minute pallido-maculatis; laminis anguste triangularibus, a
vaginam + 15 mm. latis, pallidis: scapo decurvo, quam folia brewer
bracteis foliaceis omnino occulto: inflorescentia laxe bipinnatim
paniculata, pendula (ex cl. P. C. Standley); axi gracili, leviter flexuos?

‘ Mez in DC. Mon. Phan. ix. 703 (1896).

STUDIES IN THE BROMELIACEAE,—II 47

glabro; bracteis primariis maximis, patentibus, quam spicae axillares
3-+-plo longioribus, ad 26 cm. longis, 2 cm. latis, spicis involventibus,
pulchre fulgideque rubris: spicis lanceolatis, bene complanatis, valde
stipitatis, 6 em. longis, 2 cm. latis, dense ad 8-floris, pallide viridibus;
thachi valde complanata, angulata; bracteis florigeris anguste lanceo-
latis, acuminatis, rhachin obtegentibus, sepalis aequantibus, ad 35
mm. longis sed saepe multo brevioribus, dorso valde impressis,
laevibus vel apice paulo nervosis, glabris: sepalis anguste lanceolatis,
acuminatis, valde carinatis, aequaliter liberis, laevibus, glabris;
petalis et partibus genitalibus ignotis: capsulis acutis, quam sepala
brevioribus. Pl. 1X, figs. 1-4.

..HONDURAS: Comayacvua: on oak tree in pine forest, El Achote, near
Siguatepeque, alt. 1500 m., 1928, Standley 56177 (FM, rypx; phot. G).

This is distinguished at once from all other species of Tillandsia
by its pendulous scalariform inflorescence and its exceedingly long
primary bracts. In fact it is rather difficult to say to what it is
most nearly related. Its occurrence in Central America would argue
ts probable inclusion in the section Platystachys, and on the basis
of the key in Mez’s Monograph it comes nearest to 7. Leiboldiana
in that section,

Unlike any other species of Tillandsia known to the writer, 7’.
Standleyi has a peculiar growth at the base of each spike on the side
toward the axis of the inflorescence (Pl. IX, figs. 1-2). Various
facts would indicate that this growth functions as an aid to spreading
the spikes at maturity. The deep groove left in the basal floral
bract which is next the axis (Pl. IX, fig. 4) and the total absence
ot “ny gtoove in the one on the opposite side indicate that before
qaturity the spikes were erect and very closely appressed to the axis.
The extremely shrunken condition of the growth indicates that it
Was formerly very turgid, and this with its spreading position evidently
es 't well adapted to pushing the spike away from the axis of the
nee. It is not possible to ascertain the morphological
Mature of this growth with any certainty until fresh material is
obtainable,
bose figure of the habit is in some slight degree a restoration, since
ab ame was much folded over on the herbarium sheet.
_ pleasure in naming this remarkable species after Mr. F. ©.
dley, whose name is written so large in Central American botany.
a triticea Burch. The discovery of 7. triticea in central
i prac Messrs. Killip and Smith is particularly noteworthy, since
The . tely extends the range of this species into the Andean region.
Species is of very wide range in eastern South America, but

48 SMITH

strangely enough has never before been recorded from the north-
western part of the continent which is the center of distribution
of almost all of its near relatives.

PERU: Junin: epiphytic in dense forest, Pichis Trail, San Nicolas, alt.
ca. 1100 m., 1929, H. P. Killip & A. C. Smith 26055 (GQ).

Vriesia sparsiflora, spec. nov., acaulis, pergracilis, semimetralis:
foliis rosulatis, 20 cm. longis, subtus minute perobscureque lepidotis;
vaginis late ellipticis, pulchre violaceis; laminis ligulatis, abrupte
acutis, apiculatis, 5-20 mm. latis: scapo erecto, gracili, foliis bene
longiore; scapi bracteis anguste ellipticis, acuminatis, supremis quam
internodia distincte brevioribus: inflorescentia perlaxe bipinnatim

XU eC:

parvis: spicis erectis, perlonge stipitatis, bracteis sterilibus multis
auctis, 2-3-floris sed flori apicali semper sterili abortivo; bracteis
florigeris ovato-acutis, membranaceis, nervatis, sepalis subduplo
brevioribus et cum eis secunde versis, haud carinatis: floribus patenti-
bus, secundis, ad 35 mm. longis; sepalis anguste ellipticis, i
obtusis, nervatis, carinatis; petalis biligulatis, verisimiliter citrims;
staminibus petalis subaequantibus. PI. X, figs. 1-2.

BRAZIL: Sio Pauto: Estacdo Biologica, Alto da Serra, 1926, Hoehne in
Inst. Biol. 17928 (SP).

This graceful little species comes next to Vriesia Philippo-Coburg
in Mez’s treatment, but differs from it in its much smaller size, its
inconspicuous primary bracts, and the remarkably long sterile bee
of its spikes. Owing to the small number of fertile flowers on a spike
(never more than 2), it is not obvious at first that the flowers are
secund. But careful examination shows that in several cases the
floral bracts are twisted so as to lie almost entirely above the flower,
thus indicating that the flower ‘also must have turned to allow such
movement.

EXPLANATION OF PLATES.
Piate IX.

TitLaNpsta Sranpueyi L, B. Smith (Standley 56177), habit X he
Same, spike and primary bract x 1%.
Same, sepal :
Same, basal floral bract from side next the axis X 1. 5
TILLANDSIA MICRANTHA Bak. (W. E. Broadway), flower X 2
Same, sepal :
Same, petal and stamen X 5.
Same, pistil x 5,
Same, branch x 1,

O° MEER OF 89

GROUP OF ASPLENIUM FRAGILE 49

PLATE X.

Fig. 1. Vriesta sparsirtora L. B. Smith (F. C. Hoehne in Inst. Biol. S.
P. 17982), habit x \4.
Fig. 2. Same, petal and stamens x 1. :
“ 3. QuEsNeLIA Hoesnetr L. B. Smith (L. B. Smith 1968), habit x \4.
“4. Same, flower
“5. Same, petal and stamens x 1.
“6. GLOMEROPITCAIRNIA ERECTIFLORA Mez (W. E. Broadway), flower
and floral bract x 1. ES
Fig. 7. Same, petal and stamen x 1.

Piate XI.
Fig. I. RONNBERGIA MARANTOIDES L. B. Smith (H. M. Curran 142), habit
xX :

XM.
Fig. 2. Same, flower 1.
‘ 3, Same, petal and stamen X 2.
4. Streprocatyx Lanatus L. B. Smith (H. M. Curran 138), branch
ie

section of the inflorescence < 1
Fig. 7.
of panicle x 1,
Fig. 8. Same, sepal x 2.
9. Same, petal and stamens x 2.

V. THE GROUP OF ASPLENIUM FRAGILE
IN SOUTH AMERICA.

By C. A. WEATHERBY.

The group of Asplenium fragile consists, in South America, of
three closely related and critical species. All three share the char-
‘ters of habit, large number of pinnae, glandularity of frond, prolif-
“tation of the lower axils, and blackish-reticulate spores.'! Probably
mall, and almost certainly in the material of A. fragile, var. lomense

ed, the stipes of the proliferous fronds frequently become greatly
elongated, with the lamina poorly developed, or lacking altogether.
After the lamina, or what there is of it, has withered, the stipe re-
mains, apparently living, until the young plant which develops

‘One can easily be misled as to the character of the spores in herbarium
Specimens. : * hi

50 WEATHERBY

from the gemma at its summit has become well rooted, giving quite
the appearance of a flagelliform stolon.

Since the three species have been confused in herbaria, the following
attempt to define them and the citation of specimens belonging to
them, may be of service. I am much indebted to the authorities of
the United States National Museum for the loan of specimens, and
to those of the Royal Botanic Garden, Kew, for fragments of the
type of Asplenium Gilliesti. These fragments, and photographs,
in the Gray Herbarium, of the two plants cited by Hieronymus in
the original publication of A. Lorentzii, have been used, in addition
to the descriptions, in determining the application of the two names
concerned. A. fragile is employed in the usually accepted sense, for
the plant which seems most nearly to agree with Presl’s description
and figure.

Margins of the broadly flabelliform-rhombic, pale green and
(when mature) rather thick pinnae dentate with sharply
acute or acuminate teeth; median cells of the scales elon-
gate-oblong or linear-oblong, generally with small lumina;
spores nearly spherical, 50 uw in diameter

Margins of pinnae entire, crenate, or with mostly blunt, short-
mucronate teeth; median cells of the scales broadly oblong

square, usually with conspicuous lumina; spores chiefly
psoid, 30-50

1, A. Gilliesii.

Pinnae trapezoid-ovate or trapezoid-oblong, not at all or only
gh r in well-developed fronds,
ate at 1

dentate with mostly blunt teeth.................--- 2. A, Lorenizii.
Pinnae flabe form-rhombic, in well developed fronds com-
monly auriculate or shallowly or sometimes deeply bi-tr-

th
Stipe and rachis brown; pinnae bright green and rather
thin, their margins or those of their lobes entire.
3a. A. fragile, var. lomense.

i. AspLentum Griurestt Hook. Exot. Fl. ad t. 208 (1827). A.
Gilliesianum Hook. & Grev. Icon. Fil. t. 73 (1829). A. debile Fée 10
Ménm. 28, t. 35, f. 2 (1865) (see Christensen Ark. Bot. xx, pt. 7 8
(1926)).—Peru: Chicla, April 21-23, 1882, Ball (G); Oroya, July
14, 1914, Rose, no. 18,702 (U. S,); Ollantaytambo, 3000 m., Apt
26, 1915, Cook & Gilbert, nos. 340 and 342 (U. S.); among volcanle

near Tinta, 3500 m., April 15, 1915, Cook & Gilbert,

(U. S.); Tarata, Espinosa (G.). Bottvta: La Paz, 11,500 ft., fart
1901, R. 8. Williams, no. 2596 (mixed with A. fragile; U. 8.); 18%

> 45 (G.),

GROUP OF ASPLENIUM FRAGILE 51

Bang, no. 121 (G., U. S.); Jan. 25, 1906, Buchtien, no. 165 (U. S.);
Pozna, 4200 m., May, 1908, Buchtien, no. 1138 (U. S.); Comancha,
Rose, no. 18,878 (U. S.); Sorata, 7500 ft., Sept. 3, 1901, R. S.
Wiliams, no. 2597 (U. S.).

may perhaps be explained by the fact that the type material (so far,
at least, as evidenced by the fragment seen by me) is young and has
immature spores. Ball (Journ. Linn. Soc. xxii. 64 (1885)) and Chris-
tensen (Ark. Bot. xx, pt. 7. 14 (1926)) have pointed out that the
characters of A. Gillicsii are not altogether constant and that forms
seemingly intermediate between it and A. fragile are found. However,
itis usually rather readily distinguishable by the ear-marks enumer-
ated in the key and seems quite worth maintaining—as is done by both
the authors referred to.

2, AspLENIUM Lorentz Hieron. Engl. Bot. Jahrb. xxii. 375
(1896). —Boxrvra : Im Walde, Unduavi, Nord-Yungas, 3300 m.,
Buehtien, nos. 408 (G.), 901 and 2642 (U. S.); Oct., 1885, Rusby,
no. 405 (G.); Pinos bei Tarija, 2300 m., March 12, 1904, Fiebrig, no.
“974 (U.S.), in part.
A. Lorentzii appears to vary toward A, Gulliesii, but differs from
that species in the structure of its scales and the shape of its spores,
and from both A. Gilliesii and A. fragile in the texture and shape of
Pitnae, which, as Hieronymous remarks, resemble, when well
‘veloped, those of small specimens of the group of A. lunulatum.
On these characters it is here kept up.

3, ASPLENIUM FRAGILE Presl, Tent. Pterid. 108 (1836). “A.
toloniferum Bory” Presl, Rel. Haenk. i. 44, t. 6, f. 4 (1825), not

(1804). 4. rhomboideum Brack. Wilkes Expl. Exp. xvi. 156,
ni rm (1854). —Vernezueta: Péramo de Timoteo, Mérida, 3000 m.,
M » #1, 1922, Jahn, no. 865 (U. S.). Coromsta: without locality,
orl, no. 326 (G.). Ecuapor: Cuenca, Sept. 16, 1918, Rose &
no a 22,734 AU. S.); Sept. 17-24, 1918, Rose, Pachano & Rose,
35 4 839 (G., U, S.); in dumetis Cotocallas, Mille, no. 109 (fronds

m. tall; G., U. S.); ravines at Chuquipollo, Dee. 8, 1855, Couthouy,
he. Peru: Panticalla Pass, 3000 m., July 14 and 17, 1915,
May 4a. ibert nos. 1825 and 1885 (U. $.); La Quimua, 12,000 ft.
Wi. @ 1922, Macbride & Featherstone, no. 2025 (U. S.); [Bajos]
+ No. 44 (type or cotype of A. rhomboideum Brack. U. S23

88 exp.
on
Wall of reforested terrace, Ollantaytambo, 3000 m., May 11,

1)
9, Cook & Gilbert, no. 661 (U. $.). Boravia: Sorata, 10,000 ft.,

52 WEATHERBY

Feb., 1886, Rusby, no. 404 (G., U. S.); Pelichuco, 10,000 ft., May 1,
1902, R. S. Williams, no. 2595 (U. S.).

I cannot understand what is meant by A. fragile, var. stoloniferum
Rosenst. Meddel. Rijks Herb. Leyden no. 19, 11 (1913), said to be
based on A. stoloniferum Presl. There is properly no A. stoloniferum
Presl; that author, in the Reliqueae Haenkeanae, misapplied the
name A, stoloniferum Bory to the plant which, recognizing his error,
he later called A. fragile.

Dr. Winifred Robinson (Bull. Torr. Bot. Club xl. 209 (1913)
makes the surprising statement that the type locality of A. rhombo-
dewm Brack. is the Hawaiian Islands. It is, of course, actually
Bajios, Peru, and no other collection is cited by Brackenridge. The
identity of his plant is quite clear from his excellent figure and the
Wilkes Expedition specimen in the National Herbarium. The
Hawaiian plant in question (which is closely related to, but hardly
conspecific with, the Andean) he seems to have referred to A. Menzte-
sit Hook. Mettenius, Hooker, and Christensen were quite correct in
referring A. rhomboideum to A. fragile. I cannot separate the two
even varietally and believe that Hooker was right in supposing that
Pres] had described a dwarfed and Brackenridge a more luxuriant
phase of the same species.

ASPLENIUM FRAGILE Presl, var. LOMENSE Weatherby, Contr. Gray
Herb. Ixxxy. 13 (1929).—Cuite: under rock in bottom of gulch
above springs, in moist and sheltered situation, Aguada del Panul,
Pee Antofogasta, Dept. Taltal, Dec. 4, 1925, Johnston, no. 5421

Asplenium peruvianum Desv. Mém. Soc. Linnéenne Paris vi. 271
(1827) was assigned to this group by Mettenius, probably correctly,
but I cannot be sure of its identity. Desvaux’s description is very
brief and not very helpful. A photograph, in the Gray Herbarium,
of his type specimen at the Paris Museum, shows a single small and
seemingly depauperate frond, broken in the middle and at the tip,
and earing seven pairs of pinnae. The better developed of these
exhibit 4 strong tendency toward the quadrilateral outline characte
istic of A. Lorentzii, but are more coarsely toothed and more auricir
late than in that species. Desvaux’s binomial, based, as it 18, 0
material insufficient in a critical group, had best remain in the limbo
of species dubiae, at least until someone thoroughly familiar with the
points involved can examine the actual type. With it may =
A. tenue Presl Rel. Haenk. i. 44, t. 6, f. 5 (1825), if, indeed, it is °
this affinity.

NEW SPERMATOPHYTES 5a

VI. NEW SPERMATOPHYTES FROM MEXICO
AND ARGENTINA.

By Ivan M. Jounston.

Struthanthus Hunnewellii, sp. nov., fruticosus erectus villo-
sulus sparse ascendenterque ramosus 3-5 dm. altus; ramis juventate
dense conspicueque villosulis canescentibus, maturitate glabres-
centibus fuscescentibus 2-4 mm. crassis; foliis alternis lanceolatis
vel lanceolato-ovatis vel ellipticis coriaceis laeto-viridibus integer-
rimis inconspicue villosulis concoloribus evidenter costatis sed ob-
scurissime nervatis, maturitate plus minusve glabratis 8-15 mm.
latis 2-4 em. longis apice acutis vel rariter obtusis, basi obtusis vel
latissime acutis in petiolum 1-2 mm. longum villosum contractis;
floribus pluribus dioeciis (femineis tantum cognitis) ad apices ped-
unculorum congestis; pedunculis ex axillis foliorum maturorum
frumpentibus 1-2.5 cm. longis villosulis ascendentibus tantum
summum ad apicem bracteatis vel floriferis; bracteis conspicuis
foliaceis villosulis 3-10 mm. longis 1-3 mm. latis lanceolatis tarde
deciduis, exterioribus majoribus fructibus longioribus; corolla floris
feminini 2.5 mm. longa extus hispidulo-villosula, lobis anguste lineari-
lanceolatis ca, 0.7 mm. latis acutis; staminoideis supra medium

Pedunculate, conspicuously bracteose, capitate flower-clusters. Its
closest Telatives appear to be S. alni Bartlett, from Oaxaca, and 8.

Fesemble S. Hunnewellii in having very shortly pedunculate (or even
subsessile) cymose-capitate inflorescences and also a tendency for
leafy bracts. The former species is glabrous. The latter, although
distinctly pubescent is much less obviously so than in the species
here proposed. In §. Hunnewellii all the collar-like bracts, borne
lust below the flower-attachment, produce on their abaxial side an
“vident lanceolate herbaceous blade. In the two related species the
r-like bracts of only the outermost flowers of each cluster develop
Pronounced blades. ‘The bracts of the inner flowers of the cluster

54 JOHNSTON

quite lack them. The blades of the bracts of the outermost flowers
are leaf-like in size and form. In S. Hunnewellii, however, the
bract-blades are all much smaller and proportionally more elongate
than the cauline leaves.

It is a pleasure to associate with this highly interesting plant the
name of Francis Welles Hunnewell, for many years the Curator of
the Herbarium of the New England Botanical Club, who collected the
species on a recent journey in Mexico. Mr. Hunnewell’s material
was obtained while he was motoring over the recently constructed
highway between Mexico City and Cuernavaca. It was through a
study of his collections that this well marked species was first brought
to my attention.

Hysterionica caurina, sp. nov., herbacea erecta 1-4 dm. alta;

setosis, margine integris vel obscurissime sinuatis paullo incrassatis
ciliatis, juventate plus minusve evanescenter glanduliferis; inferiori-
bus mox deciduis oblanceolatis sub apice late acuto usque ad basem
versus gradatim attenuatis; mediis majoribus persistentibus lanceo-
latis vel oblongo-lanceolatis 3-4.5 cm. longis 5-10 mm. latis sessilibus
basi rotundis vel obtusis apice late acutis a basi 3(—5)-costatis;
superioribus mediis similibus sed aliquantum minoribus; pedunculis
30 mm. longis terminalibus vel ex axillis foliorum superiorum
reductorum erumpentibus pilosis medium bracteolas duas subulatas
1-3 mm. longas saepe gerentibus; involucro capituli 7-10 mm. dia-
metro; tegulis ca. 50 linearibus acutis ca. 5 mm. longis ca. 0.3 mm.
latis minute glanduliferis lobos corollae disci vix superantibus 1
costa setosis 3-4 seriatis imbricatis subaequilongis, exterioribus
viridibus ceteris stramineis; ligulis linearibus flavis ca. 100 pistilifer's,
tubo (faucibus haud differentiatis) ca. 1.5 mm. longis, limbo ¢.
® mm, longo ca. 0.5 mm. lato; corollis disci 3-3.5 mm. longis herma-
Phroditis flavescentibus, lobis deltoideis ca. 0.3 mm. longis cTasse
marginatis, tubo ca. 1 mm. longo extus sub lente minute sparseque
pubescentibus, faucibus cylindricis ca. 2 mm. longis quam tubo duplo
crassioribus antheris 1.3-1.5 mm. longis basi rotundis apice acutls;
stgmato breviter lobato; pappo albo duplicato, parte interiore
setis 15-20 antrorse barbellatis 3-3.5 mm. longis sistente, patte
exteriore paleis anguste oblongis vel linearibus 0.30.5 mm. longs
apice laceratis composita; achaeniis valde compressis ca. 1.2 mm.
longis 0.5-0.6 mm. latis apice truncatis, lateribus antrorse appressedue
Sparsim pubescentibus—ARGENTINA: common along roads, Perico,
Prov. Jujuy, Jan. 19, 1930, L. R. Parodi 9052 (ryre, Gray Herb.);

|

NEW SPERMATOPHYTES 55

common, E] Suncho, Prov. Catamarca, Sept. 11, 1915, P. Jdrgensen
1279 (G).

A very distinct species related to H. jasionoides Willd., differing
in its smaller heads, very much more conspicuously heteromorphic
pappus, branching stems, very short peduncles, and much larger
broader greener and less firm leaves. The species is known only from
northwestern Argentina.

Contrrp. Gray Hers. XCV. Piate |

CUPHEA OBTUSIFOLIA Koehne.

Contris. Gray Herp. XCV.

.
i

.

DETERMINE: Se
INED AT THE Gray HERBARIUM

2

nents

Puate II,

ys Plantx in ¢ivitate argentina Misiones collect.

CupHea oBiiqua Bacigalupi.

Contris. Gray Hers. XCV. E III
LATE :

‘ ‘ j
CuPHEA PARANENSIS Bacigalupi.

Contris. GRAY Hers. XCYV. Puate IV,

CuPHEA TUBE IF MIS Koehne.

Contris. Gray Hers. XCV. PuaTE V.

saan
Ea)

Da

. wk c&. pe eee

©. @.
raed
PLANT A:

ib
43-992 oTaTE oe ai a

Baar tgevie. serene
Go, Jk, (90,
7 T

@

CurHEA TENUIPES Bacigalupi.

rip. Gray Hers. XCV. PLatTe VI.

Fig. 1 (upper). San GaLLan ISLAND FROM SOUTHEAST.
Fig. 2 (lower JPPER RIDGES OF SAN GALLAN ISLAND.

Contris. Gray Hers. XCV. PuatTe VII.

Fig. 3 (upper). Soranum MurPHYI Sp. nov.
ig. 4 (lower). PoLYACHYRUS NESITES Sp. NOV.

Contris. Gray Hers. XCV. Puate VIII.

Contris. Gray Hers. XCV. 7 Pua EX,

Figs. 1-4. Trtuanpsra Stanpeyi L. B. Sm.
Figs. 5-9. T. micranTua Bak.

Contris. Gray Hers. XCV. Puate X.

Figs. 1-2. Vrresra sparsiFLora L. B. Sm.
Figs. 3-5. Quesnetia Horennet L. B. Sm.
Figs. 6-7. GLOMEROPITCAIRNIA ERECTIFLORA Mez.

Contris. Gray Hers. XCV. PLaTE XI

. RONNBERGIA MARANTOIDES L. B. Sm.
. SrreprocaLyx LANATus L. B. Sm.

: Ararococcus pectinatus L. B. Sm.
Figs. 7-9. SrrerrocaLyx Currant L, B. Sm.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

XCVI.
ISSUED AUG 15 1931

I. “oe apa to a general treatment of the Eupa-
eik

By B. L. Rosrnson.

PUBLISHED BY
THE GRAY HERBARIUM OF HARVARD UNIVERSITY

CAMBRIDGE, MASS., U. S. A
1931

/
RV & YL /00

RECORDS PRELIMINARY TO A GENERAL TREATMENT
F THE EUPATORIEAE—IX.

study and identification. In this recently examined material it has
been possible to find some thirty plants which appear to merit record
as New species or varieties. These are characterized below, a few
notes relating to the synonymy and ranges of previously known species
being interspersed. The sources of the material examined are indi-
cated in each case and the methods of citation, etc., are so closely
similar to those of the earlier publications of the series as to require
no explanation at this point.

Stevia (§ Multiaristatae) andina, spec. nov., erecta herbacea
dense breviterque pilosa et superne glandulari-puberula 5 dm. vel
ultra altitudine ; caule tereti brevissime griseo-velutino usque ad in-
florescentiam folioso; foliis oppositis late ovatis obtusis crenatis ad
asin versus integris vel cuneatis et sessilibus vel subabrupte con-
tactis et in petiolum distinctum ca. 5-7 mm. longum cuneatim an-
gustatis; lamina 2.5-5 cm. longa 2.3-3.5 cm. lata membranacea supra
puberula subtus pallidiori praecipue in nervis venisque dense griseo-
pubescent supra basin 3-nervata; i i

longa et 2.5 dm. diametro glandulari-puberula; pedicellis maturis
plerumque involucro longioribus; capitulis 5-floris 1 cm. longis; in-
Volucri squamis lanceolato-linearibus acutis glandulari-puberulis 5-6

longis —Ecuapor: Province Loja: quebrada near Loja, Poortmann,
pop 138 (Par.), inflorescence only; between Loja and San Lucas, alt.
+00-2600 m., A. 8. Hiteheock, no. 21,475 (r¥PE, in Gray Herb., iso-
ype, N. Y.); near Tablon de Ofia, Rose, Pachano & Rose, no. 23,091
(Gr,); near Las Juntas, Rose, Pachano & Rose, no. 23,230 (Gr.).

’

4 ROBINSON

Peru: Chachapoyas, Mathews, no. 3033 (K.) and Mathews without
number.
This plant has been associated with S. crenata Benth. and is without
doubt a close relative of that species. However, the type material of
S. crenata, namely Hartweg’s no. 1089 from Guapulo near Quito has
narrower, less pubescent and much less definitely petioled leaves,
which are crenate only above the middle, while in S. andina the tooth-
ing begins distinctly below the middle. It is clear therefore that
these are readily recognizable as different plants and unless some
evidence of intergradation can be demonstrated they must be treated
as separate species. :

8. Bertholdii, spec. nov., herbacea perennis suberecta 4 dm.
alta longe patenterque albo-pilosa; caule tereti brunneo curvato-

paullo scabratis; pappo coroniformi e squamulis scariosis plus pong
connatis 0.4 mm. longis sistente omnino exaristatis.—EcUADOR:

lage of Sacaranga, Berthold Seemann, no. 658 (K., phot. Gr.). Mixed

wns.
: » Spec. nov., a basi plus minusve decumbente suberect#
herbacea perennis 3-5 dm. alta ; caule tereti flexuoso fo josissim¢
purpurascenti usque ad inflorescentiam simplici vel paucirame?
basin versus ca. 3 mm. diametro glabrato in parte media 0 )
et_brevissime griseo-puberulo (pilis minutis tenuissimis @

RECORDS OF THE EUPATORIEAE—IX 5

in parte superiori et inflorescentia minute cum glandulis brevissime
stipitatis ornato; internodiis inferioribus 1-3 cm. longis, mediis et
superioribus multo brevioribus; foliis (infimis oppositis ceteris plerisque
alternis) sessilibus lineari-lanceolatis acutis integerrimis vel ad apicem
versus obscure crenulatis a basi 3-nerviis subconcoloribus utrinque
subglabris et minute nigro-puncticulatis plerisque 2-3.5 em. longis et

5 mm. latis dense imbricatis; inflorescentia terminali fastigiatim
ramosa corymbosa planiuscula 3-6 cm. diametro multicapitulata
revissime et densissime glandulari-puberula; pedicellis inferioribus
usque ad 2 em. longis; capitulis ca. 10.5 mm. longis 5-floris; involucri
squamis lineari-lanceolatis 1(-3)-nerviis acutis vel acutiusculis stra-
mineo-viridibus dorso dense glandulari-puberulis ca. 7 mm. longis 1.4
mm. latis; corollae tubo ca. 1.5 mm. longo et faucibus paullo ampli-
atis 4 mm. longis purpureis; lobis limbi ca. 1.8 mm. longis ut videtur
pallidis ad apicem versus paullo hispidulis; achaeniis homomorphis
glaberrimis brunneo-stramineis 3-3.8 mm. longis quoque aristis 20-24 / ~~
barbellatis ca. 7 mm. longis munito.—ARGENTINE REPUBLIC: Prov.

~,

i

smaller heads and practically glabrous foliage.

8. eclipes, spec. nov., ut videtur herbacea perennis patenter villosa
vel subsetulosa; caule subrobusto usque ad 4-5 mm. diametro fistuloso
flexuoso subtereti post exsiccationem paullo angulato-costato viridi
vel purpureo-tincto; internodiis plerisque 6-9 mm. longis; foliis

6 ROBINSON

longo; faucibus 3 mm. longis; adelphocarpiis coronula annulari brevis-
sima ornatis et cum aristulis 1-2 brevissimis (vix 0.2 mm. longis)
munitis; idiocarpio sine coronula sed aristulis 1-2 etiam plus reductis
munito.—Bo.ivia: Dept. Tarija; Pinos near Tarija, alt. 2300 m.,
Mar. 10, 1904, K. Fiebrig, no. 3126 (rypE in Gray Herb., isotype,
Field Mus.).

This plant was distributed as “Stevia cf. Schultzii Hieron.” S.
Schultzii Hieron., it may here be explained, was founded on two
elements, namely a plant of Tucuman coinciding with the much
earlier S. breviaristata H. & A. and a specimen of Mandon’s from the
Sorata region which proves identical with the plant on which Rusby
had already founded his S. boliviensis. To neither of these elements
does the southern Bolivian plant here in question make any close
approach.

S. ELaTIOR HBK. The varieties of this species, as they occur
north of the isthmus of Panama, have been already discussed (Con-
trib. Gray Herb. xe. 13, 97-98 (1930), and consisted, as may be here
recalled of (1) Var. TYPICA, with pauci-aristate adelphocarps and sub-
sessile leaves; (2) Var. PopopHYLLA, with pauci-aristate adelphocarps
and conspicuously petiolate leaves; and (3) Var. pissouuta, with all
its achenes exaristate.

'S most southern and so far as we now know quite isolated manr
festation of the species may be put on record as follows:

RECORDS OF THE EUPATORIEAE—IX 7

Var. austrina, var. nov., laxe decumbens quam var. typica paullo
gracilior et hirsutior; involucri squamis extimis dorso hirsutis, pilis
plerisque longis albis articulatis attenuatis maturitate eglandulosis;
achaeniis 4 (adelphocarpiis) quoque 5-aristatis; achaenio quinto
(idiocarpio) exaristato.—S. elatior (as det. by Hieron.) in Ktze., Rev.
Gen. iii. pt. 2, 180 (1898), not precisely HBK.—Botivia: Dept.
Santa Cruz: Sierra de Santa Cruz, 1600 m., May, 1892, Dr. O. Kuntze
(type, N. Y., isotype, U. S., phot. Gr.).

8. Hilarii, spec. nov., herbacea erecta gracilis perennis 3-4 dm.
alta; caule tereti brunneo brevissime puberulo foliosissimo 2-2.5 mm.
diametro; internodiis plerisque 5-12 mm. longis; foliis oppositis ses-
silibus patentibus lanceolatis vel anguste oblongis subacutis crenato-
serratis basi integris cuneatis utrinque glabris (sed supra glandulis
sessilibus adsparsis) subtus paullo pallidioribus nigro-punctatis fere
a basi trinervatis; foliis caulinis 1.6—2 cm. longis 6-9 mm. latis plerisque
in axillis proliferis; foliis rameis multo minoribus saepe integris ali-
quando alternis; inflorescentia terminali glandulari-puberula laxius-
cula; pedicellis maturis gracillimis usque ad 1 cm. longis; capitulis
S-floris 1 em. longis saepe nutantibus; involucri squamis oblongis
acutis pilosiusculis et brevissime glandulari-puberulis ca. 6 mm. longis
et 1.5 mm. latis saepius ad apicem versus purpureo-tinctis; corollis
ut videtur rubro-purpureis; tubo proprio ca. 1.5 mm. longo in fauces
ampliatas subaequilongas gradatim transienti; dentibus limbi ca.
2mm. longis et 1 mm. latis apice rotundatis; achaeniis 3.6 mm. longis
fere isomorphis quoque cum aristis 6-8 ca. 5 mm. longis et cum squam-
ulis paucis brevissimis coronato.—BraziL: Minas Geraes, 1816-1821,
A, de Saint-Hilaire, no. 1023 Cat. D. no. 275 (TYPE, Par., phot. and
fragm. Gr.); also Cat. C! no. 367 (Par.).

This plant is attractive on account of its slender erect and evenly
leafy stems which are virgate almost to the inflorescence. The leaves
are mostly proliferous in the axils and at anthesis those on the un-
developed secondary axes are about half the length of the cauline,
Hus giving the plant somewhat the look of bearing leaves of alternating
‘es. The species is manifestly close to the real S. veronicae DC.,
but differs in its greater stature, shorter internodes, proliferous axils,

rter pubescence and much more open inflorescence.

8. kuhnioides Rusby in herb., herbacea perennis gracilis suberecta
. alta; caule tereti flexuoso vix 1.5 mm. diam«
‘tropurpureo breviter crispeque piloso; internodiis plerisque 1-2.5
‘Mm. longis; foliis oppositis distincte petiolatis ovatis acutis crenato-
‘errulatis basi integris abrupte contractis deinde cuneatis utrinque
Pubescentibus subtus paullo pallidioribus, lamina 1.7-2.3 cm. longa

8 ROBINSON

1-1.5 em. lata, textura membranacea; petiolo 4-5 mm. longo sursum
paullo alato; glomerulis 2-3 densis, terminali breviter, lateralibus
longius pedunculatis; capitulis 7-10 mm. longis sessilibus; involueri
squamis lanceolatis acutis vinaceis dorso subpatenter vel crispe pilo-
siusculis; pilis albis saltim maturitate eglandularibus; corollae tubo
et faucibus purpureis, lobis limbi albidis pilosis; achaeniis heteromor-
phis; adelphocarpiis plerisque 2-3-aristatis et cum squamulis paucis
coronatis; aristis inaequalibus, longioribus corollas subaequantibus;
idiocarpio cum squamis solis coronato exaristato.—Bo.rvia: Cor-
dillera Real, on the road from Okara to Ancoma, alt. about 3350 m.,
Apr. 29, 1926, G. H. H. Tate, no. 870 (type, N. Y. Bot. Gard., phot.
Gr.).

This plant, as yet known only from rather imperfect material,
collected by the Ladew Expedition and presented to the New York
Botanical Garden by the American Museum of Natural History, was
most kindly placed at the disposal of the writer for inclusion in the
present paper by Dr. H. H. Rusby who had already detected its
novelty.

niis 1-3-aristatis 3.6 mm. longis; aristis aliis corollam subaequantibus,
aliis in aristulas multo breviores deminutis.—PERU: Dept. Huanuco-
southwestern slope, alt. about 2750 m., July 8-22, 1922, M:
Featherstone, no. 1382 (tyPE, Field Mus., isotype Gr.). |

In the heads examined this species showed a moderate and eet
heterocar py, thus the awns on the different achenes of the same hea
were in number about as follows: 3-2-114-114-114. The pet
said to grow in dense masses and to be known by the vern
name “utcuisha.” Its fine chalky-white sessile glands, if comet?
should give a ready mark of recognition.

RECORDS OF THE EUPATORIEAE—IX 9

8. Macbridei, spec. nov., suffruticosa erecta usque ad 1.3 m. alta
saltim aliquando ab caudice erecto lignoso ramoso pluricaulis; caulibus
teretibus ad basin versus lignescentibus usque a mm. diametro
glabrescentibus superne adscendenter ramosis atropurpureis foliosis
puberulis; foliis oppositis et (praecipue supra mediam partem) alternis
anguste ovatis acutiusculis vel obtusis crenatis vel crenato-serratis
ad basin versus integris cuneatis breviter petiolatis vel (rameis) sessili-
bus chartaceo-subcoriaceis supra brevissime griseo-puberulis subtus
multo pallidioribus subalbescentibus glaberrimis vel (juventate)
obsolete in nervis puberulis fere a basi 3-nervatis 3-4 cm. longis 1-2
cm. latis; petiolo ca. 4 mm. longo; foliis rameis multo minoribus ob-
longo- vel lanceolato-linearibus; capitulis ca. 12 mm. longis sessilibus
in glomerulos corymbosos aggregatis; inflorescentia densissime gland-
ulari-puberula; involucri squamis linearibus subacutis ca. 7 mm.
longis atrorubris dense glandulari-puberulis; corollae tubo proprio
gracili 1.5 mm. longo glandulari-puberulo; faucibus rubris 4.5 mm.
longis; dentibus limbi 1.5-2 mm. longis albis dorso paullo hispidulis;
achaeniis vel homomorphis et pappo coronulato exaristato ornatis
vel rarius heteromorphis et multiaristatis.

A Peruvian species with two sharply marked pappus-varieties,
name]

Var. a. typica, homocarpa; achaeniis omnibus exaristatis solum
‘um pappo brevi squamuloso 0.2-0.4 mm. longo coronatis.—PERU:
Dept. Huanuco: on “ grass-shrub slope,” Mito, alt. about 2750 m.,
July 8-22, 1922, Macbride & Featherstone, no. 1613 (TYPE, Field Mus.,
‘sotype Gr.); Dept. Lima: steep grassy northwestern slope, Rio Blanco,
alt. about 3660 m., May 8-19, 1922, Macbride & Featherstone, no. 663
(Field Mus., Gr.); also in somewhat lower and more slender form (as

from somewhat less favorable environment) but without essential
alteration of character, and unfortunately quite without record of
og 1839-1840, Gay, no. 355 (Par., phot. and fragm. Gr

. :

vam corolla multo brevioribus 1-3 mm. longis instructo.—PERu:
without more exact locality, 1839-1840, Gay, no. 379 (Par., phot. and
. Gr).

Awnless forms or varieties of the normally few-awned Stevias are
‘ather numerous and have often been recorded. It is much less usual
a an awnless form of a normally multiaristate species. In this

“Tuvian species we seem to have a manifestation of reverse tendency,
'8 to say an occasional multiaristate development on the part

of @ species usually exaristate.

10 ROBINSON

S. ophiomaches, spec. nov., suberecta vel curvato-adscendens
subherbacea vel ad basin versus lignescens dense breviterque incurvo-
pilosa infra mediam partem saepius simplex et delapsu foliorum mox
denudata supra patenter ramosa et foliosa 4—6 dm. alta; radice fibrosa;
caule tereti brunneo usque ad 4 mm. diametro; internodiis plerisque
1-2 em. longis; foliis oppositis oblanceolatis utroque angustatis ad
apicem verum rotundatis ad basin versus in petiolum paullo mar-
ginatum 4-7 mm. longum gradatim angustatis supra mediam partem
obscure crenatis a basi 3-nervatis utrinque breviter pubescentibus
textura membranaceis; lamina plerumque 1.5-3.5 em. longa 5-8 mm.
lata; corymbis compositis terminalibus et lateralibus convexis incurvo-
puberulis; capitulis pro genere parvis plerisque 7-8 mm. longis 5-
floris breviter pedicellatis; involucri squamis substramineis breviter
acutatis 4-5 mm. longis dorso cum pilis brevibus incurvis eglandulari-
bus et cum glandulis paucis sessilibus ornatis; corollis albis sublaevi-
bus ex involucro bene exsertis; achaeniis heteromorphis ca. 2 mm.
longis; adelphocarpiis 4, quoque cum aristis corollam aequantibus
3-5 et cum coronula vix 0.2 mm. alta munitis; idiocarpio coronulato
exaristato.—Co.LomB1a: Dept. Norte de Santander: Prov. of Ocaia:
open space by roadside, Santa Maria Chica, 16 Nov. 1879, Kalbreyer,
no. 1240 (TPE, in herb. Royal Gard. Kew, phot. and small fragm.
Gr., isotype Brl.); also in open sunny dry spaces, common and some-
what glutinous, alt. 2125-2750 m., Jan. 1878, Kalbreyer, no. 605 (kK,
phot. Gr., Brl.). Flowers said to be fragrant. Plant said to be
employed medically as an antidote for snake-bite—hence the dis-
tinctive if somewhat fanciful specific name. The species is clearly
related to the widely distributed S. rhombifolia HBK. and might at
first sight be thought merely a stunted paramo form of that plant,
but its relatively narrower as well as much smaller leaves and more
stiffly spreading branches give the Colombian species here described
a characteristic and individual habit and suggest that it is to be Te
garded as a separate specific entity. :

S. PALLIDA (Sch.-Bip.) Hieron. and S. PURPURASCENS (Sch.-Bip:)
Hieron. in Engl. Bot. Jahrb. xxi. 328 (1895). These binomials, 8°
obscurely put forward as to escape the notice of the compilers ae
Index Kewensis, resulted from an attempt on the part of Hite
to give specific status to two alleged plant-entities doubtfully attach
by Schultz (Linnaea xxv. 275) to S. elongata HBK. Somewhat pre”
longed effort to check, compare, and differentiate more clearly the
tangled elements here involved in the work respectively of Kunth,
Schultz, and Hieronymus, has led the writer to the following coneli-
sions.

RECORDS OF THE EUPATORIEAE—IX 11

S. elongata HBK. is duly shown by a somewhat fragmentary but
happily still extant and convincingly authentic specimen in the Hum-
boldt and Bonpland series at Paris. This specimen shows that the
species was somewhat misleadingly characterized from the start, for
its branches, while in some cases alternate, as stated in the original
description, are at other nodes opposite, and the “flores” [doubtless
used for heads], stated by Kunth to be “longe pedunculati,” are in
fact subsessile or shortly pedicelled in rather dense glomerules. The
latter, however, are in fact rather conspicuously stalked and it may
have been this fact that suggested the “longe pedunculati,”’ an ex-
pression perhaps inadvertently or by clerical error applied to the
“flores. ’’

When Schultz drew up his revision of Stevia he merely mentioned
S. elongata and made no effort to characterize it. Indeed, he gave no
evidence that he had been able to examine the typical form. How-
ever, he appended to the species a “8. Moritziana Schultz Bip. an
nova species?’”’ Close comparison of the described features of this
var. Moritziana and of the original description of S. elongata discloses
no differential trait of significance that was not founded upon some
misapprehension or which does not break down when the material
M question is critically examined.

Under this by no means satisfactorily separated “8. Moritiziana”
Schultz further indicated two minor phases, which he called “I. Pal-
lida” and “IT. Purpurascens” without assigning them to any precise
taxonomic rank. Of these, the former was said to be less glandulif-
erous and to have the phyllaries and flowers pale, while the latter was
described as strongly glandular and said to have the upper part of the
Phyllaries and the flowers “tantum ad tubi partem superiorem”’
purplish

Though the original S. elongata HBK. is stated to have come from
‘entral Colombia, both forms of Schultz’s “8. Moritziana” came from
Western Venezuela, having been collected by Moritz and by Linden
in the mountains of Mérida. When Hieronymus in 1895 was describ-
ng the Bolivian S. Stuebelii he made incidental mention of Schultz’s
‘wo forms of his “8. Moritziana” and though adding no clear or con-
“Meing differential traits raised them each to specific rank.

Ae writer has studied the pertinent material of this aggregate at
Paris, Berlin, and Kew and is forced to conclude that the first cited
Part of Schultz’s “I. Pallida”’ (namely Moritz no. 1376) and the first
mentioned element in his “II. Purpurascens” (namely Moritz no.
7376b) do not differ in any essential point from the typical S. elongata

In all three of these plants the inflorescence and dorsal sur-

12 ROBINSON

face of the rather gradually pointed phyllaries are closely beset with
stipitate glands. :

Under his “I. Pallida,” however, Schultz mentioned a second speci-
men—a plant cited thus “Venezuela, prov. Merida, Culata, 9000’
Septbr. 1842 Linden (flores albi).”” In the Schultz herbarium now
at the Museum of Natural History in Paris there is a small specimen of
Linden’s no, 483, referred to his subcategory pallida and bearing pre-
cisely the data cited. This plant, however, differs from the Moritz
specimens nos. 1376 and 1376b in having the inflorescence less glandu-
lar and the phyllaries relatively broader and more abruptly pointed
and furthermore beset on their dorsal surface with a rather sparse
incurved or subappressed pilosity, the individual hairs being delicate,
jointed, attenuate and glandless but being interspersed with a few
scattered sessile or subsessile glands. It is to be inferred that this
plant of Linden’s was the one which formed the basis of Schultz’s
brief description of his “I. Pallida’” and which presumably suggested
its name.

Thus it will be seen that the name pallida as here employed by
Schultz cannot be maintained on the basis of its first-mentioned speci-
men, Moritz no. 1376, for that proves to be in fact the more glandular
form and to have colored corollas and finally to be indistinguishable
from the much earlier S. elongata HBK. Thus, if pallida is here
to be maintained in any status it must be taken as resting upon the
second-mentioned element, namely Linden’s specimen bearing i
Schultz’s herbarium the identical data cited as follows “Venezuela,
prov. Merida, Culata, 9000’ Septbr. 1842, Linden (flores albi)” and
further bearing at Paris the number 483. This conclusion is reach
for the double reason that this plant of Linden’s, being less glandular
and having paler flowers, best agrees with Schultz’s brief character
and the name he selected, while furthermore it is the sole remaming
element after the extraneous element, Moritz no. 1376, has been fe
moved to an earlier named plant-entity.

Unhappily, the writer when examining (necessarily with some
haste) the respresentation of Stevia in the Botanical Museum at
Berlin-Dahlem in the summer of 1927 was unable to find any precise
equivalent of the Linden plant numbered 483 at Paris. Thus .
would appear probable that Hieronymus, when raising Schultz's
pallida to specific rank did not have before him its typical —
This inference is strengthened by the fact that in the very brief :
criptive phrases which Hieronymus applies to his S. pallida (
Bip.) Hieron. some do not apply to the plant shown by Linden’s 2°
483. Thus he speaks of the stems as being subsimple and the leaves

RECORDS OF THE EUPATORIEAE—IX 13

subentire or but remotely serrate, while the leaves of the Linden speci-
men at Paris are very clearly and by no means remotely serrate and
the plant proves to be one which branches rather freely.

However, it is to be remembered that the name pallida of Schultz
must be applied solely with reference to the identity of its own type
and no confusion of extraneous material during the transfer of this
name to new position or rank can be allowed to effect any change in
the plant-identity covered by the name.

Though the plant represented by Linden no. 483 does not appear to
have been subsequently collected in Venezuela, excellent and copious
specimens of what seems to be identical material have been collected
ala years by Pennell, Killip and others in various parts of Colom-

ia.

To summarize, it may be said that, so far as the author has been
able to ascertain, S. purpurascens (Sch.-Bip.) Hieron. falls into the
synonymy of S. elongata HBK., while S. pallida (Sch.-Bip.) Hieron.
(as determined by its actual type and not by extraneous material
studied at the time of its transfer to specific rank) may be appropri-
ately used for the plant of both Venezuela and Colombia somewhat
fully described under this name on page 34 of the present paper.

S. PUBERULA Hook. Bot. Misc. ii. 225 (1831). This Peruvian
species, as originally described, was stated to have a pappus of about
ll awns. When last at Kew, the writer was most kindly permitted
to dissect a head from the type material. This proved anisocarpic,
four of the achenes being each 9-10-awned and the fifth being provided
merely with a scale-pappus and two short awnlets not a quarter the
length of the awns in the other florets. Through the kindness of the
director of the Museum of Natural History at Paris the writer has
been able to borrow and examine with greater care two sheets of this
Species collected by Martinet at Matucana, Peru. In these some of

e heads proved essentially isocarpic while others, adjacent on the
same inflorescence, were found to be strongly anisocarpic. It is there-
fore evident that heterocarpy is here a mere tendency, sometimes
manifesting itself partially. No better proof is needed that as a differ-
“ntial trait heterocarpy should be employed in Stevia only with the
sreatest care,

8. Purdiei, spec. nov., herbacea perennis suberecta vel curvato-
adscendens 4-6 dm. vel ultra altitudine breviter crispeque pilosa a
media parte adscendenter ramosa; radice e fibris longis fuscis tenuibus
Sistente:; caule gracili tereti rubro-brunneo ca. 2 mm. diametro; inter-
nods 1-3.5 em. longis; ramis patenter adscendentibus plerisque sim~
Plicibus 1-1.5 dm. longis; foliis oppositis (supremis paucis alternis)

14 ROBINSON

sessilibus lanceolatis vel anguste subovatis utroque acutis supra med-
iam partem inconspicue crenato-serratis supra pilis albis sparsis basi
incrassatis instructis subtus paullo pallidioribus punctatis in nervis
venisque crispe pilosis a basi 3-nervatis 2-2.5 cm. longis 5-8 mm.
latis textura membranaceis; caulinis saepe in axillis proliferis; rameis
minoribus subintegris; glomerulis densis 1-2.5 cm. diametro in caule
ramisque terminalibus; capitulis subsessilibus 10-11 mm. longis 5-
floris; involucri squamis lanceolato-linearibus attenuatis peracutis
dorso cum glandulis breviter stipatatis numerosis et cum pilis longiori-
bus albis crispis eglandulosis ornatis ca. 7.5 mm. longis 1.2 mm. latis;
corollae tubo proprio ca. 1 mm. longo; faucibus cylindratis 3.5 mm.
longis glabris ut videtur atropurpureis; dentibus limbi pallidis paullo
hispidulis; achaeniis gracilibus ca. 4 mm. longis; pappo coroniformi

: mm. alto scarioso margine eroso-denticulato.—CoLOMBIA:
Dept. Magdalena: St. Miguel, Nevada de Santa Marta, Nov. 1844,
William Purdie (rypr, in herb. Royal Gard. Kew, phot. Gr.); identical
material on a second sheet in the herbarium at Kew bears the data,
Purdie, St. Martha, 1845, and a third sheet in the Gray Herbarium
bears like material said to have been collected by Purdie at Santa
Marta in 1849, Possibly all three are parts of the same collection,
of which the date was somewhat obscurely written and may have been
erroneously copied. One of the sheets in the Kew herbarium is anno-
tated in hand of Bentham as “Stevia sp. nov.,” but though thus
correctly interpreted the plant does not appear to have received a name
even In manuscript.

RECORDS OF THE EUPATORIEAE—IX 15

homomorphis 3.5 mm. longis laeviusculis; aristis pappi 19-22 albis
vel ochroleucis paullulo barbellatis 6.5 mm. longis—Uruauay: sands
of the Rio Negro opposite Mercedes, April 1867, C. Fruchart, no. 281.

Specimens of this number have long lain undetermined in the her-
baria at Kew and Paris. Persistent efforts have failed to place them
satisfactorily in any hitherto described species. Though the material
available for study is of slight extent it is sufficient to permit adequate
diagnosis and therefore seems to merit published record. The ver
sharply incurved-serrate obovate-elliptical leaves are quite distinctive.
The sheet of this species at the herbarium of the Royal Botanic Gar-
dens at Kew may be regarded as the type.

8. (§ Multiaristatae) samaipatensis, spec. nov., herbacea per-
ennis erecta virgata habitu S. serratac similis 2-3 dm. alta; caule tereti
2mm. diametro minute incurvo-puberulo brunnescenti; foliis oppositis
et alternis sessilibus saepe ab axillo proliferis anguste oblanceolatis
obtusis ad apicem versus crenato-serratis basi longe attenuatis inte-
gris textura subcoriaceis utrinque nigro-puncticulatis inconspicue
puberulis 1.6-1.9 cm. longis 3-4 mm. latis; corymbo terminali fasti-
giatim ramoso denso planiusculo 4-5 cm. diametro; capitulis sessilibus
vel brevissime pedicellatis 8-9 mm. longis 5-floris; involucri squamis
anguste oblongis abrupte acutatis incurvo-puberulis saepe purpureo-
tnetis ca. 4.5 mm. longis; corollis in tubo et faucibus atropurpureis
puberulis; dentibus limbi albis vel albidis; achaeniis heteromorphis
‘a. <./ mm. longis; adelphocarpiis 9-11-aristatis; aristis ca. 5 mm.
longis corollas subaequantibus ad basin versus scarioso-dilatatis ;
idiocarpio coronula e squamis ca. 0.7 mm. longis plus minusve connatis
omnato exaristato.—Boxivia: Dept. Santa Cruz: Prov. Vallegrande:
Samaipata, alt. 2000 m., 16 Mar. 1920, J. Steinbach, no. 3761 (TYPE
m Bot. Mus.. Berlin-Dahlem, phot. and small fragm. Gr.).

From the habitally similar S. serrata of more northern latitudes

lant may be readily distinguished by its thicker leaves and
multiaristate adelphocarps. It is doubtless more nearly related to
S. mercedensis, but is clearly distinct, having slightly smaller heads,
shorter achenes, more virgate habit, thicker leaves and denser
corymb.,

8. grandidentata Sch.-Bip. var. subeglandulosa Hieron. in Engl. Bot.

tt in 1884 for a plant of Mexico. Renaming the Bolivian plant
‘S'S. soratensis, Hieronymus (I. c. xxviii. 560) failed to transfer his

16 . ROBINSON

already indicated var. subeglandulosa. This is here given what is
believed to be its correct nomenclatural status.

8. totorensis, spec. nov., herbacea erecta perennis 1.5—-4 dm. alta;
caule tereti gracili vix 2 mm. crasso folioso brunneo minute scab-
riusculo-puberulo; internodiis plerisque 1-2.5 cm. longis; foliis op-
positis sessilibus conduplicatis recurvatis ovatis acutis serratis basi
acutis utrinque glabriusculis exserto-reticulatis textura coriaceis
2.5-3.2 em. longis 1.5-2 cm. latis, summis minoribus subintegris;
glomerulis paucis subdensis in corymbo planiusculo dispositis; capi-
tulis 5-floris ca. 1 cm. longis sessilibus vel subsessilibus; involucri
squamis ad apicem versus violaceis acutis rigidiusculis minute granu-
latis (cum glandulis brevissime stipitatis dense ornatis); corollae tubo
proprio gracili et faucibus 3.7 mm. longis subcylindratis rubro-vio-
laceis glandulis minutis subsessilibus munitis; achaeniis heteromorphis
2.3-2.7 mm. longis; adelphocarpiis 4 quoque 3-5-aristato et cum
squamulis etiam coronato; aristis purpureis vel purpurascentibus
corollas subaequantibus; idiocarpio cum squamulis et 2-3 aristulis
multo brevioribus munito.—Bourvtia: Dept. Cochabamba: Prov.
Totora: Bucona, alt. 3000 m., Mar. 28, 1920, J. Steinbach, no. 3953
(TyPE in Bot. Mus. Berlin, phot. and small fragm. Gr.)

This Bolivian plant recalls in some respects S. myriadenia Sch.-
Bip. of Brazil, but has more closely netted leaves, which are almost
sessile instead of distinctly petioled, denser inflorescence, and some
other differences rendering it pretty clearly a distinct species.

S. uRticaEFouIA Billb. in Thunb. Pl. Bras. Dec. I. 13 (181%).
In the Brazilian specimens of this species which have been examined
and which may be presumed to represent the typical form of the plan t
the corolla-throat is a fairly deep purple. A plant of the Bolivian
uplands closely resembles this Brazilian form in habit, foliage, i-
florescence and pappus, but differs decidedly in having the throat of the
corollas much paler and of a yellowish or greenish color, This may
be put on record as follows: ;

Var. pallidifiora, var. nov., corollae faucibus pallidis lutescentibus
vel viridescentibus; aliter var. typicae simillima.—Listed by Rusby
Bull. N. Y. Bot. Gard. iv. 377, as S. urticifolia Thunb.—Bouvs:
without stated locality, but probably from the Yungas region, Bang,

x U8

RECORDS OF THE EUPATORIEAE—IX 17

subalternis) sessilibus utroque acutis basi integra cuneata excepta
serratis utrinque parce pilosis subconcoloribus a basi trinervatis
textura membranaceis 2-3 cm. longis 6-10 mm. latis; inflorescentia
subirregulariter ramosis; glomerulis multi- vel pauci-capitulatis densis
vel laxiusculis; capitulis 16 mm. longis 5-floris sessilibus vel breviter
pedicellatis; involucri squamis anguste oblongis breviter acutatis
8-8.5 mm. longis 1.8 mm. latis, extimis dorsaliter glandulari-puberulis;
glandulis minimis stipitatis; corollae tubo proprio ca. 1.5 mm. longo;
faucibus cylindratis 4 mm. longis roseis; dentibus limbi albis vel
pallide roseis oblongis 2.5 mm. longis; achaeniis heteromorphis,
eorum quattuor (adelphocarpiis) cum aristis stramineis barbellatis 10-
16 apice munitis; achaenio quinto (idiocarpio) solum squamis brevi-
bus scariosis coronulato.—PERv: Carumas, Prov. Moquegua, alt.
3200 m., Feb. 21—Mar. 6, 1925, Dr. A. Weberbauer, no. 7293 (TYPE,
Field Mus., isotype, Gr.).

From the mostly erect and essentially herbaceous S. puberula
Hook. this species differs in its lanceolate rather than oval leaves,
its somewhat trailing habit and distinctly woody though slender
stem, also in its more irregular inflorescence. From S. Bangii Rusby
and §. tapacariensis Hieron. the plant here described differs in its
larger heads and glandular-puberulent phyllaries, as well as in its
more woody stem.

FLEIscHMANNIA URENIFoLIA (Hook. & Arn.) Benth. & Hook. f.
€x Hemsl. Biol. Cent.-Am. Bot. ii. 91 (1881) as urenaefolia. This

exican species is found to vary remarkably in several ways. Its
pubescence js singularly inconstant on different codrdinate parts
of the same individual. Even more striking differences are shown
in the contour and lobing of the leaves. Even in the type specimen
some of the leaves are but slightly lobed while others are trifoliolate
to the rachis. In a notable collection of plants secured by Mrs. Ynes
Mexia in little known parts of Jalisco two very unlike plants of this
sfnus were distributed under nos. 1684 and 1684-a. At first these
Were taken to be distinct species but on closer study it seems best to

t them as follows:

Var. %. typica, foliis paucis vel nullis indivisis ovatis vel ovato-
oblongis paullo trilobatis et dentatis basi acutiusculis vel rotundatis
lo Plerisque ad rachin trifoliolatis cum foliolis dentatis et saepe
Hoes Phania ? urenifolia Hook. & Arn. Bot. Beech. 297 (1840).

netst urenifolia Walp. Rep. vi. 106 (1846-7). Helogyne
werifolia Walp. Rep. iv. 457 (1848) by typographical error. Phania
ota Benth. & Hook. f. Gen. Pl. ii. 243 (1873). Fleischmannia
“vuaefolia Benth. & Hook. f. ex Hemsl. Biol. Cent.-Am. Bot. ii. 91

18 ROBINSON

(1881). F. Langlassei Robinson, Proc. Am. Acad. xli. 273 (1905).—
Jatisco: wet cliffs, Barranca of Beltran, Pringle, no. 5492 (Gr.);
steep wooded slope, inclined to be prostrate, Hacienda del Ototal,
east of San Sebastian, Sierra Madre Mountains, Mezia, no. 1684
(Gr.). Michoacan or GUERRERO: on moist rocks, Arroyo de Barabas,
alt. 1000 m., Langlassé, no. 27 (Gr.). Cutapas: on rocks along creek
east of Montserrate, Purpus, no. 218 (Gr.). STaTE UNKNOWN:
Beechy (K., phot. Gr.)

Var. 8. Mexiae, var. nov., foliis omnibus late ovatis vel suborbicu-
laribus indivisis inciso-lobulatis et dentatis basi rotundatis vel sae-
pius cordatis. Mexico: Jalisco: herb growing in crevices of roc
near water, steep wooded slope of narrow cafion, Hacienda del Ototal,
east of San Sebastian, alt. 1500 m., Mrs. Ynes Mezia, no. 1
(TYPE, in Gray Herb.).

No floral differences have been detected between these in foliage
strikingly unlike plants. However, the outer phyllaries in var.
Meviae are a little narrower and more acute and the leaves generally
more pubescent than in var. typica. 4

Symphyopappus leptophlebius, spec. nov., fruticosus glabert-

e runnels

RECORDS OF THE EUPATORIEAE—IX 19

On the other hand its inflorescence instead of being, as in most species
of Symphyopappus corymbose, is subracemosely paniculate as in
Eupatorium crenulatum Spreng. The pappus-bristles are united at
base into a distinct ring, which sometimes disarticulates from the
summit of the achene. Therefore the plant must be referred to
Symphyopappus rather than to Eupatorium if these two very closely
telated genera are kept apart.

8. Lymansmithii, spec. nov., fruticosus uni- vel pluri-caulis 2-3
m. altus glaber lucidus vernicosus viscidus; caulibus erectis ad-
scendenter ramosis subteretibus; ramis ramulisque purpureo-brunneis;
foliis oppositis petiolatis ovato-oblongis utrog subabrupte angustatis
apice obtusis vel rotundatis basi acuminate cuneata excepta grosse
serratis pinnatim 5-nervatis utrinque glabris supra lucidissimis subtus
vix pallidioribus sublucidis textura chartaceo-subcoriaceis 5-8 cm.
longis 2.5-4.3 em. latis ; petiolo 8-12 mm. longis; corymbis compositis
planiusculis terminalibus plerisque 1-2 dm. diametro; capitulis 5-
floris ca. 1 cm. longis 3-4 mm. diametro; involucri cylindrati 3-4-
seriatim imbricati squamis ca. 15 (saepe in seriebus 5 erectis dispositis)
obtusis pallidis ciliolatis dorso glabris pulcherrime 3-5-nervatis

on campo, at the Biological Station in Alto da Serra, alt. 800-900 m.,
lat. 28° 47’ South, long. 46° 19’ West, Feb. 15, 1929, Dr. Lyman B.
Smith, no. 1914 (rypr, in Gray Herb.).
species as to its heads rather closely simulates S. compressus
(Gar dn.) Robinson but has somewhat flatter corymbs and quite differ-
‘nt foliage, its leaves being shorter, relatively broader (usually
‘bout half as wide as long), much more obtuse, fewer-toothed and in
general more glutinous. From S. cuneatus (DC.) Bak. this new
Species differs in its thinner leaves of ovate rather than obovate
tendency, less bluntly toothed, less gradually cuneate at base and
~' Much less conspicuous reticulation. It is a pleasure to name
this striking shrub for its discoverer, who during a recent journey to
Brazil, undertaken primarily to obtain further knowledge of his
‘Pecialt » the Bromeliaceae, found opportunity to gather helpful
material in many other groups.
Eupatorium angulare, nom. nov. LE. fistulosum Robinson,
- Bost. Soc. Nat. Hist. xxxi. 249 (1904), not Barratt [in a leafiet
to accompany exsiccatae, and] in Wood, Class-book, ed. 2, 314 (1847).

20 ROBINSON

The earlier homonym, obscurely put forth and omitted from the body
of the Index Kewensis, was overlooked by the writer when naming
the Costa Rican plant in 1904. A serious revival of Barratt’s name has
not seemed probable and until now there has appeared to be no special
reason to rename the Central American plant bearing the later hom-
onym. However, the recent ruling regarding homonyms by the
International Congress at Cambridge, though reached contrary to
the advice and best judgment of many experienced taxonomists,
seems to place the obligation to rename all later homonyms rather
than to await necessity as some few of the long discarded earlier ones
may from time to time be revived. This decision is one of several
popular with nomenclatural reformers in which an imposition of rules
produces greater change than could have arisen had the matter been
left to usage. The new name here chosen alludes, of course, to the
unusual form of the palmately angled leaves.

E. cERASIFOLIUM (Sch.-Bip) Bak. in Mart. Fl. Bras. vi. pt. 2, 308
(1876). Conoclinium cerasifolium Sch.-Bip. ex Bak. 1. ¢. in synon.
This species was founded on a plant collected by Spruce at San Gabriel
da Cachoeira near the Cataracts of the Rio Negro. No further mate-
rial of this species has been found by the writer in any of the herbaria
which he has examined and no rediscovery of the plant has been re-
ported so far as known to the writer. It therefore seems well wo
while to give published record to a second station for this little known
species, as follows: in open [places], at Cucuhy, Rio Negro, State
Amazonas, alt. 120 m., Feb. 1930, Holt & Gehriger, no. 341 (Gr.).
This new station appears to be about 125 km. north of the type
locality. The specimen corresponds closely in all important details
with Baker’s description and with a photograph (Gr.) of the type
(K.). The involucre closely resembles that of E. macrophyllum L.
of § Hebeclinium but the disk is neither hairy as in this section no
much elevated in the middle as in § Conoclinium with which the spec
was associated by Schultz. On technical grounds it should be
ferred to § Subimbricata.

E. conntvens Rusby, Mem. Torr. Bot. Club, vi. pt. 1, 57 (1896).
The range of this plant, hitherto known only from Bolivia, ca? now
be extended to central Peru for a specimen recently submitted PY
the United States National Herbarium to the writer for study and
identification proves to agree closely with Bolivian material am
well marked species but comes from Peru and has the following da
of collection: shrub 9-15 dm. high with lavender florets, grow!DS va
rocky cliff, at Huanucayo, Prov. Junin, alt. 3400 m., Apr- 13 ee
Paul Ledig, no. 51 (Gr.).

RECORDS OF THE EUPATORIEAE—IX 21

E. (§ Eximbricata) lasiolepis, spec. nov., erectum herbaceum
molliter breviterque piloso-subvelutinum griseum; caule tereti 4 dm.
vel ultra altitudine paullo flexuoso albo-medulloso virgato; foliis al-
tenis parvis saepe deflexis oblanceolato-oblongis obtusis dentatis
vel crenulatis vel (supremis) subintegris ad basin versus cuneatis et
integris sessilibus saepe in axillis proliferis utrinque subvelutinis
unicostatis (cum vel absque nervis lateralibus obscuris) ca. 1.8-2.4
cm. longis 5-7 mm. latis; inflorescentia corymbosa planiuscula e
ramis alternis saepe elongatis cum foliis paucis parvis remotis instructis
prope apicem solum capituliferis sistente, axe principi abortivo;
bracteis linearibus; ramulis flexuosis; capitulis pedicellatis ca. 18-
floris ca. 1 cm. altis ad apices ramulorum inflorescentiae in corym-
bulos paullo convexos densiusculos aggregatis; involucri squamis ca.
13 oblanceolato-oblongis maturitate vix imbricatis apice rotundatis
saltim supra mediam partem densissime velutinis fulvido-griseis ca.
7mm. longis; corollis tubulatis sursum paullo ampliatis 5 mm. longis
ut videtur pallidis; dentibus limbi deltoideis brevibus dorso granu-
latis; achaeniis 3.2 mm. longis in faciebus et praecipue in angulis his-
pidulis; pappi setis ca. 40 albidis vix scabratis corollam subaequanti-

—Braziu: State of Rio Grande do Sul, A. de Saint-Hilaire, no.
950, Cat. C? no. 2737 (rype, Par., phot. and fragm. Gr.); no. 984,
Cat. C? no. 2737 (Par., phot. and fragm. Gr.).

This plant, originally determined as Trichogonia laxa Gardn., but
clearly distinct and having simple not plumose pappus, in habit
recalls EF. alternifolium var. Burchellii Bak., which however has much
larger leaves, smaller heads, narrower, less obtuse and less velvety
phyllaries and much smaller achenes (about 2.2 mm. long).

E. (§ Subimbricata) Rimbachii, spec. nov., arboreum usque ad
10 m. altum griseo-fulvide arachnoideo-tomentosum; ramulis sub-
‘ngulatis; internodiis plerisque 1.5-3 em. longis denique glabratis;
foliis Oppositis breviter petiolatis oblongo-lanceolatis, vel -ovatis
gradatim acuminatis vel attenuatis (sed ad apicem verum paullulo
totundatis) basi acutis integris margine revolutis pinnatim 5-nerviis
vel fere regulariter penniveniis supra subglabris in costa venisque
Principibus solis puberulis subtus multo pallidioribus cum indumento

Atulvido-griseo arachnoideo-tomentoso laxe indutis 9-11 cm. longis
2.54 em. latis textura firmiter membranaceis; petiolo ca. 1 cm. longo;
inflorescentia (immatura) terminali densa rotundata ca. 6 cm. dia-
metro verisimiliter ad maturitatem corymboso-expansa et planiuscula,;
“apitulis sessilibus saepe binis vel ternis in ramulis paniculae 10-
floris ca. 9 mm. longis (submaturis) et 5 mm. diametro; involucri
Squamis ca. 15 ca. 4—5-seriatim imbricatis, extimis brevissimis ovatis

22 ROBINSON

subacutis, intermediis et intimis gradatim longioribus ovato-oblongis
apice rotundatis dorso viridibus leviter striatis laxe pilosis margine
ciliatis; corollis albis gracilibus sine faucibus distincte ‘ampliatis ca.
5 mm. longis glabris, dentibus limbi acutis ca. 0.5 mm. longis; antheris
apice bene appendiculatis; achaeniis (immaturis) ca. 2.5 mm. longis
sparse cum glandulis subsessilibus ornatis; pappi setis ca. 40 albis
‘sublaevibus corollas subaequantibus.—Ecuapor: inner slope of the
eastern cordillera, alt. 3400 m., A. Rimbach, no. 31 (TYPE MATERIAL
in Field Museum and Gray Herbarium).

This species is obviously related to E. hypargyrum Robinson, but
may be readily distinguished by the fact that its intermediate and inner
phyllaries are rounded instead of abruptly pointed at the tip, by its
white instead of violet corollas, and by the fact that the leaves instead
of being wholly glabrous on the upper surface are distinctly though
finely pilose along the costa and chief veins. Finally the arachnoid
tomentum on the lower surface of the leaves is in E. Rimbachii dis-
tinctly fulvescent especially along the midrib and chief lateral veins,
while in E. hypargyrum the lower surface is silvery-pannose. Ac-
cording to a field-note of the collector the wood of E. Rimbachit

attenuatis sine faucibus distinctis; dentibus limbi reflexis ca. ss el
longis; achaeniis gracilibus 2 mm. longis deorsum gradatim dect

RECORDS OF THE EUPATORIEAE—IX 23

centibus; pappi setis ca. 20 tenuiter capillaribus albis laeviusculis
corollas longitudine subaequantibus.—PeERvu: without further in-
dication of locality, 1839-1840, Gay, without number (Par., phot.
and small fragm. Gr.).

This plant, which séems to have been hitherto undescribed, is
unfortunately known only from the upper part of a flowering stem.
This bears five pairs of opposite ascending branches, each having 2-3
pairs of small opposite grayish leaves of harsh texture and being ter-
minated by dense subglobose or oblate glomerules of short-pedicelled
heads. The main cauline leaves have fallen away. They may well
have been larger but were presumably similar in form to the rameal.
The latter, viewed against a strong light, are seen to have a fine
translucent network of veinlets. Beneath they are closely punctate
and shortly scurfy. Repeated efforts have failed to place this plant
in any species thus far described. From those like EZ. Cookii Robin-
son and EF. pycnocephalum Less. which it somewhat approaches in
habitand inflorescence, it differs widely in indumentand in its much less
imbricated involucre. The specimen is unfortunately unprovided
with any detailed data of collection, being merely stated to have
come from Peru. Presumably it came from the Cuzco region.

Mikania (§ Corymbosae) Allartii, spec. nov., gracilis volubilis
brevissime patenterque purpureo-setulosa; caulibus subteretibus

o- diametro atropurpureis; foliis oppositis petiolatis anguste
ovatis caudato-acuminatis cuspidate denticulatis basi rotundatis ob-
Scure supra viridibus dense brevissimeq tulosis subtus paullo pallid
loribus molliter pubescenti-subvelutinis penniveniis vel obscure a parte

eriori laminae penninervatis 4.5-7 cm. longis 2-3 em. latis; petiolo
gracili 8-10 mm. longo, eis ejusdem jugi a linea transversa paullo
Incrassata conjunctis; appendicibus stipuliformibus nullis; corymbis
Tamos terminantibus sessilibus basi foliaceo-bracteatis valde convexis
6-10 em. diametro; capitulis ca. 13 mm. longis sessilibus ad apices
inflorescentiae ramulorum ternatim vel quinatim aggregatis; bracteolis
oblanceolatis vel oblongis acutis subpetiolatis; involucri squamis

eari-oblongis ca. 9 mm. longis dorso rotundatis sparse fusco-setulosis
“pice rotundatis dense setulosis; corollis glaberrimis ut videtur
‘iridibus; tubo proprio exacte tubulato 4 mm. longo; faucibus brevibus
turbinatis vix 1 mm. longis; dentibus limbi lanceolatis acutis ca.
2 mm. longis; achaeniis paullo granulatis aliter laevibus ca. 5.5 mm.
longis deorsum paullo attenuatis; pappi setis ca. 75 salmoneis vix scab-
tatis corollam aequantibus.—VENEZUELA: Los Venados near Caracas,
- 1924, 4. Allart, no. 112 (U. S., phot. and fragm. Gr.).
Among the Mikanias heretofore known from Venezuela this may

24 ROBINSON

be placed near M. trinitaria DC. and M. Johnstonii Robinson. From
both of these, however, as well as from M. parviflora (Aubl.) Karst.
of French Guiana and M. latifolia J. E. Sm. of the Lesser Antilles,
it is readily separable by its smaller relatively narrower leaves, which
are densely and permanently setulose on the upper surface and sordid-
ly pubescent beneath.

M. (§ Spiciformes) Archeri, spec. nov., paullo lignescens volu-
bilis usque ad 3 m. adscendens; caule tereti laevi 3 mm. vel ultra
diametro flexuoso; medulla alba; internodiis 8-10 cm. longis; foliis
oppositis petiolatis oblongis integris acuminatis basi cuneatis paullo
supra basin pinnatim nervatis utrinque glabris textura firmiter cori-
aceis 8-14 em. longis 2.7-3.8 cm. latis subtus ornate reticulatis; venu-
lis vix prominulentibus sed per colorem pallidum conspicuis; petiolis
crassis supra canaliculatis 1-1.5 cm. longis; eis ejusdem jugi basi a
annulo suberoso-incrassato conjunctis; paniculis pyramidatis saepe
lateralibus 4-18 cm. longis puberulis; capitulis sessilibus spicatim
secus ramulos inflorescentiae dispositis 4.5 mm. longis (valde immatur-
is); involucri squamis anguste oblongis ca. 3.5 mm. longis brunnels
apice rotundatis et ciliolatis; corollis ad limbum versus erubescentibus
ca. 3.5 mm. longis fere glabris; tubo proprio fauces longitudine sub-
aequante; achaeniis (valde immaturis) ut videtur glabris; papp!
setis ca. 30.—CoLomB1a: Dept. Antioquia: Medellfn, alt. about 1500
m., Jan. 4, 1931, W. A. Archer, no. 1318 (Gr.).

This interesting species, recently received for determination from
the United States National Herbarium, immediately recalls by the
thickness and general form of its leaves M. pachydictya Robinson.
However, it is clearly a distinct plant, for it has much smaller and
more closely set heads, and its leaves when more closely examined
show several differences. Thus the base is cuneate instead of round-
ed. The veinlets are scarcely at all prominulent, but are on the lower
surface curiously bordered on both sides by pale tissue, so that each
veinlet appears as a narrow pale band practically flush with the
leaf-surface.

_M. (§ Corymbosae) Dusenii, spec. nov., herbacea gracilis Rael
bilis primo aspectu glabra sed in partibus junioribus et in foliis by
marginem apicemque versus obscure puberulis; caule subteretl plus
minusve costato-angulato vix 1.5 mm. diametro brunnescenti; intel:

odiis usque ad 1.5 dm. longis; foliis oppositis graciliter petiolats
elongate lanceolato-ovatis perattenuatis et acutissimis in ieee
paullo undulatis et cuspidate subremoteque denticulatis bast S46)"
tatis, lobis basilaribus triangularibus acutis vel subacutis, simu '
orbiculari patenti vel clauso; lamina 6-9 cm. longa 2-3 ¢™- late

RECORDS OF THE EUPATORIEAE—IX 25

laete virdi subconcolori a basi 3-nervata textura firmiter membran-
acea; petiolis flexuosis 1.5-2.5 cm. longis, eis ejusdem jugi basi ab
annulo crassiusculo transverso conjunctis, appendicibus stipuliformi-
bus nullis; corymbis longipedunculatis oppositis axillaribus plerisque
24 cm. diametro duplice subumbellatim divisis puberulis; involucri
squamis elliptico-ovatis acutis ad apicem versus plus minusve eroso-
denticulatis dorso pallidissimis sparse hirtellis vel subglabris ca. 3.7
mm. longis 1.3 mm. latis; corolla alba glabra; tubo proprio gracili 1.7
mm. longo; faucibus campanulatis ca. 1.2 mm. altis; dentibus limbi
late deltoideis ca. 0.7 mm. altis; achaeniis (immaturis) glabris ca. 1.2
mm. longis; pappi setis ca. 40 tenuiter capillaribus albis vix scabratis
non apice clavellatis—Braziu: Parand: Castro, in grassy somewhat
marshy places. alt. 840 m., 29 Mar. 1916, Dusén, no. 18,040 (TyPE,
in Gray Herb.). ArGEntTINa: Chaco: Colonia Bebitez, in grassy
a of ravines and bottom lands, Mar. 1931, A. G. Schulz, no. 185
r.).

This plant, apparently first collected and distributed by the late
- Per Dusén, was by him regarded as a variety of M. scandens, a
name vaguely applied in a very comprehensive way to a group of
small-headed and cordate-leaved Mikanias of § Corymbosae. As
already pointed out by the writer, Contrib. Gray Herb. Ixiv. 23 (1922),
€ real M. scandens (L.) Willd. is a plant which appears to be re-
stricted to Atlantic North America and which has purplish corollas
and relatively long and narrow linear-oblanceolate and acute phylla-
es. Dusén’s plant here characterized has distinctly broader and
ovate-elliptic phyllaries, white corollas (with proper tube exceeding
the throat, which is not the case in the North American M. scandens),
and highly characteristic narrow and very gradually pointed leaves,
which are Sagittate rather than hastate in tendency and have a small
tound often closed sinus at the base.
_ +t has thus far proved impossible to place this plant satisfactorily
w1 any of the rather numerous Mikanias which have been described
from the South American countries. From all these it appears very
‘eadily distinguishable on account of its rather striking foliage.
© Dusén’s Brazilian material of this interesting plant was being
held for further study, the writer was much pleased to receive (for
®xamination and identification) through the courtesy of Mr. Angel
- Cabrera an Argentine specimen seemingly identical in all respects.
came from the Chaco, where it had been collected by Mr. A. G.
Schulz in 1931,
There is still much need of close monographic work before the
Tather humerous, variable and excessively perplexing plants of this

26 ROBINSON

affinity can be satisfactorily delimited and keyed. It is quite possible
that the one here characterized may be found to intergrade with M.
micrantha HBK., M. periplocifolia H. & A., M. cynanchifolia H. & A.
or M. umbellifera Gardn., but in the absence of such evidence it
seems best placed on record as a species. To refer it, as suggested by
Dusén’s label, to the North American M. scandens would be to sur-
render all hope of clarity in the classification of these small-headed
and cordate-leaved Mikanias of § Corymbosae.

M. (§ Thyrsigera) Guilleminii, spec. nov., praeter inflorescen-
‘tiam obscure glandulari-puberulam glaberrima verisimiliter scandens
et plus minusve lignescens; caulibus teretibus gracilibus post exsicca-
tionem multicostulatis albo-medullosis; internodiis ca. 7 cm. longis;

longis 2-6 cm. diametro oppositirameis subdensis; ramulis pedicellis-
que cum glandulis fi til } 13] ae:

and the inflorescence in consequence decidedly denser. The phyl-
laries are broader and more herbaceous. From the study of a photo-
graph of the type of De Candolle’s M. laevis var. angustior it ae
probable that the plant of Guillemin may represent it, but eve?

RECORDS OF THE EUPATORIEAE—IX 27

this be true it is still no less evident that the plant is so different from
the typical M. laevis (or more properly M. trinervis) that its specific
separation becomes essential to proper taxonomy.

M. Widgrenii, spec. nov., volubilis verisimiliter alte scandens

que 2~4-denticulatis (dentibus ca. 0.2-0.3 mm. longis inter se ca. 5
mm. distantibus) basi patenter cordatis et ad insertionem petioli
breviter acuminatis a basi 3-nervatis (nervis lateralibus mox bifur-
catis) textura tenuiter membranaceis glabris viridibus (subtus paullo
pallidioribus) post exsiccationem modice fuscescentibus; paniculis
terminalibus compositis; partialibus lateralibus oppositis in pedunculo
3-6 cm. longo elevatis ovoideo-subpyramidatis ca. 5-12 cm. altis et
6-8 em. diametro; ramulis pedicellisque brunneo-puberulis (pilis
Incurvis); pedicellis 3 mm. longis; capitulis ca. 8 mm. longis; involucri
squamis lanceolatis brunneo-viridibus tenuibus dorso subglabris 3-
hervatis ca. 5 mm. longis 1—1.2 mm. latis; corolla glabra; tubo proprio
racili 2 mm. longo; faucibus campanulatis ca. 1 mm. altis; dentibus
limbi late ovatis ca. 0.6 mm. longis; achaeniis nigris 2.5 mm. longis
cum angulis pallidis ornatis; pappi setis erubescentibus corollam
subaequantibus vix scabratis.—BraziL: Minas Geraes, Widgren, 1845
(Par., phot. Gr.).

: i puzzling plant, while slightly recalling some forms of the widely
distributed and variable M. micrantha HBK. cannot be satisfactorily
included in that species, since its inflorescence is of a different character,
the heads being more openly and paniculately disposed and the shape
of the inflorescence being somewhat more elongated and subpyram-
dal than is the case in M. micrantha where the heads are in flattish
or but moderately convex corymbs. Nor has it been found possible
hog ce the plant satisfactorily in any other of the now numerous
fh an species. The lateral lobes of the leaves are rounded or but
slightly angle-pointed, not hastately developed. Their teeth are re-

d on each side to 2-4 obscure little cusps and the texture of the
leay » even at full maturity of the achenes, is unusually thin and
delica _As to the form of its inflorescence the species is about
intermediate between § Thyrsigerae and § Corymbosae.

28 : ROBINSON

THE STEVIAS OF COLOMBIA

TuHoucH few in number the Colombian Stevias have never been
carefully listed, keyed or mutually contrasted. In consequence
their identification has long been a matter of no small difficulty and
much of the Colombian material of the genus found in herbaria has
been incompletely or often erroneously labelled. A revision of such
specimens as have been available to the writer leads to the recogni-
tion of some ten species as occurring naturally within the limits of
the coun

In the way of brief generalization it may be said that of these
species two, namely S. ophiomaches and S. Purdiei are endemics, here
presented as new to science. The Colombian record of a third species,
S. Wageneri, rests upon an isolated and surprisingly remote station
in southwestern Colombia for a plant hitherto recorded only from
the neighborhood of Caracas in northern Venezuela. 8S. Lehmannii,
a little known species of western Colombia, apparently reappears mm
southern Mexico. It is still rather doubtfully separable from S. elatior
of which possibly it may prove only a form with denser inflorescence
and awnless achenes.

Of the remaining Colombian species, four, namely the diffusely
panicled S. elatior, the vernicose and shrubby S. lucida, the virgate
herbaceous S. serrata with narrow, crowded and mostly alternate
leaves, and the variable but always rather small-headed S. rhombifolia,
are among the commonest species of the genus and are found in fairly
characteristic form in Colombia as they are both to the northward in
Central America and Mexico and to the eastward in Venezuela.

S. elongata HBK. constitutes something of a problem. The type
material in Paris is not very complete or characteristic. It shows only
the uppermost part of the stem with not very satisfactorily developed
inflorescence. After examining this specimen and comparing It WI
seemingly related material from Mexico and Venezuela, it has bee
impossible for the writer to escape the conviction that it is merely @”
awnless form or variety of the rather widely distributed and mostly
awned plant later described from Venezuela as S. caracasana De.
and from western Mexico as S. elliptica H. & A. Strangely enough
the type of S. elongata, though stated to have come from the vieity
of Bogoté, has not been very precisely or convincingly matched by any
subsequently collected material from that region, though a -—
from Guadalupe (Ariste-Joseph, no. A76) in the United States Natio
Herbarium, while having somewhat more prolonged glomeru
shows in most of its technical characters a fairly close ap.

STEVIAS OF COLOMBIA 29

Much better and more copious material of this rachis is needed be-
fore final conclusions can be reached regarding i

The recently recognized Colombian occurrence i whee’ is believed to
represent S. pallida (Sch.-Bip.) Hieron. has’'made it necessary in this
connection to review the mixed material on which this species was -
_ based and to determine the logical application of the name. A dis-
cussion of this matter will be found on pages 10-13.

Key To THE STEVIAS OF COLOMB
a. Inflorescence lax; oy eo ‘, aay oh, Giialiltys or often
much exceeding the ee et by
b. sg 1 are coroniform, ex ristat ied
oe pedicels fully ni $s the longi of the involucre;
s borne singly; leaves white-tomentose benea me

ot CCU sk cbc us hs cece ek et S. Wageneri.
¢. Pedicels (even at full maturity) scarcely = — i
ab 2 Gi heads clustered in co glom

osely pubescent and da rk tokted feoaath: 2. 'g. Lehmannii.
b. Pac of salen and awns, the latter about equalling the

0 8 Cee 68 8 wee ee ee be 6 8 le S66 O86 oo Be Bree e 8 8S ee ee ee

d. Herbaceous or suffruticose, not verni vase
é. Leaves $e lanceolate or oblanceolate vt 5 watbodis oblong or

Par ae ae ate eA ee
OWA ewe ee ee ee ee

b 25 wm
thiarech ee byrne ee oe ae i nieschen
g. Leaves acutish; heads subsessile in close Laue
at the ends of elongated ascending branches;
phyllaries attenuate, 6-7 mm. long; corolla-throat

f. Leaves ov ib na ew ak pos sa eee ke orwee «>

7. S. Purdiet.

. With some sessile glands. .. .t.
1. P dase of firmish texture, more or less thick-
n the median part, rarely over 5 mm. long;
corollas mostly white; pedicels short but usually

oe Pe PR ete oe cso 8. S. rhombifolia.
i. Phyllaries thinly membranaceous, loosely villous,
not thickened in "the median part eiegt near
base, m 7 mm. long; corolla throat
enetaagatr beet - more y gree -white; 9. S. pallida

eeeeeee
eee ee ee ee oe ee ASU R eS

30 ROBINSON

1. S. Wageneri Hieron. Weak-stemmed and often decumbent
perennial herb, closely glandular-puberulent; leaves opposite, petiolate,
ovate, obtuse, crenate-dentate except at the abruptly contracted then

throat pale yellow or greenish, the lobes white or nearly so; achenes
crowned by scales only.—Hieron. in Engl. Bot. Jahrb. xxviii. 562
(1901).

Ex Cauca: along stream below “San Isidro,” ,Purace, on the Cordillera
Central, alt. 1900-2100 m., Pennell & Killip, no. 6407 (Gr.).

[ VENEZUELA.

A species previously known only from the neighborhood of Cara-
cas yet closely matched by this geographically remote Colombian
material.

2. 8. Lehmannii Hieron. Closely related to the preceding but
somewhat taller (4-6 dm. high) and coarser; leaves larger, clearly
dark-punctate and much less canescent beneath; panicles or at least
their component parts flattish-topped; pedicels even at full maturity
not much exceeding the involucre; corollas said to be roseate; achenes
crowned with short scales, exaristate-—Hieron. in Engl. Bot. Jahrb.
xxviii. 562 (1901); Robinson, Contrib. Gray Herb. xc. 98 (1930).

Et Cauca: Popayan, alt. 1700-2400 m., Lehmann, no. 5199 (Bri., Gr., N. Y.)
Et Vatux: El Saladito above Cali on road to Buenaventura, on the Western

Cordillera, alt. 1600 m., Pittier, no. 750 (U. S., Field Mus.).
[Also apparently in southern Mexico.|}

_3. 8. elatior HBK. Of similar habit but with diffuse panicle and
(in Colombia) mostly subsessile leaves of broadly rhombic-ovate of
even subdeltoid-ovate form; heads mostly 10-12 mm. long; ¢0
with tube and throat usually deep purple, occasionally varying (©
pale greenish yellow, and lobes pink or white; achenes usually hetero-
morphous; the adelphocarps (mostly 4) bearing 3-5 slender purple
awns about equalling the corollas.—Nov. Gen. et Spee. iv. 144 Ceo
Robinson, Contrib. Gray Herb. xc, 97 (1930), which see for synony™Y
and varietal subdivision.

. Macpauena: Santa Marta, 1849, Purdie (Gr., where mixed with a bit oe

Pea gragg possibly S, ata). NORTE DE SANTANDER: on 08 *9000-
m, no. 555 (K., BM., Gen.); between Pamplona and La Isla, alt- iibolat

2500 m., Killip & Smith, no. 20,745 (Gr.). Cunpinamarca: Bogoté, Hu

STEVIAS OF COLOMBIA 31

& Bonpland (Par., phot. Gr.); meadow at base of mountain south of Sibate,
alt. 2900 m., Pennell, no. 2463 (Gr., N. Y., U. S.). BoyacA: Duitama,
Toro, no. 13 (N. Y.). Ex i :
2300 m., Lehmann, no. 3710 (Gr., K.). NariNo: in bushy woods on clay soil
near Yacuanquer, alt. 2500 m., Lehmann, no. 663 (Gr., K.). WrrsotT sTaTED
LOCALITY: 5 : (K.); ine
K. where stated to have come from the unlocated “ Pérama de la Culata’’).

EXICO, GUATEMALA, VENEZUELA; to the southward replaced in Ecuador
and Peru by multiaristate species of similar habit, but re-appearing in Botivia
in a doubtful variety.]

4. §. lucida Lag. Shrub, mostly 0.6-3 m. high, the young parts
vernicose; leaves opposite, slender-petioled, bright green, ovate-
lanceolate to oblong, acuminate, serrate, 3.5-7(-11) em. long, 1.5-
3.5 em. wide; heads sessile or nearly so in corymbosely disposed
dense glomerules; phyllaries narrow, acute, very viscid and vernicose;
corolla-tube and throat usually purple or rose-colored, but sometimes
yellowish or greenish white, the lobes white or pale pink; achenes
uniform, each having a crown of short erose or dentate irregularly
connate scales.—Gen. et Spec. Noy. 28 (1816); Robinson, Contrib.
Gray Herb. xe. 110 (1930), which see for more detailed statement of
lit. synon., and vars. S. glutinosa HBK. Nov. Gen et Spec. iv. 148,
t. 353 (1820)

Maqpatena: Riohacha, Schlim, no. 804 (K.). Norre pr SanTanpéR:
shrub 2.4-3.6 m. high, Pamplona, alt. 2300-2400 m., Killip & Smith, no.
3778 (Gr., N. Y.). Sanranper: San Pedro, alt. 2440-3050 m., Kalbreyer, no.

(Brl., K.); open hillsides near La Baja, alt. 2700-3700 m., Killip & Smith,
tes. 17,171 (Gr. N. Y.), 18,193 (Gr., N. Y.), and 18,393 (Gr. N. Y.); near

tas, alt. 3100-3750 m., Killip & Smith, nos. 17,247 (Gr.. N. Y.), 17,321

re ( * P. r egas alt 300—

Kil & hs tp i .
tp & Smith, no. 16,866 (Gr. N. Y.). CunprnaMarca: mountains about

» Humboldt & Bonpla (P ot. Gr.), Holton, no. 308 (K., N. Y.),
ont 1167 (Par.), Stubel, no. 157 (Brl.), Rusby & Pen 1292
Ugh; > U Mus.) and 1293 (N. Y.), AristeJoseph, n
oe fragm , Pring, no. 118 (Mo.); in open woods t Sibate and
San alt m., Lehmann, no. 2564 (Gr., K.); dry ef Keine
cliff Bogota, alt. 2 900 m., Pennell, no. 2312 (N. Y.); on sandy
K . the Eastern Cordillera at Facatativa, alt. 2650 m., André, 7 (Gr.,
(Gr A ramo d nion, alt. a 2300 m., André, no. 2931
rocky jg, Sanitic cliffs, alt. 2300-2500 m., Killip, no. 6852 (U. S.); open

¥ citis, below Coconuco, Pennell, no. 7170 (Gr., N. Y.). WiTHOUT LocaL-

la 08 (KJ; Lehmann, bo. 4737. (K
\ Both no. 4 .); Lehmann, no. as
Forma Guaremata, Costa Rica, PANAMA, western and northern VENE-

R
y
S
is)

f Of this Species only the typical variety (itself subject to minor and
vmal changes) is as yet known from Colombia. Triana reports
that 1t—doubtless its resinous and somewhat balsamic foliage—s
“ployed by the natives as a compress for the healing of bruises.
. 8. serrata Cay. Erect virgate perennial herb, mostly 3-6 dm.

32 ROBINSON

high; stem terete, incurved-puberulent, simple to the fastigiately
branched corymb, very leafy; leaves chiefly alternate, oblanceolate
to spatulate or sublinear, crenate-serrate from near the middle to the
apex, gradually narrowed and entire toward the base, firmly membra-
naceous, often conduplicate and recurved-spreading, sessile or shortly
petioled, mostly 2-3 cm. long and 4-6 mm. wide; corymb rather dense,
moderately convex, mostly 5-10 cm. in diameter, its branches re-
peatedly divided, shortly pilose; heads pedicelled, 8-10 mm. long;
phyllaries pale green, incurved- or appressed-puberulent and sprinkled
with conspicuous sessile glands; corollas white or pale pink; achenes
normally heteromorphous; adelphocarps (mostly 4 in number) 2-3-
awned; isocarp mostly awnless but crowned with scale-pappus.—lc.
33, t. 355 (1797); Robinson, Contrib. Gray Herb. xc. 122 (1930),
which see for lit. and synon. S. ivaefolia B? bogotensis DC. Prod. v
118 (1836
_Counpinamarca: Triana, no. 1163 (N. Y., K., Par., fragm. Gr.); on 0

hillside, Usme, Dawe no. 119 (K.); Tequendama. alt. 2400 m., Bros. A
r, NO.

aire & Arthu 7 Seve oe a of Bogotdé, Goudot (Gen ye Bro. A le
Joseph, no oy, v6 (C. , frag , also unnumbered (Gr.); San Cristo ristobal,
Pring, no Mo.); so est of a Cruces, Pennell, no. 3193 (Gr. . N. Y.).

Eu Caen: “eld | in Ponta Cations, Coconuco, Pennell & Killip, no. 6477
[Mexico, CenrraL America ; VENEZUELA, Ecuapor.]

6. 8. ophiomaches Robinson. Almost herbaceous curved-ascend-
ing or nearly erect perennial, usually simple below the middle, rather
stiffly and spreadingly branched above, 4-6 dm. high; leaves opposite,
oblanceolate, narrowed toward an obtuse or rounded tip., obscurely
crenate on the distal half, cuneate and entire toward the base, 3-
nerved, shortly pubescent; blade 1.5-3.5 em. long 5-8 mm. wide,
membranaceous in texture; petiole slightly margined, 4-7 mm. long;
corymbs composite, terminal and lateral, convex, incurved-pubert-
lent; heads mostly but 7-8 mm. long, shortly pedicelled; phyllaries
substramineous, 4-5 mm. long, short-pointed, covered on
with short incurved glandless hairs and bearing also a few sca
sessile glands; corollas white, much exserted; achenes heteromorphous;
adelphocarps 4, each 3-5-awned and bearing a short scale crown
idiocarp also coronulate but awnless.—Contrib. Gray Herb. xevi-
10 (1931).

Fe DE SANTANDER: open ous by roadside, Santa Maria

isotype Kalbreyer, no. 1240 (rypx, in herb. K., phot. here Ge
m7 . ce in open. bones, ee and Sana 2125-
eyer, no. 605 (K, dry space

Flowers said to be fragrant; plant used for snake-bite.

STEVIAS OF COLOMBIA 33

7. 8. Purdiei, Robinson. Suberect herbaceous perennial 4-6 dm.
high, shortly and crisply hairy, upwardly branched from the middle;
stem slender, terete, reddish brown; internodes 1-3.5 cm. long; leaves
(except a few of the uppermost) opposite, sessile, lanceolate or narrow-
ly subovate, pointed at each end, inconspicuously crenate-serrate
from the middle, above covered with scattered white hairs (thickened
at the base), beneath slightly paler, punctate, crisped-pilose on nerves
and chief veins, 3-nerved from the base, 2-2.5 cm. long, 5-8 mm.
wide, membranaceous, often proliferous in the axils; glomerules
dense, 1-2.5 cm. in diameter, terminal on stem and elongated branches;
heads subsessile 10-11 mm. long; phyllaries lance-linear, attenuate,
very acute, dorsally beset with many small short-stiped glands and a
few longer crisped glandless hairs; corolla limb white or pink, the throat

] 1 ae a

andtube probabl slender. } , surmounted

J Purple, ’ i;
by a short crown of scarious erose partially connate scales.—Contrib.
Gray Herb. xevi. 13 (1931).
Macpatena: St. Miguel, Nevada de Santa Marta, William Purdie (K.,
phot. Gr.)
8. §. rhombifolia HBK. Upright perennial, herbaceous to suffru-
ticose or sometimes distinctly shrubby, mostly 0.6-1 m. high; stem
terete or nearly so, often purple, incurved- or crisped-pubescent;
leafy to the inflorescence; leaves (except a few of the uppermost) op-
posite, thombic-ovate, subacute to acuminate, crenate-serrate from
about the middle, either sessile by a narrowed base or shortly petioled,
‘parigly to copiously pubescent on both surfaces, usually not much
paler beneath, mostly 2-6 em. long, 1.5-2.5 cm. wide; panicle a com-
Pound convex corymb; heads (small for the genus) mostly 6-9 mm.
high, sessile or short-pedicelled; phyllaries narrowly oblong, shortly
pomnted, usually pale green though sometimes purple-tinged, incurved-
puberulent with glandless hairs but sprinkled with sessile glands;
corollas white or more rarely roseate to (at least as to throat and tube)
purple; achenes bearing a scale pappus with or without 1-4 awns.
~ Noy. Gen. et Spec. iv. 143 (1820); Robinson, Contrib. Gray Herb.
Po (1930), which see for further lit., synon., and varietal sub-
ion,

This widely distributed and variable species exhibits in Colombia
* Same inconstancy in degree of pubescence already remarked in
.“xico. Of the more definite varieties dependent upon pappus-dis-
Unctions two are to be readily recognized in Colombia.
(Ba %. typica. Achenes normally heteromorphous; adelphocarps
r ) each bearing both scales and 1-4 (mostly 2-3) awns.—S. ca-
“asana Klatt in Engl. Bot. Jahrb. viii. 33 (1887), not DC. S.

34 ROBINSON

Benthamiana var. pesarensis Hieron. in Engl. Bot Jahrb. xxviii. 561
1901).

Y.); in clearing, Onaca, alt. 760 m., H. H. Smith, no. 612 in part (N. Y.).
: Ora

Norve pe SANTANDER: borders of forest, Prov. Ocafia, Aalbreyer, no. 1234
K.). CunpinaMaArca: Cipaquira, Prov. of Bogota, 2 m., Triana,
, Par.); Boqueron de Bogota, alt. 2700 m., André, no. 745 (N. Y.);
open slope, alt. 1800-1 : ‘e Redondo,” south of Quetame, Pennell,
(N. Y., U.$.), a form with leaves unusually hairy beneath and thus
approaching the Mexican S. nepetaefolia HBK.; dry bushy mountain slope,
alt. 2: 2900 Pennell, 306 (Gr., U. S., N. Y.) a alt
; ? , Pennell, no. 2520 (N. Y.); Sabana, alt m., inaire &
Arthur, no. 77 ( ) Lu Cauca: El Chorro, about the River Tulud, Holton
no Ag ing on loess formation near Hacienda Sotara,
‘ , Lehmann, no. 3672 (Gr., K.); in de rubby thickets, Al

Pa, P

Gr.); rocky bank, alt. 2300-2800 m., between “Canaan” and Coconueo,
Killip. no. 6819 (U. 8.) ARINA: on dry clayey soil, about Yacuanquer near
Pasto, alt. 2500 m., Lehmann, no. 558 (Gr.).

(Mexico, GuaTEMALA, VENEZUELA, Ecuapor.]

Var. 8. stephanocoma Sch.-Bip. Closely similar in habit and in
variability of foliage; achenes homomorphous, all exaristate an
crowned merely with a short scale-pappus.—Linnaea, xxv. 279 (1853);
Robinson, 1. c. 134 (1930). _S. compacta Benth. Pl. Hartw. 197 (1845).

Macpa.ena: Sierras de Maracaybo, Sta. Marta, Purdie (K.); in hillside
clearing, alt. 760 m., H. H. Smith, no. 612 in part (Gr., N. Y., Field Mus.,
u, 8. ; Norte DE SANTANDER: shrub, 1.5 m. high, on savannas, alt. ig
Ocafia, Schlim, no. 117 (Gen.); woods along streams, Culagé Valley north of
Toledo, alt. 1500-2100 m., Killip & Smith, no. 20,139 (Gr., N. Y.). SaNTAN-

Las Vegas, alt. 2000-2500 m.

Hartweg, no. 1091 (K., phot. Gr., N. Y., phot. Gr.); on open slope, alt. 1800-
2100 m., “Monte Redondo,” south of Gusta Pennell, no. 1816 (Gr.).
NariNa: Proy, Pasto, alt. 1500 m., T’riana, no. 1164 (Par.).

(Mexico, Cenrrat America, Venezurta, Ecuapor, PEerv.]

Varying much in the degree of its lignescence. Some collectors
refer to it as an herb, while others report it decidedly shrubby.

9. S. pallida (Sch.-Bip.) Hieron. Herbaceous or suffrutescent
perennial 0.5-1 m. high, upwardly branched and finely crisped- of
incurved-pubescent; root a fascicle of tough dark fibres; stem terete,
thickish (sometimes 6 mm. in diam.) and a little woody at base, pul
plish-brown, leafy; leaves opposite, subsessile or shortly petioles
blade rhombic-ovate, acute or acutish, rather sharply serrate except
toward the cuneately narrowed base, pale green and shortly incurve®
puberulent above, beneath even paler, dark-punctate, and pubes
on nerves and chief veins, 2.5-3.5(-5) em. long, 1.4-1.8(-2.4) cm:
wide; corymbs convex and compound; glomerules dense, mostly | 1
cm. in diameter, terminating the ascending branches; heads about |

STEVIAS OF COLOMBIA 30

mm. long, sessile or nearly so; phyllaries 7 mm. long, 1.3-1.6 mm.
wide, oblong, acute, thin, dorsally crisped-hirtellous, the hairs white,

white to rather deep purple, its lobes white to roseate; achenes black,
slender, 4 mm. long surmounted by a short crown of scarious scales
(0.8 mm. long), either all awnless or 1-3 achenes in each head bearing
7 awns each.—Hieron. in Engl. Bot. Jahrb. xxi. 328 (1895), at least
in its significant part, that is, as to the name-bringing synonym. (See
page 10.) 8. elongata B. Moritziana I. pallida Sch.-Bip. Linnaea,
mv. 275 (1853)
pink, stigmas pink, vicinity of Vetas, alt. 3
: (Gr, N. Y.); open rocky hillside, woody herb, corolla and styles
light pink, Killip & Smith, no. 17,269 (Gr., N. Y.); herb,
corollas deep pink [to purple on the throat], the lobes and styles paler, edge of
‘dramo 3400-3 m., Killip & Smith, no. 17,398 (Gr., N. Y.).
CUNDINAMARCA: herb, field, base of mountain, alt. 2800-2900 m., Zipaquira,
ennell, no, 2519 (Gr., N. Y.).
[WEsTERN VENEZUELA. |
This plant, early gathered in western Venezuela by horticultural
fovectors, has long been one of the most vaguely and inadequately
r bed members of the genus. Its precise identity and obscure
Publication are discussed on page 10. Copious and excellently pre-
ted specimens, secured by Pennell and by Killip & Smith in central
and northeastern Colombia, now happily permit it to be put upon
far more complete and satisfactory record. The species in habit
recalls some forms of S. rhombifolia HBK. but has larger heads and
other differences indicated in the key. From the also nearly related
8. elongata HBK. it may be distinguished by its smaller and much
less petiolate lower leaves and by the quite different and glandless
Puberulence of the phyllaries. :
10. 8. elongata) HBK. Erect herbaceous perennial, sometimes
tly stoloniferous; stems terete, purple, at least when young
Closely beset with a short spreading gland-tipped puberulence, but
ming sometimes also some longer scattered glandless hairs; leaves
a (the uppermost sessile or nearly so, the middle or at least the
€r borne on rather long cuneately winged petioles), serrate or
CTenate except toward the base, ovate, acuminate to obtuse, mem-
us, green on both surfaces though a little paler beneath, 3-
nerved from a point distinctly above the contracted basal portion,
3-7 cm. long, 2-5(-6) em. wide, inconspicuously pubescent
long along the nerves; floral leaves much reduced; heads 9-11 mm.
Sessile or nearly so in rather dense corymbosely disposed glom-

SanTanp&r: herb. 6-9 dm. high, thicket, corolla-tube greenish-white, lobes
ici 100-3250 hs

Mostly
at least

36 ROBINSON

erules terminating upright or ascending branches; phyllaries ob-
long to lanceolate, sharp-pointed, pale green or more often purple-
tinged, closely beset with short spreading at length stiffish gland-
ipped hairs; corollas pale rose, lilac or white; achenes (in var. typica,
the only form as yet known from Colombia) homomorphous, bearing
a crown of short scarious scales but all exaristate-—Nov. Gen. et Spec.
iv. 144 (1820); DC. Prod. v. 116 (1836); Robinson, Contrib. Gray
Herb. xc. 140 (1930).

CunpinaMarca: near Bogotd, alt. about 2500 m., Humboldt & i.
(Par., Brl.); also at the mines of Zipaquira, alt. about 2700 m., Hum &

on ace. to Kunth in HBK. I. ce.

(Mexico, WESTERN VENEZUELA.|

Never closely matched by material subsequently collected about
Bogot&, this species has for more than a century remained something
of a mystery. In essential features the original, unfortunately frag-
mentary, material is not separable from awnless forms occasionally
found in Mexico and Venezuela of a plant which usually has awned
adelphocarps. Of this aggregate, S. elongata is the earliest name for
the awnless and S. caracasana DC.., for the aristate element and the
latter, as pointed out by the writer, |. c. 119, may appropriately be
reduced to varietal rank. As yet the awned condition has not been

cal condition is not yet clear, therefore the precise status 0
rather striking form can not be satisfactorily determined without
further study of better and much more copious material.

TRANSFERRED OR REDUCED TO SYNONYMY.

S. Benthamiana var. pesarensis Hieron. in Engl. Bot. Jahrb. xxv.
561 (1901) = S. rHomprrouia HBK. (var. ryPIca).

S. caracasana Klatt in Engl. Bot. Jahrb. viii. 33 (1887), not DC. =
S. rHomBIFoLIA HBK. (var. rypica).

S. compacta Benth. Pl. Hartw. 197 (1845) = S. RHOMBIFOLIA var.
STEPHANOCOMA Sch.-Bip.

S. elongata 8. Moritziana I. pallida Sch.-Bip. Linnaea, **¥- sad
(1853) as to type = S. paiva (Sch.-Bip.) Hieron. F

S. glutinosa HBK. Nov. Gen. et Spec. iv. 148, t. 353 (1820). weg
LucIDA Lag.

S. inaefolia 8? bogotensis DC. Prod. v. 118 (1836) = S. SeRRATA CAV:

STEVIAS OF VENEZUELA a7

THE STEVIAS OF VENEZUELA.

Or the seven species and one variety of Stevia at present known
from Venezuela none is endemic. ‘Their occurrence in that country
would seem to be of the nature of outlying and often surprisingly
remote and isolated eastern stations in South America for plants
prevailingly of Cordilleran distribution. S. elatior HBK., S. lucida
lag., S. serrata Cav., and S. rhombifolia HBK., together with its
awnless var. stephanocoma Sch.-Bip., are all common plants of the
uplands from Mexico to Colombia. S. Wageneri Hieron., hitherto
recorded only from the Caracas region in northern Venezuela, re-
appears without change of form in southwestern Colombia. 4.
pallida (Sch.-Bip.) Hieron. and S. elongata HBK. are plants inade-
quately known, at least in their Venezuelan occurrence, and each of
them seems to be restricted in that country to the mountains about
Mérida. Of these plants much more Venezuelan material must be
assembled before it will be possible either to speak with any confidence
regarding their eastern geographic limits or indeed to be quite certain
of the precise identity of the Venezuelan forms and those assigned to
the same species further to the westward.

The literature relating to the Venezuelan Stevias is slight. Hum-
boldt & Bonpland, though early visiting portions of the country on
their notable journey of exploration, do not appear to have collected

ere, with the exception of the previously described S. lucida Lag.,
‘ny of the score or more of Stevias which they record from tropical
America, De Candolle, Prod. v. 119 (1836), described as S. cara-
“sana a plant of Vargas collected near Caracas. In 1853 Schultz-
Bipontinus in the course of a treatment of Stevia, which he contrib-

Me Venezuelan plants earlier classified by Schultz as subvarieties or
ms of the Colombian S. elongata. Some years later Hieronymus

HBK.) which had been collected several times in the neighborhood
of Caracas.

38 ROBINSON

The only attempt to list the Venezuelan Stevias is that of Knuth,
who in his “Initia Florae venezuelensis”’ published in Fedde, Rep.
Spec. Nov. Beihefte xliii. 696-698 (1928) cites nine species and some
half dozen subspecific, varietal or formal manifestations of the genus.
This work, the sole enumeration as yet published of the Venezuelan
flora, will be found very helpful because it brings together biblio-
graphical references to exceedingly scattered literature. However,
concerning the taxonomic validity or even in some instances the no-
menclatural availability of the names it assembles, it should be used
with due caution since it contains many errors of the kind inevitable
in any catalogue of names drawn without close scrutiny from the
works of many authors, of quite different periods, unequal ability and
diverse classificatory methods.

the “S. dissoluta Schlecht.” of Knuth’s enumeration rests,
as to its Venezuelan occurrence, on precisely the plant characterized
by Hieronymus as S. Wageneri, a species independently included in
Knuth’s list and quite distinct from the original Mexican S. dissoluta
of Schlecht.

“SS. urticifolia Thunb.,”” as employed by Knuth, covers material
collected by Pittier and by Jahn, which was in part S. Wagenert, @

S. rhombifolia var. stephanocoma Sch.-Bip. and as to the rest »-
elatior HBK., while the real S. urticaefolia Billb. ex Thunb. 1s @
Brazilian plant as yet unknown from Venezuela. :

Knuth unfortunately attributes to Klotzsch several names whi
should certainly be assigned to Schultz-Bipontinus, whose author-
ship of the paper on Stevia in Linnaea xxv. 268-292 is clearly and
adequately indicated by the expression “auctore C. H. Schultz Dip.
beneath the name of the genus.

Knuth appears to have overlooked the certainly rather obscure
note of Hieronymus in Engl. Bot. Jahrb. xxi. 328 (1895) in which
he raised to specific rank the plants briefly characterized by Schult
(not Klotzsch) as subvarieties or forms of S. elongata under the name
pallida and purpurascens.

As already explained (Contrib. Gray Herb. Ixxx. 11) the name 8.
punctata Sch.-Bip., employed by Knuth, has no nomenclatural validity.

Key To THE STEVIAS OF VENEZUELA.
a. Inflorescence diffusely paniculate; pedicels li -
equalling or mo
often much docading the involucre. .. .b.
- Pappus exaristate; lower leaves on petioles more than half Wagener.
vate, as the blade; corolla-tube greenish-white. oka
rs) % aR Qo -f ee Bae wat

.
equal-

b. Pa th
ine the corolla; leaves (even the lower in Venezuelan S. elatior.
forms) subsessile; corolla-tube mostly reddish-purple. - - ao

STEVIAS. OF VENEZUELA 39

a. Inflorescence much more compact; heads sessile or on pedicels
normally shorter than the involucre... .c.
c. Fruticose; leaves and young parts vernicose; pappus exaris- :
c. Herbaceous or suffruticose, not conspicuously vernicose. . . .
d. Phyllaries incurved-puberulent with glandless hairs but
usually sprinkled with sessile glands... .e.
e. Leaves alternate or scattered, mostly linear-oblance-

Re ae ib ea be ee ee ee 4. S. serrata.
e. Leaves (except sometimes the uppermost) opposite,
ovate... .f.
f. Heads (small for the genus) mostly 6-8 mm. long;
phyllaries rarely over 5 long, normally som
what thi n th an line at least toward
the base, mostly obtuse or shortly pointed; corol- i aes
me thoatiy white... coi. oA ee 5. S. rhombifolia.

J. Heads somewhat larger, 9-11 mm. long; phyllaries
hin and of softer texture, usually purple-tinged
toward the gradually pointed tip, mostly 6-7 mm.
a a eeertee err re eRe a | S. pallida.
d. Phyllaries dorsally closely beset with short straight
spreading at length stiffish gland-tipped hairs or stipi-
glands 7. S. elongata.

1. 8. Wageneri Hieron. (See p. 30.) Slender species with small
long-pedicelled heads, coroniform pappus, and opposite petiolate

(1930). “S. urticifolia Thunb.” as used by Knuth, |. ¢. 698 in part
(i.e. for pl. of Fed. Dist.), not S. urticaefolia Billb. ex Thunb., which

s Brazilian and wholly distinct.

RR RDERAL Dist.: on savannas about Caracas, alt. 1200 m., Wagener, no. 407
taing j,oPPer Catuche wood, above 1400 m., Pittier, no. 7335 (U. 8.); —

-)} on s, Sil ge
phot. Gr., P ar.); La Ciéne a, Silla de Caracas, alt. 2000— m., Pittver, ms
$320 atl S.); Cerro del Galipén. alt. 1370-1830 m., Eggers, no. 13,5

[Also in the Central Cordillera of CoLoMBIA.]

2. 8. elatior HBK. (See also p. 30.) Also diffuse as to inflores-
y uce, but (in the Venezuelan form) having leaves (mostly 3-5 cm.
sessile or nearly so, rather coarsely pubescent but not conspicu-
ously arachnoid-canescent beneath; heads larger (12 mm. long at
yy) at least some of the achenes in each head 1-3-awned.—
Con, cen. et Spec. iv. 144 (1820); Knuth, |. e. 697 (1928); Robinson,
ontrib. Gray Herb. xc. 97 (1930), which see for further lit., synon.,

40 ROBINSON

and varietal division. S. sessilifolia Willd. and S. rhombifolia Willd.,
quite useless names, unpublished by their author but much later
brought out by Schultz, I. c., 270, thus needlessly increasing the burden
of synonymy. “S. urticifolia” as used by Knuth, |. c. 698 in part
(i. e. as to Jahn, no 533), not S. urticaefolia Billb. ex Thunb.

Feperat Dist.: near Galipan, alt. 1220 m., Otto, no. 929 (Brl.). ARaGua:
grassy openings in the mountains, Colonia Tovar and on the Péramo de la
Culata, Moritz, no. 398 (K., Brl.); Colonia Tovar, alt. about 1375 m., j
nos. 622 (Gr., N. Y., K.) and 622 b. (Gr.), the latter a form with the leaves
slightly eens at base. TrusILLo: near Agua Obispo, alt. about 1830 m.,
Funck & Schlim, no. £ :
1 m., Linden, no. 348, ace. to Schultz, |. c. (the same number at Kew w.
labeled, probably erroneously, as from Caracas); Péramo de La Sal, alt. 1525
m., Jahn, no. 55 ra U.8.);

EXICO, GUATEMALA, CoLoMBIA, and (in a doubtful variety) Bourv1A.]

3. S. lucida Lag. (See also p. 31.) Upright glutinous and
vernicose shrub with opposite petiolate oblong serrate leaves and
densely corymbose heads; achenes of like form, all exaristate but
having a crown of more or less connate scales.—Gen. et Spec. Nov.
28 (1816); Sch.-Bip. 1. c. 288 (1853), including the unimportant
foliar variants «. latifolia Sch.-Bip. and B. angustifolia Sch.-Bip. |. ¢
289; Pittier, Pl. usual. Venez. 210 (1926); Knuth, |. c. 697 (1928).
S. fastigiata Willd. ex Sch.-Bip. |. c. 289 (1853), another name useless-
ly published in synonymy.

Le: ‘
2640 m., Piitier, no. 8333 (Gr.); Pico de Galipan, Pittier, no. 6208, acc. to Kauth,
1. c, Mérua: Pdéramo de Mucuchies, Moritz, no. 368 (K.); Sierra Nevada Sh
Mérida, Linden, ace. to Sch.-Bip. 1. e.; Paéramo de los Apartaderos, alt.

. no. 546 (Gr.); San Rafael, Mérida, G no. . 8. f a
Gr.). Probably from near Mérida, Moritz, no. 1377 (Bri., K.), the number ®
Bry Schultz founded his 8. angustifolia. Without precise locality: Lansverg

rl.).

c.;

According to Pittier, 1. c. 210, this species is in Venezuela called
chilca, a name applied to several other plants also.
4. 8. serrata Cav. (See also p. 31.) Erect virgate very leaty-

Herb. xe. 122 (1930), which see for further lit., synon., distrib
varietal subdivisions. §. punctata Sch.-Bip. Linnaea, xxv. 286
Knuth in Fedde’s Rep. Spec. Nov. Beibl. xliii. 697 (1928).
bados, Zumbadar, et), alt. 2440 mr Lkdens no, 698 (Par, Cen. ema 86

[Mexico, Cenrrat America, Cotompia and Ecvapor.]

STEVIAS OF VENEZUELA 41

5. §. rhombifolia HBK. (See also pl. 33.) Upright perennial
herb or undershrub with opposite rhombic-ovate membranaceous
leaves; heads small (mostly 6-8 mm. long) in compact corymbously
disposed glomerules; phyllaries obtuse or shortly pointed, mostly
pale green or stramineous, tending to be thickened in the middle,
incurved-puberulent with glandless hairs.—Nov. Gen. et Spec. iv.
143 (1820); Robinson, Contrib. Gray Herb. xe. 132 (1930), which see
for further lit., synon., distrib. and vars.

The commonest, most widely distributed, and variable of the
pauciaristate Stevias, this species though abundant in the cordil-
leras from Mexico to Ecuador has been reported but rarely from
Venezuela, where however at least three of its pappus-varieties occur,
as follows. All these without discrimination in Venezuela bear the
vernacular name molinillo.

Var. a. typica. Achenes (all crowned with short scales) normally
heteromorphous; adelphocarps 3-4, each about 3-awned; awns about
equalling the corollas; idiocarp awnless.—Robinson, |. c._ S. uitensis
HBK. Nov. Gen. et Spec. Nov. iv. 145 (1820); Kuntze, Rev. Gen.
366 (1891).

Var. 8. uniaristata (DC.) Sch.-Bip. Pappus-awns present but
few and scattered, rarely more than one or two in a head and only in
some of the heads of the inflorescence.—Linnaea, xxv. 279 (1853).
8. uniaristata DC. Prod. v. 120 (1836).

Ménipa: savannas, alt. about 1525 m., Linden, no. 323 [in part] acc. to
Sch.-Bip., I. ¢.; hillsides, alt. 1800-2000 m., Agua Caliente, Tabay, Gehriger,
no. 532 in part (N. Y., small fragm. Gr.).

(Mexico, }

Var. y. stephanocoma Sch.-Bip. Achenes alike and awnless,
‘towned merely by a scale-pappus.—Linnaea, xxv. 279 (1853); Robin-
son, Contrib. Gray Herb. xc. 134 (1930). S. compacta Benth. PI.

- 197 (1845).

ARagua: Open grassy sl lonia T Moritz, no. 1918 (BM.); near
Colonia Tovar, alt. about 1700 Tolle, nos. 619 (Gr., K.) and 620 (Gr.,
TRANDA: Las Mostazas, alt. 963 m., Allart, no. 200 (U S.,N.¥.,

fragm. Gr.) 'T : ah 27
, T.). Trusgm.o: PA de Jabén, alt. 3200 m., Jahn, no.
(U.8,), Méripa: savannas, alt. about 1698 m., Linden, no. 323 in part (Gen.,

42 ROBINSON

Bip. l. c.) var. 6.; Mérida, Moritz, no. 1375 (Brl.). State DOUBTFUL: ‘
is (B 5, While the label here reads Caracas, the plant was probably
lected in Mérida.

{Mexico to northern Prrv.]

6. S. pallida (Sch.-Bip.) Hieron. (For history and discussion
of this species see p. 10.) In habit not very dissimilar to the pre-
ceding species, but with slightly larger heads (mostly 9-11 mm. long),
thinner, broader, and more gradually pointed usually purple-tinged
phyllaries, and achenes at full maturity about 4 mm. long, either all
merely coronulate or in part each 1-3-awned.—Hieron. in Engl.
Bot. Jahrb. xxi. 328 (1895), where raised to specific rank rather ob-
securely in notes upon quite a different plant from a region remote.
_S. elongata 8. Moritziana I. Pallida Sch.-Bip. Linnaea, xxv. 274 (1853);
Knuth, |. c. 697 (1928).

Méria: Culata, alt. about 2750 m., Linden, no. 483 (Par., sk. and notes
Gr., also K. and BM., where erroneously labeled as from Caracas.)

This imperfectly known plant of western Venezuela appears to be
identical with better and more copious material in recent years col-
lected in Colombia. (See p. 34.)

7. 8. elongata HBK. (See p. 35.) Similar in habit and foliage
to the two preceding species but to be readily distinguished by its fine
and copious glandular puberulence, which covers the stem, branches,
inflorescence, and dorsal surface of the phyllaries——Nov. Gen. et
Spec. iv. 144 (1820); Sch.-Bip. 1. c. 275 (1853) as to his subvariety oF
form purpurascens; Robinson, Contrib. Gray Herb. xc. 140 (1930),
where affin., distrib., vars., etc. are discussed. Both pappus-forms of
this species occur in Venezuela, namely

Var. a. typica. Achenes uniform, all exaristate and crowned
merely with a scale-pappus.—Lit., etc., as above.

Mériva: Péramo de Mucuchies, Moritz, nos. 1376 (Brl., sk. and notes Gr.)
and 1376b (Brl., phot. Gr.). Srate UNCERTAIN: “Caracas,” Linden, no. 475
bis (K.), probably labelled inaccurately and in reality from Mérida.

Mexico, CoLoMBia.

Var. 8. caracasana (DC.) Robinson. Achenes normally hetero
morphous, 3-4 of them in each head 2-3-awned, the (1-2) idiocarps
merely coronulate with scale-pappus.—Contrib. Gray Herb. x¢- =
(1930). S. caracasana DC. Prod. v. 119 (1836); Sch.-Bip. |. ©
(1853); Knuth, 1. c. 696 (1928). a

FEDERAL : . Ge;
war) rt eae t 8 2° pate a al Ge) lng
Caracas, Birschel (Gr.). AraGua: Colonia Tovar, alt. 1000-1100 m» *©-
no. 621 (Gr., N. Y.). Ménrrpa: alt. 2440 m., Linden, no. 476 (Pars Gr).
small fragm. Gr.); Péramo Santo Domingo, alt. 3600 m., Jahn, no. 1098 (Gr.

y

STEVIAS OF ECUADOR 43

TRANSFERRED OR REDUCED TO SYNONYMY.

8. caracasana DC. Prod. v. 119 (1836) = S. ELONGATA var. CARA-
casaNna (DC.) Robinson.

S. compacta Benth. Pl. Hartw. 197 (1845) = S. RHOMBIFOLIA var.
STEPHANOCOMA Sch.-Bip.

“S. dissoluta Schlecht,” as used by Knuth in Fedde Rep. Spec. Nov.
Beibl. xliii. 698 (1928) = S. Wacenenrr Hieron., while the original
. dissoluta Schlecht. applies to a Mexican exaristate variety of S.
evaivor.

8. elatior var. coronata Sch.-Bip. Linnaea, xxv. 271 (1853) = S.
Wacenerr Hieron.

S. elongata 8. Moritziana I. Pallida Sch.-Bip. Linnaea, xxv. 274
(1853) = §. patuma (Sch.-Bip.) Hieron.

S. elongata B. Moritziana II. Purpurascens Sch.-Bip. |. ec. = S.
ELONGATA HBK. (var. rypica).

S. punctata (Jacq.) Sch.-Bip. Linnaea, xxv. 286 (1853) and Knuth,
l. e. 697 (1928) = S. serrata Cav.

S. quitensis HBK. Nov. Gen. et Spec. iv. 145 (1820) and Ktze. Rev.
Gen. 366 (1891) = S. rHomBrrouia HBK.

8. rhombifolia Willd. ex Sch.-Bip. Linnaea, xxv. 270 (1853) = S.
ELATIOR HBK.

S. sessilifolia Willd. ex Sch.-Bip. l. c. = S. eLatior HBK.

8. uniaristata DC. Prod. v. 120 (1836) = S. RHOMBIFOLIA var.
UNIARISTATA (DC.) Sch.-Bip.

‘S. urticifolia Thunb.” as used by Knuth in Fedde Rep. Spec. Nov.
Beibl. xliii. 698 (1928), as to both plants of Pittier = S. WAGENERI
Heiron., as to Jahn, no. 553 = S, eLation HBK., and as to Jahn, no.
27 = S. RHOMBIFOLIA Var. STEPHANOCOMA Sch.-Bip., while the true
S. urticaefolia Billb. in Thunb, is a Brazilian plant not as yet known
to in Venezuela.

THE STEVIAS OF ECUADOR.

Watte not numerous the Stevias of Ecuador have a certain phyto-
8eographic interest, for this country is the meeting place for three quite
different tendencies in the genus. To the north of Ecuador the species
ro Aad known belong to what has been called Eustena and have

exceedingly few exceptions) a pappus consisting of a cycle of
which subequal often connate scales with or without 1-5(-7) awns
tw, about equal the corollas. To the south of the country, however,
‘Wo other for the most part very easily differentiated pappus-tenden-

44 ROBINSON

cies prevail and are illustrated by species which have (1) a short
coroniform pappus of conspicuously unequal scales, and (2) a multi-
aristate pappus, at least the adelphocarps having for the most part
9-20 awns each.

In Ecuador all three of these pappus-forms occur. T hus far little
tendency on their part to break down or intergrade has been observed
or recorded. Nevertheless, it seems best to base the keys in this genus
as far as feasible upon characters of habit, foliage, or inflorescence
rather than upon the distinctions of pappus. As in the case of the
Venezuelan species, very much more material, especially from southern
parts of the country, is needed before any satisfactory treatment of
the Ecuadorian Stevias can be accomplished. The one here presented
is obviously sketchy and can aim only at a usable, if pretty artificial,
key for the distinction of such Ecuadorian members of the genus as
have come to the attention of the writer during several years’ study
of the group.

Key To THe EcuaporIAN SPECIES OF STEVIA.

a. Leaves scattered, narrowly oblanceolate........--+---+.°° 1, S. serrata.
a. Leaves (except sometimes the uppermost) distinctly opposite,
roader, oblong to ovate... .5.
b. Pappus merely coroniform or at most 1-5-awned....¢.
c. Leaves closely sessile by a rounded somewhat amplexi- f
Sr aS ee a teed ees 2. S. tunguraguensis.
c. Leaves somewhat petioled or at least narrowed to a more
or less cuneate base... .d.
d. ves rhombic-lanceolate, mostly less than half as
ong; inflorescence beset with stipitate
ands... .¢.
é. aiid ie cbr el Sane rect igre ree much
paler (almost whitened) beneath, acute to acumin- :
ate, eply serrate. <<. 1.55.45. cine eee 3. S. dianthoide.
e. Leaves somewhat firm in texture but scarcely paler
1

a EM ae Ree Eee ae eee Be aE

oo involuc ete : 7 pod

inc -p ent with glan airs (co

monly interspersed with some sessile glands). .5 S. rhombifolia
J. Inflorescence loose; pedi (at maturity) a

SHAS. oi. eee .
b. Pappus multiaristate; awns (of the adelphocarps) mostly
12-18 and about equalling the corollas. .. .9.

.f
Pek eae Oe eS eee

STEVIAS OF ECUADOR 45

1. 8. serrata Cav. (See pp. 31 and 40.) Erect virgate perennial
of characteristic habit; leaves mostly alternate, narrowly oblanceolate,
crenate-serrate from about the middle, often proliferous in the axils;
heads short-pedicelled in a rather dense convex fastigiately branched
corymb; corollas white; adelphocarps mostly 3-awned.—Ic. 33, t.
355 (1797); Robinson, Contrib. Gray Herb. xc.. 122 (1930), which see
for further lit., synon., distrib., and varietal subdivision. “8. canes-
cens HBK.” as used by Hieron. in Engl. Bot. Jahrb. xxix. 5 (1900).

PicuincHa: on banks of the Machdngara River near Quito, Jameson
(U. 8. fragm. Gr.); in the Chillo Valley and on banks of the Machéngara

(K.); “Verde Cruz,” alt. 2650 m., Bro. Firmin, no. 313 (Gr.).
HWESTERN UNITED States, Mexico, CENTRAL AMERICA, COLOMBIA,
and VENEZUELA.]

2. §. tunguraguensis Hieron. Upright perennial herb, sometimes
1.5 m. high, leafy and usually simple to the inflorescence; stem terete,
purplish, sparingly pubescent above, glabrate; leaves opposite, sessile
by a broadish and sometimes amplexicaul base, crenate-serrate from
below the middle to the gradually narrowed but at the tip often
obtusish apex, 3(-5)-nerved from the base, dull green and scantily
hirtellous or glabrate on both surfaces, becoming about 3 cm. long and

as wide; inflorescence terminal, mostly compound, its branches
opposite, spreading-ascending, terminating in dense corymbously dis-
Posed glomerules; heads sessile or nearly so, about 12 mm. long;
Phyllaries 6.5-7 mm. long, acutish, often purple-tinged toward the tip,
sparingly pubescent on the back; corollas said to be light rose; achenes
homomorphous, each crowned with a scale-pappus, the scales con-
Spicuously unequal, tending to be longer on one side and sometimes
48 Mauch as 1.5 mm. in length.—Hieron. in Engl. Bot. Jahrb. xxvm.
563 (1901),

y bi ms |-
son Tanguay al TSS Lekman, no. 5000 (By phot Cr).

An endemic species, to which Spruce, no. 5042 (Gen.), from the
Andes of Ecuador (distributed as a new species and given an unpub-

ed and now homonymic name), though slightly more pubescent,
*Ppears clearly referable. :
3. 8, dianthoidea Hieron. Erect virgate herbaceous perennial;
eteta terete, somewhat hirsute-villous but at length glabrate below,

~~ Viscous-glandular in the inflorescence, 2-4 dm. high; leaves
°PPosite, sessile by a narrowed base, subrhombic-lanceolate, acute to
sutninate, serrate above the middle and perceptibly callous-margined,
“nuire toward the base, dark-punctate, green and sparingly pubescent

46 ROBINSON

above, much paler, indeed whitish and somewhat villous beneath, at
most 4 em. long and 1.5 em. wide, of firmish membranaceous texture;
inflorescence terminal, flattish, of small rather closely aggregated
cymes of sessile or subsessile heads; phyllaries 6-6.5 mm. long, acute,
purple, sparingly glandular-puberulent on the back; corollas deep
purple; achenes all crowned with a pappus of very short scales (scarcely
0.3 mm. long), this being supplemented in 1-2 achenes of each head
by a short awn scarcely 1 mm. in length.—Hieron. in Engl. Bot.
Jahrb. xxix. 4 (1900

PROVINCE NOT INDICATED: in bushy thickets at the eastern base of Mt.
Piels, Sodiro, no. 5/4 (Brl., phot. Gr.).

Indigenous to Ecuador. A plant striking on account of the pale,
almost white lower surface of the leaves.

4. §. anisostemma Turcz. Decumbent, curved-ascending, or sub-
erect perennial herb, subsimple or several-stemmed almost from the
base; stems terete, dark purple, sparingly to rather densely villous
and toward the summit glandular-puberulent, 2-6 dm. high; leaves
opposite, rhombic- or subelliptic-lanceolate, obtusish, crenate from
the middle, entire toward the narrow sessile or subsessile base, 3-
nerved, sparingly pubescent on both surfaces, of subcoriaceous texture,
scarcely paler beneath, 3.5-4.5 cm. long, 8-18 mm. wide; the crena-
tures thickened and slightly callous on the margin; heads 11 mm. long,
sessile or very shortly pedicelled in dense solitary and terminal or
corymbously disposed glomerules; phyllaries lanceolate, acute, dark
violet (sometimes almost nigrescent), glandular-puberulent, about 6
mm. long; corolla-tube and throat deep purple, the limb roseate t
white; achenes homomorphous or nearly so, each crowned by erose
scarious scales about 0.7 mm. long and at one side usually 2(1-3) short
awns (1-2 mm. in length).—Bull. Soc. Imp. Nat. Mose. xxiv. Pl 1,
167 (1851).

Picut : Qui - U. S., where ul
ee Clie, Jenene, TO ee oo
nterandean thickets, Sodiro (Gen., small fragm. :

An as yet little known endemic species from which the preceding
one is but doubtfully separable. a

5. 8. rhombifolia HBK. (See pp. 33 and 41.) Erect perent
herb or somewhat frutescent plant, mostly 0.5-1 m. high; stem io
incurved- or crisped-puberulent; leaves opposite (except the ee
most), rhombic-ovate, crenate-serrate, membranaceous, sessile by ¢
narrowed base or shortly petioled, hirtellous to rather densely a
bescent on both surfaces, mostly 3-7 cm. long, 1.5-3.5 cm. wide; ge
small, mostly 6-9 mm. long, in dense corymbously disposed glom

STEVIAS OF ECUADOR 47

—Nov. Gen. et Spec. iv. 143 (1820); Robinson, Contrib. Gray Herb.
xc. 132 (1930), which see for additional lit., synon., vars., ete.

Var. typica. Achenes normally heteromorphous, 3-4 in each head
bearing both a coroniform scale-pappus and 1-5 (mostly 2-3) awns
each, the others having the scale-pappus but being exaristate.

Forma normalis. Corollas white or nearly so.—S. quitensis HBK.
Nov. Gen. et Spec. iv. 145 (1820); Jameson, Syn. Pl. Aeq. ii. 79 (1865).

PIcHINCHA: open places between Quito and Mt. ‘‘Burro Potrera,” alt.
about 2560 m., Humboldt & Bonpland (Par., phot. Gr.); along the torrent of
Machingara near Quito, Sodiro (U. S.).

(Mexico, Centra, America, CoLomBia, and VENEZUELA.]

Forma colorata Robinson. Corolla-throat and tube a vivid rose
or even purple, the limb pink to white.—Contrib. Gray Herb. xe. 133
(1930). “\S. quitensis Humb. et Kunth” as used by Benth. Pl. Hartw.
197 (1845), not HBK. S. Benthamiana Hieron. in Engl. Bot. Jahrb.
Xxvili. 561 (1901).

_ PicHincHa: near the bridge of Gudpulo, Hartweg, no. 1090 (K., phot. Gr.);
Ge randean thickets at Gudpulo, Father Mille, no. 473 (U. 58., small fragm.

[Mex1co, Cotomsia.] .

Var. untartstata (DC.) Sch.-Bip., with obsolescent awn-pappus,
and var. STEPHANOCOMA Sch.-Bip., with achenes consistently exaris-
tate, have as yet been unrecorded from Ecuador. That they occur
and will ultimately be discovered in the country seems highly probable.

6. 8. Bertholdii Robinson. Suberect perennial herb, 4 dm. high,
spreadingly long-pilose; leaves opposite, shortly petiolate or the upper
Sessile, ovate, acutish or obtuse, crenate except toward the rounded
or rather abruptly narrowed base, dull green and setulose above, a
little paler and pilose on the nerves and veins beneath, 3-4 cm. long,
2-3 cm. wide; inflorescence an ovoid panicle, its spreading-ascending
branches terminating in fastigiately divided cymes, the heads (about
12 mm. long) on pedicels about the length of the involucre and often
falsely racemose (their pedicels opposite the corresponding bractlets) ;
Phyllaries linear-oblong, shortly pointed, 8 mm. long; corollas deep
Purple in throat and tube, the limb white; achenes homomorphous,
eo ate with a short scale-pappus (about 0.4 mm. long).—Contrib.

Tay Herb. xevi. 4 (1931).

Tous (2): Vi no. 658 in part (K., phot.
Gr the trie Santeria satel oa the ‘owing habitally similar but tech-

quite different multiaristate species.

7. 8. andina Robinson. Erect perennial (5 dm. or more in height)
with rather closely the habit of S. elatior of more northern regions;

48 ROBINSON

stem very shortly grayish-velvety or sometimes subglabrate, the in-
florescence finely glandular-puberulent; leaves opposite, broadly ovate,
obtuse, crenate, cuneately narrowed and entire toward the base; limb
2.5-5 em. long, 2.3-3.5 em. wide, membranaceous, puberulous above,
paler and especially on the nerves and chief veins grayish-pubescent
beneath, subsessile or often raised on a narrow cuneate-winged petiole;
inflorescence ample, terminal, paniculate; pedicels usually exceeding
the involucre; heads about 1 cm. long; phyllaries lance-linear, acute,
glandular-puberulent, 5-6 mm. long; corollas reddish-purple as to
throat and tube, the lobes white; achenes heteromorphous; adelpho-
carps each with an annulus surmounted by 9-12 awns about equalling
the corolla; idiocarp with short scale-pappus and usually a single awn.
—Contrib. Gray Herb. xevi. 3 (1931).

Losa: quebrada near the city of Loja, Poortmann, no. 138 (Par.); between
Loja and San Lucas, alt. 2100-2600 m., Hitchcock, no. 21,475 (Gr., N. Y.);
vicinity of Las Juntas, Rose, Pachano ose, no. 23,230 (Gr.). AguaY:
vicinity of Tablén de Ofia, Rose, Pachano & Rose, no. 23,091 (Gr.).

{[NoRTHERN PERU.]

8. S. crenata Benth. Suberect or curved-ascending perennial
herb, finely pubescent und toward the summit minutely and densely
glandular-puberulent; leaves opposite, elliptic-lanceolate, narrowed
to an obtusish apex, crenate above the middle, cuneate and entire
almost from the middle to the subsessile base, obscurely puberulous
above, nearly glabrous beneath, subchartaceous, 3.5-4.9 cm. long,
1.5-1.8 cm. wide, the margin callous-thickened especially around the
blunt teeth; heads about 1 cm. long, borne in glomerules of varymé
size and density, sometimes sessile or nearly so, sometimes on slender
pedicels 4-6 mm. in length, terminating opposite ascending rather
elongated branches of the inflorescence; phyllaries acute, finely gland-
ular-puberulent; corollas as to throat and tube purple, the limb white
or nearly so; achenes heteromorphous, the adelphocarps bearing each
8-11 awns, the idiocarp exaristate but with a short crown of about 8
ovate to linear erose scales.—P1. Hartw. 197 (1845); Jameson, Syn. Pl
’ Aeq. ii, 78 (1865), at least as to compiled char. but probably not as "
all his material so referred.

PicHINCHA: i ., phot. Gr. 4
Y., phot. ee te ion La 396 (Par.
phot. Gr.).

The typical material of this species had the glomerules rather loose
and but 2-5-headed. On the other hand the material subsequently
collected by Rivet has more compact and numerously headed glomer”
ules. Though at first supposed to be a distinct plant, gad

STEVIAS OF ECUADOR 49

material agrees in all other respects so precisely with Hartweg’s
original specimens that the two collections may with confidence be
regarded as conspecific.

IMPERFECTLY KNOWN SPECIES.

A Stevia, represented by the collection of J. N. & G. Rose, no.
23,825 (N. Y.), from Huigra in the southwestern part of the Province
of Tungurahua is a plant not as yet satisfactorily referable to any of
the here recognized Ecuadorian species. Unfortunately the material
seen—a single lateral branch of the inflorescence—scarcely justifies
its description as a novelty in a group so intricate and imperfectly
understood as the Andean stevias. It has an irregular coroniform
pappus of erose scales, some of them a little longer and slightly aristu-
late at the tip. Its leaves are rhombic-ovate, sharply serrate, sessile
by a narrow base, and membranaceous in texture. The corollas have
the throat and tube an almost orange red and the conspicuously
villous limb yellowish-white, though these colors (quite unusual in the
genus) may be solely the result of conditions during the drying.

tter and more copious material is necessary in order to place this
plant with any proper confidence.

TRANSFERRED TO SYNONYMY.
S. Benthamiana Hieron. in Engl. Bot. Jahrb. xxviii. 561 (1901) =
S. RHOMBIFOLIA forma COLORATA Robinson.
8. canescens Kunth” as used by Hieron. 1. c. xxix. 5 (1900), not
Precisely of HBK. = S. serrata Cav.
“S. quitensis Kunth” as used by Benth. Pl. Hartw. 197 (1845),
hot quite of HBK. = S. rHomBrroira forma coLoraTA Robinson.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

A MONOGRAPH OF THE GENUS PEREZIA,
SECTION ACOURTIA,
WITH A

PROVISIONAL KEY TO THE SECTION EUPEREZIA.

By Rimo BaciGALupPt.

PUBLISHED BY
THE GRAY HERBARIUM OF HARVARD UNIVERSITY
CAMBRIDGE, MASS., U. S. A.

1931,

A MONOGRAPH OF THE GENUS PEREZIA, SECTION
ACOURTIA, WITH A PROVISIONAL KEY TO
THE SECTION EUPEREZIA.

By Rimo BaciGaurt.
(Plates I-V IT.)

TuE need for a monographic treatment of the genus Perezia was
brought to the writer’s attention by Professor B. L. Robinson. The
first species of the group was described one hundred and fifty years
ago and the genus as such has been in existence since 1811. To it,
there have been referred more than 150 species; yet most of these
have never been keyed or mutually contrasted. Not since 1838, when
De Candolle in the seventh volume of his Prodromus enumerated
(under no less than six separate generic names) the then known ele-
ments of what today is called Perezia has the group been subjected
to any general recension. At that comparatively early date only forty
species had been described and of these, De Candolle reduced some
ten to synonymy. Even of the remainder, no synopsis or key was

these Philippi species has been woefully absent from any of the
American or European herbaria visited by the writer. Consequently,
oe of revising § Euperezia must be postponed until the types In

hile can be examined by the monographer. Incidently, it may be
added that more abundant collections, particularly of Chilean species,
ke 4 great desideratum. In spite of these and a number of other

**stacles, the writer has been able to prepare a key to those South

4 BACIGALUPI

American species which are available in the Gray Herbarium. While
relatively few in number, these represent the more widely ranging and
therefore more commonly encountered species. For this reason, it is
hoped that such a partial key will be more useful than would at first
seem probable.

On the other hand, the mainly Mexican § Acourtia, though present-
ing problems of its own, is happily almost completely represented
at the Gray Herbarium, and the few species which were not available
there were very obligingly forwarded from the National Herbarium
at the Smithsonian Institution in Washington and from the herbarium
of Pomona College. It was therefore feasible to attempt a complete
monographic treatment of § Acourtia which here is considered to
consist of 44 species and 8 varieties. Of these, 4 species and as many
varieties are proposed as new.

To Professor B. L. Robinson, under whose scholarly supervision
this work was prosecuted and to whose advice and encouragement
whatever merit this paper may possess is largely due, I tender my
most heartfelt thanks. I would also express my appreciation of Dr.
I. M. Johnston’s friendly interest and stimulation. For help in the
matter of photographing types, I am indebted to Dr. L. B. Smith.
Dr. Maxon of the United States National Museum and Professor
Munz of Pomona College were kind enough to forward types indis-
pensible to the monographic study of the Mexican Perezias. Professor
Briquet of the Conservatoire de Botanique in Geneva similarly
authorized the transmittal of specimens which proved of invaluable
aid. For assistance in bibliographical matters, I wish finally to &
press my gratefulness to Miss Ruth D. Sanderson, Librarian of the
Gray Herbarium.

Taxonomic History oF THE GENUS.

During the thirty years prior to the publication of the generic rai
Perezia, seven of the South American species had been described and
published, four of them under the soon defunct genus Perdiciwin and
three temporarily referred to Chactanthera. Of these the first, de-
scribed in 1781, was Perdicium magellanicum Linn. f. Supplement a
later to be made, appropriately enough, the type of the. new .
Perezia. The Linnaean type is labelled merely “Forster,” indica 8
certainly that the Fuegian plant had been collected by Johann aerrottes
and his young son Georg, the botanists on board the “ Resolution
during Cook’s second expedition (1772-75). It reached the la
herbarium through Abraham Back, “Linné’s most intimate -
from whom he received occasional gifts of plants.”

MONOGRAPH OF THE GENUS PEREZIA 5

In 1790, M. Vahl, Skrivt. af N. H. Selskab. i. pt. 2, pp. 8-15, pl. 5,
6, and 7, described a number of Perdicia, three of which (P. recurvatum,
P. lactucoides, and P. squarrosum) were later to be included in Perezia.
The specimens had been collected in 1776 at Montevideo and the
Straits of Magellan by Commerson, who accompanied the French
naval expedition headed by de Bougainville, charged with the task of
officially surrendering sovereignty over the Falkland Islands (“Iles
Malouines”) in favor of the Spaniards who had shortly prior thereto
reclaimed them.

The extensive travels of Alexander von Humboldt and Aimé Bon-
pland next brought to light three more species of the nomencla-
torially still non-existent genus Perezia from the high Andes of Peru
and Ecuador. These were described and admirably illustrated in
Humboldt and Bonpland’s well known earlier work, Pl. Aequin. ii.
(1809), 146, 168, 170, as Chactanthera pungens, C. multiflora and C.
pinnatifida.
__ Inan article entitled “ Disertacion sobre un Orden Nuevo de Plantas

de la Clase de las Compuestas,” published in the Amenidades Natur-
ales de las Espafias, i. (1811) 26-43, Mariano Lagasca characterized
several new genera, many of them still current today, in his newly
Proposed “order” of bilabiate composites, the Chaenanthophorae.
Among these was Perezia, based on Perdicium magellanicum Linn. f.
and commemorating Lorenzo Perez of Toledo, a pharmacist and writer
on materia medica in the sixteenth century. It is interesting to note
that though Lagasca definitely based the genus on the Linnaean
Species, and doubtfully referred the three species of Perdicium de-
scribed by Vahl to Perezia, he refrained from actually making any
binomials under the newly erected genus.

That Augustin Pyramus de Candolle in 1812 was unaware of La-
gasca’s publication of Perezia a year earlier is clearly indicated in the
_ foreword to his third “Mémoire” on the composites, devoted to the

Labiatiflores,” Ann. Mus. Par. xix. (1812) 59-72, wherein he states
that, although he had been in friendly communication with Lagasca for
“me time, untoward political conditions in Spain had interrupted
that Correspondence. It is not surprising, therefore, that no mention
: Lagasca’s Perezia is found in the tentative generic division of the
oo Composites that follows and that, of the two genera proposed

ch refer to the present-day Perezia, one, Clarionea, attributed to -—
Carlier unpublished communication by Lagasca, amounted to an in-
arent renaming of Perezia, since it was, like this genus, based on
Linnaean species. The other, Homoianthus, attributed to Bon-

6 . BACIGALUPI

pland and: based on species separated from Chaetanthera because of
their homomorphously hermaphrodite flowers clearly refers to Chaetan-
thera. multiflora, C. pungens, and C. pinnatifida. Any possible doubt
is at once dissipated by the diagnostic illustrations accompanying the
genus Homoianthus which delineate the heads of the first two Chaetan-
theras mentioned.

Kunth, in H. B. K. Nov. Gen. et Sp. iv. (1820) 12-15, 309, pub-
lished the generic name Homanthis, a renaming of Homoianthus Bonpl.
To his new genus, Kunth relegated Chactanthera multiflora, C. pun-
gens and C. pinnatifida..

The first Mexican Perezias were described by La Llave and Lexarza
in their pioneer pamphlet, Novorum Vegetabilium Descriptiones, i.
(1824) 25-27. The binomials published were Perezia turbinata, P.
fruticosa, P. moschata and Perdicium cordatum. Although it is quite
likely that each of these names is properly referable to one or another
of the Mexican Perezias, the indefiniteness of the diagnoses and the
absence of authentic specimens make their application a matter of
more than ordinary difficulty. As a result, only one of these binomials,
' P. turbinata, is in use today, and even here, its application is by 2°
means certain. It is to be noted that in this Mexican publication,
species are described for the first time under the generic name Perezid.

A partial review of the genus Perezia was included by Cassini in his
Dictionnaire des Sciences Naturelles, xxxviii. 454 et seq. in 1825. He
emphasized the fact that Perezia antedates Clarionea, and finding the
latter name now available, employed it in a new sense, basing his new
genus on Perdicium lactucoides Vahl. The newly erected Clarvonea
differed from Perezia in its fimbrillose-pilose receptacle, since that of
Perdicium magellanicum Linn. f. was described as naked. Pi
of the Dictionnaire, he described Homoianthus echinulatus which 1s,
as he suspected, conspecific with Perdicium recurvatum of Vabl.

In Linnaea v. (1830) pp. 13-24, Lessing treated Perezia the
larger sense, reducing the generic segregates of De Candolle, Cassin,
and Bonpland to synonymy, thus being the first to formulate the
present-day generally recognized more or less homogeneous gene
unit finally crystallized by Bentham and Hooker. Besides giving dis
criminating recognition to the species of his predecessors, he P :
three of his own, P. squarrosa, P. cubataensis, and P. laevis. He we
evidently struck by the strong divergence in habit of the a
species of this genus which are here considered to comprise # _
defined section, for he based a new genus, Dumerilia, on @ 8 “s
of Perezia collected by Humboldt in Mexico. Unfortunately, #*

ta

MONOGRAPH OF THE GENUS PEREZIA 7
specific identity of Dumerilia Humboldtii has not yet been clearly
ascertainable.

From the Mexican collections of Sesse, Mocifio, and Cervantes in
1788 secured by Lambert, David Don, Trans. Linn. Soc. xvi. (1830)
200-212, launched the genus Acourtia, first describing the plant known
today as Perezia formosa. The Mexican Perezia reticulata was here
first described as a Proustia. Under Clarionia (evidently an inad-
vertence for Clarionea) were described two Chilean species, the now
current P. spathulata and P. virens, and the Mexican species P. run-
cinata. The Chilean specimens were among those collected by Ruiz
and Pavén in 1777-1778. From the collections in the Chilean Andes
of Dr. Gillies, Don, Phil. Mag. xi. (1832) 388, published, under Cla-
rionia, two of the today better represented species of that region,
P. carthamoides and P. pilifera. In the same year, Lessing, in his
Synopsis Generum Compositarum (pp. 407-413), supplemented his
previous account of Perezia by the diagnoses of seven new species
secured during the explorations of Poeppig in Chile. From collec-
tions in the Andes by Gillies, Bridges, Mathews, Tweedie and others,
Hooker and Arnott, Comp. Bot. Mag. i. (1835) 33-34, added three
More species to this now adolescent genus.

In 1838 appeared the portion of De Candolle’s Prodromus treating
the Mutisioid Composites, in which three Mexican species (Acourtia
rigida, A. hebeclada and Dumerilia Alamani), one Chilean (Perezia
Gayana), and another (Acourtia microcephala) from California (col-
lected by Douglas) first saw the light of publication. Most of the
segregates were here maintained, the 30 species of Perezia recognized
being distributed among no less than six separate genera. These were

ased on the character of the receptacle and the outline and serial
disposition of the phyllaries, characters at best of only specific value,
for subsequently collected and described material has served to show

at the distinctions upon which the segregates were based are very
artificial and in no sense generic. Notwithstanding the work of
Endlicher in his Genera Plantarum (pp. 492-3), issued in final form
in 1840, in which he, like Lessing before him, reduced all the South
“meérican segregates, considering them at best sections and sub-
Sections, prominent Chilean workers on the group, among the fore-
ost of whom were Julio Remy, who contributed the family Composi-
‘ae to Claudio Gay’s Flora Chilena, and Rudolf Philippi adopted and
Persisted in using the segregates Clarionea, Homoianthus, and Perezva
in the restricted sense.

Conservative as his treatment of the South American segregates

8 BACIGALUPI

was, Endlicher maintained the genus Acourtia of D. Don for the Mex-
ican species. If habit alone be the criterion, the Mexican group cer-
tainly seems to deserve generic segregation, but unfortunately,
truly diagnostic distinctions are difficult to find and still more
difficult to maintain. Among those used by Endlicher, especially
emphasized were the pappus-characters, and it requires only a short
acquaintance with the Mexican Perezias to discover that the pappus-
bristles are in no case penicillate, in very few species uniseriate and
by no means consistently white as opposed to “flavidus.”’ The differ-
ences in the character of the epigynous disc emphasized between the
two groups is one that will not bear scrutiny, while a few of the Mexi-
can species, contrary to Endlicher’s assertions, do have short-bristly
receptacles. The only difference mentioned by Endlicher that has
not been disproved is the fact that in some (Endlicher would have it
true for all) of the South American species, the marginal corollas appear
radiate owing to a slight exaggeration of the ligulate lip, a distinction
which only partially applies.

Baillon, Hist. des Pl. viii. (1882) 100-101, made the radical move of
relegating the genus Proustia to Perezia, treating it as a section of the
latter. Schultz Bipontinus, in Seem. Bot. Herald (1856) 315, had
earlier referred the Mexican Perezias treated by him to Trivis. The
work of O. Hoffmann in Die Natiirlichen Pflanzenfamilien (hereafter

scription of new species. Chief among the contributors of sup
novelties was Rudolf Philippi who published, between 1856 and 18%,

Chlor. And. i. (1855) 39-43, who described many of the cespitose
Andean species of Bolivia and Peru from his own collections and those
of d’Orbigny; Baker, in Mart. Fl. Bras. vi. pt. 3 (1884) 380; and more
recently Otto Kuntze, Rusby, and Spegazzini. “*
e limits of the genus were finally set down in definite form by
Bentham and Hooker in their monumental and still fundame?

MONOGRAPH OF THE GENUS PEREZIA 9

Genera Plantarum ii. (1873) 500-501, their delimitation having been
accepted by all subsequent workers in the group save Philippi and
retained essentially unchanged in Engler and Prantl’s Die Natiir-
lichen Pflanzenfamilien iv. part 5 (1893) 349, by O. Hoffmann.

The only fairly general and usable treatment of the South American
species of Perezia that has appeared since their enumeration ia De
Candolle’s Prodromus is that in Carlos Reiche’s Flora of Chile iv.
(1905) 435-455. While quite evidently a compiled work and marred
by many objectionable features, its key makes it an indispensable
reference work since, as yet, the botanist who would try to identify a
Chilean Perezia has nowhere else to turn unless he be willing to read
through the overponderous mass of diagnoses which is his almost
futile alternative.

Very few specimens of the Mexican species of Perezia came to the
attention of the early workers on the genus. As already mentioned,
until 1838 only twelve had been described. Of five of these, the exact
identity never has been and probably never will be known with cer-
tainty. A new impetus was furnished by the explorations along the
Mexican boundary and in northern Mexico in the late 40’s and early
50’s by such collectors as Fendler, Wright, and Thurber which
brought new Perezias to the attention of Asa Gray. Three were
described from Fendler’s collections in Pl. Fendl. iv. (1849) 111,
five from those of Wright in Pl. Wright, i. (1852) 127, and another from

urber’s collections in Mem. Amer. Acad. ser. 2. v. (1855) 324.
Subsequent to this, in Proc. Amer. Acad. xv. (1880) 40-41 and loc.
Cit. xix. (1883) 58-61, five more were added from the collections of
Parry and Palmer, Thomas Coulter, Dugés, Galeotti, Linden, Lieb-
mann and Ghiesbreght. A tentative key, separating groups of species
rather then individual species, published by Gray as a footnote in the

eniodical last mentioned, is the only key to the Mexican Perezias
hitherto advanced. :

The work of Gray was followed by that of Watson, who contributed

species, and of Rose, who published another. Between 1892 and

1909 several of the specimens received at the Gray Herbarium from
that prince of collectors, C. G. Pringle, as well as a few from the
collections of E. W. Nelson, Palmer, and Langlassé, formed the basis
for ten additional interesting and highly local novelties characterized
— separately or in collaboration by Robinson and Greenman.
craig , the subsequent explorations of Purpus disclosed several more

ch were described by T. S. Brandegee.

preceding brief account reviews in more or less summary

10 BACIGALUPI

fashion the most important literature of the genus. Among those
contributors not previously mentioned, most of wliom treated Perezia
floristically rather than taxonomically but a few of whom contributed
one or more species each, are Lamarck, Sprengel, Meyen, Grisebach,
Franchet, Gaudichaud, Walpers, d’Urville, J. D. Hooker, Hieronymus,
Hemsley, Bartlett, Dusén, Skottsberg, de Wildeman, and Blake.

GEOGRAPHICAL DISTRIBUTION.

The Mutisieae, of which Perezia is one of the larger genera, form
a preponderantly American group, denizens, with the notable excep-
tion of some species of Mutisia, of extra-tropical realms, either of the
higher mountains in the tropics or of the more temperate plains of
South America. Of the 57 genera admitted by O. Hoffmann in “Die
Natiirlichen Pflanzenfamilien,” 40 are confined to the Americaa
continent and the Antilles, while the whole Subtribe Nassauvinae is
peculiarly New World. The remaining 17 genera are largely restricted
to central and southern Africa, and to a lesser degree, southern and
eastern Asia.

Perezia is a strictly extratropical, for the greater part, austral
American genus. Geographically, as well as taxonomically, the genus
falls into two well marked groups. The species of the section Acourtia
range from the coastal region of southern California to El Salvador
and eastward to central Texas. They are restricted, for the most
part, to the forested middle and higher slopes of the Mexican Sierra
Madre and to the high plateau region east of the mountains in centra
and northern Mexico. Between El Salvador and southern Colombia
occurs a gap of some 1300 miles in the distribution of Perezta. The
§ Euperezia is restricted to the middle and higher slopes of the Andes,
ranging from southern Colombia, through Ecuador, Peru, Bolivia,
and Chile to Tierra del Fuego, one of the Fuegian species even reac
the Falkland Islands. ;

By far the greater proportion of the species with this Andean dis-
tribution are depressed, acaulescent, more or less cespitose and, ™ "
few cases, matted forms, some of the southernmost developing
greatly thickened, coriaceous, much reduced, highly revolute linear
leaves suggesting adaptation to more or less extreme xerophytic ea
ditions. A limited few are found on the higher desert ranges whi
parallel the Andes south of Bolivia in the Argentine provinces of Jujuy,
Salta, Catamarca, Tucumén, and Cérdoba. Yet another small at
of species, forming a habitally and, to a lesser degree, taxonomica”
distinct unit, are found, isolated from this essentially Andean grouP

MONOGRAPH OF THE GENUS PEREZIA 11

and restricted to the plains from southern Brazil to the Rio Negro
region in central Argentina. These are so few in number that they
may be most conveniently enumerated at this point. They are P.
cubataensis, P. squarrosa, P. sonchifolia, P. Kingii, and P. laevs.

Of the Andean types, the most widely distributed is P. multiflora
with a northwesterly-southeasterly range of some 2600 miles from
southern Colombia to Tacna in Chile, throughout western Bolivia
and southward in the arid heights of the Argentine Republic to the
province of Cérdoba. The ranges of Perezia pungens and P. pinnati-
fida extend from Ecuador through Peru to Bolivia. P. recurvata is
found in the extreme southern Andes in Fuegia and on the Falkland
Islands. For the most part, the remaining species, certainly more
than three-fourths of the § Euperezia, are much more restricted in
their distribution, many being extremely local. Regions of high
endemism occur in the central Chilean Andes, in Fuegia, and in the
higher portions of the Bolivian and southern Peruvian Andes at
elevations close to the snow-line.

ndemism is a marked feature of the Mexican Perezias as well.
Of the 44 species admitted in the monograph of this section, fully
twenty are as yet known either only from the original collections or
from a very narrowly circumscribed area adjacent to the type lo-
cality . Nine are from the mountain slopes of the states of Nayarit and
Jalisco, a region of generally recognized high endemism. Three more
are from the upland fastnesses of Michoacén, two from Baja Cali-
fornia, and three from the rugged region between Cuernavaca and
northern Oaxaca, another focal point for endemism. :

The more widely ranging species are those of the plateau region
fast of the Sierra Madre in northern and central Mexico. The most
*asterly ranging of these is the scapiform P. runcinata which reaches

¢ Gulf coast in the region about the mouth of the Rio Grande. It
anges from south central Texas through Nuevo Leén and Coahuila
southward to Hidalgo, occurring largely on the lower portions of the
Rag area as well as on the coastal plain. The ranges of P. nana
entire , Wrightii, approximately coextensive, are the largest of the
the he group. These are species of the isolated ridges which break up

high central Mexican plateau and whose geographic ranges extend

ing P. nana) eastward to west central Texas and southward to
high uis Potosi. The varieties of P. rigida are found only on the
ms ah Plateau close to the base of the Sierra Madre as well as in the
i = Cordillera, from central Chihuahua to Michoacién. P. Thurber,

12 BACIGALUPI

a Cordilleran species, extends in a narrow belt from southernmost
Arizona to Guanaxuato. A group of species, comprising P. reticulata,
P. Alamani, P. hebeclada, P. Dugesii and P. thyrsoidea, exhibit a more
southerly distribution, occupying the lower mountain slopes that rise
from the high southern plateau stretching from Guanaxuato to Oaxaca.
P. nudicaulis, the most southerly of the § Acourtia, is thus far known
only from Guatemala and El Salvador. The few remaining which
have not been accounted for are of restricted and widely isolated oc-
currence not coinciding with any of the general ranges just indicated.

GENERIC CONCEPT AND RELATIONSHIPS.

The vicissitudes through which Perczia passed during the early
years of its existence have already been outlined in the section de-
voted to the taxonomic history of the genus. The concept here fol-
lowed is that which has been adopted by workers on the group since
the interpretations of Lessing, Endlicher and more especially of Asa
Gray were ultimately expressed in full and definite form by Bentham
and Hooker in their Genera Plantarum. Even though the emphasis
was laid upon sets of characters other than those used by these authors,
the subsequent work of O. Hoffmann in Die Natiirlichen Pflanzen-
familien corroborated fully this earlier interpretation.

Most satisfactory of the characters which may be used to differ-
entiate the genera are the form of the corollas and the character of
the apex of and the distribution of the stigmatic papillae on the style-
branches. The usefulness of this last character, that of the distribution
of the papillae on the style-branches, was recognized by Hoffmann.
On these characters, the tribe is divided by Hoffmann into tir
sufficiently natural subtribes. By making use of additional dis-
tinctions (e. g. the heads homogamous as opposed to heterogamous,
anthers long-caudate or not, style-branches connate, connivent to the
tip, or shortly spreading only at the apex, as opposed to divided )
Bentham and Hooker succeeded in dividing the mutisioid genera int
as many as five perhaps somewhat more natural subtribes. Hof-
mann’s treatment, however, has the virtue of greater convenience
and clarity.

Perezia is relegated by these authors to the same and perhaps
closely interrelated of the subtribes—the Nassauvieae of Bentham am
Hoo er, OF the Nassauvinae of Hoffmann. The members of this
subtribe possess flattened, widely spreading and recurve style-
branches whose stigmatic papillae are largely confined to the apex,
more rarely run down the margins of the style-branches for 2 51°"

most

MONOGRAPH OF THE GENUS PEREZIA 13

distance as they sometimes do in Perezia. The corollas of the homo-
gamously hermaphrodite flowers are always distinctly divided into
the two highly dissimilar “lips.’’ Occasionally the marginal corollas
are tadiatiform. This combination of characters quite effectively
separates the Nassauvinae from either of the two other subtribes, the
Gochnatinae and Gerberinae of Hoffmann, as well as from the other
four subtribes of Bentham and Hooker.

Of the thirteen genera of the Nassauvinae, one, Polyachyrus, may
be at once separated from the others by its 2-flowered heads, since the
involucres of the other genera enclose from 4 to many flowers. Of
these genera, two are distinguished by their sessile heads. These are
Nassauvia and Triptilion. Of the remaining genera with pedunculate
heads, two, Jungia and Moscharia, are at once separated by their
chafty receptacles, Leuceria and Oxyphyllum have plumose pappus-
bristles, leaving the group of genera with simple pappus-bristles, of
which Perezia is one, still to be accounted for. Three of these have
tuneate style-branches. These are Perezia, T rivis, and Cleanthes.
Trizis is distinguished by its consistently yellow corollas and mostly
|-2-seriate involucre, Cleanthes by its mostly white corollas and uni-
serlate involucre, Perezia by its mostly white, blue, purple to pink
corollas and 3- to more usually many-seriate involucre. Leunisia, a
genus with yellow corollas and many-seried involucre, 1s distinguished

m these three genera by its rounded style-tips.

CLASSIFICATORY DIFFICULTIES.

Since the main portion of these studies on Perezia is devoted to a
ea eamhic treatment of § Acourtia, it is to this section that the
llowing few remarks have reference.

ugh the species in most cases are, in their vegetative characters,

ly differentiated, the discovery of trustworthy diagnostic char-
acters proved difficult. In the search for these, receptacle, achenes,
Pappus, and corollas were in turn carefully studied. The corolla, as
Was to be expected, exhibited very little variation. In only one case,
Be : P - collina, was the corolla of any value as a keying character,
ere it is anomalous for the genus. The receptacles, pappus. and

€s, on the other hand, were not nearly so stereotyped but the
.rences were found to be too inconsistent to be considered of any
et diagnostic value. The shape of the achenes, in the great
tty of the species, is highly variable, depending largely upon

. ‘Mount of compression they have had to withstand from their

bors. Immature achenes were almost always found to be more

14 BACIGALUPI

glandular or hirsutulous than the fully matured ones in any given
species. The same difference was found to apply to the receptacle.
The variations in the pappus were found to be few and in no case
usable, consisting mostly of color-differences, probably due in part to
age, and in the number of series of the bristles, a character thought to
be of generic importance by early writers. The pappus in a single
flower could be found to vary from 1-seriate on the portions of the
often very uneven epigynous disc that were narrowest, to 3-seriate
where it happened to be especially wide. Only the phyllaries remained
to be exploited and, fortunately, these were found to be most useful
in dividing the section into its primary species-groups. Except in the
case of only three species, the character of the apex of the phyllaries
is consistently either on the one hand rounded to obtuse or on the
other acute. The acute apex is that characteristic of the majority
of the species. In a few of these, the edges of the phyllaries taper
gradually from near the base, forming a very acute apex with or with-
out a slight terminal acumination. This type of phyllary has been
characterized as lance-subulate. The majority of the remaining spe
cies have the phyllaries of an oblong type, the sides being parallel or
even narrowing to the base, almost always with an acuminate tp.
Many of the species have the phyllaries either consistently glabrous
or glandular-puberulent. In the case of the three species falling not
clearly into either of the two main categories (i. e. with phyllaries
either unmistakably acute or obtuse), supplementary characters have
been found useful. Among these is that of the number of florets ™
the heads, first used by Asa Gray, a character here used to a very
linited extent and only when the disparity in number was consider-
able. Once these primary groups were successfully separated, the
mutual contrasting of the species so thrown together proved a rela-
tively easy task.

Economic Uses.

Mexican Indians, however, various species of Perezia have been in -
regard because of their medicinal properties. According to José pre
rez, Dat. Mat. Med. i. (1894) 63-78, the Indians used an infusion of the

MONOGRAPH OF THE GENUS PEREZIA 15

parently being those most sought. Definite reference was made by
Ramfrez to P. adnata (P. Alamani var. adnata), while on the label of a
specimen of P. oxylepis, Schaffner made a note of its purgative prop-

es.

According to a plant-label of Palmer’s, the matted woolly hairs at
the base of the stem of P. Wrightii were used by the Indians as a
styptic.

D1aGnoses OF NOVELTIES.

In order to preserve unity of presentation, it seems best to describe
here the novelties of § Acourtia treated in the monograph before their
presentation therein.

Perezia scapiformis, spec. nov. Herba perennis scapiformis cau-
libus floriferis sparse bracteatis sat altis rigidis insuper corymboso-
tamosis et foliis radicalibus ex rhizomate erecto vel horizontali sericeo-
tomentoso; radices fibrosae diametro 1.5 mm.; folia omnia radicalia
Petiolata rigida chartaceo-coriacea pallide viridia saepe purpureo-
tincta haud reticulata venis lateralibus haud prominentibus supra

Totundata circa 30-flora, 10-13 mm. alta; involucri squamae
— tigidae induratae subcoriaceae pallide virides —-
Iman BS 3 * 1 ° 4 ] atae

oblongne utroque leviter angustatae; corolla brevis glabriuscula,
f. ; pappi setae molles tenuissimae sordide albae 2-seriales
apicem versus parum clavellato-incrassatae; achenia circa 4 mm.

MEXICo:
Pry : Oaxaca: dry, hills) Las Sedas, elev. about 2000 m.,
mingle, no. 6015 (TYPE, Gr Ber gatesees N.’ Y.). PUEBLA: Tehuacin, Purpus,

16 BACIGALUPI

no. 5613 (Gr., N. Y.). PunsBua or Oaxaca: Cerro de Matzize, Purpus, no.
3923 (Gr., N. Y.).

Heretofore referred to Dr. Gray’s P. nudicaulis but quite distinct
from that species. In P. nudicaulis, all the phyllaries, even the outer-
most, are blunt and rounded at the apex, narrowly scarious on the
margins, glabrous dorsally and occasionally obscurely ciliolate toward
the tips. In P. scapiformis, many of the outer phyllaries are acuminate-
tipped, and these, as well as the blunt to narrowed and obtusish inner
phyllaries, are somewhat apiculate and more or less arachnoid on the
backs and margins. In addition to these distinctions in the phyllaries
(later emphasized in the key to the species), there are a number of
others to which attention may here be appropriately called. The
leaves of P. nudicaulis (particularly the terminal lobe) are usually
much wider, regularly triangular- to sinuate-dentate and runcinately
pinnatifid below the middle. Those of P. scapiformis are more or less
lyrately lobed, the segments ascending rather than retrorse, the mar-
gin ranging from very unequally and divaricately dentate to subpec-
tinately denticulate. Again, the heads of P. nudicaulis are 11-13-
flowered; those of P. scapiformis are about 30-flowered.

Perezia lobulata, spec. nov. (P. patens var. y. Gray, Pl. Wright.
i. (1852) 127.) Herba perennis altitudine ignota verisimiliter sat alta
ac robusta. Caulis teres glaberrimus laevis paullo flexuosus stramin-
eus vel saturate purpureus diametro 2.5-5 mm.; folia sessilia oblonga
rigida tenui-coriacea utrinque scabrida crebre reticulata supra parum
glaucescentia apice leviter acuminata basi haud angustata cordato-
amplexicaulia, alia auricula foliorum majorum ampliatorum saepe
super aliam excedens, crebre et saepe inaequaliter spinuloso-denticu-
lata ad basem versus interdum retrorso-denticulata vel grosse runcima-
to-dentata majora sinuate runcinato-lobulata vel runcinato-lobata
apicibus loborum plerumque acuminato-dentatis, ea caulis 9-11 ™.
longa, 27-40 mm. lata, ea ramulorum 4-6.5 cm. longa, 12-20 mm-
lata; rami divaricatim patentes 16-45 cm. longi, foliosi graciles
ad apices versus magis magisque densius pilulis articulatis crispisqué
pubescentes et valdius flexuosi; capitula 2-3 in cymulas ad apices
ramulorum foliosorum versus glomerulata 15-flora ca. 18 mm. longa;
involucrum campanulatum 11-14 mm. longum; involucri squamae
rigidiusculae ad basem versus induratae 4-5-seriales pallide virides
ad apicem versus saepe atropurpureotinctae magis minusve sericeor
ciliolatae, exteriores lanceolato-ovatae interdum leviter acumine®
interiores attenuatae oblongo-lanceolatae acutae vel acutiusculae haut

acuminatae apice saepe producto et involuto incurvatoque Tam

MONOGRAPH OF THE GENUS PEREZIA 17

obtusiusculae, 10-12 mm. longae; corolla rosea ca. 11 mm. longa tubo
insuper sparse glanduloso-punctulato lobis duobus linearibus inter-
ioribus ad basem versus glanduloso-ciliolatis; pappus rigidiusculus
2-3-serialig sordide albidus; achenia lineari-fusiformia atrobrunnea
dense glanduloso-punctulata ca. 5 mm. longa, disco epigyno angustis-
simo; receptaculum paullo concavum haud profunde scrobiculatum
sparsissime glandulosum ceterum glabrum. (Plate I.

MEXICO: Oaxaca: mountains, San Juan del Estado, alt. about 1920 m.
lwius C. Smith, no. 888 (rypB, Gr.). No STATE SPECIFIED: “Mexico. Fi.
rose, lemon-scented,”’ Galeotti, no. 2001 (type of P. patens var. y- Gray).

A species of the habit of and related to P. carpholepis (Sch. Bip.)
Gray, but to be distinguished from it by its thicker, coriaceous,
runcinate-lobulate to runcinately lobed cauline leaves, its much nar-
rower, more attenuate, acutish, often involute and incurved inner
phyllaries and 15-flowered heads.

Perezia dissiticeps, spec. nov. Herba perennis glabra ca. 10 dm.
alta; caulis simplex teretiusculus sat foliosus glaber laevis striato-

gustata, 2.5-6 cm. longa, 7-18 mm. lata; inflorescentiae ramuli
rein glabri subnudi bracteas minimas sparsissimas adscendentes
ceolato-subulatas minutim ciliolatas gerentes; pedunculi divaricati,
5-6 (plerumque 2-4) cm. longi, ad apicem summum minutim
slandulosi; capitula 14~18-flora, 12-15 mm. longa; involucrum (3-)
tseriale campanulatum 7-8 mm. longum; involucri squamae dorso
slabratae sericeo-ciliolatae, exteriores patentes lanceolato-oblongae,
Interiores oblongae ad basem versus valde incrassatae et parum
— apice purpureotincta sat variabiles obtusae saepe mu-~
a — vel acutiusculae; corolla rosea, 8 mm. longa, segments

basem versus et fauce tubi resinoso-punctulatis; pappus mollis sat

a punctulata, ca. 3 mm. longa; receptaculum planiusculum

Profunde alveolatum resinoso-punctulatum. (Plate III.)

Qo. 5018 CO: San Luis Porosf: Sierra Tablén, Minas de San Rafael, Purpus,
(typ, Gr. and isotype, N. Y.).

18 BACIGALUPI

This novelty is most closely related to P. Wrightit Gray. It may
be distinguished from that species by its absolutely glabrous, glauces-
cent and somewhat shining leaves and its distant, more abundantly
flowered heads not in glomerulate cymules.

Perezia aspera, spec. nov. Herba perennis rigidiuscula altitudine
4 dm. vel ultra; caulis simplex insuper ramosus teretiusculus diam.
1.5-2 mm. striatus glaber laevis in axillis foliorum tantum pilis pallide
brunneis sericeo-tomentosus flexuosus stramineus plerumque purpur-
eotinctus; folia sessilia vel subsessilia rigido-coriacea oblongo-lanceolata
apice rigido-mucronata basi cordato-amplexicaulia sat reticulata
glabriuscula divaricatim et irregulariter et pungenter dentata minutim
calloso-ciliolata, 4.5-5 cm. longa, 1.5-2 cm. lata, superiora minora
oblonga, 2-3 em. longa, 7-10 mm. lata; capitula 1-3 subcongesta
2-2.3 cm. longa 30-flora ad apices ramulorum elongatorum foliosorum
adscendentium gesta; pedunculi 8-20 mm. longi, plerumque bracteas
9-10 mm. longas adscendenti-patentes anguste lanceolatas vel linear-
subulatas glabras gerentes; involucrum campanulatum basi obtusum
ca. 3-seriale 20-23 mm. longum, pappum aequante; involucri squamae
basi tantum imbricatae rigidaeque haud induratae 2-2.3 mm. latae
attenuato-subulatae acutissimae glabrae striatae pallide virides ad
apices purpureotinctas versus interdum minutissime ciliolatae inter-
iores 17-19 mm. longae, exteriores rigidiores 8-10 mm. longae; corollae
segmenta ad bases versus sparse glanduloso-ciliolata, corolla ceterum
glabra rosea, 14-15 mm. longa; pappus mollis sordide albidus sets
tenuissimis 2-3-serialibus discum epigynum infundibuliformem coro-
nantibus; achenia linearia minutim resinosa ca. 5 mm. longa; T&
ceptaculum convexum profunde alveolatum membranis inter alveolos
fimbriato-glandulosis. (Plate IV.)

MEXICO: Coanuzza: Saltillo, alt. 1650-1800 meters (coll. Bro. Nil, n0- 55,
Dec., 1908), Arséne, no. 3395 (rrpx, Gr.).

(as opposed to 40-60-flowered) heads, in not having an acute rer
the involucre and in the absence of bristles on the receptnre

corollae plerumque glabratae vel rarius sparsissime glan
punctulatae; ceterum ut apud varietatem typicam. (Plate V.)

MONOGRAPH OF THE GENUS PEREZIA 19

MEXICO: Guanaxvuato: Guanaxuato, Dugés. Jatisco: Rio Blanco, Oct.,
1886, Palmer, no. 661 (Gr., N. Y.); cool hillsides near Guadalajara, Pringle, no.
1860 (tvpn, Gr. and isotype, N. Y.); rocky bluffs of a barranca near Guadalajara,
dey. 1525 m., Pringle, no. 9950 (N. Y.).

Perezta RicIpA (DC.) Gray var. acuminata, var. nov. Folia ac
habitus ut apud varietatem linearifoliam; involucri squamae angustae
praesertim intimae plerumque oblanceolatae supra partem mediam
paullo ampliatae sericeo-ciliatae apice acutae abrupte acuminatae;

i in capitulis 9; corollae ut apud varietatem linearifoliam;
achenia late fusiformia, 6 mm. longa, 1-1.5 mm. lata. (Plate VI.)

MEXICO: Jaxisco: rocky bluffs of a barranca near Guadalajara, elev.
1525 m., Pringle, no. 9950 (TyPE, Gr.).

PEREZIA MONTANA Rose var. intermedia, var. nov. Inflorescentia
e cymulis laxis 2-3-capitulis composita; pedunculi nudiusculi squamam
unicam gerentes vel nudi, glabri vel glabriusculi; involucrum campanu-
latum basi rotundatum vel obtusum, 9-10 mm. longum; involucri
squamae interiores oblongae basi haud angustatae apice obtusae vel
Totundatae interdum minutissime mucronulatae; ceterum ut apud
varietatem typicam.

qo: Duranco: Santiago Papasquiaro, Palmer (1896), no. 59 (TYPE,

Seemingly intermediate between P. montana var. typica and P
ngida var. typica. The phyllaries and involucres are almost identically
those of the latter species, while the general habit conforms very
closely with that of P. montana. x

REZIA Ducest Gray var. pilulosa, var. nov. A Perezia Dugesu
Yar. typica foliis prominentius reticulatis valde grossius inaequalius-
que dentatis obovato-spathulatis 9-11.5 cm. longis 5-7 mm. latis,
— densius glanduloso-punctulata et acheniis dense adpresso-
utis differt. (Plate VII.)
veins and veinlets of the lower leaf-surfaces are often much
More strikingly hirsutulous than in var. typica.

MEXICO: Oaxaca: . Lucius C. Smith,
m. 873 (ape, Gr), Vous Guise: Orazaba, Muller, 1855 (N.Y)
aoe course of these studies, specimens were examined from the
ee of the New York Botanical Garden and the herbarium _
States College, while a few types were forwarded from the Unite

S National Herbarium. These have been so indicated. Unless
i tely stated to the contrary, the specimens quoted without evi-

© of source are those of the Gray Herbarium. In the citation of

20 BACIGALUPI

specimens examined, the names of the herbaria referred to are ab-
breviated as follows: Gr. = Gray Herbarium; P. C. = Herbarium of
Pomona College; N. Y. = Herbarium of the New York Botanical
Garden; U. S. = United States National Herbarium.

Taxonomic ACCOUNT.

Perezia Lagasca, Amen. Nat. i. (1811) 31.

Heads homogamous, mostly many-flowered, occurring singly on a
scapiform stem or in loose to dense, bracteate to leafy, paniculate,
corymbose or thyrsoid cymes; phyllaries in 3—many series, broadly
oblong or ovate to linear-subulate, acute to obtuse, often mucronu-
late, the inner sometimes with very narrow to wide scarious margins,
the outer occasionally pungently ciliate-denticulate to ciliate-dentate,
rarely pinnatisect, often merely ciliate to ciliolate, green, often red-
dish- or purplish-tinged, herbaceous and lax to indurate, rigid or
subcoriaceous, loosely to densely imbricated, glabrous to densely
glandular, less often woolly or arachnoid; flowers all hermaphrodite
and fertile; corollas pink, purple, blue or white (yellow only in 2 species),
glabrous or the upper portion of the tube often resinous-punctulate to
stipitate-glandular, three of the segments coalesced to form the outer
(lower) ligulate lip, that of the marginal corollas often longer and wider
than that of the others, all terminally 3-dentate, the teeth sometime
minutely papillose, two inner (upper) segments free or readily sepat
able, terminally often minutely papillose, equalling the ligulate lip ot
somewhat to much shorter, erect to highly cirrhiform, often almost to
the base (in P. collina, the segments 5, subequal, linear) ; each of the
anthers with two long tails from the base, tipped by elongated, more
or less connate, somewhat rigid appendages colored above like the
corolla; style terete, glabrous, deeply divided at the summit into two
recurved branches, these flattened, with the stigmatic papillae col-
lected at the truncate apices so that these appear penicillate, or coca
ti ing a short distance down the margins of the flattened stylat
branches; achenes linear-cylindric to fusiform, sometimes costulate,

to abundantly hirsutulous as well with stiff upwardly di ected wer
trichomes (these, in one species, shortly forked at the sumo")
rostrate but

greyish-brown to dark brown or rarely even black, never Trost
crowned by a very slightly expanded to relatively broad disc up?
which the pappus stands; pappus bright white to tawny oF tet
purplish, of more or less abundant, simple, scabrous, soft to somew

rigid bristles, usually in 1-2 (sometimes partly in 3) series, sometime’

MONOGRAPH OF THE GENUS PEREZIA 21

fused at their bases; receptacle merely areolate to scrobiculate or even
alveolar, sometimes with the membranes between the alveolae
fmbrillate or glandular-fimbrillate, more rarely short-bristly, often
merely resinous-punctulate. _

West American, mostly short to dwarfed Cordilleran perennials,
or more rarely annuals, with herbaceous to subligneous aerial shoots,
the entire to pinnately parted leaves in a basal rosette and the
flowering stem scapiform, or the plants tall, caulescent, the leaves
altenate, mostly sessile, often crowded, sometimes densely so; the
more or less woody caudex often covered by a dense felty mass of
rust-colored wool.

KEY TO THE SECTIONS.

Flowers homomorphous, the ligulate lip of the marginal flowers
never any broader or longer than that of the inner ones; Hear

§ Acourtia.

Os a ie ee et

tials, always with the base of the stem more or less rusty-woolly, for
"et part tall and robust, the sometimes several usually very
Stems simple below the terminal leafy or leafy-bracteate inflores-

| wail ¢ ‘thate, mostly sessile, usually ample leaves thin-ch
and thick-coriaceous, generally finely reticulate, mostly oblong

ri uate, acute to obtuse, undivided, usually spinulose-denticulate
tale pe ose-dentate; or much less usually scapiform, the moderately
almost naked flowering stems mostly paniculately to less usually

22 BACIGALUPI

corymbosely branched above, the leaves basal, more or less rosulate,
semi-erect, petiolate, mostly pinnately to lyrately divided; heads
homogamous, often glomerate, 4-60 (or more) -flowered (mostly 8-15-
flowered); involucres obconic, turbinate or campanulate, rarely
widely hemispheric and truncate at the base, 3—-8-seried; phyllaries
mostly more or less rigid, indurate at least at the base, entire, often
sericeo-ciliate, more rarely arachnoid to tomentose on the margins,
mostly glabrous or glandular-puberulent, the outer mostly ovatish,
the inner, in general, oblong to lance-subulate, the tips sometimes
mucronulate, rounded and obtuse to attenuate-acuminate and very
acute, stramineous to green, very often purplish on the exposed
portions; corollas homomorphous, mostly lavender-pink and glabrous,
the upper portion of the tube often sparingly resinous-punctulate,
the inner 2-partite lip quite or almost as long as the ligulate outer
one, very little, if at all, cirrhiform; anthers, achenes, pappus an
receptacle as characterized in the generic diagnosis. The scabrous
processes of the pappus-bristles are not infrequently denser toward the
tips so that they appear slightly clavellate-thickened above—Acourtia
D. Don, Trans. Linn. Soe. xvi. (1830) 203; Clarionea § Palesia D. Don,
in part, loc. cit., 207; Damerilia Less. Linnaea v. (1830) 13, not Lag:
nor DC.; Prowhé § Thelecarpaea DC. Prodr. vii. (1838) 27.

Key To THE SPECIEs.
A. True foliage arising from the eaudex, the flowering stems
pecans but often iarinig bracteal or scale-like leaves. . . . B.
B. Phy oe es acuminate, lanceolate or narrowly oblapoealate
C. "Phyllsries glandular-puberulent, mostly irregularly cilio-
late, linear- meee t . inner setaceous-acuminate; :
flowering ste 1(-3)-head Weck veces ce eee 1. P. runcinaia.
C, — er short-acuminate, the
rte oblanceolate, the outer obovate; eeabeleh ay ireeell
B. Phyllaries obtusish to rounded at the apex, oblong to
ear, only the shortest outermost occasionally very

D. At least the short outermost phyllaries somewhat apicu-
late and often shortly cuspidate-acuminate; the larger
phyllaries ray semstme pubescent at least on the
margin, scariou
E. Inner ‘phyllaries at most mucronulate from the blunt
apex, sericeous on the margin, otherwise glabrous,
uniformly thin-membranaceous; heads at anthes# mbratilis.
saviowiy: campanulate, 15-20 mm. high........-- P. u
E. Inner phyllaries, as well as the outie, mostly aii
from a blunt to narrowed and obtusish pont pe
pecially the — arachnoid on | the back and m
all firm and su at anthesis Prowl ‘forms.
campanulate, 10-13 mm. bie te kagiet 1S ee ae _ P. scapif

MONOGRAPH OF THE GENUS PEREZIA 23

D. a “me phyllaries, even the erage blunt and rounded
e ‘ain narrowly scarious the margins, glabrous,
mes obscurely ciliolate toward the tipes5 7. 5. P. nudicaulis.
A, True foliage mostly on the flowering stems and branches;
eaulescent, often with many leafy shoots from the

aie
F. Cauline leaves runcinately lobed, those of the branches
often retrorsely denticulate ‘to coarsely run ncinate-
dentate toward the base, gid and shincearianenes ner-
wit

somewhat involute tip.........--.--.++sesser steers . P. lobulata.

F. Cauline leaves not lobed, the dentations never rire;
phyllaries neither favoliies nor incurved at the tip :
G. Phyllaries squarrose at tip, herbaceous........--+-+-> 7. P. Lozani.

G. Phyllarie not squarrose, scarious to indurate, more or

ess Ti
H, Bis liavice obtuse to rounded at the apex, often aor
ome rai never in the leas

oblong to o vane er er 8. P. carpholepis.
I. Inflorescence “essential y single and terminal... .J.
J. He 4-6 y 5)-flowered; leaves not cor-
San daaplun: 0 othe eolate to oblong-spatulate,
arely

aggre-
orp ars d-topped ada th
g into a round-topped pyr dae P. reticulata.

J. Heads 8-35-flowered; leaves more or less cordate-
nee: at least the cauline not narrowed to

K. Heads 30-35-flowered, broadly campanulate, at

least 2 cm. broad, 2.5-3 cm. high, o rigid
ge plus iorses MP, wat

K. as A cor gig narrowly . cone a
onic, at

(usually 6-10 ram} broad... ..
riaceous or ’ rigid-cartilagin-
ous, glabrous

M. Heads 19-22-flo wered; leaves rigid-car-
tilaginous, spiny-toothed, - teeth nar-
3.5 mm. lon :
rowly triangular, 2-3.5 mm. Jong. var. inter iB
M. Heads D-h4-flewareds leaves thickly cori-
ose-d ae

phyllaries often with a more or less
rusty tomentum on the margin 1 P. platy phylla.

ear-O

times a age Spar : phyllaries often

minutely glandular-ciliolate oF less bee
sericeo <7 te eae ee et 17. P. rigida.

often ilia
L. Leaves neither thick-coriaceous nor rigid-

acutish to ac

BACIGALUPI

cart tr mostly a = chartaceous
to less often rigid-charta
O. Leaves Futesta: heads 2 pe " glomerulate,

a , on rigid, slender but no

pumas cordate- to eben

Oe ae ee a

tip, neither lance-subulate nor attenuate, though
1 Baa
; = = thick-coriaceous to rigid-cartilaginous,
gia
7. Phyllaries oblong-lanceolate, acute, the more
umerous basal densely 1 sabeicateds
gradating into the want bee lt Ye uamae;
base of the involucre acute, more or less 0 ob-
conic; soba rigid-cartilaginous. spiny-
oo = spicuously bluish-green, glaucous

M1. P. patens.

P. Wrightii.

3. P. Coulteri.

R. Phyllaries not at all acuminate, not even at the very

OWwered 6.2245 ilu, 5. montana var. typica.

often broadly so, bloc: (if
at all) ebro the basal neither more numer-
0

phyllaries often with a a or less =
tomentum on the margin........---- 1

U. Leaves elliptic- to iene ten
= little spatulate; phyllaries often minutely

P. platyphylia.

dular-ciliolate or less often Soren P. rigida.

Peer ee
Oe Oe a eee ee 6 i oe ee we a eee

_y laginous, poet thin and chartaceous
quite glabrous, glaucous to glaucescent

MONOGRAPH OF THE GENUS PEREZIA 25

at least i the span a little shiny above;

ecien 138

; heads not app ig

) em. long th. ap: dissiticeps.

VY. Leaves mostly glandulaspaberdiant, nt ‘at all

ot more than 3 series;
1-

. Leaves sessile, not narrowed at the

oblong to ovate,

base,
the lower not infrequently Saw
sagittate-clasping...... Ee P. Wrightiv.
c

lanceolate, mostly narrowe to the some-
— abruptly aware base, neither at all

auriculate nor clas
R. Sy Sora sy mucronat

ings 45 eee es 13. P. Coultert.
S aprelite to acuminate,

orem subulate, attenuate or even caudate-

a wAt | ae ‘the inner phyllaries caudate-attenuate,
‘so slender tips not rigid and somewhat tortu

¥. Whole plant covered by a dense whitish to-
mentum; heads borne singly on a a
leafy pedunele; phyllaries scarious......-- . P. tomentosa.
Y. Plant + most vi scid-puberulent, bright nie
rymbif

in decent N ie Mla nce a
phyllaries peels more or me ghar?

Rualecs var. urolepis.

X. None of the dhe ag 8 he caudate-attenuate. .. .Z.

Z. Plant 1-3
near the bas

]

— ra anched ig
es terml

ated b
“acy relatively large, canal Sas auteonaet

a. Leaves mors or less cuneate at the base, not
7 mm

in width, mostly nar-
soca amber-colored

a. Leaves ee rounded to eeoital at the

g to be orbicular, 2 2-4(-5) cm.

wide; ones more or less densely and
sheets atipi eo meuldla®. oe oe ees 20. P. nana.

Z. Plant tales, meee robust, usually ma es
b. Heads 3.5-5 ¢

ed above. . so. ia. ee bee OL P. Wislizeni.

b. ~— or upper branchlets terminated by an
orescence of smaller ~ much narrower

sess.

C. At Teast the j inner

payee lance-subulate

¢tenuated triangle in out-
_ line) very gradually narrowed. . . .

. Heads congested

in a dense very le sl

elongated cylindric thyrse; involuc

= obconic; peduncles “oa

22. P. thyrsoidea.

eee ee ee
foe le eee a ee eee

BACIGALUPI

d. Inflorescence not a leafy cylindric thyrse .. . .¢.
e. Peduncles densely squamose, th ee
scales more 0 i i

acutish to acute at the raion VS 3
f. Head 3 . high, glabrous
to glabrate; involucres at anth

conduplicate, linear-oblong, es-

ava glabrous save on the

Pei? PA eae Pte 23. P. formosa.
q. Piyilarice attenuate-acuminate,

ae beneath on the vein-

me hasan arcuate nor —

f. kore aot more than 2.5 cm. high,

to woolly, not at all con-

gested, obconic at least at the
base; phyllaries a serious, in-
be

=]
a)
Qu.
ae
So
ic)
i
=
2
>
a

A? oe ne iT laries peta ei Te

chnoid, at least on the mar-

Ve ae
s, gs. oval to lance-ovate, ir-

: ea
i. Leaves

gent-denta
ielee: of the thickish pedun-
cles soon becoming leaf-like;
ememnrenee not ake —
oblanceolate- to Bi
oblong, finely and regularly

- g leaf-like below; nflores-
ence widely open-paniculate,

10-12-hea ri toed Senge rare ere
e. Peduncles, even when sascha squa-
imbrica
— c to oes turbinate or
aed at the base... .
j. dnvolnares aa rage bie, mostly
te) nang gee much ex-

ceeded by cee vn stems =
with the decurrent eaf-bases. .

P. longifolia.

P. lepidopod

P. Purpusit.

rachnolep's-

98, P. platypler®

MONOGRAPH OF THE GENUS PEREZIA

j. Involucres 14-22 m m. high, —
equalling or little pratt 2 by th

NT leaves not decurrent. . . 'E.

k. Inflorescence a broad-topped or

a and more or less arene
ee ee

z Tflorescene op ay nicula a

e than ve eaded; phy!

NEO Wg OF ta.

m. high.....-+--«+--
L. Phyllaries subulate-acumina
n 4-5 series, fairly well im-
brieated; heads often 28 mm
hgh. chun cee
Cs vs senda not “ance subulate, a al-
, from la anceola ate to ob-

—- a mucronate-acum!
n. At “een ae jnnermost phyllaries atten-
uate-acuminate. . . -0-
. Hea 2 4-6-flowered; Lg oe 1-1.5
wide, minute Ce ag
i a ets

ie ewes 32.

P. hebeclada.

P. turbinata.

_ P. Thurbert.

_ P. Pringle.

BACIGALUPI

-obconic;
SORE 555 i. Sins, iO O35. P. Dugesii.

ed the stem; phyllaries often 2 mm.
r more broad, nidbrats to glandu-
cetiaet pubes ss yes 36. P. Alamani.
q. Cymes lax to i caneun at a
events well exceeding the sedueal
terminal foliage; —e) not at al
aaeate to “pe a. oe

mostly 2-3 mm. broad; leaves
oe not crowded, :
juke seeks cee 37. P. Palmer.

a Phyllaries s objong-ianceolate to

“chiar , 2 mm. or less
a Phyllaries soblong-lanceolate;
fruiting _ ds

uU. Phyillaries very acute,

adually attenuate-acu-

minate, 1.5-2 mm. broad,

lance-oblong.......-- 38. P. oxyleps.
u. Phyllaries acute, not grad-
ually but rather abruptly
narrowed at the acuml-
nate tip, 1-1.7 mm. br oad,
indurate; leaves aaa
sean, pect.

wide, not amplexicaul;
vom co at on bE wae

eaetvoeeae

shor
ere base, cordate-
amplexicaul; py
laries 1-1.7 mm. wi ee granaifoli

t. Phyllaries oblong-oblanceo-

MONOGRAPH OF THE GENUS PEREZIA 29

late, abruptly eh and
often very sharply acumi-
nate hartecsdue @ not in-
durate; fruiting heads 10-13

mm. high... 53.3 6355 41. P. microcephala.
r, Phyllaries ‘not F gianduler, dorsally
gla ee to deciduous-lanate. . . .w.

rollas at anthesis “deeply
ed into 5 linear segments;
cabs: ben or te ee

6 eee Se ee BE a he ee eee

42. P. collina.

z.
x. Fruiting heads 16-22 mm.
high; payee at most 2
Y. Gober portion ‘of the slen-
der stem, except ors sa
sparse _ highly
bract-like Teavek. dest,
tute of foliage....-.- nudiuscula.

P. rigida var. acuminata,
aa. tne lanier hemi-
spheric -campanu-
late, fulvous-tomen-
ose; h —_

merulate lea
Shorain ‘spatula, :
cuate. .43. michoacana.

ar
2. oe obconic at in
basal phyllar-

ting into uncu-
lar scales; heads not glo-
merulate; foliage blu-
i een, conspicuous-
ly glaucous. e .
15. P. montana,var. typica.
x. Fruiting heads 2.5-3 em
turbinate; broadest phyllar-

30 BACIGALUPI

ies 3.5-4 mm. broad, oblong-
lanceolate to oblanceolate;

them with a petiole 0.5-2
MN OUR a ce 44. P. cuernavacana.

1. P. runcinata Lag. ex D. Don, Trans. Linn. Soe. xvi. (1830) 207,
in syn.; Clarionea runcinata (Lag.) D. Don, loe. cit.

Acaulescent perennial from an elongated, erect, rather sparingly
woolly caudex, this giving rise at its lower end to fascicles of elongated
fusiform tuberous roots 8-15 cm. long; aerial shoots borne laterally on
the caudex; leaves rosulate, the leaf-blades chartaceous, oblong-
oblanceolate in outline, pinnately parted into rounded, spreading,
often undulate lobes, irregularly spinulose-dentate, glandular-scabrous,
especially beneath, hispidulous on the midrib and veins beneath,
15-17 cm. long, 4.5-7 em. wide, narrowly decurrent and gradually
decreasing along the petiole, this 1.5-7.5 cm. long, usually pilose at the
base; scapes rigid, slender, usually bracteate above, glandular-scabrous,
from much exceeding (33 cm. long) to about equalling (8-20 cm. long)
the leaves, sometimes divided at the summit; bracts scale-like, subu-
late, as much as 5 mm. long, mostly shorter; heads 20-25 mm. high,
about 40-flowered; involucres campanulate, 15-18 mm. high; phyl-
laries linear-lanceolate, the short outer ones tending toward ovate,
gradually to somewhat abruptly acuminate, glandular-puberulent,

7-10 mm. long; receptacle moderately alveolate, in immature heads
densely bristly, becoming less so.
South central Texas, southward in Mexico to Hidalgo.

UNITED STATES: Texas: bluffs above Austin, April 1, 1848, Wes on
09, in part; Mt. Burnell, Austin, Hlihu Hall, no. 374 (Gr., N.Y. P.-C).
the Rio Grande, 1848, Wright, no. 409, in part; stony hills of the F beasesss

z . e

well Crossing road, Brewster County, V. L..Cor no. 2807; Uvalde,
western Texas, Palmer, no. 769 ( 1879-1880) $ st phasis hills near Eagle iy
- elev., Pringle, no. 8280 (Gr., N. Y., P. C.); Starr County, July Croft
R. Runyon (Gr., N. Y.); San Diego [Duval Co.], 1885-1886, Mary Gr., N-

Y.); along Corpus Christi Bay, Nueces County, Heller, no. 1SST 9

Pope. : 76 (Gr
CO: Tamautipas: “prope Matamoros,” Berlandier, no- 31 Nuevo
N. Y.); vicinity of Victoria, alt. about 320 m., Palmer, no. 473 (1907). Bishop’
ae limestone hills near Monterey, elev. 615 m., Pringle, no. 1 741; no. 44
ill, Monterey, Feb. 6, 1847, Gregg; Cerralvo, May 29, 1847, Gregg B™

MONOGRAPH OF THE GENUS PEREZIA oO

(Gr. N. Y.). Coanuma: Saltillo, Palmer, no. 86 (1898) (Gr., N. Y.). _Dv-
Rango?: woods, Ramos-Carrizitos, Dec., 1852, Thurber, no. 860. San Luts
Porosfi: Agua Media, Purpus, no. 47 Hipatgo: Iximiquilpdn, dry slopes,
July, 1905, Purpus, no. 1167, in part (Gr., P. C.); same locality, rocky slopes,
Sept., 1905, Purpus, no. 1167, in part (Gr., N. Y.).

According to Dr. Gregg’s label, locally known in Nuevo Leén as
cardito, which signifies “ little thistle.”

In no specimen seen by the writer were the leaf-segments reflexed;
hence the leaves cannot truly be characterized as runcinate, the
specific name notwithstanding.

2. P. nudiuscula Robinson, Proc. Amer. Acad. xliv. (1909) 625.

Basal portion of stem as yet unknown; stems erect, evidently scapi-
form, rigid, slender, teretish, finely to obscurely striate, smooth,
purple-tinged, glabrous except for small tufts of fulvous tomentum in

e axils of the very sparse bract-like leaves, these sessile, linear-
spatulate to linear, glabrous, pale green and purplish-tinged, rather
fim and thick, ascending to erect, the larger subrevolute, sparsely
denticulate, especially toward the often subamplexicaul base, 2-4 cm.

ng, 2~5 mm. wide; inflorescence laxly corymbose-paniculate, widely
spreading, the very slender, sparsely subulate-squamose peduncles
12-65 mm. long; lower inflorescence-branches ascending, subtended
by linear bracts 6-9 mm. long; scales sparse, subulate, appressed to
merely ascending, 1-3 mm. long; heads ultimately 15-20 mm. high,
12-13-flowered, the larger as much as 20-flowered; involucre campanu-
late, 8-11 mm. long; phyllaries in 4 (occasionally 5) series, indurate
only at the very base, thin and cartilaginous, all but the innermost

a va long, densely glandular-punctulate and hispidulous with as-
one whitish trichomes, the epigynous disc light yellow and funnel-

N. :e ICO: Nayarit: Tepic, Palmer (1892), no. 2018 (vype, Gr. and isotype,

More complete specimens of this interesting species are much to be
- The general appearance of the type and isotype suggests

32 BACIGALUPI

that it most likely is one of the scapiform Perezias related to P.
nudicaulis.

. P. umbratilis Robinson & Greenman, Proc. Amer. Acad.
xxxii. (1896) 50.

Acaulescent perennial with aerial shoot-producing caudices and
numerous stout fibrous roots from a horizontally creeping rhizome;
caudices 8-12 mm. thick, sessile on the rhizome or raised 2-4 cm.
above it on a slender stipe-like base, multicipital, bearing in the axils
of the radical leaves, but especially at the top, tufts of sordid white
sericeous hairs; leaves all radical; petioles 1.5-5.5 em. long, ultimately
glabrous, very narrowly alate with the decurrent base of the blade;
leaf-blades thin-membranaceous, deciduous-pubescent with long rela-
tively sparse weak silky hairs similar to those of the caudex, ultimately
glabrate to glabrous, not at all reticulate, elliptic-oblong in outline,
lyrately pinnatifid, 7-11.5 cm. long, 3.5-4.5 cm. wide, 5-7-lobed, the
large terminal lobe ovate to orbicular-ovate, slightly produced at the
blunt and mucronulate apex, sinuate-dentate, each tooth sharply mu-
cronulate; lower segments very shallowly sinuate, distantly mucronate-
denticulate; scapes 2-4 from a caudex, erect but very lax and slender,
17-30 cm. high, divided above into 2-3 single-headed peduncles,
obscurely angled, darkened (in drying?), smooth, glabrate to sparingly
deciduous-pubescent with rather long weak silky hairs and puberulent
with whitish crisped hairs above, bearing minute, thin, appressed,
subulate, usually puberulent bracts, these more abundant and more
closely approximated on the ultimate (2-4.5 cm. long) peduncles;
heads narrowly campanulate, 15-20 mm. high, about 18-flowered;
phyllaries in 5-6 series, uniformly thin-membranaceous, deep purplish,
glabrous dorsally, more or less sericeous on the margins, appearing
oblong when imbricated but narrowed toward the base, obtuse ©
rounded at the apex, often mucronulate, the basal ovate-oblong, only
a few of the very outermost acutish, rarely the slightest bit acuminate;
corollas lavender-pink, glabrous, 12 mm. long; pappus veTy tawny ©

rownish, very fine and soft, 1-2-seriate; achenes immature) -
brown, subcylindric, copiously covered with gemi-appressed ase
white hairs, stipitate-glandular just below the short light-coh
epigynous disc; receptacle highly convex, densely white-bristly-

MEXICO: Oaxaca: in shade, Tomellin Cafion, elev. 920 m., Pringle, 0°
5966 (TPE, Gr.).

4. P. scapiformis Bacigalupi (see p. 15).

Acaulescent perennial with rigid, relatively tall, scapose, ©" |
bosely branching flowering stems and radical leaves from a0 erect

=

bret

MONOGRAPH OF THE GENUS PEREZIA 33

horizontally creeping, deciduously silky-tomentose rootstock; roots
fibrous, 1.5 mm. thick; leaves all from the basal caudex, petiolate,
rigid, chartaceo-coriaceous, pale green often tinged with purple, not
at all reticulate, the lateral veins of the upper surface hardly notice-
able, often slightly scabrid above, deciduously long-sericeous on the
midribs, otherwise glabrous, the midribs, too, at length glabrous, from
very unequally and divaricately mucronate-dentate to subpectinately
denticulate, often ciliolate as well, the blades sometimes undivided,
abicular-obovate to oval-obovate, 3-5 cm. long and 3-3.5 em. wide,
or more usually oblanceolate in outline and more or less lyrately lobed
or parted, 6-10 cm. long and 18-30 mm. wide; petioles 1.5-7 em. long,
mostly between 1.5 and 2.5 em. in length, narrowly winged, especially
above, by the decurrent leaf-blade, deciduously sericeo-lanate ;
flowering stems erect, rigid, 1.3-2.3 mm. thick, naked except for much
reduced, sparse bracts, 2.5-6 dm. tall, striate, smooth, glabrous above
the very base, paniculately to corymbosely branched in the upper
to fifth to form the inflorescence; bracts of the scape from a broad
base, Iinear-lanceolate, more or less arachnoid; heads laxly disposed on
divaricate, somewhat arachnoid peduncles 1.5-3 cm. long, broadly
campanulate, rounded at the base, about 30-flowered, 10-13 mm.
high; phyllaries in 4 series, rigid, firm and subcoriaceous, light green,
often dark purplish on the exposed portions, especially the mostly
“tuminate outer lanceolate to ovatish, arachnoid on the backs and
margins, all for the most part more or less apiculate, the inner from a
blunt to narrowed and obtusish apex, these oblong, somewhat nar-
towed at both ends; corolla very small, the segments especially short,
sather widened at the throat, 6 mm. long, probably lavender-pink,
perolly glabrous; pappus soft and fine, slightly tawny, not dense,
9 series, more densely scabrous toward the apex and hence appearing
. though slightly clavellate-thickened outward; achenes about 4 mm.
, linear-oblong, dark brown, whitened by the dense coat of up-
wardly directed short stiff hairs, very slightly glandular above, the
*pigynous disc very short and very slightly expanded; receptacle
— very shallowly scrobiculate, sparingly glandular and
ent.

Dube

MEXICO: oO AXA * P. ingle
: ca: dry cal hills, Las Sedas, elev. 2000 m., £ ringre,
S13 (a (tyre, Gr. ih Rooney, N Y.).. Purest: Tehuacin, Purpus, no.
Gr, NY Y.). Pursia or Oaxaca: Cerro de Mataize, Purpus, no. 3923

ri nudicaulis Gray, Pl. Wright. i. (1852) 127.
caulescent perennial, the underground portion a horizontally

34 BACIGALUPI

creeping rhizome giving rise to thick fibrous roots and erect caudices,
these in turn bearing the aerial shoots; caudices more or less densely
brownish-tomentose, 7-17 mm. thick; leaves on pubescent petioles
1-10 em. long, thin-membranaceous, rarely somewhat coriaceous,
somewhat rigid in the dried state, pubescent with long weak hairs on
the basal part of the lower midrib, otherwise glabrous, obscurely and
not closely reticulate, rather regularly, closely and prominently tri-
angular- to sinuate-dentate, the teeth mucro-tipped, somewhat
roughened to decidedly scabrous above, lance-elliptic in outline,
runcinately pinnatifid below, 5 (-7)-lobed, the blades most usually
9-18 cm. long, 4-7 cm. wide, the undivided terminal lobe comprising
at least half the length of the leaf, broadly ovate to lance-ovate, often
truncate at the base, occasionally, in very much reduced leaves, only
the terminal lobe present; the lowest lobes decurrent for 1-2 cm.
along the petiole; scapes 1-2, erect, rather slender, 2-6 dm. tall, striate,
glabrate, smooth, bracteate, paniculately branched above into a 20-
60-headed cyme; bracts of the scape scale-like, sparse, lance-subulate,
pubescent, or at length glabrate, 3-12 mm. long; heads often closely
approximated but not congested, 8-12 (mostly 10) mm. high, 11-13-
flowered; involucres slenderly campanulate (when not crushed), 7-8
mm. high; phyllaries oblong, the inner linear-oblong, not narrow
toward the base, the outer broadly oblong, almost oblong-ovate, all
bright green and glabrate, entire, scarious on the margin, blunt
rounded at the sometimes obscurely ciliolate tips, imbricated in 3-4
series; corollas lavender-pink, glabrous; pappus purplish-bronze, ¢-
pecially toward the base, very fine and soft, in 1 series; achenes linear-
cylindric, 5 mm. long, striate, evenly but sparsely appressed-hispidu-
lous, the epigynous disc very slightly developed; receptacle flattish,
moderately alveolate, the ridges between the pits fimbriate and rather
sparsely bristly.

MALA: Chee ion of TYPE)
ALA: No DEPARTMENT SPECIFIED: Skinner Napebess te) Paige 3

aec. . Se ; he

John Donnell Smith, no. 2364. Basa Verapaz: Sierra de las Minas, N. Y,)
m., opposite El Rancho, Jan. 5, 1908, W. A. Kellerman, no. Mak m.
(leaves somewhat coriaceous). Santa Rosa: Buena Vista, alt. 1

6. P. lobulata Bacigalupi (see p. 16). P. patens var. 7. GT
Pl. Wright. i. (1852) 127.

Erect, leafy, robust, evidently tall perennial; stem terete, quite

I Re Te ee ee ee ne

MONOGRAPH OF THE GENUS PEREZIA 35

smooth and glabrous, slightly flexuous above, stramineous to dark
purplish, 2.5-5 mm. thick; leaves sessile, oblong, rigid and thin-
coriaceous, scabrid, closely reticulate-veiny, the slightest bit glauces-
cent above, shortly acuminate, not at all narrowed toward the cordate-
clasping base, the ample auricles of the larger leaves frequently over-
lapping, closely and sometimes unequally spinulose-denticulate, often
tetrorse-denticulate to coarsely runcinate-dentate toward the base,
the larger cauline leaves 9-11 em. long, 27-40 mm. wide, sinuately
runcinate-lobulate to runcinately lobed, the tips of the lobes usually
acuminate-dentate; leaves of the branches 4-6.5 cm. long, 12-20 mm.
wide; the leafy inflorescence-bearing branchlets spreading almost at
right angles from the main axis, 16-45 cm. long, slender, gradually
more and more pubescent toward the tips with crisped articulated
hairs and becoming more pronouncedly flexuous; the glomerulate 2-3-
headed cymules on short lateral leafy-bracted common peduncles,
isolated from one another by the leafy portions of the branchlet-tips;
heads 15-flowered, about 18 mm. high; involucre campanulate, 11-14
mn, high; phyllaries somewhat rigid, indurate at their bases, in 4-5
series, pale green, often tinged with dark purple toward the tips, more
or less sericeo-ciliolate, the outer lanceolate-ovate, sometimes shortly
acuminate, the inner attenuated, oblong-lanceolate, acute or acutish
but not acuminate, 10-12 mm. long, the apex often produced, involute
and incurved, more rarely somewhat obtusish; corolla lavender-pink,
about 11 mm. long, the upper portion of the tube sparsely glandular-
punctulate and the 2 inner segments glandular-ciliolate toward the

*; Pappus somewhat rigid, in 2-3 series, tawny; achenes linear-
tusifo , dark brown, densely glandular-punctulate, about 5 mm. long,
‘umounted by a very narrow epigynous disc; receptacles slightly
Concave, shallowly scrobiculate, glabrous save for a few amber-

red glands.

MEXICO: Oaxaca: mountains, San Juan del Estado, alt. 1938 m., L. C.
h, ho. O STATE sPEcIFIED: ‘Mexico. Fl. rose, lemon-

ho. 888 (Typr).
“ented,” Galeotti, no. 2001 (type of P. patens var. y. Gray).

ag of stem teretish, striate, flexuous, more or less scabrid and
i ‘

‘tamineous and purplish-tinged; leaves thick-chartaceous, sessile,
“rdate-amplexicaul, mostly shortly acuminate and acute, oblong-
Vate, some of the u permost more nearly oblong, unequally spinulose-

tate to -denticulate, scabrous, the midribs often purplish, ob-

36 BACIGALUPI

scurely reticulate, dark green and slightly puberulent on the midrib
above, evenly glandular and densely hirtellous with crisped articulated
hairs on the prominent reticulations beneath, 3-7 em. long, 17-32
mm. wide, those immediately subtending the cymules smaller;
cymules few-headed, lax to subcongested at the tips of the terminal
leafy branchlets, forming a corymbose to paniculate cyme about 28
em. wide; heads on densely glandular-pubescent peduncles 3-20 mm.
long, 23-25-flowered, 12-17 mm. high; involucre hemispheric-cam-
panulate, 9-10 mm. high; phyllaries distinctly herbaceous, greenish
to purplish, rigid only centrally toward their bases, mostly oblong,
the apices highly variable, from truncatish-rounded or rounded and
mucronate to tapering and acute, the outermost often quite foliaceous,
lanceolate or oblong-oblanceolate, all more or less squarrose and dense-
ly glandular-ciliolate, glabrate below and more or less hirsutulous
toward the apex dorsally, densely papillose-glandular toward the tip
on the inner reflexed exposed surface, the bases well imbricated, in
5-6 series; corollas essentially glabrous, lavender-pink; pappus tawny,
somewhat rigid, mainly uniseriate; achenes fusiform-cylindric, cos-
tulate, 4-5 mm. long, covered with amber-colored subsessile glands.

MEXICO: Himatoo: under dry cliffs between Metepec and Zontecomate
Pot elev. about 2600 m., Pringle, no. 8871 (rrPx, Gr. and isotypes, N. Y»

A well marked and apparently little collected species.

8. P. carpholepis (Sch. Bip.) Gray, Proc. Amer. Acad. xix. (1883)
60; Acourtia carpholepis Sch. Bip. ex Gray, in syn., loc. cit.; P. patens
var. 8. Gray, Pl. Wright. i. (1852) 127.

Leafy and evidently divaricately branched, probably somewhat
scandent perennial; upper portion of stem flexuous, obscurely angled,
slightly scabrid to smooth, stramineous, often purplish-tinged; upper
branchlets leafy, flexuous, stramineous to purplish, smooth to scabric,
minutely crisped-puberulent, especially toward the tips, and often
with minute articulated purplish trichomes as well; leaves chartaceos;
sessile, oblong or lance-ovate, deeply auriculate- to subsagittate-
clasping, acute, shortly to gradually acuminate, closely and somewhat
unequally spinulose-dentate, more especially toward the base, to
spinulose-denticulate, beneath somewhat prominently reticulate,
scabrid, puberulent and glandular-punctulate, above not at all prom’
nently reticulate, glabrate save at the base of the midrib, 5-10 c™
long, 2-5 em. wide; inflorescence-bearing branchlets arising sa
axils of the leaves toward the branch-tips, the cymules subconges"”
2-4-headed; ultimate peduncles fairly densely pubescent and glandu-

SE

MONOGRAPH OF THE GENUS PEREZIA 37

lar; heads 10-11-flowered, about 15 mm. high when fully developed;
involucre widely obconic-campanulate, 8-9 mm. high; phyllaries
glabrous dorsally, obtuse, often conspicuously purple-tipped, some-
what indurate, slightly to copiously sericeo-ciliate, loosely imbricated
in 4-5 series, the outer basal spreading, ovate to suborbicular-ovate,
the inner oblong; corollas lavender-pink; pappus moderately rigid,
biseriate, some of the bristles very slightly clavellate-thickened above,
the bristle-bases usually highly fused; immature achenes subcylindric,
densely studded with amber-colored glands, the epigynous disc slight;
receptacle convex, only slightly scrobiculate, glabrate to puberulent.

MEXICO: Pursia: Chapulco, Dec., 1841, Liebmann, no. 351. Cuxtapas:
Linden, no. 439; in pine forests, Ghiesbreght, no. 525. All these specimens
are co-types, being cited after the original diagnosis by Gray.

9. P. reticulata (Lag.) Gray, Pl. Wright. i. (1852) 128; Proustia
reticulata Lag. ex D. Don, Trans. Linn. Soc. xvi. (1830) 200; Perdicium
serrulatum Sesse & Moc. ex D. Don, in syn., loc. cit.

Rigid, simple, erect, suffruticose perennial, 1-1.5 m. tall; stems
glabrous, more or less sulcate, densely leafy; leaves sessile or rarely
having petioles 1-2 mm. long, inclined to be ascending and imbricated,
coriaceous, cuneate-obovate to elliptic-oblanceolate, mostly gradually
narrowed to the base and strictly exauriculate, obtuse and rounded
at the apex, sometimes abruptly acuminate and mucroaulate, rarely
truncate to slightly sagittate or the least bit auriculate at the base,
finely reticulate with raised veins on both surfaces, acutely sinuate-
dentate, often less closely so toward the base, yellowish green,
slightly darker above, glandular-puncticulate and minutely hirsutulous
on the reticulations beneath, glabrate to glabrous above, 6-12 em. long,
1.5-5 em. wide; termina] lax to dense usually pyramidal thyrse com-
Posed of glomerulate 3-10-headed cymules, the ultimate peduncles
Pubescent; involucre cylindric-campanulate, 4—5-seried, 4—6-flowered,
$-10 mm. h
at the apex, glabrate dorsally, erose to fimbrillate-ciliolate; corollas
lavender-pink; exposed portion of pappus about equalling the involucre
m length, the bristles the slightest bit clavellate-thickened toward the
*pex; mature achenes costulate, fusiform-cylindric, 4-5 mm. long,
“piously stipitate-glandular; receptacle sparingly pilose.
fro plateau and mountainous region of south central Mexico

m San Luis Potosf to Oaxaca.

P CO: San Luis Porosf: “ex convalli San Luis Potosi, in umbrosis
in 7, uasco,”’ Schaffner, no. 374. GuanaxuaTo: hills near Acambaro,
m, ye Ho 4324 (Gr., N. Y.). MicwoacAn: vicinity of Morelia, alt. 1950
: no. 3155. Micuoackn or Guerrero: Sierra Madre, alt. 1200

38 BACIGALUPI

oronow
Pringle, no. 1 E : 3096;
Hacienda del Sr. Pesado, Orizaba, Sept. 1850, Fred. Muiller (N. Y.); Orizaba,
Botteri, no. 495; Muiller in 1855 (N. Y.). Purswa: vicinity of Puebla, alt.
1 :

m., us C. Smith, no. Pee itt near Oaxaca, alt. about m.,
Pringle, no. 5777. O STATE SPECIFIED: Fred. Muiiller (comm. H. Schlum-
berger) 1853, no. 1183 (N. Y.).
10. P. Nelsonii Robinson, Proc. Amer. Acad. xxxv. (1900) 342.
Erect, moderately leafy perennial 8-15 dm. tall; stems apparently
several from the caudex, slightly decumbent at the base, soon erect,
unbranched below the inflorescence, rigid or flexuous, striate, smooth,
purplish, glabrous except for tufts of fulvous tomentum partly hidden
in the leaf-axils; cauline leaves ample, firm, subcoriaceous, reticulate,
especially above, paler and somewhat scabrous below, oblong-oblance-
olate to spatulate, auricular-clasping, shallowly sinuate-dentate, the
teeth very acute and of unequal length, usually mucronate at the
rounded apex; inflorescence naked, diffuse, 5-10-headed, the ascending
rigid peduncles often 10-17 cm. long, bearing sparse, appressed, lance-
attenuate bracts; heads 2.5-3 cm. long, about as wide above, 30-
flowered; involucre broadly campanulate, 15-18 mm. high; phy llaries
well imbricated in 6-7 series, green, purplish on the exposed portions,
indurate, glabrous dorsally, sometimes ciliolate, the outer ovatish, a
trifle longer than broad, the outermost sometimes the slightest bit
apiculate, the rest broadly oblong, at least 2 mm. wide, obtuse, with
or without a small mucro at the apex, only 2 or 3 of the very innermost
lance-oblong and acute; corollas 1.5 em. long, lavender-pink; pappus
soft, not dense, interruptedly biseriate; achenes subcylindric, often
more or less compressed, appressed-hispidulous, covered with amber
colored sessile glands as well, 5-6 mm. long.
Apparently localized in Jalisco, Mexico.
MEXICO ) me ue

elson, no. . i j ?
Mace Oriental alt 3300 metry oly si Yee Mera
(Gr., N. Y.); Real Alto, rocky slope near crest of La Bufa, Sierra Madre,
2500 meters, Ynez Meria, no. 1623 (N. Y.).

11. P. patens Gray, Pl. Wright. i. (1852) 127; Trinis patens
Sch. Bip. in Seemann Bot. Voy. Herald (1856) 315; Acourtia formos®
Sch. Bip. loc. cit. and pl. 56, not D. Don nor DC. a !);

Erect perennial up to 3.5 m. tall (acc. to Mrs. Mexia’s a
stems very smooth, glabrous, glaucescent to glaucous, straw-colo

MONOGRAPH OF THE GENUS PEREZIA 39

streaked with maroon; leaves rigid-chartaceous, sessile to subsessile,
broadly oblong to elliptic-oblong, the uppermost inclined to be ovatish,
rounded to acutish at the apex, broadly auriculate-clasping at the
base, loosely reticulate, more prominently so beneath, glaucous,
especially beneath, or sometimes glaucous only on the lower midrib,
smooth, glabrous, closely spinulose-denticulate, mucronulate, 4-16 em.

ng, 1.5-6.5 em. wide; inflorescence terminal, the one-headed, corym-
bosely disposed peduncles very slender and lax, subulate-squamose
and sericeo-pubescent above; involucres narrowly campanulate, 9-15-
flowered (20, fide Gray), 11-13 mm. long, at center about 6 mm. broad;
phyllaries rigid-chartaceous, the outer ovatish, the rest mostly oblong
to lance-oblong, maroon-tipped, glabrate dorsally, sericeo-ciliate, ob-
tuse, often the slightest bit carinate and mucronulate; corollas “dark
maroon”’ or claret-colored; achenes minutely scurfy-glandular above,
at anthesis 2.5 mm. long; receptacle glabrate.

MEXICO: Jautsco: Real Alto, trail to El Tajo de Santiago, densely wooded
damp north slope, abundant, alt. 2500 m., Sierra Madre Occidental, Ynez

. 1749 (Gr eh EFINITE STATE UNKNOWN: “Northwestern
Mexico,” Seemann (T E). Seemann collected in southern Sinaloa, southern
Durango, Nayarit and northern Jalisco.

12. P. Wrightii Gray, Pl. Wright. i. (1852) 127; P. Wrightii var.
subpuberula Gray, loc. cit. 128; P. arizonica Gray, Bot. Calif. i. (1876)
422; P. Coulteri Gray, Proc. Amer. Acad. xv. (1880) 40, in part; P.
Schaffneri Gray; Urbina, Cat. de Pl. Mex. (1897) 197, nom. nud.;
illustration in Armstrong, West. Wild Fl. 537 (1915).

Erect, leafy, simple or more commonly branching perennial about
0 cm. tall, varying from 30 em. to 1 meter or more in height; stems
smoothish and glabrous to glandular-scabrid, striate, stramineous or

green, often varying to purplish; leaves sessile, oblong-lanceolate

'o elliptic-oblong or lance-ovate, chartaceous or rarely rigid, mostly
tinutely scabrid on the more prominent reticulations above and more
~ less copiously glandular-puberulent and scabrous beneath, some-
varying to glabrate, the smaller floral leaves closely and regu-

larly repand-dentate, tending to a truncate base, the larger coarsely
and irregularly dentate, tending to be cordate- or subsagittate-am-
Plexicaul, the smaller 2.5-3.5 cm. long, 5-15 mm. wide, the larger
13 cm. ong, 2-5.5 em. wide; inflorescence compound-corymbose,
many-headed, the ramifications and peduncles usually densely glandu-

“pubescent; heads in subglomerate cymules, 8-11-flowered, 10-15
0 mm. high; involucres 2-3-seriate, 6-7 mm. high; phyllaries

“ar-oblong to lance-oblong, the tips varying from obtusish to acutish,
very rarely the slightest bit acuminate, chartaceous, sericeo-ciliate,

40 BACIGALUPI

dorsally glabrate to glandular-puberulent; corollas lavender-pink,
more or less copiously and shortly stipitate-glandular; pappus mod-
erately stiff, usually bright white, in 2 series; mature achenes linear-
fusiform, costellate, glandular-puberulent, 5-6 mm. long, slightly to
barely expanded into a disc above; receptacle sparingly covered with
short thick-stalked amber-colored glands

From southern Utah, south central New Mexico and western Texas
southward in Mexico to San Luis Potosf.

UNITED STATES: Urau: near St. George, Parry, no. 141. ARIZONA:

E. Jon . W. Jones,
no. 151; Skull Valley, alt, about. 1300 x m., eee ce 1903, M. E. Jones (P. C.);
Wiekenburg alt. about 640 m., May 5, 1903, M. E. Jones, (P. C.); near Mt.

ai » April 30 1909, Ss. ‘4 hoes (p. C); south of Fort Whipple, Coues &
Palmer, -’ Santa Catalina eon oak May 30, 1881, Pringle; —
slopes, Na li 1 Sone near 9g L. N. Goodding, 1096 (N. Y.); near Fort
Huachuca, ents nos. 2796 & 2797; gravelly slopes, Caliuro Mo iW.
Price (Py - N. Goodding, no. 1072 (N. Y.); near Bishes, Ma 35, fassdonr ag

rice C EW Mexico: Apache Pass, (Gran L
in 1881; Duck Creek Hills, Upper Gila (nw. Grant Oe. , “May 26, < ra
Gr (Gr., P. C.); Mangas ‘etn Grant Co., ae en no. hont

ona Ana Co., alt. 1770 m ooton, Y. ); wit ay
more specific Pecans R. Vasey, in 1881. Texas: El aS a Nabors 12, 1
M. pu C.); Fort Davis, Girard, in 1880; low pues at northe ak end
ied lag Eos near Sierra Blanca, dspet

pet's Sutton Co. nag ag no. 2810; border of Salt Basin, ci thicket rear Bs od
ey Aug. 1914, Ks

weatward; also on n the Rio Grande,” Wright, no, 411 exe) ‘ils mets te

Pecos and t e Limpia, is, Wright, m 412 (t r. sub

O: Sonora : Vrantiren alt. about 1400 no. c ¥. Hartman, 00.
Las Playas, Thurber, no. 316. CHIHUAHUA: rocky mountain cafions, aa s

Juarez, Elmer (N foothills of the Sierra Madre, near
Juarez, E. no, 6341; 8 Eulalia Plains, E. Wilkinson, in
vieinity of Chihuahua, alt. abot Gat 1300 sopra Pal? no. 143 (1908) (Gr,

, =r f

Rts AH : aS, Pain mer, no. 767 (1880); mountains nort oO
clova, — no. 768 (1880); ae a de la Plog raed Purpus, nw, Corr’ 180
3 nan

> . oje AN' * m E. W. =
Nvevo Lr6én: El Carine, near Monterey, Pringle, nos. Y. ‘ San
iguPowe moceee Cm Oro, Palmer, no. 387 Cab) (Gry » a7, in

Si: in region o ax Tate Potcad, Ps mer, nO. ‘hafiner,
(1878) ; “ex convalli San Luis ee is eabroale pope Tloscala (2), S
no. 376 (type of P, Schaffneri

MONOGRAPH OF THE GENUS PEREZIA 41

According to Margaret Armstrong (West. Wild FI. 536), the local
name (southwestern U.S.) “ Brown-foot” alludes to the mass of brown
woolly hairs covering the caudex of this and the other North American
species of Perezia.

13. P. Coulteri Gray, Proc. Amer. Acad. xv. (1880) 40; loc. cit.
xix. (1883) 59.

Unbranched, erect perennial about 7.5 dm. high; stem rigid, sulcate
to striate, glandular-puberulent, leafy to within a short distance of the
inflorescence; leaves subsessile, chartaceous, often somewhat rigid,
oblong-lanceolate to -oblanceolate, glandular-puberulent, especially
on the lower surface, prominently to weakly reticulate, repand-
denticulate to subentire, often subrevolute, acute to obtuse at the
apex, often abruptly rounded to truncatish at the base, 5-10 cm. long,
11-26 mm. wide; inflorescence corymbose, widely triangular, with a
spread of 6-15 cm. and slightly less high; heads 10—20-flowered;
involucres (2—)3-seriate, 6-10 mm. high; phyllaries membranaceo-
chartaceous, linear-lanceolate, very little imbricated, slightly sericeo-
ciliate to eciliate, subpuberulent to glabrate dorsally; corollas lavender-
pink; mature achenes glandular-puberulent, striate, 5-6 mm. long;
receptacle only slightly alveolate, obscurely glandular and very
sparingly pilose.

South central Mexico and northwestern Guatemala; perhaps occur-
ming in an uninterrupted range between these two regions.

MEXICO: San Luts Potosi: hillsides, San José Pass, Pringle, no. 3178 (Gr.,
Y.). Hipaueo: Zimapan, Coulter, no. 234 (TYPE).

GUATEMALA: Huraverenanco: Chaculd, on sparsely wooded calcare-
ous slopes, growing below oaks, alt. 1600 meters, Caec. & Ed. Seler, no. 3069.

In his original diagnosis (first citation above), Gray confused a
form of P. Wrightii with this species. In the second citation given,
oo P. Coulteri, basing it entirely on the Coulter specimen
cited,

14. P. dissiticeps Bacigalupi (see p. 17).
labrous perennial about 10 dm. tall; stem simple, teretish, leafy,
glabrous, smooth, striate-angular, buff-brown, glaucescent, very
Slightly flexuous and laxly branching above to form the inflorescence;
Ves sessile, thin-chartaceous, reticulate, pale green, glaucescent to
glaucous, especially below and more especially on the midribs, slightly
uning above, the cauline oblong, obtuse and often mucronulate,
slightly to not at all narrowed below, unequally spinulose-dentate,
55-9 cm, long, 16-25 mm. wide, those of the flowering branches
oblanceolate, spinulose-denticulate, narrowed to the shortly rounded

42 BACIGALUPI

to subsagittate base, 2.5-6 cm. long, 7-18 mm. wide; inflorescence-
branchlets slender, glabrous, almost naked, the very sparse ascending
bracts very small, 1-2 mm. long, lance-subulate, minutely ciliolate,
one or two of the lower 5-10 mm. long; peduncles divaricate, 1.5-6
(mostly 2-4) em. long, minutely glandular only at the very top; heads
14-18-flowered, 12-15 mm. high; involucre (3—)4-seried, campanulate,
7-8 mm. high; phyllaries dorsally glabrate, sericeo-ciliolate, the outer
spreading, lanceolate-oblong, the inner oblong, strikingly thickened
and very slightly narrowed toward the base, purplish-tinged on the
margin toward the obtuse and somewhat mucronulate to somewhat
acutish apex; corolla lavender-pink, 8 mm. long, the throat of the tube
and the lower parts of the segments resinous-punctulate; pappus soft,
whitish, in 1-2 series, the relatively sparse bristles sometimes very
slightly thickened toward or at the apex; achenes brown, linear-
oblong, sparsely to copiously resinous-punctulate, about 3 mm. long;
receptacle flattish, quite deeply scrobiculate, the slender ridges
between the alveolae resinous-punctulate.

MEXICO: San Luts Porosf: Sierra Tabl6n, Minas de San Rafael, Purpus,
no. 5018 (Typr, Gr. and isotype, N. Y.).

15. P. montana Rose.

Erect, slender, leafy, glabrous and glaucous perennial 9-15 dm. high,
becoming attenuated above; stem sometimes somewhat decumbent
at the base, soon erect and rigid, often slightly flexuous at the very
slender top, striate, smooth, glabrous, often glaucescent to quite
glaucous, light green, commonly reddish-tinged; leaves rigid-cart-
laginous, thin, closely reticulate, glabrous, grey-green and conspicu-
ously glaucous, oblong to linear-oblong, not at all narrowed toward
the very broadly cordate- to subhastate-clasping base, the auricles
usually overlapping and often greatly exaggerated, acute or very
shortly acuminate, 5-15 em. long, 18-30 mm. wide, very coarsely and
unequally spinulose- to spinose-dentate, the acute triangular tee
2-3.5 mm. long; inflorescence a more or less lax, divaricately branched,
leafy-bracteate cyme, the branches very slender, smooth and glabrous,
the glabrous ultimate peduncles rarely naked or bearing a single scale,
more usually bearing several sparse to moderately crowded lance-
subulate, cuspidate, entire to serrate-denticulate, often ciliolate scales;
heads 18-25 mm. high, 19-22-flowered; involucre turbinate, acute at
the base, 12-15 mm. high (in var. 8. campanulate and rounded hed
obtuse at the base, 9-10 mm. high); phyllaries in 5-7 series, gi
laginous, greenish to stramineous, often purplish, the short sot
ones densely imbricated, merging (in var. a.) with the peduncular

MONOGRAPH OF THE GENUS PEREZIA 43

scales, very rigid, narrowly ovate-lanceolate, cuspidate, more or less
copiously sericeo-ciliate, the inner oblong-lanceolate (var. «.) to oblong
(var. 8.), much more loosely imbricated, becoming much less densely
ciliate than the basal ones, often ultimately eciliate, sometimes rigid
only below, the often narrowly scarious apex varying (in var. «.) from
very shortly acuminate to merely acute or (in var. 8.) obtuse to
rounded and sometimes very minutely mucronulate; corollas lavender-
pink, 12 mm. long, essentially glabrous; pappus slightly stiff, fine, a
little tawny, in 1-2 series; mature achenes widely fusiform, costulate,
more narrowed above just below the slightly expanded disc than
toward the base, glabrous to hispidulous, 6-7 mm. long; receptacle
flattish, very shallowly to somewhat deeply scrobiculate, essentially
glabrous, the fimbriolate ridges between the areolae sometimes
glandular.

Key To VARIETIES.

Var. a. typica. Inflorescence widely paniculate and abundantly
leafy-bracteate, the branches divaricate and very slender, the closely
‘pproximated peduncular scales gradating into the basal phyllaries;
the wiry peduncles mostly 2.5-3.5 em. long, divergent, the heads
Widely separated, not arranged in definite cymules; involucres acute
at the base, obconic-campanulate, 12-15 mm. high; inner phyllaries
oblong-lanceolate, very shortly acuminate to merely acute.—P.
montana Rose, Contrib. U. S. Nat. Herb. i. (1891) 105, pl. viii.

Var. 8. intermedia Bacigalupi (see p. 19). Inflorescence made up
of more or less definite, lax, 2-3-headed cymules, less diffuse than in
Yar. typica; peduncles smooth, glabrate to glabrous, naked or more
sally bearing, at or about the middle, an ovate-lanceolate, glabrous,

ng or ascending bract (very rarely another much smaller one

44 BACIGALUPI

just below the involucre), this campanulate, obtuse to rounded at the
base, 9-10 mm. high; inner phyllaries oblong, not at all narrowed
at the base, sounded and blunt to obtuse at the apex, sometimes tipped
with a very minute mucro; otherwise as in variety typica

re seamen Duranoo: Santiago Papasquiaro, Palmer (1896), no. 59 (TYPE,

it. F. Ras ne Gray, Mem. Amer. Acad. ser. 2, iv.. [Pl. Fendl.]
(1849) 11
Rigid, ack perennial 6.5-18 dm. tall, apparently simple below the
inflorescence; stem slightly flexuous, striate, often slightly glaucous,
smooth, glabrous, often purplish; leaves sessile, rigid, coriaceous,
glabrous, often slightly glaucous, finely reticulate, the lower midrib
often very prominent, the cauline imbricate, orbicular- to widely
oblong-obovate, more usually closely and sharply triangular-dentate,
the teeth sometimes spinulose, sagittately to cordately clasping,
sometimes shortly mucronate at the widely rounded apex, 6-12 cm
long, 5-10 cm. wide, those of the inflorescence becoming lance-ovate,
finely and intermittently spinulose-denticulate to entire, the upper-
most often merely subhastate to shortly rounded at the base, 1-3.5
em. long; inflorescence made up of corymbosely disposed glomerulate
cymules, even the ultimate peduncles quite glabrous; heads at anthe-
sis 15-18 mm. high, 10-12-flowered; involucres turbinate-campanu-
late, 9-12 (mostly 10) mm. high; phyllaries broadly (slightly less than
2 mm. wide) to more narrowly (about 1 mm. wide) oblong, the outer-
most Scatics, the broader inclined to be quite obtuse, the narrower
somewhat acutish, all indurate, glabrous to glabrate dorsally, more oF
less copiously rusty-tomentulose on the edges; corollas lavender-pink;
mature achenes fusiform-cylindric, somewhat incurved at the base,
weakly striate, appressed-hirsutulous, glandular as well when young,
mm. long, only slightly expanded at the summit; pappus rather
stiff, mostly in 2 series; receptacle very shallowly alveolate, very
sparingly glandular.
Sierra Madre of Mexico from Chihuahua to Michoacan.
west of Chi-

O: Carmuanva: Cosiq uiriachi, in the a "Chihual has Pringle,

ntains
no. 766 (N. Y.); Santa Eulalia Mountains, Pri le, no. 583; “ é
Chihuahua,” Palmer, tg 299 (1885). Duniwoo! Sandia Whee Pringle,
i i Z AS ga

17. .. rigida « we) Gray.

MONOGRAPH OF THE GENUS PEREZIA 45

Erect, robust, leafy, apparently simple perennial 6-18 dm. high;
stems smooth, glabrous, more or less striate,-in more robust specimens
suleate as well, often slightly flexuous and tinged with purple above,
usually somewhat glaucous; leaves highly variable in outline, the
lower from spatulate-oblong to oblong or elliptic-oblong (or in vars.
8. and y. linear-oblong and broadest above the middle, ascending to
spreading, arcuate and strongly conduplicate), the base truncatish to
cordate-amplexicaul, the apex obtuse to merely obtusish, usually some-
what mucronate, subentire to coarsely and usually unequally spinulose-
dentate, rigid, thick-coriaceous, closely reticulate, smooth, glabrous,
very often somewhat though not conspicuously glaucous, 6-14 cm.
long, 2.5-6.5 cm. wide; the upper more usually from linear or oblong
to lance-oblong, often acute, closely spinulose-denticulate to sub-
entire, 4-10(—15). cm. long, 1.5-3(—6) cm. wide (the linear ones 4-10
mm. wide), otherwise as in the basal leaves; floral leaves quite entire,
not at all cordate at the base, sometimes oblanceolate; inflorescence
corymbose to widely paniculate, the ultimate peduncles often spar-
ingly puberulent; heads closely approximated into cymules at the
periphery of the inflorescence, 9-11(-14)-flowered, 14-20 mm. high;
involucre usually 4-seried, turbinate- to oblong-campanulate; phyl-

es indurate, moderately well imbricated, more usually minutely
glandular-ciliolate, less usually sericeo-ciliate, dorsally glabrate to
slightly glandular or merely puberulent, the outer ovatish, the others
oblong, the tips mostly obtuse, sometimes barely acutish, rarely
mucronulate (in var. y. abruptly acuminate); corollas lavender-pink,
more usually the upper portion of the tube glandular-punctulate, the

portion of the segments, particularly the 2 linear inner ones,
glandular-ciliolate, less usually glabrate; pappus stiffish, very often
slightly clavellate-thickened toward the apex, in 2 series; achenes sub-
dric (in var. y. broadly fusiform and costulate), somewhat nar-
Towed above, surmounted by the slightly expanded epigynous dise,
densely glandular, 4-5 mm. long; receptacle not deeply alveolate,
slabrous to sparingly glandular.

Ph Key To VARIETIES.

Vllaries obtuse to more rarely acutish, not at all acuminate,
‘ra, | dular-ciliolate; mature achenes subcylindric, 4-5
Hi baie not more than 0.5 mm. wide.

usually 10-flowered, varying from 9-11-flowered; up-

Per cauline leaves broadly oblong or elliptic-oblong, neither

arcuate nor conduplicate; corollas glandular-punctulate, the :
Hea i uents glandular-ciliolate toward the base....---..-;. Var. «. typica.

er above the middle than below, somewhat arcua

46 BACIGALUPI

and strongly conduplicate; corolla mostly glabrate, with
only an occasional gland present...............---- at B. linearifolia.

form, 6 mm. long, 1-1.5 mm. wide; leaves and corollas as in
Var. y. acuminata.

Var. a. typica. Leaves broadly oblong or elliptic-oblong, the lower
often widely spatulate-oblong, 6-14 cm. long, 2.5-6.5 em. wide, those
of the inflorescence 1-3 cm. long, often varying to lance-oblong or
oblanceolate; heads 9-11-flowered, mostly 10-flowered; phyllaries
obtuse, often tipped with a minute mucro, mostly widely oblong and
densely glandular-ciliolate, the outer sometimes sparingly sericeo-
ciliate; tube of the corolla glandular-punctulate above, the seg-
ments, particularly the 2 linear inner ones glandular-ciliolate at the
base; mature achenes subcylindric, 4-5 mm. long, not more than 0.5
mm. wide.—P. rigida (DC.) Gray, Pl. Wright. i. (1852) 127; Acourtia
rigida DC. Prodr. vii. (1838) 66; A. formosa Hook. & Arn. Bot. Voy.
Sulph. (1845) 122, not D. Don.

Sierra Madre of Mexico from Durango to Guanaxuato.

MEXICO: Duranoo: vicinity of Durango, Palmer (1896), no. 611, in part.
San Luts Porosf: region of San Luis Potosi, alt. 1830-2440 m., Parry & Palmer
(1878), no. 558; “ex convalli San Luis Potosi, in umbrosis prope San Migue-
lito,” Schaffner, no. 375 (Gr. N. Y.). Guanaxvuato: Moro Leén, Feb., 1898,

. ; near Guanaxuato (by inference from original diagnosis),
(fragment of type of Acourtia rigida DC.).

Var. 8. linearifolia Bacigalupi (see p. 18). General habit of variety
typica; leaves linear-oblong, slightly wider above the middle than
below, not narrowed toward the base, the lower about 14 cm. long,
2.54 cm. wide, tending to be arcuate and conduplicate, this tendency
of the leaves increasing upward, the upper cauline narrowly oblong-
linear to linear, strongly conduplicate, 6-9 cm. long, 8-24 mm. wide;
heads 14-flowered; phyllaries glabrate to slightly glandular-pu-
berulent, oblong to lance-oblong, obtuse to acutish, for the most
minutely but densely glandular-ciliolate, more rarely sparsely 9°
shaggily sericeo-ciliate; corollas mostly glabrate exteriorly with only
an occasional gland present; achenes as in var. typica.

.
.

MEXICO: Guanaxvato: Guanaxuato, Dug?s. Jatisco: Rio Blanco, Oct.,
366, P . 661 (Gr., N. Y.); cool hillsides near Guadalajara, Pring
no. 1860 (rrPE, in Gr. and isotype, N. Y.); rocky bluffs of a barranca
lajara, elev. 1525 m., Pringle, no. 9950 (N. Y.).
Var. y. acuminata Bacigalupi (see p. 19). Leaves all are
oblong, a little wider above the middle, not narrowed below t0 |
base, very strongly conduplicate, arcuate, especially toward the UPS

MONOGRAPH OF THE GENUS PEREZIA 47

divergent to ascending, 5-14 cm. long, 8-32 mm. wide, like those of
var. linearifolia in all other particulars; heads 9-flowered; involucre
narrowly obconic-campanulate, mostly 9 mm. long, varying in length
from 8-10 mm.; phyllaries narrower than in the preceding varieties,
all but the outermost narrowly oblanceolate in outline, gradually
narrowed toward the base from above the middle, more or less copi-
ously sericeo-ciliate, shortly acuminate and acute; corollas as in var.
linearifolia; mature achenes minutely glandular, costulate, widely
fusiform, 6 mm. long, 1—-1.5 mm. wide.

MEXICO: Jauisco: rocky bluffs of a barranca near Guadalajara, elev. 1525
m., Pringle, no. 9950 (rypE, Gr.).

18. P. tomentosa T. S. Brandegee, Zoe v. (1908) 258.

Simple, leafy, erect, densely deciduous-tomentose perennial 2-4
dm. tall; stem terete, striate, the dense, rusty tomentum of the caudex
clothing its somewhat decumbent base for about 5 cm., the remaining
portion densely floccose-lanate with a more nearly whitish and some-
What arachnoid tomentum; leaves sessile, closely approximate and
imbricate, completely covering the stem, for the most part, above the
base, oval to oblong-ovate, shortly acuminate, serrate-dentate, the
teeth acute and mostly spinulose, densely white-tomentose beneath,
arachnoid on the slightly nitidulous and reticulate upper surface, the
uppermost tending to be revolute, 3-5 cm. long, 13-27 mm. wide;
heads occurring singly at the thickened end of the axis and short
2.5-3 cm. long) ascending peduncular branches inserted just below the
axis-tip, about 50-flowered, hemispheric, 2-3 em. thick above, 1.5-2
em. high, sometimes, according to the author, sessile in the axils of the
upper leaves; phyllaries in 3-4 series, 1-1.5 cm. long, the outer oblong-
laneeo te, acuminate, tomentose, indurate below, the inner becomin
Scarious and less tomentose, linear-lanceolate, stramineous, the inner-
Most arachnoid, at least toward the purplish, caudate-attenuate,
glandular-punctate and, at least in the dried specimen, recurved or
‘ortuous tips; corollas lavender-pink, about 1 cm. long, the outer
ligule and the two linear segments of the inner lip strikingly glandular-

: olate; pappus soft and fine, with a silky lustre, biseriate only on the
Wider Portions of the scarcely enlarged epigynous disc of the oblong-
cylindric (5 mm. long) resinous-papillose achenes.

2639 RICO: A aes dry, sunny hills, Esperanza, Sept., 1907, Purpus, no.
‘ A well-marked and apparently little-collected species from a region
YN means rarely visited by botanical collectors.

48 BACIGALUPI

19. P. Parryi Gray, Proc. Amer. Acad. xv. (1880) 40.

Erect, caulescent perennial subequally branched from near the base;
stems many, slender, rigid, leafy from near the base to the very sum-
mit, slightly to deeply striate, glandular-puberulent, scabrous, be-
coming glabrate and smoothish, inclined to be flexuous, 10-30 cm. long;
leaves sessile, ascending, cuneate-oblong to cuneate-obovate, rarely
rounded at the base, irregularly and rather saliently pungent-dentate,
sometimes less so or not at all at the base, glandular-punctate and
smoothish to glandular-scabrous, subcoriaceous, 18-30 mm. long,
9-16 mm. wide; heads terminating the branches, borne singly, im-
mediately subtended by the uppermost leaves, 18-23 mm. high, 15-18-
flowered; involucre subtruncate to rounded at the base, campanulate;
phyllaries well imbricated, indurate, the outer ovate, 2-4 mm. broad,
mostly abruptly cuspidate-acuminate, dorsally glandular to glabrate,
sometimes glandular-ciliolate, the inner oblong, acute, about 1.5 mm.
wide; corollas lavender-pink; mature achenes sometimes sparsely
covered with subsessile glands, otherwise smooth and glabrous,
slenderly subcylindric, 5-7 mm. long; receptacle somewhat deeply
alveolate, sparsely glandular, otherwise glabrous.

Central Mexico from southern Coahuila to Querétaro.

MEXICO: Coanurta: Saltillo, Parry, no. 41 (1878); same locality, Palmer
(1898), no. 278 (Gr., N. Y.); Sierra de Pérras, Purpus, no. 1129 (Gr., N. Y.)-

CaTECaS: plains, Sean, Lloyd & Kirkwood, no. 151. San Lois Potosi:

% nN

Potost, Parry & Palmer, no. 545, in 1878 (tveE). QueRéraRo: along road
Cadereyta, near San Juan del Rio, Rose, Painter & Rose, no. 9620 (Gr., N. Y.).

rust-colored tomentum; stems densely glandular-scabrous, U timately
smooth and glabrous, often somewhat sulcate and flexuous; leaves
sessile, rhombic-orbicular to suborbicular, rhomboidal to rounded at
the base, rounded to subattenuate at the apex, coarsely, acutely and
irregularly spinulose-dentate, raised-reticulate on both surfaces, from
copiously glandular-scabrous and glandular-ciliolate to smooth and
glabrous, yellow-green to glaucescent, 2-4(-5) cm. wide, mostly as
long as wide; heads single at the stem-tips, leaf-subtended
“sessile”), deeply campanulate, 19-28 (mostly 20-25) mm. high, the
smaller heads about 15-flowered, the larger 20-24-flowered; phyllaries
mostly in 4(-5) series, well imbricated, chartaceo-coriaceous, the

MONOGRAPH OF THE GENUS PEREZIA 49

outermost ovate, obtuse to acuminate and mucronate, the others
oblong-lanceolate, acuminate, often slightly scarious-edged, sericeo-
ciliolate to merely glandular-ciliolate, glandular-puberulent (especially
the outer) dorsally, corollas lavender-pink; achenes subcylindric,
prominently costate, 6-8 mm. long, more or less densely beset, espe-
cially above, with shortly stipitate glands; receptacle deeply alveolate,
the alveolae separated from one another by a brownish membrane,
this fimbrillate at the top.

Southeastern Arizona, es New Mexico and southwestern
Texas southward in Mexico to San Luis Potosf.

UNITED STATES: Arizona: Benson, May 7, 1882, G. W. Dunn; koe tn

more specific locality, Lemmon, no. 83 (1 880); “southern Arizona, ” Palmer ;
no. 144 (1867); Wil t, elev. 78 shea no. C., N.. Y.); Fort
Verde, Aug. 27, 1887, Mearns (N. "Y); tH achuca 1892, Wilcox

2 Y;
(N. Y.); Agua Verde Creek, ater no. C148 3 (N. Y.); Tucson, Griffiths, no.
2061 (N. Y.); Mt. Desert, Chiricahua Maat alt. 1920 m., Blumer, no.
2194 (Gr., N. Y.). hax Me aga gee ins of the upper Gila, May eid 1880,
E. L. Greene (Gr., N. Y.); Gila Valley, June, ala x Gree ne (P. C.); Nogal,

Lincoln County, Josephine Skehan, no. 26 ( Y., P. C.); Organ Moun-
a County, April 29, 1899, “n O ton; mesa ne S
Dona Ana County, 300 m., Sept., 97, Wooton (N. Y.); Fillmore
nhyon, an Mountains, April 29, 1899, Wooton (N. Y.); “chiefly in
Valley of the Rio Grande Lig ana,’ Parry, Bigelow, Wright, & Schott
(Mex, Bound. Sur v.) (N. Y.); Rincon, A 30, E. Jor (Pot).
May 16, 1890, M. "E. Jones (P. C.); - ry ‘hills and plains, aint ne Rusby,

no. 181 (N. Y.); Gila River bottom near Cliff, Grant County, alt. 1370 m., Met-
me no. eh (N. Y.); Florida Sctiednk A. Isabel Mulford, he 9%
AS: ( :

T E "gun April 15, 188 Jone. C.); in arroyo on 8. P. .
ra ages es west of Sierra Blanca a, Hudspeth County, Ferris & Duncan, no.
rf he ag Pig Pass, Thurber (1850), no. 93; Barstow, T’racy

Y.); flats near Big Spring, Howard ‘County, May 9,

1901, H. Eggert; Fort Stockton, H. C. Hanson, no. 565 (Gr., N. Y); Marfa,
5, E. N. Plank (N. Y.); Tom Greene County, 1878, F. Tweedy;

without more specific ionality, Wright, no. 410 (Gr.

: Sonora: San Pedro, Hactingh: no. 830. CHIHUAHUA: calcareous

: i300 mn Palm er (1008 no. 186 (G ,N. ye Santa Eulalia po reer
ou no. 13 (Gr., NX. ¥ 5. a HUILA: ‘Saltillo, Palmer (1898), no. 85 (Gr.,
«& P, San Luis Porosf:r region of San Luis Potosi, alt. 1850-2430 m arry
N. ier (1878), no. 546; “ex convalli Sia Luis Potosf, Schaffner, no. 373 (Gr.,
UNITED STATES AND MEXICO: “along branches of the San Pedro,

4} Sept. Stony hills, near El Paso; March,” Wright, no. 1418.

M tt Holly” is the local Southwestern name, according to
arraret Armstrong (West. Wild Fl. 536).
21. P. Wislizeni Gray.
as mple, erect, leafy, glabrous perennial 3.5-7 or more dm. tall;
ems unbranched or rarely divided at the top, teretish, striate,

ed

50 BACIGALUPI
glabrous, smooth, often reddish-tinged and somewhat glaucous; leaves
ascending, approximate, often imbricated, sessile, coriaceous and
rigid, often very thick, oblong-obovate or spatulate-oblong to oblance-
olate, narrowed to a base varying from narrowly truncatish to shortly
rounded and semi-amplexicaul, rounded to obtuse and usually shortly
mucronate at the apex, very closely and minutely spinulose-denticu-
late to irregularly and serrately sinuate-dentate, the teeth 1.5-2 mm.
long, glabrous and smooth save for the close and prominently raised
reticulations on either surface, often glaucescent, 5-14 em. long, 2-5.5
em. wide; upper leaves entire, often somewhat amplexicaul, oblong to
elliptic-spatulate, crowded and subimbricate up to the involucre or
sparse, 2-4 cm. long; single head sessile on the usually enlarged stem-
summit, widely hemispheric, truncate at the base, 50-60(—“80”)-
flowered, 3-4 cm. high, 3.5-5 em. thick; phyllaries more or less coria-
ceous, in 5-7 series, entire to somewhat sericeo-ciliate, glabrous dor-
sally, from dark purplish throughout to only at the tips, the inner
oblong-lanceolate, acute and long-acuminate to widely oblanceolate
and obtuse, with or without a small mucro, the outermost ovate-
lanceolate, ovate, oblong-lanceolate or oblanceolate; corollas lavender-
pink, 17-20 mm. long; pappus soft to rather rigid, somewhat tawny,
some of the bristles obscurely clavellate-thickened toward the apex,
biseriate, most often slightly exceeded by the corolla; achenes oblong-
cylindric to linear-cylindric, deeply striate, smooth and glabrous to
very sparingly glandular-punctate, 8-11 mm. long; receptacle (fide
Gray) subalveolate and glabrous.

Key To VARIETIES.

Upper portion of the stem naked save for reduced and_ sparse
bractlike leaves; cauline foliage thin-coriaceous, minutely
spinulose-denticulate; outer basal phyllaries ovate or more :

OR er a 5 es i Var. a. typrca-

Upper portion of the stem more or less densely leafy-bracteate
u x

“
E
°
=
“s
al
md
i,
i=)
=
2
3
®

larly and serrately sinuate-dentate, the teeth 1.5-2 mm. long; phala.

Var. x. typica. Apparently the more commonly collected and more
northerly variety.—P. Wislizeni Gray, Mem. Amer. Acad. ser. 2; 1:
[Pl. Fendl. (1849)] 111; P. Wislizeni subsp. minor Blake, Contrib.
U.S. Nat. Herb. xxii. (1924) 653. The latter is based on a specimen
with an underdeveloped and insect-ridden (in the field?) head.

-MEXICO: Cumuanva: Llanos, in the Sierra Madre west of Chih am
Wislizenus, no. 198 (TYPE); pine plains, base of the Sierra Madre, P ringle,

1307 (Gr., N. Y.); roadside near San Julian, alt. 2135-2440 m., 4. V
no. 4925. Duranao: city of Durango and vicinity, Palmer (1896),

no. 783;

MONOGRAPH OF THE GENUS PEREZIA 51

without more definite locality, alt. 1000 meters, P. [bafta Garcia, no. 376 (type
of P. Wislizent subsp. minor Blake).

Var. 6 megacephala Gray. Apparently a more southerly variety. —
P. Wishizent var. megacephala, Gray, Proc. Amer. Acad. xxii. (1887)
433; P. megacephala Greenman, Proc. Amer. Acad. xli. (1905) 269.

MEXICO: Nayarir: Tepic, Palmer (1892), no. 1962 (Gr., N. Y.); Tepic.
M. E. Jones, no. 23358 (P. C.). Jatisco: Rio Blanco, Palmer (1886), no,
655 (rypz, Gr. and isotype, N. Y.); gravelly hillsides near Guadalajara,
Pringle, no. 2290 (N. Y.).

22, P. thyrsoidea Gray, in Torr. Bot. Mex. Bound. (1859) 104.

Densely and amply leafy perennial 6-15 dm. high; stem apparently
simple, robust, copiously glandular-pubescent, scabrid; leaves ap-
proximate, usually imbricate-ascending, sessile, all save the upper-
most shortly sagittate-clasping at the somewhat rounded base, elliptic-
to oblong-ovate, obtusish to shortly acuminate and pungent at the tip,
igid, subcoriaceous, conspicuously reticulate, especially beneath,
densely glandular-hispidulous, especially on the veins beneath, scabrid,
glandular-ciliolate, the upper closely spinulose-serrulate, the lower
much more remotely and irregularly so, the floral leaves 25-70 mm.
long, 15-35 mm. wide, the lower 12 em. long or longer and about half
as wide; inflorescence a dense, strict, elongated cylindric thyrse, each
24-headed cymule nearly equalled by a subtending leaf; ultimate
peduncles bearing imbricated linear-lanceolate densely glandular
bracts; heads 20-25 mm. long, 10-12-flowered; involucre obconic;
phyllaries attenuate-subulate, very acute, in 3-4 series, densely and
uniformly glandular-puberulent, not indurate though somewhat rigid

‘low; corollas lavender-pink; immature achenes narrowly subcylin-
tric, slightly more narrowed above, densely glandular-puberulent,

mm. long; receptacle somewhat deeply alveolate, glandular-

puberulent.

AL
” } . - ops hills near
1175 (Gr. ngs ran (Gr., N. Y.); Acambaro, Caec. & Ed. Seler, no.
oe P. formosa (D. Don) Gray, Proc. Amer. Acad. xix. (1883) 58;
eh formosa D. Don, Trans. Linn. Soc. xvi. (1830) 204; P. tur-
turby Gray, Pl. Wright. i. (1852) 126, not La Llave & Lex.;
binata Sch. Bip. in Seem. Bot. Voy. Herald (1856) 315; Acourtia
mecrocephala Sch. Bip. in Seem. in syn., loc. cit. and pl. 55.

Erect, leafy » apparently simple perennial at least 6-9 dm. high;
“ems rigid, teretish, the slightest bit flexuous and quite striate above,

52 BACIGALUPI

smooth and glabrous, often purplish-tinged, leafy up to the inflores-
cence; leaves closely approximated, the bases subimbricate, rigid-
coriaceous, sessile, smooth save for the raised and close reticulations,
rarely a little scabrid-puberulent on the larger reticulations beneath,
linear-oblong to lance-oblong, cordate-clasping, attenuate and acute
to obtusish and mucronulate at the apex, often ciliolate with minute
callosities, strongly conduplicate and arcuate, irregularly and moder-
ately closely spinulose-dentate, the major dentations 1-2 mm. long
and very acute, the upper cauline leaves 9-15 cm. long, 2-3 em. wide,
gradually becoming shorter upward; inflorescence condensed, corym-
bose-glomerate, 4-5-headed; peduncles becoming gradually thickened
above, merging into the base of the involucre, 10-12 mm. long,
densely imbricate-squamose, the scales appressed, glabrous, lance-
subulate; heads 25-30-flowered (fide D. Don and Gray), 2.5-3.5 cm.
high; involucres obconic-turbinate, very acute at the base, spreading
above, 2-3 mm. wide; phyllaries stiffish-chartaceous, lance-oblong,
stramineous, purplish toward the very acute and shortly acuminate
apex, readily disarticulating, well imbricated in 5-7 series; pappus
rather soft, only a few of the bristles noticeably, and then very slight-
ly, clavellate-thickened toward the apex.

MEXICO: Durango: city of Durango and vicinity, Palmer, no. 611, in
part (1896) (immature specimen, referred here provisionally). No DEFINITE
STATE SPECIFIED: ‘ . Mexico” (i. e. northwest of the City of Mexico),
Seemann. (Seemann collected in southern Durango and Sinaloa, in Nayarit
and northern Jalisco.)

The achenes in the only mature specimen seen by the writer were
badly insect-eaten. Material at hand was moreover too scant to
warrant an examination of the receptacle.

Don described the achenes as teretish, papillose-scabrous, @ half
inch long, with the epigynous disc very slightly dilated. The recep-
tacle, according to this author, is glabrous and scrobiculate.

24. P. longifolia Blake, Journ. Wash. Acad. Sci. xviii. (1928) 36;
P. foliosa M. E. Jones, Contrib. West. Bot. no. 15 (1929) 154, not of
Rusby, 1896.

Erect, robust, leafy perennial at least 60 cm. tall; stems usually
several from the caudex, simple, herbaceous and relatively slender
(3-6 mm. thick), terete, usually hollow, very finely striate, Very
smooth and glabrous, purple-tinged and glaucescent; leaves sessile,
ascending, overlapping, chartaceo-coriaceous, scabrid to gris é
lance-elliptic to nearly linear-elliptic, shortly cordate-clasping, close y
unequally and somewhat serrately spinulose-denticulate (the eine
between the teeth rounded), callose-ciliolate, entire at the acumina

MONOGRAPH OF THE GENUS PEREZIA 53

and very acute apex, somewhat prominently reticulate, especially
beneath, roughened with minute hairs on the reticulations above,
paler and roughish-hispidulous on the reticulations beneath, 15-27
em. long, 3.5-7 cm. wide; inflorescence 2—3-headed or the heads some-
times subsessile in the axils of the upper leaves; peduncles 0.5-3.5 cm.
long, glabrous to puberulent, densely appressed-squamose; scales
imbricate, sometimes spreading, lance-subulate, entire, glabrous,
0.5-1 em. long; heads 3-3.5 em. high, 45-55-flowered; involucre tur-
binate, acutish to very acute at the base, about 2.5-3 em. high; phyl-
laries glabrous, thinly membranaceo-chartaceous, oblong-lanceolate,
attenuate-acuminate and acerose-tipped, strongly imbricate, in about
7 series, stramineous and often suffused with lavender-pink, gradating
into the peduncular scales; corollas lavender-pink, 2.5 cm. long,
glabrous; pappus moderately stiff, mostly in 2 series, the bristles not
appreciably thickened above; achenes narrowly cylindric, 8 mm. long,
densely glandular-punctate and appressed-hispid, infundibular-spread-
ing above to form the epigynous disc; receptacle flattish, not deeply
serobiculate, glandular-punctate.

MEXICO: Jaxisco: Calabazas, 1925, Reko, no. 4872 (rrpx, U. 8.). Naya-
nT: “on prairies,” La Barranca, M. E. Jones, no. 23359 (type of P. foliosa
Jones, P. C. The number of the type is doubtless erroneously cited in the
diagnosis as no. 23358); Tepic, Palmer (1892), no. 1972 (distributed

a8 °. turbinata La Llave & Lex.).
25. P. lepidopoda Robinson, Proc. Amer. Acad. xliii. (1907) 47.
Tall, erect, slender, very leafy, essentially glabrous perennial
Probably 12 dm. or more tall; upper 8 dm. of the stem (portion com-
prising the type specimen) slender, rigid, not more than 3.5 mm. thick,
state, very smooth and glabrous, dark-purplish; leaves sessile,
crowded, rigid and thin-coriaceous, linear-oblong or -oblanceolate to
linear, attenuate and very acute, sagittately to hastately clasping,
reticulate, glabrous, arcuate to reflexed, conduplicate, yellow-green
above, gr ay-green below, very saliently spinose-dentate, 6-14 em. long,
6-22 mm. wide, the longer (2-4 mm. long) triangular-subulate spread-
ing teeth usually alternating with others less than half as long; upper
peduncular branches widely divaricate, the lower spreading almost
ght angles to the axis, striate, sometimes slightly pubescent,
14 em. long, distally somewhat appressed- and imbricate-squamose,
me scales toward the base of the peduncles becoming distant and leaf-
®; the upper acicular-subulate, minutely puberulent and ciliolate,
jn ting into the phyllaries, those toward the base linear-lanceolate,
a. to entire, conduplicate and arcuate, sometimes subhastate,
‘m. long; heads (immature) turbinate, many-flowered, 2.5—-3 cm.

54 BACIGALUPI

high; phyllaries lance-subulate, very acute, striate, greenish to stra-
mineous and purplish-tinged, glabrate to slightly puberulent, minutely
ciliolate, in about 7 series, easily disarticulating, the outer slightly
spreading; receptacle somewhat convex, glabrate; flowers and fruit
too immature.

MEXICO: Moretos: valley near Cuernavaca, elev. 1220 m., Pringle, no.
9253 (TYPE).

In spite of the immature condition of the heads, it is plainly dis-
tinct from its most closely related species, P. arachnolepis.

26. P. arachnolepis Robinson, Proc. Amer. Acad. xliii. (1907) 47.

Slender, erect, moderately leafy perennial, simple below, branching
above, 10-15 dm. high; stem more or less decumbent and tortuous at
the base, soon rigid, straight and erect, very smooth, terete, striate
below, glabrous, for the most part purplish, the lower leaf-axils often
flocculent-tomentose with the tawny tomentum characteristic of the
caudex; leaves sessile, oblanceolate- to linear-oblong, cordate-clasping
to subhastate at the base, acuminate on an acutish to obtuse apex,
closely and more or less regularly spinulose-denticulate, scabrid, thin-
coriaceous, firm and rigid, finely reticulate, glabrous and dark green
above, minutely hispidulous on the larger reticulations of the grey-
green lower surface, 5.5-17 cm. long, 1.5-6 cm. wide; inflorescence 4
widely branching, lax, paniculate, 10—12-headed cyme; peduncles very
slender, acicular-squamose and slightly arachnoid above, more oF ess
naked below, 3.5-15 cm. long; heads 2-2.3 cm. high, about 15-flowered;
involucre obconic-turbinate, acute at the base; phyllaries gradating
into the peduncular scales, broadly lance-subulate, attenuate to 4
very acute apex, rigid except at the slender apex, arachnoid on the
margins, glabrate dorsally, in 6-7 series; corollas lavender-pink, gla-
brous, 13 mm. long; pappus fine and soft, mostly biseriate; immature
achenes subcylindric, densely appressed-hirsute, 3.5—4 mm. long.

MEXICO: Jautsco: canyons, Chapala Mountains, near Guadslae
dg args 2935 (Type); barranca of Rio Blanco, near Guadalajara, Pring’

27, P. Purpusii T. S. Brandegee, Univ. Calif, Publ. Bot. iv. (1911)
194

Leafy caulescent perennial of unknown height; stems and branches
striate, becoming almost sulcate above, yellowish to purplish, glabrous
save for small tufts of tawny wool concealed in all but the up
leaf-axils, occasionally puberulent; leaves sessile, lance-ovate, or ob-
long to oval, irregularly pungently dentate, the larger dentation®
somewhat acuminate, the smaller oval leaves usually spinulose-dentice”

MONOGRAPH OF THE GENUS PEREZIA 55

late as well, very slightly cordate at the base, rounded at the apex and
occasionally mucronulate, 25-55 mm. long, 14-35 mm. wide, the
others quite cordate at the base, obtusish to acuminate at the apex,
30-70 mm. long, 13-32 mm. wide, all essentially glabrous, bright green
and reticulate-veiny on both surfaces, smoothish to scabrid, especially
on the upper surface, subcoriaceous; inflorescence 1—2(—“ 4”’)-headed,
terminal, the heads solitary on somewhat divaricate, densely squamose
peduncles 6-10 em. long; upper scales imbricated, often densely so,
triangular-subulate to narrowly subulate, 5-6 mm. long, their margins
becoming gradually more conspicuously sericeo-lanate upward, the
lower becoming leaf-like and spreading, lanceolate, often long-acumin-
ate, shortly ciliate to eciliate, 6-12 mm. long; heads 20-23 mm. high,
40-60-flowered; involucre obconic to turbinate-campanulate, its acut-
ish base confluent with the somewhat expanded summit of the ped-
uncle, 17-18 mm. high; phyllaries lance-subulate, mostly narrowed to a
fine point, glabrate to puberulent dorsally, flocculent-lanate to arach-
noid on the margins, fairly well imbricated in at least 4—5 series, especi-
ally the lower tending toward narrowly subulate and gradating into the
peduncular scales; corollas lavender-pink; pappus copious, biserial,
very fine and soft; achenes (immature) densely and minutely glandu-
lar, subeylindric, 3.5—4 mm. long, dilated above into a short epigynous
; receptacle alveolate, moderately densely bristly.
MEXICO: San Luts Potosf: Minas de San Rafael, Purpus, no. 4787 (iso-
type, Gr.); Guadalcazar, Dugés, in 1883 (distributed as P. turbinata).
28. P. platyptera Robinson, Proc. Amer. Acad. xliv. (1909) 626.
Erect, leafy, moderately robust perennial 15 dm. tall; stem striate,
slightly flexuous, stramineous, very smooth and glabrous, alate with
the strikingly decurrent leaf-bases, these about 1 cm. wide at the top,
gradually decreasing in width and shortly rounded off at the base,
persisting for about two internodes, usually sparingly denticulate,
like the leaves rigid and coriaceous, finely reticulate, smooth an
glabrous; leaves spreading to somewhat erect, oblong-elliptic, mostly
‘cute and acuminate, cuneate toward the decurrent base, spinulose-
dentate, 10-12 cm. long, 3-4 cm. wide; inflorescence a widely corym-
b cyme with the cymules subcongested at the flat broad top;
Tanches of the inflorescence leafy-bracteate, the bracts lanceolate,
the larger about 3 cm. long, decreasing upward, conspicuously de-
Current, often sparingly denticulate only toward the base; heads 14-15-
flowered, 15 mm. high; involucre obconic-campanulate, 9-10 mm.
i ; Phyllaries rigid save at the acute, attenuate tips, narrowly
hee-subulate, in 4-5 series, the basal ones very much shorter and

56 BACIGALUPI

much more closely imbricated than the almost totally exposed inner
ones, all minutely glandular-punctate; corolla lavender-pink, glandu-
lar-papillose externally, about 1 cm. long; pappus moderately rigid,
dense, biseriate, bright white, many of the bristles clavellate-thickened
toward the apex; immature achenes oblong-cylindric, glandular-
punctate, 3-4 mm. long, slightly expanded above into the short,
widely funnelform epigynous disc; receptacle flattish, very shallowly
scrobiculate, glandular-punctate.

MEXICO: Micnoacdn orn GuERRERO: in clayey soil, in the Sierra Madre,
alt. 1700 meters, Langlassé, no. 773 (TYPE).

The only species thus far known with alate stems.

- 29. P. hebeclada (DC.) Gray, Pl. Wright. i. (1852) 127; Acourtia
hebeclada DC., Prod. vii. (1838) 66; Deless., Ic. Sel. iv. (1839) 41, pl.
95; Eupatorium affine Mair., in syn. ex DC., loc. cit.

Rigid, erect, leafy, apparently simple perennial, 9-12 dm. high;
stems slightly striate to sulcate, glandular-puberulent, more or less
scabrous; leaves sessile, coriaceous, ascending, reticulate, especially
beneath, cordate-clasping, the lower cauline elliptic-oblong to spatu-
late-oblong, auriculate-clasping, abruptly subacuminate, irregularly
spinulose-dentate, glandular-scabrous, sometimes becoming glabrous
above, very shortly bristly-glandular on the veins and prominent mid-
rib beneath, 10-20 cm. long, 3.5-9 cm. wide, the upper lance-elliptic
to -ovate or even linear-lanceolate, often less scabrous, 3.5-6.5 cm.
long, 8-30 mm. wide; thyrse leafy-bracteate, corymbose, from short
and spreading to thickly cylindric in form, 20-60-headed, the pe
duncles densely glandular, the cymules more or less glomerate; heads
20-23 mm. high, 16-30-flowered; involucres turbinate to turbinate-
campanulate, 14-16 mm. high; phyllaries in 4—6 series, moderately
well imbricated, greenish, indurate only below, chartaceous toward the
tip, the outermost often linear and distinctly herbaceous, recurved-
spreading, the inner erect, subulate, copiously glandular dorsally ?
often glandular-ciliolate; corollas lavender-pink; achenes fusifo ee
cylindric, minutely glandular-punctate, 7 mm. long; pappus-bristles
relatively few, mostly in one series, often very slightly clavellate-
thickened toward the apex; receptacle moderately deeply alveolate,
sparingly glandular.

MEXICO: : Pedri ico, Pringle, We
4360 ae e ¥ PC) : emir arora Taipan elev. 2440 m., Pringle, no.

. naib

9949 ( .). Morexos: mountain side, near Cuernavaca, elev. uebl,

Pringle, no. 7178. Purnia: Cerro Tepoxuchitl, alt. 2330 m., vicint
Arséne, no. 2273; barranca near OS Ba Alamos, on the roa Vera Cruz,
alt. 2170 m., vicinity of Puebla, Arséne, no. 2084 (Gr., N. Y-).

MONOGRAPH OF THE GENUS PEREZIA 57

Teposcolula, alt. 2440 m., F. Lépez, no. 63. State NoT SPECIFIED: “Mex.,”
Alaman (fragment of TYPE).

P, hebeclada (DC.) Gray var. urolepis Robinson, Proc. Amer-
Acad. xliv. (1909) 625.

Differs from the typical form by having the phyllaries less rigid,
the outer much elongated, almost equalling the attenuate, subherba-
ceous inner in length, caudate-attenuate, the slender tips lax, some-
what spreading and tortuous; heads 23-25 mm. high; otherwise un-
differentiated from the typical form.
i Hipauco: Sierra de Pachuca, elev. 2900 m., Pringle, no. 13975

30. P. aspera Bacigalupi (see p. 18).

Rather rigid slender perennial, branching above, at least 4 dm. tall;
stem simple below, teretish, slender (diam. 1.5-2 mm.), striate, gla-
brous except for small tufts of a light brown silky tomentum in the
leaf-axils, smooth, flexuous, straw-colored and purplish-tinged; leaves
sessile or subsessile, rigid-coriaceous, oblong-lanceolate, pungently and
rigidly mucronate, cordate-amplexicaul, quite prominently reticulate,
glabrate, divaricately, irregularly and pungently toothed, minutely
callose-ciliolate, 4.5-5 cm. long, 1.5-2 em. wide, the upper smaller,
oblong, 2-3 cm. long, 7-10 mm. wide; heads 1-2, subcongested at the
ups of the elongated, ascending, leafy upper branchlets, 2-2.3 cm.
high, 30-flowered; peduncles 8-20 mm. long, usually bearing glabrous,
ascending-spreading, narrowly lanceolate to linear-subulate bracts
+10 mm. long; involucre campanulate, obtuse at the base, in about 3
series, 20-23 mm. long, equalling the pappus; phyllaries imbricated
only at the bases, rigid, not indurate, 2-2.3 mm. wide, attenuate-
subulate, very acute, glabrous, striate, -light green, purplish toward

e apices, sometimes very minutely ciliolate, the inner ones 17-19 mm.
long, the outer more rigid, 8-10 mm. long; corolla-lobes_ sparsely
glandular-ciliolate toward the base, the lavender-pink corolla other-
eg glabrous, 14-15 mm. long; pappus soft, tawny, the bristles very

e, in 2-3 series; achenes linear, minutely resinous-punctulate, about
Jean long, the epigynous disc funnelform; receptacle convex, deeply

Veolate, the membranaceous partitions between tha alveolae fim-
briolate-glandular.

MEXICO: C : i . (coll. of Frére Nil, no. 55,
Dee, 1908), Avsine, no, 3395. ane cae .

31. P. pinetorum T. S. Brandegee, Zoe v. (1901) 105. 2
» leafy perennial, branching above, 10 dm. high; stem rigid,

teretish, Shallowly grooved, stramineous, crisped-puberulent with

58 BACIGALUPI

articulated hairs; inflorescence-branchlets spreading, the slightest bit
flexuous, leafy up to the peduncles, striate-angular, densely glandular-
hirtellous and crisped-puberulent with hairs less evidently articulated
than those of the stem; leaves sessile, thin-chartaceous, ciliolate,
uniformly a little scabrid and puberulent on the not at all prominent
reticulations above, rather copiously and uniformly glandular-
hirtellous and articulately hirsutulous on the prominent veins and
reticulations beneath, the cauline oval-oblong, acuminate, broadly
auriculate-clasping, unequally spinulose-dentate, 18-20 cm. long,
9-10 em. wide, those of the branchlets lance-oblong, spinulose-dentic-
ulate, shortly acuminate, cordate-clasping, the auricles broad and
truncatish-rounded, 4-6 cm. long, 16-20 mm. wide; inflorescence a
widely divergent, leafy, paniculate cyme, the 1—5-headed cymules not
at all congested; peduncles 1.5-3 cm. long, bearing a few, spreading,
lance-subulate, glandular scales 2.5-4 mm. long; heads campanulate,
15-20-flowered, 1.5-2 cm. high; phyllaries rigid-herbaceous, attenuate-
subulate from a broadish base, gradually and evenly narrowed to a
very acute tip, loosely imbricated in about 3. series, glandular-
puberulent dorsally, fimbriolate-ciliolate on the narrow scarious mar-
gin, the longest inner ones 14-15 mm. long; corollas sparingly stipitate-
glandular exteriorly, lavender-pink, 11-13 mm. long; pappus soft,
biseriate, white, the bristles very fine and not thickened toward the
apex; achenes linear-cylindric, 5 mm. long, densely covered with
shortly stipitate amber glands and rather densely spreading-hirsute
as well; receptacle shallowly scrobiculate, slightly convex, moderately
densely beset with amber-colored shortly stipitate glands.

MEXICO: Basa Cauirornia: Sierra de la Laguna, Jan. 22, 1899, T. 8
Brandegee (isotype, Gr.).

Sierra de la Laguna is the northern continuation of the Sierra de la
Victoria, in the Cape San Luis region at the southern extremity of the
peninsula of Lower California.

i P. turbinata La Llave & Lexarza, Nov. Veg. Descr. i. (1824)

Erect, apparently simple, glabrate, more or less leafy perennial
sometimes 12.5 dm. tall; stems slender, especially above, conspicuously
roughened below to smoothish and glaucescent above, slightly angl
to somewhat sulcate; leaves sessile, chartaceous to chartaceo-coma
ceous, oblong to cuneate-oblong, closely spinulose-denticulate,

MONOGRAPH OF THE GENUS PEREZIA 59

or linear-lanceolate to linear, attenuate, 1.5-5 cm. long; inflorescence
openly paniculate, the heads usually well separated by the long
slender peduncles, these bearing remote lance-subulate to subulate
scales or mostly naked, glabrate to glabrous, angled; heads 20-28 mm.
high, 20-30-flowered; involucres widely obconic to turbinate-campanu-
late and somewhat rounded at the base, 15-22 mm. high; phyllaries
fairly well imbricated in 4—5 series, the outer lance-attenuate, the inner
subulate, often the slightest bit acuminate as well, chartaceous, not
rigid, often slightly scarious-edged, occasionally purplish, glabrous
dorsally, some rarely sericeo-ciliolate ; corollas lavender-pink; imma-
ture achenes scabrous with upwardly appressed hairs, shortly stipitate-
glandular as well, 5-6 mm. long; pappus soft; receptacle fairly deeply
alveolate, stipitate-glandular.

MEXICO: Sra
common, Valley of Mexico, Schaffner, no. 226; open wood, elev. 2440 m.,
Salto de Agua, Purpus, Ino. 1558 (Gr., N. Y., P. C.); dry slopes, Ixtaccihuatl,
2100-2400 m., Purpus, no. 63 (1903); Cofiode Mountains, near Mexico
Schmitz, no. 701.

In Pl. Wright. i. (1852) 126, the specimens which Gray referred to

P. turbinata should have been referred to P. formosa (D. Don) Gray,
as he later admitted in Proc. Amer. Acad. xix. (1883) 58. It is the
Interpretation Gray gave to P. turbinata in the latter publication that
is here followed.
Following Gray’s earlier interpretation in Pl. Wright., Schultz
Bipontinus, who sought to refer the genus Perezia to Trixis, founded
the binomial Trixis turbinata on specimens properly referable to
Perezia formosa (D. Don) Gray. (Trizxis turbinata Sch. Bip. in Seem.
Bot. Voy. Herald (1856) 315, pl. 55.)

33. P. Thurberi Gray, Mem. Amer. Acad. n. s. v. (1855) 324 [PI.
Nov. Thurb.].

Erect, usually simple, densely leafy perennial, 4-15 dm. tall; stems
Tobust, shallowly suleate below to striate above, densely glandular-
Puberulent, occasionally scabrid above; leaves sessile, chartaceo-
coriaceous, the larger oval- to obovate-elliptic, more usually short-

winate and rarely pungent at the often rounded tip, shortly
Sagittate- or more rarely shallowly cordate-clasping, the floral leaves
oblong-elliptic to lance-attenuate, merely abruptly rounded at the

» all closely acerose-denticulate, less often saliently and irregularly

te, densely glandular-puberulent, scabrid, conspicuously reticu-

late beneath, the cauline 5-18 em. long, 3-10 cm. wide, the floral 1.5-4
ie long, 6-20 mm. wide; subcongested cymes of the usually elongated
‘0 triangular thyrsiform inflorescence exceeding the leafy bracts; heads

60 BACIGALUPI

at anthesis 16-20 mm. high, 4—6-flowered; involucre obconic-campanu-
late, 7-10 mm. long; phyllaries oblong-oblanceolate, the outer some-
times somewhat lanceolate, all attenuate-acuminate, minutely glandu-
lar-puberulent, somewhat indurate below; corollas lavender-pink,
upper part of the tube and the segments glandular-papillose; achenes
subcylindric to subfusiform, striate, glandular, 5-6 mm. long; pappus
rather rigid, only an occasional bristle very slightly clavellate-thick-
ened above; receptacle glandular.

Southeastern Arizona and adjacent New Mexico, southward along
the Sierra Madre to Guanaxuato.

UNITED STATES: Arizona: Santa Rita Mountains, June 3, 1884, Pringle

(N. Y.) [leaves abnormally narrow]; Wilgus Ranch, alt. 1830 m., Chiricahua
Mountains, J. C. Blumer, no. 1771 (Gr., N. Y.); dry slopes, Dixie Cafion,
O ’

According to Dugés, this species is known in Guanaxuato as “Cal-
zadilla” and “Cola de Zorra.

The last specimen cited, collected by Dugés, has heads decidedly
immature but doubtless belongs here. The lower leaves show a strong
tendency to adnation, a feature noted neither in the other specimens
of P. Thurbert at hand, nor, except for P. Alamani, in those of any
other species of the section Acourtia. It is possible that Lessing had
just such a form of P. Thurberi before him when he described Dumer-
dia Humboldtii. ee

34. P. Pringlei Robinson & Greenman, Proc. Amer. Acad. XxiX.
(1894) 388.

Erect, apparently strict and simple, tall perennial (definite .
ra

coarsely

the dark, blue-green, upper surface, fulvous-pubescent with articur

MONOGRAPH OF THE GENUS PEREZIA 61

lated hairs on the grey-green lower surface, 7-11 cm. long, 4-5 em.
wide, the upper inclined to be narrower, lance-oblong and quite
arcuate and conduplicate, more densely pubescent beneath, mostly
4-7 em. long; inflorescence moderately leafy, narrowly racemiform,
in the specimen at hand (type) not more than 12 em. wide, 29 cm.
long; heads for the most part single, on rigid, ascending, pubescent,
leafy-bracted peduncles 2.5-6 cm. long, very broadly campanulate,
about 50-flowered, mostly 2.5 cm. wide and 3 cm. high; phyllaries well
imbricated in 6-7 series, acute, the inner oblong-lanceolate and atten-
uate-acuminate, the outer ovate-lanceolate and much more abruptly
acuminate, all glabrous, striate dorsally, shaggily sericeo-ciliate, rigid,
greenish, deep purplish on the exposed portions; corollas not exserted
beyond the involucre, lavender-pink, essentially glabrous, 14 mm. long,
the two lips very short, exceeded by the anther-column by as much as
2mm.; pappus relatively soft, whitish, biseriate, included within the
involucre; immature achenes cylindric, densely appressed-hirsute, 4-5
mm. long, expanded above into a funnelform epigynous disc; recepta-
cle slightly convex, shallowly scrobiculate, essentially glabrous.

MEXICO: Micuoac&w: hills near Morelia, Pringle, no. 5464 (TYPE).

A striking species as yet too little collected.

35. P. Dugesii Gray.

Slender, laxly and widely branching, flexible, evidently somewhat
scandent perennial 18-30 dm. high; stems striate, glandular-scabrid,
often ultimately quite smooth, greenish to stramineous, sometimes
purplish-red; leaves sessile, submembranaceous to chartaceous, very
closely spinulose-denticulate or shallowly triangular-dentate with the

*th terminated by spinules to rather coarsely and irregularly dentate,
minutely callose-ciliolate, moderately to quite prominently reticulate,
more prominently so beneath, especially in the forms with thicker
leaves, glabrous and smooth to scabrid above, smoothish to scabrid,
glabrate to crisped-puberulent and glandular-punctate beneath, broad-
ly lanceolate to oblong-lanceolate or obovate-spatulate, cordate to
Subsagittate at the base, the two auricles often overlapping, attenuate-
acuminate and very acute to obtuse at the apex, 7-15 cm. long,
25-7 m. wide; upper branchlets very slender and flexible, gradually

ming more sharply flexuous outward; inflorescences several to
ay » Isolated, separated by the leafy terminal portions of the branch-

ts, consisting of 2-4-headed glomerulate cymules, the axis and pe-
on.» Seniculate, leaf-bearing at the nodes, from sparingly to densely
“lsped-pubescent; heads 5-7-flowered; involucres very slender,
Cylindric-obconic, 9-12 mm. high; phyllaries for the most part oblong,

62 BACIGALUPI

glabrous dorsally, chartaceous, rigid only at the base, sericeo-ciliate to
eciliate, the very short outer narrow, acute, the widest narrowly ovate-
lanceolate, well imbricated in 3-4 series, the inner exposed for most
of their lengths, in but 2-3 series, for the most part abruptly acuminate
from an obtuse tip, sometimes more gradually narrowed and acuminate;
corollas from sparsely to rather densely and very minutely glandular-
papillose exteriorly, lavender-pink; pappus soft to somewhat rigid,
intermittently biseriate, usually exceeding the innermost phyllaries by

mm.; achenes densely glandular-punctate and very sparingly to
very densely hirsute, narrowly fusiform, rather more conspicuously
narrowed above than below, 6-7 mm. long, the epigynous disc very
narrow.

Key To VARIETIES.

irregular! : A
gularly dentate; achenes very densely appressed. ie . p. pibsioen.
Var. a. typica. Leaves only moderately reticulate, more promi-

nently so beneath than above, very closely spinulose-denticulate to

shallowly triangular-dentate with the teeth terminated by spinules,
broadly lanceolate to oblong-lanceolate, mostly attenuate-acuminate
and very acute at the apex, 7-15 cm. long, 2.5-5.5 em. wide; corolla
sparsely and very minutely glandular-papillose exteriorly; achenes very
sparingly appressed-hirsute.—P. Dugesii Gray, Proc. Amer. Acad. xix.
(1883) 60.

2220 m., Pringle, no. $253 (N. Y., labelled “Aster pauciflorus,’

label confusion); Nov. 22, 1892, Pringle, no. 5254 (see below). PUEBLA:

del Monte, Purpus, no. 2977. Vera Cruz: hills near Jalapa, elev. 1220 ms
Pringle, no. 8131 (Gr., P. C.); Orizaba, Mohr.

The additional data, in Pringle’s own hand, on the label of his no.
5254, above cited, is as follows: “warm wooded ledges, mts. near LAXt
Chapala.” Lake Chapala, however, is situated between Jalisco and
Michoacfén, more than 150 miles west of the state of Mexico (the
printed label bears the inscription “State of Mexico”’). Specimens

ing numbers immediately preceding and following no. 5254 were
found to have come unquestionably from the region just south
city of Mexico. Perhaps “Lake Chapala” here is a lapsus calamet 19°
Lake Chaleo.

:
:
|

MONOGRAPH OF THE GENUS PEREZIA 63

Specimens referable to P. oxylepis have been confused with P.
Dugesi var. typica. The inflorescences of the two species are quite
dissimilar. P. oxylepis has the corymbose, many-headed terminal
inflorescence typical of the majority of the species of Perezia § Acour-
tia, while P. Dugesii has several isolated inflorescences which termin-
ate the upper lateral branches as well as the central axis. Moreover,
the very acute attenuate-acuminate and copiously glandular phyl-
laries and wider 15-21-flowered heads of P. oxylepis at once distin-

It.

Var. 8. pilulosa Bacigalupi (see p. 19). Leaves prominently re-
ticulate, especially beneath, coarsely and irregularly dentate, obovate-
spatulate, rounded and obtuse to very shortly acuminate at the apex,
11.5 em. long, 5-7 cm. wide; corollas more densely glandular-
Papillose externally than in var. typica; achenes very densely ap-
pressed-hirsute. The veins and veinlets of the lower leaf-surfaces
are often much more strikingly hirsutulous than in var. typica.

MEXICO: Oaxaca: below Jayacatlan, elev. 1067 m., L. C. Smith, no. 373
(tvPz, Gr.). Vera Cruz: Orizaba, Muiller, in 1855 (N. Y.).

36. P, Alamani (DC.) Hemsl.

Erect, densely leafy perennial 3-6 dm. high; stem rigid, simple,
at least below, teretish, leafy to the summit, in the dried specimen
more or less costulate-angular, glandular-puberulent, scabrid, stra-
mineous, less commonly purplish above; leaves more or less imbri-
tated, sessile, glandular-scabrid and moderately reticulate on both
Surlaces, often somewhat nitidulous above, serrate-dentate to widely
tmangular-dentate, rigid-chartaceous to -coriaceous, ovate to oblong
or oval, truncate-cordate at the base to distinctly cordate-clasping,
the lower cauline more or less adnate to the stem below, the auricles

; inflorescence many-headed, flat-corymbose, densely congested,
very little to not at all exceeding the terminal leaves; heads 7—12-
flowered; involucres cylindric-campanulate, 9-14 mm. high, somewhat
®cutish at the base; phyllaries widely ovate-lanceolate to oblong,
densely glandular-puberulent to glabrate, eciliolate to sericeo-ciliate
or fulvous-arachnoid on the margins, mostly somewhat acuminate,
i 9-5 series; corollas lavender-pink, the three terminal teeth of the
Outer ligulate lip and the tips of the two segments of the inner lip
mlnutely papillose, occasionally the inner segments, lower portion of

late lip and the summit of the tube more or less covered with
very minute amber-colored outgrowths; pappus quite stiff, not greatly
Nie the involucre, 1-2-seriate; immature achenes glandular-
Puberulent, about 4 mm. long; receptacle shallowly scrobiculate,
“bristly and sparingly glandular to glabrate.

64 BACIGALUPI

Key To VARIETIES.

a. Involucre 12-14 mm. high; phyllaries glabrate, often sericeo-

ciliate, relatively loosely imbricated in 3-4 series, the inner

often exposed for more than half their length........... Var. a. typica.
a. Involucre 9-10 mm 1 igl “t 1 y al 2 gl j J} I 1 7 4 ]

more or less fulvous-arachnoid, relatively well imbricated

in 5 series, the inner often with less than half their length

expose : :
b. Cauline leaves ovate- to lance-oblong, membranaceo-coria-

ceous, , g thei te to lance-
ovate; stem divaricately branched at the top, each :
branch terminated by an inflorescence.............. Var. y. oolepis.

Var. a. typica. Stem simple below, the inflorescence-bearing
branches sometimes as many as 7, often widely divaricate; leaves
rigid-coriaceous, ovate, lance-ovate or rarely lance-elliptic, serrate-
dentate, sometimes coarsely so, to triangular-dentate, the upper
little, if at all, adnate, the larger cauline adnate for not more than 2.5
mm. and usually truncate-cordate to very shallowly cordate-clasping
at the base, 4-5 cm. long, 2-3 cm. wide; heads 7-10-flowered; involu-
cres 12-14 mm. high; phyllaries glabrate to densely glandular dorsally,
occasionally sericeo-ciliate, imbricated in 3 to mostly 4 series, the outer
not wider than broadly lanceolate; corollas, especially the inner seg-
ments, somewhat more densely covered with amber-colored excres-
cences than in the other varieties; receptacle glabrate—Dumerilia
Alamani DC. Prod. vii. (1838) 67; Mem. ix. (1838) 41, pl. 17; Pereata
Alamani Hemsl. Biol. Centr.-Amer. Bot. ii. (1881) 255.

Cuyamaloya Station, elev. 2290 m.. Pringle, no. 10270 (Gr. N. ¥.). Mextco:
‘Valle de Toluca,” near Tenancingo, Schaffner, Sept., 1854; same
collector, Oct. 1, 1854 EBLA: Esperanza, alt. 2450 m., Arsene, no.
(Gr., N. Y.); grassy hills, Esperanza, Purpus, no. 2631 (Gr., N. Y.). No stat®
SPECIFIED: “Mexico,” oa (fragment of authentic material of De Can

MEXICO: Guanaxvato: Guanaxuato, D2 Hipateo: rocky hills,
2

adnation decreasing, the uppermost often merely shortly rounded at
the base; inflorescence single and terminal; heads 9-12-flowerets
involucres 9-10 mm. long; phyllaries, especially the ovate-lance?

|
j
|

MONOGRAPH OF THE GENUS PEREZIA 65

outer, glandular-puberulent, well imbricated, the oblong inner be-
coming glabrate dorsally, all more or less fulvous-arachnoid on the
margins, in 5 series; corollas glabrous except for occasional amber-
colored outgrowths at the summit of the corrolla-tube; receptacle
shallowly scrobiculate, sparingly glandular, short-bristly when young.
—P. adnata Gray, Pl. Wright. i. (1852) 127.

MEXICO: Micwoac&n: forests of pine near Morelia, “Terre froide,”’ Sept.
1844, Ghiesbrecht, no. 378 (rypE of P. adnata); Quinceo, vicinity of Morelia,
elev. 2800 meters, Arsene, no. 5800.

Var. y. oolepis (Bartlett), comb. nov. Leaves coriaceous, the cau-
line broadly oval, 10-11.5 cm. long, 7-7.5 em. wide, the lower ovate-
spatulate, the uppermost ovate to lance-ovate; stem divaricately
branching above, each branch terminated by an inflorescence; phyl-
laries, especially the outer, inclined to be broader than in the other
Varieties.—Perezia adnata Gray, var. oolepis Bartlett, Proc. Amer.
Acad. xliv. (1909) 637.

MEXICO: Srate or Mexico: rocky hills, Tultenango, Pringle, no. 3244
(Gr. N. Y.); hills near Tultenango Station, elev. 2530 m., Pringle, no. 9945
(Gr., N. Y.).

There is some probability that this variety may be the same as
Perezia fruticosa of La Llave & Lexarza, since the above specimens
fit, in a general way, the diagnosis [Nov. Veg. Descr. (1824) 26] of
that species. In view of the very general nature of the diagnosis,
this inference must, unfortunately, remain merely conjectural.

Bartlett [Proc. Amer. Acad. xliv. (1909) 636-7] pointed out several
characters which might be used to distinguish from P. Alamani, both

- adnata and P. adnata var. oolepis. Specimens inserted in the
Gray Herbarium since 1909 largely invalidate many of the distinctions
set forth in Bartlett’s publication. Arséne’s number 5800, which
matches very closely Gray’s P. adnata, has glandular-puberulent phyl-
laries varying from 14 to 20 innumber. The characteristic numbers of
Phyllaries attributed by Bartlett to P. adnata, P. adnata var. oolepis
and P. Alamani are 28, 21 and 14 respectively, those of the last being

T distinguished as typically “paene glabris.” Yet, Purpus, no.
2631, Which agrees with all other characters of P. Alamani, has its
pyivies, which number 13, copiously viscid-glandular with stalked

An attempt has been made by Bartlett to use the number of pappus-
bristles as a character to help distinguish these entities. This
Whi cter is not any more reliable here than elsewhere in the genus.

ile the number of pappus-bristles in the flowers of any given head

66 BACIGALUPI

often tends to be approximately the same, there will be found to be
no uniformity in this respect when the flowers are chosen at random
even from neighboring heads.

In this same publication, the inner lip of the corolla of P. Alamani
is said to be papillose-pubescent on the outside while that of both P.
adnata and P. adnata var. oolepis is described as glabrous. To be
sure, in many of the specimens cited by Bartlett under P. Alamani,
the inner lip of the corolla, seen under a binocular which magnifies
somewhat more highly than a X 20 hand-lens, may be seen to be more
or less covered with semi-appressed amber-colored projections which
are easily detachable with a needle. These are most abundant at the
summit of the corolla-tube and occasionally are found some distance up
on the ligulate outer lip. An examination of material of both P. adnata
and P. adnata var. oolepis, however, shows that these amber-colored
projections are very often present on the corollas of these two entities
as well, though here much more restricted toward the bases of both
the inner and outer lips. This difference, then, proves to be one of de-
gree rather than an absolute one, and it is further noteworthy that this
corolla character is by no means confined to this species.

An interesting case of teratology was noted in the pappus from
several heads taken from Purpus, no. 2631. In these, the ring of
pappus-bristles was interrupted and, at the point of interruption,
replaced by a bract-like structure of a texture closely simulating that
of the inner phyllaries. The margins of some of some of these had the
scabrous bristliness characteristic of the pappus-hairs. In one head,
three of the seven flowers present exhibited this irregularity. That
this phenomenon is by no means of uncommon occurrence is indicated
by the following statement, quoted from the general introductory
remarks on the Compositae in O. Penzig’s “ Pflanzen-Teratologie,” 1-
ed. 1 (1894) 55; ii. ed. 2 (1921) 469:“. . . an Stelle des Pappus
treten gemeinhin lineare oder lanzettliche (oft mehr als fiinf!) Kelch-
blittchen . . .”

37. P. Palmeri Wats. Proc. Amer. Acad. xxiv. (1889) 58.

“Tall and branching” leafy perennial; stems more slender and not
so rigid as in most of its congeners, glandular-puberulent, scabrid,
slightly flexuous, very slightly striate, at least above; leaves sessl®,
thin-chartaceous, moderately reticulate, the midrib prominent, &
pecially on the lower side, oblong-lanceolate, attenuate-acuminat®,
shallowly cordate-clasping, glandular-scabrid, especially above, very
acutely pungent-dentate and denticulate, the longer teeth not more

an 2 mm. long, 11-15 em. long, 3.5-4.5 cm. wide, the lower doubt-

MONOGRAPH OF THE GENUS PEREZIA 67

less larger; floral leaves similar, linear-lanceolate, 2-6 cm. long; in-
florescence diffusely paniculate, the heads not glomerate, 13-18 mm.
high, about 10-flowered, on densely glandular-puberulent peduncles
1-3 em. long; involucres widely campanulate, 10 mm. high; phyllaries
not indurate, greenish, well imbricated, glandular-puberulent dorsally,
the margins inclined to be scarious and fimbrillate-ciliolate, the outer
ovatish, sharply acuminate, often exceeding 2 mm. in width, the inner
spatulate-oblong, broadest above the middle where 2 mm. wide,
abruptly acuminate; corollas “lilac”; pappus not rigid, biseriate;
achenes (immature) subcylindric, densely stipitate-glandular, spar-
ingly hirsutulous as well, 4 mm. long; receptacle shallowly alveolate,
stipitate-glandular.
ange as yet only imperfectly known; the type specimen from
north central Lower California.
MEXICO: Basa CauirornrA: mountain sides near Los Angeles Bay, Gulf
of California, Palmer, no. 527, collected in 1887 (rypr, Gr. and isotype, N. Y.).
38. P. oxylepis (Sch. Bip.) Gray, Proc. Amer. Acad. xv. (1880) 41;
? Acourtia oxylepis Sch. Bip. ex Gray, in syn., loc. cit.
tect, apparently simple, leafy, aromatic and more or less glandular
perennial; stems rigid, only the slightest bit flexuous above, more or
less striate, densely glandular-puberulent with an admixture of crisped
hairs, somewhat scabrid, tawny-floccose in the lower leaf-axils; leaves
sessile, thin-chartaceous, scabrid above, glandular-punctate and
‘isped-puberulent on the midribs and veins beneath, oblong or lance-
oblong, spinulose-denticulate, cordate-amplexicaul to subsagittate-
amplexicaul, acute to less commonly obtuse at the apex, the lower
usually clasping the stem so as to enclose completely the tomentum
in the axils, these lower leaves only somewhat prominently reticulate
eath and often irregularly subserrate-dentate, the middle cauline
+16 em. long, 3-6 cm. wide, those of the inflorescence 3-5 em. long,
1-2 em, wide; inflorescence widely and loosely branched; each densely
leafy, copiously glandular branchlet terminated by a subcongested to
Somewhat lax corymbiform cyme; ultimate peduncles bearing spread-
/8 or recurved lance-subulate bractlets; heads 15-18 mm. high,
(12- ace. to Gray) 15-21-flowered; involucre campanulate, 11-13 mm.
high; phyllaries in 3-4 series, thin, not indurate, copiously stipitate-
glandular, often purplish-tinged, the outer ovate-lanceolate, the inner
hnee-oblong, gradually attenuate-acuminate, very acute; corollas
ender-pink, very sparsely glandular-punctate exteriorly; pappus
moderately stiff, biseriate; mature achenes narrowly fusiform, densely
jPitate-glandular as well as sparsely appressed-hispid, about 6 mm.
ng; receptacle convex, scrobiculate, stipitate-glandular.

OS BACIGALUPI

MEXICO: San Luts Porosgi: in the region of San Luis Potosi, alt. 1830-2440
m., Parry & Palmer (1878), no. 547, in part (TYPE); ‘‘in umbrosis prope Penas-
co, ex convalli San Luis Potosi,” Schaffner, no. 377 (Gr., N. Y.).. GUANAXUATO:
mountain valleys, Jaral, W. Schumann (1886), no. 120.

This species has been confused with P. Dugesii Gray. The distine-
tions between these two species are discussed in conjunction with the
treatment of the latter.

On the label of his no. 377, Schaffner made the following notation:
“Eine ausgezeichnete im frischen Zustande. . .etwasklebrig. . -
sehr aromatische Pflanze.”’ Inasmuch as the dried specimen in this
instance gave not the slightest evidence of its ever having been an
aromatic plant, it is quite probable that the exudations of many of the
other more glandular species of Perezia § Acourtia are, in the fresh
state, equally aromatic.

39. P. Seemannii Gray, Pl. Wright. i. (1852) 127; Tris Scemannii
Sch. Bip. in Seem. Bot. Voy. Herald (1856) 315; Acourtia Seemannu
Sch. Bip. in Seem. loc. cit., in syn. and pl. 54. :

Erect, leafy, rigid, probably tall perennial, simple below the 1-
florescence; upper portion of stem terete, striate, smooth, glabrous,

ard

inflorescence entire, narrowly elliptic to spatulate-oblong,
6-13 mm. long; inflorescence wide and open, made up of laxly disposed
corymbiform cymes, the 8-12-flowered heads not at all congested, on
slender, subulate-squamose peduncles 1-2 cm. long; involucres Nar
rowly obconic, acute at the base, 8-9 mm. long; phyllaries indurate,
oblong-lanceolate, acuminate and acute, glandular-punctate dorsaly,
glandular-ciliolate, in 3-4 series, the lower well imbricated, the lower-
most cuspidate; corollas lavender-pink, very sparingly §
papillose externally; pappus moderately stiff, uniseriate,
mm. beyond the fruiting involucre; achenes 3-4 mm.
oblong, moderately to densely glandular-papillose, somew.
rowed at the neck; epigynous dise slight; receptacle slightly conve®
scrobiculate, glandular-papillose.

MEXICO: Srare Nor sprcirimp: Seemann, “N. W. Mexico,” in the a
Madre (Tryp). Seemann collected in southern Durango and Sinaloa
Nayarit and in northern Jalisco.

MONOGRAPH OF THE GENUS PEREZIA 69

40. P. grandifolia Wats. Proc. Amer. Acad. xxv. (1890) 156.

Erect, robust, leafy perennial, simple below the inflorescence, 18-30
dm. high; stem stout, rigid, terete, suleate and 1 cm. thick below,
smooth, glabrous, greenish to stramineous, often purplish-tinged,
becoming flexuous and angled upward; leaves sessile, subcoriaceous,
finely reticulate, glabrous, nitidulous and, save for the reticulations,
smooth above, crisped-puberulent on the reticulations, glandular-
punctulate and slightly scabrid beneath, the larger cauline obovate
with a short and wide spatulate base to oval-ovate, repandly serrate-
dentate toward the broadly auriculate, cordate-clasping base, be-
coming repandly spinulose-denticulate toward the rounded and some-
times mucronulate apex, 27-32 cm. long, 17-20 cm. wide, the upper
from oblong to elliptic-oblong, auriculate-clasping, glabrous except for
the sessile glands beneath, 3-8 cm. long, 2.5-4.5 cm. wide; floral leaves
1.5-2 em. long, oblong to oblanceolate, often apiculate, the lower
auriculate; inflorescence composed of corymbiform laxish cymules:
disposed at the periphery of a more or less hemispheric cyme 2.5-3
dm. wide, the slender ramifications of the cymules densely glandular-
puberulent, the ultimate peduncles bearing somewhat spreading,
glandular, lance-subulate scales; heads 13-16 mm. high, 10-12-flow-
ered; involucre obconic-cylindric, 8-10 mm. high; phyllaries mostly
in 4 series, thin, indurate below, densely glandular-punctulate dor-
sally, glandular-ciliolate, often purplish, oblong-lanceolate, the inner
~4./mm. wide, shortly acuminate; corollas lavender-pink, the upper
Portion of the tube densely glandular-punctulate; pappus stiff and
very scabrous, slightly clavellate-thickened toward the apex, sub-
unseriate; achenes cylindric, copiously glandular-papillose, 4-5 mm.
long, the epigynous dise very slight to quite broad; receptacle flattish,
slightly scrobiculate, glandular-punctate.
1858 ICO: JaLisco: cool, roy, hillsides near vhsnctear tie ae no.
elev, 1525 ta, Pringle, nor bods (Gro N.Y). Navanir: Tepic, M. B. Jones,
nO. 23356 (N. Y) , no. T., IN. X.)- ; ?

A species with strikingly large leaves, the largest thus far known in
~ genus. Watson described the achenes as glabrous; all those ex-
“mined, while certainly destitute of trichomes, were found to be
densely glandular-punctate. ;

41. P. microcephala (DC.) Gray, Pl. Wright. i. (1852) 127;

‘ourtia microcephala DC. Prod. vii. (1838) 66; Perezia sericophylla
Millsp, & Nutt. Field Mus. Publ. Bot. v. (1923) 297 (incidental men-
“ton); illustration in Jepson, Man. Fl. Pl. Calif. (1925) 1013.

70 BACIGALUPI

Erect, stout, usually very leafy perennial, 6-12 dm. high; stem rigid,
sometimes slightly flexuous, more or less densely glandular-scabrid,
striate; leaves sessile, oblong-ovate to elliptic-oblong, the larger
usually deeply cordate-clasping, the base sometimes varying from
subattenuate to shortly truncatish, mostly obtusish at the apex,
occasionally mucronate, chartaceous to chartaceo-coriaceous, irregu-
larly spinulose-denticulate, only the principal veins prominent be-
neath, densely glandular-scabrid to -scabrous on both surfaces, 5-15
cm. long, 2.5-6 em. wide; inflorescence ample, primarily loosely
paniculate, the bracts leaf-like, the ultimate peduncles usually densely
glandular; heads approximated into cymules at the tips of the in-
florescence-branchlets, 10-13 mm. high, 10-20-flowered; phyllaries
oblong-oblanceolate, abruptly acuminate on the broadened apex,
densely glandular-puberulent dorsally, often glandular-ciliolate, loose-
ly imbricate, green, not indurate, usually in 3 series; corollas lavender-
pink; achenes more or less cylindric, bristly glandular-puberulent,
— mm. long; receptacle very shallowly alveolate, glandular-puberu-
ent.

Coastal southern California from San Luis Obispo County south-
ward, doubtless extending some distance into Lower California.

CALIFORNIA: San Luis Opisro Co.: Templeton, Abrams, no. 5045.
ANTA Barpara Co.: Mountain Drive, Santa Barbara, Abrams, no. 4151
(N. Y.); hills back of Santa Barbara, Rothrock, no. 121; oak pig _—

~)

TE “s : lonia F
railroad, Linda Vista, Macbride & Payson, no. 806; chaparral, Howard
Canyon, La Jolla, F. E. & E. 8. Clements, no. 259 (Gr., N. Y.); on sandy hills
ito, Spencer, no. 904 (N. Y.); San Diego, D. Cleveland, in 1874; idem,
1875; idem, without date (N. Y.);San Di o, T. S. Brandegee, no. 1654; ne
north of San ego, Chandler, no. 5356 (N. Y.); near San Diego, Palmer,
1875; same locality, Palmer (1875), no. 203 (N. Y.).

42. P. collina Wats. Proc. Amer. Acad. xxvi. (1891) 144.

Tall, leafy, erect, stout and rigid perennial, apparently simple below
the inflorescence; stem teretish, angular-striate, slightly a
above, very smooth and glabrous, stramineous. and purple-tn :
leaves rigid-coriaceous, subsessile, glabrous, sometimes slightly rou
ened on the close reticulations, widely elliptic-oblanceolate, 9-15 cm.
long, 28-60mm. wide, coarsely and unequally serrate-dentate, the t

MONOGRAPH OF THE GENUS PEREZIA 71

tipped by a spinule, strongly conduplicate, arcuate, ascending only
toward the base, more or less abruptly narrowed at the apex and acu-
minate, very acute, rounded to subhastate at the not strongly nar-
rowed base, the lower only about half-clasping the stem, the upper
hardly or not at all clasping; floral leaves much reduced, denticulate to
subentire; inflorescence a more or less leafy paniculate cyme, the
cymules glomerulate and often densely crowded at the axis-apex, the
ultimate peduncles usually pubescent, subtended by a rigid, very
narrowly spatulate to lance-subulate sometimes pungent bract;
fruiting heads 15 mm. high, consistently 8-flowered; involucres cylin-
dric-eampanulate, acutish at the base, about 12 mm. high; phyllaries
indurate save at the apices of the innermost, glabrate dorsally, shortly
acuminate and acute, obscurely fimbriolate-ciliolate to subentire,
more rarely glandular-ciliolate, stramineous to slightly purplish,
imbricated in 5-6 series, the outer subulate-lanceolate, rigid, often
pungently tipped, the inner linear-oblong, not at all narrowed below,
1 mm. wide; corollas with the outer lip (ligulate in other species of
Perezia) ultimately parted almost to the base into its 3 component

ear segments, these and the 2 inner ones subequal; pappus soft,
whitish, 2-3-seriate, not thickened apically; achenes 5-6 mm. long,
very slenderly cylindric, sometimes slightly narrowed above, densely
glandular-punctulate and hirsutulous with stiff ascending white
trichomes; receptacle slightly convex, rather deeply alveolate, densely
gland-studded.

MEXICO: Jatisco: hills near Guadalajara, Pringle, no. 2123 (TYPE).

_ The corolla in this species is anomalous for Perezia, simulating, in
its 5-partite character, that of the not too closely related genus
Gochnatia. If, however, the corollas from young flowers (all too few
on the type specimen) be examined, the outer ligulate lip will be seen
to be still unparted, though the lines along which the ligule is later to
split may be clearly discerned. In florets a little more advanced, the
Segments will be seen to have separated below, though still coalesced
toward the tips

Per,

te BACIGALUPI

in length, while that of P. rigida var. acuminata is not more than 9-10
mm. long.

43. P, michoacana Robinson, Proc. Amer. Acad. xxvii. (1892) 179.

Erect, simple, leafy perennial 12-18 dm. tall; stem rigid, glabrous,
striate, dark purplish, striate-angled above; leaves sessile, subcoria-
ceous, moderately reticulate, especially below, the cauline glabrous
save for a sparse minute pubescence on the midrib and veins beneath,
obovate-spatulate, closely, very acutely and somewhat irregularly
denticulate, auriculate-clasping, rounded at the apex, 14-15.5 em.
long, 6.5-7.5 cm. wide, the upper inclined to be shortly acuminate and
acutish, spatulate-oblong, entirely glabrous, more evenly callous-
denticulate, 4-6 cm. long, 1.5-2.5 em. wide, those of the inflorescence
much sparser, lanceolate to lance-subulate, entire; inflorescence widely
and loosely corymbose, the ultimate peduncles 1-2 cm. long, fulvous-
tomentulose; heads relatively few, not glomerulate, 17-20 mm. high,
40-50-flowered; involucre hemispheric-campanulate, 1.5 cm. high;
phyllaries dark purplish on the exposed portions, fulvous-tomentose,
especially on the margins, 5-6-seried, the short outer ovatish, the
inner narrowly elliptic-oblong, about 11 mm. long, all and especially
the outermost abruptly acuminate and acute; corollas lavender-pink;
pappus not very rigid, a few of the bristles the slightest bit clavellate-
thickened toward the apex; immature achenes subcylindric, copiously
covered with dark brown sessile glands; receptacle shallowly alveolate,
studded with sessile glands, crisped-pilose as well.

Apparently localized in Michoacdn, Mexico.

MEXICO: Micnoac4y: hills near Patzcuaro, Pringle, no. 3988 (T¥P#, Gr.
and isotype, N. Y.).

. P. cuernavacana Robinson & Greenman, Proc. Amer. Acad.

Xxxii. (1896) 50. :

Erect, leafy perennial, for the most part glabrous, 6-9 dm. high,
branching above; stems sometimes several from the caudex, slender,
striate-angular, the least bit flexuous, stramineous to purplish, smoot),
glabrous except for tufts of light brown tomentum in the leaf-axils;

Ce a ee ee ee ee ea

MONOGRAPH OF THE GENUS PEREZIA 73

peduncles 3-8 cm. long, or at the tips of the axis and branchlets;
scales rigid, not imbricated, spreading, acicular-subulate and shaggily
ciliolate, gradating into the phyllaries and sometimes into the more
foliaceous bracts below when these are present; heads 2.5-3 em. high,
about 30-flowered; involucre turbinate, 20-22 mm. high; phyllaries in
about 7 series, thin-coriaceous, striate, bright green, purplish and
sericeo-ciliate with long, weak, tawny hairs toward the tips, somewhat
scarious on the margin except toward the obtuse, mucronulate tip,
the basal well imbricated, ovatish to lance-ovate, 3-8 mm. long,
those of the 3 innermost series more loosely imbricated, oblong-
lanceolate to oblanceolate, 13-16 mm. long, the broadest 3.5-4 mm.
wide; corolla lavender-pink, the tube often sparingly glandular-
punctulate above; pappus soft and fine, tawny, the bristles slightly
clavellate-thickened above, mostly biseriate, often triseriate on the
wider portions of the scarcely enlarged funnelform epigynous disc;
mature achenes oblong-cylindric, 7-8 mm. long, costulate, appressed-
hirsutulous; receptacle glabrate.

MEXICO: Moretos: woodlands above Cuernavaca, elev. 2140 m., Pringle,
no. 6196 (TyPE, Gr. and isotype, N. Y.); pine woodlands above Cuernavaca,
elev. 1980 m., Pringle, no. 9947 (Gr., N. Y.).

A very handsome and apparently local species.

DovuBTFUL OR UNCERTAIN SPECIES.

AcoURTIA ForMosA DC. Prod. vii. (1838) 66, not of D. Don. Gray
‘roc, Amer. Acad. xix. 61) refers this, with some misgiving, to P.
rgida (DC.) Gray. A bit of material of De Candolle’s Acourtia for-
mosa, unfortunately, does not settle the question today any more than
it did when Gray had the same fragment before him. The broad
Phyllaries, which are a trifle tawny-arachnoid on the margins, are
*r too acute and even shortly acuminate for P. rigida, and the in-
a are a little too broad. It may represent a phase of P. platy-

Domermia Humsotpti Less. Linnaea v. (1830) 13; Perezva Hum-
Polat Gray, PI. Wright. i. (1852) 128. Lessing described this species
from a specimen of Humboldt’s in Willdenow’s herbarium. It is only
doub y one of the Mexican Perezias. Among the characters of
Lessing’ 8 redefined genus (De Candolle, Ann. Mus. Par. xix. 64, 1812,
taken up Lagasca’s name in erecting the earlier Dumerilia, which
a phase of J ungia) are the following: achenes beaked, papillose, the
“Pigynous disc large; pappus 1-seriate, paleaceous, the paleae linear,
“errulate. No member of the genus Perezia has beaked achenes, nor

isa

74 BACIGALUPI

is the pappus ever paleaceous. Were it not for these objections, the
specific diagnosis would seem to fit the unbranched P. Alamamni var.
adnata reasonably well, the mention of adnate leaf-bases (foliis “semi-
amplexicaulibus profunde cordatis, auriculis apice parum decurrenti-
bus et adnatis basi liberis”) being especially striking. Another possi-
bility is that Lessing may have had an aberrant form of P. Thurberi
with adnate leaf-bases similar to that of the Dugés specimen quoted
under this species from Guanaxuato. The phyllaries, described as
glabrate, from ovate to obovate or oblong-obovate, would seem, how-
ever, to exclude this interpretation.

Perpicrum corpatum La Llave & Lexarza, Nov. Veg. Deser. i.
(1824) 27. Herb with a strict, herbaceous, simple, terete, glabrous,
dark purple stem 3 feet high; leaves cordate-oblong, alternate, sub-
imbricate, the lower oval, the upper acute, membranaceous-coriaceous,
very minutely dentate, glabrous on both surfaces, not veiny save for
the thick, dark purple midrib; inflorescence corymbose; flowers labiate,
purple, odorless. “Habitat in montibus del Desierto Mexico viciniis,
floretque novembri.”

The very inadequate nature of the diagnosis gives no clue whatever
as to the probable identity of this species. De Candolle (Prod. vii. 66)
doubtfully referred it to his Acourtia hebeclada, and Ramirez [Dat.
Mat. Med. Mex. i. (1894) 65] comes likewise, though still question-
ingly, to the same conclusion, but it would seem that if the plant here
described were that very glandular species, some mention of its viscid
nature would have forced itself into the diagnosis. Moreover, 4
pointed out under Perezia moschata, the description of this last species
seems to fit P. hebeclada much better than does that of Perdicium cor
datum.

Perezia capitata Wats. Proc. Amer. Acad. xxv. (1890) 156 = GocH-
NATIA GLOMERIFLORA Gray, Proc. Amer. Acad. xix. (1883) 57, as
pointed out by Blake in Standley’s Trees and Shrubs of Mexico,
Contrib. U. S. Nat. Herb. xxiii. (1926) 1636.

Perezia foliosa M. E. Jones, Contrib. West. Bot. no. 15. (1929) 154,
not Rusby, 1896 = Perezia LONGIFOLIA Blake, Journ. Wash. Acad.
Sci. xviii. (1928) 36. :

Prrezia FRuTICOsA La Llave & Lexarza, Nov. Veg. Descr. }- (1824)
26. Described from the mountains about Morelia (“Vallisoletum a
Valledolid, the old name for Morelia) as an alpine shrub, about thé
height of a man; stem branching, striate, becoming purplish; ¢@
leaves alternate, amplexicaul, suborbicular, coriaceous,
nitidous, obtuse, serrate, the teeth spinose; floral leaves acute,

~~ ee

MONOGRAPH OF THE GENUS PEREZIA 75

heads (“Flores”?) corymbose, very crowded; involucre cylindric,
many-flowered, imbricated; phyllaries somewhat reddish, ovate, the
margins woolly; marginal flowers radiate, pink, equal; inner flowers
small, bilabiate, trifid.

Ina very general way, this scant description seems to fit P. Alamani
var. oolepis better than any other described Mexican Perezia, though
the mention of two types of corolla, an outer radiate series and inner
smaller ones, offers an obstacle too great to overlook. Is it not possible
that these authors were deceived by the centripetal order of flowering
which is the rule in the family Compositae?

An additional source of doubt is the fact that the authors of P.
fruticosa made no mention of the shortly adnate leaf-bases so striking
in the varieties of P. Alamani.

Perezia MoscHata La Llave & Lexarza, Nov. Veg. Descr. i.
(1824) 26. Described, likewise from the mountains about Morelia,
as a plant with a simple, erect stem 3-4 feet high, the upper leaves
and flowers forming a very long thyrse; leaves alternate, sessile, sub-
mbricate, ovate, ample, “scarioso-membranacea,” reticulate, smooth,
serrate, the teeth and apex acute, the upper gradually smaller; thyrse
oblong, a foot or more lon , the pedunculate heads congested into
corymbs; corymbs axillary, many-headed, bracteate; involucre oblong,
mmbricate, the phyllaries not scarious, linear, acute, reflexed at the
apex; outer radiate corollas violet, 18 in number; inner corollas bi-
labiate, similar to those of the other Perezias. ‘The whole plant exudes
4 strong odor of musk, whence the name.

Here agian we find two kinds of flowers described. But for this
fact (and it may be that the authors did not take into account the
centripetal order of flowering!), this description might plausibly fit P.

clada, a very glandular species having, not infrequently, a very
large thyrse. It is not improbable that a species as glandular as P.

distantly related species, P. orylepis, Schaffner, on the plant label of
his no. 377, wrote “Eine. . . im frischen Zustande. . . etwas
Klebrig - . . sehr aromatische Pflanze.” Yet the specimen has re-
tained none of the aromatic odor which evidently was so striking when

© plant was freshly gathered. As to the phyllaries with reflexed
*Pices, those of P. hebeclada, especially the often linear outer ones,
become recurved-spreading, and in some cases the slender tips have
become decidedly reflexed. At all events, the contention here is
merely that the description would seem to fit P. hebeclada more closely
than any other Mexican Perezia.

76 BACIGALUPI

Perezia paniculata Gray, Proc. Amer. Acad. xxi. (1886) 393, based
on Palmer, no. 279 (1885) = VERNONIA SERRATULOIDES H. B. K. Nov.
Gen. et Spec. iv. (1820) 33.

PEREZIA TURBINATA La Llave & Lexarza, Nov. Veg. Descr. i.
(1824) 25. The original diagnosis calls for acute ovate leaves and short
peduncles. Although Dr. Gray, Proc. Amer. Acad. xix. (1883) 58,
described the leaves as “ovatis oblongisve,” none of the leaves in the
two sheets of Schaffner’s no. 226, the number cited by Gray, can be
said to be ovate, being in fact mostly oblong, a few even somewhat
narrowed toward the base, while the reduced upper leaves are lance-
attenuate. Most of the cauline leaves, moreover, are decidedly obtuse
to rounded at the apex. Of the 9 peduncles present, one is as short as
1 em., the rest measuring 23 mm., 32 mm., mm., 50 mm., two 55
mm., and two more 65 mm. in length, all exceptionally long for this
group. More fundamental still is the fact that it is hard to reconcile
a head with radiate outer corollas (corollulae”) with those of any of
the known members of the § Acourtia. Dr. Gray raised this very ob-
jection to P. fruticosa La Llave & Lex. (Proc. Amer. Acad. xix. 61);
yet, in the case of P. turbinata, these very corolla-characters seem t
meet with no objection. ‘

In spite of the uncertainty still attending the identity of this species,
it has seemed best, inasmuch as a more positive interpretation of the
species seems improbable, to follow that of Dr. Gray.

Perezia vanillosma (Wright) Molt. & Gomez, Anal. de la Soc. de
Hist. Nat. Madrid xix. (1890) 268 = Proustia vantLLosma Wright
in Sauv. Fl. Cub. (1873) 83. In transferring this Cuban Proustia to
Perezia, the authors followed Baillon [Hist. des Pl. viii. (1882) 101]
in reducing Proustia as a section under Perezia. Gomez makes this
species the type of a new sub-section (Vanillosma G-M) under §
Proustia. :

Although, as Hoffmann (Engl. & Prantl, Natiir. Pflanzenfam. 1°
Abt. 5, 343) indicates, Proustia vanillosma differs from the other
species of Proustia in having a smaller portion of the st Je-branch
covered with papillae and in its almost truncate stigmas, it is hardl y
to be referred to Perezia. Habitally at least, it is quite distinct, bemé
a subscandent, very branching shrubby plant with pairs of short, not
pungent, slightly to strongly recurved, spine-like processes si
each of the nodes and elliptic-oblong, shortly petiolate leaves si
em. long. The specimen at hand (Wright, no. 3616, the isotype) ®
the Gray Herbarium is in the fruiting stage; hence an examination
and comparison of the stylar characters was not feasible.

MONOGRAPH OF THE GENUS PEREZIA 1

Perezia vernonioides Gray, Proc. Amer. Acad. xxii. (1887) 433,
based on Palmer, no. 745 (1886) = VERNONIA JALISCANA Gleason,
Bull. N. Y. Bot. Gard. iv. (1906) 198. According to the international
rules of nomenclature, however, Gray’s earlier specific name must be
takenup. The binomial Vernonia vernonioides (Gray), comb. nov.,
is therefore proposed.

OUSTIA MEXICANA Lag. ex D. Don, Trans. Linn. Soc. xvi. (1830)
201. Erect branches sulcate, glandular; leaves amplexicaul, cordate-
ovate or oblong, acuminate, sharply and closely spinulose-denticulate,
occasionally subserrate, membranaceous, reticulate-veiny, rough-
papillose on both surfaces, green, the lower surfaces, like the branches,
covered with glistening, resinous glands, an inch or two long, one inch
wide; heads fasciculate-corymbose; ped lar bracts ovate-lanceolate,
acuminate, somewhat recurved, minutely glandular; involucre tubu-
lar-oblong, many-bracted, copiously glandular; phyllaries ovate-
lanceolate, appressed, acuminate; flowers 5, white; achenes teretish,
papillose-scabrous; pappus white, the scabrous bristles united at the
very base, subsimple at the apex.

This description, as Dr. Gray himself intimated (Proc. Amer. Acad.
xix. 59) “answers not badly to P. Thurbert of which the outermost
bracts are ‘ovato-lanceolatis acuminatis.’ ” Inasmuch as there is
nothing in the diagnosis definitely to exclude P. Thurberi and much
that agrees with that species, it is quite possible that the plant Don had
before him was this Grayian species.

Lessing (Syn. Comp. 408), and De Candolle (Prodr. vii. 67) who
followed him, must both be incorrect in referring Proustia mexicana
of Lagasca to Lessing’s Dumerilia Humboldtii, since it is quite evident,
from a perusal of the two diagnoses, that Lessing described quite a
different plant.

LatiFor1a Hook. & Arn. Bot. Beech. (1840) 300. In the
‘upplement to the “Botany of Capt. Beechey’s Voyage,” on P- 437,
€ authors of this binomial referred it to Acourtia formosa D. Don
and of De Candolle. But these are most certainly not the same.
Gray, Pl, Wright. i. (1852) 127, referred Trizxis latifolia to P. rigida
(DC.) Gray; later (Proc. Amer. Acad. xix. 61, 1883) he again referred
cer - rigida, though here not without reservation. In the original
45 is of Trizis latifolia, the leaves are described as obovate,
th Mches long and 3 inches wide, the cymules as laxly corymbose and
® phyllaries as lanceolate and “ spinoso-acuminatis.” The peduncles
we squamose, the acuminate subspinulose scales “ gradually passing
‘nto the spinescent scales of the involucre.” This combination o

78 BACIGALUPI

characters definitely excludes Trixis latifolia from either P. formosa
or P. rigida, nor does it give any clue to its probable identity, inasmuch
as none of the Perezias so far described from Mexico has spinescent
phyllaries.

§ Euperezia Gray, Syn. Fl. N. Am. ed. 1, i. pt. 2 (1884) 408.

Mostly acaulescent often cespitose single-headed annual or perennial
herbs with the pinnately sinuate to parted often spinose-ciliate leaves
radical, or with the leaves restricted to the base of the stem, or (much
more rarely) taller and truly caulescent, branching and several-headed
above with the cauline leaves alternate, denticulate to dentate, often
spinose-ciliate, not infrequently pinnatifid to pinnately parted; herb-
age glabrous to glandular-puberulent, smooth to scabrid; phyllaries
for the most part not indurate, lax to rigid, the margins of the inner
very often scarious, often widely so, the outer sometimes pungently
ciliate-denticulate to -dentate, rarely pinnatisect, usually well im-
bricated; corollas mostly blue or white, varying to purple, rose and,
in 2 species, yellow, the marginal corollas often with the ligulate “ lip”
somewhat broader and longer than that of the inner ones, the heads
thus appearing radiate; otherwise as in the generic diagnosis.—Perezva
Lagasca, Amen. Nat. i. (1811) 31; Chaetanthera Humb. & Bonpl. PI.
Aequin. ii. (1809) 146, 168 and 170, not Ruiz & Pav.; Clarionea Lag.
ex DC. Ann. Mus. Paris, xix. (1812) 65; Homoianthus Bonpl. ex DC.
loc. cit.; Homanthis H. B. K. Nov. Gen. et Sp. iv. (1820) 12; Droza
Cass. Dict. Sci. Nat. xxxiv. (1825) 217; Homoeanthus Spreng. Syst. 1.
(1826) 503; Clariona Spreng. loc. cit. 504; Clarionia D. Don, Trans.
Linn. Soc. xvi. (1830) 204; Clarionia § Palesia D. Don, loc. cit. 207,
in part; Clarionella DC. ex Steud. Nom. i. (ed. 2, 1840) 377; Clarionema
Phil. Linnaea xxviii. (1856) 717.

ProvistonaL Key To THE CoMMONER SPECIES.
a. Inflorescences paniculate-branching, at least in mature plants,
and often os plants typically caulescent... ..
6. Phyllaries with scarious margins, neither linear-lanceolate

saliently spin
to oblong-lanceolate.............0++000+: ie
e. Heads not congested, disposed in an elongate inflor-

| MONOGRAPH OF THE GENUS PEREZIA 79

escence; cauline leaves not saliently spiny; phyl-
laries widely oblon;
af Inflorescence a diffusely branched, usuall
ed paniculate cyme, the heads aggregated

ortly SS RE ices Peis . cubataensis.
f. Inflorescence narro Ue edi aded; basal leaves"
sige s, especially above, s subcoriaceous us, close-
ares the teeth attenuated into

aif in iary PEOCARHEB. . 5. Fi ed 4. P, squarrosa.
d. a 7 iggy in naseas rene few (2-5)-headed, not

ese
g. Heads at anthesis narrowly campanulate, 2-2.5 ¢
igh; phyllaries oblong-lanceolate, at least the
inner entire; oe ichomes (when present) not
forked at the s i

h, Phyllaries dino rphic, the outer calyculate and
foliaceous, like gee cauline leaves Sosy Bini

wide; corolla 22 mm. long............ . P. carduncelloides.
h. Phyllaries not dim able the outer at lh but
slightly more faliancous than the inner; plants
at least in Pep viscid-pubescent; ligules not
arcuate. ...
i. Inner hvtlaine oblong-lanceolate or narrower,
their central portions not coriaceous, light

green, often ptly nuate-mucronate,
the scarious margin not very broad; cauli
leaves not crowded; wide;
Corolla 29 mini. lone oss io et se See ss . pungens.
t. Inner phyllaries es mths their central
portions coriaceous, dark green, with very
broad scarious margins, Gertnating in a
_ stiff callous fo de
j. Flowering stem exceeding the basal leaves,
co osely branched at he summit in

the short floweri — eigen: ‘eee ria
ant son es 5 mm.
lon

wide; corolla 20 mm. long.......... 7. P. multicapitata.
j. Flowering s e or nes at all exceeding
h leaves; plant not rigid. Here may

g. Heads at anthesis broadly campanulate, 8-15 mm.
igh porliats ries obovate-oblanceolate, more or less

bri ht mike forked at the very summit...... P. sonchifolia.
b. yy Sooterdles without scariou tas Sea tony aioe — east the inner
far-lanceolate, fr cnciaree cauline leaves cnet orm
a, Inflo PYOCONCE CHTUBRS i ob oe ee vee P. Ga a perfoliata).
ke Aqocence reduced to a single head ao stem. ...
Achene-trichomes with a dark brown waxy ng lg rollas
k yellow; leaves entire, the basal Hines oblanceoiate: “tt P. lactucoides.
Athene-iriehom es (when present) without a waxy
Tollas mostly y ene to blue, more rarely violet or enk
(vello ow in P, 1

80

BACIGALUPI

l. sfueig gin exceeding the basal leaves—when by as little
or even — shorter (P. pilifera), then the

plants vdeprease
scape-like, not infre quently bearing bract-like leaves,

cas
2
ar}
mn
Dm
_
is]
3
a
5
B

m. iaab inten seither pinnatifid nor pinnately parted,
wb most ee cenate; achenes shortly stipitate-
n. "Havel Taatlaces narrowly oblanceolate to some-
jab spatulate, ee ak aes Boigaaeaty nor coria-
eous; sca. em. tall, wiry....
0. Baant leaves cer obiancedats Pr cite at
ends, the larger repand-dentate and 8-15
not more than 8 mm. wide; achenes
pen ti except for the glands; plant’ glabrous
PACE ac s8stls es pi yw alte vase cre wien MaRS 11.
o. Basal leaves oblong-lanceolate to cuneate-obovate,
ciliolate and som etimes weakly dentate, 3-6 cm.
ong. mm. wide; achenes sparsely tawny

— as well as glandular; scape glandular,

DOREY GOCE i OS ie

n. Leah hades linear, entire, rigid-coriaceous, mo

less hs ig aga Soars 2-7 (-9) em. tall a
p. Leaves Joniicnweay beset with stiff callous pro-
cesses 1.5-2 mm. long, apparently on the margin
“a Am oe from the surface of the stron: ly revo-
e; achenes glabrous (acc. to ion)

q- rridy paras if at all revolute, bristly
n the actua in; s glabrous,
particularly the inner Sositains achenes
rsely canescent-strigose, pata above;

22— gel i rere a oc
ately parted, oblanceo-
late to linear in vatthne : sahenes sericeous to str rigose,
usually so densely as ba mask the short-stipitate

gla: when present... .

r. Achenes whitishstrigose heads narrowly pigs
late, 1-1.5 em. wide; peduncles 1-3.5 cm. long,
mostly little ¢ to rarely not at all exceeding the
dense rigi ee basal leaves... .

‘ ;

not ciliate, the p

bristle; peduncles very slide ©

at all exceeding the basal ee corollas mh

* ot
ual
oa
a
of
see
:

8. Leaves cuneate-oblanceolate in outline, pinna poe
, the more or less rounded-obtuse lobes

P. Mandoni.

He laurifolia.

P. recurvata.

15. P. linearis.

. P, pilifers-

MONOGRAPH OF THE GENUJ PEREZIA 81

fresh) ig VOUOW csc sec ere ewes "U7. P. virens.

u. Corollas white; sca ag et Ht more than twice
as long as the basal | leaves; phyllaries ieee
18

t. Basal leaves not ani er fsx; sep with a
strong tendency to Heong pectinately parted
P. pedicularidifolia.
l. Peduncles not exceeding the basal leaves, often equalling
them—if slightly lon nee a mene then the leaves more

)
v. Peduncles about eq cuttin os basal leaves (not infre-
» aide longer or ‘dor by 1 cm.), bearing foliage

w. Heads mo stly 2.5-3 cm. high; plants 1.5-2.5 dm. tall;
uter ae always spinulose-ciliate. ...

eo] a 1
Soe @

attenuate: not at all c riaceous;
achenes slenderly cylindric, conspicuously gland-

ular, rather sparsely strigose with light brown :
tricho ee pk kes 4 oe a 22. 2 atacamensis.

-acaulescent or but su song plants
with relatively. rom heads; plants 0 of high “ailtpies

BS
=]

. coerulescens.

spin neg C iS Nak s bo Vee OG MES POOR Shee
z. Leaves narrowly spatulate-obovate, not dissected,
at ty precio, often un
2a. FP; ee (P. pygmaea).
y. Phyllaries glandular-puberulent or = nom

punctate; Sere glandular-pu
6. pee nceolate, palais lobed ‘ahees am gs

Oe 6 Ce eee el es ew eck & + OS ee ole es

Contrip. GRAY Hers. XCVII. PuateE I

¢. G PRINGLB.
PLANTA MEXICAN
S06
stat ¢? caxack—
6015 Peresia nudicaulis, Gray.

Dey catmarecien bite, tan Soden, 4.600 ft

PeRrezra SCAPIFORMIS Bacigalupi.

Contris. Gray Hers. XCVII. Piate II.

PeREZIA LOBULATA Bacigalupi.

Contris. Gray Hers. XCVII. Puate III.

PLANT MEXICANA PURPUSIANA
SAN LUis POTOSE
Skat 08 aN OP

Pacis toll

PEREZIA DissiTIcEePs Bacigalupi.

Contris. Gray Hers. XCVII.

Puate IV.

[Selon

Ria abated oe Sl

hel ass”

»

,
Meth met Be

Perezia ASPERA Bacigalupi.

Contris. Gray Hers. XCVIT. PLaTE V.

DETERM RE GRAY WERBARIUM

r
ae 2 pas) arn
Re

a) pan folie Breage! La

Prrezia rtgipa (DC.) Gray var. LINEARIFOLIA Bacigalupi.

Contris. Gray Hers. XCVII. Puate VI,

, ems
SPLATT TEE MRT

retorted

i

Hertel

min
Gare “a “ bere t

C. G. PRINGLE,
PLANTA, MEXICANA.
1902.

STATE OF JaLreco
9950 Perezia rigiday Gray.

hy We af 4 hartanen nese Buatatabenn, LO00 fe

al

Perezia rigipa (DC.) Gray var. acuminata Bacigalupi.

Plants of Oaxaca.
c “OMLBOTED BY Rev. Lucivs © Sarr.

ContTris. Gray Hers. XCVII. Puate VII.

te RP a Poi

Prerezta Duaesu Gray var. pILULOSA Bacigalupi.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

XCVII.

ISSUED APR 14 1932

STUDIES IN THE BROMELIACEAE.— III.

1. Notes preliminary to a Revision of the Bromeliaceac.... .
2. Provisional key to the Genus Guzmania with notes on new or
mriCal BDOGIES. | 6 ce

By Lyman B. Smiru.

PUBLISHED B :
THE GRAY HERBARIUM OF HARVARD t UNIVERSITY
CAMBRIDGE, MASS., U. S.

oases Sra me gia eed iat asec wens ce eae aS Nie nm

STUDIES IN THE BROMELIACEAE.—III.
1, NoTES PRELIMINARY TO A REVISION OF THE BROMELIACEAE.

THE largest sources of material for the present paper have been
the collections of the Gray Herbarium (G), the Field Museum of
Natural History (FM), and the United States National Museum (US).
These have been studied intensively in the preparation of a treatment
of the Bromeliaceae for the projected flora of Peru undertaken by the
botanists of the Field Museum. Constant use has been made of the
photographs of types in the Berlin Herbarium (B) taken by Mr. J.
Francis Macbride under the Rockefeller Foundation Fund. Since
nearly all the recent types based on the specimens of Weberbauer,
Ule, and Herrera are at Berlin, these photographs have proved ex-
tremely valuable in establishing the relationships of the various
species and in constructing keys. It is only justice to say that the
quality of the photographs themselves leaves little to be desired.

All novelties and new combinations from the Peruvian material are
treated in the present paper, since it is considered advisable to publish
them before their inclusion in the flora. At the same time it is pos-
sible to discuss the species in more detail than could be done in a
strictly floristic work. In making certain new combinations it has
been necessary to take the liberty of using manuscript or herbarium
names of Mez, but in each instance their origin is duly acknowledged.
_ Ina study of the complete ranges of the Peruvian species of Bromel-
‘aceae, they have been found to illustrate certain geographic tendencies
of the family very well. The most striking single feature is their
‘pparently high degree of endemism, well over half the Peruvian
species not being known outside the country. This degree of endemism
is partially discounted by the large area involved in Peru and by the
incompleteness of collecting in tropical America generally, but on the

ole it is not unreasonably high for such a generally mountainous
country. Of the large Peruvian genera, Puya is the most strongly
endemic and Pitcairnia next, while Tillandsia shows by far the strong-
est extra limital character. The Peruvian endemic species of these
three genera follow the same tendencies, those of Puya being largely
restricted to a single locality each, those of Pitcairnia sometimes from
Several localities each, and those of Tillandsia often having a wide
‘ange within the country. These genera show practically the same
relation to each other when viewed as a whole, Puya having the
smallest range as a genus and only one widely distributed species, P.

floccosa Morr., Piteairnia having a larger range and several widely

4 SMITH

distributed species, and Tillandsia having the largest range of all, or
practically the same as the whole family, and a large number of wide-
ranging species.

The remaining Peruvian genera are too poorly represented to be
significant in considering endemism, but taken by tribes they illustrate
extra limital affinities well. Those with the superior or partly su-
perior ovary, Pitcairnieae and Tillandsieae are strongly Andean in
affinities, while those with the completely inferior and indehiscent
ovary, Bromelieae, are derived almost exclusively from Brazil and in
Peru are limited for the most part to the northeastern lowland.

The following table summarizes briefly the geographic affinities of
the Peruvian Bromeliaceae:

Extra limital distribution of
Peruvian species:

5

E es

oN HO ee "ay

moe fee a Soe 8883

Sag SEPese Ss Passe Se

° S Re me So oO 0 OC 8 mb

eH MOD aO tm OF oO 40
Ae POR San 32 30 - ---- 206 3 ee
oo Lanne i. of 8 DS 1 i Ae ie ee ee
o Devtetehmin ct 0) ance el hee am
4, Pitesirain:;: 45, 20°20) Le eH 2 2? See
5. Tillandsia .. 68 34 883531416147 5 76 43
CVs 6. deo oie ee
7. Guzmania DA AG os parle 2 2D tee
8. Catopsis fo Od ln = hee
9. Bromelia......, 8.9 2h ol eles
10. Greigia......... Pils eh ee
1]. Aregelia......., fo Puke went ote
12. Streptocalyx 04038 6 i a le ee TT
13. Aechmea... .... 2 ee eee koe
MADRE eS BS Oo de ee co ee
1 DiS eR a a a a

ioe
Total Bromeliaceae. .175 113 2110 3 5 4 18 21 218 10 129 68

Thanks to a generous loan of material from the Riks a
Stockholm (S) and from the Biological Institute of Sao Paulo, }

STUDIES IN THE BROMELIACEAE.—III 5

possible to publish several new species of Brazilian Bromeliaceae.
There are also scattering notes on other collections. For the sake of
convenience, genera and species have been arranged in alphabetical
order. In the case of Guzmania the subject has grown so in preparation
that it seems best to give it separate treatment.

Aechmea ferruginea, spec. nov., terrestris, acaulis: foliis 6-7 dm.
longis, subtus punctulato-lepidotis; vagina magna, atro-brunnea;
lamina lineari, 3 cm. lata, spinuloso-serrata, spinis 1-2 mm. longis:
scapo gracili, erecto, foliis multo brevioribus; scapi bracteis lanceolatis,
acuminatis, membranaceis, internodia subaequantibus: inflorescentia
foliis breviori, tripinnatim paniculata, pyramidali, 17 cm. longa,
stellato-ferrugineo-lepidota, superne densa, inferne laxa; bracteis
primariis eis scapi similibus, infimis exceptis, quam rami axillares
brevioribus: ramis laxe florigeris, suberectis vel patentibus; bracteis
florigeris minutis, aciculosis: floribus gracile pedicellatis, 2 cm. longis,
pedicellis ad 8 mm. longis; sepalis 4 mm. longis, inermibus, rubris,
valde asymmetricis, latere dextro in alam membranaceam quam pars
centralis majorem producto; petalis albis, ad apicem rubenti-caeruleis,
13 mm. longis, basi ligulis binis auctis; placentis loculis apice affixis,
ovulis obtusis. Pl. I, figs. 1-3.

PERU: Junin: in dense forest, Pichis Trail, Dos de Mayo, alt. 1700-1900
m., 1929, Killip & Smith 25815 (US, rrpx; phot. G).

This species belongs to the small subgenus Podaechmea, and presents
the second instance of the extension of the group down into South
America. Its classification is extremely simple since it differs from the
other four species of the subgenus in having unarmed sepals. Its
scales as seen under the hand-lens have a very striking form, which is
more like that in Pitcairnia ferruginea R. & P. than in any other

Tomeliad known to me. ;

Aechmea recurvata (Klotzsch), comb. nov. Macrochordium re-
cureatum Klotach, in Allg. Gartenz. xxiv. 393 (1856). Hohenbergva
legrelliana Bak. in Saund. Ref. Bot. iv. t. 285 (1871). Ortgiesia
palleolata Morr. Cat. 2 (1871). 0. tillandsioides var. subexserta Regel
in Gartenfl. xxiv. 188 (1875). Aechmea Legrelliana Bak. in Journ. Bot.
%vl. 236 (1879). Portea Legrelliana Benth. & Hook. Gen. in. 662
(1883). Ortgiesia Legrelliana Bak. Brom. 19 (1889).

Aregelia eleutheropetala (Ule) Mez in herb. Berol., comb. nov.

tularium eleutheropetalum Ule, Verhandl. Bot. Ver. Brandenb.
Xlviii. 131 (1907). According to Ule this species can not be included
Many of the genera as given by Mez in 1896, unless Aregelia, Euni-
dularium and Canistrum are all merged under Nidularvwm. Un-

6 SMITH

fortunately in his discussion Ule does not indicate directly why the
species differs in character from each of the three genera, but gives
its points of similarity so that the distinctions must be drawn by
inference. In comparing Ule’s description with Mez’s key it is at
once evident that the species is excluded from Canistrum on the basis
of naked petals. Examination of fragments of the isotype material
from the Goeldi Museum in Para confirm this point (t. I, figs. 4-6).
Apparently Ule excludes the species from Nidularium in the sense of

ez because of its free petals, while slighting a better distinction with
the words “floribus . . . . subpedicellatis.” To be sure the out-
lines of ovary and pedicel merge without abrupt distinction, but when
the pedicel is as long as the ovary the flower may be considered quite
thoroughly pedicellate as opposed to the sessile flowers of Nidularium
(t. I, fig. 5).

Though admitting that the species resembles Aregelia in habit Ule
seems to exclude it on the grounds of free petals and compound in-
florescence. Checking Mez’s generic description for Aregelia shows
that the character of joined petals is usual but not absolute. The
character of compound inflorescence, however, can not be reconciled
with Aregelia. An examination of photographs of the type and isotype
material shows no indication of a compound inflorescence, but both
appear as in typical Aregelia. Consequently it seems that Ule must
have made some miscalculation in his description. This impression 1s
strengthened by the fact that he gives the inflorescence as 300-500-
flowered when it can not be much more than a tenth of that figure.

In Mez’s key, Aregelia eleutheropetala comes nearest to A. ¢
centrica from which it is distinguished by having the floral bracts
straight and the sepals acute rather than acuminate.

Dyckia Dusenii, spec. nov., valida, ad 7 dm. alta: foliis 3-4 dm.

STUDIES IN THE BROMELIACEAE.—III 7

ultra tubum petaleo-stamineum inter sese connatis; antheris sub-
linearibus, acuminatis. PI. IT.
RAZIL: Parand: Porto Amazonas, alt. 800 m., 1916, Dusén 18081 (S,
tyPE; phot. G); Tamandaré, alt. 870 m., 1914, G. Jénsson 1029a (S); Serrinha,
1909, Dusén 8686, 8996 (S).
The outstanding characters of this species are its simple inflor-
escence, large floral bracts, and wing-carinate sepals.
Greigia columbiana, spec. nov., e fragmentis solum cognita:
foliis ultra 4 dm. longis; vagina ignota; lamina peranguste lanceolata,
ad 3 em. lata, glabra, toto margine spinosa, spinis validis, atro-casta-
neis, 4-5 mm. longis: inflorescentia subpauciflora, densa; bracteis ex-
terioribus basi late ovatis, atro-castaneis, dissite lepidotis, integris,
late hyalino-marginatis, apici anguste triangularibus, pallidis, grosse
denseque serratis: bracteis florigeris lineari-lanceolatis, pungenti-
acuminatis, ad 4 em. longis, sepalis brevioribus, integris, mucroni
terminali dorso inserto, parti inferiori straminea, membranacea, parti
superiori atro-castanea, subcoriacea, dissite lepidota: floribus 5-6 cm.
longis; sepalis bracteis florigeris similibus; corolla sepalis 10-15 mm.
longiore. PI. I, figs. 7-8.
COLOMBIA: Norte pr SANTANDER: Paramo de Romeral, alt. 3800-4200
m., 1927, Killip & Smith 18689 (G, TYPE).
Greigia Macbrideana, spec. nov., e fragmentis solum cognita, ad
. alta: foliis 3-7 dm. longis, subtus adpresse albo-lepidotis;
spinis 3-4 mm.

lepidotis: bracteis florigeris, lineari-lanceolatis, 35 mm. longis, sepala
Superantibus, integris, pungentibus, punctato-lepidotis, parti inferiori
8 minea, membranacea, parti superiori atro-castanea, subcoriacea :
floribus 4 cm. longis; sepalis bracteis florigeris similibus, 2 cm. longis;
corolla 27 mm. longa, rubenti, tubo lobis subaequilongo, lobis erectis;
Staminibus styloque inclusis. Pl. I, figs. 9-11.

iat : Huanuco: wet places, Tambo de Vaca, alt. ca. 3700 m., 1923,

€ 4442 (FM, rypx; phot. G).

; The two above-mentioned species of Greigia can not be interpolated
n the key given by Mez,' since neither the material itself nor the notes
*ecompanying it indicate whether the inflorescence is lateral or

a (igomeliacene Mez, in Monographiae Phanerogamarum, De Candolle ix.

8 SMITH

terminal. The following key serves to show their relationships as
nearly as possible without the complete plant:

1. Outer bracts of the inflorescence entire.
2. Sepals 25 mm. long, mucro dorsal. Chile............ 1. G. Landbeckii.
2. Sepals 15 mm. long, mucro terminal. Chile............2 }
1. Outer bracts of the inflorescense definitely serrulate.
3. mm. long, ecarinate, white. Chile, Peru..... 3. G. sphacelata.
3. Sepals scarcely over mm. long, often carinate, green or
more or less castaneous.
4, -blades finely and remotely serrulate.
5. Bracts wholly or in large part dark-castaneous.
6. wholly castaneous, with a few short teeth at
apex. re aes ce ek en 4. G. Sodiroana.
6. Bracts partly castaneous and partly stramineous, wit
at least the upper half coarsely and closely serrate.
7. Bracts and sepals stramineous at apex. Vene-
pei eee wae Uiviie core ey Geom 5. G. albo-rosea.
7. Bracts and sepals stramineous at base. Peru.
6. G@. Macbrideana.
5. Bracts and sepals green. Costa Rica.............. 7. G. sylvicola.
4. Leaf-blades coarsely and rather closely serrate throughout.

apex. :
ae ec | cs Le 9. G. columbiana.

qj AMAICA: without further locality, N. L. Britton 2313 (NY, TYPH; phot.
The narrowly cylindric long-stipitate spikes and strongly nerved
short floral bracts distinguish this species of Hohenbergia from all
others. ;
Pitcairnia Calderonii, Standley & Smith, spec. nov., acaulis,
semimetralis, rhizomate foliorum delapsorum vaginis obtecto bulbum
efformante: foliis dimorphis, alteris persistentibus, fere setiformibus,
brunneis, uncinato-serratis, alteris foliaceis, tempore sicco deciduis;
lamina anguste lanceolata, ad 3 dm. longa, 16 mm. lata, integra,
supra glabra, subtus floccose albo-lepidota: scapo valido, erect,

STUDIES IN THE BROMELIACEAE.—IlI 9

floecose lepidoto; scapi bracteis foliaceis, internodia multo superanti-
bus: inflorescentia simplicissima, floccose lepidota, 2 dm. longa, apici
densa, inferne laxa: bracteis florigeris lanceolatis, acuminatis, magnis,
infimis sepala superantibus: floribus erectis, ad 6 em. longis, brevissime
pedicellatis; sepalis anguste lanceolatis, acuminatis, 28 mm. longis,
valde carinatis, dorso omnino albido-lepidotis; petalis rubris, eligu-
latis, stamina subaequantibus; antheris ad 8 mm. longis; ovario ad
% longitudinis supero. PI. III, figs. 1-2.

SALVADOR: Sonsonats: Volcano of Izalco, 1931, Salvador Caldéron 2595
(G, TypE).

This species is closely related to P. splendens and P. Karwinskyana.
It differs from the former in its smaller sepals and narrow rather lax
inflorescence, and from the latter in its very short flower-pedicels.

Pitcairnia cuzcoensis, spec. nov., ca. 9 dm. alta: foliis dimorphis,
exterioribus ad vaginas atro-brunneas squamiformes pungenti-mu-
ronatas reductis, interioribus viridibus, super vaginam angustatis
sed non distincte petiolatis, basi excepta integris, lineari-lanceolatis,
5-6 dm. longis, 25 mm. latis, supra glabris, subtus minute albo-
puberulis: scapo erecto, gracili, dissite furfuraceo; scapi bracteis
anguste triangularibus, acuminatis, superioribus quam internodii
brevioribus: inflorescentia paupere paniculata; axi dissite pallido-
furfuraceo: bracteis florigeris elliptico-lanceolatis, acutis vel acumi- ~
natis, pedicellos subaequantibus: floribus suberectis vel patentibus,

mm. longis, gracilibus, pedicellis gracilibus, 10 mm. longis; sepalis
‘nguste triangularibus, acuminatis, dissite punctulato-lepidotis;
petalis rubro-scarlatinis, nudis; staminibus exsertis; ovario ad 34
longitudinis supero. PI. III, figs. 3-5.

ERU: * mi i ichaca, Marca-
ais Valley, provinge of Guispieanchi alt 1700 my 1920, Weberhuer 7885
(FM, TYPE; a G).

Pitcairnia pectinata, spec. nov., metralis: foliis 1 m. vel ultra

» Spinose petiolatis; lamina lanceolata, integra, 8 cm. lata: scapo

ot.

Periores multo brevioribus: inflorescentia simplicissima, dense sub-
lata, axi bracteisque ferrugineo-lanatis: bracteis florigeris eis
ier al similibus, pectinato-serratis: floribus pallido-rubris (Archer),
og glabris, breviter pedicellatis, solum perjuvenilibus cognitis ;
lis angustissime triangularibus, ad 28 mm. longis; petalis eligulatis;
*vatio ca. 14 supero. PI. III, figs. 6-7.

An LOMBIA: Anrio 1500 m., 1931, W. A.
: quia: La Sierra, Medellin, alt. ca. m., :
1872 (G, rex). -

10 SMITH

This species is unusual for Pzteairnza in having the floral bracts
serrate instead of entire

Pitcairnia subpetiolata Bak. A study of the type material
indicates that Pitcairnia sessiliflora Rusby should be included in this
species. Consequently the range of P. subpetiolata is extended as
follows:

BOLIVIA: La Paz: province of Nor Yungas, Polo-Polo near Coroico, alt.
1100 m., 1912, Buchtien 3675 (NY, phot. CG). InpEeFiniTE: M. Bang (NY,
phot. G; Type of P. sessiliflora).

Pitcairnia truncata, spec. nov., fere metralis: foliis fasciculatis,
dimorphis, alteris setiformibus, margine valde spinosis, badiis, alteris
foliaceis, petiolatis, laminis lineari-lanceolatis, acuminatis, 4-5 dm.
longis, mm. latis, remote minuteque spinuloso-serratis, supra
glabris, subtus dense minuteque puberulis: scapo erecto, dissite fur-
furaceo-lepidoto; scapi bracteis anguste triangularibus, filiformi-
caudatis, submembranaceis, integris: inflorescentia laxe bipinnatimque
paniculata, pyramidata, albo-furfuracea; ramis patentibus, rectis, ad
7 cm. longis (non omnino evolutis), basi sterili brevissimi: bractels
florigeris ellipticis, apiculatis, pedicellis brevioribus et bases suas 1n-
volventibus: floribus gracilibus, ca. 4 cm. longis, plerumque patenti-

us, pedicellis gracilibus, 6 mm. longis; sepalis oblongis, truncatis,

0 mm. longis, carinatis, paullo asymmetricis; petalis fulgide rubris,
ligulatis, stamina subaequantibus; stylo exserto, ovario ultra \y
supero. PI. ITI, figs. 8-9.

PERU: Ayacucuo: wooded hillside, Cearrapa, between Huanta and Rio
Apurimac, alt. 1500 m., 1929, Killip & Smith 22414 (G, TYPz).

The outstanding character of this species is its short truncate
sepals, a very uncommon form in Pitcairnia.

sone) 37 mm. longis, lamina haud distincta, intus nudis; staminibus
styloque inclusis, ovario pyramidali. PI. III, figs. 10-11.

STUDIES IN THE BROMELIACEAE.—III 11

‘PERU: Junin: shrubby rocky canyon side, Cabello, hacienda 9 miles above
Huertas, alt. ca. 2700 m., 1922, Macbride & Featherstone 1339 (FM, Type; G).

Further collections of this species will doubtless demonstrate that
the inflorescence may be simple at times. The inflorescence in the
present specimen might easily be mistaken for the simple type, so
insignificant are the lateral racemes.

Puya gracilis, spec. nov., ca. 8 dm. alta: foliis anguste triangulari-
bus, 6-7 dm. longis, 18 mm. latis, ferrugineo-lepidotis, furfuraceis,
spinis gracilibus, uncinatis, 7 mm. longis: scapo erecto, glabro, 5 mm.
in diametro; scapi bracteis remotis, ovatis, spinuloso-serratis, infimis
longissime caudatis: inflorescentia elongate racemosa, 35 cm. longa,
omnino minute stellato-lepidota: bracteis florigeris membranaceis,
ovato-apiculatis, plerumque quam sepala brevioribus: floribus erectis
vel suberectis, gracilibus, pedicellis gracilibus, 5-7 mm. longis; sepalis
anguste ovatis, acuminatis, 30 mm. longis; petalis tubuloso-erectis,
intus nudis, stamina subaequantibus; stylo exserto. Pl. UI, figs.
2-13.

PERU: without definite locality, Weberbauer 6474 (FM, Type; phot. G).

At first glance this plant scarcely seems to be a member of the genus
Puya so slender are all its proportions, yet the flowers show the
characteristic features of the genus. The individual flowers are more
ike those of P. longisepala of the subgenus Pitcairniopsis than like
those of any other species known to the writer. This resemblance
brings up the point of the phylogenetic value of the divisions which
Mez has given under Puya. His subgenus Eupuya seems to be logical
and to contain closely related species, but his subgenera Piteairniop-
ss and Pourretia, based solely on whether the inflorescence is simple
or compound, appear decidedly artificial. Further collections and
study probably will give rise to a more logical system of classification

don the characters of the flowers and their bracts.

Puya llatensis, spec. nov., ca. 2 m. alta: foliis anguste triangulari-
bus, 1 m. Jon is, utrinque aequaliter pallide adpresso-lepidotis, supra
glabrescentibus, spinis 7-8 mm. longis: scapo basi 15 cm. in diametro
(Macbride & F eatherstone): inflorescentia composita; ramis ad apicem
Usque fertilibus, dense strobiliformibus, ovoideis vel ellipsoideis, quam

€ primariae longioribus: bracteis florigeris late ellipticis, apice
tcuminulatis, integris, dense ferrugineo-lanatis, sepala ad / aequanti-
bus, late convexis, ecarinatis: floribus 7-8 cm. longis, pedicellis crassis,
‘5 mm. longis; sepalis late ellipticis, 30 mm. longis, dense ferrugineo-
lanatis; petalis 65 mm. longis, intus nudis, lamina elliptica; stamini-
bus styloque inclusis, ovario pyramidali. PI. III, figs. 14-16.

12 SMITH

PERU: Hvanuco: steep canyon slope, Llata, alt. ca. 2300 m., 1922, Mac-
dride & Featherstone 2300 (FM, Tyrx; G).

=a | ie aa ue

Theabovespecimen was d I g
of the inflorescence. Although incomplete the material was sufficient
to indicate that the specimen could not be included satisfactorily in
any known species.

Puya Macbridei, spec. nov., 1 m. alta: foliis anguste triangulari-
bus, 7-8 dm. longis, 25 mm. latis, supra glabris, subtus minute ad-
presso-lepidotis, apice longe caudatis, spinis validis, atro-brunneis,
uncinatis, 6 mm. longis: scapo erecto, tereti, glabro; scapi bracteis
deciduis: inflorescentia elongata, laxe racemosa, omnino minute
cinereo-tomentosa: bracteis anguste lanceolatis, acuminatis, pedi-
cellis brevioribus, crassis, reflexis: floribus erectis vel suberectis, pedi-
cellis validiusculis, sursum incrassatis, 2 cm. longis; sepalis anguste
lanceolatis, 35 mm. longis, 7 mm. latis; petalis 6 em. longis, spiraliter
contortis, intus nudis. PI. III, figs. 17-18.

PERU: Ancacus: gravel river bluffs, Recuay, alt. ca. 3000 m., 1922, Mac-
bride & Featherstone 2516 (FM, Type = (3),

This species is closely allied to P. stenostele Harms, from which it
differs chiefly in the smaller floral bracts. :

Puya pectinata, spec. nov., saxicola, verisimiliter metralis:
foliis anguste triangularibus, subtus dense adpresseque pallido-

PERU: : i i ck crevices,
Pr steeadh a ee sy 1008, Macbride 2080 (RM, taen; phot. G).

The specimen described above is somewhat immature so that it
is probable that all the parts attain slightly larger dimensions than
given here. The photograph accompanying the specimen 18 ae
evidently not of the specimen itself, and is so difficult to correlate
that it seems best not to consider it in describing the species.

Puya Penduliflora, spec. nov., ad 6 dm. alta vel paulo ultra:

STUDIES IN THE BROMELIACEAE,—III 13

foliis interioribus solum cognitis, his anguste triangularibus, spinuloso-
serratis, supra glabris, subtus minute dissiteque lepidotis: scapo
erecto, €e comparatione gracili; scapi bracteis ex ovato longe lami-
natis, serrulatis, submembranaceis, rubris: inflorescentia ad 25 cm.
longa, paupere paniculata, omnino minute albo-tomentosa; bracteis
primariis ovatis, acutis, serrulatis, submembranaceis, rubris, nervatis,
quam racemi axillares multo brevioribus: racemis laxe paucifloris,
apice sterilibus, basi complanatis, lateralibus haud ultra 9 cm. longis;
bracteis florigeris primariis similibus sed integris, quam pedicelli
paullo longioribus: floribus perpendulis, ca. 5 cm. longis, ad 13 mm.
longe pedicellatis, sepalis anguste lanceolatis, 35 mm. longis, acumi-
natis; petalis oblongis, 45 mm. longis, intus nudis, (ex sicco) violaceo-
Virentibus; staminibus styloque petalis brevioribus sed post anthesin
paulo exsertis. Pl. IV, figs. 3-4.

BOLIVIA: La Paz: province of La Recaja, Hacienda Casana on the road
to Tipuani, alt. 1400 m., 1922, O. Buchtien 7182 (US, ryrx; phot. G).

It is difficult to say whether this species should be included in the
subgenus Eupuya or Pitcairniopsis. Its branches are sterile at apex,
but this sterile portion is short and bud-like just as in some specimens
of such species as P. caerulea, P. violacea, and P. floccosa. Probably
it would be best to consider Eupuya as comprising only such species
as have the sterile apex of the branches definitely elongated, and con-
sequently to include P. penduliflora in Pitcairniopsis. In this case its
relations would be with P. Pearcei from which it differs in its short
racemes and serrate primary bracts.

Puya stipitata, spec. nov., ca. 2 m. alta: foliis anguste triangulari-
bus, 5-6 dm. longis, 18 mm. latis, supra glabris, subtus membrana
“inerea e lepidibus formata obtectis, spinis 5-6 mm. longis: scapo
valido, erecto; scapi bracteis foliaceis elongatis: inflorescentia panicu-
ata; racemis ad apicem usque fertilibus, dense strobilatis, 9-10 cm.
longis, 3 cm. latis, patentibus vel reflexis, omnino dense pallideque
lanatis, stipite 20-25 mm. longo: bracteis primariis anguste ovatis,
‘cutis, stipite paulo longioribus: bracteis florigeris late ovatis, abrupte
‘cutis, sepala ad 14 aequantibus, integris, late convexis, ecarinatis:
Horibus erectis, 35 mm. longis, obscure crasseque ad 7 mm. pedicel-

ain Sepalis late ellipticis, obtusis, 20 mm. longis; petalis caeruleo-
us, 28 mm. longis, intus nudis; staminibus styloque petala
‘equantibus, ovario pyramidali. PI. IV, figs. 5-6.

PERU: Hvanuco: rock illsi i blo 7 miles north-
,. ; ss-shrub hillsides, Cani, pueblo
G. of Mito, alt. ca. 2800 m., 1923, Macbride 3436 (FM, TrPE; G), 3854 (FM,

14 SMITH

In Mez’s key to Puya in DeCandolle’s Monographiae, this species
would come out to P. venusta but differs in having the floral bracts
only half as long as the sepals. Judging by the descriptions of the
more recent species, it seems most nearly related to P. gummifera
Mez & Sodiro, but has the woolly covering of the bracts persistent and
the racemes very definitely stipitate.

Streptocalyx Williamsii, spec. nov., e fragmentis solum cognita:
foliorum laminis linearibus, acuminatis, 6 dm. vel ultra longis, 3 em.
latis, subtus dense adpresseque albo-lepidotis: scapo ignoto: in-
florescentia cylindrica, ad 3 dm. longa, bipinnatim paniculata, utrinque
dense albo-farinosa; bracteis primariis amplis, spicas axillares aequan-
tibus vel superantibus, late ovatis, acutis, lacinato-denticulatis,
chartaceis, fulgide rubris: spicis laxe 2—5-floris, distichis, apice sterili-
bus, infimis distincte stipitatis; rhachi angusta, subrecta; bracteis
florigeris late ovatis, pungenti-mucronatis, integris, ovarium aequan-
tibus vel superantibus sed non omnino obtegentibus, nervatis, ad 12
mm. longis: floribus 3 em. longis, suberectis; sepalis liberis, mucro-
natis, valde asymmetricis, 17 mm. longis; petalis 26 mm. longis, intus
nudis, (ex sicco) purpureis, quam stylus staminaque longioribus;
ovario oblongo, valde angulato. Pl. IV, fig. 7.

PERU: Loreto: forest, La Victoria on the Amazon River, 1929, L. Williams
2722 (FM, Tyre; phot. G).

In the general form of the inflorescence this species resembles
Streptocalyx Poeppigii, Vallerandi, and juruana. It differs from 8S.
Poeppigit and S. Vallerandi on account of its large floral bracts and
from S. juruana on account of its slender nearly straight: spike-
rhachis.

Thecophyllum capituligera (Griseb.), comb. nov. Tillandsia
capituligera Griseb. Cat. Cuba 254 (1866). 7. fastuosa André, Rev.
Hort. Ix. 568 (1888). Guzmania fastuosa André ex Mez in DC. Mon.
ix. 926 (1896). G. capituligera Mez in DC. Mon. ix. 926 (1896).

BA: Oriente: Loma del Gato, Wright 3275 (G, 1soTyrE); Gran yt
. 191 ICA: Harris in herb. bot.

Hitchcock 17633 (G, NY). COLOMBIA: CUNDINAMARCA: Barroblanco, peor!
se alt. 2900 m., 1876, André 1746 yy. ISOTYPE of Tilla’ re
osa;

phot. G).
Pennell & Killip 5747 (G).

In studying the isotype material of Tillandsia capituligera inate
and 7. fastuosa André it was impossible to find differences betwee
the two which seemed of specific value. At the same time it was note

|
|
|
|

STUDIES IN THE BROMELIACEAE.—III 15

that André in his Broméliacées gave an excellent plate (t. 37) of 7.
fastuosa and showed the petals (fig. 3) as free and bearing two scales
each.at base. These characters plus the polystichous-flowered spikes
would exclude the species from Tillandsia and Guzmania and make
ita Thecophyllum

Tillandsia boliviensis Baker emend. L. B. Smith, acaulis videtur
sed basi extrema ignota, 2 dm. alta: foliis fasciculatim rosulatis, in-
florescentiam subaequantibus, dense lepidibus cinereis paullo pruinose
obtectis, margine lepidibus magnis paleaceis patentibus auctis;
vagina vix distincta; lamina anguste triangulari, subulato-acuminata:
scapo erecto, brevi, foliis fere omnino obtecto; scapi vaginis imbri-
eatis, lanceolatis, apiculatis vel infimis breviter caudatis, dense
cinereo-lepidotis: inflorescentia e spica terminali 8 cm. longa et una
laterali 4-5 em. longa composita; bracteis primariis eis scapi similibus,
spicis subduplo brevioribus: spicis oblongo-lanceolatis, 4-10-floris;
bracteis florigeris imbricatis sed ob formam angustam rhachin haud
obtegentibus, lanceolatis, acutis, ad 30 mm. longis, sepala superanti-
us, non carinatis, adpresse cinereo-lepidotis; rhachi glabra, alatim
angulosa, leviter geniculata: floribus erectis, brevissime pedicellatis;
sepalis liberis, anguste lanceolatis, acutis, 21 mm. longis, glabris,
hervatis, posticis alatim carinatis; petalis Negron ignotis:
ie brevi, bracteam subaequanti. PI. IV, figs

OLIVIA: La Paz: vicinity of La Paz, alt. 3,300 m., 1890, Bang 159a in
a (NY, phot. G).

The original description of this species was evidently compounded
from two sheets bearing the same number. One of these sheets is
Mllandsia paraensis Mez, the other after exhaustive checking with
all species of Tillandsia and Vriesia appears to be a distinct species.
Consequently the species is redescribed here in order to clarify its
identity, Owing to the absence of good flowers it is impossible to
ascertain its exact relationships but the spikes have much the same
general appearance as those of Tillandsia patula Mez and the leaves
‘an not be distinguished from those of 7. Lorentziana Griseb.

andsia exserta Fernald, Bot. Gaz. xx. 537 (1895). T. cinerea
Mez in DC. Mon. ix. 679 (1896). PI. IV, figs. 10-11.

M not T.
sateen) Srwatoa: re aca 1898, ree foe 381 (G, ii ee 1926, J. G.
rsa O4? 6477 (G); vi sean of "Topolobampo, 1910, an Standley & Russell

45 (G, US); Altata, 1903, C. A. Purpus gis Calif.); Yerba Buena,
Me. ity of Oulinean, 1004, T. S. Brandegee (Un Calif.); Los a 1927,

E. Jones 2 23458 (Pomona College); vicinity, ‘of Gu adalup 1910, ee

Sandley Ul 147, itv of Mazatlan, 1910, Foe, Standley
Russel] istes teh. 14755 (US); vicinity o

16 SMITH

A study of the material cited above leaves no doubt that Till-
andsia exserta and T. cinerea are identical. Evidently 7. exserta was
published just too late for inclusion in Mez’s monograph of the Bromel-
taceae, and until now has never been correlated with that treatment.
It is generally rather dangerous to try to predict ranges in the Bro-
meliaceae, but in this case it seems likely that 7. exserta is another
example of the strong endemism of the flora of northwestern Mexico.

Tillandsia Harmsiana, spec. nov., florifera 6 dm. alta vel ultra,
verisimiliter multo altior: foliis anguste triangularibus, acuminatis,
6 dm. longis, basi 4 cm. latis, omnino dense adpresseque cinereo-
lepidotis: scapo erecto, valido; scapi bracteis dense imbricatis, ovatis
cum laminis longis angustis recurvatisque: inflorescentia bipinnatim
paniculata, e spicis ca. 7 composita, dense cylindrica, 30 em. longa,
infra 5 cm. lata; axe spicis bracteisque omnino obtecto; bracteis
primariis ovatis, acutis, erectis, spicarum bases arte vaginantibus:
spicis rectis, stricte erectis, lineari-lanceolatis, acutis, 11-15 em.
longis, 3 cm. latis, 12-20-floris; rhachi dense lepidota, ad florum re-
ceptionem valde excavata: bracteis florigeris imbricatis, ovatis,
acutis, 3-4 cm. longis, quam sepala longioribus sed rhachin non ob-
tegentibus, carinatis, dissite lepidotis, aureo-brunneis, basi purpurels:
floribus stipitatis, 5 cm. longis; sepalis liberis, anguste ovatis, 3 em
longis, glabris, ecarinatis; petalis angustis, purpureis, intus callis
binis verticalibus auriculatis auctis ; staminibus petala subaequantl-
bus; stylo exserto. Pl. IV, figs. 12-15.

Pt cee Pag ep focky shrubby slopes, Mito, alt. ca. 3000 m., 1923,

The spikes of this species closely resemble that of 7. lateritua
André, and the petals have pronounced lateral folds as in 7. lateritia
and with the summits of the folds definitely auricled so that at first
glance they might be taken for the scales found in Vriesia. :

Tillandsia subandina (Ule) Mez in ms. comb. nov. Cipuropsis
subandina Ule, Verhandl. bot. Ver. Brandenb. xlviii. 149 (1907).

Vriesia altodaserrae, spec. nov., metralis vel ultra: foliis dense
rosulatis, 6-7 dm. longis, minutissime punctulato-lepidotis; hd ec
atro-brunnea; lamina subtriangulari, pallide viridi, basi 5-6 cm. lata:

orum brevioribus: racemis spiciformibus, 2-3 dm. longis, laxe Uma
floris, basi a prophyllis 1-2 auctis; bracteis florigeris late e

STUDIES IN THE BROMELIACEAE.—III 17

obtusis, per anthesin patentibus flores involventibus, fulgide aureis,
ad 28 mm. longis, quam sepala distincte brevioribus, apice paulo
punctulato-lepidoto excepto glabris, laevibus vel leviter venosis, haud
carinatis; rhachi gracili, glabra, flexuosa, sulcata: floribus distichis,
per anthesin patentibus, ad 7 mm. longe stipitatis; sepalis anguste
ellipticis, obtusis, ad 26 mm. longis, stramineis, valde nervatis; petalis

34 mm. longis, staminibus subaequantibus, basi a ligulis binis
auctis. Pl. V, figs. 1-2.

BRAZIL: Sio Pavto: epiphytic or terrestrial in dense forest, Estagéo
Biologica, Alto da Serra, alt. 800-900 m., 1929, L. B. Smith 1875 (G, Type),
1926 (G), 1924, F. C. Hoehne 9476 (G, SP), 1875, Mosén 3249 (S); Praia de

i 7 (8). ParanA:

Dusén 14414 (S)

In habit this species closely resembles Vriesia procera and V.
glutinosa, but can be distinguished from both of them even when it is
in fruit by the long narrowly triangular blades of the scape-bracts and
lower primary bracts. The flowers of all specimens so far examined
would indicate that it belongs in a group different from that of V.
procera and glutinosa on account of the slightly included stamens.
Consequently its systematic postition probably should be next to V.
reticulata from which it differs in its acuminate leaf-blades. The plant
is very beatiful when growing. The branches are bright yellow, the
leaves and the blades of the large bracts are light green, and the
broad bases of the bracts are bright red.

riesia Dusenii, spec. nov., semimetralis: foliis dense rosulatis,
minutissime punctato-lepidotis, ad 3 dm. longis; vagina elliptica,
brunnea; lamina lineari, acuminata, basi 15 mm. lata: scapo erecto,
gracillimo, glabro; scapi vaginis foliaceis, dense imbricatis: inflores-
centia graciliter cylindrica, bipinnata; bracteis primariis a basi ovata
longe acuminatis, quam spicae axillares multo longioribus: spicis laxe
2-3-floris, stricte erectis, basi prophyllo parvo auctis; bracteis flori-
eris ovatis, acutis, sepalis duplo brevioribus, carinatis, nervatis,
dissite lepidotis: floribus subsessilibus, subpatentibus; sepalis lanceo-
latis, acutis, 14 mm. longis, carinatis, nervatis, dissite lepidotis;
petalis aureis, 25 mm. longis, in laminam subrhomboideam dilatatis,
a ligulis binis triangularibus alte insertis auctis; stylo staminibusque
Petalis brevioribus. Pl. V, figs. 3-4.

BRAZIL: wale , 1910, Dusén 10712B (S, TyPE;

poe G); Antonina, Hin Debeor orien + virgin forest, 1914, Dusén

This species with its spikes largely concealed by the bracts has a

18 SMITH

habit so unlike that of any known Vriesia, that were it not for the
indubitable scales on the petals one would be inclined to seek its re-
lations in the vicinity of Tillandsia biflora. Even the scales are un-
usual in being inserted high up near the junction of the blade and
claw instead of at the base.

Vriesia Schippii, spec. nov., ca. 4 dm. alta: foliis rosulatis, ad 3
dm. longis, minutissime punctulato-lepidotis, subtus vittis transver-
salibus atro-brunneis ornatis; vagina magna, elliptica; lamina ligulata,
apice acuta, basi ad 3 cm. vel paulo ultra lata: scapo erecto, gracili,
glabro; scapi vaginis imbricatis, e basi ovata acuminatis, inflatis,
pallide brunneis, apice lepidotis, cetera glabris, sublaevibus: in-
florescentia simplicissima, secunda, laxe pauciflora; bracteis florigeris
eis scapi similibus, ad 43 mm. longis, non carinatis, sepala longe super-
antibus, nullo modo imbricatis, cum floribus secunde versis; rhachi
gracili, flexuosa: floribus ad 5 mm. longe stipitatis; sepalis late el-
lipticis, 17 mm. longis, glabris, leviter venosis; petalis ignotis: capsula
anguste ellipsoidea, 3 cm. longa. PI. V, figs. 5-6.

British Honpuras: epiphytic in mountain forest near Middlesex, alt. 600
m., W. A. Schipp S82 (FM, Trpx; phot. G).

Vriesia macrochlamys Mez & Wercklé is probably the nearest
relative of the above species, but differs in its obtuse floral bracts and
larger sepals.

2. ProvistonaL Key To tHe Genus GuzMANta with Notes
ON NEW OR CRITICAL SPECIES.

’ Mez in DC. Mon. Phan. ix. (1896).
* Mez in Bull. Herb. Boiss. ser. 2, iti. 131 (1903).

STUDIES IN THE BROMELIACEAE,—IIlI 19

The more recent work of Harms in the 2nd edition of the Pflanzen-
familien' is not adopted here because it is felt that the segregates from
Guzmania, namely Caraguata, Massangea, and Schlumbergera are
made on characters that are not of sufficient value to be generic.
Caraguata is distinguished from Guzmania on the basis of the outline
of the spike. Massangea is distinguished by having the sepals “ peta-
loid” and highly fused, yet to be consistent it would be necessary to
set off Guzmania Weberbaucri and G. Scherzeriana as a segregate from
Schlumbergera on the same basis. Schlumbergera itself seems the
least tenable of all, its one distinction being a compound inflorescence.
is character is a very convenient one in keying species, but is
searcely of phylogenetic value since it does not correlate with any
characters of the flowers themselves and since at least one species, (.
oezli, may have either a compound or a simple inflorescence. Con-
sequently Guzmania is here understood to include all members of the
Tillandsieae with fused or agglutinated naked petals, polystichous-
tanked spikes or racemes, and either non-petaloid sepals or else
rosulate leaves. This last distinction is for the purpose of excluding
Sodiroa, which seems to be a rather weak genus and on further study
may well follow Caraguata, Massangea, and Schlumbergera into the
card,

1. Jnflorescense simple.
2. In ‘Scan eyathifor rm, the outer bracts much enlarged.
3. Inn a r bracts differing i in shape from the outer, much re-
vd

e
4, Inner isi spatareatl persistent, exceeding the sepals:
MESIOPORCENCE BCADONE . go eu. «oo eee 1. G. lingulata.
- ne inet straight. " muembranaceous, soon disintegrat-
ing: inflorescence sunk in the leaf-rosette.
5. Inner bracts elliptic, obtuse, shorter than the sepals.
2. G. sanguinea,
, _ Inner bracts triangular, exceeding the sepals... .3. G. crateriflora.
3. Inner bracts Soniler to the ou
KG the inflorescence: inflorescence 6 cm.
OP MOTS 1ONR es oe ee ee 4. G. minor.
6. Leaves inflorescence not
ate es ean the inflorescence: ig ag A
2. Inflorescence elongate, spicate, the outer bracts not at all
onspicu
7. ae be 4 thin, pponarscie Bere
8. Infloreacenioe sterile toward ape
lant long-caulescent: faves narrowly ——

caulescens.
9. Plant acaulescen
shy apc or acuminate
11. orescence few-flowered: leaves not over 7mm
oa , eps ak sae eee 7. G. angustifolia.

‘ Harms, in Engl. & Prantl, Nat. Pflanzenf. 2 Aufl., xva. 65 ( 1930).

'

20 SMITH

11. Inflorescence many-flowered: leaves 12-50 mm.

wide.
TZ: Sil thin or chartaceous, not strongly
epl
13. Bracts distinctly gg saa than the fos
flowers up to 6 em. long........... G. Besoromeen
13. Bracts about sanalias: he flowers: tower
not over 3 emvlong 4.3 i eS, G. monostachia.

12. Leaves rigid, a tipucaenk strong. jepiiet
ues Fusrstehberpiiele
10. Leaves rounded-apiculate.
14. Inflorescence few-flowered: sepals 25 mm. one
11. G. fusispica.
14. Inflorescence many-flowered: sepals 12 mm. long.
12. G. apiculata.
8. Inflorescence fertile throughou
15. Sepals broadly obtuse: tects not dark-farinose.
16. Sepals 15 mm. lon
17. - sae = few-flowered: sepals owe!
Be es Cringe ine se ene . G, Melinonis.
1h Inflorescence many-flowered: sepals 8 ae
Me fo eee 14. G. platysepala,
oe ‘ay rae OI —25 mm
8. Upper floral tracks ‘acute or narrowly ssid 9
. G. bracteosa.
18. Upper floral bracts broadly rounded.
19. Floral bracts all obtuse: leaves cube
G. erythrolepis.
19. Lower floral bracts broadly acute: gare
densely pale-appressed-lepidote below
17. G. nicaraguensis.
5. Sepals acuminate: bracts dark-farinose....... 18. G. pita ose
fe Flozal bracts stout, coriaceou
. Floral bracts little if at all shorter than the sepals: at
ia t the upper part of the inflorescence densely strobi-

oi. ee bracts erect. lax
escence lax below, sterile toward apex. . --- - 19. G. laxa.
22. Inflorescence strobilate and nde throughowk:
23. Scape-bracts with a lea lade.
24. Up Upper nie bracts Is
Hlowe ser ch econ the bracts: Gas conten:
2 als
© Heer toting tec a
ont Upper £ floral bracts obtiise. 65.5. 35. 22. G. mucronala.
23. Scape-bracts without a leaflike blade.
26. Floral bracts strongly ne
27. shes al bracts _br acs ‘obi sigs

26. Floral barks even, not at all nerved.
28. phd = bracts equaling or exceeding the

29. Floral bracts broadly triangular gto oriostachya-

STUDIES IN THE BROMELIACEAE.—III 21

|. Inflorescence pene d.
‘31. od bracts ¢ equaling or longer than the sepals.
eth strongly recurved: primary bracts —
SILC eG Gals A eptdle eta ley ee a cs rs . G. Osyana.

34. Inflorescence short, capitate, e, few-flowered.. . .30. G. Cornuaulti.
34. — a elongate, cylindric or nearly so, many-

35. ‘Floral bracts apes One a 31. G. Harrisii.
on Floral brac
oes pals acumin. ee 38 mm. long. 32. G. nce

: 36. Serials abruptly acute, 22 mm. long... ee 33. G. colum
33. Spikes more or less elongate, not corym
37. Inflorescence short and compact, 0 oa pressed
ge
38. Spikes exceeding the primary bracts at maturity.
39. Spikes ovoid, acute: sepals connate for 2-3 mm.
. Morreniana.
39. Spikes ellipsoid: sepals connate for 10 “mm.
5. G. glomerata.
38. Spikes shorter than the primary bracts at maturity.
. A = 36. G. densiflora.
37. ab ead vane or less elongate, interrupted or at
least the lo reflexed.
40. pai za te rif feligbtne ine lax stan the inflorescence

41. Floral bracts acute.
42. —— ry bracts ample, nearly or quite conceal-
the ikes
43. imary bracts straight: sepals connate for
nearly half their len — jeg aie Neots G. densiflora.)
at a
Oepaa ken ee ar 0 toler.
42. Primary tenets narrow, not at all cogealing
the sa al —— eee . G. calamifolia.

sca Foe Gani brac se
r than
the ce pendulous < eee : ag ne 39. G. Bakert.
44. Spikes erect or spreading, but not nodding.
~ Mere long-stipi
. Spikes abbreviated, ovoid or obovoid.
47, a poe and bracts finely and irregulary ue
cross-striate 40. G. ent.

wb ele eee eee oe 8 ee ee

22 SMITH

sae hg elongate
Floral ants not at all carinate,
strongly imbricate: sepals 15 mm.
lon. eee Doe ae ee 42. G. Alisonii.
48, Floral bracts carinate toward apex,
barely imbricate: sepals 24 mm. long.
43. G. Splitgerberi.
45. Spikes sessile or very or ninco
49. Floral bracts not at all e
50. Sepals free or equally peso
51. Floral bracts carinate toward —
(43. G. Splitgerberi.)
51. Floral bracts not at all carinate.
52. Spikes few, remote......... 44, G. ventricosa.
62. Spikes ma — massed at apex of in-
ores
53. Floral rests ae the
s, 13 mm. long..... 45. G. brevispatha.
53. Floral bracts beer flat, 21 mm.
Sp cu el ulieanur aan, 46. G. tarapotina.
50. _— — for 6 mm. axially and 3
apne sinus, densely
0te 2.43 4 clas pleiosticha.
49. Flora bracts duaaty emarginate: pte
or nearly 60... . is s6 65 ee 48. G. polycephala.
40. Spikes aad har tealbewanaie lax: flowers mostly a :
ee Pe Pere iy AN, Geer ee . G. panniculata.
31. Floral bracts shorter than the sepals.
54. Spikes abbreviated, or if gs on then —,
55. “Spikes pendulous: sepals 37 mm. long....... 50. G. Lehmanniana.
55 ect or tei ‘ ‘if slightly nodding then
e ac af uch short
ree amply. genre me ‘ “—
. c rT an the tora
“Seeertcrmmieme se
57. Flo e he flor
ns atc bes = ine 52. C. Pennell
56. Inf wali nce not more than bipinn.
58. es, or at sah the lowest, debnitely stipitate.
i. Spikes 2-3-flow: m. lon one
. Petals varie: more than ice as long as
the sepals: sepals 25 mm. long......---- 53. G. virescens.
60. Petals oe catalensereet only about 5 mm. longer
e se mm. shes
sees “54. @. Donnellsmithii
59. Spikes 4-00 -flowered.
61. Spikes several to many: inflorescence elo: sor
6 connate for 9-10 mm.: cakes rt tout

+66 eee oo ee eee & ee re S

ome c ite
. Sepals acuminate........- (32. G. megastachya-)
65. Sepals obtuse.
66. Axis oe the roi concealed
by t

flowers pote * Png pedal G. Dussii.

STUDIES IN THE BROMELIACEAR.—III 23

66. Axis of the inflorescence not at all
concealed: flowers subsessile
57. G. Andreana,
64. Spikes globose or ellipsoid.
67. Sepals equally connate for 3 mm.
G. straminea.
Depale (ree. oesks eae oe sas 59. G. Ekmanii.

sit short-stipitate, —
apex, ca. 15-flow G. elongata,
69. Spikes conspicuously stipltate, le
througho
70. Axis of hufeaniecs and spike-
rhachis. slender: spikes 3-6-

ere
70. Axis of inflorescence and spike-
— stout: spikes many-
HOWSIOG ck it 62. G. multiflora.
68. Spikes dense aot yeas ee all
pressed one against anot
va 8 tne subcylindric . : 63. G. Van a

Oeic bee cee eee ees

72. Leaves fics cinereous-lepidote...... é. lepidota.

73. Spikes all spr g.
24, Spikes globows. «2. ps5. - ss 66. G. sapien
74, Spikes subcylindric..... Re (76. G. patula. )
. Spikes mostly erect.
75. Floral bracts obtuse. ..........5.: 67. G. —
75. Floral bracts acuminate....... 68. G. Goudotiana.
58. Spikes all e.
76. lpdcnaeains pinnate
77. Nigger not more than
8. Spikes subglobose, oat owe OB mm. long. ee
G. acorifolia.
78. Spikes ellipsoid, 60 mm. long........ (60. G. elongata.)
77. Sepals 23-38 mm. lo
79. Floral bracts acuminate oh hee (32. G. megastachya.)
79. Floral bracts _ . lene ee
e
80 Rapa connate for 3 thei Tenet mike.
80. Sepals agar ge ht ri ool Sha
their len spi
oe Spikes lips, S-12flowered, . (59. e ini
Spikes corymbose, 2 -3-flowered. . (53. G cens.)
et Inflorescence digitate.
orescence sca .
83. Leaves kets with irregular transverse
lines: floral bracts even.....--.--+-++-+ 70. G. vittata
83. Leaves without transverse lines: ral
Tracts NOVVOd 3. op ee ee oe i: G, compacta,
lorescence mduiar. 3. 72. G. Eduardi.

82. Inf subsessile an
54. ed elongate te and ta the flowers = st all touching one

24 SMITH

84. og connate for 8-10 m
. Floral

bracts and raise te 2. Pees 73. G. brasiliensis.
5B, Floral bracts and sepals ahiibs See: 74. G. Scherzeriana
84. Sepals not connate fe: ira than 3
ag Sepals 25-3

wal bracts greene 7 acute, 25 mm. long.
75. G. Faweettit.

87. siesta bracts oblong or elliptic, obtuse, only 15 mm.
CO acca a ey Veg aN i NS a ea aut (59. G. Ekmanii.)

86. Bepla not more than 16 mm. lon
88. (on aR lage lax, at least ihe serene shorter
e internodes.

30, Some of the floral bracts slightly imbricate. . . 76. G. patula,

89. None of the floral bracts at all imbricate..... 77. G. Roezli.
88. ace togete dense, all exceeding the si tapnodiele
90. Sepals acute, lanceolate..............--- 78. G. Plumierii.
90. Se ales use.
91. Beoale Gonnate for 8mm... 2... 6 ess (60. G. elongata.)
91. Sepals free

92. Floral bracts 12 mm. long, slightly shorter
cer Lr sepals: inflorescence —s y
Sh pe er ce G. condensata.
92. Floral ats 17 mm. long, much eck
n the sepals: inflorescence — bi-
WO ee at eae _ G. costaricensis.

1. Guzmania lingulata (L.) Mez. The Antilles and Nicaragua
(ace. to Mez in ms.) south to Ecuador, Bolivia, and the Brazilian
state of Matto Grosso. Mez in DC. Mon. Phan. ix. 899 (1896).
Tillandsia lingulata L. Sp. Pl. 286 (1753). Caraguata lingulata var.
cardinalis André, Illustr. Hort. xxvii. 35, t. 374 (1880). Guzmania
oo var. cardinalis André ex Mez in DC. Mon. Phan. ix . 900

DOR: Guayas: Teresita, 3 km. west of Bucay, alt. 270 m., 1923.

ECUADO
Hitchcock 20406 (G). BRAZIL: Matto Grosso: Capio Secco, 1894, Lind-
man A2359 (S, phot. G).

The above citations constitute large extensions of range for the
species. The variety cardinalis does not seem sufficiently distinct t
warrant maintaining it. ;

2. G. sanguinea André. Southern Colombia. André ex Mez
DC. Mon. Phan. ix. 901 (1896). Caraguata sanguinea André in Rev.
Hort. lv. 468 (1883).

On account of its prominent scape and floral bracts it is probable
that G. sanguinea var. erecta André ex Mez} or Caraguata myriostignt
as André tentatively called it, is specifically distinct from ty] pical
at bias However, the specimens upon which the variety is based

, Mez

van are rg tag eng po (1896).

STUDIES IN THE BROMELIACEAE.—III 25

are so imperfect that it has been impossible to include them in the
foregoing key.

3. G. crateriflora Mez & Wercklé. Costa Rica, Cocos Island (to
Costa Rica). Bull. Herb. Boiss. ser. 2, v. 110 (1905). Tillandsia sp.
Stewart in Proc. Calif. Acad. Sci. ser. 4, i. 388 (1912). PI. V, fig. 7.

Cocos Istanp: 1889, Snodgrass & Heller 962 (G); Wafer Bay, 1930, H. K.
Svenson 322, 434 (G).

4. G. minor Mez. Nicaragua, Costa Rica, Panama, Brazil. Mez
in DC. Mon. Phan. ix. 901 (1896).

BRAZIL: Par&: epiphytic in woods, Tapand, near Para (Belém), 1929,
Killip & Smith 30349 (US).

5. G. brachycephala (Bak.) Mez. Peru. Mez in DC. Mon.
Phan. ix. 902 (1896). Tillandsia brachycephala Bak. in Journ. Bot.
xxvi. 40 (1888).

6. G. caulescens Mez & Sodiro. Ecuador. Bull. Herb. Boiss. ser.
2, v. 112 (1905).

7. G. angustifolia (Bak.) Wittm. Costa Rica, Colombia. Wittm.
in Engl. Bot. Jahrb. xi. 62 (1889). Caraguata angustifolia Bak. in
Gard. Chron. New Series, xxii. 616 (1884).

8. G. Berteroniana (R. & S.) Mez. Porto Rico. Mez in DC.
Mon. Phan. ix. 904 (1896), as G. Berteroana. Caraguata Berteron-
tana Schult. f. in R. & S. Syst. vii. 1229 (1830).

. G. monostachia (L.) Rusby. Antilles, Costa Rica, Panama,
Venezuela, Colombia, Ecuador, Peru, Bolivia. Rusby ex Mez in DC.
Mon. Phan. ix. 905 (1896), as G. monostachya. Renealmia mono-
stachia L. Sp. Pl. 287 (1753).

10. G. Fuerstenbergiana (Kirchh. & Wittm.) Wittm. Ecuador.
Wittm. in Engl. Bot. Jahrb. xi. 61 (1889). Caraguata Fuerstenbergana -
Kirchh. & Wittm. in Berl. Gartenz. ii. 299 (1888).

ll. G. fusispica Mez & Sodiro. Ecuador. Bull. Herb. Boiss. ser.
2, v. 112 (1905).

ok €modia paulo superantibus: inflorescentia simplicissima, poly-
Sticha, multiflora, dense strobiliformi, 8 em. longa, fusiformi, ad
glabra: bracteis florigeris suborbicularibus,
is, quam sepala multo

longioribus: floribus breve stipitatis; sepalis ellipticis, 12 mm. longis,

26 SMITH

late acutis, aequaliter ad 3 mm. connatis, membranaceis, striatis:
capsulis 3 em. longis; coma rubro-brunnea. PI. V, figs. 8-9.

PERU: Junin: epiphytic, Hacienda Schunke, La Merced, alt. ca. 1300 m.,
1923, Macbride 5711 (FM, Typx; phot. G).

13. G. Melinonis Regel. French Guiana. Gartenfl. xxxiv. 116
(1885), as G. Melinoki, which however was a typographical error
according to Mez in DC. Mon. Phan. ix. 908 (1896).

14. G. platysepala Mez & Baker. Nicaragua. Bull. Torr. Bot.
Club, xxx. 437 (1903).

15. G. bracteosa André. Ecuador. André ex Mez in DC. Mon.
Phan. ix. 908 (1896). Caraguata bracteosa André, Rev. Hort. lx. 565
(1888).
16. G. erythrolepis Brongn. Cuba, Jamaica, Porto Rico. Brongn.
ex Planch. Flor. d. Serres, ser. 2, i. 25, t. 1089 (1856).

17. G. nicaraguensis Mez & Baker. Nicaragua. Bull. Torr. Bot.
Club, xxx. 436 (1903).

18. G. calothyrsus (Beer) Mez. Peru. Mez in DC. Mon. Phan.
ix. 910 (1896). Anoplophytum calothyrsus Beer, Brom. 263 (1857).

19. G. laxa Mez & Sodiro. Ecuador. Bull. Herb. Boiss. ser. 2, VY:
111 (1905).

20. G. conifera André. Ecuador, Peru. André ex Mez in DC.
Mon. Phan. ix. 911 (1896). Caraguata conifera André, Rev. Hort.
Ix. 565 (1888).

PERU: Junin: Hacienda Schunke, La Merced, alt. ca. 1300 m., 1923, Mac:
bride 5615 (FM, G); same, above San Ramon, alt. 1400-1700 m., Killip &
Smith 24872 (US).

21. G. Devansayana Morr. Ecuador. Belg. Hortic. xxxiii. 119, t-
_ 8-9 (1883).

22. G. mucronata (Griseb.) Mez. Venezuela. Mez in DC. Mon.
Phan. ix. 912 (1896). Tillandsia mucronata Griseb. in Goett. Nachr.
20 (1864).

23. G. strobilantha (R. & P.) Mez. Peru. Mez in DC. Mon.
Phan. ix. 913 (1896). Bonapartea strobilantha R. & P. Fl. Peruv. 1
39, t. 263 (1802). Guzmania parviflora Ule, Verhandl. Bot. Ver.
Brandenb. xlviii. 146 (1907).

PERU: Loreto: epi ie i i as de Ponasa, alt.
1000 m., 1903, Ule 5p (B. wren of G. porvifora Uie, phot G).

Since no good distinction has been found between Guzmania part
flora and G. strobilantha and since an annotation label of Mez on the
former indicates that he considers them probably identical, the twe

STUDIES IN THE BROMELIACEAE.—III 2

species are merged in this treatment. This merging extends the range
of G. strobilantha from central to northeastern Peru.

24. G. pallida, spec. nov., verisimiliter acaulis, ad 6 dm. alta:
foliis 4 dm. longis; vagina ovata, minutissime dissiteque punctulato-
lepidota; lamina ligulata, 35 mm. lata, rotundato-apiculata, supra
glabra: scapo glabro, suleato; scapi bracteis remotis, ovato-acutis,
submembranaceis, nullo modo foliaceis, minutissime obscureque
punctulato-lepidotis: inflorescentia simplicissima, ellipsoidea, dense
strobiliformi, 75 mm. longa, 35 mm. lata: bracteis florigeris late ovatis,
acutis, coriaceis, basi excepta valde striatis, sepala bene superantibus:
floribus brevissime pedicellatis; sepalis ovatis, acutis, basi ad 3 mm.
connatis, 15 mm. longis, minutissime obscureque punctulato-lepi-
dotis; petalis 25 mm. longis, flavis (ex sicco), lamina patenti, elliptica,
obtusa; staminibus styloque e petalorum fauce exsertis. Pl. V, figs.
10-11.

COLOMBIA: Maapa.ena: epiphytic in forest on banks of the Rio Gaira,
alt. ca. 1200 m., 1898-99, H. H. Smith 2342 (NY, Trpx; phot. G).

25. G. coriostachya (Griseb.) Mez. Venezuela, Colombia. Mez
in DC. Mon. Phan. ix. 914 (1896). Caraguata coriostachya Griseb. in
Goett. Nachr. 21 (1864). Tillandsia nigrescens André, Rev. Hort. Ix.
568 (1888).

COLOMBIA: Macpaena: Santa Marta, 1898-99, H. H. Smith 2853 (NY,
phot. G). Sanranper: terrestrial in woods, vicinity of Vegas, alt. 2600—

m., 1926, Killip & Smith 15958 (G). Ex Vai: epiphytic in forest
above La Cumbre, alt. 2000-2200 m., 1922, Pennell & Killip 5793 (G). _Na-
RINO: Altaquer, alt. 1400 m., 1876, André 3327 (K, TPE of Tillandsia nigres-
cens André; phot. G).

André, and Mez following his lead, have confused 7° illandsia nigres-
cens with Bonapartea strobilantha R. & P., but an examination of the
type of the former species shows that it has the triangular even bracts
of Guzmania coriostachya and should be included under that species.

IS view is further strengthened by the occurrence of Guzmanra
Plea at several stations (as cited above) between its type
ccality and that of Tillandsia nigrescens. co

26. G. Michelii Mez. Colombia. Bull. Herb. Boiss. ser. 2, m.
226 (1903).

27. G. strobilifera Mez & Wercklé. Costa Rica. Bull. Herb.

Iss. ser. 2, v. 110 (1905).

G. musaica (Linden) Mez. Colombia. Mez in DC. Mon.

en ix. 898 (1896). Tillandsia musaica Linden in Cat. Expos. Gand.

28 SMITH

29. G. Osyana (Morr.) Mez. Ecuador. Mez in DC. Mon. Phan.
ix. 914 (1896). Caraguata Osyana Morr. in Belg. Hortic. xxxv. 254, t.
16-17 (1885).

30. G. Cornuaulti André. Colombia. André ex Mezin DC. Mon.
Phan. ix. 925 (1896). Tillandsia Cornuaulti André, Rev. Hort. Ix.
568 (1888).

31. G. Harrisii Mez. Jamaica. Mez in DC. Mon. Phan. ix. 927
(1896).

32. G. megastachya (Bak.) Mez. Lesser Antilles. Mez in DC.
Mon. Phan. ix. 928 (1896). Tillandsia megastachya Bak. in Journ.
Bot. xxvi. 46 (1888).

Sr. Vincent: Morne A Gavon, alt. 600 m., 1890, H. H. & G. W. Smith 1682
(NY, phot. G).

The above specimen has the floral bracts slightly shorter than the
sepals, but is typical in other respects. Contrary to Mez’s description,
the sepals in this species are somewhat connate at base (PI. V, fig. 12).

33. G. columnaris Mez & Sodiro. Ecuador. Bull. Herb. Boiss.
ser. 2, v. 113 (1905).

34. G. Morreniana (Linden) Mez. Peru. Mez in DC. Mon.
Phan. ix. 932 (1896). Massangea Morreniana Linden in Cat. Expos.
Brux. (1880).

PERU: Cuzco: epiphytic, slopes of Media Naranja, Urubamba Basin, alt.
2000 m., F. L. Herrera 2039 (FM. G).

Prior to this collection the country whence the species was derived
was in doubt.

35. G. glomerata Mez & Wercklé. Costa Rica. Fedde Rep.
Spec. Nov. xiv. 256 (1916). ass

36. G. densiflora Mez. Colombia. Bull. Herb. Boiss. ser- 2)

226 (1903).

37. G. Sodiroana Mez. Ecuador. Bull. Herb. Boiss. ser. 2,
114 (1905). ;

38. G. calamifolia André. Colombia. André ex Mez in DC.

Mon. Phan. ix. 931 (1896). Caraguata acorifolia André, Rev. Bett
lx. 566 (1888), not Guzmania acorifolia (Griseb.) Mez (1896).

39. G. Bakeri (Wittm.) Mez. Colombia. Mez in DC. Mom
Phan. ix. 933 (1896). Caraguata Bakeri Wittm. in Engl. Bot. Jahrb.
xi. 59 (1889).

_ 40. G. Lindeni (André) Mez. Peru. Mez in DC. Mon. Phan.
- ee (1896). Massangea Lindeni André, Ill. Hort. xxv. 55, t- 30?

STUDIES IN THE BROMELIACEAE.—IIl 29

41. G. Killipiana, spec. nov., terrestris, ad 1 m. alta: foliis 8-9
dm. longis, subtus obscure punctulatolepidotis: vagina haud distincta,
badia; lamina ligulata, acuminata, em. lata: scapo pervalido,
erecto, glabro; scapi bracteis nbyicnt tis, late ovatis, acuminatis,
pungentibus, flavo-viridibus, dorso longitudinaliter rubro-striatis:
inflorescentia laxe bipinnata; axe valido, angulato, glabro; bracteis
primariis eis scapi similibus, quam spicae brevioribus vel infimis
paulo longioribus: spicis valide ad 1 em. longe stipitatis, dense stro-
biliformibus, ellipsoideis vel obovoideis, 30-40-floris, 7 cm. longis,
4-5 cm. latis; bracteis florigeris 3 cm. longis, sepalis multo longiori-
bus, anguste ellipticis, obtusis vel minute apiculatis, fere planis,
coriaceis, striatis, subglabris: floribus subsessilibus, 38 mm. longis;
sepalis ovatis, acuminatis, 16 mm. longis, subliberis, carinatis, cori-
aceis, punctatis; petalis albis, erectis; staminibus inclusis. Pl. VI
st 1-2.

: Junin: Pichis Trail, Enefias, alt. 1600-1900 m., Killip & Smith
26680 ( (US, pas phot. G); Pichis Trail, Yapas, alt. 1350-1600 m., Killip &
Smith 25561 (US).
42. G. Altsonii L. B. Smith. British Guiana. Contrib. Gray
Herb. Ixxxix. 7 (1930).
43. G. Splitgerberi Mez. Dutch Guiana. Mez in DC. Mon.
Phan. ix. 930 (1896
44. G. ventricosa (Griseb.) Mez. Venezuela. Mez in DC. Mon
So ix. 929 (1896). Tillandsia ventricosa Griseb. in Goett. Nechy.
1864).
ae brevispatha Mez. Peru. Fedde Rep. Spec. Nov. iii. 45
G. — Ule. Peru. Verhandl. Bot. Ver. Brandenb.
slviii, 147 (190
47. G. sities (Griseb.) Mez. Venezuela. Mez in DC. Mon.
Cot ix. 930 (1896). Tillandsia pleiosticha Griseb. Goett. Nachr.
64).

48. G. andes vere a & Wercklé. Costa Rica. Fedde Rep.
Spee. Nov. xiv. 254 (19
i Ppanniculata pe Peru. Bull. Herb. Boiss. ser. 2, v. 116

D).

50. G. Lehmanniana (Wittm.) Mez. Colombia. Mez in DC.
Mon. “ting ix. 934 (1896). Schlumbergeria Lehmanniana Wittm. in
Engl. Bot. Jahrb. xi. 60 (1889).

51. G, candelabrum André. Colombia. André ex Mez in DC.
Mon. Phan. ix. 935 (1896). Caraguata candelabrum André, Rev.
Hort. lx. 566 (1888).

30 SMITH

COLOMBIA: Sanranper: epiphytic in woods, vicinity of Las Vegas, alt.
2600-3000 m., Killip & Smith 15873 (G).

52. G. (?) Pennellii, spec. nov., e fragmentis solum cognita:
foliis verisimiliter rosulatis, 55 cm. longis, subtus minutissime punc-
tato-lepidotis, supra glabris; vagina late ovata, brunnea; lamina
ligulata, acuta, 25 mm. lata: inflorescentia ample tripinnatim panicu-
lata, glabra; bracteis primariis parvis, quam rami axillares multo
brevioribus, ovato-acutis; ramis 7 cm. vel ultra longis, laxiuscule
florigeris: spicis paucifloris, apice sterilibus, longe stipitatis; bracteis
florigeris ellipticis, 5 mm. longis, quam pedicelli subduplo breviori-
bus, obtusis, nervatis, submembranaceis: floribus suberectis, longe
gracileque pedicellatis; petalis (ex sicco) albis, 15 mm. longis, stylo
staminibusque longioribus, lamina ovata, obtusa; sepalis anguste
ellipticis, obtusis, nervatis, 8 mm. longis, aequaliter ad 3 mm. con-
natis: capsula angustissime subclavata, 17 mm. longa. PI. VI, fig. 3.

COLOMBIA: Bourvar: terrestrial in shrub zone, below Paramo de Cha-
quire, Cordillera Occidental, alt. 2800-3100 m., 1918, Pennell 4344 (NY, phot.

The petals on the above specimen are so old and shriveled that it
has been impossible to ascertain whether they are free or fused much
less whether any scales are present. Consequently it is not certam
that it is a Guzmania and not a Thecophyllum, but its unusual habital
characters immediately distinguish it from any known species
either genus.

3. G. virescens (Hook. f.) Mez. Venezuela or Colombia? Mez
in DC. Mon. Phan. ix. 943 (1896). Puya virescens Hook. f. in Bot.
Mag. lIxxxiii. t. 4991 (1857). Caraguata Beleana André, Rev. Hort.
Ixiii. 114 (1891). Guzmania Beleana André ex Mez in DC. Mon.
Phan. ix. 944 (1896).

54. G. Donnellsmithii Mez. Costa Rica. Bot. Gaz. xXxxV- 9

903).
55. G. Weberbaueri Mez. Ecuador, Peru. Bull. Herb. Boiss.
ser. 2, v. 114 (1905)

: : ; afios and
Cashuren, 8 hours east of Bain fe {300° 1800 ine 1025, 4. Fees
21861 (G).

56. G. Dussii Mez. Guadeloupe, W. I. Mez in DC. Mon. Phan.
ix. 923 (1896). Thecophyllum Dussii Mez, Bull. Herb. Boiss. se: 2,
iii. 131 (1903). PI. V, figs. 13-14. ‘

In reérecting the genus Thecophyllum on the basis of free Ped
bearing scales, Mez transferred! to it all species which he had trea

* Mez, Bull. Herb. Boiss. ser. 2, iii. 130 (1903).

STUDIES IN THE BROMELIACEAE.—III 31

under the section of the same name in the genus Guzmania.
section, Z'hecophyllum was characterized by its extremely abbreviated
secondary axes, and Mez’s action at the time implies that he considered
this as correlating perfectly with the petal characters. But he did
not mention having seen the petal structure of any but 7. ororiense
among these transferred species, and later he described species of
Thecophyllum with well developed secondary axes. Consequently it
is possible that better material of these transferred species will show
that some should have been kept in Guzmania as in the case of G.
Dussit.

57. G. Andreana (Morr.) Mez. Colombia. Mez in DC. Mon.
fee ix. 936 (1896). Caraguata Andreana Morr. Rev. Hort. lvi. 247
1884),

. G. straminea (C. Koch) Mez. Colombia? Mez in DC. Mon.
Phan. ix. 937 (1896). Anoplophytum stramineum C. Koch in Ind.
Sem. Hort. Berol. 7 (1856).

59. G. Ekmanii Harms. Haiti. In synonymy as Schlumbergeria
Ekmanii Harms, Notizblatt, x. 804 (1929).

60. G. elongata Mez & Sodiro. Ecuador. Bull. Herb. Boiss. ser.
2, v. 115 (1905).

61. G. gracilior (André) Mez. Colombia. Mez in DC. Mon.
Phan. ix. 937 (1896). Caraguata Van Volxemi var. gracilior André,
Brom. André, 54 (1889).

G. multiflora André. Colombia, Ecuador. André ex Mez in
DC. Mon. Phan. ix. 939 (1896). Caraguata multiflora André, Rev.
Hort. lx. 566 (1888).

- G. Van Volxemi André. Colombia. André ex Mez in DC.
Mon. Phan. ix. 938 (1896). Caraguata Van Volzemi André, Ill. Hort.
Xxv. 139, t. 326 (1878).

64. G. mitis, spec. nov., terrestris, metralis vel ultra, acaulis:
foliis rosulatis, 5-7 dm. longis, minutissime punctato-lepidotis;
Vagina late ovata; lamina ligulata, acuta, 3-5 em. lata: scapo erecto,
valido, glabro, scapi bracteis foliaceis, densis, internodia longe super-
antibus: inflorescentia bipinnatim paniculata, anguste cylindrica, 26
rm. longa, 5 cm. crassa, glabra; bracteis primariis ovatis vel lanceo-
latis, suberectis, infimis quam spicae multo longioribus: spicis globosis
vel late ovoideis, breviter stipitatis, 25 mm. longis, dense 12-15-
floris, infimis remotis, supremis dense aggregatis; bracteis florigeris
late Ovatis, late rotundatis, quam sepala multo brevioribus: floribus
Subsessilibus; sepalis 10 mm. longis, obovatis, obtusis, ad 2 mm.
ronhatis, nervatis; petalis albis, 18 mm. longis, staminibus styloque
longioribus. Pl. VI, figs. 4-5.

32 SMITH

COLOMBIA: Norte DE SANTANDER: dense woods, Pica-Pica Valley above
Tapata ‘nari of T ee): alt. 2100-2400 io Paget Killip if Bite 20195 (G,
TYPE); SANTANDER: dense wet woods, wes slope o t San Vicente

near Charta, alt. 2500-2700 m., 1927, Killip. & Smith 18990 ta).

65. G. lepidota André. Ecuador. André ex Mez in DC. Mon.
Phan. ix. 941 (1896). Caraguata lepidota André, Rev. Hort. lx. 566
(1888).

66. G. sphaeroidea André. Colombia. André ex Mez in DC.
Mon. Phan. ix. 942 (1896). Caraguata sphaeroidea André, Rev. Hort.
Ix. 566 (1888).

COLOMBIA: Sanranper: epiphytic in dense woods, Mesa de los Santos,
alt. 1500 m., 1926, Killip & Smith 15291 (G).

67. G. Zahnii (Hook. f.) Mez. Costa Rica. Mez in DC. Mon.
Phan. ix. 940 (1896). Caraguata Zahnii Hook. f. in Bot. Mag. xcix. t.
6059 (1873).

68. G. Goudotiana Mez. Colombia. Mez in DC. Mon. Phan.
ix. 942 (1896).

69. G. acorifolia (Griseb.) Mez. Venezuela. Mez in DC. Mon.
ass). 945 (1896). Tillandsia acorifolia Griseb. Gale Nacehr. 19

1864

G. vittata (Mart.) Mez. Brazil. Mez in DC. Mon. Phan. ix
a rip Bonapartea vittata Mart. in R. & S. Syst. vii. 1198 (1830).
71. G. compacta Mez. Nicaragua, Costa Rica, Colombia. Mez
in DC. Mon. Phan. ix. 947 (1896). Guzmania capitulata Mez &
Wercklé, Fedde Rep. Spec. Nov. xiv. 255 (1916).

COLOMBIA: Ex Vauiu: wooded cliffs of Dagua Valley, Cordoba, alt. 80-
100 m., 1922, Killip 5059 (G).

72. G. Eduardi André. Colombia. André ex Mez in DC. Mon.
Phan. ix. 947 (1896). Caraguata Morreniana André, Rev. Hort. lix.
12 (1887), not Guzmania Morreniana (Linden) Mez, (1896).

73. G. brasiliensis Ule. Brazil. Verhandl. Bot. Ver. Brandenb.
xlviii. 147 (1907).

74. G. Scherzeriana Mez. Costa Rica. Mez in DC. Mon. Phan.
ix. 949 (1896
fish "a Fawcettii Mez. Jamaica. Mez in DC. Mon. Phan. ix. 951

76. G. patula Mez & Wercklé. Costa Rica, Colombia. Fedde
Rep. Spec. Nov. xiv. 255 (1916).

7
COLOMBIA: T ie é > 1900 ms, Bh
Pennell 3396 (NY, p zea a” in forest, Libano, alt. 1500- 800

STUDIES IN THE BROMELIACEAE.—III 33

77. G. Roezli (Morr.) Mez. British Guiana, Colombia, Peru.
Mez in DC. Mon. Phan. ix. 948 (1896), as G. Roezlii. Schlumbergeria
Roezli Morr. in Belg. Hort. xxviii. 311 (1878

gears Ex VauLte: epiphytic in forest, eager fiat Dagua Valley, <n

“aire Killip 11474 (G). PERU: San Mar rat an Roque,
1350-1500. m. 0, L. Williams 7236 (G); Ac aouee: Sarisicel — vom
hillside. ee Huanta and Rio ‘Apurimac, alt. 750-1 1929,
Killip . & Smith 22686 (G).

Mez expressed doubt that the plants collected in Peru could be
the same species as those in British Guiana, but further material from
both localities shows no essential differences and an intermediate
station in Colombia makes their equality all the more plausible.

8. G. Plumierii (Griseb.) Mez. Lesser Antilles. Mez in DC.
Mon. Phan. ix. 950 (1896), as G. Plumieri. Brocchinia Plumierii
Griseb. Fl. Br. W. Ind. 593 (1864).

LESSER ANTILLES: Guapetoups: Savane 4 Mulets, Matoula, 1893,
1894, Pére Duss 3327/3443 A phot. G); St. Kirr’s: Summit of Mt. Misery
Britton & Cowell 553 (NY, phot. G).

The above specimens are characteristic of the species except that
the inflorescence is bipinnate instead of tripinnate. Consequently it
seems best to transfer the emphasis to other characters in keying G.
Plumieri and G. Fawcettii.

79. G. condensata Mez & Wercklé. Costa Rica. Bull. Herb.
Boiss. ser. 2, iii. 228 (1903).

80. G. costaricensis Mez & Wercklé. Costa Rica. Fedde Rep.
Spec. Nov. xvi. 78 (1919)

EXCLUDED OR NEWLY REDUCED TO SYNONMY:
Caraguata os Bot. Reg. xiii. sub t. 1068 (1827) = GuzmManta
R. & P. (see p. 19).
¢. ia. André i in Rev. Hort. lxiii. 114 (1891) = G. viRESCENS
(Hook. f.) Mez (see p. 30).

C. myriostigm a André, in synonymy as C. sanguinea var. erecta
André, Brom. Rides 46 (1889) = Guzmanta sP., but not G. sanguinea
André (see p. 24).

sanguinea var. erecta André, see C. myriostig
G. Beleana André ex Mez in nee Mon. Phan. ix. in O44 (1896) =
VIRESCENS (Hook. f.) Mez (see p. 30).
G. capitulata Mez & Wercklé, Fedde Rep. Spec. Nov. xiv. 255 (1916)
= G. compacta Mez (see . 32).
G. oo neere (Griseb.) Mez in DC. Mon. Phan. ix. 926 (1896) =
TaEcorpayiium CAPITULIGERA (Griseb.) L. B. Smith (see p. 14).

34 SMITH

G. clavata (Lam.) Urban, Fedde Rep. Spec. Nov. xv. 99 (1917) =
G. monostacuta (L.) Rusby. This plant as described by Lamarck is
obviously the same as G. Monostacuta (L.) Rusby, and remains so
in spite of Urban’s evident intention of interpreting it otherwise. I
can account for all the characters which Urban cites from Lamarck’s
description except the one important one: “ bracteis serratis.”” There
is no such character in Lamarck’s description, but there is a passage:
Es a ae embriqué d’écailles serrées,”’ in which Urban may
have mistaken “serrées’”’ to mean serrate instead of imbricate. If the
bracts are truly serrate the plant could not be a Guzmania or even a
member of the Tillandsieae.

G. complanata Wittm. Mededell. Rijks Herb. xxix. 92 (1916).
This plant is described as having distichous-ranked spikes and con-
sequently if it does not prove to be a Tillandsia it should constitute a
new genus rather than be included in Guzmania.

G. fastuosa André ex Mez in DC. Mon. Phan. ix. 926 (1896) =
THECOPHYLLUM CAPITULIGERA (Griseb.) L. B. Smith (see p. 14).

G. obtusa Rusby, Mem. N. Y. Bot. Gard. vii. 212 (1927) = VRIESIA
HELICONIOIDES Lindl.

G. parviflora Ule, Verhandl. Bot. Ver. Brandenb. xlviii. 146 (1907)
= G. sTROBILANTHA R. & P. (see p. 26). :

G. Peacockii (Morr.) Mez in DC. Mon. Phan. ix. 915 (1896). This
species is so imperfectly known that it is impossible to ascertain its
position within the genus.

G. sanguinea var. erecta André ex Mez in DC. Mon. Phan. ix. 901
(1896) = Guzmanta sp., but not G. sanguinea André (see p. 24).

Massangea Morr. in Belg. Hort. xxvii. 59, 199, t. 8-9 (1877) =
Guzmanta R. & P. (see p. 19).

Schlumbergera Morr. in Belg. Hort. xxxiii. 46, t. 4-6 (1883) =
Guzmanta R. & P. (see p. 19).

Thecophyllum Dussii Mez, Bull. Herb. Boiss. ser. 2, iii. 131 (1903)
= G. Dusst Mez (see p. 30).

Tillandsia capituligera Griseb. Cat. Cub. 254 (1866) = THECO
PHYLLUM CAPITULIGERA (Griseb.) L. B. Smith (see p. 14).

T. nigrescens André, Rev. Hort. Ix. 568 (1888) = G. CORIOSTACHYA
(Griseb.) Mez (see p. 27).

STUDIES IN THE BROMELIACEAE.—III 35

EXPLANATION OF PLATEs.
Puate I.

Fig. 1. ARCHMEA FERRUGINEA L. - aac (Killip & Smith 25815), flower X 1-
2. Same, com section of flow

3. Same, se 1;

4, Deerar ts | ELEUTHEROPETALA Sees Mez (Ule 6304), flower X 1.

5. Same, sagittal section of ovary X

6. Same, sepal x

7. Grecia conumBraNA L. B. Smith (Killip & Smith 18689), outer bract

8. Same, sepal Xx
9. GREIGIA coin EANA by 2 Bently (Macbride 4442), inflorescence X 1.
10. Same, dorsal view of sep
ll. Same, lateral view of s ce al 3
12. HonEenpercia BRirroNnIANA a B. Smith (N. L. Britton 2313), spike X 1.
13, Same, flower 1
14, Same, sepal x 2.
Puate II.

Dyckx1a Dusen L. B. Spon di 18081 ie — with collector’s sketches
m fresh materi

Prats III.
Fig. 1. Prrcarrnra CaLpEroni Standley & Smith (Caldéron 2595), flower
and bract K 14.
2. Same, sepal x 1.
3. Prrcarrnra cuzcoEnsis L. B. Smith (Weberbauer 7825), flower and bract
Same, sepal X 1.
5 Same, petal I,
Prre CAIRNIA PEcTINATA L. B. Smith (W. A. Archer), flower and bract Xx

7. hg sepal X 1.
8. Prre Rinark TruNcaTA L. B. Smith (Killip & Smith 22414), flower and
1.

9. Same, sepal x 1.

10. Poya eiuemeues L. B. Smith (Macbride & Featherstone 1339), flower
and bract x G

ll, Same, se

x
13 oem cuscras L. B. Smith (Weberbauer 6474), flower and bract X 1.
ie, se al x
a de LLATENSIS L. B. Smith (Macbride & Featherstone 2300), floral

15. Same, flower < 1.
16. Same, se’ pal x I fl d
17. Puya pease L. B. Smith (Macbride & Featherstone 2516), flower an

brac
18. Same, t sepal X 1. :
Puate IV.
Fig.l. Poya ae L. B. Smith (Macbride 2920), spike X 1.
2 tne sepal X 1
: aura PENDULIFLORA L. B. Smith (Buchtien 7182), branch X 4.
ame, sepal X 1.
: ag erie L. B. Smith (Macbride 3436), flower and bract X 1.
P sepal X 1

a rx}
PWN ODI NO Oh whe

Me
oR

SF St oo bo

SMITH

SrREPTOCALYX Wi.uiaMsu L. B. Smith (L. Williams 2722), branch X 1.
TILLANDSIA BOLIVIENSIS nox (Bang 159a in part), inflorescence X 4.

TILLANDSIA EXSERTA Fern: aid (F. H. Lamb 381), spike X 1.
Same, partially united posterior sepals X 1.
TILLAN eae Harmstana L. B. Smith (Macbride 3272), inflorescence X \4.
— flowe he
Same, sepa al X
Cpe petal ° 4
PLATE V

me gested ALTODASERRAE L. B. Smith (L. B. Smith 1875), spike X yy

Same, flower x
ae RIESIA ‘Doseva L. B. Smith (Dusén 10712B), spike X 1.

ame,
VRIESIA eager; L. B. Smith (W. A. Schipp S82), inflorescence X Ys.
Same, cal
GUZMANIA CRATERIFLORA Mez & Wercklé (H. K. Svenson 322), expanded

coro

GuzMantA aprcuLata L. B. Smith (Macbride 5711), floral bract X 1.
Same, calyx X 1.
GUZMANIA PALLIDA L. B. ‘eae (H. H. Smith 2342), floral bract x 1.
Same, expanded cal
GUZMANIA MEGASTACHYA (Bak. ) Mez (Duss 3405, 3726), calyx X i.
Guzman1a Duss Mez (Duss 3326), spike X %.
Same, expanded corolla x 1.

PuaTe VI

a & Guzmanta Kruurrrana L. B. Smith (Killip & Smith 25630), spike X

Same, flower X 1.
UZMANIA Pennewiu L. B. Smith (Pennell 4344), branch X 1.
Guzmania Mitts L. B. Smith (Killip & Smith 20195), spike X 1.

Same, calyx X f
Map, showing relative r casscenmea of the Bromeliaceae as a whole and 0
the genus Guzman

Contris. Gray Hers. CXVIII. PiateE I,

Pigs. 1-3, Ancumea rerruainea L. B. Sm.; 4-6, AREGELIA conta
PETALA (Ule) Mez; 7-8, GREIGIA COLUMBIANA L. B. Sm.; 9 II, 1, Mac-
BRIDEANA L. B. Sm.; 12-14, Honenperaia Brirroniana L. B. Sm.

Contris. Gray Hers. CXVIII.

Y PRE

Puate II,

Plantae Brasilicases
APO on Dy: Fine.

C 7 ‘
Parana ae Ama Lemad , om puctat bus
ed ‘ : ¢. fot mh Ae

“4 bert poe

POADN

Dyckra Dusenu L. B. Sm.

Contris. Gray Hers. CXVIII.

PuLatre IIT,

l— of , Pare AIRNIA CaLDERONII Standley & Smith; 3-5, P. cuzcoEensis
L. is Sm, 5 , P. PECTINA TA L. B. Sm.; 8-9, P. rruncata L. B. Sm.; 10-11
Poya DE thie i B.S Sm.; 12-13, P. GRACILIS L. B. Sm.; 14- 16

LLATENSIs L, B. Sm.; 17-18, P. Macsrivet L. B. §

Contris. Gray Hers. CXVIII. Puate IV

. Figs. 1-2, . Sm.; 3-4, P. peNpuLIFLoRA L. B. Sm.;
m P. stiprrata L. B. Sm.; 7, Streprocatyx Wiuiiamsi L. B. Sm. ; 8-9,
| ZHANDSIA BOLIVIENSIS Bak.; 10-11, T. EXSERTA Fernald; 12-15, T. HarM-
SIANA L B. Sm.

Puya pectinata L. B
B.S

Puate V,

Conrris. Gray Hers. CXVIIT.

B. Sm.;
; 8-9,

MEGASTACHYA

Mez & Werklé

; +4, V. Dusenu L.
, G. paLuipa L. B. Sm.; 12, G.

Sm.; 7, GuzMANIA CRATERIFLORA

¢
~
2)
faa)
J
<)
< S
8 ~
2 =
—
Nh mM
< | w
Aro x
Cgdreb
& re
Phe ;
<j me
A<Qno
Lapa .
ny :
Ln | ies |
a
Hk
Pa
eee —
NR as
a ae 8)
am
Eat a
1» OS

Contris. Gray Hers. CXVIII. P VI
LATE :

‘ Y: 4 oy F ;
6 \ ™ 4}:
' tis, 2 ae ee $
’ w, at - - 4
’ Se ea La / ;
‘ . : f
tee ‘ a B s
imits of Bromeliaceae: ! ag Pal
oe SF 4
; a
lmite of Guzmania; —-~._. + if ly Le
Stations of Guzmania: e ; ae Prd
Definite station: —> vel tL es
z fot *
egional report: » ‘ / “7
UE 40
120° . : a a
110 100 i ar as i i f- h

Fi rs . ’ ie
5G iE AURMANLA Kinurerana L. B. Sm.; 3, G. Pennewut I. B. Sm.;
*MiTIs L. B. Sm.; 6, Distribution of family and genus. :

Reprinted from Ruopora, Vol, 34, May and June, 1932.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

XCIX,

BOTANICAL EVIDENCE OF POST-PLEISTOCENE MARINE
CONNECTION BETWEEN HUDSON BAY AND
THE ST. LAWRENCE BASIN.

By Davin Porter.

Dates oF Issuer
er es ee 4 May, 1932
Met 3 June, 1932

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY—NO. XCIX

BOTANICAL EVIDENCE OF A POST-PLEISTOCENE
MARINE CONNECTION BETWEEN HUDSON BAY AND
THE ST. LAWRENCE BASIN

Davin Porrer

INTRODUCTION

FLoristic studies about the region of Hudson Bay have suggested

“ases of discontinuous distribution are fairly common, but the agencies
Tesponsible for these phenomena are not always the same.

Itis the purpose of this paper to examine the various environmental
agencies which might effect the distribution of these maritime plants
and to determine, if possible, or at least to suggest, which factor has

Most important. The writer wishes to express his sincere
*PPreciation of the never-failing kindness and assistance of all mem-
of the staff of the Gray Herbarium. Particular thanks are due,
also, to Mr. C. M. Pomerat for his help in the field-work. Above all
S the writer indebted to Professor M. L. Fernald, whose inspiration
nd guidance have made this work possible.

70 Rhodora [May

Tue DisTrRIBUTION OF CERTAIN HALOPHYTIC PLANTS
OF THE Hupson Bay REGION

Of the numerous plants collected by the writer around the southern
region of Hudson Bay, during the summer of 1929, the following
species were found representing this peculiar distribution: Zan-
nichellia palustris L. var. major (Boenn.) Koch; Glawx maritima L.
var. obtusifolia Fernald; Juncus Gerardi Loisel.; Carex maritima O. F.

ueller; Carex norvegica Willd.; Carer glareosa Wahlenb. var. am-
phigena Fernald; Plantago juncoides Lam. var. decipiens (Barneoud)
Fernald; Poa eminens J.S. Presl and Scirpus rufus (Hudson) Schrad.

Associated with these plants were found the following indifferent
halophytes, whose occurrence in this region may or may not have
been brought about by the same agencies as effected the distribution
of the above-mentioned halophytes: Potamogeton filiformis Pers.;
Triglochin maritima L.; Triglochin palustris L.; Scirpus americanus
Pers.; Juncus balticus Willd. var. littoralis Engelm.; Potentilla An-
serina L.; Myriophyllum ezxalbescens Fernald; Lathyrus maritimus
(L.) Bigel.; Arenaria peploides L. and Mertensia maritima (L.) S. F.
Gray. To the above list should be added Bidens hyperborea Greene,
an estuarian species, and Zostera marina L., a plant confined strictly
to salt water. Maps 1-6 give the geographic ranges! of the strict

The geographic ranges are based w ens in the Gray Her snd
the Herbarium of the New England Boabiiical Club, and upon the naaide ng. publica-

Alcock, F. J., List of Plants Collected Along the Churchill River between Missi Falls
: raed Mouth of the Little Churchill River. Geol. Surv. Can., Summ. Rep. for
Bell, John, On the Plants of Manitoulin Island. Geol. Surv. Can., Rep. of Prog.,

Bell, Robert, Report on Exploration of the Churchill and Nelson Rivers and around
s and Island 1. Surv. Can., Rep. for 1878-79, pp. 1-72 C.

Bell, ll, Robert, List of Plants Callout 1880, Geol. Surv. Can., Rep. for 1879-80, P-

obert, List of _—s Collected in Hudson Straits. Geol. Surv. Can., AnP-

Rept. xi. = PP. 34-37 M.

Dowling, D , List of Plants Collected at the Mouth of the Ekwan and Albany
Rivers. cae pga Can., Ann. arse xiv. pt. F, p tions to
+ The Distributi of Plants in hater in Some 0 161-177.

md, A ts Rela
Physical ee bona Geological Conditions. Can. Nat., n. s. res (1860) pp.
Drummond, an Argument for the

anadian Plan
Marine Origin of the Erie Clays. Can. Nat., n. s. vii. qee pp. 217-223. inlet,
Fernald, M. L., The Botanical Evidence of Marine Conditions in Hamilton

Repor sede ada and the Colony of Newfoundland in the Labrado -
of the Lords of the ‘Judicial Committee of the Privy Coun cil, delivered

, 1927, London, ° :
Henderson, A. , Agricultural Resources of Abitibi. Rept. Bur. Mines, Ont., xiv. pt. 1,

1932} Potter,—Post-Pleistocene Marine Connection 71

ae a =
gan = . 4 , " oe. ee om

Map 1, Geographic Range of Carex norvegica;
2, of Zannichellia palustris, var. major.

ry
i i ins 4
W amen. i Z
{TBs ay
iM ase = ae Baas
\ a i UW » s
ES oy jase ee oe
agg ~~ aa
Po ane j je BS (Re
Hay SES }
a ae | ; '
tre
aes r - ie we
} H
! i

Rhodora

Nc
.
4
/
}
% SRY
.
ant

Map 3, Geographic Range of Poa eminens;
|4 (lower), of Glau maritima, var. obtusifolia.

{May

.
oe"
*

1932]

\ ‘ .
‘ ‘ ‘a
"
\ ; 2 |i (th
‘ i f
H i
: !
i |
A .
§ i
oo ia ;
«
Vie \ ae % iH g
1) Bae Ss Ge
iM Bi sae pan PE A
a} ;
re Hi - c iS
us. Nef : a ‘4
‘Pee .
} Ny dane Bet EMA Meg | Secs DN ‘
? j H sm,
\ ; heey are: Ww
; j aE :
3 / 7 :
i fl +
E eet ers eee meee | , i
~ | Bee! ater i ee aE:
“> L Es f ee ee
“as : H :
: H i H ‘
si 7 i i i \
a | . tr
\
\ H
j
er as i
ae ae
~~ + f° 3
‘ H :
\ H di
a 03
\ i 2 i ~
, i
i
i i
;
i
es C
wl
\ See Soe = g
ees, 7 é
ff ~%. =—_
; ee
al — Jo
ALS fe
j .
bpd Fe
>
' To ; H
aa Gy Be f
A ! reais 2 walle
Tae LY
‘ ! , maaehveas
. j
a . | i '
ae se fan Se
Z {Pte © Sexi ©
‘“ i 88 gees que
a, ' 7 :
= i : at
aa i f Tr
“) [hte Pe \ \
; i
4 H - fo“)
\
o la
i

Potter,—Post-Pleistocene Marine Connection

‘
a
ee
20

Aon 8

res

Map 5, Geographic Range of Carex maritima;

6, of Scirpus rufus.

wt
peoo*’

74 Rhodora [May

halophyteslistedabove. In addition to their occurrence in the Hudson
Bay region, it will be noted that in the cases of Carex norvegica (MAP
1) and Zannichellia palustris var. major (MAP 2) they are restricted
to the Atlantic seaboard, extending at least from southern Labrador
south along the New England coast, and in the case of Zannichellia
palustris var. major south to Florida. Poa eminens (Map 3) and
Glaux maritima var. obtusifolia (Map 4) have northeastern ranges
somewhat similar to those mentioned above, but they appear also on
the Pacific coast. The distribution of Carex maritima (maP 5),
Scirpus rufus (MAP 6) and Juncus Gerardi (Map 7) varies from the
strictly maritime, in that the two former species occur in the region
_ just north and west of Lake Winnipeg, while the latter has been
reported from the Finger Lakes district of New York and at the
southern tip of Lake Michigan, as well as from the Pacific coast.

In the cases of Plantago juncoides var. decipiens (Map 8) and
Carex glareosa var. amphigena (Map 9), both found farther north
along the Labrador coast, they occur also on the southwestern coast
of Greenland. Zostera marina (Map 10) occurs from Labrador and
the lower St. Lawrence south to North Carolina, in James Bay,
along the north Pacific coast and on the southwestern coast of Green-
land.

Bidens hyperborea (Map 11), an estuarian species, has been studied
by Fassett (29) and Fernald (31) and the data regarding its distri-
bution are taken from their p papers. It is interesting to note that this
plant occurs exclusively on fresh tidal mud in estuaries from eastert
Massachusetts to Quebec, then j jumps to Rupert River, James Bay.

greets A a a ig of Canadian Plants. Geol. and Nat. Hist. Surv. Can., 1883-
acoun, 5. ene of Plant Distribution. Geol. Surv. Can., Rept. of Prog.
1875-1876.

Macoun, J. M., List of Plants in Low's Report of the Mistassini Expedition. Geol.
and Nat. Hist. Surv. Can., Ann. Rept., 1885, p a

Macoun, J. M., List of Plants in Low’s nv gs on Exploration in James Bay 42
Country East of Hudson Bay. Geol. Surv. Can., Ann. Rept., iii. Pncas pt. J. +

Macoun, J. M., List of Plants in Low’s Report on “fonnas Geol. Surv. Can., Rep

Richardson, Jas., Report on ag Country North of Lake St. John. Geol. Surv. Cam
Rept. of Prog., 1870-71, p
St. John, H., A Botanical Pesos of the North Shore of the Gulf of St. Lawrenc®
Can. Dept. Mem. No. 126, 1922.
Svenson, H. K., Studies on Interior ovwiiztaaa of Maritime Plants. I. R#op0R*
05-11

Tyrell, J, B., Pianite Coliseted Begwaen, ca Athabasca and the West Coast of Hud-
n Bay, north shore of Hudson Straits and Fort Churchill. Geol. Surv. Can., AD
Rept. pod eg D. 205 F Red
B, Geographic Limits of oasis of Plants in the Basin of the
River z pri North. Proc, Boston Soc. Nat. Hist. xxv. (1890) pp. 140-172.

} re
SF Gen Y i
ip OF fe oe ee ;
mS
oe a
saat ae .
f id j BN et
4 !
~ ? i ren Be
‘ | ee RRR spon, en sll Sega Bh SOAS
; |
ie p. z i. a
98 imine oe
‘i i oy rt ~

Map 7, Geographic Range of Juncus Gerardi;
8, of Plantago juncoides, var. decipiens.

.
y
.,
‘a

CLOILA / Wy BY
Ng 7 SALE!
he Za (3 are Nie,
’

XK Mi a IB .
d fs i, (5) a r\ on q
[Nee rae

<2

e 5
2 iT $e ; s) ,
Sa lg 4 \A AS
\ " : 4s / Uy
“] q ee I 7) yf
L . : aw Wp le
ici hae y
; Se ee ee oe:
e of Carex gl . amphigena
*

“J

~J

~

‘y memeen ged.
i Fad a
w : Ew 10;
me i H

Oe Bg Oy Hee EA EE Se ee, Res eee Teer Pein Tabi “See = aetinee Rea

soe? A ee av.
\ 17 \
\ i
\ i 2 i ‘
' i — Y a
: Boe.
i | R
y . i )
i * K
Hg Yas?
fa. — g 4
Wat m
iM 22 em tee c
1
\ :
~. 7 poof} ra é

tm.

7 a h

Map 10, aeteae Range of Zostera marina;
of Bidens hyperborea.

78 Rhodora [May

In no case have these plants been reported as growing along the
shores of Hudson Straits or along the northernmost Labrador coast.
Four of them have been rarely found north of Hamilton Inlet; two
reach their northern limits in Hamilton Inlet; and five are not known
north of the Straits of Belle Isle.

Several expeditions! have recently gone to Labrador and, as each
expedition has had among its members a trained botanist, whose
work was to collect and study the flora at points touched, it is reason-
able to assume that, since most of these plants have not been reported
from far north on the coast, they do not usually occur north of
the southern corner of Labrador (except as above stated); and their
appearance along the shores of Hudson and James Bays has not been
due to a migration along the outer coast of Labrador, thence by
way of Hudson Straits to James Bay. Again, in no case have these
plants been reported from the region between James Bay and the St.
Lawrence River, the most inland point for any of them (Bidens
hyperborea) along the St. Lawrence being northeast of Montreal.
The absence of these plants from inland areas is to be expected, since
they are strictly or primarily maritime, and it indicates that the
edaphic conditions of the land lying between the St. Lawrence Basin
and James Bay are at this time such as not to favor the presence of
halophytic species. It is true, however, that much of the country
lying north and northwest of the St. Lawrence River has not been
carefully botanized and future exploration in this area may possibly
reveal the existence of some of these species. If such should prove t
be the case, the present problem would be more easily solved as will
be later pointed out.

PosstsLE Causes oF THIS DiscontTINUoUS DISTRIBUTION

The agencies which might have brought about this peculiar dis
tribution are as follows:

1Th

ports from some of the recent expeditions referred to are as follows:
Fernald, M. L., A Botanical Expedition to Newfoundland and Southern Lace
Ruopora, xiii. Ppp. 109-157 (1911).
Fernald, M. L. and Sornborger, J. D., Some Recent Additions to the Labrador Flom
@ Naturalist, xiii. pp. 89-107 (1899).
Bishop, Harlow, List of Plants Soliecest i the Austin Labrador Expedition, sn
Unpublished list at the Gray Herba: : vak,
Delabarre, E. B., Botanical Report of the Brown-Harvard Expedition to =
Labrador. Bull. Geog. Soc. Phila. iii. pp. 167-201 (1902). prador,
Wetmore, R. H., Plants of the Hamilton Inlet and Lake Melville Region, os
Ruopora, xxv, pp. 4~12 (1923), orthera
Woodworth, R. H., List of Plants Collected on the Iselin Expedition to ”
Labrador, 1926. Unpublished list at the Gray Herbarium.

1932] Potter,—Post-Pleistocene Marine Connection 79

1. Driftless areas which may have harbored these plants during
the Wisconsin glaciation.

2. Dispersal of these plants since the recession of the last ice sheet:

a. by means of wind.

b. by means of animals, other than birds.
c. by means of birds.

d. by means of water.

3. Migration of these plants along the shores of a marine connection
between the St. Lawrence Basin and James Bay.

These agencies will now be considered in detail.

1. Driftless Areas. That driftless areas within the broad region
invaded by Pleistocene ice occur is acknowledged by most geologists.
The possibility of these areas harboring plants during the period of
glaciation is unquestionable, for not only is an analagous condition
now prevailing in the case of Greenland with its continental glacier,
but, in addition, Fernald (30), in his remarkable studies upon the
flora of Newfoundland and the Gaspé Peninsula, has definitely
proved that in these regions such was the case. This fact is conceded
by the great Canadian Pleistocene geologist, A. P. Coleman (19).

Many explorations have been made by geologists of the Canadian
Government in the Northeastern part of Canada and the findings
of such men as Bell, Low, Chalmers, Coleman, Tyrell and others
largely agree concerning the glaciation of this area. Nowhere along
the coasts or on the islands of either James Bay or southern Hudson
Bay have areas been discovered which escaped the ruthless work of
the Wisconsin ice sheet. The writer visited the region of James
Bay during the summer of 1929 and found evidence of glaciation at
every point touched: all along the rivers entering James Bay and along
the southern and eastern coasts of the Bay itself. The same conditions
were found on Charlton Island as also on a number of smaller islands
which were explored. Mr. A. E. Porsild, one of the botanists for
the Canadian Government, visited a number of the islands in James
Bay, including Agamaski, the Twins and one large island off the
Coast at Fort George, and reported verbally to the writer that all
these islands had been severely scoured by ice. In discussing Strut-
ton and Trodely islands, Low (41) speaks of many large boulders
strewn over the land.

Thus, the evidence is quite conclusive that driftless areas do not
exist in the neighborhood of James Bay, and some other explanation

80 Rhodora [May

must be sought to account for the presence of the southern halophytic
plants on the shores of this great inland sea.

2a. Wind. Wind plays an important role in the dissemination
of plants and plant parts. Seeds, or even plants, are often caught up
by winds and carried varying distances before they are dropped or
before they strike some obstacle which impedes their progress. Warm-
ing (64) claims that relatively heavy fruits may be carried by wind up
to distances of at least sixteen miles. Many plants have their seeds
modified for dispersal by this means and in the case of the dandelion,
Small (51) claims“ . . . that so long as the relative humidity
of the air remains above 0.77 and so long as the fruit does not en-
counter an obstacle, a horizontal wind of 1.97 m.p.h. is sufficient for
its dispersal to any distance.” It is quite conceivable that a seed
having once been dropped can again be picked up by wind and this
process be repeated over and over again until long distances have
been traversed. Again it is possible that a seed falling upon favorable
soil might germinate and in succeeding years its seeds be blown to
a new region and this process continue until long distances be crossed.
There are, however, many elements of chance in such a hazardous
mode of transportation, and the mortality in terms of successful
germination and maturity, would of necessity be tremendous. 10
the case of plants establishing intermediate stations between two
points it is difficult to apply this method of migration to halophytes
over areas not favorable to salt-loving plants.

Fernald (30), in his studies upon the flora of Newfoundland, has
furnished us a most striking example in point. To quote Professor
Fernald, speaking of the absence from southwestern Newfoundland
of many wind-dispersed species of Cape Breton: “ . - - the
distance across Cabot Strait, the shortest route from the southwestern
mainland to Newfoundland is fully 70 miles, and, although this does
not seem a forbidding gap, the fact remains that very many common
Canadian species with fine spores or with the seeds plumose, feathery
or otherwise adapted for wind-transportation, have failed to cross
from Cape Breton to southwestern Newfoundland.” Svenson (56),
in his studies upon the interior distribution of halophytes, reached
the conclusion that wind was not a predominant factor in Z
dispersal. Thus, the influence of wind as a primary factor in seed-
dissemination (especially in the case of halophytes) over long distances
has, perhaps, been overestimated and a more plausible cause for the
occurrence of these plants in the Hudson Bay region must be sought.

1932] Potter,—Post-Pleistocene Marine Connection 81

2b. Animals (exclusive of birds). Animals other than birds are
responsible for local dissemination of plants and plant-parts. Those
animals with which we need to concern ourselves are restricted to
the mammals, for probably fish, amphibia, and reptiles play a very
minor part in carrying seeds or plant-parts over such distances as
are concerned in the present problem. Many plants have their seeds
or fruits so modified as to allow them to cling to the fur of mammals
and thus be carried along. In the species of plants under consideration
no such modifications are present and the possibility of the seeds
of such plants as Glawx maritima var. obtusifolia and Zannichellia
palustris var. major being transported in this manner is very slight.
Human travel has not been great between these two regions and we
can scarcely look to this source as the means of introduction. Had
these plants any food value or other economic significance, human
migrations would of necessity have to be investigated.

2c. Birds. With birds, as with wind, there is no question but that
they effect seed-dispersal locally. Concerning the distribution of
plants or plant parts over long distances by birds, the question is
highly debatable. Warming (63) sums up this matter in referring to

ud Anderson’s work, supplemented by the conclusions of Winge,
as follows: “For a number of consecutive years, thousands of birds,
picked up dead at the Danish lighthouses have been sent to the
Zoological Museum at Copenhagen, and notes on these dead birds
have for many years been published annually by H. Winge. :

is eminent zoologist writes to me,” continues Warming, “as
follows: ‘In one of the first years, the contents of the stomachs were
systematically examined, later on only occasionally, but the stomach
as always proved to be empty. . . . Though I have had
thousands of dead migratory birds between my hands and have made
a habit of examining every single one, I have not as yet found any
seeds adhering to the feathers, beaks, or feet.’ ”

Commenting upon the above observations, Warming continues:
i the above observations are made by so careful and eminent an
Mvestigator, I must consequently believe that birds at least very
seldom carry seeds and other larger reproductive organs, and
small plants, across great distances, and the indisputable evidences
of birds carrying seeds either in them or adhering to them mentioned
in books evidently apply to birds shot at or not far from, their daily
haunts, and not to such as have just made a long journey.”” This

82 Rhodora [May

conclusion seems to be wholly sound and again forces us to seek a
better explanation.

2d. Water. If the plants in question spread from the Gulf of St.
Lawrence after the land to the north and northwest became freed
from its burden of ice, only two now existing avenues of water were
open for dispersal of seeds; namely, either the marine water along
the coast of Labrador and thence through Hudson Straits, or the
fresh water ponds and streams lying to the northwest of the St.
Lawrence. As to the route via the North Atlantic and Hudson
Straits, it seems wholly improbable that the dispersal took place by
this route, for Guppy (34) has shown that only a few seeds have
sufficient buoyancy to keep afloat in ocean-drift for more than a few
days and that in the cooler regions of the globe, seed-drift is usually
very scanty. Had seeds been carried short distances along the coast
by this agency and in favorable spots established themselves, and
had their progeny been transported in the same manner a short
distance farther north, there establishing themselves and so on until
the goal had been reached, it would be expected that more traces of
the parent stock would be found along the outer coast of Labrador.
As has been pointed out, no trace of some of these plants has ever
been found north of the Straits of Belle Isle, while others are unknown
north of Hamilton Inlet; and the few northern stations of the re-
maining species are exceeding localized.

Separating the St. Lawrence Basin and James Bay there is 4
divide which results in a double drainage system for this interior
region. A large portion of this area drains to the south into the St.
Lawrence River, while the remainder is drained to the north. After
the recession of the Wisconsin ice-sheet, all the territory in north-
eastern America stood at a much lower level than at present; yet it
is possible that the divide existed in approximately the same relative
proportions as at present, and if so, it is difficult to understand how
water flowing south could be of assistance in the dispersal of seeds
towards the north. If this intermediate water was fresh, migration
for any great distance along its shores for strictly halophytic plan's
would be improbable. On the other hand, if the divide separating
these two regions was much lower in proportion to the two adjacent
areas, our problem would be simplified, for, as is soon to be pol
out, such a low basin would allow a marine connection betwee? the
two areas and thus permit migration of the plants in question along
the shores of this inland sea.

1932] Potter,—Post-Pleistocene Marine Connection 83

Since the factors of driftless areas, dispersal by means of wind,
animals, and water do not solve the present problem of seed dispersal,
it is necessary to consider one more agency which, if proved to have
existed, will help us out of the present dilemma. This last is the
possibility of migration along the shores of a marine connection
between the St. Lawrence Basin and James Bay.

THE CHAMPLAIN SUBMERGENCE

After the recession of the last ice sheet, as just noted, the north-
eastern portion of the North American Continent lay at a much
lower level than is the ae at present. In support of this statement
Taylor (61) says that: ; . nothing within the realm of
Pleistocene geology is more cheegiy established than the depression
of land during the growth of the Wisconsin ice sheet in the region
of the Great Lakes. . . . ” This lowering of the Continental
land mass resulted in the drowning of the St. Lawrence and Ottawa
valleys. Chalmers, Dyer, Kindle, Bell, Low, Keele, Coleman, Taylor,
Dowling, Tyrell and others all agree that such a submergence took
place. Additional evidence of a marine invasion is supplied by the
presence in these formerly submerged areas of large deposits of
marine clays, many of which bear numerous fossils, such as: Mya
arenaria Linn., Mya truncata Linn., Saxicava arctica Linn., Macoma
calearea Gmel., Macoma balthica Linn., var. groenlandica Beck.,
~ edulis Linn., Serripes groenlandicus Gmel., Leda pernula

.. localities have been reported for these marine shells and the
following Taste 1 gives a selected list, including only those which
show the highest elevation. The localities listed in TABLE 1 have been
Plotted on the accompanying map and are indicated by solid black
circles, together with the elevation in feet above sea level. These
Points have been connected and the enclosed area filled in by vertical

es. Thus, a clearer conception may be gained of the extent of
this so-called Champlain Submergence as based upon the actual
finding of marine fossils.

An interesting point to note in this connection is that these fossils
are not evenly distributed throughout the known region of sub-
Mergence. Many areas occur at much lower altitudes than those
given in the above list, which are barren as concerns fossil shells.
Chalmers (14) has pointed out that no marine shells have been found

| iii oi :

i i i ee
alll (I il he iil I Wes

| | if ul | i

a |

}

M tase comet sen ome
———

ra

athal | =

9 ’ a — =

\ —

f

+8

eiop ouy

AYN |

Se OO ew eh Ne em we 6 ele ee oe ae ae oa ele

WO ee
Glen ie Counts, (ot. 32
piace ae ee

C0 & ieee Wee Bows ee we «

4 miles above Ha Ha Bay on River Ona-
DenChDAeNINS. |.

C20 SOR RA AO e Oe eee 6 8 6 ee a ee eo ee

TABLE |
MARINE DEPOSITS

Approx.

x
tude Longitude
0

Height in
feet above
sea level

320-350
440

marine shells

marine shells

marine shells

marine shells

marine shells
marine shells

marine shells

marine shells
marine shells

Reinecke (48)

Richardson (49)
Ells (23
Wilson (68)

Coleman (20)
Chalmers (17)
Richardson (49)
Richardson (50)
Chalmers (15)
Low (42)

Low (42)
Chalmers (14)

[2861

WOoTj,o9UUO’) dUlLLe yy 9U900}SI9[ J-}SOg—‘19}} 0g

S8

86 Rhodora [May

in the beds representing Leda clays and Saxicava sands of New
Brunswick. He goes on to say that only a few localities on the west
et of Prince Edward Island have Pleistocene marine fossils:

though a large area now forming dry land must have been
satel the sea, which was doubtless then, as at present, inhabited by
marine anima

Again Ells (24), in speaking of the lower Ottawa and St. Lawrence
valleys, says: ; . throughout the district there are great
expanses of marine dave which show no trace of organisms, while
very often the overlying sands and gravels hold great quantities of
marine fossils.” Thus it may be safely concluded that the absence
of marine fossils does not necessarily mean the absence of marine
waters.

The extent of this post-pleistocene marine submergence has never
been definitely shown and at present there is much disagreement
among geologists concerning the matter. The presence of marine
fossils is reliable evidence as to depth of submergence, but, in addition
to this, many raised beaches have been discovered and described.
Whether these are of marine or lacustrine origin is still a moot question.

TABLE 2 gives a list of some of these raised beaches together with
their approximate location and their altitudes.

a be ae eed

a i Rie Re PES NNER Nr a EEE Te © See ee eta

i a a ai Ty

1932] Potter,—Post-Pleistocene Marine Connection 87

TABLE 2

RAISED BEACHES—-PROBABLY MARINE

Height in
Approx. Approx. feet above
Location Latitude Longitude _ sea level Authority

ee ee 43° 40’ So 1B). 262 Fairchild (28)
CN 5. oe 43° 50’ (6-1. 320 Fairchild (28)
Me ek 44° 15’ (6° 0” 400 Fairchild (28)
SS IR es ee 44° 25 16° 50’ 450 ather (44)
Tair os 44° 25’ fi 0 © B10 Mather (44)
Ce ee ee Agr 74° 45’ 701 Wilson (66)
MEMWOY 56... eee 46° 0’ 74° 45’ 663 ilson (66)
eee 45° 30’ 75° 40’ 90 Mather (44)
De ee 44° 10’ C20! Seo Taylor (57)
ee re 44° 0’ 17° 45’ 309 Taylor (57)
ky 44° 10’ 77° 30' 320 aylor
Baie St. Paul.......... 47° 30’ 70° 30’ = 500 Mandsley (43)
Sweetsburg............ 45°30! |... 73° 40 10- halmers (13)
Frelighsburg........... 45° 0’ 72°50’ 475-785 Chalmers (13)
Shefford Mt 45° 25’ 72°30" 20 halmers (13)
De A haties....) so 2. bucks 46° 50’ 71° 0’ 540-550 Chalmers (12)
Renal 46° 40’ rie Of 20 halmers (12)
St. Henedine........... 46° 30’ Ao Oo 70 halmers (12, 16)

ee 46° 25’ 1 60 Chalmers (12)
St. Joseph Lake........ 46° 10’ 71° 30’ = 860 Chalmers (12)
mde ata... 46° 20’ M1” 40 «= - 805 Chalmers (12)

MN re ey 230! PT? 30%: -. S50 Chalm
Lake Memphremagog... 45° 20’ 72° 0’ 865-990 Chalmers (12, 14)
mete. ° 40’ 74° 20’ 1000 lls (24)

wiston, N. Y.... 2310! 9°. 0’... 386 Spencer (53)
Watertown, N. Y....... 44° 0’ 15° 60’ 730 Spencer (53)
A SERS ee 44° 15/ 75°10’ 972 Spen
Malone, N. Y......:... © 50’ 74° 15’ 1100-1200 Chalmers (11)
St. Anne de Beaupré 47° 0’ a 0 Chalmers (11)
North of Quebec. ...... ° 50’ 71°10’ 560 Chalmers (11)
Lake Maskinong....... 46° 20’ 73° 30’ = 560 alme
MCs a 45° 45’ 74° 0’ 900 Chalmers (11)

gsmere Mt. (North of

Ottawa)..... ‘ peru 45° 30’. 76° 46' 965 Chalmers (11)
North side of Ottawa

River just above Allu- :

Teebte 16. 46° 50’ 76° 50°... 800 Chalmers (11)
Brulé Rapid........... 46° 30’ 76° 0’ 700 Keele (38)
Georgeville............ "10° 72 a0 Spencer (53)
Magoon Point... ...... ‘a? ar O80 ge ed (53)
ee 45°60’ 72° 0 800 halmers (16)

It will be noted from the list in TABLE 2 that these raised beaches
Tange from a few feet above sea level to 1200 feet. Their localities
are indicated on the accompanying map by hollow circles, with
their altitudes. These points have been connected and the area
included filled in by horizontal lines.

Marine fossils have not been found in the Champlain, St. Lawrence,
and Ottawa regions above an altitude of approximately 735 feet.

88 Rhodora [May

This fact has led some geologists to search for some theory other than
marine invasion to account for the non-fossiliferous raised beaches
The theory of Glacial Dams has received considerable support. In
discussing the Iroquois beach as having been formed by an arm of
the sea, Coleman (20) states, “This conclusion is a very natural
one and tends toward simplicity by avoiding the assumption of an
ice dam; but the finding of fresh water shells in the Iroquois beach
near Toronto seems conclusive as to the character of the water,
which could hardly remain fresh or even brackish with an opening
seventy or eighty miles wide and four hundred feet deep into the
inland sea formed by the enlarged Gulf of St. Lawrence.” Spencer
(54) is opposed to this view and he states: “As we ascend to the
elevation of the higher beaches, the question of glacial dams becomes
more and more difficult, for we must assume them to have been
hundreds of miles long and at enormous altitudes, damming up
waters which had the proportions of inland seas.” He then goes on to
say, “I am compelled to assume the initial plane of the Algonquin
beach to be at sea level.” Later in the same paper Spencer says,
Shs there is additional evidence; for crustaceans of marine
species have so adapted themselves as to still live in the depths of
Lake Superior (55, 52), as also maritime plants along its shores.” (36)

Conceding the existence of the glacial dams, which would account for
the higher beaches to the north of the Great Lakes, they would not
account for the occurrence of the raised beaches found along the
valley of the St. Lawrence River. As Chalmers (13) has stated, “ All
the shore-lines noted face the open plain of the St. Lawrence valley
- + + + No barriers exist or could have existed, capable of
holding in a body of fresh water at heights sufficient to allow the
formation of these shore-lines; and the only reasonable theory as to
their origin seems to be that they were formed along the margin of
a sea which occupied the St. Lawrence valley in the Pleistocene
period.” Dr. Ells (24) agrees with Chalmers, for he says “The
underlying clays at the higher levels up to nearly one thousand
feet, and in places to a height of considerably more than this, are
apparently continuous with those of undoubted marine origin to the
north and east, and the inference naturally follows that all the de-
posits were laid down by the same agencies. The absence of marine
organisms over a large part of the area is only negative evidence
to the contrary, and there are certain facts that go far to estab

jr
q
q
:
.

|
j

1932] Potter,—Post-Pleistocene Marine Connection 89

this theory of their marine deposition.” In the same article, Dr.
Ells continues: “There is no break apparent in the deposition of
these beds, from the nodule-bearing clays near Ottawa to the most
northerly outcrops on the Gatineau and the Liavre where the clays
are, in so far as yet known, all barren.”

The question might well be asked as to the relative ages of the
different beaches under discussion, and so far as known they seem
to be all of post-glacial origin. Taylor (58) has summed up this
matter as follows: “I express again, and with increased confidence,
the belief that the Iroquois beach and the highest beaches in the
lower St. Lawrence, Champlain, Huron, and Ottawa valleys, and in
the basins of Lakes Huron, Michigan, and Superior, and also in the
valley of the Red river of the North, are all one continuous shore
line of the sea.”

If these conclusions of Spencer, Ells and Taylor be correct, it
would follow that this marine invasion just after the recession of the
last ice-sheet was much greater in extent than is indicated by the
altitudes at which marine fossils have been found. The same factors
which are responsible for the lack of marine fossils in areas unquestion-
ably once covered by the sea, as was pointed out earlier in this paper,
may account for the lack of fossils in these raised beaches. Chalmers
(14) has suggested “That the arenaceous clays and the sand which
together constitute the stratified deposits lying below the uppermost
shore line of the Pleistocene submergence, contain minerals destructive
to shells and tests of marine animals.

Finally, in James Bay today no living marine mollusks are to be
found, except perhaps in the northern part, owing probably to the
muddy and brackish nature of the water (9). A similar condition
might well have existed during the Champlain Submergence in which

€ waters were muddy and brackish, not permitting the existence
of marine life, yet allowing the existence of plants of maritime habitats
along its shores.

If, then, these raised beaches owe their origin to marine water (and
there seems to be considerable evidence in support of this hypothesis)
the invasion by the sea would have reached the Lake Temiskaming
tegion, involving also Lakes Ontario and Nipissing. Reference to
the accompanying large map will show the area involved, which is
indicated by horizontal lines.

(To be continued)

BOTANICAL EVIDENCE OF A POST-PLEISTOCENE
MARINE CONNECTION BETWEEN HUDSON BAY AND
THE ST. LAWRENCE BASIN

Davin Porrer
(Continued from p. 89)
Marine SUBMERGENCE IN THE Hupson Bay ReEGIon

Bell, Low, Wilson, Tyrell and others have conclusively proven
that a post-glacial submergence occurred in the Hudson Bay region,
which covered a vast territory to the east, west, and south. This
area is covered by a mantle of marine clays containing many fossils
which are identical with those reported from the St. Lawrence Basin.

TABLE 3 gives a list of some of the localities where these marine
fossils have been found. The stations of TABLE 3 have been plotted
on the accompanying Jarge map and are indicated by solid black
circles. Based upon the known fossiliferous areas, the extent of the
marine invasion from James Bay has been shown to reach nearly to
that of the St. Lawrence Basin. The area indicated is probably not

102 Rhodora [JUNE

TABLE 3
Height in feet
Location above sea level Authority
Round Bay, 125 miles from the
mouth of Moose River.......... 300 marine shells Bell (5)
Missinabi Lake and along Missinabi
gna Moose Rivers... .........: 300 marine shells Bell (8)
McLean (46)
Nelson river, 54 miles upstream.... 200 marine shells Bell (3, 4)
Nelson river, above the third Lime-
BEONB TADICE es ee osc 200 marine shells Bell (4)
Churchill river, 60 miles from its
ee marine shells M (3)
enog LSE ARN SN OER i tee 450-500 marineshells Bell (6, 7)
Little Churchill river............. ine shells Alcoc
PP os is 350 marine shells McInnes (45)
Mammamemmatawa............. 380 marine silts | Williams (65)

inclusive enough to show the outer margin of submergence, for Low
(41), in his work along the east coast of James Bay, discovered
sediments and terraces probably of marine origin up to at least 675
eet. In discussing this situation he states: “The evidence of strati-
fied deposits of marine sands and clays along the valleys, near the
mouths of the rivers on the east side of Hudson Bay, show that a
subsidence of the land over 500 feet (and probably 700) took place
after the period of glaciation.” Bell (8) further supports this con-
tention, for he says: “On the islands and shores all along the Eastmain
coast, the ‘raised’ beaches are very conspicuous at all heights up to
about 300 feet immediately near the sea, but, no doubt, higher ones
would be found further inland.”

The problem of finding fossil-bearing clays and sands in this region
and farther southward is a difficult one due to the ground-cover,
which is for the most part boggy or heavily wooded, with considerable
accumulation of peat. Even though much of this area has been
explored by Canadian geologists, their work has been for the most
part concerned with the economic aspects of geology and Pleistocene
phenomena have not been thoroughly investigated. Thus, wi the
difficulties of exploration coupled with lack of interest, it is not
surprising that relatively few localities of fossiliferous clays oF sands
have been reported. For the majority of places given in TABLE 3, _—
are indebted to the early Canadian geologist, Bell, who, in speaking
of the Churchill River says (3): “As the bank continued with the
same characteristics for a long distance upstream, I have no doubt
that shells may be found at a greater distance inland than that at
which they were observed by myself.”

1932] Potter,—Post-Pleistocene Marine Connection 103

It would seem, therefore, justifiable to extend the southern boundary
of the marine invasion from the head of Hudson Bay, at least to
within a comparatively short distance from the divide. This has
been done on the map representing this submergence and the area
involved indicated by horizontal lines.

Discussion OF THE AREA Across THE HeIcut oF LAND
BETWEEN JAMES Bay AND THE OTTAWA VALLEY

Reference to the accompanying map will show that the region
under discussion includes the Lake Temiskaming and Lake Abitibi
districts, the two areas separated by the divide, the lowest point
of which is 914 feet! above sea level and occurs just southwest of
Lake Abitibi. Examination of the superficial covering of this inter-
mediate area has shown that the entire region up to altitudes of
1000 feet (35) is overlain by clays which were water-laid, for Coleman
(21) has shown that this region was once covered by a vast body of
water, which he designated as Lake Ojibway.

On the large map certain elevations are shown which are based
upon Fairchild’s isobases and which, according to him, are theoretical
elevations of marine waters (26, 27). These localities are represented
by solid black triangles. TABLE 4 lists the above regions together
with their altitudes and approximate locations.

TABLE 4
; Approx. Height in feet
Location it Pace Longivade stots & sea level
ES ee aes 48° 30’ 72° 15’ 1000
et see lives 47° 0! 72° 0’ 970
Jacques Cattior River i Oo 47° 0! 71° 30’ 950
Saint Lawrence Valley, Quebec. 46° 50’ 71° 10’ 925
Montmorency, River Laval... .... 47° 0! 1. 940
Dn Tichee 47° 10’ 70° 50’ 925
Mes ke 45° 30° 73° 50’ 840
North River, Sainte ee an 46° 0’ 74° 0’
West River, La Chute............. 45° 45’ 74° 30’ 825
SS ee a 45° 30’ 75° 50’
SN ee aie 46° 30’ 78° 50’ 700
en le re ae 46° 20’ 79° 30’ 675

The 700-foot isobase extends from the Lake Temiskaming region
Southeastward along the Ottawa River and passes through Ottawa
city. Since marine fossils have been found in the Grenville District,
Ontario, up to 735 feet above sea-level (22), it must follow, if Fair-

‘Elevation taken from “Altitudes in Canada,” James White, Commission of Con-
Servation, Canada.

104 Rhodora | JUNE

child’s isobase for 700 is correct, that the marine waters extended
up to and included Lake Temiskaming, since at the present time the
altitude of the lake is about 590 feet above sea level. Again, the
topography of the Lake Temiskaming region is such as to make this
highly probable. The area lies wholly within the Laurentian Plateau
and may be divided into three sections: (1), rocky uplands; (2),
clay belt; and (3), linear valleys. The linear valleys have been
brought about by erosion along planes of faulting, and are probably
pre-glacial, for as Wilson says:“ . . . they cannot possibly be
of post-Glacial origin because stream dissection since the Glacial
epoch has been almost insignificant and the valleys are themselves
occupied by glacial drift deposited by the ice-sheets. It is also very
improbable that they are the result of glacial denudation, for they
have no relationship to the character of the rocks they traverse, and
they trend, in some cases, at right angles to the direction of ice move-
ment. Since the valleys are neither Glacial nor post-Glacial in their
origin, it follows, a priori, that they are pre-glacial valleys.” The
depression occupied by Lake Temiskaming and the Ottawa river
between the lake and the village of Mattawa has a length of about
100 miles and a depth, in places, at present only 100 feet above sea
level.

The 800-foot isobase extends from the region just north of Lake
Temiskaming southeastward, crossing the St. Lawrence just west of
Montreal, thence to Sherbrook where the line turns northeast, passing
just north of Tring Junction in the Chaudiére valley. Between this
line and the 900-foot isobase occur many raised beaches, of which
the following may be mentioned: that lying slightly to the north-
east of Allumette Island at 800 feet; that at Lake Maskinonge at
865 feet; and at St. Jerome at 900 feet. In addition, still higher
beaches have been described at Kingsmere Mountain (north of
Ottawa) at 965 feet, and at La Chute at 1000 feet above sea level.

The extension of the northern margin of the previously discussed
Champlain Submergence toward the north would thus seem justified,
if Fairchild’s isobases are substantially correct and ‘these raised
beaches are of marine origin. In further support of this extensiop
of sea-margin, Taylor says in discussing the limit of post-glacial
submergence in the highlands east of Georgian Bay:“ . - - the
facts show clearly that the same water that filled the ancient channels
in the southern highlands extended far to the north and west. It

|

1932] Potter,—Post-Pleistocene Marine Connection 105

evidently covered all the lowlands of this region and, as indicated by
the altitude of the shore line, made a strait over Lake Nipissing at
least twenty-five miles wide and five hundred feet deep, and probably
another farther north over the height of land to Hudson Bay.” In
a later paper Taylor says: “There can be no doubt of the recent
presence of wide waters at high levels over Lake Nipissing and the
headwaters of the Mattawa river. At the five places seen in the
Ottawa valley, however, no clear and certain evidence of high level
submergence was found, except, perhaps, the thin silts and clays
overlying the drift south of Mattawa, up to about eight hundred
feet. This limit for such a deposit would seem to imply a contem-
porary water surface at a still higher level, and it is more than prob-
able that for a comparatively brief period such an eastward extension
actually existed.” One more note from Taylor’s researches may not
be amiss. In speaking of the history of the Great Lakes he says
(60): “At their highest level, the Great Lakes had open connection
with waters to the east through a broad strait at Nipissing, and it
how seems possible that they had another to the northeast. It is not
yet proved that these connections were with the ocean, but I believe
that the evidence tends more and more strongly toward that con-
clusion.” Coleman (21), in his discussion of Lake Ojibway, admits
the possibility of this lake having been united to the sea; for he says
“. . . whether any portion of the bed of Lake Ojibway was
covered by the sea is uncertain.”

DURATION OF THE SUBMERGENCES

Antevs (2) has estimated about 30,000 years as necessary for the
retreat of the ice sheet from the terminal moraines in New York
State to Cochrane, Ontario. The retreat from Stony Lake, Ontario
(which on the map included in this paper would be located approxi-
mately at latitude 44° 30’, longitude 72°) to Mattawa, Ontario,
required 13,000 years. According to his map (p. 164), the marine
Stage in the St. Lawrence Basin was inaugurated at the time the ice
sheet stood at Stony Lake. He concludes his studies with the fol-
lowing statement: ““ . . . the last ice sheets had their greatest
€xtent and began to wane about 40,000 years ago. This figure may
be less than 10,000 years too large or too sma wii et this
is approximately correct it may be safely concluded that the ice

isappeared from the Hudson Bay region about 35,000 years ago.

106 Rhodora [JUNE

Since the retreat of ice from the terminal moraines of New York
State to Stony Lake, Ontario, consumed 17,000 years, it follows
that the Champlain submergence was inaugurated about 16,000
years B. C. This marine invasion was probably of considerable
duration. Taylor (62) has shown in his paper entitled “ New Facts
on the Niagara Gorge” that the outlet at North Bay, Ontario, was
closed by uplift about 3000 years ago. Mather (44) in his studies on
the Champlain Sea in the Lake Ontario basin states: “That the up-
lift of the St. Lawrence region, in greater part at least, lagged con-
siderably after the removal of ice.” In further support of the length
of life of the Champlain Submergence, Coleman says (20); “This
inland sea must have existed for a long period because clay-forming
sediments accumulated to a thickness of one hundred and fifty feet
or more in some parts of the basin and filled up all the hollows and
formed a fairly level sea-floor.” Keele (38) substantiates this view,
stating that the clay deposits of the St. Lawrence basin vary in thick-
ness up to a maximum of about 200 feet.

In the Hudson Bay region marine fossils have been found up to
approximately 500 feet above sea level. The evidence of raised
beaches in this area further suggests that the marine waters stood
about 700 feet higher than is the case at present. Uplift has thus
taken place and, according to Bell (8), at the rate of five to ten feet
per century. If uplift has been more or less uniform, the time con-
sumed since this differential elevation began would be between
5000 and 14,000 years. That this land-elevation has been more oF
less uniform seems to be the case, for Ells (25) found that the thick-
ness of moss and peat increased toward the south as follows: “Eight
miles west of Moose Factory the moss and peat was two to three
feet thick. Ten miles farther south, two to four feet, forty miles
south, four and a half to five feet, sixty miles to eighty miles south,
five and a half to six feet and ninety miles south, six to eight feet.
It would be fairly safe to conclude, therefore, that at one time during
the existence of these two marine invasions they were contempe-
raneous.

As pointed out earlier in this paper, the Lake Temiskaming and
Lake Abitibi regions were also inundated, although perhaps for 4
shorter period of time than for the two submergences just discussé 4
Cooke (22) concludes from his studies of the Lake Temiskaming

1932] Potter,—Post-Pleistocene Marine Connection 107

region that the waters covering this area must have existed at least
2500 years. Furthermore, the clays of the Abitibi district (Lake
Ojibway) are, according to Knight, Barrows, Hopkins, and Parsons
(39), about seventy feet in thickness. If there is a correlation be-
tween the depth of deposition of clay and the length of life of a body
of water, it follows that the waters over the height of land existed
for a considerable length of time, since as already pointed out, the
length of life of the Champlain Submergence was comparatively
long, with a clay deposition of about two hundred feet

From the foregoing discussion it would seem probable that a water-
connection existed for at least a short period of time between the St.
Lawrence Basin and Hudson Bay.

So far no marine fossils have been Saeisd in this intermediate
region, although in places diligent search has been made (10). The
absence of fossils, however, may be due to the great influx of fresh
waters, in this area from the lakes to the west, through the strait
at Nipissing and also from the retreating ice front. This influx
would have made the waters of the invasion brackish and might
have so modified the salt constituents as to have forbidden the
presence of marine forms, or at least have reduced their numbers.
This fact, supported by the comparatively short life of the probable
marine connection, the erosion and the leaching out of the deposits
and the possible presence of certain chemicals, which would militate
against the preservation of shells, may account for the absence of
fossiliferous clays and sands. Furthermore, as already noted, no
living marine mollusks have been found in James Bay even though
the connection with the sea is direct and the waters of the Bay brack-
ish (9

From the foregoing studies, it seems evident that it is justifiable
to extend the northern margin of the St. Lawrence marine invasion
to at least the region of Lake Temiskaming, and the southern margin
of the Hudson Bay marine submergence to a point a few miles from
the present height of land. Even if the intermediate region was not
covered by marine waters, the two sea-margins were close enough
to allow ordinary agents of local seed-dispersal to become operative
and to bridge the gap, so that plants growing along the shores of the
southern waters might establish themselves in the maritime habitats
of the northern invasion.

108 Rhodora [JUNE

SUMMARY

If this theory of a marine connection between the St. Lawrence
Basin and James Bay is correct, a pathway was open after the re-
cession of the Wisconsin Ice Sheet for the migration of plants of
maritime requirements from the south to James Bay along the
shores of these inland seas. Two sources of evidence would aid
tremendously in establishing this theory: namely, the finding of
marine fossils in the intermediate zone and the persistence of some
maritime plants in this interior region. As stated above, no fos-
siliferous marine clays have been found north of Mattawa until the
marine deposits are met about one hundred and twenty-five miles
south of James Bay.

In only one case have any halophytes been reported from the re-
gion under discussion and in this case the identification is open to doubt
(35). This lack of halophytes in the interior is more or less to be
expected, for unquestionably considerable erosion and leaching of
the soil has occurred since the end of the marine stage, which would
have changed the edaphic conditions enough to have forced these
plants to abandon the region.

Olsson-Seffer (47) has shown that each species of plant has a maxi-
mum salt requirement to which it is very accurately adapted and
that this maximum cannot be overstepped without fatal results to
the plant. Thus, in the case of Glaux maritima, 2.7 per cent is the
maximum; with Juncus Gerardi, 2.2 per cent; with Triglochin maritima,
2.6 per cent, and with Elymus arenarius, 2.6 per cent. It naturally
follows that there must be a minimum salt requirement for halophytes,
and since the maximum in most cases is relatively slight, the leaching
out of this amount after land-elevation took place would not require
a very long period of time. As was noted earlier in this paper, the
halophytes under discussion do not now occur in much of the region
which was known to have been submerged and whose soil about oe
margin of the sea unquestionably supported these plants for a ume
after land-elevation. Svenson (56) has pointed out that the occurrence
of several of these maritime plants about the Finger Lakes region of
New York State is due to the presence of salt springs, rather than to
salt deposits during the Champlain Submergence.

Nevertheless, the occurrence of raised beaches all along the St
Lawrence Basin at altitudes up to 1000 feet, the raised beaches
along the east coast of James Bay, the isobases of Fairchild, the depth

1932] Potter,—Post-Pleistocene Marine Connection 109

of clay-deposits and the occurrence of maritime plants in the Hudson
Bay region all lend strong support to the theory that there existed
after the Wisconsin glaciation a marine connection between Hudson
Bay and the St. Lawrence Basin, and Guppy (34) is probably cor-
rect in saying: “The witness of the living plant is often quite as in-
sistent as the testimony of the rocks.”

CONCLUSION

From the above studies the following statements seem justified:

1. A marine invasion occurred in the St. Lawrence Basin after
the recession of the Wisconsin Ice Sheet, which probably extended
north to include the Lake Temiskaming region.

2. A similar invasion of the sea occurred in the Hudson Bay region
contemporaneous with that of the St. Lawrence Basin, which ex-
tended southward to within a few miles of the height of land.

Evidence seems to indicate that a possible marine connection
existed between these two known submergences involving the Lake
Abitibi area.

4. Driftless areas have never been found in the Hudson Bay region
so that it is improbable that these plants existed within the district
during the Wisconsin glaciation.

5. Wind, water and animals do not seem to be the chief factors which
brought about the introduction of these halophytes into the Hudson
Bay area.

6. The above marine connection (if it existed) would have offered
suitable conditions for the migration of halophytic plants along its
shores

‘. Pndoubtediy the northern margin of the Champlain Sea and
the southern margin of the Hudson Bay inundation were relatively
close and, even though a marine connection may have been lacking,
the factors effecting local plant-distribution may have been sufficient
to have bridged the slight gap and thus account for the occurrence of
the plants in question in the region of Hudson Bay.

BIBLIOGRAPHY
1. Alcock, F. J., Lower Churchill River, Manitoba. Geol. Surv. Can.,

ummary Rep., 1915, pp. 133-136.
2. Antevs, En Ernst, T: oi be Lad (Giaciati ion. Am. Geog. Soc., Research Series,

3 Bell, Robert lorations of the Churchill and Nelson Rivers
and around Report iad Island Lakes. Geol. Surv. Can., Rep. of Prog.,
1879, pp. 1-72 C.

110

oe

Rhodora [JUNE

. Bell, Robert, Report on the Country pegwest Lake eS Fie Hud-

son’s Bay. Geol. Surv. Can., Rep. of Prog., 1878, 1-31

5. Bell, Robert, Report on Exploration in 1875 between James Bay

jor)

er

co

S38 &

and
es Superior and Huron. Geol. Surv. Can., Rep. of Prog., \875-
1876, pp. 294-341.

. Bell, Robert, Report on the Country between Lake Superior and the

Albany River. Geol. Sur v. Can., Rep. of Prog., 1871-1872, pp. 101-

Bell, Robert, Report on Ay asap of Portions of the Attawapishkat and
Alban ny Rivers, Lonely Lake to James Bay. Geol. Surv. Can., Ann
Rep., li. (1886) pp. 34-38 G.

. Bell, Robert, Report on an Exploration of the East Coast of Hudson’s
1-

Bay. Geol. Surv. Can., Rep. of Prog., 1877, pp.
Bell, Robert, bod gen of the Rising . the Land around Hudson Bay. Am.
Jour. Ag . ser. 4, 1. pp. 219-2

; para AE, , Report on the Geology of the Area included by the hi
v. Can., Ann. Rep.

and Tamis amiskaming Map Sheets. Geol. Sur
ol 7) Pi 1, p. 23.
» Pleistocene pouty of the St. Lawrence Valley. Geol.
vere ‘an. p. x. 74 A,

. Chalmers, R.., Havent on oie Surface Geology of New k and

Adinnent Provinces. Geol. Surv. Ca an., Ann. Rep. vi 895) pp.
96-97 A.

. Chalmers, R., Eastern Townships of Quebec. Geol. Surv. Can., Ann.

Rep. ix. (1896) pp. 81-82 A.

. Chalmers, R., Report on the Surface mag crea and Auriferous Deposits

Southeastern Quebec. Geol. Surv. Can , Ann. Rep. x. (1897) fe
1-69 J.

- Chalmers, R., Report on the Surface Geology of Northern New Bruns-

i and Southeastern Canada. Geol. Surv. Can., Ann. Rep. ii. (1886)

: Chalmers, R., " Surface Geology of the Southern Pipi ee = Province of

Quebec. Geol. Surv Can., Summary Rep., 1903, p 143 A.

. Chalmers, R., Artesian B Borings, urine one po fete nt Beaches

- Ontario, Geol. Surv. Can., Ann. Rep., N. 8. xv. (1 1902-3) pp. 270-

- Chalmers, R., Report on the Surface Geology shown on the ree

and Andover r Quarter-Sheet maps, New Brunswick. Geol. Surv. Can.,
. (1899) pp. 30-33 M.

» tvep. XIl
. Coleman, A. P., Ice Ages: Recent and Ancient. Macmillan (1926) B a
; irc 88 A. Pa: — and Trek Water Beaches of Ontario.

. (1901) pp. 129-146.
Coleman, A. 7. Lake Ojibway: a of the Glacial Lakes. Ont. Bur.

aming County, Quebec.

; . 12
. Ells, R. W., Report on i rn of Ar enteuil, Ottawa and a -

Rep. xii. (1899) pp.
: sata Pe W., Sands and Clays of the Ottawa Basin. Bull. Geol. Soc. Am.
os . W., Exploration: Cochrane - Jam —— Temiskaming and
ern Ontario Railway Commissi Ree

“ue to (1912).
; Fairchild. © H. L., Pos ost Glacial C Continental’ Uplift. Science, N. 8. xlvii. no.
15-617

: Pp

Fairchild, H. L., Post-Glacial U ca. Bull

y <4. plift of Northeastern Ameri
Geol. Soe. Am. xxix, (1918) pp. 187-238.

pe aN le a EC SE ae SE EN ae eae

1932] Potter,—Post-Pleistocene Marine Connection 111

. Fairchild, H. L., Gilbert Gulf fongct: W. aters in the Ontario Basin). Bull.
718.

Soc. Am. xvii. (1906) p

; Fassett, N. C., The Vounticn. of ‘the Estuari ies of Northeastern North
1928

merica. Proc. Boston Soc. Nat. Hist. xxxix

A
. Fernald, M. L. A Botanical Expedition to Wonioandinas and southern
Labra 44-145.

abrador. RuHopora xiii. (1911) pp. 1

. Fernald, M. L., Some Relationships of the Flor as of the Northern Hemi-

sphere. Proc. Internat. Cong. Plant Sciences, il, (1929) pp. 1487-1507.

: Fernald, M. Persistence of Plants in Unglaciated Areas of Boreal

America. Mem Am. Aca = ies and Sci. xv*. (1925) pp. 239-342.
Broce of / Canada (1863) p
euppy , Plants ey Boods, and Currents in the West Indies and

Hone, Archibald, Agricultural Resources of Abitibi. Rep. Ont. Bur.
Mines, xiv. p. 241,

? Hitchcock, C. H., The Distribution of Maritime Plants in North America.
A. s., 18
; eat W.F. and Mandsley, J. B., La Motte and Fourniére a Geol.

Surv. Can., Dept. of Mines, Sum mary Rep., 1925, pt. C, p

2-77.
; Keele, i. “Northern Portions of Pontiac and Ottawa pc Quebec.
9-229.

an., Summary Rep., 1916, pp. 2

, Knight C. W., B Barrows, A. G., Hopkins, P.E. ea Parsons, A. L., ge

Night Hawk Gold Area. Ont. Sar: Mines, Ann. Rep. xxxvi

Low, A. P., Report on an Exploration - Part of the South Shore of Hud-
son, Strait and of Ungava Bay. Geol. Surv. Can., Ann. Rep. xi. (1898)

. Low RB; Report on Exploration in James’ Bay and Country East of

Hudson Bay. Geol. and Nat. Hist. Surv. Can., Ann. Rep. iii. (1887) p

i. A. P., Report on the Geology and Economic Materials of ag ener
ern Portion of Portneuf, re proe . and yong si ency Count Geol.

urv. Can., ; pe p. 5 L.
- Mandsley, J. B., St. bebae re A C bee eat District, Quebec. Geol.

Surv. Can., Mem . No. 1

- Mather, K. F., he Cheunen: Sea in the Lake Ontario Basin. Journ.

Geol. xxv. dei7y RP: 542, 554.

- McInnes, W., The Win nisk River. Geol. Surv. Can., Summary Rep.,
90. 00-108 A.

Tt
Missinaibi, Opaz atika and Mattagami fits , Ont. Geo talon On.

Bot. 00) p
Reinecke, L., Road Material Surveys i in 1915. Geol. Surv. Can., Mem.

» Pp. 8- :
Richardson, James, Exploration on Lake St. John. Geol. Surv. Can.,
p. of Pro og., 1857.

: Richardson, , James, Report on the North Shore of the oe er St. Lawrence.
Geo 66-69, 305-31

an., Rep. of Prog., 1

- Small, Jasna Me ag and Development of the Vanuoninn New Phytolo-

gist, xvii, (1918
Smi ith, 8. I. The Crustacea of ace dige — of the United States,
i 1872-

- Spencer, J. W., The Tea Shore hers of the Adirondacks. Bull.
Geol. So 88-495

Ol, Soe. Am., ii. pp.

54, Spencer, oN, Pegg ag of the Algonquin neg and the Birth of

Lake Huron. Am, Jot r, Sci., ser. 3, xii. (1891) pp. 12-21.

for]

112 Rhodora [June

5. Stimpson, W., On the Deep Water Fauna of Lake Michigan, Am. Nat.
iv. (1870) p . 403.
6. Svenson, - ae. Studies on Interior Distribution of Maritime Plants.
HODOR. xxix, pp. 6.
fe ep Fm F, i Gilbert Gulf. U.S. Geol. Surv., Mon. No. 53 (1915) pp.
445-446,

8. Taylor, F. B., Notes on the Quarternary —— of the Mattawa and
Ottawa Valleys. Am. Geol. xviii. (1896) p
9. Taylor, F. B., The Limit = Post-Glacial ‘bree we in the Highlands
ast of Georgian Bay m. Geol. xiv. pp. 273-289.

60. Taylor, F. B., The reels Bait at Nipissing. Bull. Geol. Soc. Am. vy.
(1898) pas 620-626.

Taylor, F. B., Evidence of Recurrent Depression and Resilience in the
Region - apd Creat. ‘Sakon Pap. Mich. Acad. Sci. Arts and Letters,

2. Taylor, F. B., Sow Facts on i Niagara Gorge. Pap. Mich. Acad. Sci.

63. Warming, oe Botany of the Aa alba ii. 676-677 (1903).
arming, E v9 , Ueber Grénlands Vegetation. Englers’ Bot. Jahrbuch, x.
364.

65. Williams, M, ¥: Palaeozoic Straitigraphy of Pagwachuan, Lower Keno-
gami, and Lower oes Rivers. Geol. Surv, Can., Summary Rept.,
8-25

1920 ap
66. Wilson, M. E., Geology and Mineral Deposits of a Part of Amherst Town-
ship, Quebec. Geol Surv. Can., Mem. No. 113 (1919) p.
" conegaphes C ‘ounty, Quebec. Geol. Surv. Can., Mem.
18

~]
=
ie
i? 2]
°
St
r=
5

owe ilson, M. E., Southwestern Portion of the Bu ckingham Map Area, |
Qu eol. Surv. Can. , Summary Rept., 1915, pp. 156-162. |
ARK UNIVE
Woresiter Tee iicate.

CONTRIBUTIONS FROM THE GRAY HERBARIUM
OF HARVARD UNIVERSITY.

C,

ISSUED AUG 9 1932

:, pears preliminary to a general treatment of the Eupa-
rieae—X

By B. L. Rosrnson.

PUBLISHED BY
THE GRAY HERBARIUM OF ogee ag as UNIVERSITY
CAMBRIDGE, MASS., U. S.
1932

RECORDS PRELIMINARY TO A GENERAL TREATMENT
F THE EUPATORIEAE—X. BUS ESO

THERE are here put on record such novelties as have come to the
attention of the writer in his further studies of the Compositae-
Eupatorieae in the course of the past year. To these have been added
a few notes on distribution and synonymy. One new species, which is
not of the writer’s, namely Stevia setifera Rusby, hitherto unpublished,
is with Dr. Rusby’s kind permission here put into print so that it
may be duly included in the succeeding revision of the Bolivian
members of the genus. Finally, two plants of the Revillagigedo
Islands, already described in the Proceedings of the California Acad-
emy of Sciences, are here again put on record in order that they may
be provided with the Latin diagnoses now essential under the Inter-
national Rules of Botanical Nomenclature. The sources of the
material studied are individually indicated, the well known abbre-
viations for the different herbaria being the same as those repeatedly
employed in earlier papers of this series.

Stevia Banc Rusby, Mem. Torr. Bot. Club, iv. 209 (1895).
In its strikingly unlike yet clearly intergrading forms this constitutes
one of the commonest and at the same time most perplexing Stevias
of the Bolivian Andes. It varies conspicuously in stature, habit,
length of internodes, spread of branches, breadth and toothing of
leaves, and in the number of pappus-awns. While these features
vary somewhat independently, most of the as yet available material
of the species (taken in a broad sense) may be divided into two fairly
Tecognizable tendencies, as follows:

Var. x. typica, erecta vel suberecta saepius 2.54 dm. alta plerum-
que multicaulis vel a basi adscendenter ramosa; ramis gracilibus
rectiusculis; foliis lanceolatis vel anguste ovato-oblongis plerisque
acutis et argute serratis saltim duplo vel triplo longioribus quam
latis.—For lit., synon., exsicc., etc., see S. Bangii p. 63.

Var. 8. dyscrita, var. nov., humilior saepius 1-2 dm. alta; caule
flexuoso patenter ramoso; ramis saepius flexuosis vel recurvatis;
foliis ovatis obtuse serratis dimidio longitudinis latioribus.—Botiv1a:
near the city of La Paz, alt. 3750 m., Dr. Otto Buchtien, nos. 186 in
part (typE, in Gray Herb.), 3017 (U. S., N. Y.), 3018 at least in large
part (U. S.) and 3019 (N. Y., U.S.); on stony hills above La Paz, alt.
3700-3800 m., Pennell, no. 14,255 (Gr., N. Y.); on hills, same locality,
alt. 4000 m., Mandon, no. 246 bis (Gr., N. Y.).

This variety clearly connects on the one hand with the typical
upright and narrow-leaved S. Bangii and on the other passes con-

+ ROBINSON

tinuously to dwarf high-alpine forms closely simulating’ both S.
Mandonii and S. tarijensis. While the union of these three species
into a confessedly polymorphous aggregate would be a_ possible
solution of the taxonomic problem here presented, they are really
too unlike to be convincingly so treated. At least until their geo-
graphic relationships are better understood, it seems best to let these
species stand. The variety dyscrita may be said eminently to justify
the name bestowed upon it, which may be rendered as “ difficult to
define” or “hard to determine.”

S. (§ Breviaristatae) clivicola, spec. nov., herbacea_perennis
suberecta vel paullo basi decumbens 3-4 dm. alta minute incurvo-

terminate fragments. Of these one has sessile elliptical obtuse entire
upper leaves (7-11 mm. long), small heads, obtusish phyllaries with
subappressed puberulence, and fully developed pappus-awns about
equalling the corolla. The other fragment has shortly petioled
crenate-serrate leaves, acute dorsally glandular-puberulent phy. llanies,
and awnless achenes,

Curiously enough, this number of the Mandon plants was referred
without question by Schultz-Bipontinus, Bull. Soc. Bot. Fr. xii 8!
(1865) and Linnaea, xxxiv. 535 (1865-66), to S. suboctoaristata Lag.»

RECORDS OF THE EUPATORIEAE—X 5

an alleged Peruvian species said to resemble in habit the Mexican S.
ovata but to have 7—9-awned pappus. Certainly no part of the con-
fused material which reached the Gray Herbarium under this Mandon
number in any way justifies this interpretation of Schultz. No one of
the three elements has either habital likeness to S. ovata or exhibits a
multiaristate pappus.

8. (§ Multiaristatae) discolor, spec. nov., verisimiliter auger
perennis (basi ignota) erecta; caule tereti lednieo-perare a.4mm.
diametro folioso superne breviter crispeque puberulo tardius datieta’
internodiis plerisque 3-6 cm. longis; foliis (supremis exceptis) op-
positis subsessilibus anguste ovatis crenato-serratis vel subintegris ad
apicem verum obtusum versus setae angustatis basi in partem
petioliformem contractis conspicue discoloribus supra viridibus
glabris subtus multo pallidioribus fere albidis glabris vel in costa
obscure puberulis conspicue punctatis textura firmiter membranaceis
4-5 em. vel ultra longis 1.5-2 cm. vel ultra latis, parte petioliformi

primo et in ramulis adscendentibus gracilibus terminales aggregatis;
glomerulis 2.5-3.5 cm. diametro saepius hemisphaericis vel subglo-
bosis; corollis in tubo proprio gracili ca. 2 mm. longo et in faucibus
cylindricis ca. 4.5 mm. longis purpureis cum glandulis sparsis sub-
sessilibus ornatis; dentibus limbi lanceolatis ca. 1.5 mm. longis
albidis; achaeniis gracilibus 5 mm. longis laeviusculis heteromorphis,
eorum 3 (adelphocarpiis) 8—9-aristatis; aristis ca. 5.5 mm. longis
paullo hispidulis; achaeniis 2 reliquis (idiocarpiis) canal vel
aristulis perbrevibus vix 0.2-0.3 mm. longis coronatis exaristatis.—
Bouvia: Prov. Larecaja: “Viciniis Sorata, Iminapi, Milipaya, &c.
in silvulis, dumosis, &c., Reg. temp., 2650-3300 m. Jan.—April,
1859,” Mandon, no. 242 pro parte (ryPE in Gray Herb., isotype Brl. ).
This multiaristate plant was distributed with typical S. boliviensis
Sch.-Bip. ex Rusby which has pauciaristate achenes, as well as much
broader, less discolorous, much more toothed and decidedly petiolate
eaves. From the label it is clear that material distributed under
number came from several different stations. The writer is
unable to distinguish the S. Schultzii of Hieronymus from S. boli-
tensis Sch.-Bip. ex Rusby. Certainly the multiaristate plant here
described is quite distinct from both.
8. (§ Multiaristatae) Herrerae, spec. nov., habitu et inflorescentia
S. andinae similis, differt caule foliinqne glabriusculis nec sub-
Velutinis, foliis rhombeo-ovatis acutis nec obtusis, serratis nec crenatis.
Herba 6-9 dm. alta erecta superne glandulari-puberula aliter glabra vel

6 ROBINSON

subglabra; caule (basi paullo lignescenti) subtereti purpurascente brun-
neo usque ad inflorescentiam foliifero sed inferne aliquando delapsu
foliorum nudato; internodiis plerisque 1—2.5 cm. longis; foliis oppositis
breviter petiolatis rhombeo-ovatis utroque acutis serratis utrinque
glabris 3-3.5 cm. longis ca. 1.5 cm. latis subtus pallidioribus punctatis
tenuiter reticulatis a basi 3-nervatis; panicula magna ovoidea folioso-
bracteata 3-4 dm. alta. 2 dm. crassa; pedicellis involucrum subae-
quantibus vel eo 2-3-plo longioribus; capitulis 5-floris ca. 1 cm. longis;
involucri squamis anguste oblongis acutis vel acuminatis purpureo-
tinctis dorso brevissime (saepe obscure) glandulari-puberulis; corollae
tubo et faucibus vix distinguendis gradatim ampliatis ca. 4.5 mm.
longis purpureis; limbi lobis oblongis obtusis 1.8-2.5 mm. longis albis
dorso hispidulis; achaeniis heteromorphis, eorum 3—4 (adelphocarpiis)
8-9-aristatis (aristis purpureis corollam subaequantibus); achaeniis
ceteris 1-2 (idiocarpiis) 0-1-aristatis——Prrvu: Dep. Cuzco: Valle de
San Miguel, Machupicchu, alt. 2400 m., Herrera (type in Field Mus.,
phot. and fragm. Gr.). For further exsiccatae see p. 28.

It is a pleasure to name this attractive novelty for its discoverer,
Professor F. L. Herrera of Cuzco, who by his personal explorations
and by his extensive publications has done much to extend scientific
knowledge of the flora of the Department of Cuzco.

8. (§ Multiaristatae) Hoppii, spec. nov., verisimiliter herbacea
et perennis (basi ignota); caulibus subteretibus gracillimis (ca. 1 mm.
crassis) juventate glanduloso-puberulis; internodiis plerisque 1-3 cm.
longis; foliis (supremis reductis exceptis) oppositis lanceolato-ovatis
obtusiusculis crenato-serratis basi cuneatim contractis integris sub-
sessilibus utrinque viridibus concoloribus plerisque 2.5-3.5 em. longis
10-17 mm. latis textura tenuiter membranaceis; corymbis terminali-
bus planiusculis 2.5~-4.5 cm. diametro; capitulis (ca. 14 mm. longis)
breviter pedicellatis; involucri squamis oblongo-linearibus viridibus
tenuibus glandulari-puberulis 7.5-8 mm. longis; extimis acutiusculls
integris; intimis saepius ad apicem subtridentatis (mucro medio 1
apice retuso) ; corollae tubo et faucibus vinaceo-purpureis cum glandu-
lis sessilibus instructis; lobis limbi oblongis pallidis; achaeniis hetero
morphis 4-5 mm. longis; adelphocarpiis 4; aristis pappi 14-16 scabridis
purpureo-tinctis ca. 7 mm. longis; idiocarpio exaristato in summo cum
corona erosa ca. 0.5 mm. alta ornato.—PeErvu: Arequipa, 0D rocks,
alt. 2500 m., June 8, 1925, Werner Hopp, nos. 3 (TYPE, Brl., phot. Gr.)
and 12 (Brl., phot. Gr.). This species in habit somewhat approaches
S. puberula Hook., but has narrower, thinner, nearly sessile inci
and peculiar slightly trifid inner phyllaries. From S. PRlippen
Hieron. of northern Chile it differs in its more slender stems @”
ovate-lanceolate, instead of oblanceolate leaves.

La A

RECORDS OF THE EUPATORIEAE—X ‘

S. (§ Multiaristatae) Pearcei, spec. nov., erecta verisimiliter
herbacea et perennis, ca. 0.3-1.8 m. alta (basi ignota); caule tereti
urpureo brevissime incurvo-puberulo albo-medulloso folioso; foliis
oppositis petiolatis, caulinis ovatis ad apicem obtusiusculum gradatim
angustatis crenato-serrulatis supra puberulis subtus incurvo-puberulis
utrinque viridibus textura membranaceis ca. 4 cm. longis et 3 em.
latis; petiolo gracili ca. 1 cm. longo; capitulis ca. 9 mm. longis ses-
silibus in glomerulos plerumque 1-2.5 cm. diametro ad apices ramorum
elongatorum paniculae approximatos congestis; involucri squamis
lineari-lanceolatis viridibus sed saepe purpureo-tinctis 6 mm. longis
acutis et saepe mucronatis dorso incurvo-puberulis et cum glandulis
sessilibus ornatis; corollis albis, tubo proprio 1 mm. longo et faucibus
subeylindratis 3.5 mm. longis et limbi lobis ovalibus apice rotundatis
omnibus cum glandulis sessilibus sparsis instructis; achaeniis hetero-
morphis; 4 adelphocarpiis quibusque 6-7-aristatis, aristis corollas
paullo superantibus, idiocarpio calvo.—PErRv: Dep. Huanuco: common
on waste ground, dry hillsides, etc., Huanuco, Mufia, &c., R. Pearce,
no. 104 (K., phot. and small fragm. Gr.).

8. (§ Breviaristatae) Pennellii, spec. nov., habitu et pappo
S. anisostemmati subsimilis, differt foliis grossius et acutius serratis
nec crenato-serratis, capitulis majoribus, achaeniis longioribus.
Herbacea perennis decumbens; caule tereti gracili folioso atropur-
pureo ca. 2 mm. diametro incurvo-puberulo et ad corymbum versus
etiam glandulari-puberulo; internodiis plerisque 1.5-5 em. longis;
foliis oppositis subsessilibus sed caulinis mediis et inferioribus ad
basin versus in partem integram petioliformem cuneato-alatam
subabrupte angustatis, margine serratis (dentibus subacutis minutis-
sime ciliolatis); foliis ovatis textura subcoriaceo-membranaceis
utrinque glabris vel juventate obsolete hirtellis ca. 3 cm. longis ca.
1.7-1.9 em. latis acutis vel (inferioribus) obtusis; corymbo terminali
trifido; glomerulis campanulatis planiusculis densis 1.5-2.5 cm.
diametro; capitulis 12-15 mm. longis 5-floris sessilibus; involucri
Squamis atropurpureis oblongo-lanceolatis acutis dorso g andulari-
puberulis; corollae tubo proprio purpureo gracili 1 mm. longo ; faucibus
distincte ampliatis subcylindratis ca. 4 mm. longis purpureis; denti-
bus limbi 5 ovato-oblongis subacutis albis vel roseo-albidis 2-3 mm.
longis; achaeniis homomorphis gracilibus 5-6 mm. longis sublaevibus;
Pappi e squamis paucis scariis exaristatis acutis vix 0.7 mm. longis
et squamis 1-2 aristuliferis ca. 2 mm. longis composito.—PERU:
Dept. of Cuzco: grassy slope, edge of paramo, alt. 3800-3900 m.,
Paso de Tres Cruces, Cerro de Cusilluyoc, May 3, 1925, Dr. Francis
W. Pennell, no. 13,883 (rypx, in Gray Herb., isotype Field Mus.).

8 ROBINSON

This species must be attractive by reason of its dark purple involucre
and conspicuous white or pink corolla-limb. It'is a pleasure to name
it for its discoverer who has contributed by his indefatigable explora-
tion very extensively to the knowledge of the cordilleran floras of
South America.

S. Pururpprana Hieron. in Engl. Bot. Jahrb. xl. 364 (1908). 8.
menthaefolia Phil. Fl. Atac. 203 (p. 29 of reprint) 1860, not Sch.-
Bip. As pointed out in Contrib. Gray Herb. Ixxxv. 123 (1929) this
species of northern Chile presents merely the broader-leaved form of
a plant which without other change of importance displays a con-
siderable and continuous variation in leaf-breadth and which, as it
passes to its narrower-leaved forms, becomes what Philippi, |. ¢.,
described as his S. hyssopifolia, a name maintained by Robinson,
Contrib. Gray Herb. Ixxxv. 122 (1929), who was then following the
Vienna Rules which still permitted the use of a later homonym in
cases where the earlier had been completely abandoned. However,
in accordance with the subsequent vote passed at the Cambridge
International Congress, it becomes needful to alter this procedure
and give up the name 8. hyssopifolia Phil. (1860) since it was ante-
dated by S. hyssopifolia HBK. (1820) even though the latter has at
present no validity. Apparently the next available name for the
species of northern Chile is S. Philippiana Hieron., which must
therefore be taken up as above both for the form with oblong-lanceo-
late leaves which Philippi called S. menthaefolia and the linear-
leaved form which he named S. hyssopifolia. :

A pubescence-variation is perceptible and was indicated by Robin-
son, |. c. With corrected nomenclature the varieties under the specific

Supra vix puberulis infra multo pallidioribus glabris punctatis 3-3.6
em. longis 1-1.5 em. latis; inflorescentiae fastigiatim ramosae ram

|
;
|
:
|

RECORDS OF THE EUPATORIEAE—X 9

glandulari-puberulis; capitulis sessilibus vel breviter pedicellatis ca.
15 mm. longis; involucri squamis lanceolato-linearibus acutis vel
attenuatis dorso vix puberulis sed cum glandulis sparsis sessilibus
ornatis ca. 7 mm. longis; corollis albis inferne paullo glandulari-
puberulis, lobis limbi ca. 2 mm. longis; achaeniis heteromorphis;
adelphocarpiis 3, quoque 10—12-aristato, aristis corollas subaequanti-
bus scabridis purpurascentibus; idiocarpiis 2, eorum altero exaristato
squamulis solis coronato, altero squamulis et 2-3 aristulis inaequalibus
corolla distincte brevioribus ornato.—PrErvu: Dep. Puno: in high
meadows with scattered shrubs and locally called “educhuifia,”
Cuyocuyo in the Prov. Sandia, alt. 3700 m., Weberbauer, no. 850 (Brl.,
phot. and fragm. Gr.).

8S. (§ Multiaristatae) sarensis, spec. nov., herbacea erecta
perennis superne adscendenter ramosa, habitu S. breviaristatae
similis sed heterocarpa et multiaristata; radice densissime fasciculata
fibris ex caudice brevi lignoso plerisque simplicibus paullo incrassatis;
caule tereti folioso ca. 4 mm. diametro pallide brunneo pilis longius-
culis moniliformibus albis crispis ornato; internodiis inferioribus
saepe brevibus (6-14 mm. longis), mediis et superioribus longioribus
(2.5-8.5 cm. longis); foliis oppositis rhombeo-ovatis vel anguste
deltoideo-ovatis crenato-serratis (dentibus numerosis parvis 1-2 mm.
altis) apice gradatim acutatis ad basin versus integris primo abrupte
contractis deinde acuminate in petiolum alatum gradatim angustatis
(basi petioliformi inclusa) 6-9 cm. longis 2-4 em. latis textura
membranaceis vix discoloribus utrinque pilosis supra basin 3-nervatis;
laminae parte latiori plerumque 5-7 cm. longo, parte petioliformi
saepius 1.5-2 cm. longa; capitulis 5-floris sessilibus vel subsessilibus
in glomerulos paniculatim dispositos aggregatis 1 cm. longis; involucri
squamis linearibus acutis vel acuminatis dorso breviter laxeque
pilosis; corollis (post exsiccationem) tubo ca. 2 mm. longo et faucibus
ca. 2.5 mm. longis sparse granulatis flavido-brunneis, dentibus limbi
ca. 0.8 mm. longis albidis hispidulis; achaeniis gracilibus nigrescenti-
bus 2.6-3.5 mm. longis in angulis paullo hispidulis heteromorphis;
eorum 4 (adelphocarpiis) ca. 11-aristatis; aristis flavido-albidis breviter
barbellatis ca. 6 mm. longis; achaenio quinto (idiocarpio) inaequaliter
ca. 4-aristatis et cum squamulis (vix 0.4 mm. longis) coronato, aristo
uno corollam subaequante, ceteris multo brevioribus (3-4 mm. longis).
—Botivia: Department Santa Cruz: Province Sara: in clearing,
Buena Vista, alt. 450 m., Apr. 7, 1925, Steinbach, no. 7048 (TYPE,
in Gray Herb., isotypes in herb. Bot. Mus. Berlin-Dahlem, Gen., K.,
Field Mus., Mo.); Buena Vista, alt. 500 m., July.4, 1924, Steinbach,
no. 6117 (K., Gen., Brl., phot. Gr.); clearings, Buena Vista, alt. 450

10 ROBINSON

m., Mar. 1915, Steinbach, no. 1143 (Brl., phot. Gr.); low woods,
Buena Vista, alt. 800 m., Dec. 9, 1920, Steinbach, no. 5169 p. p. (hb.
Osten).

This species of § M ultiaristatae, with heterocarpous character and
many-awned adelphocarps, distinctly recalls the common Argentine
S. breviaristata Hook. & Arn., but of course is to be readily separated
by its very different pappus as well as in various minor traits, It is
perhaps even more closely related to S. menthaefolia Sch.-Bip. to
which its variety of laxer inflorescence was referred by Hieronymus
when determining the plants of Kuntze, but it has a looser and
usually longer pubescence, somewhat thinner leaves with a more
deltoid tendency to the expanded part of the blade and a much
longer petioliform base. The leaves are also much more disposed
to darken to brown in drying, while in the Brazilian plant they remain

at present best treated as distinct species and of course they are
rather widely separated geographically.

S. sarensis shows two varietal tendencies readily distinguishable
Mm most cases.

Var. a. typica, capitulis plerisque sessilibus vel obscure breviter
pedicellatis in glomerulos ad apices ramulorum paniculae subcorym-
bosae aggregatis.—Distribution and exsiccatae as above stated. ae

Var. 6. dissiticeps, var. nov., inflorescentia laxiori capitulis
plerisque plus minusve evidenter pedicellatis in ramulorum paniculae
subracemose dispositis.—S. menthifolia of Kuntze’s Rev. Gen. ill.
180 (1893), not S. menthaefolia Seh.-Bip.—Botivra: Dept. Santa
Cruz: Prov. Cerecado: in clearings, Bafiado del Piray, alt. 450 M.,
July 16, 1924, Steinbach, no. 6248 (ryPx, in herb. Bot. Mus. Berlin-
Dahlem, phot. and small fragm. Gr.); Prov. Sara: moist woods of the
Pampa, Buena Vista, alt. 500 m., Dec. 9, 1920, “forma raudales
sobre arbustos de 2-4 metros,” Steinbach, no. 5169 p. p. (Gr., N. Y.);
Prov. Sara: Pampobosque, alt. 450 m., Apr. 10, 1916, Steinbach, no.
2790 (Brl.); Camino, Proy. Buenavista, alt. 450 m., Mar. 1919,
Steinbach, no. 1126 (Brl.); Dept. Santa Cruz, Apr. 1892, Kuntze
(A, Y.. 0. S., phot. Gr.). :

S. (§ Multiaristatae) setifera Rusby, in herb., herbacea perennis
vel fortasse frutescens (basi ignota); caule tereti medulloso purpure?
breviter incurvo-puberulo ; inflorescentia dense glandulari-puberula :
foliis oppositis, caulinis petiolatis ovatis vel ovato-lanceolatis acutls
vel attenuatis serratis (basi rotundata vel gradatim angustata integra
excepta) membranaceis supra incurvo-puberulis subtus pallidioribus
et laxe in nervis Pilosis ca. 4 em. longis et 2 cm. latis, rameis minoribus

RECORDS OF THE EUPATORIEAE—X 11

lanceolatis ; capitulis 12-15 mm. dongs sessilibus vel subsessilibus
in glomerulos minores dispositis; his aut in glomerulos majores
campanulatos vel atau aggregatis aut ad apices inflores-
centiae compositae corymboso-paniculatae versus sublaxe segregatis;
involucri squamis anguste oblongis acutis vel obtusiusculis saepe pur-
pureo-tinctis dense cum glandulis breviter stipitatis et saepe pilis
paucis patentibus eglandulosis ornatis; corollae tubo proprio (ca. 2
mm. longo) et faucibus subcylindricis (ca. 5 mm. longis) purpureis a
glandulis subsessilibus instructis; dentibus limbi albis vel fortasse
roseis; achaeniis heteromorphis, 3 adelphocarpiis 7—11-aristatis,
achaenio intergradienti unico 3-4-aristato, idiocarpio corona scariosa
erosa vel dentata instructo.—Botivia: Dep. Larecaja: Prov. Murillo:
dry hillside, Pongo, alt. about 3500 m., Dr. O. E. White, no. 168
(TyPE, in herb. N. Y. Bot. Gard., isotype, Gr.).

For citation of further exsiccatae see p. 61.

Dr. H. H. Rusby, finding this novelty among the collections secured
by the Mulford Expedition, has most kindly permitted its publication
in the present paper.

. SORATENSIS Hieron. in Engl. Bot. Jahrb. xxviii. 560 (1901).
this species three varieties may now be distinguished as follows:

ar. «. typica, involucri squamis dorso dense patenterque glandu-
lari-puberulis; pappo e squamis brevibus solis plus minusve connatis
composito omnino exaristato.—For distribution and exsiccatae see
pages 26 and 47.

Var. 8. SUBEGLANDULOSA Hieron., involucri squamis dorso cum
pilis albis eglandularibus incurvis vel crispis ornatis cum vel absque
glandulis paucis sessilibus vel subsessilibus intermixtis; pappo ut
apud var. typicam squamiformi et exaristato.—Hieron. in Engl. Bot.
Jahrb. xl. 359 (1908).—For further literature, synonymy, distribution
and exsiccatae see page 48.

r. y. Mecoyensis, var. nov., involucri squamus dorso ut apud
var subeglandulosam cum pilis albis i incurvis vel crispis eglandularibus
ornatis; pappo cujusque achaenii e squamis brevibus plus minusve
connatis et aristis corollas subaequantibus duabus composito.—
Bottvia: Dept. Tarija: Red Hills, Mecoya, alt. 2745-3050 m.,
es 1864, Pearce (typ, in herb. Kew, phot. and small fragm.

r.).

Eupatorium (§ Cylindrocephala) Bartlettii, spec. nov., fruti-
Cosum glaberrimum; ramis ramulisque oppositis Pr ae teretibus
Post exsiccationem tenuiter multicostulatis; foliis oppositis graciliter
petiolatis oblongis vel ovato-oblongis ad apicem verum obtusiusculum
vel rotundatum acuminatis integris vel paullo crenato-undulatis basi

12 ROBINSON

San Antonio, May 6, 1931, H. H. Bartlett, no. 13,068 (rypn, in her-
barium of the University of Michigan, phot. and fragm. Gr.).

So far as known to the writer there are but two other Central
American Eupatoriums of § Cylindrocephala which have about 5-
flowered heads. These are E. sexangulare (Klatt) Robinson and E.
psoraleum Robinson. Both of these have their branches (or at least
branchlets) distinctly angled and their leaf-blades acute to acuminate
at base, thus being distinguishable from the species here described.

It is a pleasure to associate with this interesting plant the name of
its discoverer, Professor H. H. Bartlett of the University of Michigan,
whose courage and diligence in penetrating several of the less explored
parts of Mexico and Central America have brought to scientific
attention this and other novelties. ;

E. (§ Cylindrocephala) bathyphlebium, spec. nov., ut videtur
suffruticosum vel vere fruticosum; ramulis teretibus foliosis brevissume
fusco-puberulis; foliis oppositis subsessilibus lanceolatis basi subpetioli-

involucri cylindrati squamis ca. 5-seriatim imbricatis brunnescentibus
apice rotundatis dorso glaberrimis vel obsolete granulato-puberulis;
flosculis ex involucro vix exsertis; corollis ca. 4 mm. longis cylindratis;
achaeniis deorsum decrescentibus 3 mm. longis atrobrunneis in an

op A eae a et eS CR cis Beha a

RECORDS OF THE EUPATORIEAE—X 13

alludes to the veins of the leaves, these being conspicuously depressed
on the upper surface.

E. (§ Subimbricata) cuchabense, spec. nov., herbaceum vel
suffruticosum erectum ad 1 m. altum post exsiccationem fuscescens;
caule subtereti medulloso mox glabrato post exsiccationem paullo
costato-angulato; internodiis plerisque 3.5-14 cm. longis; foliis
oppositis ovato-lanceolatis caudato-attenuatis et acuminatis argute
serrulatis (dentibus vix 0.5 mm. altis cuspiformibus inter se 3-6 mm.
distantibus) basi gradatim attenuatis subsessilibus anguste perfoliatis
10-16 cm. longis 3-4 em. latis tenuibus subregulariter penniveniis
utrinque glabriusculis vel glaberrimis; inflorescentia terminali sub-
pyramidata multicapitulata minute fulvescente crispe puberula;
ramulis et pedicellis saepe arcuatis; capitulis ca. 40-floris ca. 8 mm.
altis primo in glomerulos congestis tardius bene pedicellatis; involucri
squamis ca. 3-seriatim laxe imbricatis plerisque anguste lanceolatis
vel linearibus dorso striatis acutis vel acutiusculis subglabris; corollis
gracillimis ad limbum versus laete purpureis; dentibus limbi brevissi-
mis; styli ramis filiformibus; achaeniis glabris nigrescentibus ca. 1.8
mm. longis; pappi setis tenuissimis lutescente albidis corollam ae-
quantibus; disco paullo convexo ut videtur glaberrimo.—Ecuapor:
Province of Esmeraldas, on the Cuchabi River, August, 1904, Father
Sodiro, no. 15 (rypE, in Bot. Mus. Berlin-Dahlem, phot. and small

ag. Gr.).

This plant so closely resembles in general habit E. nemorosum
Klatt that it was at first regarded as a possible form of that species.
However, its heads are about 40-flowered, while in E. nemorosum
they are mostly 100-200-flowered. The plant is here much smoother,

e indument even on the inflorescence being merely a minute fulves-
cent crisped or incurved puberulence, while in E. nemorosum it is a
longer, more spreading and much paler pilosity. The leaves are here
less definitely provided with a winged petioliform base and are much
more finely and sharply toothed than in E. nemorosum. The corollas
of E. cuchabense are clearly purple while in E. nemorosum they are
white or nearly so. Finally no hairiness, even of the shortest and most
obscure, has been found on the disk in E. cuchabense while in E
nemorosum it is ordinarily to be observed. Therefore, there seems no
Teason to doubt that the plant represents a distinct species.

E. (§ Subimbricata) fragile, spec. nov., habitu laxum, aspectu
Primo glabrum sed superne obscure minuteque puberulum ; caule
gracili tereti fragilissimo primo purpureo ad basin lignescenti; interno-
diis plerisque 1-9 cm. longis; ramis oppositis longis flexuosis foliosis;
foliis oppositis graciliter petiolatis subchartaceo-membranaceis ovatis

14 ROBINSON

acutis grosse serratis basi subtruncatis juventate pubescentibus
maturitate solum obscure in nervis venisque hirtellis supra aetate
rugulosis sublucidis subtus vix pallidioribus opacis impunctatis a
basi 3-nervatis 3-4.4 em. longis 1.5-2.3 em. latis; petiolo plerumque
8-14 mm. longo; corymbis ramos terminantibus convexis subdensis
paucicapitulatis; capitulis ca. 20-floris pedicellatis ca. 7 mm. longis;
involucri squamis valde inaequalibus ca. 3-seriatim imbricatis post
exsiccationem atropurpureis, extimis brevibus lanceolatis acutis
hirtellis, intermediis anguste ovato-oblongis acutis, intimis longioribus
lineari-oblongis apice rotundatis glabris; corollis albis glabris 3.2-3.7
mm. longis; achaeniis gracilibus atrobrunneis 2.5 mm. longis, angulis
pallidis sursum scabratis; pappi setis 25-30 delicatulis albis corolla
brevioribus.—Prrvu: sunny cafion rock-ledges, alt. about 2750 m.,
very fragile, from a woody caudex, Mito, July 8-22, 1922, Macbride
& Featherstone, no. 1585 (rypx, in Field Mus., isotype Gr.). Style-
branches and phyllaries said in nature to be pink, though after drying
the former are most often a yellowish or brownish white and the
latter a dark violet purple. Perhaps as close to E. callacatense Hieron.
as to any Peruvian Eupatorium heretofore described, but readily
distinguished by its acute to acuminate outer phyllaries and its
firmer, smaller, at length bullate and nearly glabrate leaves.

E. (§ Subimbricata) gentianoides, comb. nov., Brickellia
gentianoides Robinson, Contrib. Gray Herb. lxviii. 42 (1923). Further
and manifestly conspecific material of this plant, shown by the sub-
sequent and more copious collections of Bornmiiller, no. 515, and of
Dusén, no. 14,611, makes clear the fact that the ribs of the achenes
examined in the original specimen (collected by Saint-Hilaire) were
more humerous than ordinarily occur in the species. The later
examined material of the plant has often 5-angled achenes, quite
typical of Eupatorium, and wholly lacks the intermediate ribs. — It
thus appears that the greater number of ribs first observed in the Saint-
Hilaire material was of an exceptional and non-characteristic nature.
The occurrence of intermediate secondary ribs is not without pre

E. (§ Subimbricata) hosanense, spec. nov., fruticosum gla-
berrimum 2-3 m. altum ; caule subtereti brunneo post exsiccationem

RECORDS OF THE EUPATORIEAE—X 15

paullo costulato; internodiis plerisque 3-4.5 cm. longum; foliis
oppositis graciliter petiolatis ovatis crenato-serratis vel argute ser-
ratis obtusis vel breviter acuminatis basi obtusis vel rotundatis vel
subcordatis supra basin pinnatim 5-nervatis textura firmiter mem-
branaceis utrinque glabris viridibus subtus paullo pallidioribus 5-9
cm. longis 3.5-6 cm. latis; petiolo plerumque 2.5-5.5 em. longo;
panicula terminali corymbiformi convexa 15 cm. diametro densius-
cula; capitulis ca. 30-floris ca. 8 mm. longis graciliter pedicellatis;
involucri brevis campanulati ca. 3-seriatim imbricati squamis ca.
17 ovalibus obtusis glabris dorso striatis margine subscareosis minute
erosis; corollis ca. 4 mm. longis roseo-albis; achaeniis gracilibus
glabris 2.5 mm. longis; pappi setis ca. 20 albis tenuissimis corollam
aequantibus.—Botivia: Department Santa Cruz: Rio Blanco, Cerro
Hosana, alt. 1200 m., Aug. 12, 1917, J. Steinbach, no. 3368 (TYPE, in
herb. Bot. Mus. Berlin-Dahlem, phot. and small fragm. Gr.). This
species appears to have no close relative among the Bolivian Eupa-
toriums thus far recorded, nor has it been possible to place the plant
satisfactorily in any of the species of the adjacent countries.

E. inulaefolium HBK. forma lasiophlebium, forma nova,
habitu formis typicae et suaveolenti simile differt indumento; caule
Tamisque dense pilosis, pilis tenuissimis moniliformibus attenuatis
purpureo-articulatis patentibus vel paullo dejectis; foliis tenuibus
utrinque viridibus supra sparse papilloso-hirtellis subtus in nervis
venisque lasio-pilosis in areolis glabris—BraziL: in campo, Serra do
Itatiaia, ca. 2100 m., May 20, 1902, P. Dusén, no. 1161 (TPE, in
herb. Bot. Mus. Berlin-Dahlem). Though closely sharing the habit,
inflorescence, and floral characters of the common, widely distributed,
and variable species E. inulaefolium HBK. (as broadly interpreted
by Hieronymus) this plant has a marked difference of indument.
The hairs of its stem and branches are much longer, more clearly
jointed, and purple-tinged, while on the lower surface of the leaves
only the nerves, veins, and veinlets are hairy, the enclosed areolae
being entirely glabrous, which is rarely if ever the case in the typical
form or in forma suaveolens. ‘

E. (§ Eximbricata) pacificum Robinson, fruticosum breviter
pilosum; caulibus erectis vel saepius curvato-adscendentibus superne
famosis brunneis subteretibus; medulla alba; internodiis 3-10 cm.
longis; pilis brevibus patentibus gracilibus attenuatis nodulosis ;
foliis oppositis graciliter petiolatis deltoideo-ovatis acuminatis pa-
tenter dentatis sed ad basem late cordatam vel subcordatam versus
integris utrinque viridibus primo aspectu glabris sed in nervis venisque
Pilosiusculis subtus paullo pallidioribus tenuiter immerseque reti-

16 ROBINSON

culatis membranaceis 2.5-9 cm. longis 2-7 em. latis, dentibus limbi
1.5-3 mm. altis 2-6 mm. basi latis apice subacutis vel obtusis vel
rotundatis; petiolo patenter pilosiusculo 1-5 cm. longo; corymbis
compositis oppositirameis foliaceo-bracteatis; capitulis ca. 40-floris;
involucri squamis subaequalibus lanceolato-linearibus acutis 2-3-
seriatim laxe imbricatis maturitate ca. 5 mm. longis, exterioribus
dense pilosiusculis; corollis albis laeviusculis, tubo proprio gracili 2
mm. longo; faucibus anguste campanulatis 1—-1.3 mm. altis; achaeniis
nigris argute angulatis lucidulis 1.5-2 mm. longis sursum et deorsum
paullo attenuatis praecipue in angulis saepe hispidulis; pappi setis ca.
20 albis tenuibus vix scabratis corollam subaequantibus.—Proc.
Calif. Acad. Sci. xx. 99 (1931)—Socorro Istanp, May 27-July 3,
1903, F. E. Barkelew (Univ. Calif.); shrub at summit of Socorro
Island, May 8, 1925, Herbert L. Mason, no. 1666 (ryPE, Calif. Acad.
Sci., isoTyPE, Gray Herb.).

For some years this species has been known to the writer from Mr.
Barkelew’s specimens which were too immature to justify description.
The plant has again been collected by Mr. Mason. His material is
so overmature as to have shed nearly all its florets. However, it
is clearly identical with that previously obtained by Barkelew and
together the two collections yield characters sufficient to show the plant
an undescribed endemic species. In many respects it is similar
to E. pazcuarense HBK. and E. conspicwum Kunth & Bouché, both
of southern-central Mexico. From the former it differs in its much
greater lignescence and its more deltoid-ovate distinctly cordate
leaves which are 3-nerved from the very base. From the latter it
may be readily distinguished by its somewhat coarser and more
spreading pubescence, shorter petioles and by its leaf-blades which
are nerved from the very base, much less toothed toward the base,
and not disposed to be decurrent upon the petiole.

E. (§ Campanuloclinium) phlebodes, spec. noy., ut videtur
herbaceum robustum erectum virgatum 7 dm. vel ultra altitudine
(basi ignotum); caule tereti post exsiccationem costulato 7 nm.
diametro folioso glandulari-puberulo et sparse setuloso; foliis oppositis
et saltim in parte superiori caulis alternis sessilibus late ovatis apice
subacutis margine crenato-serratis basi cordatis vel etegyi

utrinque scabris fere a basi 5-nervatis venosissimis textura

puberulo; pedicellis saepe curvatis vel paullo flexuosis plerisqué
1-2.5 cm. longis, capitulis maturis ca. 1.5 cm. altis et crassis; involuct!
Squamis subaequalibus ca. 2-seriatim laxe imbricatis oblanceolatis

RECORDS OF THE EUPATORIEAE—X li

acutis ca. 1 cm. longis et 2-3 mm. latis dorso praesertim ad basin
versus hispidulis; corollis gracilibus limbo brevi hispidulis; achaeniis
atrobrunneis deorsum attenuatis saepe curvatis ubique paullo granu-
latis ca. 5.5 mm. longis; pappi setis corollam subaequantibus salmoneo-
albidis barbellatis——Paracuay: being a portion of inharmonious
material distributed in Dr. Hassler’s important series as no. 10,178,
said to have been collected by Mr. 7. Rojas, on the plateau and slopes
of the Sierra de Amambay in February, 1908.

This number (10,178) was the one on which Dr. Hassler in Fedde
Rep. Spec. Nov. xi. 173 (1912) based his E. estrellense which he clearly
described as a small-leaved plant of quite different nature and which
he later reduced, in Fedde Rep. Spec. Nov. xiv. 289 ( 1916), to a variety
of E. lysimachiotdes Chod. with which it agrees in all essentials. Of
no. 10,178 the writer has examined in all three sheets. Of these
one, in the herbarium of Mr. Cornelio Osten of Montevideo, cor-
responds perfectly with Dr. Hassler’s characterization of his E
estrellense, while the other two, of which one is in the herbarium of
the British Museum of Natural History and the other in Mr. Osten’s
herbarium, are both of the exceedingly different plant here described
as E. phlebodes. Of these two sheets the one at the British Museum
may be selected as the TyPE (phot. in herb. Gray), the one in Mr.
Osten’s herbarium thus becoming an isotype.

As there evidently has been some confusion in the sorting, number-
ing, or labeling of these sheets of no. 10,178 there is naturally doubt
whether the data on the label applies. However, there can be a
Teasonable assumption that both elements distributed under this
number were collected by Mr. Rojas on his memorable journey into
northern Paraguay in 1907-8.

- TUCUMANENSE Lillo & Robinson, Contrib. Gray Herb. xc. 32
(1930). When characterized this species was known only from the
Province of Tucumén in Argentina. It is now possible to record a
Bolivian station as follows: Rio Blanco, Cerro Hosana, Dept. Santa
Cruz, Boliva, alt. 1300 m., Aug. 12, 1917, a plant 2-3 m. high, with
pinkish white flowers, Steinbach, no. 3367 (Brl.).

ikania (§ Racemosae) chocensis, spec. nov., fruticosa sub-
scandens ad 6-7 m. longa glaberrima; caule tereti post exsiccationem
Costulato; internodiis 1-1.4 dm. longis; foliis oppositis petiolatis
oblongis integris caudato-acuminatis basi acutis pallide viridibus
supra basin pinnatim 5-nervatis subtus paullo pallidioribus 10-11
cm. longis ca. 4 cm. latis; nervis venisque supra subimpressis subtus
Prominulentibus; petiolo brunneo ca. 2 cm. longo; appendicibus
stipuliformibus amplexicaulibus lunulatis ca. 2-3 mm. altis crassius-

18 ROBINSON

culis margine cuspidato-denticulatis; panicula terminali magna
diffusa; capitulis maturis 6-7 mm. longis graciliter pedicellatis race-
mose in ramis ramulisque paniculae dispositis; bracteola ad basin
pedicelli affixa subulata brevi saepe recurvata; involucri squamis
oblongis apice rotundatis ciliolatis dorso brunnescentibus vix nervatis
ca. 3 mm. longis; corollis albis glabris tubo proprio gracili 1.5 mm.
longo; faucibus subcylindrati-campanulatis ca. 1.5 mm. longis;
dentibus limbi ca. 0.5 mm. longis; achaeniis atrobrunneis glabris
2.5 mm. longis; pappi setis ca. 25 lutescenti-albidis corollam subae-
quantibus.—Cotompsta: Intendencia del Chocé: between La Oveja
and Quibdé, Apr. 1, 2, 1931, W. A. Archer, no. 1706 (rypx, U. S.
Nat. Herb., phot. and fragm. Gr.).

Said to be common and to clamber twenty feet and to bear the
local name of “Blumella.” This species, belonging to the relatively
small series of Racemosae, may be placed near the little known WM.
fragrans Klatt of Antioquia, but differs clearly in its caudate-acumin-
ate leaves of paler and duller aspect and somewhat different venation,
also by its stipular appendages—structures not present in M. fragrans
so far as n.

M. punctata Klatt, Bull. Soc. Bot. Belg. xxxi. 195 (“ 1892” [1893]).
This name, in recent years, has been applied to a fairly common an
widely distributed species of Mikania ranging from Mexico to Bolivia
and Brazil. The writer noticed some years ago that this somewhat
widely used name was, in fact, a later homonym of a Mikania punctata
Gardner of Brazil (Gard. in Hook. Lond. Journ. iv. 118 (1845)), but
as the latter was generally reduced to synonymy it was not felt
needful on its account to displace the later name, for the Vienna
Rules did not insist upon the renaming of a later homonym if the
earlier one had been universally relegated to synonymy. The later
rules of nomenclature resulting from the International Congress
at Cambridge require such renaming of the later homonym, eve?
where the earlier has been submerged in synonymy. ‘Therefore 1t
becomes necessary to rename M. punctata Klatt. For this purpose It
's possible to raise M. boliviensis Lingelsheim from synonymy, as it
seems to have been applied to a plant certainly conspecific with ig
punctata Klatt. It is true that the types from Mexico and Bolivia,
if closely compared, show some minor differences, but these break
wholly when the material, now available for the species as a whole,
is critically examined. be

Hieronymus once referred to the southern Brazilian M. vitif olta
DC. Lehmann material from Colombia which is clearly identical with
M. punctata Klatt. After examining the type of the DeCandollean

a RS RON. CE on Re ae Oe em

RECORDS OF THE EUPATORIEAE—X 19

species at Geneva the writer is forced to conclude that it is indeed
doubtfully distinct from the more northern plant which has long
been passing as M. punctata Klatt. However, though the inflorescence
is very similar, the leaf-contour is rather conspicuously different,
being in M. vitifolia more numerously and sharply lobed. Futhermore,
M. vitifolia is geographically removed. Therefore, until some con-
vineing evidence of intergradation has been observed, it seems best
to keep these plants separate and adopt for the Cordilleran species
the first available name, that is, M. boliviensis Lingelsheim.

BRICKELLIA PENINSULARIS Brandegee, Zoe, v. 160 (1903); Robinson,
Mem. Gray Herb. i. 108, fig. 84 (1917). In describing this species
in my monograph of the genus some years ago I pointed out the fact
that material from the island of Socorro, collected by Barkelew, was
of more compact habit. At the time the observed differences did
not justify any varietal or formal separation of this insular strain.
However, the now much more copious material secured by Mr.
Mason on both Socorro and Clarion Islands discloses several de-
partures from the mainland type. While these distinctions would be
very weak if used for specific segregation, it would seem that they
fully warrant a recognition of the following varieties.

Var. typica, foliis ovatis vel saepe oblongo-ovatis plerisque saltim
quartam partem longioribus quam latis modice tenuiter membranaceis
utrinque tenuiter piloso-hirtellis; pedunculis plerisque 2-7 cm. longis;
corollis in summa parte post exsiccationem plerumque nigrescentibus.
—Lower Catirornia: Cape Region, Brandegee, Purpus; Puerto
Escondido, Johnston, no. 4126 (Gr.). .

Var. amphithalassa Robinson, foliis subdeltoideo-ovatis vel fere
orbicularibus vix longioribus quam latis firmius membranaceis vel
subcoriaceis utrinque tenuissime et densissime puberulo-hirtellis;
pedunculis plerisque 8-24 mm. longis; corollis flavidis post exsicca-
tionem vix nigrescentibus.—Contrib. Calif. Acad. Sci. xx. 100 (1931).
—REVILLAGIGEDO IsLaNnps: Socorro Island: Barkelew, no. 198 (Gr.,
Mo., Univ. Wyo., N. Y., Arn. Arb.); a shrub at Benner’s Cove,
May 3, 1925, Mason, no. 1598 (Calif. Acad., Gr.). Clarion Island:
shrub 0.6-1.5 m. high, branched from the base, branches erect, very
brittle, herbage grayish green, forming dominant cover and giving
grayish aspect to the hillsides, Apr. 26, 1925, Mason, no. 1555 (TYPE,
Calif. Acad., tsorypx, Gr.).

20 ROBINSON

THE STEVIAS OF PERU.

Tue Genus Stevia and its special classificatory difficulties were
discussed in some detail by the writer in a former paper (Contrib.
Gray Herb. xc. 36-58). Though its pappus is subject to great vari-
ability and at times to disconcerting transitions, it forms nevertheless
a fairly convenient means of dividing the genus into several for the
most part readily recognizable, if not quite mutually exclusive,
sections. Of these three occur in Peru. The first may appropriately
be called § Eustevia since it is nomenclaturally the typical series of
the genus. In it the pappus is prevailingly of a symmetrical nature
being either exaristate, and composed of short subequal often connate
scales alone, or it may be pauciaristate and made up of few (usually
3-5, more rarely 1-8) awns, with or without some scales, the awns
being about as long as the corolla. These pappus-forms are illustrated
in Contrib. Gray Herb. xc. Pl. I. Figs. I-IV. The transition between
the awned and awnless conditions seems very facile for both are
often found in the same species. Therefore this distinction has little
if any classificatory value as a group character. The actual differential
traits are that 1) its pappus tends strongly to be symmetrical, 2)
its awns, when present, are few (rarely over 5), and 3) its awns when
they occur are about the length of the corolla.

pappus to be found in this section are illustrated by Robinson, |. ¢.
Figs. V-VI and XII-XIII. ‘

In the third section the pappus is symmetrical and multiaristate
(at least in the adelphocarps), the awns being normally 10-25 (-30)
in number and about equalling the corolla in length. This sectio?
was called Multiaristatae by DeCandolle, Prod. v. 122 (1836) and is
appropriately typified by S. multiaristata Spreng.

or a satisfactory phylogenetic interpretation arrangemey
of these sections we must await a very much better knowledge . of
en

the sections, it may be noticed that Stevia is ordinarily placed be

Ageratum and Eupatorium. This being the case and Stevia, as ©

THE STEVIAS OF PERU 21

pappus, being somewhat intermediate between these genera, it is
clear that § Eustevia, with its pappus of few and subsymmetrically
arranged scales, should be placed next (i. e. on the side toward)
Ageratum and that § Multiaristatae, which at-times suggests actual
transition to Eupatorium should be placed as close as feasible to
that genus.

The literature of the Peruvian Stevias is of slight extent and ex-
ceedingly scattered. In no work hitherto published have more than
four members of the genus been attributed to the country and no
attempt has thus far been made to contrast or key any of the Peruvian
species

The treatment here presented, though the result of considerable
study and the examination of most of the pertinent types, as well as
the Stevia material in several of the most important herbaria both
in Europe and in America, is confessedly tentative. While it has
seemed best to distinguish on more or less convincing evidence no
less than twenty-four species of Stevia as native in Peru, several of
these are still very inadequately known—in some cases from a single
specimen each. On this account it is often impossible to indicate
the natural ranges or to estimate the degree of variability occurring

in the entities characterized and likely to obliterate at least some of
their distinctions, when these plants have received the needful rep-
resentation in herbaria and, if feasible, comparison in living material.

However, notwithstanding its limitations from lack of adequate
material, the present paper not only brings together information
previously available only after prolonged search in scattered literature,

ut furnishes what has not heretofore existed, namely a specific
key to the Stevias of Peru in which the more obvious of their distinc-
tions are brought out. It is therefore hoped that it will prove helpful
and stimulating to further study of this difficult but highly interesting
Key To THE STEVIAS OF PERU.
a. Pa 1 sometimes —
ppus gun a ptaet way 6 rg Piven y fey Pree e nc
times pausarstat with at least the adelphocarps (i) 3-5
ews the awns about equalling the corollas § 1 Ev-
b. via. lax, mature —— as long as or often much
longer than the involucre....

. — merely coroniform Be rae. i ous cal ....-1. 8. urticaefolia.
about t equalling the cotellek 0 eee _.2. 8. cathartica.
b. Inflorescence or its parts denser; pedicels none or shorter
the involucre....

d. Leafblad > onset of the cuneately winged petiole)
rowdy eae

22 ROBINSON

é. oo in 2—5-headed fascicles; these rather loosely
mbose at the ends of elongated panisie- bras
Re PPOVAUO. oc Gn es cee cece S. pabloensis.

bar nal pappus-sca
d. mh a rhombieovate to chit rarely more than
as wide as ei
tf: Deda doenally covered with short incurved o
glandless hairs with some sessile glands inter.
ad Gabe Phe cU hex b ws anicigs ee aren ales? 4, S. rhombifolia.
rs Pie are dorsally beset with short straight spreading
gland-tipped hairs or Soa gla: “pt oo oe
g. Leaves gradually narro to an acute apex but
abruptly contracted te oe abteaiah or rounded or :
We er ORB Cn. cc cs vse cn ees . soratensis.
g. Leaves rhombic-ovate to lanceolate-oblong or nar-
rowly elliptical, cuneate at base... .h.
full maturity finely white-punctate
above; heads in y fastigia tely peanchiod clusters;
phyllaries green; corolla-tube and throat — :
. 8. leucosticta.
- Leaves not spent unctate; heads in dense m
headed glomerules; corolla-tube and hrcet? red
to deep nani ogo cvtasii ia eee S. Macbridei.
a. Pappus asymmetrical, composed of conspicuously unequal
— 12 of = longer "han the others and often awn-
tippe wned scales rarely more then half as tole as
_ the corollas § 2 BREVIARISTATAE
i. Heads 12-15 mm. long; ves rhombic-ovate, nearly gla;
brous, about 2 em. wide, firmly membranaceous. .... - - S. Pennellit.
i. Heads about 10 mm. long; leaves subdeltoid-ovaie, ily oli
on surfaces, thin-membranaceous...9. S. galeopsidifolia.
a. Pappus (at least of the Aspon pigtired ah symmetrical, co so
of many, normally 10-25 (30) alee about equalling the
_ corollas ere citieis
j. Inflorescence lax; mature pedicels about equalling or often
exc ing the involuer a
k. mee (in well si ag pairs): 6 aa ad.

TOOUIY OVATE ONO ODUISE... 6. c.g ee ee ee
l. Stem wad tacves el os. “ nearly so; leaves jaime
acutish to acuminate..............+-+-+: 11 S. Herrerae.
_ k. Leaves (crowded) linear...............0.-++: 1 ristala.
7. peat or its parts much denser; heads sessile, sub-
a sessile or on Aimee shorter than the pebkrsieig ‘
. Leaves merous and crowde ihe ‘man
cauline often oneiverouat itis asi _ 8. polyphylla.
m. Leaves lanceolate-ovate to ohana. obovate or sah
orbicu » oh
n, hayes or greets ec pane yy dagbege or more in Mier ke ht ae
ves amplexicaul by a broadly petio ae fos
of the ova a or — — a are 48: melissaefolta-
plexic

raidilie p ber a wiehay retuse aS) Pe ae te 15. S. Hopp

THE STEVIAS OF PERU 23

p. Inner phyllaries entire or nearl ee
q. Outer phyllaries dorsally covered 1 with fourvel
crisped white glandless hairs (some sessile
— being te interspersed) or rarely sub-
glabrous. ...
r. Stem-leaves “narrowly, ovat ovate : huge

ac -
nate, about 12-toothed on each side;
P yllaries one " attenuate; corollas

r. Stem-leaves broadly ovate, mostly about two-
as wide as long....u.
u. eae 12-15 mm. one in rather loose sub-
giate glomerules; stems in clumps
ae a woody alae, loosely gcse
re st n dec a :
the adelp sopping mostly 9-12, seatealy
equalling the corollas................ 19. S. puberula.
u. Heads 9-11 mm. go practically sessile in
dense glomerules; stem a oie Be
parently single and erec
or - hocarps 6-8, slightly Pinte Mo the

bates boo Cavey messy ees 20. S. Pearcei.
. Outer riwvierids gin tt beset with short PES
ing gland-ti mae (a sng venice hairs
sometimes ak
v. Leaves ix anselats to nacrowly sc renter
lly less than a wide as long... .
w. Leaves membranac sac anaes but ams
manently shorkuliias: on both. surfaces, ;
subconcolorous: 2022004605 ee es 21. S. Weberbaueri.
w. Leaves chartaceo-subcoriaceous, glabrous

or nearly so, conspicuousl er beneath.
é a iy paler | vr nga beard var. anomala.
v. Leaves more eh ped —* usually mo
than half as wide ee»
xz. Leaves shactoenien ‘eas serrate, the
teeth directed forward, ay over 0.5
mm. igh UVa eee meus sery Wenn wee 22. S. Lechleri.
x. Leaves membranaceous, dasa b and often
doubly ¢ pater tee unded
the largest as ha as 2mm. ~ ig ep

n. Dwarf, matted, not over 1 dm. high, with short pr
cumbent or prostrate stems and small sessile e ot =
tical leaves (about 1 em. long).................. 24. S. Mandonii,

or

24 ROBINSON

1. 8S. urticaefolia Billb. Slender erect ascendingly much branched
herbaceous perennial, 7-12 dm. high, sometimes a little woody at
the base, sparingly pilose; leaves opposite (except the uppermost),
short-petioled, rhombic-ovate, usually acute at both ends, membrana-
ceous; panicle diffuse; corollas in the typical variety (the only one
thus far seen from Peru) with violet-purple tube and throat, the
teeth white or roseate; achenes alike, awnless, surmounted by a crown
of short often somewhat connate scales.—Billb. in Thunb. PI. Bras.
Dee. i. 13 (1817); Herrera, Estud. Fl. Dep. Cuzco, 187 (1930).

Cuzco: Prov. Quispicanchi: Distr. Oropesa, at Huasao, alt. 3200 m., Her-
rera, no. 893 (Bri.

[Braziu, Bortrvra.]

2. §. cathartica Poepp. Suffruticose, erect, 4 dm. or more in
height; stems slender, leafy, unbranched up to the rounded to ovoid
loose leafy-bracted inflorescence; leaves opposite, ovate-oblong,
narrowed at both ends, serrate, pubescent, the cauline very short-
petioled; the branches of the panicle corymbose at the summit;
mature pedicels about equalling the involucres; involucral scales
acute, dorsally pubescent; pappus said to be 5-awned, the awns
equalling the corollas and alternating with as many short scales.
Nov. Gen. ac Spee. iii. 55 (1845).

DEPARTMENT NOT ASCERTAINED: Cassapi, on subandine calcareous moun-
tains, Poeppig (Nat. Hist. Mus. Vienna, phot. Gr.).

Known as yet only from the type-specimen, this species appears
suspiciously close to awn-bearing forms of S. urticaefolia occurring
in Bolivia and Brazil. However, the leaves are somewhat narrower
and more oblong in outline and the cauline more shortly petioled.
Rediscovery and more copious material will be essential before the
real status of this plant can be confidently determined.

3. 8. pabloensis Hieron. Perennial herb or undershrub, erect,
0.7-1.5 m. high, sparingly hirtellous to villous below, glandular-
puberulent in the inflorescence; leaves opposite, the main cauline
with the expanded portion of the blade broadly ovate, obtusish,
crenate-serrate, half to more than two-thirds as wide as long, sparingly
glandular-villous on both surfaces, raised on a slender cunea
winged petioliform base; panicle-branches ascending; heads 10-12
mm. long, sessile to shortly pedicelled; corollas deep purple as to tube
and throat, the limb roseate to white ; achenes alike, crowned by .
short awnless scale-pappus.—Hieron. in Engl. Bot. Jahrb. x1. 364
(1908).

Casamarca: below S i in a gorge,

m., Weberboucr, 30.3008 GA ee ee

alt.

THE STEVIAS OF PERU 25

4, $. rhombifolia HBK. Erect perennial, herbaceous or at times
frutescent, from incurved-puberulent to somewhat conspicuously
and densely pilose; leaves (except the uppermost) opposite, sessile or
short-petioled, rhombic-ovate, acutish to acuminate; heads small,
6-9 mm. long, sessile or shortly pedicellate in dense corymbously
disposed glomerules; phyllaries narrowly oblong, acute, usually pale
green, incurved-puberulent; corollas mostly white or pink, rarely
a rather deep purple; achenes crowned by a scale-pappus with or
without 1-4 awns about equalling the corollas.—Nov. Gen. et Spec.
iv. 143 (1820); Robinson, Contrib. Gray Herb. xc. 132 (1930), which
see for further literature, synonymy and varietal subdivision.

A highly variable species, probably the most widespread and
abundant of the genus though relatively rare in the Southern Hemi-
sphere. The typical variety with most of its achenes 3 (1-4)-awned
has thus far not been verified as occurring in Peru, where, however,
the following varieties are of at least occasional occurrence.

Var. uniaristata (DC.) Sch.-Bip. Most of the achenes awnless
and crowned merely with a scale-pappus, but a few single awns to
be found in each inflorescence.—Linnaea, xxv. 279 (1853); Robinson,
Contrib. Gray Herb. xc. 134 (1930), xevi. 41 (1931). S. wniaristata
DC. Prod. v. 120 (1836).

Perv: without coal Mathews (BM., Gen.).

(Mexico, VENEZUELA.]

This variety, though early given a certain nomenclatural status,
is probably only the expression of an occasional intermediate tendency

tween the ha oblugssihis variety of the species and the
following awnless

Var. Se nAosema cd, -Bip. Achenes all exaristate, crowned
merely by a scale-pappus.—Linnaea, xxv. 279 (1853); Robinson,
Contrib. Gray Herb. xc. 134 (1930), xevi. 34, 41 (1931).

Amazonas: Chachapoyas, Mathews ms ., phot. Gr., Gen.).

Cuzco: rales of the Rio Apurimac, Herrera a (Field Mus.).

EPARTMENT NOT INDICATED: Mateon

(Mexico, CENTRAL AMERICA, Vewuaviekis Coomera and Ecvapor.]

5. 8. soratensis Hieron. Erect perennial, herbaceous or nearly
so, 4-10 dm. high, sparingly glandular-pilose; gb opposite, petio-
late, lanceolate-ovate, gradually narrowed fro siting base to an
acute or attenuate apex, serrate-dentate, asia unded or even
subcordate at base; petiole about 1 cm. long, carey if at all winged;
heads (about 1 cm. long) sessile in dense convex usly dis-
posed glomerules; phyllaries (about 7 mm. long) densely beset Saeally

26 ROBINSON

with stoutish-stiped glands, often purple-tinged; corollas deep purple;
achenes unlike, four of them bearing a short scale-pappus, the fifth
(idiocarp) wholly destitute of pappus.—Hieron. in Engl. Bot. Jahrb.
xxviii. 560 (1901).

Cuzco: Paucartambo, alt. 3000-3100 m., Weberbauer, no. 6920 (Gr., Brl.);
Checacupe, on open rocky bank, alt. 3400-3450 m., Pennell, no. 13,536 (Gr.,
N. Y., Field Mus.).

[Borrv1a.]

With its copious deep purple flowers this must be a rather handsome
species. The leaves, of a shape rather unusual in the genus, are
commonly (though not always) reflexed, the point being toward the
ground. The Peruvian material examined agrees admirably with the
typical (i. e. the more glandular) form of this Bolivian species.

6. S. leucosticta Robinson. Weak, rather slender, loosely and
flexuously branched glabrous perennial herb, 8-10 dm. high; leaves
(except the bracteal) opposite, sessile or subsessile, lanceolate, acute
or gradually narrowed to a rounded tip, shallowly crenate-serrate
to subentire, finely sprinkled with sessile at length chalky-white
glands; heads (about 12 mm. long) sessile or nearly so, somewhat
fastigiately clustered at the ends of long slender flexuous panicle-
branches; phyllaries lanceolate-linear, acute, green, gland-sprinkled;
achenes 1-3-awned, bearing also some scales and often an awn or
two much shorter than the others.—Contrib. Gray Herb. xevi. 8
(1931).

HvaAnvwco: on a southwestern slope, alt. about 2750 m., Macbride & Feather-
stone, no. 1382 (Field Mus., Gr.).

7. S. Macbridei Robinson. Erect perennial herb, a little woody
toward the base, finely glandular-puberulent, 6-12 dm. high; leaves
opposite or in part alternate, oblong-ovate, obtuse, crenate-serrate
to subentire, shortly petioled or subsessile, glabrous or inconspicuously
hirtellous above, paler and glabrous or obscurely hirtellous on the
nerves beneath, at full maturity apt to become chartaceo-membra-
naceous; heads (about 12 mm. long) sessile in dense corymbously
disposed clusters; phyllaries mostly dark purple, about 7 mm. long;
acute, densely glandular-hispid; corolla-throat dark red, the lobes
white—Contrib. Gray Herb. xevi. 9 (1931).

Var. a. typica. Achenes alike, all awnless and crowned only by
a scale-pappus (0.2-0.4 mm. long).

Fegth onw00; on ‘‘grass-shrub slope,” Mite, alt. about 2750 m., Macbride &
eatherstone, no. 1613 (Field Mus., Gr.)

IMA ,

~s P_grassy northwestern slop , Rio Blanco, alt. about 3660 m-.
a ide & Featherstone, no. 663 (Field Mus., Gr.); open_hillsides, Rie
laneo, alt. 3000-3500 m., Killip d Smith, no. 21,768 (Gr., N. Y.)-

THE STEVIAS OF PERU 27

N: Tarma, Espesto, no. 127 (Brl.); open rocky hillside, Mantaro Canyon,
uth of Huaneayo, between Viques and Ingahuasi, Killip & Smith, no. 22,158
% War EXACT Papper Gay, no. 355 (Par., phot. and fragm. Gr.);
Wilkes . Sood (U. &.).

Var.$. anomala Robinson, 1.c. Heterocarpous, the achenes unlike,
four of them 7-10-awned, the awns about equalling the corollas, the
fifth achene (idiocarp) crowned by scales (0.2-0.4 mm. long) and
1-3 awnlets much shorter than the corollas.

Lima: Oroya, Townsend, no. 1521 (U. S8.).

Junfn: open rocky hills ide, rege Canyon, south of Huancayo, alt. 3150
m., Killip & Smith, no. 22, 158 in part (Gr
; ae Ollantaytambo, ‘alt. about 3000 m., Cook & Gilbert, no. 363 (U. S.,

ge ENT NOT INDICATED: Gay (Par., phot. and fragm. Gr.); Pifiasniocj,
Panticalla E Pass, alt. 3000 m., Cook & Gilbert, no. 1849 (U.S., fragm. Gr.).

This is a species in which the usually practicable and convenient
Dcton between §§ Eustevia and Multiaristatae breaks down.
Without apparent change in any of its other traits this plant occurs
either in the homocarpous awnless state or the heterocarpous condi-
tion with many-awned adelphocarps.

8. S. Pennellii Robinson. Decumbent perennial herb about 3-4
dm. high; stem purple, hirtellous, glandular-puberulent in the in-
florescence; leaves opposite, petiolate, obtuse to acutish, serrate,
ciliolate, the expanded part of the blade raised on a cuneately winged
petioliform base, firmly membranaceous, 3-nerved, nearly or quite
glabrous on both surfaces; heads (12-15 mm. long) sessile in dense
somewhat corymbously disposed clusters; phyllaries dark purple,
oblong-lanceolate, acute, delicately glandular-puberulent; corollas
with short proper tube and much longer subcylindric throat both
dark purple, the lobes of the limb pinkish-white; achenes subequal;
pappus of few short awnless scales and 1-2 shortly awned scales
(about 2 mm. in length).—Contrib. Gray Herb. ec. 7 (1932).

Cuzco: on grassy slope, Paso de Tres Cruces, Cerro de Cusilluyoe, Pennell,
no. 13,883 (Gr., Field Mus.).

9. S. galeopsidifolia Hieron. Upright perennial herb, finely
pilose; leaves opposite, slender-petioled, the blades ebddtcrlovate,
acute, coarsely crenate-dentate, nearly truncate to subcordate at
base; phyllaries pale green, subscarious, nearly smooth; achenes all
bearing a short scale-pappus, the scales at first appearing subequal
but at full maturity definitely unequal, the shorter about 0.4 mm.
long, but 1-2 of the others nearly twice this length and slightly
aristulate.— Hieron. in Engl. Bot. Jahrb. xxii. 719 (1897).

28 ROBINSON

Cuzco: Ollantaytambo, alt. about 3000 m., Cook & Gilbert, nos. 408 (U. S.,
fragm. Gr.) and 414 (U. S., fragm. Gr.).

[Botrv1a.]

10. 8. andina Robinson. Herbaceous perennial, 5-9 dm. high,
pilose; leaves opposite, petiolate (or raised on a slender cuneate
petioliform base of the blade) broadly ovate, obtuse; inflorescence
a lax compound corymb; pedicels at maturity usually longer than the
involucre; corollas with red-purple tube and throat and white lobes;
achenes unlike, four of them 9-12-awned, the fifth awnless or with
but a single awn.—Contrib. Gray Herb. xevi. 3 (1931). 8. elatior
as used by Hieron. in Engl. Bot. Jahrb. xxxvi. 463 (1905), not HBK.

Amazonas: Chachapoyas, Mathews (K., phot. Gr.).

Casamarca: Callate, von J elski, no. 669 (Brl.).

[Ecuapor.]

11. 8. Herrerae Robinson. Habit of the preceding but with stem
and leaves nearly or quite glabrous; leaves rhombic-ovate, acutish
to acuminate; inflorescence tending to be even more lax than in the
preceding, pedicels at full maturity often 2-3 times as long as the
involucre; corollas with throat deep purple and lobes white; achenes
unlike, adelphocarps (3-4) each 8-9-awned, idiocarps (1-2) awnless
or 1-awned.—Contrib. Gray Herb. c. 5 (1932).

Cuzco: Valle de San Miguel, Machupicchu, alt. 2400 m., Herrera (Field

us., phot. and fragm. Gr.); Valley of the Apurimac, Herrera (Field Mus.),
the latter (exhibiting only inflorescence and the uppermost leaves) is doubtful
but probably represents only a slightly puberulent but scarcely glandular form
of this species with less diffuse j orescence.

AYACUCHO: /ealrapa, between Huanta and Rio Apurimac, alt. 2800 m.,
on wooded hillside, Killip & Smith, no. 22,288 (Gr., N. SS

WITHOUT INDICATION OF DEPARTMENT: Mathews (N. Y.).

12. S. multiaristata Spreng. Erect very leafy-stemmed perennial,
4-6 dm. high, grayish-puberulent with fine white incurved or crisped
glandless hairs but the inflorescence and phyllaries closely beset with
short straight gland-tipped hairs; leaves lanceolate-linear, acute at
both ends, entire or obscurely few-toothed; achenes alike or nearly
so, 18-25-awned.—Syst. iii. 449 (1826); Robinson, Contrib. Gray
Herb. xe. 71 (1930), which see for further lit., synon. etc.

Peru: without locality, Mathews “ex herb. Lambert” (K.).

[Uruguay and eastern ARGENTINA. ]

Unfortunately the Mathews specimen which stands for the Peruvian
occurrence of this species is quite without more exact data. Until
the plant, otherwise known only from Uruguay and eastern Argentina,

been rediscovered in Peru, its presence as a native in the country
must be regarded as doubtful.

THE STEVIAS OF PERU 29

13. 8. polyphylla DC. Slender erect virgate perennial, about
3 dm. high; stem densely incurved-puberulent (the hairs glandless),
very leafy; leaves (often proliferous in the axils) linear, entire, shortly
and closely pubescent-tomentellous and punctate; heads sessile in
corymbously disposed glomerules; phyllaries broadly oblong, slightly
toothed at the rounded summit, dorsally spreading-pubescent;
corollas purplish; achenes unlike, the adelphocarps 16—20-awned, the
idiocarp (single) crowned with linear scales and about 4 awns.—
Prod. v. 123 (1836).

Prrv ? “inter Peruvianas et Panamenses herb. Haenke.” (Gen., phot. and
trac. Gr.),

This plant, excellently preserved in the Prodromus herbarium at
the Geneva Conservatory of Botany, seems never to have been re-
discovered in nature. In habit it rather closely simulates the Argen-
tine S. mercedensis Hieron. but has almost twice as many pappus-
awns and has broader, blunter and somewhat toothed phyllaries.

14. 8. melissaefolia (Lam.) Sch.-Bip. Perennial herb or under-
shrub; stem decumbent, curved-ascending, often flexuous, very leafy,
densely glandular-puberulent ; leaves opposite, sessile by a narrowed
but somewhat amplexicaul base, membranaceous, oval to more often
obovate, obtuse or rounded at apex, crenate-serrate; heads sessile
in hemispherical to subglobose glomerules; these single or clustered
(mostly by 3’s) at the ends of the branches; corolla-throat purple,
the lobes white or pink; achenes unlike; adelphocarps (usually 4)
9-12 awned; idiocarp (mostly single) awnless but crowned by a scale-
pappus.—Linnaea, xxv. 291 (1853). Eupatorium melissaefolium Lam.
Enel. ii. 411 (1786). Mikania melissaefolia (Lam.) Willd. Spec. Pl. .
ili. 1747 (1804). Nothites latifolia Cass. Dict. xxxv. 163 (1825).
Stevia dodecachaeta DC. Prod. v. 122 (1836). Nothites melissaefolia
(Lam.) DC. 1. ¢. 186 (1836).

IMA: Abadia (Gen., phot. Gr.); stony ground on slopes of mountains, on
granite, alt. 300-750 m., Weberbauer, no. 35 (Brl., fragm. Gr.); on stony ground
of the loma formation, Amancaes Mountains, alt. 200-400 m., Weberbauer
no. 1591 (Brl., fragm. Gr.); sandy lomas along the sea, Lurin, Macbride, no.
we * Avent ao te 7. the 7 ase road, southwest of Matucana,

Witvoct spe Op tascecae: Dombey acc. to Lamarck.

15. $. Hoppii Robinson. Probably herbaceous and perennial (the

ase unknown); stems very slender, erect, glandular-puberulent
When young; leaves (except the uppermost) opposite, subsessile,
lanceolate-ovate, obtusish, crenate-serrate; heads (about 14 mm. long)
shortly pedicelled in flattish somewhat fastigiate terminal corymbs;

30 ROBINSON

phyllaries thin, green, glandular-puberulent, the outer acute, entire,
the inner usually subtrifid at the apex, a small mucro being perceptible
in the middle of a somewhat retuse summit ; corollas claret-colored at
least as to the subcylindrical throat, the lobes pale, probably pink;
achenes unlike, the 4 adelphocarps 14-16-awned, the idiocarp awnless
and crowned merely by an erose scarious border (0.5 mm. high).—
Contrib. Gray Herb. c. 6 (1932).

ArEQuIPA: on rocks, alt. 2500 m., Arequipa, Hopp, nos. 3 (Brl., phot. Gr.)
and 12 (Brl., phot. Gr.).

While not strongly marked, this very slender plant of the Are-
quipa region cannot be as yet connected satisfactorily with any other
species.

16. 8. petiolata (Cass.) Sch.-Bip. Smoothish much branched
herb; leaves opposite and (the upper) alternate, short-petioled, oval
to elliptical or obovate, obtuse to rounded at the tip, serrate from
about the middle, subglabrous to finely incurved-puberulent at least
on the nerves beneath, 2.5 cm. long, 8-15 mm. wide, cuneate and
entire toward the base; branches many, ascending, their terminal
glomerules of heads together forming a convex or flattish compound
corymb; heads sessile or nearly so, about 1 em. long; phyllaries
oblong-lanceolate, obtuse, very finely (sometimes obsoletely) in-
curved-puberulent; corollas apparently white; achenes unlike, the
4 adelphocarps 12-16-awned, the idiocarp awnless but crowned by
a scale-pappus.—Linnaea, xxv. 291 (1853). Nothites petiolata Cass.
Dict. xxxv. 166 (1825); DC. Prod. v. 186 (1836). Stevia ———?
Ball, Journ. Linn. Soc. xxii. 43 (1885).

Lima: near the city of Lima, Dombey ace. to Cassini (Dombey material of

0
Hist. at Paris bearing the number 507 (Par., phot. and fragm. Gr.), being
presumably “hick of the original collection from which Cassini drew his
; Chicla, in stony places, alt. 3 . r.}.
Cuzco: Proy. Canchis: Sicuani, alt. 3550 m., Hicken, acc. to Herrera, Con-
trib. Fl. Dep. Cuzeo 198 (1921), also Chlor. Gaze. 85 (1926). The writer has

id no opportunity to verify this record, first published by Hicken, Apunt.
Hist. Nat. i. 176 (1909).

WITHOUT EXACT LOCALITY: von Besser (Brl., trac. Gr.).

m.,

Under the genus Stevia, to which the plant manifestly belongs,
this species 1s very unhappily, indeed misleadingly, named. Its

in Stevia, where petioles are often much more conspicuously de-
veloped.

li a as aad a Ei

THE STEVIAS OF PERU 31

17. 8. cuzcoensis Hieron. Herbaceous or suffruticose, shortly
incurved-puberulent, usually branched from the base, the branches
often virgate, leafy, erect or ascending, 2-5 dm. high; leaves (except
sometimes the uppermost) opposite, rhombic-ovate to -lanceolate,
acutish, serrate except for the cuneately narrowed entire shortly
petioled base, 0.54 em long, 10-18 (-20) mm. wide; corymbs rather
dense; heads 10-12 mm. long, subsessile in closely clustered glomerules;
phyllaries green or often purple, acutish to rounded at apex, incurved-
puberulent; corollas mostly with deep purple tube and throat and
with roseate lobes; achenes unlike, the (3-)4 adelphocarps each
12-18-awned, the awns purple, about equalling the corollas, the 1
(-2) idiocarps crowned by a scale-pappus (1—1.5 mm. high).—Hieron.
in Engl. Bot. Jahrb. xl. 363-364 (1908); Herrera, Contrib. Fl. Dep.
Cuzco, 198, also Chlor. Cuze. 85 (1926) and Estud. Fl. Dep. Cuze.
187 (1930).

Cuzco: hills, Sacsahuamdn, alt. 3500-3600 m., Weberbauer, no. 4851 (Brl.,
phot. Gr.), same locality, Herrera, no. 2353 (Field Mus., fragm. Gr.), Pennell,
no. 13,553 (Gr., Field Mus.); rocky slope of cafion, San Sebastian, Pennell, no.

i erro

: f : .» Cook & Gilbert, nos. 362
(U. S., fragm. Gr.) and 489 (U.S., fragm. Gr.); Pennell, no. 13,675 (Gr., Field
Mus.), likewise from Ollanta: tambo, probably consisting of weak lateral
branches, also doubtfully referred to this species.

18. $8. punensis Robinson. Erect herbaceous perennial, smooth-
ish, 2-2.5 dm. high; stem (denuded below by the falling of the leaves)
very leafy above; leaves opposite, the cauline shortly petiolate,
thombic-ovate, acuminate, serrulate, cuneate and entire toward the
base, 3-3.6 cm. long, 1-1.5 em. wide; inflorescence subfastigiately
branched, the branchlets very finely glandular-puberulent; heads
sessile or shortly pedicelled, 15 mm. long; phyllaries acute to attenuate,
almost glabrous dorsally except for a few scattered sessile glands;
corollas white; achenes unlike, the 3 adelphocarps each 10-12-awned,
the awns purplish and about equalling the corollas, of the 2 idiocarps
1 crowned only with scales, the other with scales and about 3 unequal
awns all much shorter than the corollas—Contrib. Gray Herb. c.
8 (1932).

0: Prov. Sandia: Cuyocuyo, in high loosely oboe meadows, alt.

Pon
37 . and fragm. Gr.).
00 m., Weberbauer, no. 850 (Brl., phot. and fr. ears was referred with

ae ROBINSON

19. S. puberula Hook. Decumbent or suberect herb or under-
shrub, 0.3-2 m. high, finely puberulent with white incurved or crisped
glandless hairs; leaves (except sometimes the upper ones) opposite,
shortly petioled, the mature cauline broadly ovate to suborbicular,
crenate-serrate; heads (about 15 mm. long) sessile or shortly pedicelled
in dense corymbously disposed glomerules; phyllaries (8 mm. long)
incurved-puberulent to smoothish except for sessile glands; corollas
white or pink (Barclay); achenes unlike, the 4 adelphocarps each
about 10-awned, the single idiocarp crowned by a scarious border of
connate scales and 1-2 awnlets (1—-1.2 mm. long).—Bot. Miscel. ii.
225 (1831); DC. Prod. v. 123 (1836).

Lima: Valley of Canta, Cruckshanks (K., phot. Gr.), Wilkes Exped. (U. 8.)

open rocky banks along Rio Chillon, above de a oa ro 14,367
t Noee
Yan

Gr.), e
F ei yrs nos. 157 (Gr., N. Y., Field Mus.), 314 (Gr., Field Mus.) and 482

ARTMENT NOT INDICATED: Serra de Puruchucas, Barclay, no. 2266 (K.).

20. S. Pearcei Robinson. Tall erect herbaceous perennial with

H : sThat » KC,
Pearee,no. 104 (phot. and small fragin- Gr)” unnee™ Mule

A species named for its discoverer, Richard William Pearce, an
English horticultural explorer employed by Veitch.

21. 8. Weberbaueri Robinson. Procumbent or ascending under-
shrub, glandular-puberulent above, 3 dm. high; leaves opposite,
sessile or nearly so, lanceolate to narrowly rhombic-ovate, acute to
obtuse, serrate except for the entire cuneate base, membranaceous,
2-3 cm. long, 6-10 mm. wide, shortly puberulent; heads 14-16 mm.
long, sessile or nearly so in dense corymbously disposed clusters;
phyllaries narrowly oblong, shortly pointed, 8-8.5 mm. long, dorsally
glandular-puberulent; corollas slender, roseate with paler lobes;

THE STEVIAS OF PERU 33

achenes unlike, the 4 adelphocarps each 10—-16-awned (awns about
equalling the corollas), the idiocarp awnless, coronulate with short
more or less connate scales.—Contrib. Gray Herb. xevi. 16 (1931).

Mogvunreua: rainy-green formation, alt. 3200 m., Carumas, W. eberbauer, no.
7293 (Field Mus., Gr.).

Lima: Rio Blanca, open hillsides, alt. 3000-3500 m., Killip & Smith, no.
21,673 (Gr.); near Oroya, alt. 3000-4000 m., Kalenborn, no. 18 (U. 8., Mo.).

22. §. Lechleri Hieron. Perennial, herbaceous or perhaps suf-
fruticose; stem simple to the inflorescence, leafy, glabrate below,
finely glandular-puberulent above; leaves opposite, broadly though
rhombically ovate, obtusish, crenate-serrate except for the cuneate
petioliform basal portion, nearly glabrous except for scattered hairs
on the nerves beneath; heads sessile in close clusters ; these terminating
the curved-ascending branches of the rather loose inflorescence;
phyllaries 6 mm. long, acutish, subscarious, purplish toward the tip,
dorsally glandular-puberulent; achenes unlike, the 3 adelphocarps
6-11-awned, the 2 idiocarps exaristate but surmounted by a coroni-
form pappus.—Hieron. in Engl. Bot. Jahrb. xxii. 727 (1897).

DEPARTMENT NOT INDICATED: in grassy places, Tabina, Lechler, no. 1882
(Brl., phot. Gr.).

With this little known plant, it has not been found possible exactly
to match any subsequently collected material. A more satisfactory
knowledge of the species must await its rediscovery and the examina-
tion of further specimens.

23. 8. cajabambensis Hieron. Suffruticose, 5-6 dm. high, erect
or curved-ascending; leaves (except the uppermost) opposite, the
lower cauline broadly ovate, obtuse, coarsely and irregularly crenate-
dentate except for the entire cuneate petioliform base, membranaceous,
above sparingly and very shortly puberulent and sprinkled with
sessile glands, beneath sparsely villous; heads closely sessile to dis-
tinctly pedicelled in rather small loosely and corymbously disposed
Sometimes subfasciculate clusters; phyllaries narrow, acutish to
shortly acuminate, about 9 mm. long, densely beset on the back
with stipitate glands; corollas probably roseate; achenes unlike, the
4 adelphocarps each 8-10-awned, the idiocarp awnless and crowned
by somewhat connate scales.—Hieron. in Engl. Bot. Jahrb. xl. 367
(1908).

ANcacus: below the Hacienda Cajabamba, Weberbauer, no. 3124 (Brl.,
phot. Gr.),

Lima: Huamantanga, Mathews, no. 563 (Gr., N. Y.).

This is very close to S. puberula Hook. and may well prove only
a glandular form of that plant. However, until these species are

34 ROBINSON

better known and transition has been demonstrated, they may be
maintained.

24. §. Mandonii Sch.-Bip. Dwarf matted high Andean perennial;
stems prostrate or ascending, rarely over 1 dm. long, leafy; leaves
opposite, sessile or nearly so, elliptical, obtuse or mostly rounded at
tip, shallowly 1-3-toothed on each side above the middle, entire
and cuneately narrowed toward the base, subcoriaceous, 1-1.5 cm.
long, 5-7 mm. wide; heads (about 12 mm. long) closely clustered in
single terminal sessile glomerules; corollas with red-purple tube and
throat and pink or white lobes; achenes unlike, the 4 adelphocarps
each about 19-awned, the idiocarp about 9-awned and crowned also
by 1-3 short scales.—Bull. Soc. Bot. Fr. xii. 81 (1865) and Linnaea,
xxxiv. 535 (1865-6), both without char.; Robinson, Contrib. Gray
Herb. Ixxvii. 6 (1926), where characterized.

Puno: in open mats, common, alt. 3700 m., Weberbauer, no. 431 (Brl., phot.
Gr.) ; Juliaca, alt. about 3675 m., R. S. Williams, no. 2504 (N. Y., fragm. Gr.);
Chuquibambilla, rocky clay slope over limestone, alt. 3900-4000 m., Pennell,
no. 13,372 (Gr., N. Y., Field Mus.).

[Botrvia.]}

EXCLUDED, TRANSFERRED OR DOUBTFUL SPECIES.

S. BREvIPAPPOSA Hieron. in Engl. Bot. Jahrb. xxii. 718 (1897).
This species, of which the writer has examined authentic material
both at New York and Berlin, seems to him to be conspecific with
S. BOLIVIENSIS Sch.-Bip. S. brevipapposa is reported by Herrera,
Chlor. Cuze. 183 (1926) and Estud. Fl. Dep. Cuzc., 186 (1930), as
occurring near the city of Cuzco. The writer has had no opportunity
to verify this record.

S. cortacea Lag. Gen. et Spec. Nov. 28 (1816). This species was
described as follows: “pappo paleaceo mutico: foliis lato-lanceolatis
argute serratis oppositis ternisve subcoriaceis.”’

“ Pedunculus communis terminalis, apice umbellatus, radiis 3-5
apice corymbosis, corymbis multifloris, compactis, foliosis. F olia
floralia ad pedunculorum pedicellorumque basim linearia, acutissima
integerrima.”’

“Hab. in Peruviae Regno, ubi legit. D. Ludovicus Neé. b ”

No Stevia corresponding to the brief character here given is known
to the writer, nor so far as he is aware has any such plant been sub-
sequently reported from Peru. Unfortunately there is evidence
that the collections of Neé were geographically confused. So the
localities attributed to his specimens are always subject to doubt.
Under these circumstances it has seemed best to omit S. coriacea
from the foregoing treatment of the Peruvian Stevias.

THE STEVIAS OF PERU 30

S. CRENATA Benth. Pl. Hartw. 197 (1845). The Peruvian plant
referred to this species by Hieronymus in Engl. Bot. Jahrb. xxxvi.
463 (1905), namely von Jelski, no. 668, has not been available for
study during the preparation of the present paper. The writer has
seen no evidence that the real Ecuadorian S. crenata Benth. extends
to any part of Peru and is disposed to think that the plant so recorded
by Hieronymus may well have been a form of the later characterized
and habitally similar S$. andina Robinson.

S. ELATIOR HBK. Nov. Gen. et Spec. iv. 144 (1820). The multi-
aristate plant referred to this species by Hieronymus in Engl. Bot.
Jahrb. xxxvi. 463 (1905), namely von Jelski, no. 669, is S. ANDINA
Robinson.

S. TRACHELIOIDES DC. Prod. v. 115 (1836). This plant of Mexico
appears to the writer to be indistinguishable from the earlier S.
MONARDAEFOLIA HBK. By an unfortunate error the name S. tra-
chelioides DC. was used by Gray when attempting to identify some
rather poorly prepared specimens of Peruvian plants collected by
the Wilkes Expedition. Apparently from this early misidentification
the binomial S. trachelioides has to some extent persisted on herbarium
sheets and is applied to some Peruvian plant by Professor Herrera
in his Estud. Fl. Dep. Cuzc. 187 (1930).

EXsICCATAE OF PERUVIAN STEVIAS

Abadia Dombey
—— melissaefolia (Lam.) Sch.- 507 petiolata (Cass.) Sch.-Bip.
sy one anaes —— melissaefolia (Lam.) Sch.-
Ball sip. hoe
—— petiolata (Cass.) Sch.-Bip. —— petiolata, (Cass.) Sch.-Bip.
Barclay © Espesto
2266 puberula Hook. 127 Macbridei Robinson
Besser, von rapt
— i -Bi 355 Macbridei inson
Re Veen Sey —— Macbridei v. anomala Rob-
Cook & Gilbert inson
362 cuzcoensis Hieron. 'ienke
363 Macbridei v. anomala Rob-
inson — polyphylla DC.
408 galeopsidifolia Hieron.
414 galeopsidifolia Hieron. _ Herrera
489 cuzcoensis Hieron. 893 urticaefolia Billb.
1849 Macbridei v. anomala Rob- 1475 cuzcoensis Hieron.
inson 353 cuzcoensis Hieron.

: —— Herrerae Robinson
Cru —— rhombifolia_v. stephanoco-
—— puberula Hook. Bip.

ee aa

36 ROBINSON

—— multiaristata Spreng.

Hicken
petiolata (Cass.) Sch.-Bip. ——— rhombifolia v. stephanoco-
ma ip.
Hopp rhombifolia Be uniaristata |
3 Hoppii Robinson (DC.) Sch.-Bip. |
|

12 Hoppii Robinson Pear

Jelski, von 104 Pearcei Robinson
669 andina Robinson Pennell
13,372 Mandonii Sch.-Bip.

Kalenborn :
18 Weberbaueri Robinson ratensis Hieron.

a 13,553 cuzcoensis Hiero
Killip & Smith 13,622 cuzcoensis Hieron
21,673 Weberbaueri —— 13,675 cuzcoensis Hieron
21 768 Macebridei Robin 13,883 Pennellii Robinson

22;158 (in part) M bridei Robin-
(i ft rt) Macbridei Robin 14,367 puberula Hook.

22,158 ior part) — v. anom-
22,288 feeling Hotinsen

Lechler
1882 Lechleri Hieron.
Macbride
5965 oo (Lam.) Sch.-

Macbride & Sc olga

McLean
rhombifolia _v. stephanoco-
Sch.-Bip.

Martinet
—— puberula Hook.

Math
563 ca jabambensis Hieron.
— n

-—— Herrerae Robinson

Poeppig
—— cathartica Poepp.

Townsend
1521 Macbridei v. anomala Rob-
inson

Weberbauer
35 melissaefolia (Lam.) Sch.-
ip.
198 —— (Lam.) Sch.-
431 Mandonii Sch. -
850 punensis Robinso
1591 melissefolia (Laud Sch.-

3124 eajabambensis igi.
3860 pabloensis Hier

atensis on.
7293 Weberbaueri Robinson

Wilkes Exped.
—— Macbridei Robinson
—— puberula Hook.

Williams, R. 5.
2504 Mandonii Sch.-Bip.

THE STEVIAS OF BOLIVIA.

For some ten or twelve years there has been accumulating at the
Gray Herbarium a considerable mass of undetermined Bolivian

Stevias. This has been due on the one hand to accessions, received ef

purchase or exchange, and on the other to loans of Eupatorieae ™m0s

THE STEVIAS OF BOLIVIA of

kindly permitted by several of the larger herbaria, both of Europe
and America.

Attempts to classify this unnamed material have been made at
different times, but have, until recently, led to exceedingly unsatis-
factory results. It became clear that the species were numerous,
Bolivia having, in fact, a richer Stevia-flora than that of any other
country except Mexico. Some of these species were evidently un-
described and several were represented only by tantalizingly frag-
mentary specimens, sufficient to show their probable novelty, but
inadequate for use as taxonomic types. Though a high endemism was
to be expected, there was unavoidable suspicion that some of these
plants might well be merely the Bolivian representatives of species
or varieties already based on extra-limital material. The literature
relating to the Bolivian Stevias was scanty and furnished no keys and
few carefully worked out contrasts even among the better known
species.

On the whole it seemed best to defer further attempts at revision
of the Bolivian members of the genus until those of surrounding
countries, notably of Peru, Argentina and Paraguay, could be given
at least preliminary elaboration. Thus, for some years, the best
that could be accomplished was to make such preparations as were
feasible. In the first place, needful photographs of the types of the
Bolivian Stevias were obtained as far as possible by purchase and
exchange. During the writer’s journeys to Europe in 1927 and 1928
he was able to visit, at least briefly, several of the leading herbaria
of the world, notably those of Kew, the British Museum (both in-
cluding the collections of Bridges and of Pearce), the Museum” of
Natural History at Paris (rich in the collections of Mandon and
including the herbarium of Schultz-Bipontinus), the herbaria of
both the Conservatory and University at Geneva, the Botanical
Museum at Zurich, and the Botanical Museum at Berlin-Dahlem,
where Hieronymus worked and most of his numerous types are
preserved. At these establishments the Stevia material was examined
and many dissections, kindly allowed by the management, were
made and recorded by sketches and notes. Through lack of time
the work had to be done hurriedly, but loans of much critical material
were amiably permitted so that it could receive more leisurely study
at Cambridge. ‘

The larger American herbaria have shown also great readiness to
Place their Stevia material at the service of the writer. Especially
important in this connection was the New York Botanical Garden,
with its unexcelled representation of the notable Bolivian collections

4

38 ROBINSON

of Rusby, Bang, R. S. Williams, and of the Mulford Expedition, as
well as the herbarium of Kuntze, including the types of the Bolivian
species based on his collections. The United States National Her-
barium lent, together with much other material, an extended series
of the Bolivian Stevias collected by Dr. Otto Buchtien, Dr. Erik Asp-
lund and Mr. Ernesto Giinther. Mr. Cornelio Osten of Montevideo
was so kind as to lend the entire representation of the Eupatorieae
in his large private herbarium, exceptionally rich in extra-tropical
South American plants. Dr. H. H. Rusby with great generosity
lent for study, criticism, and publication if desired, the types of
several Stevia novelties which he had observed. The Missouri Bo-
tanical Garden and the Field Museum of Natural History have also
lent their Stevia material both of Peru and Bolivia.

From these various sources it seemed likely that the writer had
seen a high percentage of the Stevia specimens thus far deposited in
herbaria. But, as above indicated, preliminary studies of the genus
as represented in the surrounding countries had to be made. These
have now been completed and, in part, published, at least as to those
: the Argentine Republic, Paraguay and (in the present paper) of

eru

Notwithstanding this preliminary work, extending over some years,
an elaboration of the Bolivian Stevias has proved, during the last
few months, a task of a very discouraging nature. In no country
previously studied did.the Stevias show such perplexing inconstancy
and transition. Thus, for instance, between such very diverse types
as those of the upright, narrow-leaved, mesophytic S. Bangi and
the low, matted, elliptic-leaved alpine xerophyte S. Mandonii, @
practically complete series of intermediates can be demonstrated.
The monographer is thus confronted by a serious dilemma: on the
one hand, the extreme elements are here far too unlike to be con-
vineingly united as conspecific ecological phases of the same plant,
and on the other hand, their separation seems, at best, vague and
artificial.

The treatment of the Bolivian Stevias here presented is confessedly
unsatisfactory, but it is put forth in the confident hope that it will
prove useful. Its key should be followed with all due caution, and
the user should bear in mind that at almost any point he is likely to
encounter exceptions of a perplexing nature. If he becomes disgusted,
let him try to identify his Bolivian Stevias by reference to previous
literature, and, if he has the same experience as the writer, he will
return to the key with the feeling that even an obscurely blazed trail
is better than no guidance at all through this taxonomic jungle.

THE STEVIAS OF BOLIVIA 39

Of the 41 species and 7 (non-typical) varieties here recognized, no
less than 39,.that is to say about 81%, are, to our present knowledge,
endemic to the country. Of the rest only 3 extend into Peru and 6
into northern Argentina. Strangely enough, even the Bolivian
representatives of such widely distributed species as S. urticaefolia
and S. elatior are endemic varieties.

Notwithstanding this high endemism and the exceptionally rich
representation of Stevia in Bolivia, the country cannot be regarded

a natural center or distributional area for the genus. This be-
comes evident on study of internal distribution. The Stevias fall
with astonishingly few exceptions into three well nigh mutually
exclusive groups, as follows: 1) those of central and western Bolivia
(Departments Cochabamba and La Paz); 2) those of extreme southern
Bolivia (Departments Potosi, Tarija and Gran Chaco); and 3) those of
eastern Bolivia (Department Santa Cruz). To the surprise of the writer,
he has been thus far unable to find any conclusively identical elements
between the known Stevia floras of Bolivia, on the one hand, and of
Paraguay or Brazil, on the other. Obviously this must be due solely
to our very imperfect knowledge of campestrine eastern Bolivia and
of the adjacent portions of Brazil and Para

It has seemed best to include in the present treatment some two
or three species, founded by Hieronymus upon plants of Fiebrig and
attributed by him to Bolivia, which, however, coming from Toldos
were, as it now appears, collected just across the southern frontier
of the country, and actually in the Argentine Republic. Their
inclusion seems justifiable on the ground that future exploration is
almost certain to show their presence on both sides of this botanically
artificial, and even politically somewhat vague line.

In conclusion it must be said that the extent of Bolivia is little
realized by most Europeans and North Americans. For the former
it may be stated that the country about equals the joint area of France,
Germany and Italy; and for the latter that it somewhat surpasses
the combined extent of California, Nevada, Utah and Arizona. With
the highest appreciation of the courage, energy and diligence of those
who have taken part in its botanical exploration, and the value, as
well as the extent, of their services, it must be admitted that we as
yet know relatively little of the floral wealth of this great country
with its vast diversity of altitude, humidity and varied terrain.

Key To THE Stevias OF BOLIVIA.
A. Pappus either awnless and composed of short sub-equal
roam somewhat connate scales or (at least in ‘thie adelpho-
ps) a Fe awns about as long as the corollas
$i 1 hoeuve.

40 ROBINSON

B. priciecupenas loose; —— ne cheer ig equalling or often
much exceeding the in a

C. Laer asatty alternate) "nee iaisegiade pee 21. S. Fiebrigii.
C. Leaves (chiefly o woe ovate or rhombic. .
D. Achenes Saale homomorphic and all awnless, but

occasionally het sonoma th 1-2 achenes each
1—2-awned; corolla-throat pairs e.
precise v. pallidiflora.
D. Achenes heteromorphic; adelphocarps, 2-4-aw
corolla-throat deep rose to purple.

aves membranaceous, 2.2-3.2 cm. m. wide, shortly
setose-pilose on the veins Daonth: the veinlets not
exserte

F. Suberec ts main cauline leaves rhombic-ovate, the
blade gradually narrowed to the base; outer
phyllaries glandular- Saiber dices

ety oe the blade erm or epee Cie
cte a scarcely sy: gia petiole. a ee elatior v. austrina,
E. Leaves Maboorinneca us, 1-1.8 ¢ ide
= y hirtellous on the veins (beneath. thie yeiniets :
Pewee Ur Vice Lose ak os pe os ORR S. filipes

B. fatloonencs or its - arts much more sich heads ses-
e or on pedicels s vena are the involuc ae fA
G. Achenes of the sa d alike as to ae «Ff
H. Tappan incr ag er short subequal oie without
Sf "Phyllatied —t with short pees rea hag gland-
ip heads densely clustered... .
J. Leaves “3 most 2.5 em fe “6 12 a wide,
baa scurely 3-nerved, short-pilose on both sur-
nding wee ee fe pe baa pe uenry ar nan tons 4, S. tunariensis.
becoming 5 cm. long and 2.5-3 ¢
with 3 strongly exserted nerves ; the saslaee we

abrous
aries subappressed- or incurved-puberulent
with glandless hairs; heads more loosely clustered
oe somewhat scattered along the inflorescence-
WU e ugh ceecen ee ee nin ay Sateen: 6. S. neglecta
H; 1 wr both scales and awn
low, decumbent; it on ‘elliptic-oblong,
btuse, We ee 7. S. Chamaedrys.
K. Plant erect, 3 dm. or more in height; leaves acute,

mS

. Leaves deltoid-ovate, about two-thirds as wide a
ng: F phyllaries beset with straight ee
net gear, ee ee eee
. Leaves ay preg meg tag and about two-
thir vcd pg wide g but phyllaries seat
ee ee iret

. S. glomerata.

mS

46 to. pappu
or 4 . bors bearing a ton us,
the fifth calvous. . ' Papp

THE STEVIAS OF BOLIVIA

41

N. Leaves ovate, petiolate, serrate............ 10. S.
N. Leaves lanceolate, subsessile, entire........... 112.4, ome.
0.

M. Some of the achenes awne
. Leaves elliptic-obovate, mostly obtuse or even
— ed at tip; stems subsimple, sero nake
S.

d
OVE a et ae ee obovata v. aristifera.

0. Leave ovate or lanceolate, acute or south; stems
usually much branched and leafy to the su
P, Leaves small, mostly less than 1.5 em. wi de

Q. Leaves ovate, subcoriaceous, ws acamesge rd
mee ate-lanceolate, sharply ser

COLUATG DASE Se ae eee 14.
R. Sarees. distinetly petioled or sessile by a narrow

ase.

S. Leaves suborbieular-ovate, Phineas or barely

acutis ere may ught occasional

longer-awned forms of dae: hed thern Bo-

S. Leaves ovate or more often rhombic-ovate
or -lanceolate, acute or attenuate... .7.

T. Heads mostly pedicelled, borne in loose

glomerules or fastigiately branched inflor-

escences.

U. Leaves sparingl and shortly crisped-

or ek igh esi ose on nerves and chief

veins neat. corollas not dark-

2. S. Bridgesit.

S. camachensis.

POU re ss eet 31. S. boliviensis.

U. Leaves copio ae setose-villous beneath;
— rollas dark-punctate toward the

MG ee ci ver ieee 15. S.

omerules.
V. Leaves firmish chartaceo-me at ed
reek og veins and veinlets exse:

cap Hosts nearly or quite sessile in dense
gl 1 te

aces.
WwW. Corolins reddish-violet; phyllari
nsely granulated ted with s ecealle
= usually purplish; eee 2-3.5
OOM 6 A

W. Corollas —— ehowieng pale green,
A debi ck Oa is urved-puberulent,
d-dot nave: 5-7 em. long.

16. S. totorensis.

17. S. reclinata.

Leaves membranaceous, not prominent-
ly retic
. Glomerules clustered at the a of
the branches and s

surrounding leaves....18. S. ere ae

42 ROBINSON

X. Glomerules Eeeunailate, loosely dis-
osed in nearly naked corymbs, not
equalled by the surrounding leaves

Y. Leaves ovate-lanceolate, sessile by
adual rchiaks wed base; involu-
cre half ¢ wo-thirds as long as

short spreading Ene oF ed
PRES POR F alamechse's:

narrowe a scarcely einen
+9 involuere nearly as long
as the florets; phyllaries spread-

glandlesshairs.............. S. kuhnioides.
A. Pappus Sabi least of the adelphoearps) ri goers of scales eth
1-3 longer and firmer om Sen mostly unequal and much
shoctar | isha the etna mane true a not normally present... Z.
Z. Inflorescence a diffuse panicle; pedicels equalling or exceed-
ing the involucre. .. .
a. Leaves (mostly slieriaie) linear to linear-lanceolate,

ee AN ON OB Wide... 21. S. Fiebrigit.
a. Leaves (chiefly opposite) rhombic-lanceolate, 2-3 a
kk hk . S. yaconensis.
Z. pretaier ga Bid its parts much pees pedicels oe -
normally ee than the Inv olucre....
b. Leavy via or nearly so, ite 2 as much as 1.5 cm.

c Leaves ecaeslaie. attenuate; plant suberect, 3-4 dm.
PING CCE seh ooh « cnt jwelnn bi «cae 23. S. clivicola.
c. — obovate to — pom or obtuse at tip;

nt decumbent, 1-2 Ee eae 24. S. obovata.
b. Strleaves plete petiolate or provided with a cuneate
usually a orm base, the blade Shore fet than 2
cm. wide....
d. Many-ste

S MASS Ce Cee ROME CEs Saude «<4 vei. sss 4 cues . S. Benderi.

= in _ ait

é. ves Pai se acid v labrous on both surfaces;
Southern Bo er EES SEO TIEEE 26. S. bermejensis.
e. Leaves = ney pie ewhat —— either above or (at

least one vf the nerves) b
J. ina pare of the leaf-blade i in mature cauline
eaves passing abruptly into - very narrowly
ah r wingless petiole
q: taares y me from above the base . the
lade, slightly carneous; Southern Bolivia. . . .h.
h. Fr hyllaries sparingly setulose or villulose, not .
_ Stipitate-glandular... .7
. Leav ves suborhicular-ovate, obtuse or neeey
acu at —_ 5 em. long; florets muc
Spice eran eT
. a ne often e-tin y ;
“tires weap me ge

THE STEVIAS OF BOLIVIA 43

nish-stramineous, rather abruptly
Spt nee. oe 58 BE 28. SS. eclipes.
h. Phyllaries (at least the outer) beset with
stipitate glands; geo deep rose to violet.
29. S. calderillensis.
g. Leaves 3-nerved from the ve ase of the
blade, thin-membranaceous; Cantal Bolivia

0. S.
f. Expanded part of the leaf-blade in mature cauline
sie vee: gradually Renin below to a cuneate
joliform base... .j
9: Phyllanes obtuse, erose-dentic ulate at summi
sparsely ee vebernient sareelie
Southern Bolivia............ ecm oe
j. Phyllaries acute Jf acuminate, finely incurved-
or rims asd uberulent or subglabrous; West-

galeopsidifolia.

ern ee Bouvias souk ee oisa D. Sudice 31. S. boliviensis.
A, Pappus (at least of the adelphoearps) multiaristate, the ( (6-)
long a llas ULTIARISTATAE. .. .k.

k, Leaves narrowly oblanceolate to linear, 7-20 times as long
L Teaven (in Bolivian forms) linear or nearly so, acutish,
membranaceous; internodes mostly 1.5-2.5 ¢ m. long:
slemonegics disposed in an cb spreadingly Nese!
corymb; heads i 1 cm. long; achenes 3.2 mm. long.
2. S. mercedensis.
l. Leaves agree aide: thickish; internodes mostly less than
le Sue close y pac in a rope
cory; poe 8-9 mm. long; achenes about 2

ene
PURSE Wt are eee ere es er a S. iasisitpciai
k, Leaves: lanceolate to broadly we or elliptical, ais
-5 times as long as wide.
m ‘Oa phyllaries closely biert a short-st coe glands

es small, even the main cauline only at cm. long,
; oriaceous, obtuse; eticulated ;
ricer te prominent on both surfaces.............. 34. S. Kuntzet.

n. Leaves larger, the main cauline mostly 4-6 cm. long,

etioled, ME DTARANOOUS, acute; veinlets not con- :
spicuously ¢ "oh ay MEER ig ee ee .. 85. S. selifera.
m. Outer + ge illu
2 - eit mop pe hairs; pre ‘f oe grenlige ses-

0. Petiolee together with the contracted pociensare P bases
the main cauline leaves mostly 2-3 cm. nme . S. sarensis.
0. Petioles together with se contracted petioli
main ca

° uline leaves about 5-8 mm. ig A
Figs Song rag oa (1-3 mm. in length) in a
flat ttish prin get a ri d ond
oncolorous, nigrescen

ee ee eee
ee e in paniculately disposed glomerules;

1 vent discol lmost white beneath.
seasil Ww.
eaves strongly olorous, alm $0 disualey.

44 ROBINSON

serrate, membranaceous or nearly so............. 9. S. Bangit.

ate, slightly fleshy or subcoriaceous. . . .r.
r. Glomerules disposed in a more or less expanded
corymb, the heads much exceeding the mostly
lanceolate or narrowly oblong surrounding

3
8
B.
sg
* Ba
=
eo
ao
oO
=
o
mn
jan
a
°
i=]

1. 8. urticaefolia Billb. For description see p. 24 —Billb. in
Thunb. Pl. Bras. Dec. i. 13 (1817).

Of this species the typical variety, with awnless pappus and purple
corolla-throat, does not seem yet to have been found in Bolivia,
though the following varieties occur there at least sparingly.

Var. pallidifiora Robinson. Pappus a short erose or toothed
crown of connate scales, uniform in all 5 achenes or in 1-2 supple-
mented by 1-2 awns; corolla-throat yellowish or greenish.—Contrib.
Gray Herb. xevi. 16 (1831). Listed by Rusby, Bull. N. Y. Bot.
Gard. iv. 377 (1907), as S. urticifolia Thunb.

La Paz: Prov. Larecaja: sunny slopes at Hacienda Casana, Tipuani Valley,
alt. 1400 m., Buchtien, no. 7544 (U. 8., Gr.).

DEPARTMENT NOT INDICATED: without data of collection, Bang, no. 2877
By N. Y.); at Nequejahuira, alt. about 2440 m., Tate, no. 669 (N. Y., phot.
T.).

with a short scale-pappus and 1-3 awns (about equalling the corollas
or some of them shorter).—S. Clausseni var. boliviensis Hieron. in
Engl. Bot. Jahrb. xxii. 723 (1897).

CocHaBaMBa: Prov. Chapare: t Barbechos Incachaca, alt. 2250 m.,
Steinbach, no. 9147 (Gr., Mo). Be Y

DEPARTMENT NoT INDICATED: alt. 2600 m., Kuntze (Brl., phot. Gr.; N. ¥+
phot. Gr.).

2. §. elatior var. austrina Robinson. Loosely decumbent and
curved-ascending, hirsute with long white jointed hairs (at maturity

THE STEVIAS OF BOLIVIA 45

glandless); the lax inflorescence both hirsute and glandular-puberu-
lent; corolla-throat dark, probably purple, the lobes pink or white;
achenes unlike, the 4 adelphocarps each 5-awned (their awns about
as long as the corollas), the idiocarp awnless.—Contrib. Gray Herb.
xevi. 7 (1931). 8. elatior HBK. as reported by Hieron. in Ktze. Rev.
Gen. iii. pt. 2, 180 (1898), but by no means typical.

a hte Cruz: Sierra de Santa Cruz, alt. 1600 m., Kunize (N. Y., U. S., phot.
ir).

This puzzling plant differs strikingly from any form of the common
S. elatior HBK. known to the writer from north of the equator,
being weaker, more decumbent, and much more coarsely pubescent.
However, its characters are not really strong and it may stand pro-
visionally as a geographically isolated variety of the widespread
and more northern plant.

3. §. filipes Rusby. Slender perennial herb, slightly lignescent
toward the mostly inclined base; the loosely paniculate inflorescence
with filiform dark purple glandular-puberulent branches and pedicels;
internodes short, mostly exceeded by the leaves; leaves opposite,
the main cauline rhombic-ovate, crenate-serrate except on the con-
tracted cuneate-winged petioliform base; blades subcoriaceous,
sparingly hirtellous on both surfaces, plane or nearly so above,
slightly paler and conspicuously reticulate with prominent nerves
and veins beneath; heads about 1 cm. long; phyllaries linear-lanceo-
late, acute, dark purple, finely glandular-puberulent; corolla-throat
deep purple, the lobes nearly white; achenes unlike, the 4 adelpho-
carps each 3-4-awned, the idiocarp awnless, surmounted by a shallow
erose or denticulate cup of connate scales (about 0.3 mm. high),—
Bull. N. Y. Bot. Gard. viii. 126 (1912).

La Paz: Prov. Caupolican: on hills near A lo, alt. 1830 m., R. S. Williams,
no. 124 (N. Y., phot. Gr.; U. 8., fragm. 754 4g

4. 8. tunariensis Hieron. Curved-ascending perennial herb with
slender leafy crisped-puberulent to shortly pilose at length glabrate
stem; leaves opposite, sessile or nearly so, acute, serrate except at the
entire cuneate base, membranaceous, shortly crisped-pilose on both
surfaces, 1.5-2.5 em. long, 8-11 mm. wide; heads about 1 cm. long,
sessile in dense pedunculate glomerules; these rather closely clustered
in a small terminal corymb; achenes alike, crowned by a short hyaline
laciniate-toothed border of connate scales.—Hieron. in Engl. Bot.
Jahrb. xxii. 713 (1897).

Cocuapamsa: Proy. Tunari: at Tunari, alt. 3000 m., Kuntze (N. Y., phot.
Gr., fragm. Brl., tracing Gr.).

46 ROBINSON

5. 8. Stuebelii Hieron. Erect opposite-branched glandular-
puberulent perennial herb; leaves opposite, subsessile or shortly
petiolate, lanceolate-ovate, narrowed to an elongate but at tip ob-
tusish apex, crenate-serrate except for the entire cuneate base, 3-
ribbed from above the base, sparingly puberulent but finally gla-
brescent; heads sessile or shortly pedicelled in small glomerules;
these disposed in somewhat fastigiately and irregularly branched
terminal corymbs; phyllaries dark purple, narrowly oblong-lanceolate,
obtuse to shortly pointed, dorsally beset with spreading straight
gland-tipped hairs; corolla-throat purple, the lobes much paler,
probably roseate, hispid on the back; achenes alike, crowned by a
laciniate border of more or less connate scales (0.5-0.8 mm. high).—
Hieron. in Engl. Bot. Jahrb. xxi. 328 (1895).

La Paz: Prov. Yungas: above Taca in the valley of the Yungas, Sitibel, no.
52d (Brl., phot. and fragm. Gr.). :

6. S. neglecta Rusby. ‘all perennial herb; stem terete, bright
red to brownish-purple, pithy, incurved-puberulent, becoming as
much as 7 mm. thick; internodes mostly 3-8 cm. long; leaves opposite,
thombic-ovate, acute, serrate, the main cauline rather abruptly

(about 9 mm. long) mostly short-pedicelled, in small rather loose
terminal glomerules or somewhat scattered along the branchlets of
subfastigiate panicles; phyllaries linear-oblong, acute, substramineous,
incurved-puberulent with glandless hairs interspersed with some
sessile glands; corollas with slender fuscous proper tube and ochro-

Mem. Torr. Bot. Club, iv. 209 (1895). S. stenocephala as used by
Rusby, Mem. Torr. Bot. Club, iii. no. 3, p. 52 (1893), not Sch.-Bip.
ms inal — of La Paz, alt. 3050 m., Bang, no. 611 (Gr., N. Y., U.S.
7. S. Chamaedrys Griseb. Low perennial herb, several-stemmed
or branched almost from the usually decumbent base; stems slender,
flexuous, leafy, mostly 1-2 dm. long; leaves opposite, sessile or nearly
so, rhombic-ovate to elliptic-oblong, obtuse to rounded at tip, shallow-
ly crenate to entire, 1-2 cm. long, 5-8 (-12) mm. wide, membra-
naceous; heads (about 12 mm. long) sessile or subsessile in terminal
solitary or clustered glomerules; phyllaries lanceolate-oblong, acute,
dorsally pubescent; corollas with purple tube and throat, the lim
pink or white; achenes crowned by a scale-pappus (0.4 mm. high)
with or without 2-3 delicate bristle-like awns (sometimes as long as

THE STEVIAS OF BOLIVIA 47

the corollas).—Goett. Abh. xxiv. 167 (1879); Hieron. in Engl. Bot.
Jahrb. xxii. 729 (1897); Robinson, Contrib. Gray Herb. xc. 68 (1930).
Tarisa [?]: alt. 3000 m., Fiebrig, no. 3423 in part (Brl., phot. and small
fragm. Gr.).
[ARGENTINA. ]
8. 8. glomerata Hieron. Suffruticose, probably 4 dm. or more in
height; stems leafy, glandular-puberulent and villous; leaves opposite,
sessile, acute, crenate-serrate, the main cauline ovate or subrhombic-
ovate, more or less cuneate and entire at the base, puberulent above,
somewhat pubescent beneath, 3-4 em. long, 1.8-2.5 cm. wide; heads
sessile or nearly so in close glomerules; these shortly stalked, caly-
culate-bracteate, and disposed in a dense rounded terminal corymb;
phyllaries lanceolate-linear, acuminate, 8 mm. long, pale and greenish-
substramineous toward the base, purple near the tip, densely beset
dorsally with gland-tipped hairs; corollas deep purple; achenes
crowned with short connate scales and about 2 unequal purple awns.—
Hieron. in Engl. Bot. Jahrb. xl. 357 (1908).

TariJa: Prov. Arce: Camacho, Fiebrig, no. 2868a in part (Brl., phot. Gr.).
TMENT [proba ij OT IFIED: Calderi re alt. 3000 m.,

DEPaAR bly Tarija] NoT vER
Fiebrig, no. 3423 in part (Brl.), poorly developed and doubtful.

9. 8. triaristata Hieron. An erect perennial herb, 6 dm. or more
in height, very similar to the preceding but having the pedicels and
branches of the corymb somewhat lanate, the stem-leaves much more
definitely petioled, and the achenes all 3-awned as well as crowned by
a short scale-pappus.—Hieron. in Engl. Bot. Jahrb. xl. 358 (1908).

SouruerN Bonrvia [probably Tarija]: Fiebrig, no. 3541 (Brl., phot. Gr.).

10. S. soratensis Hieron. For description see p. 25 —Hieron.
in Engl. Bot. Jahrb. xxviii. 560 (1901). S. grandidentata Sch.-Bip.
Bull. Soc. Bot. Fr. xii. 81 (1865) and Linnaea, xxxiv. 535 (1865-66),
without char. in either place; first described by Hieron. 1. ¢. xxii.
714 (1897) but then antedated by the homonym of Sch.-Bip. in Klatt,
Leopoldina, xx. 75 (1884). S. micropappa Sch.-Bip. Bull. Soc. Bot.
Fr. xii. 81 (1865), micropappea Sch.-Bip. Linnaea, xxxiv. 535 (1865-6),
and stenocephala Sch.-Bip. ll. cc—all being nomina nuda uselessly
put on printed record merely to become burdens on synonymy.
S. glandulifera as used by Klatt, |. c., but not of Schlecht.

Three varieties may be distinguished as follows.

Var. x. typica. Phyllaries densely beset with spreading gland-
tipped hairs; pappus awnless.

: ' ja: laces in copses, Quincocirea, vicinity of
hai is sciiieia Ga alt. 2650 m., ronsas no. 245 in part (Gr,

48 ROBINSON

N. Y., Gen.); also at Sorata, Giinther, no. 8, being no. 12,015 of Buchtien’s
distrib. (U. S., fragm. Gr.).
{PERv.]

Var. 8. subeglandulosa Hieron. Phyllaries dorsally covered
with incurved or crisped white glandless hairs, a few scattered sub-
sessile glands being sometimes interspersed; pappus awnless.—Hieron.
in Engl. Bot. Jahrb. xl. 359 (1908); Robinson, Contrib. Gray Herb.
xevi. 15 (1931). S. grandidentata as used by Rusby, Mem. Torr.
Bot. Club, vi. 55 (1896), but of Sch.-Bip. only in small part. S.
grandidentata var. subeglandulosa Hieron. in Engl. Bot. Jahrb. xxii.
798 (1897).

La Prov. Larecaja: dry places in copses, Quincocirea, vicinity of
Sorata, in the temperate region, alt. 2650 m., Mandon, no. 245 in part (N. Y.).
Prov. Murillo: Rio Palca Valley, base of Mt. Illimani, alt. 2300 m., Bro. Julio,
no. 70 in part (U. S., fragm. Gr.).
no Vicinity of Cochabamba, Bang, no. 1149 (Gr., N. Y., U.S.,

Var. y. mecoyensis Robinson. Phyllaries as in the preceding
variety, but achenes each provided with 2 well developed awns as
well as the usual scale-pappus.—Contrib. Gray Herb. ec. 11 (1932).

Tara: Red Hills, Mecoya, alt. 2745-3050 m., April 1864, Pearce (K., phot.
and small fragm. Gr.).

11. S. cardiatica Perkins. Slender, suffruticose, erect, 3 dm. or
more in height, branched from near the base, densely glandular-
puberulent; leaves opposite below, scattered above, subsessile,
lanceolate, attenuate, entire, membranaceous, glandular-puberulent;
heads (about 13 mm. long) shortly pedicelled or subsessile in dense
terminal solitary or corymbously disposed glomerules; phyllaries
dorsally beset with sessile glands and spreading glandless hairs;
corollas with tube and throat deep purple, the lobes pale (probably
white or nearly so in nature); achenes unlike but all of them awnless,
the 4 adelphocarps crowned by a toothed border of more or less
connate scales (about 0.4 mm. high), the idiocarp destitute of pappus-
—Perkins in Engl. Bot. Jahrb. xlix. 222 (1913).

Porost: Quechisla . 20° 30’ ° 90’ ound, alt.
3450- _ joes aged = Bu. toy A Bgl laeel ray Fs no.
89 in part, that is to say as to element with narrow entire leaves (Gr.).

Though this species was originally described as having 3-flowered
heads, they are, at least the heads (even of the type-material) so
far as examined by the writer, 5-flowered, as is normal throughout
most of the genus. The plant is reported by Bender as being valued
locally as a remedy for heart disease.

THE STEVIAS OF BOLIVIA 49

12. §. Bridgesii Rusby. Erect or curved-ascending, copiously
branched, probably herbaceous and perennial (the base unknown);
stem vather weak, pithy, green to somewhat purplish above, densely
glandular-puberulent, leafy; leaves opposite or (many of the upper)
alternate, oblong-lanceolate, acute, serrate except at the cuneate
base, sessile or shortly petioled, hispid-puberulent on both surfaces
(the hairs short, white, slightly thickened at base, at first gland-
tipped, soon glandless), mostly 2-3 cm. long and 6-10 (-12) mm.
wide, the rameal smaller, often entire; heads (12-12.5 mm. long)
sessile or shortly pedicelled in subfastigiate corymbously or panic-
ulately disposed glomerules; phyllaries lanceolate-linear, attenuate,
about 8 mm. long, spreading-hispidulous (the hairs white, stout-
based, fugaciously if at all gland-tipped); corollas with deep purple
tube and throat, the lobes white or roseate; achenes unlike, the 4
adelphocarps crowned by scales and 2-3 awns (mostly about equalling
the corollas), the idiocarp bearing a scale-pappus only.—Bull. N. Y.
Bot. Gard. iv. 377 (1907).

CocHaBaMBA: Vicinity of Cochabamba, ee no. 2047 - i ? Se Gr.).

WITHOUT INDICATION OF DEPARTMENT: Bri acc. to Rus

In habit this suspiciously approaches some forms of de Argentine
S. vaga Griseb., particularly the one described by Hieronymus in
Engl. Bot. Jahrb. xxii. 709 (1897) as S. diversipapposa var. longiaris-
tata, but that has somewhat smaller heads, a less copious inflores-
cence, and a pubescence inclined to be incurved or crisped, and
therefore appears to be a different plant.

13. §. santacruzensis Hieron. Much branched suberect under-
shrub; stem and curved-ascending leafy branches brown, finely
puberulent, the hairs being very short, incurved, rarely and fuga-
ciously if at all gland-tipped; leaves opposite, sessile, ovate, acute to
obtusish, serrate except at the entire cuneate base, subcoriaceous,
the cauline 1.5-2.5 cm. long, 1-1.5°em. wide, the rameal smaller,
mostly 1-1.5 em. long; nerves and chief veins prominent on the lower
surface; heads (about 13 mm. long) sessile or shortly pedicelled in
rather small glomerules; these borne at the tips of the branches and
together forming an ovoid or subpyramidal inflorescence; phyllaries
oblong, acute, mucronulate, very finely puberulent, the hairs mostly
incurved and glandless but interspersed with sessile or scarcely
stiped glands; corollas deep reddish purple, the lobes of the limb
rather narrow, linear-lanceolate; achenes unlike, the 4 adelphocarps
each 5-awned, with some scales or slight scarious expansions at base,
the idiocarp awnless, bearing a toothed scarious cup (about 1 mm.
high).—Hieron. in Engl. Bot. Jahrb. xxii. 731 (1897).

50 ROBINSON
Santa Cruz: Sierra de Santa Cruz, alt. 2600 m., Kuntze (N. Y., phot. Gr.).

14. §. fruticosa Griseb. Shrub about 2 m. high, much branched
and leafy above; leaves (except some of the uppermost) opposite,
closely sessile by a rounded or subcordate entire base, acute, sharply
serrate, mostly 2-3 em. long and 1-2 cm. wide, of firm texture, the
nerves and veins prominent beneath; heads (10-11 mm. long) sub-
sessile in bracteate corymbously disposed glomerules; bracts ovate-
lanceolate, acute, subentire, equalling or often exceeding the glom-
erules; phyllaries lanceolate-oblong, acute, dorsally scabrid-puberu-
lent; achenes unlike, the 4 adelphocarps each 1-3-awned and pro-
vided also with a short crown of toothed scales, the idiocarp bearing
a similar crown (about 0.6 mm. high) but no awns.—Goett. Abh.
xxiv. 167 (1879); Hieron. in Engl. Bot. Jahrb. xxii. 731 (1897).

Tarisa: on the Cuesta del Tambo, between El Tambo and Narvaez, Hiero-
nymus & Lorentz, no. 876 (Brl., phot. Gr.).

15. 8. chacoensis R. E. Fries. Perennial herb about 1 m. high,
pilose, leafy to the rather loose inflorescence; leaves subopposite
or (especially the rameal) alternate, softly membranaceous, ovate-
lanceolate, consisting of an expanded serrate attenuate portion and
an abruptly contracted then cuneately narrowed entire petioliform
basal part, conspicuously pilose on both surfaces, yellowish green,
the largest about 1 dm. long and 3.8 cm. wide; heads sessile or more
often short-pedicelled and somewhat scattered on the branches of a
subfastigiate inflorescence; phyllaries narrowly oblong, acute and
cuspidate, dorsally covered by jointed white glandless hairs inter-
spersed with some sessile glands; corollas white; achenes unlike, the
3 adelphocarps crowned by a scale-pappus and 2-3 well developed
stramineous awns (about equalling the corollas), the 2 idiocarps
bearing a similar scale-pappus with or without a rudimentary awn
(much shorter than the corollas).—Ark. Bot. v. no. 13, p. 7, t. 3,
ff. 1-6 (1906).

OR eed teed Lost in dense forest at Tatarenda, Fries, no. 1388

16. 8. totorensis Robinson. Erect virgate perennial herb, 1.5-4
dm. high, slightly scabrid; leaves opposite, sessile, ovate, acute,
serrate except at the entire cuneately narrowed base, subglabrous,
2.5-3.2 em. long, 1.5-2 cm. wide, of rigid texture, the nerves and veins
prominent on both surfaces; glomerules few, rather dense, borne in
a flattish corymb; phyllaries (about 6.5 mm. long) lanceolate-oblong,
acute, firmish in texture, dorsally beset with subsessile glands; corollas
violet; achenes unlike, the 4 adelphocarps each crowned by scales

THE STEVIAS OF BOLIVIA apt

and 3-5 awns (as long as the corollas), the idiocarp bearing similar
scales and 2-3 awnlets much shorter than the corolla—Contrib.
Herb. xevi. 16 (1931).
CHABAMBA: Prov. Totora: alt. 3000 m., at Bucona, Steinbach, no. 3953
(Bu. .» phot. and small fragm. Gr.).
Habitally similar to but clearly not identical with the Brazilian
S. myriadenia Sch.-Bip.

17. 8. reclinata Rusby. Probably suffruticose and about 1 m.
high (the base unknown), spreading-pubescent; stem apparently
curved-ascending, leafy to the inflorescence; leaves opposite, petiolate,
ovate, acute, sharply and rather coarsely serrate except for the
abruptly contracted then gradually cuneate entire base, chartaceo-
coriaceous, slightly hirtellous, 6-7.5 cm. long, 3-4 em. wide; veins
rather closely reticulated and exserted on both surfaces; heads (about
11 mm. long) subsessile in somewhat loose often nodding glomerules;
phyllaries green, linear-lanceolate, acute, dorsally beset with delicate
white incurved or crisped glandless hairs interspersed with sessile or
subsessile glands; corollas white; achenes unlike, the 4 adelphocarps
surmounted by a very narrow scarious border together with 1-3
fully developed awns (about equalling the corollas), the idiocarp
bearing a similar scarious border but no awns.—Bull. N. Y. Bot.
Gard. viii. 127 (1912

La Paz: Prov. Caupolican: near Apolo, about 1770 m. alt., R. S. Williams,
no. 1468 (N. ¥., phot. Gr., U. 8.).

S. glanduloso-pubescens Hieron. Suffruticose, erect, up-
wardly branched above; leaves opposite, subsessile (on a cuneately
narrowed entire base), rhombic-ovate, acuminate, rather deeply
crenate-serrate, glandular-puberulent above, glandular-pubescent
beneath, the largest about 5.5 em. long, 2.5 em. wide; heads sessile
or nearly so in clustered glomerules surpassed by the foliaceous
bracts; phyllaries lanceolate-linear, obtuse, erose-denticulate at the
summit, the outer sparingly glandular-pubescent dorsally; achenes
unlike though all bearing some muticous denticulate more or less
connate scales; 3 achenes also provided with awnlet-tipped scales as
much as 1.5 mm. long; the 2 remaining achenes bearing usually at
least 1 fully developed awn (4.5 mm. long) and a shorter awnlet as
well as the muticous scales—Hieron. in Engl. Bot. Jahrb. xl. 360
(1908).

UTHERN Borivia [probably Dep. Tarija]: without indication of locality,
Pichon no. 3486 (Brl., phot. Gr.).

oe ROBINSON

19. §. cochabambensis Hieron. Erect, herbaceous perennial,
4-5 dm. high, densely spreading-pubescent, the hairs white, stoutish-
based and in the inflorescence gland-tipped; leaves ovate-lanceolate,
sessile by a gradually narrowed entire base, finely and rather sharply
serrate, acute to acuminate, pale green (sometimes purple-tinged),
the largest cauline 4-4.5 cm. long, 1.2-1.7 cm. wi e; the rameal
smaller, subentire; heads (10-14 mm. long) sessile or nearly so in
dense glomerules terminal upon the filiform ascending branches of a
loose flattish-topped corymb; phyllaries deep purple, shortly acumi-
nate, dorsally stipitate-glandular; florets much exserted; corollas as
to tube and throat deep purple, the lobes paler, probably pink or
white; achenes usually unlike, the 4 adelphocarps bearing a short
scale-pappus (about 0.5 mm. high) and 3-5 slender purple awns
(about equalling the corollas), the idiocarp crowned merely with the
scale-pappus or rarely awn-bearing like the others.—Hieron. in
Engl. Bot. Jahrb. xxii. 726 (1897).

CocuaBampa: near the city of Cochabamba, alt. 3000 m., Kunize (N. Y.,
phot. Gr.),

20. 8. kuhnioides Rusby. In most respects exceedingly similar
to the preceding, but much less glandular, the branches of the in-

La Paz: Prov. Larecaja: Cordillera Real on the road from Okara to Ancoma,
alt. about 3350 m., Tate, no. 870 (N. Y., phot. Gr.).

yellowish-green, paler beneath, the largest about 5 cm. long and 5 mm.
wide; heads loosely paniculate, borne singly on filiform pedicels
(5-10 mm. long); corollas as to tube and throat purple, limb paler,
probably pink; achenes crowned with more or less eneiate aid nothin
scales (about 0.6 mm. high), with or without 1-2 awnlets fabout ite
mm. long) or more rarely 2 fully developed awns.—Hieron. in Engl.
— xl. 365 (1908); Robinson, Contrib. Gray Herb. xe. 61

THE STEVIAS OF BOLIVIA 53

Tarisa: Red Hills, Mecoya, Pearce (K., phot. and small fragm. Gr.), a
form with part of the achenes in some of the heads 2-awned.

The type-locality for this species, namely on steep grassy slopes,
alt. 1800 m., near Toldos not far from Bermejo, where the plant was
collected by Fiebrig, no. 2330a (Brl., phot. Gr.), though cited by
Hieronymus as in Bolivia, appears from the best available maps to
be in the Argentine province of Salta slightly to the south of the
Bolivian boundary. From the earlier collection of Pearce showing
conspecific material from Mecoya, the species may, however, be
definitely included in the Bolivian flora.

22. S. yaconensis Hieron. Erect or ascending herbaceous or
suffruticose perennial, 3-9 dm. high; stem leafy, simple below, much
branched above; leaves (except some of the upper ones) opposite,
lanceolate-ovate, acute, rather coarsely serrate except at the entire
cuneately narrowed petioliform base, the largest cauline 6-8 cm.
long, nearly half as wide; inflorescence a diffusely branched panicle;
heads borne singly on filiform pedicels (1-1.5 cm. long); corollas as
to tube and throat deep purple, the limb pink; achenes unlike, the
4 adelphocarps crowned by a narrow scarious somewhat toothed
border (0.5 mm. high) and 2-3 awnlets (1-1.5 mm. long) or 2-3 fully
developed awns, the idiocarp destitute of pappus.—Hieron. in Engl.
Bot. Jahrb. xxii. 724 (1897).

An Argentine species, running into several varieties (regarding
which see Robinson, Contrib. Gray Herb. xc. 61) of which the fol-
lowing approaches so close to the frontier of Bolivia that its presence
in the flora of the country seems exceedingly probable.

Var. subeglandulosa Hieron. Pedicels and outer phyllaries cov-
ered with short incurved glandless puberulence, with or without some
sessile glands interspersed.—Hieron. in Engl. Bot. Jahrb. xl. 366
(1908); Robinson, |. c. 62 (1930).

_ Tara ? alt. 1800 m., Toldos gy ee southwest of Bermejo and probably
in the Argentine province of Salta, Fiebrig, no. 2330 (Gr.).

23. 8. clivicola Robinson. Suberect herbaceous perennial, 3-4
dm. high, finely incurved-puberulent; stem terete, brown, at anthesis
denuded below by the fall of the leaves, branched from near the
middle; rameal leaves opposite, sessile or nearly so, narrowly lanceo-
late, acute or attenuate, entire or subentire, firmly membranaceous,
3-nerved, often proliferous in the axils; heads sessile or subsessile in
dense glomerules; these terminal. on the ascending branches and
subeorymbously disposed; phyllaries deep purple, finely incurved-
puberulent on the back; corollas deep purple with roseate limb;

54 ROBINSON

achenes nearly alike, the pappus composed of an erose or denticulate
scarious border (about 0.5 mm. high) and 2-4 mostly unequal awnlets
(1-4 mm. long).—Contrib. Gray Herb. c. 4 (1932).

La Paz: Prov. Larecaja: on the Loma de Canalloquenchana, alt. 3000 m.
in the vicinity of Sorata, Mandon, no. 243 in part (Gr).

24. §. obovata Rusby. Low herbaceous perennial usually branched
from near the base, the branches curved-ascending, mostly 1-2 dm.
high; leaves sessile, chiefly opposite and crowded on the lower part of
the branches, obovate or elliptical, mostly obtuse or rounded at tip,
crenate-serrate from about the middle, chartaceo-subcoriaceous, 1.5-3
cm. long, about half as wide, the nerves and veins exserted on both
surfaces; heads (about 1 cm. long) sessile or very shortly pedicelled in
dense short-stalked glomerules; these mostly clustered by 3’s in a
small terminal corymb; phyllaries (chiefly dark purple) lanceolate,
acute to acuminate, loosely pilose dorsally, the hairs white, tapering,
glandless but interspersed with some sessile glands; corollas dark
purple, with much paler lilac or roseate lobes.—Mem. Torr. Bot.
Club, vi. 55 (1896).

A species of characteristic habit, but in pappus varying as follows:

Var. typica. All the achenes crowned by a short scale-pappus
(mostly about 0.4 mm. long) with 1-3 awnlets (about 1-1.5 (-3) mm.
long).—Robinson, Contrib. Gray Herb. xc. 17 (1930). S. obovata
Rusby, l.¢. 8. humilis Hieron. in Engl. Bot. Jahrb. xxii. 730 (1897);
Kuntze, Rev. Gen. iii. 180 (1898), merely a dwarf state.

CocHABaMBa: near the city of Cochabamba, Bang, no. 1027 (N. Y., Gr.,

: (
U.S., Field Mus., Mo.); same locality, alt. 3660 m., Cardenas, no. 59 (N. Y.).
te NOT sTaTED: Cuchicanchi Pass, Kuntze (N. Y., phot. Gr.,

Var. aristifera Robinson. Adelphocarps each crowned by a short
scale-pappus together with 3-5 well developed purple awns about
equalling the corollas.—Contrib. Gray Herb. xc. 17 (1930).

CocHaBamBa: Proy. Chapare: grassy places near Aduana, Colomi, alt.
2800 m., Steinbach, no. 8831 (Gr., Mo.). .

RR ENE RTO eT MOREE Ds APN ES TNC ao DTN SN ey ea ee ee

THE STEVIAS OF BOLIVIA 55

sally glandular-puberulent; corollas deep purple with roseate or
white limb; achenes nearly uniform, each bearing a shallow border
of connate scales (0.3 mm. long) and mostly 3 usually unequal teeth
or short awns (0.6-1.4 mm. long).—Perkins in Engl. Bot. Jahrb.
xlix. 221 (1913).

Potosi: Quechisla, alt. 3450-3500 m., Bender, no. 24 (Brl., phot. and sketches
of details, Gr.); at same locality, Cardenas, no. 89 in part (Gr.).

26. §. bermejensis Hieron. Erect or ascending perennial herb,
slightly lignescent toward the base, about 4 dm. high, nearly glabrous;
leaves opposite, rhombic-lanceolate, acute or narrowed to an obtuse
tip, crenate-serrate except for the entire at length cuneate-attenuate
petioliform base, subchartaceous, stated to be quite glabrous on both
surfaces, the largest cauline nearly 1 dm. long and 2.5 cm. wide;
heads subsessile or on short pedicels (1-2 mm. long) in dense bracteate
glomerules; these disposed in a flattish 3-branched compound corymb;
phyllaries linear-lanceolate, acute, about 5.5 mm. long, substramine-
ous, the outer gland-sprinkled, the inner glabrous; corollas pale,
probably yellowish or greenish white; achenes unlike, all bearing
at the summit a low continuous border (scarcely 0.2 mm. high),
the 4 adelphocarps provided also with 1-2 teeth or awnlets (0.5-1
mm. long), the idiocarp without these rudimentary awns.—Hieron.
in Engl. Bot. Jahrb. xl. 361 (1908).

TariJA ?

Though stated by Hieronymus to have come from Bolivia and by
him named for the Bolivian village of Bermeja, this plant was col-
lected (Fiebrig, no. 2326, Brl., phot. Gr.) at Toldos (alt. 1800 m.),
a locality which appears to be across the international boundary
from Bermeja and in the Argentine province of Salta. However, as
the station is so slightly extra-limital and the presence of the plant
within the boundaries of Bolivia so probable, the species is included
in the present treatment.

27. S. camachensis Hieron. Perennial herb, erect or ascending,
5-6 dm. high, loosely and rather shortly spreading-pubescent, leafy
to the summit; leaves (except some of the uppermost) opposite,
suborbicular-ovate, obtuse to subacuminate, coarsely crenate except
for the entire rounded base, 3-5.5 em. long, 2.4-4 cm. wide, thin,
green on both surfaces though slightly paler beneath; petiole 8-28
mm. long, scarcely winged except for a slight expansion at its attach-
ment to the blade; heads about 11 mm. long, closely sessile in dense
broadly campanulate corymbously disposed glomerules; phyllaries

56 ROBINSON

narrowly lanceolate, acute, green, sparingly incurved- or spreading-
pilose dorsally; corollas deep purple; achenes unlike, the 4 adelpho-
carps bearing a narrow scarious erose border of connate scales
(about 0.3 mm. high) and 1-3 usually unequal awnlets (0.5-3.7 mm.
long), the idiocarp destitute of pappus.—Hieron. in Engl. Bot.
Jahrb. xl. 359 (1908).

Tarisa: Camacho, Fiebrig, no. 2868a in part (Gr., U. Bi)?

28. 8. eclipes Robinson. Probably an herbaceous perennial;
stem thickish but weak, pithy and fistulose; leaves opposite, the main
cauline clearly petioled, broadly ovate, obtuse to acuminate, serrate
to coarsely crenate, 5-7 cm. long, 4-5 cm. wide, subtruncate at base,
membranaceous, green and appressed-hirtellous above, scarcely paler
and setulose-pilose on the nerves and chief veins beneath; petiole
1.5-2.5 em. long, slender below, cuneately winged toward the summit;
heads (about 9 mm. long) closely sessile in dense glomerules; these
crowded at the ends of curved-ascending branches of the compound
corymb; phyllaries greenish-substramineous, shortly pointed; corollas
white; achenes rather similar, the 4 adelphocarps bearing at the sum-
mit a low erose border and usually 2 awnlets, the idiocarp being
destitute of the scarious border but bearing similar awnlets (only
about 0.1-0.2 mm. long).—Contrib. Gray Herb. xevi. 5 (1931).

ba Face near Tarija, alt. 2300 m., Fiebrig, no. 3126 (Gr., Field Mus.,
+» Mo.).

29. 8. calderillensis Hieron. Erect upwardly branched perennial
herb, becoming 5 dm. high; leaves opposite, the main cauline deltoid-
ovate, acuminate, crenate-serrate except for the subtruncate base,
glaucescent green and sparingly puberulent above, paler and pubes-
cent on the nerves and veins beneath; petioles 1-2 cm. long, slender;
rameal leaves ovate-lanceolate to elliptic, much smaller, sessile or
nearly so, often obtuse to rounded at the tip; heads sessile or subsessile
in dense corymbously disposed glomerules; phyllaries subacute,
6.5 mm. long, densely glandular-puberulent on the back; corollas
violet-purple; achenes similar, surmounted by a low erose scarious
border and 1-3 usually unequal awnlets (1.5-2 mm. long).—Hieron.
in Engl. Bot. Jahrb. xl. 356 (1908). :

SouTHERN Bo.rvia [presumably Tarra]: Calderillo, alt. 3400 m.,
Fiebrig, no. 2959 (Bri., phot. Gr.).. is ise cesenactns

Two fragmentary bits, one of Fiebrig, no. 3425 (Gen.) and the other

separated from Fiebrig, no. 2214 (K., phot. Gr.), both lacking precise
data of collection, have been doubtfully referred to this species.

THE STEVIAS OF BOLIVIA 57

30. 8. galeopsidifolia Hieron. Erect perennial herb; stem
densely glandular-puberulent and also beset with longer spreading
attenuate jointed glandless hairs; leaves opposite, deltoid-ovate,
acuminate, coarsely crenate except for the subtruncate base, thin,
3-nerved from the very base, finely pubescent and dark green above,
somewhat paler and pilose especially along the nerves and chief
veins beneath, the largest about 8 cm. long and 5 cm. wide; petioles
1-2.5 cm. long, slender, almost wingless; heads sessile or nearly so
in hemispherical to subglobose glomerules; these aggregated in a
rather dense terminal corymb; phyllaries pale green, substramineous,
linear-oblong, tending at maturity to be blunt and erose at the tip,
dorsally almost smooth; corollas white or nearly so; achenes alike
or similar, each bearing at the summit a low denticulate scarious
border (0.4 mm. high) and 1-3 longer toothlike processes or unde-
veloped awns (0.7—1.5 mm. long).—Hieron. in Engl. Bot. Jahrb. xxii.
719 (1897).

CocHABAMBA: Prov. Ayopaya: Tunari, Kuntze (N. Y., phot. Gr., fragm.
Brl., tracing Gr.).

[Prrv.]

31. S. boliviensis Sch.-Bip. Erect perennial herb, finely and often
inconspicuously pubescent, mostly simple below and branched above;
stem usually reddish brown or purple, pithy and fistulose, leafy;
middle internodes 6-10 cm. long; leaves opposite, rhombic-ovate,
acute or acuminate, crenate-serrate except for the entire basal portion,
membranaceous, green on both surfaces, somewhat paler beneath,
the main cauline 6-9 cm. long, 2-4.5 cm. wide, the base at first grad-
ually or somewhat rapidly contracted then merging into a long,
rather broadly and cuneately winged petioliform portion; heads
(about 1 em. long) sessile or nearly so in small clusters; these some-
what fastigiately grouped at the ends of the branches in a compound
corymb; phyllaries narrowly oblong-lanceolate, acute to acuminate,
usually pale green and almost glabrous; corollas chiefly white; achenes
all surmounted by a narrow erose scarious border and bearing 1-2
awnlets (0.2-1.2 mm. long) or 1-3 of the achenes provided each
with 1-3 fully developed awns (3-5 mm. in length).—Bull. Soc.
Bot. Fr. xii. 81 (1865), and Linnaea, xxxiv. 535 (1865-66), both with-
out char.; Rusby, Mem. Torr. Bot. Club, iii. no. 3, p. 61 (1893),
where described from material collected by Bang. S. Schultzit Hieron.
in Engl. Bot. Jahrb. xxii. 721 (1897), at least as to Bolivian element.
S. brevipapposa Hieron. 1. c. 718 (1897). 8S. breviaristata as used by
Hieron. in Ktze. Rev. Gen. iii. 180 (1898), not Hook. & Arn.

58 ROBINSON

La Paz: Prov. Larecaja: in clayey soil of the temperate region in the vicinity
of Sorata, Mandon, nos. 242 in part (Gr., N. Y., Gen., Bri.) an in part
(Gr., N. Y.); Sorata, Holway, no. 542 (Gr.), Giinther, nos. 7 (U. 8.) and 25
(U. 8.); on road from Ocara to Ancoma, alt. about 3800 m., Tate, no. 874
(N. Y.); La Joya, alt. about 1800 m., Tate, no. 1049 (N. Y.); hacienda above
ro 1400 m., in subtropical region, Buchtien, nos. 561 (U.
S., Gr.), 7542 (Gr.), 7543 (U. 8., Gr.) and 7545 (U. S., Gr.). Prov. Yungas:
os. 1614 (Gr., N. Y., U. S., Field Mus., Mo.) and 1614B (N. Y.),

Bang, no. 260 (Gr., N. er Mo.). Prov. North Yungas: Unduavi, alt. about
m., ; Ny i ; : } :

693 (U.8.); Pitiguaya, valley of the Unduavi River, alt. about 1775 m., Tate,
no. 749 (N. Y.); Polo-Polo near Coroico, alt. 1100 m., Buchtien, no. 5728 (Gr.,
op Ne Leds Bee ugnaya, alt. 1200-1300 m., Buchtien, nos. 387 (Gr., Mo.),
746 (Gr.), 4414 (U.S., Gr.), 4748 (U. S., Gr.) and 4749 (U. S., fragm. Gr.).
Prov. South Yungas: Hacienda “La Flori a,” Holway, no. 65 (Gr.); Sirupaya
; alu-

ni, alt ‘
<e oinhig deep shady ravine, Cafiamina, White, no. 267 (Gr., N. Y.)
OCH

ABAMBA: near the city of Cochabamba, Kuntze (N. ¥. phot. Gr.,

‘SANTA Cruz: alt. 1600 m., Kuntze (N. Y., U. 8.). Prov. Vallegrande:
Samaipata, alt. 1200 m., Steinbach, no. 3715 (Brl., phot. Gr.).

takes place so easily that heads of both kinds sometimes occur on the
same individual. Consequently these pappus-differences, although
often striking, fail to yield any secure basis even for varietal classi-
fication.

Similarly the inflorescence, which usually consists of small but
rather dense and closely aggregated glomerules, not rarely develops a
looser condition in which the heads are borne in more open and
fastigiately branched partial corymbs, as for instance in Buchtien,
nos. 387 (Gr.) and 5611 in part (Gr.), and Tate, no. 749 (N. Y.). One
of these plants with lax inflorescence if placed by the side of others

THE STEVIAS OF BOLIVIA 59

specimens of this Bolivian plant in the Argentine S. breviaristata
Hook. & Arn., but that species has several differences which thus
far permit its pretty ready separation. Its heads are a trifle larger,
its phyllaries are much more closely incurved-puberulent, and its
corollas are normally purple at least as to tube and throat.

Of S. boliviensis, Buchtien, no. 4413 (U. S., Gr.) from Millugnaya
in North Yungas, presents a probable variety with thicker nearly
entire leaves more than usually pale beneath, but the material at
hand for study is inadequate, showing only the upper part of the plant
and not even well developed cauline leaves below the inflorescence.
As technical differences in the heads seem to be lacking, this plant
may well await further examination when more complete material is
available.

Similarly, a marked variety or possibly distinct species is represented
by specimens collected by Buchtien at Unduavi, North Yungas, alt.
3200 m., Feb. 1914 (Gr., N. Y.). These are unnumbered and were
distributed as S. menthaefolia Sch.-Bip., a wholly different Brazilian
species. They show a plant manifestly close to S. boliviensis but
differing from its typical form in having slightly thicker and smaller
leaves, at maturity practically glabrous beneath and subsessile by a
gradually narrowed blade rather than a petioliform base, purple-
tinged phyllaries, deep purple corolla-throat, and somewhat more
developed scarious portion of the pappus. No one of these distinctions
seems absolute and as the specimens at hand have only rameal and
no cauline leaves, it is felt desirable not to name or give formal descrip-
tion to the plant until it can be studied from further material showing
at least the nature of the stem foliage. The group to which S. boli-
viensis belongs is even now too difficult in the matter of specific
limits to encourage the publication of further supposed novelties
except on very complete and convincing material.

32. S. mercedensis Hieron. Slender erect or curved-ascending
perennial herb, stems mostly single and simple to the ascendingly
branched inflorescence; leaves (often proliferous in the axils) opposite
or in part alternate, sessile by a gradually narrowed base or the
larger shortly petioled, narrowly lanceolate to linear, rather obscurely
few-toothed or entire; heads sessile in turbinate or campanulate
corymbously disposed glomerules; phyllaries oblong-lanceolate, acute;
corollas as to throat usually purple, the limb pink to white; achenes
mostly unlike, the 4 adelphocarps 6-13-awned, the idiocarp with
short scale-pappus and 1-3 (-6) awns.—Hieron. in Engl. Bot. Jahrb.
xxii. 735 (1897); Robinson, Contrib. Gray Herb. xc. 73 (1930).

Of the three recorded varieties of this species two are now known
from Bolivia, namely:

60 ROBINSON

typica. Phyllaries dorsally incurved-puberulent; adel-
phocarps each 7-13-awned.—Robinson, I. ec.

Tara: hills, Mecoya, alt. 2440-2745 m., Pearce (BM., phot. and small

fragm. Gr.); slightly extralimital and beyond the southern Pee at Toldos

near Bermejo, Fiebrig, no. 3418 (Brl., phot. and small fragm. Gr
[ARGENTINA.]

Var. glanditecta Robinson. Phyllaries dorsally beset with short
preading stipitate glands; adelpl ps each 6—13-awned.—Robinson,
l. ec. 74.

Bottvia: without exact locality, Bridges (K., phot. and small fragm. Gr.).

33. 8. samaipatensis Robinson. Erect, virgate, leafy-stemmed
perennial, incurved-puberulent, 2-3 dm. high; leaves both opposite
and alternate, sessile, rather crowded and often proliferous in the
axils, narrowly oblanceolate, crenate-serrate toward the obtuse tip,
attenuate and entire toward the base, subcoriaceous; heads 8-9 mm.
long, sessile or nearly so in small glomerules; these disposed in a
dense somewhat fastigiately branched terminal corymb; phyllaries
oblong, short-pointed, incurved-puberulent on the back, often purple-
tinged; corollas as to tube and throat dark purple, the limb white or
nearly so; achenes unlike, the 4 adelphocarps each 9-11-awned, the
idiocarp awnless but surmounted by a short denticulate border of
somewhat connate scales (about 0.7 mm. high).—Contrib. Gray
Herb. xevi. 15 (1931).

s761 il ek a by tna ae aa alt. 2000 m., Steinbach, no.

34. §. Kuntzei Hieron. Frutescent, several-stemmed or branched
from near the base, about 3 dm. high; stems and branches purplish-
brown, flexuous, leafy, shortly incurved-puberulent and toward the
summit obscurely stipitate-glandular; leaves opposite and alternate,
sessile or nearly so, oblanceolate or oblong, obtusish, crenate-serrate
toward the tip, gradually narrowed and entire toward the base,
firmly chartaceo-subcoriaceous, mostly 2-3 cm. long, 4-8 mm. wide,
nearly glabrous, the nerves and veins prominent on both surfaces;
heads (about 8.5 mm. long) sessile in subhemispherical glomerules,
these clustered at the tips of the slender erect branches; corollas as
to the subcylindrical throat purple, the limb probably pale pink;
achenes unlike, the 4 adelphocarps each 7-11-awned, the idiocarp
awnless or l-awned but surmounted by a scale-pappus (0.4 mm.
high).—Hieron. in Engl. Bot. Jahrb. xxii. 733 (1897); Kuntze, Rev.
Gen. iii. pt. 2, p. 180 (1898).

THE STEVIAS OF BOLIVIA 61

Nm. Prov. Ayopaya: Tunari, alt. 3000 m., Kuntze (N. Y., phot.
Fis

35. §. setifera Rusby. Perennial, herbaceous or perhaps fru-
tescent below (the base unknown); stem terete, pithy, purple, finely
incurved-puberulent; inflorescence closely beset with stipitate glands;
leaves opposite, the cauline distinctly petioled, ovate, acutish to
attenuate, serrate except at the rounded or somewhat gradually
contracted base, membranaceous, finely incurved-puberulent above,
paler and loosely pilose beneath, about 4 cm. long and half as wide;
the rameal leaves narrower and lanceolate; heads (12-15 mm. long)
sessile or shortly pedicelled in dense glomerules; these either closely
aggregated in hemispherical clusters or scattered toward the ends of
the branches of a compound corymb; phyllaries narrowly oblong,
acute to obtusish, usually purple-tinged and beset with short spreading
stiped glands with or without some longer spreading glandless hairs;
corollas gland-sprinkled on the proper tube and purple subeylindrical
throat, the lobes white or nearly so; achenes unlike, the 3 adelpho-
carps each 7—-11-awned; 2 idiocarps surmounted by a toothed crown
of connate scales (0.4 mm. high) and awnless or not more than 3-
awned.—Contrib. Gray Herb. c. 10 (1932).

La P , Rusb
1615 in part (Field Mus.); same locality, alt. about 3300 m., Buchtien, nos.
7 3020 (N. Y., U.S.) and 3021 (U. 8.); Unduavi Valley, Bro. Julio,

U.S, 2100 ir

. Julio, no. 70 in part (U. 8., Gr.); Irupana, alt.
about 1680 m., McCarty, no. 120 (Gr., N. Y.). Pro illsi
Pongo de Quime, alt. about 3500 m., Dr. O. E. White, no. 168 (Type in herb.
N. Y. Bot. Gard., isotypes Gr., Field Mus.).

Orvro: Oruro, alt. 3900 m., Asplund, no. 5723 (no. 12,020 of Buchtien’s
distrib.), U. 8.

Dr. Rusby has kindly permitted the publication of this novelty
in the present paper. It is rather closely related to S. Kuntzei Hieron.
but has larger, thinner, more definitely petiolate and much less prom-
inently reticulate-veiny leaves.

36. 8. sarensis Robinson. Erect perennial herb, 7-10 m. high,
spreadingly branched from near the middle to the summit; stem
terete, brown, spreadingly crisped-pilose with white jointed glandless

airs; leaves opposite, the main cauline rhombic-ovate or narrowly
deltoid-ovate, gradually narrowed to an acute or sometimes obtuse
apex, crenate-serrate except for a more or less abruptly contracted
then cuneately narrowed petioliform base, membranaceous, green on
both surfaces, subappressed-pubescent above, rather densely spread-
ing-pilose particularly on the nerves and chief veins beneath, 6-9

62 ROBINSON

em. long (the petiole-like base included), 1.8-3.5 em. wide; heads
(1 em. long) sessile or distinctly pedicelled; phyllaries acute to acumi-
nate, shortly and rather loosely pilose on the back; corollas (yellowish-
brown in dried specimens) probably white or at least pale in nature;
- achenes unlike, the 4 adelphocarps each about 11-awned, the idiocarp
crowned with more or less connate scales (0.4 mm. long) and bearing
1-4 usually unequal awns.—Contrib. Gray Herb. c. 9 (1932).

Var. typica. Heads rather closely sessile in glomerules; these
terminal on the branches of a corymbiform panicle.—See p. 10.

Santa Cruz: Prov. Sara: in clearing, Buena Vista, Steinbach, nos. 1143
(Brl., phot. Gr.), 5169 in part (hb. Osten), 6117 (K., Gen., Brl., phot. Gr.),
and 7048 (Gr., Brl., Gen., K., Field Mus., Mo.).

Var. dissiticeps Robinson. Infloresence looser; heads more or
less evidently pedicelled and subracemosely arranged on the branches
of the panicle—Contrib. Gray Herb. c. 10 (1932). S. menthifolia
of Kuntze, Rev. Gen. iii. pt. 2, p. 180 (1898) not S. menthaefolia
Sch.-Bip.

Santa Cruz: Prov. Cerecado: in clearings, Bafiado del Piray, alt. 450 m.,
Steinbach, no. 6248 (Brl., phot. and small fragm. Gr.). Prov. Sara: moist
woods of the Pampa, Buena Vista, alt. 500 m., Steinbach, no. 5169 (Gr.); Pam-
pobosque, alt. 450 m., Steinbach, nos. 1126 (Brl., phot. and fragm. Gr.), 2790
(Bri., phot. and small fragm. Gr.) and 5169a ii.. th. Ys).
without indication of province or locality, Kuntze (Brl., N. Y., phot. Gr., U.8.).

37. §. melancholica Robinson. Probably an herbaceous or
suffrutescent perennial, 4 dm. or more in height (the base unknown),
minutely crisped-puberulent, blackening in drying; leaves opposite,
petiolate, at maturity often deflexed, rhombic-ovate, acutish, crenate-
serrate except at the entire cuneate base, above at first obscurely hirtel-
lous, below a little paler, glabrescent, puncticulate, firmly membrana-
ceous, 4-5.5 em. long, 2-3.6 em. wide; petiole 5-8 mm. long; corymb
flattish, rather dense, fastigiately branched; pedicels mostly about 3
mm. long; heads about 12 mm. long; phyllaries narrowly oblong,
shortly pointed, 5-6 mm. long, sparingly crisped-puberulent on the
back; corolla-throat deep purple, the lobes probably white or roseate;
achenes unlike, the mostly 3 adelphocarps each 6-8-awned and pro-
vided also with some short scales; idiocarps (usually 2) awnless but
surmounted by a scale-pappus (about 0.3 mm. high).—Contrib. Gray
Herb. xe. 15 (1930).

Bottvia: without exact locality, Bridges (BM., phot. Gr.).

38. 8. discolor Robinson. Probably an upright perennial herb
at least 6-8 dm. high (the base unknown); stem terete, leafy, above

SORT eA Ne ene eT nS eee

Fo eS OE

THE STEVIAS OF BOLIVIA 63

puberulent and beset with some sessile glands; corollas with purple
throat and whitish limb; achenes unlike, the 3 adelphocarps each
8-9-awned; the 2 idiocarps awnless but crowned by very short scales
and awnlets (0.2-0.3 mm. in length).—Contrib. Gray Herb. c. 5 (1932).

La Paz: Prov. Larecaja: in low woods and thickets at Iminapi, Milipaya,
&c., in the neighborhood of Sorata, alt. 2650-3300 m., M. andon, no. 242 in part
(Gr., Brl.).

39. S. Bangii Rusby. Several-stemmed or much branched herba-
ceous or suffrutescent perennial, finely incurved-puberulent but
not stellate-puberulent as originally described; leaves chiefly opposite,
lanceolate to ovate, serrate except at the entire cuneate sessile or
shortly petioled base, the main cauline 1.5-3 cm. long 5-12 (-18)
mm. wide, the rameal much smaller, subentire ; heads (about 1 em.
long) sessile in corymbously or paniculately disposed glomerules;
phyllaries narrowly oblong, obtuse or merely acutish, incurved-
puberulent dorsally; corollas as to tube and throat deep reddish
purple, the limb somewhat paler, probably roseate; achenes unlike,
the 4 adelphocarps each (6-) 12-15-awned, the idiocarp crowned by
a scarious border of toothed and somewhat connate scales with or
without 1-3 awns.—Mem. Torr. Bot. Club, iv. 209 (1895). 8S.
compacta as used in Bull. Torr. Bot. Club, xviii. 333 (1891) and as
doubtfully employed by Rusby, Mem. Torr. Bot. Club, iii. no. 3,
p. 51 (1893) but very unlike the real S. compacta Benth. S. tapacarien-
sis Hieron. in Engl. Bot. Jahrb. xxii. 734 (1897).

An exceedingly puzzling species, falling into the following very
unlike yet certainly intergradient varieties.

Var. typica. Erect or nearly so, several-stemmed or for the most
part uprightly branched from near the base, mostly 2.5-4 dm. high;
leaves lanceolate to narrowly ovate-oblong, normally more than
twice as long as wide.—Literature and synonymy as above.

La Paz: Prov. Murillo: neighborhood of the city of La Paz, alt. 3050 m.,
Bang, no. 86 (N. Y., Gr., U. S., Mo.); same locality, alt. 3750 m., Buchtien,

64. ROBINSON

nos. 186 in part (N. Y., Gr.) and 3018 in part (N. Y.); near La Paz, alt. about
3355, Rusby, no. 1613 (N. Y., Gr., U. S., Field Mus., Mo.), a collection in part
typical but in part broader-leaved, more spreadingly branched and passing
to the following variety.

OCHABAMBA: Prov. Tapacari: on the River Tapacari, alt. 3000 m., Kuntze
(N. Y., phot. Gr.); Cochabamba, Holway, no. 381 (Gr.).

EPARTMENT NOT INDICATED: Chiguana, Asplund, no. 5726 (U. 8S.) being
no. 12,023 of Buchtien distrib.

With this species and indeed with its narrow-leaved and more
typical variety there may be doubtfully associated an unnumbered
Bolivian plant collected by Bridges (K., BM., phot. Gr.). Closely
similar in habit, foliage, involucre, etc., it differs in having the adel-
phocarps only 6-7-awned.

Var. dyscrita Robinson. Stem mostly flexuous and decumbent,
much branched (often throughout its length), seldom over 2 dm.
high; the branches widely spreading; leaves ovate, less than twice
as long as wide.—Contrib. Gray Herb. ec. 3 (1932).

La Paz: Prov. Murillo: near the city of La Paz, alt. 3750 m., Buchtien, nos.
91 (U.8.), 186 in part (Gr.), 234 (U. 8.), 3017 (N. Y., U.S.), 3018 in part (U.
8S.) and 3019 (N. Y., U.8.); on stony hills above La Paz, alt. 3700-3800 m.,

r i 000 m.
, alt. 3400 m., Bro. Julio, no.

‘ov. Larecaja: Sorata, Giinther, nos. 9 (U. 8.), 26 (U.8.), 27 (U
being nos. 12,016, 12,018 and 12,013 respectively of the Buchtien distrib.
Prov. North Yungas 65 (U.S.).

URO: Pazha, alt. 3900 m., Asplund, no. 5725 (U. 8.) being no. 12,022 of
Buchtien series; Challapata, alt. 3900 m., Asplund no. 5724 (U. 8.) being no.
12,021 of Buchtien series.

Very near to S. Bangii, of which it may well prove a form or variety,
must be placed a still problematic plant, namely Mandon, no. 244
in part (N. Y., fragm. Gr., Brl.) from a point between “ Munaypata”’
and “ Luce Mapampa,” in clayey soil of the temperate region, alt.
2650 m., in the vicinity of Sorata in the Prov. Larecaja. Erect and
3-4 din. tall in the manner of S. Bangii var. typica, it has larger and
relatively broader leaves and very copious inflorescence, though no
significant differences in the heads have been detected. A further
investigation of this plant must await its rediscovery and the collection
of better material.

_ Certain forms of the aggregate here classified as S. Bangii, for
instance Mandon, no. 246 bis and Pennell, no. 14,255, show a close
approach on the one hand to S. tarijensis Hieron. of extreme southern
Bolivia and on the other hand to S. Mandonii Sch.-Bip. from Sorata
and the puno region of Peru.

0. S. tarijensis Hieron. Decumbent but assurgent perennial
herb, 1-2 dm. high, very finely incurved-puberulent; branches nu-
merous, widely spreading; leaves opposite, sessile or shortly petioled,

THE STEVIAS OF BOLIVIA 65

elliptic-oblong, crenate toward the rounded apex or subentire, slightly
fleshy, green on both surfaces, the main cauline caducous, about 2
cm. long and 8-10 mm. wide, the rameal smaller, mostly 1-1.5 cm.
long and 4-7 mm. wide, inflorescence a dense terminal corymb of
closely aggregated glomerules or these terminal on up-curved branches
and forming an ovoid panicle; heads (10-12 mm. long) sessile or very
shortly pedicelled in dense clusters; phyllaries oblong, obtuse, mostly
dark purple, dorsally incurved-puberulent, about 5.5 mm. ong;
corollas with dark purple tube and throat, the lobes roseate or nearly
white; achenes unlike, the 4 adelphocarps each about 15-awned, the
idiocarp awnless but surmounted by a scale-pappus (about 1 mm.
high).—Hieron. in Engl. Bot. Jahrb. xl. 362 (1908).
ARIJA: near the town of Tarija, alt. 3000 m., Fiebrig, no. 2647 (Gr.);
Escayache [Dept. Tarija ?], alt. 3500 m., Fiebrig, no. 3416 (Gr., Gen.)
coe NOT INDICATED: Bolivia but without further data, Bridges

41. 8. Mandonii Sch.-Bip. For description see p. 34 —Bull.
Soc. Bot. Fr. xii. 81 (1865) and Linnaea, xxxiv. 535 (1865-66), both
without char.; Robinson, Contrib. Gray Herb. Ixxvii. 6 (1926),
where described.

' La Paz: Prov. Larecaja: on hills near Ancohuma, in the alpine region, alt.
3800 m., Mandon, no. 246 (Gr., N. Y., Gen., Par., ., K.); Guaqui, Buchtien,
nos. 4827 (U. 8.) and 12,012 (U.S.). Isla del Sol, Yumaniy, alt. 3850, Asplund,
no. 5727 (U. 8.) being no. 12,024 of Buchtien distrib. Prov. Murillo: alt. 3750,
La Paz, Buchtien, no. 709 (Gr., U. 8., N. Y.).

SYNONYMIC OR DOUBTFUL NAMES APPLIED TO BOLIVIAN STEVIAS.

S. amplexicaulis as used by Herzog, Veg. d. Erde, xv. 144, 150
(1923), is merely a casual mention unaccompanied by description.
The name is antedated by a homonym of Hassler, applied to a small-
leaved species of Paraguay. The writer has seen no Bolivian material
in the remotest degree similar to this Paraguay species of Hassler.

S. breviaristata as used by Hieron. in Ktze. Rev. Gen. iii. pt. 2,
p. 180 (1898) not Hook. & Arn. = S. Boutvrensis Sch.-Bip.

S. brevipapposa Hieron. in Engl. Bot. Jahrb. xxii. 718 (1897)
= S. BOLIVIENSIS Sch.-Bip.

S. Clausseni var. boliviensis Hieron. in Engl. Bot. Jahrb. xxii. 723
(1897) = S. URTICAEFOLIA var. BOLIVIENSIS (Hieron.) Robinson.

S. compacta as used in Bull. Torr. Bot. Club, xviii. 333 (1891)
and as more doubtfully employed by Rusby, Mem. Torr. Bot. Club,
ili. no. 3, p. 51 (1893) but by no means of Benth. = S. Bancu Rusby.

S. elatior as used by Hieron. in Ktze. Rev. Gen. iii. pt. 2, p. 180
(1898) = S. ELATIOR var. AUsSTRINA Robinson.

66 ROBINSON

S. grandidentata Sch.-Bip. Bull. Soc. Bot. Fr. xii. 81 (1865) and
Linnaea, xxxiv. 535 (1865-66) without char., and Hieron. in Engl.
Bot. Jahrb. xxii. 714 (1897), where first described but already ante-
dated by homonym of Sch.-Bip. in Klatt, Leopoldina, xx. 75 (1884) =
S. SoRATENSIS Hieron.

S. grandidentata var. subeglandulosa Hieron. in Engl. Bot. Jahrb.
xxli. 798 (1897) = S. SORATENSIS var. SUBEGLANDULOSA Hieron.

S.
82 (1865) and Linnaea, xxxiv. 535 (1865-6). The writer has had no

Rev. Gen. iii. pt. 2, p. 180 (1898) appears to be merely a dwarf state
of S. opovata Rusby.

- MENTHAEFOLIA Sch.-Bip. Linnaea, xxv. 282 (1853). This name
(of a Brazilian species) has been at various times employed by persons
identifying Bolivian material of or closely related to S. boliviensis
Sch.-Bip., but the writer has thus far found no evidence that the true
S. menthaefolia of Brazil occurs in Bolivia.

S. micropappa Sch.-Bip. Bull Soc. Bot. Fr. xii. 81 (1865) and S.
peed Sch. Bip. Linnaea, xxxiv. 535 (1865-6) = S. soRATENSIS
eron

S. multiaristata as used by Herzog, Veg. d. Erde, xv. 143, 146 (1923).
The writer has had no opportunity to check Herzog’s application of
this name. No evidence has been seen to indicate that the real S.
multiaristata Spreng. occurs within the limits of Bolivia, though
this is by no means unlikely.

S. nepetaefolia as used by Herzog, Veg. d. Erde, xv. 143, 144, 146
(1923). The writer has every reason to infer that this record rests
upon a mistaken identification. The real S. nepetaefolia HBK. is
a Mexican plant.

S. Schultzii Hieron. in Engl. Bot. Jahrb. xxii. 721 (1897), at least
as to the Bolivian (typical) element = S. BOLIVIENSIS Sch.-Bip. as

THE STEVIAS OF BOLIVIA 67

Schultz did not describe his own species. It was further unusual that
the character of the species, when Rusby (who conscientiously and
with entire correctness attributed the origin of the binomial to Schultz)
finally described the species, was drawn from a later specimen and
not from the Mandon material seen by Schultz. However, as they
prove essentially identical and certainly conspecific, Rusby’s inter-
pretation of the Schultz name has every right to stand and Hierony-
mus’s renaming of the plant finds no justification,

S. stenocephala Sch.-Bip. Bull. Soc. Bot. Fr. xii. 81 (1865) and
Linnaea, xxxiv. 535 (1865-6) = S. soraTENsis Hieron.

S. suboctoaristata as used by Sch.-Bip. Bull. Soc. Bot. Fr. xii. 81
(1865) and Linnaea, xxxiv. 535 (1865-6), but probably not of Lag.
=. Bane Rusby. The original characterization of S. subocto-
aristata by Lagasca, stating merely that it resembled the Mexican
S. ovata but had 7-8-awned achenes, and had been collected in Peru,
is far too vague to be taken up on the basis of Bolivian material
which has no close similarity with the Mexican S. ovata and which for
the most part has more numerously awned achenes.

S. tapacariensis Hieron. in Engl. Bot. Jahrb. xxii. 734 (1897) =
S. Banam Rusby.

S. urticifolia as used by Rusby, Bull. N. Y. Bot. Gard. iv. 377
(1907) = S. URTICAEFOLIA var. PALLIDIFLORA Robinson.

EXSICCATAE OF BOLIVIAN STEVIAS

lund Buchtien

Asp
5723 setifera Rusby
5724 Bangii v. dyscrita Robinson
5725 Bangii v. dyscrita Robinson
5726 Bangii Rusby
5727 Mandonii Sch.-Bip.

1027 obovata Rusby
1149 soratensis v. subeglandulosa
2047 Bridgesii Rusby
2877 urticaefolia v._pallidiflora
obinson —
Bender

22 cardiatica Perkins
24 Benderi Perkins

91 Bangii v. dyscrita Robinson
186 in part Bangii Rusby

186 in part Bangii v. dyscrita
Robinson

189 setifera Rusby

234 Bangii v. dyscrita a

3018 in part Bangii v. dyscrita
obinson
3019 Bangii v. dyscrita Robinson
Rusby

441 ens. :
4748 boliviensis Sch.-Bip.
4749 boliviensis Sch.-Bip.

68 ROBINSON

ensis Sch.-Bip
7544 keg vinta! v. pallidifiors

Robin
12,012 Mandonit Sch, Bi
12) ii v. dyscrita Robinson
12,015 aornteniai Hieron
12/016 Bangii v. dracite Robinson
12,018 Bangii v. dyscrita Robinson
12) ,020 setifera Rusb
12; 021 Bangii v. dyscrita Robinson
2) 022 Ba angii v. dyscrita Robinson
12) 023 Bangii Rusby
12; 024 Rnadont Sch. -Bip.
Cardenas
59 obovata Rusby
89 in part Basie Perkins
89 in part cardiatica Perkins
Fiebrig
2214 in small part calderillensis

868a in art glomerata ree

- p s Griseb.
3423 in part glomerata Hieron 4
3486 elanduloso-pubescen
3541 winsintatl Hieron.
Fries, R. E.
1388 chacoénsis R. E. Fries

Giinther

7 boliviensis Sch Bip.
8 soratensis Hieron

ens
26 Bangii v. dyscrita Robinson
27 Bangii v. dyscrita Robinson

Hieronymus & Lorentz
876 fruticosa Griseb.

area
381 Ban ry
542 bok vichais Sch.-Bip.
665 boliviensis Sch.-Bip.

Julio, Bro.
57 Bangii v. dyscrita Robinson
70 im part setifera Rusb
70 in part soratensis v. sube-
andulosa
315 setifera Rusby
365 Bangii v. Reteniie Robinson

Kuntze

angii Rusby
boliviensis Sch.-Bip.

— obovata

sarensis v, dieeltieueia Rob-

inson
—— tunariensis Hier
oo Vv. _ boliviensis
eron.) Robinso

Mandon
242 in part boliviensis Sch. tet
242 in seh discolor Robinso
243 in part slivigols Hohanaan

sora

245 in part soratensis v. sube-
glandulosa H
246 Mandonii Sch.-Bi

1p.
246 bis (Poem v. dyscrita Rob-
in

McCarty
120 setifera Rusby
eae

—— Fiebrigi
—— merc hee Hieron.

Bip.
1615 in part boliviensis Sch. -Bip.
1615 in onl setifera Rusby

THE STEVIAS OF BOLIVIA 69

Stein
1126 sarensis v. dissiticeps Rob-
inson
1143 sarensis Robin
2790 oe v. lisitices Rob-

3715 boliviensis Sch.-B

5169 in part sarensis Robins

5169 in sche sarensis v. pti rennet

inson

5169a carensis v. dissiticeps Rob-

5835 Stevia n. sp.? but material
teratological

6117 sarensis Robinson

6248 — v. dissiticeps Rob-

7048 ens Robin:
8831 — Vv. Sitters Robin-

9147 Bie tain v. ist ag

n.) Robin
Stiibel
52d Stuebelii Hieron.

Ta
669 urticaefolia v. pallidiflora

inson
749 boliviensis Sch.-Bip,
870 kuhnioides Rusby
874 boliviensis Sch. “Bip.
1049 boliviensis Sch.-Bip.
White
168 setifera Rusby
267 boliviensis Sch.-Bip.

ae 8.

124 filipe ste
1468 ir Piet Rus by
andonii Sch.-Bip.