JOURNAL
OF THE
ARNOLD ARBORETUM

HARVARD UNIVERSITY

EDITORIAL BOARD

C. E. KOBUSKI, Editor I. M. JOHNSTON
I. W. BAILEY KARL SAX

VOLUME XXX

JAMAICA PLAIN, MASS.
1949

Reprinted with the permission of the
Arnold Arboretum of Harvard University

KRAUS REPRINT CORPORATION
k

1968

DATES OF ISSUE

No. 1 (pp. 1-110, 5 pl.) issued January 15, 1949.
No. 2 (pp. 111-210, 2 pl.) issued April 15, 1949.
No. 3 (pp. 211-344, 9 pl.) issued July 15, 1949,
No. 4 (pp. 345-468, 1 pl.) issued October 15, 1949.

Printed in U.S.A.

TABLE OF CONTENTS

ADDITIONAL NOTES ON DEGENERIA VITIENSIS. With one plate and one text-

NALABIEY od oa ele WA CA oy 01Te pee aren at a RN gro Ais Par ct Ogee eI 1
FURTHER CONTRIBUTIONS TO THE MORPHOLOGY OF THE DEGENERIACEAE.
With four plates and thirteen text-figures. By G. L. Swamy............. 10
ee PAPUANAE ARCHBOLDIANAE, XVIII. By E. D. Merrill and L. M.
PECTIN a ei) a ether ea aera as Sg fd) nh hea ee ane 39
er OF THE ANGIOSPERMS: NEED FOR A BROADENED OUTLOOK. By J. W.
Se RS NS cre sc te Rs SAS sls Ug enti hae eee tere Read 64
VALIDATING BOTANICAL NAMES BY REFERENCE TO PRE-1753 LITERATURE.
MBS ge Ce Cia) COs etree NE ay eat Say 5 ah gh h.4- Blase ee as A ene 71
Notes ON SoME FIjIAN WEEDS AND INTRODUCED PLants. By William
(HATER GT UGE eR cs es Se aa ade De Re ee 5
STUDIES IN THE BORAGINACEAE, XVII. By Ivan M. Johnston.............. 85
STUDIES IN THE BORAGINACEAE, XVIII. BORAGINACEAE OF THE SOUTHERN
WEST UNDIES s abe TOME LORMSLOM 5 Vos 40 deus nu 0 2 ee det
PLANTAE PAPUANAE ARCHBOLDIANAE, XIX. By Lily M. Perry.......... 139
STUDIES IN THE THEACEAE XVIII. THe West INDIAN SPECIES OF Lap-
TEAS BAU: DV OLOF CICCH EA COUUS RTs. cs ih, 4-0 :8-owb-adleee ee aha ER ee 166
THe MoRPHOLOGY AND RELATIONSHIPS OF CERCIDIPHYLLUM. With two
plates and eight text-figures. By G. L. Swamy and I. W. Bailey......... 187
THE MorpPHOLOGY AND RELATIONSHIPS OF AUSTROBAILEYA. With seven
plates and three text-figures. By J. W. Bailey and B.G. L. Swamy...... Pla

A New AGAVE FROM ARIZONA. With two plates. By Susan Delano McKelvey 227

A New GENUS OF THE ARALIACEAE. By Hui-Lin Li... 0... eee PigHl
THE GENUS ILEX IN CHINA. With nine text-figures. By Shiu-ying Hu.... 233
JOHN GrEorGE JACK. 1861-1949. With portrait. By Karl Sax............. 345
Tue Genus ILex in Cuina, II. By Shiu-ying Hu... 2.0 ee 348
THE GENUS COCCOLOBA IN CuBA. By Richard A. Howard. .............. 388
THE GENUS CaRyA IN Mexico. By Wayne E. Manning................. 425

FURTHER NOTES ON THE FERNS OF Fijt. By Edwin B. Copeland.......... 433

iv TABLE OF CONTENTS

ADDITIONAL NOTES ON THE FIJIAN SPECIES OF PEPEROMIA. By 7.G. Yuncker. 443
Tue ARNOLD ARBORETUM DURING THE FIscAL YEAR ENDED JUNE 30, 1949. 450

BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STUDENTS
Joie 1, 1048) UNE 90 108s aunts cdoht ented eeleebe wwe) eames

STAFF OF THE ARNOLD ARBORETUM 1948-49... 0000000 ce 456

JOURNAL

OF THE

ARNOLD ARBORETUM

VoLt. XXX JANUARY 1949 _, Numper 1

ADDITIONAL NOTES ON DEGENERIA VITIENSIS
A. C. SMITH
With one plate and one text-figure

WHEN THE FAMILY Degeneriaceae was proposed in 1942 (in Jour. Arnold
Arb. 23: 356-365. pl. 1-5) by Prof. I. W. Bailey and the writer, it must
have seemed to some readers that we were highly optimistic in suggesting
that a new species based on only two known collections be immediately
established as representing a monotypic family. Few plant families have
such an abrupt inception. It is more traditional to refer a genus of doubt-
ful affinities to some presumably allied family, in which it may languish
for many years or decades, until the researches of monographers exclude it
from all likely families and establish that it merits family rank. Even
then the position of the genus is suspect, for many herbarium custodians
are allergic to monotypic or small families and, rather than alter their
system, cling to unwieldy plant-groupings which have no morphological
justification. Whether the family Degeneriaceae will be accepted in her-
baria or will be thrust into a distended ‘“magnoliaceous” concept (which
already in some herbaria includes highly discordant elements) remains to
be seen

However, in 1942 Prof. Bailey and I could not bring ourselves to place
our new genus Degeneria in either the Magnoliaceae or Himantandraceae,
clearly the only existing families of its immediate alliance. The discovery
of a third collection, from southeastern Viti Levu (cf. Bull. Torrey Club
70: 537. 1943), did not add to our knowledge of the morphology of the
genus. In order to verify our conclusions or to make possible a re-
evaluation of our family concept, new material for morphological and
anatomical study was clearly desirable. Consequently, when I had an
opportunity to undertake a second collecting trip to Fiji in 1947,1 one of
1The collections upon which this study is primarily based were obtained in Fiji be-
tween April, 1947, and January, 1948, in the course of exploration under the auspices

the American Philosophical Society, and the Bache Fund of the National Academy
of Sciences. To these organizations the writer is deeply grateful.

2 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

my primary objectives was to locate, if possible, additional specimens of
Degeneria vitiensis and to obtain material for study by my colleagues
Prof. Bailey and Dr. B. G. L. Swamy. It is not within the scope of this
treatment to discuss in detail the morphological aspects of Degeneria,
now somewhat better known than in 1942, which justify its position in a
distinct family. In the following article in this Journal, Dr. Swamy ably
analyzes the characters of our genus. Here I shall refer only to the occur-
rence, habit, and habitat of Degeneria vitiensis, on the basis of my field-
study of it in 1947.

OCCURRENCE

The principal purpose of the Fijian exploration mentioned above was
the acquisition of herbarium material which would permit a better know-
ledge of the flora and which would widen the basis for future floristic
studies of the archipelago. Consequently I attempted to visit those
regions which had been comparatively neglected by previous collectors.
The first six months were spent on Viti Levu, for the most part at higher
elevations. My first objectives were the two forested mountain-ranges
which stand above the predominantly reed-covered western portion of
Viti Levu—the Mt. Evans and Mt. Koromba Ranges. During two
months in this area no specimens of Degeneria were found. Subsequently
a base was established at Nandarivatu, a now nearly deserted government
station on the northern escarpment of Viti Levu at an elevation of about
800 meters. Irom this base numerous trips southward, into the extensive
forested area drained by the upper Singatoka River, were made. During
June and July the forested area dominated by Fiji’s highest peak, Tomanivi
| Mt. Victoria], was explored without observation of Degeneria. Degener’s
type-locality is in the vicinity of a small Fijian settlement, Nauwanga,
which lies in the valley of Nandala Creek (one of the Singatoka head-
waters) a few miles south of Nandarivatu; I did not work in that imme-
diate vicinity in detail.

Due to other interests, I had relegated the search for Degeneria to a
lesser position when I began to work southward from Tomanivi onto the
Rairaimatuku Plateau, a poorly drained, uninhabited, and essentially un-
collected upland region, about 25 by 15 miles in size, occupying the center
of Viti Levu. It was, therefore, an exciting surprise to find the desired
species. Collectors will agree with me that the rediscovery of a rare plant
gives an emotional pleasure incomparably greater than its original dis-
covery, at which time it is unexpected and usually unrecognized. For my
part, I have no remembrance whatever of my first collection of Degeneria
on Vanua Levu in 1934. But the date (August 7, 1947) and the place
of my second collection of the species will not be forgotten by me. From
my headquarters in the village of Nandrau, which lies on the steep western
slope of the plateau in the Singatoka valley, I was working toward the
smaller village of Nanga, which is similarly situated a few miles to the
south. The trail between these villages 4eads through the dense forest of
the plateau a mile or two from its edge. We had left this trail and were

1949] SMITH, NOTES ON DEGENERIA VITIENSIS 3

slowly progressing through the forest toward the east, over the rounded
summit of a small elevation, when one of my Fijian assistants brought
me a strange fruit from the forest-floor. (It is fortunate for most of us
who collect in tropical forests that we have keen-eyed native helpers,
without whom we would overlook many discoveries for which our her-
barium-labels unblushingly give us the credit.) After we located the tree
—and this was not easy in the dense-foliaged rain-forest — I had plenty
of time to compose myself, for it proved to be one of those stubborn trees,
bound to others by lianas, which takes half an hour to fell. Once the
tree was down and I had reassured myself it was really a specimen of the
elusive Degeneria, we searched it thoroughly and obtained material for
numerous herbarium specimens.

This individual was in the same condition as my Vanua Levu collection,
with small fruits in the stage of development which Prof. Bailey and I had
erroneously assumed to be mature in our original study. None of my
field-crew had seen the species before, but in the village of Nanga one old
man assured us that it was very rare but that he knew it by the name of
vavaloa, which can be translated as “black shoe.” The aptitude of this
name is apparent only when one has seen the mature fruit.* In other parts
of Fiji the species appears to be known as yaranggele (in the Wainunu
region of Vanua Levu) or masiratu (in southern Naitasiri, Viti Levu).

My field-crew, now being aware of the identity of the plant and of my
desire for adequate flowering material, sought carefully for additional
specimens, but none were forthcoming on the Rairaimatuku Plateau. How-
ever, after we had returned to the vicinity of Tomanivi we discovered the
species with increasing frequency, as it became more familiar to us. On
September 2, not far from a sawmill which operates at the western base of
Tomanivi, we found individuals in full flower and with mature fruits.
Later, we found the species in some abundance very near Nandarivatu, in
the precise locations which had earlier been searched in vain. The area
immediately adjacent to Nandarivatu and Tomanivi has been examined
by several highly competent collectors, including Gibbs, im Thurn, Gilles-
pie, and Degener. That none of them except the last obtained specimens
of this genus need cause no surprise. Like most trees of the upper storey
of the rain-forest, the vavaloa mingles its branches and foliage with those
of many other species, and even when the tree is in full flower or fruit
these organs are not visible from the ground. It is only an occasional

or ignored by one unfamiliar with the plant. Eventually my Fijian
i I

believe that only a trained forester could equal them in this respect.

2I have also recorded the name vavaloa as applied, on one occasion, to elle
monosperma, which has a fairly large inequilateral mature carpel, and on other
occasion to Hernandia olivacea. There is no literal reason for the eter Raina
and my informant was probably over-enthusiastic.

4, JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Altogether I observed and made notes upon 55 trees of Degeneria during
August, September, and October of 1947. From ten of these herbarium
specimens were taken, in addition to wood samples and abundant material
in preservative. It is obvious that the observed individuals form only a
minute fraction of the total population; there must be many thousands

f Degeneria trees in the region explored. This region may be defined as
the northern part of the forest which covers the central portion of Viti
Levu, and it may be circumscribed by a semi-circle, with a radius of less
than 15 miles, lying due south of Nandarivatu. The area would fall into
the old Province of Tholo North, now divided among the four Provinces
of Mba, Nandronga & Navosa, Ra, and Naitasiri.

Degeneria vitiensis can no longer be considered rare in upland Viti Levu,
but of course it cannot be described as common when compared with some
of the characteristic species listed later in this paper. To the southward and
southeastward of the region discussed above the forest seems essentially
similar, and it is likely that the same frequency of Degeneria would be
maintained on the entire central plateau of the island and in the forests
drained by the Rewa and Navua tributaries. In favorable localities it
occurs close to sea-level, as attested by Mr. B. E. Parham’s collections of it
in southeastern Viti Levu and my own in Vanua Levu.

It seems advisable to list here all the herbarium specimens of Degeneria
vitiensis at present known; they are arranged in general from north to
south on Viti Levu. My own 1947 collections are deposited in the her-
barium of the Arnold Arboretum, and numerous duplicates will later be
distributed. Mr, Parham’s collections are in the herbarium of the Depart-
ment of Agriculture, Suva, with duplicates at the Arnold Arboretum.

Vitt Levu: Mba: Hills between Nandala and Nukunuku Creeks, along trail
sae Nandarivatu toward Lewa, alt. 750-850 m., Smith 6170, 6190; pean slopes

f Mt. Nanggaranambuluta [Lomalangil, east of Nandan vain alt. 850- m., Smith
5301, 6318; hills east of Nandala Creek, about 3 miles south of pte alt.

24, 1941; hills between Nggaliwana and Tumbeindreketi Creeks, east of the sawmill
at Navai, alt. 750-800 m., Smith 5875, 5880, 6018 (juvenile) ; southwestern slopes of
Mt. Tomanivi (Mt. Viton), alt. about 850 m., Smith 5744; Nandronga &
Navosa: Northern portion of Rairaimatuku Pineau between Nandrau and Nanga,

r Wainunu River valley, alt. 0-200 m., Smith 1754

(duplicates in pdeye baa May 7, 1934
These localities are indicated on fhe: accompanying map (fig. 1), which
also shows in an extremely approximate manner ‘the boundary between

3 Mr. Parham kindly arranged to have a Fijian assistant in the Department of Agr
culture, Apenito Gonekalou, visit this same locality on Jan. 4, 1948. Several aa

several seedlings, which were preserved in FAA and which have been ee by
Prof. Bailey and Dr. Swamy. My own several attempts to germinate seeds at Nan-
darivatu failed, and seeds which were returned to Harvard by air-mail likewise ae
unviable.

1949] SMITH. NOTES ON DEGENERIA VITIENSIS 5

forest and grassland. Actually, of course, this boundary is a fairly wide
and complex transitional zone.

178°.

mae

TAVEUNI

v pe
ee 2

Fic. 1. Sketch map of western Fiji (Lau Group omitted), showing known stations
for Degeneria vitiensis. The dotted lines show the approximate boundaries, on the
two large islands, between forested and grassland areas.

To summarize this discussion of the occurrence of Degeneria vitiensis,
field-observations lead me to suppose that it may eventually be found in
any undisturbed rain-forest area of Viti Levu and Vanua Levu between
sea-level and approximately 950 meters. It should also be sought in the
forested regions of the islands of Taveuni, Koro, Ovalau, Ngau, and
Kandavu, although these forests appear much less diverse in composition
than those of the two large islands. I have not observed the species at
higher elevations than 950 m.; from this point, to the crests of ridges and
summits of peaks the trees are smaller and frequently gnarled, with more
abundant epiphytes and cryptogams. The highest elevation in Fiji is
1323 meters, on the summit of Tomanivi, but several hills on Viti Levu
exceed 1000 meters. The highest elevation on Vanua Levu is 1030 meters,

6 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

on the summit of Mbatini. Conditions of exposure in the Vanua Levu hills
seem to be such that the truly montane forest reaches lower elevations than
it does on Viti Levu, and consequently Degeneria is not anticipated above
700 or 800 meters on Vanua Levu.

HABIT

Seedlings which could be positively assigned to Degeneria were not
observed by me personally, but fortunately those collected in Naitasiri,
mentioned above, are available for study; their salient features are dis-
cussed by Dr. Swamy in the following paper.

The juvenile plant is a slender treelet with a spreading crown and very
large leaves. Although the petioles of these leaves are no longer than
some observed on mature trees, the blades are conspicuously larger,
attaining a size of 25-45 10-14 cm. They are sometimes gradually
attenuate to the base and long-decurrent on the petiole, while the apex
may be deltoid-cuspidate and as much as 1 cm. long, a character not
observed in mature leaf-blades, which seem always to be either rounded
or lightly emarginate at the apex. The juvenile leaf-blades sometimes have
as many as 30 pairs of principal secondary nerves, whereas those of mature
plants seldom have more than 18 such pairs.

The mature tree, like most species of the Fijian rain-forest, is extremely
slender for its height, with a freely branching crown and dense foliage.
At apparently maximum growth it may be from 18 to 30 meters high, at
which stage its trunk has a breast-high diameter of only 45-70 cm. Char-
acteristically the trunk has 3-7 fairly obvious rounded buttresses, these
being usually apparent upward for about one or rarely for two meters.
The individual portrayed in Pv. 1, fig. A, is only half-grown and does not
have the buttresses well developed. The bark is dark gray and com-
paratively thick, with rather regular fissures.

Apparently our two original collections were diverse enough to cover
essentially all the dimensional foliar variation now known. It mav be
repeated that the leaf-blades of mature trees are extremely variable in
size, sometimes only 5 X 2.5 cm. (our original dimensions having been
9-27 & 3.5-13.5 cm.). The principal secondary nerves may be as few
as 8 per side (originally stated as 10-18).

Some specimens of Degeneria observed in 1947 were spectacularly
loaded with flowers, but, as mentioned above, the effect of this prolific
flowering is lost in the forest-canopy. Fully open flowers are sometimes
5 cm. in diameter, but they disintegrate and lose their petals very quickly.
At anthesis they emit a pleasing fragrance suggestive of that of some anno-
naceous flowers, e.g. Cananga odorata.

Dissection of preserved -flowers and fruits by Dr. Swamy permits the
following emendation of the original description. The sepals are very
rarely 4 in number, although 3 is certainly the characteristic number. The

etals (originally mentioned as 12 or 13 in number) are seen to be
occasionally as many as 18; they are arranged in 3-5 series, and the

1949 | SMITH, NOTES ON DEGENERIA VITIENSIS 7

maximum size now observed is 24 < 14.5 mm. (originally stated as
19 < 12mm.). The stamens may be 20-30 in number and 3- or 4-seriate;
the largest ones observed are 9.5 X 2.5 mm. (originally stated as 6 X 2.5
mm.), with thecae up to 5 mm. long. The staminodes are 10-13 in num-
ber (rather than 11 or 12) and up to 7 X 3.5 mm, in size. The carpel
is practically always single, as previously stated. However, in one dis-
section Dr. Swamy found two carpels in a flower, and one of my herbarium
specimens (no. 5555) shows a single instance of two developing fruits.
The number of ovules varies considerably beyond the 24 or 26 mentioned
originally, being as now observed 20-32, although the extreme numbers
are rare. The fruiting pedicel sometimes attains a length of 5 cm. The
mature fruit has not previously been described. The most advanced
fruits seen by me were still attached to the branchlets, but they occur far
back from the growing point, on a portion from which leaves have fallen.
Sometimes the same tree also bears half-developed fruits, which are more
distal on the branchlets but still below the flowers. Thus one may
observe specimens at full anthesis with flowers associated with the foliage,
with half-developed fruits somewhat lower on the branchlets, and with
essentially mature fruits still lower. In such cases fruits in an inter-
mediate stage of development were not found, and this may incline one
to suppose that two seasons (but not necessarily two calendar years) are
needed for fruits to reach maturity. When essentially mature and appar-
ently ready to fall from the tree, the fruits are rich pink to purple in
color, falcate-oblong-ellipsoid in shape, and up to 10.5 X 4.5 X 4.5 cm.
in size. I did not observe dehisced fruits attached to the tree, but some
which I dislodged were later observed, on the forest-floor, to have dehisced
along the ventral suture. The elaborate vascular skeletons of decayed
fruits were sometimes found, and these were always wide open along the
ventral margin. Hence we feel sure that our original mention of the fruit
as .indehiscent is inaccurate. The pericarp, except for its hard outer
layer, is very thick and fleshy, becoming greatly shriveled in drying. The
“waxy irregularly lobed appendages” which we originally described and
figured as arising from the endocarp are actually merely irregular rem-
nants of the carpellary wall. Most of the ovules appear to develop, but
only rarely are as many as 32 seeds to be found in a fruit. The fresh
seeds are up to 16 mm. long and 12 mm. broad; their outer coat is thick,
smooth, waxy, and bright orange-red in color.

HABITAT

Degeneria vitiensis is among the largest trees of the Fijian rain-forest.
It has been observed on flat land as well as on extremely steep hillsides, on
well-drained soil in undisturbed forest. The Fijian forest reaches its best
development under very equable climatic conditions, as demonstrated by
Tables 1 and 2, in which temperature and rainfall statistics are given for a
few selected localities in the forested portions of the two large islands.
The annual variation in temperature is comparatively slight, and alti-

8 JOURNAL OF THE ARNOLD ARBORETUM [VOL., Xxx

TABLE 1
Temperature statistics of selected localities in Fiji, from records
of Meterological Department, Suva

LACE, No. YEARS MEAN ABSOLUTE MAXIMUM ABSOLUTE MINIMUM
PROVINCE, COVERED TEMPERATURE ° F. AND MONTH ° F. AND MONTH
BY RECORD °F,
Suva, 56 77.1 98
Rewa, March July
Viti Levu November
Nandarivatu, 5 68.3 85 :
a, February July
Viti Levu March

TABLE 2
Rainfall statistics of selected localities in Fiji, from records
of Meteorological Department, Suva

PLACE, No. years NORMAL ANNUAL NorMAL NORMAL RAINFALL NORMAL RAINFALL
PROVINCE, COVERED RAINFALI NUMBER OF NCHES), (INCHES),
BY RECORD (INCHES) WET DAYS WETTEST MONTH DRIEST MONTH
— ea
Suva, 51-57 120.86 246.8 14.73 5.41
Rewa, March July
Viti Levu
Nasinu, 11-15 129.51 220.5 16.95 4.98
Naitasiri, April June
Viti Levu
Vunindawa, 20-33 140.10 175.6 17.16 4.14
Naitasiri, December July
Viti Levu
Nandarivatu, 19-45 137.63 176.3 25.06 3.56
Mba, February July
Viti Levu
Wainunu, 20-66 146.46 199.0 18.23 5.64
January July
Vanua Levu
Yanawai, 8 202.34 237.0 22.01 13.14
Mbua, May June
Vanua Levu
Salialevu, 11-25 213.57 183.5 20.99 13.63
Thakaundrove, October July
Vv

ua Levu

tudes up to 800 m., that of Nandarivatu, produce an average fall in
temperature of only 10°F. as compared with localities at sea-level. Rain-
fall is well distributed throughout the year in the forested areas of Fiji,
and even in the driest months four or more inches of rain may be expected.
For permission to reproduce these statistics I am greatly indebted to
Mr. Ralph Dyer, Director of the Meteorological Department in Suva.

The forest inhabited by Degeneria cannot adequately be characterized
by mention of any single species or any small group of species; it is
greatly diversified, although sometimes locally dominated by the ndakua
(Agathis vitiensis), conspicuous by its comparative great size. The more
frequently observed components of this forest, as seen in north-central
Viti Levu, may be mentioned as follows:

Jour. ARNOLD ARB. VoL. XXX PLATE I

DEGENERIA VITIENSIS Bailey & Smith

a

1949] SMITH, NOTES ON DEGENERIA VITIENSIS 9

LARGE UPPER STOREY TREES (approaching or exceeding 30 meters in
height): Agathis vitiensis (common), Podocarpus vitiensis (common),
Dacrydium sp. (frequent, no. 6244 4), Ficus obliqua, Hernandia olivacea,
Endospermum macrophyllum, Canarium vitiense, Garcinia pseudoguttifera,
Calophyllum vitiense, Elaeocarpus spp. (e.g. no. 5954), Sapotaceae spp.
(e.g. no. 6195).

TREES OF MEDIUM HEIGHT: Podocarpus sp. (asimbolo, an important
timber tree), Gironniera celtidifolia (often abundant), Lauraceae spp.
(Cinnamomum, Cryptocarya, Endiandra), Myristica castaneaefolia, X ylo-
pia no. 5868 and other Annonaceae, Elaeocarpus spp. (e.g. no. 6189).

UNDERSTOREY TREES: Gnetum Gnemon, Aglaia spp., Schefflera vitiensis,
Discocalyx spp., Diospyros spp., Couthovia collina, numerous Rubiaceae,
including Psychotria spp., Airosperma trichotomum, Readea membranacea,
and Mastixiodendron flavidum.

Suruss: Piper spp., Elatostema Archboldianum and E. no. 5976,
Cyrtandra spp., Psychotria s

VINES: LEpipremnum Pn din Piper insectifugum, Agatea violarts,
Medinilla spp., Strychnos vitiensis.

This type of rain-forest abounds in ferns, both terrestrial and epiphytic,
but lower cryptogams are comparatively few, reaching their greatest pro-
fusion at higher elevations or on exposed ridges where the above com-
ponents are generally not found.

EXPLANATION OF THE PLATE

Degeneria vitiensis I. W. Bailey & A. C. Smith. Fig. A. Trunk of medium-sized
B

tree. Fig. B. Flowering and fruiting branchlets, showing position of essentially
mature fruit below the ae x 1/5. Fig. C. Flowering branchlet, with nearly
mature flowers, xX 1/3. _D. Essentially mature fruit, showing line of ventral

suture, xX 1/3. ig. E. eee view of essentially mature fruit, x 1/3. Fig. A
from Smith 6301; Figs. B-E from Smith 5880

DEPARTMENT OF BoTANy,
U.S. Nationa, Museum,
SMITHSONIAN INSTITUTION.

4In these a a few species which have not yet been identified are referred to by my
field-num

10 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

FURTHER CONTRIBUTIONS TO THE MORPHOLOGY
F THE DEGENERIACEAE

B. G. L. Swamy
With four plates and thirteen text-figures

DURING HIS COLLECTING TRIP in the Fiji Islands in 1947,1 Dr. A. C.
Smith was able to pickle valuable material of Degeneria vitiensis I. W.
Bailey & A. C. Sm. This material includes not only leaves, nodes, shoot
apices, seedlings, etc., but also flowers and fruits in almost all stages of
development. Although most of the material was fixed in formalin-acetic-
alcohol, limitation of resources in one case forced Dr. Smith to substitute
“gin” when he secured important developmental stages of the gameto-
phytes. However, the gin-fixed material has fortunately rendered itself
suitable for interpretation, although the slides may not satisfy all the
requirements of technical perfection. The material on hand has ‘made
possible not only a verification of the original findings of Bailey and
Smith (1), but also new observations on the gametophytic and post-fer-
tilization development, vascularization of the flower, seedling structure,
etc. I sincerely thank Dr. Smith for placing the material at my disposal
and thereby providing me with the opportunity of studying it. I cannot
adequately express my gratitude and appreciation to Prof. I. W. Bailey
for his unfailing encouragement and illuminating suggestions during this
study and also for his assistance in the preparation of photomicrographs.

SECONDARY XYLEM

There is not much to be added to the anatomical characteristics of the
secondary xylem, which have already been described by Bailey and
Smith (1). However, the following points as revealed by the peripheral
regions of a large stem measuring about 60 cm. in diameter (3£ 5880) ,?
may be noted. In the inner regions of the secondary xylem, the multi-
seriate rays are predominantly 3-4-seriate, fairly high, usually with one
to three upright cells at either end as seen in tangential sections; the
uniseriate rays are infrequent, shorter than the multiseriates and com-
posed of upright cells. In the outermost region of this old specimen, the
multiseriate rays become predominantly 4—5-seriate, shorter and without
any upright cells at the margins. Uniseriate rays are eliminated. The
narrow bands of apotracheal parenchyma are distributed throughout the
secondary xylem. This situation is in contrast to the secondary xylem of
the magnoliaceous genera of the north temperate climate, where the
parenchyma is confined to the outer face of the growth rings.
1See the preceding article in this journal.
*In the few instances where it has seemed advisable to refer to a specific collection,
the number indicates Dr. Smith’s field-number. For locality, etc., see the preceding
paper in this journal.

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE ll

SEEDLING

Four seedlings of Degeneria were available for examination and all of
them possess three cotyledons. The whorl of thin, palmately veined coty-
ledons is preceded by a long hypocotyl, 6-10 cm. The first leaf is char-
acteristically emarginate, pinnately veined, and more nearly elliptic in

1 2.

Fics. 1-2. Fig. 1. Diagram showing the vascular pattern of the cotyledonary node
of Degeneria. Fig. la. Diagram to show the pattern of vascularization in the coty-
ledon, x 5. Fig. 2. Diagram showing the vascular pattern of the cotyledonary node
in Magnolia grandiflora.

12 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

outline. The subsequent leaves soon assume the elliptic to obovate-
elliptic norm of the genus

Two median vascular traces originating from a single gap in the stele
enter each cotyledon. These strands remain distinct in the petiole,
Figs. 1, la, but unite in the basal part of the lamina into a single strand
which then forms the median vein of the cotyledon. The phloem of the
strand maintains its doubleness till about half the distance in the coty-
ledon and then unites: The marginal traces of the cotyledons arise from
separate gaps in the stele, Fig. 1, and bifurcate in the basal part of the
lamina, Fig. la; the bifurcated strands diverge in the blade in a palmate
fashion. The branches of these lateral strands vascularize most of the
lower half of the blade. The remainder of the cotyledon is vascularized
by the branch system of the median strand, Fig. la.

The seedlings of the Magnoliaceae (Magnolia, Michelia, Liriodendron)
are provided with two cotyledons in the great majority of cases, Fig. 2.
The median pair of strands originates in the same manner as in Degeneria,
but the four lateral strands of the two cotyledons are derived by the bi-
furcation of two traces which arise from two gaps in the stele. Even in
those abnormal instances of tricotyledonous seedlings which are now and
then encountered in Magnolia stellata, the lateral traces of the adjacently
placed cotyledons arise in the same way. This pattern seems to be
stabilized fairly well in the family, although Magnolia Soulangeana (7)
appears to possess a nodal anatomy similar to Degeneria.

ANATOMY OF THE NODE AND PETIOLE

The first few leaves of the seedling receive three traces from a corre-
sponding number of gaps in the stele, Fig. 3A, and thus the nodes on
which these leaves are borne are trilacunar. In serially arranged trans-
verse sections of the petiole, the three traces exhibit the following changes
in their courses outward into the leaf. The median trace trifurcates,
forming a large central strand and two smaller lateral ones. The lateral
strands, from a purely descriptive point of view, appear to shift adaxially
toward the upper part of the petiole and split into four small bundles
having an inverted orientation of xylem and phloem, Fig. 3B. The lateral
traces (compare Figs. 3A and 3B) bifurcate, and the more adaxially situ-
ated products of this division subsequently divide, forming a total of six
strands, three on each side of the derivatives of the median trace. Higher
in the petiole, the large abaxial segment of the median trace and the
branches of the lateral traces converge in the form of a large arc of seven
strands, Fig. 3C. This arc is closed by the four smaller derivative strands
of the median trace, thus giving rise to an adaxially flattened eustele.

In contrast to the seedling nodes, those on the mature stem are penta-
lacunar, Fig. 4A, as reported by Bailey and Smith (1), although very
large leaves of vigorously growing saplings may receive more than five
bundles from the corresponding nodes. With increase in the number of
traces the vascularization pattern of the petiole becomes more complex.
The median trace trifurcates at the base of the petiole, Fig. 4B, the lateral

1949 ] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 13

segments of which break up into a varying number of eight to twelve
strands. These smaller strands aggregate in the upper part of the petiole,
the individual strands orienting to form a loose “inner eustele,” Fig. 4C.
The second pair of laterals also divide. At a higher level, two to four
strands from the “inner eustele” swing back into an abaxial position on
either side of the median trace, as indicated by broken arrows in Fig. 4C.
The paired adaxial segments of the first pair of laterals join the corre-

Fics. 3, 4. Fig. 3. Series of diagrams illustrating the behavior of the vascular
traces at the node and in the petiole of a seedling leaf. A represents the nodal level;
and C, basal and upper levels of the petiole. Fig. 4. Same; in the adult leaf. A
represents the nodal level; B, C, and D, successively higher levels in the petiole.

14. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

sponding ends of the now more or less opened “inner eustele” in the form
of a strap, which occupies the flattened upper side of the petiole. The
remainder of the traces unite into a large horse-shoe-shaped arc, the ends
of which approximate the lateral extremities of the strap- shaped segment,
thus forming an adaxially flattened eustele, Fig. 4D. In spite of these
complications, the fundamental pattern — trifurcation of the median trace
and the organization of the adaxial strap-shaped segment largely by the
lateral segments of the median trace —is the same in both the seedling
and the adult leaves. In the latter, the number of bundles involved is
decidedly larger and a few of the strands of the median system swing
back into an abaxial position, which situation has no counterpart in the
trilacunar seedling leaves.

ORGANOGENY

The first structures to develop on the thalamus are the calyx members.
The differentiation of the petals takes place in quick succession, the
outermost members originating first, Fig. 82. There seems to be a time
lag before the initiation of the androecium, during which period the corolla
completely envelops the broad, moun shaped floral apex, Fig. 83. The
ontogenetic differentiation of the staminal whorls is also in a centripetal

Fics. 5-9. Stages in the development of the carpel.

succession. Some of the last-formed members of the androecium later
transform themselves into staminodes. After the initiation of the androe-
cium, Fig. 84, the floral apex is rendered narrower and somewhat blunt.
This apex is used up in the organization of the solitary carpel, Fig.
85. Thus the order of development of calyx, corolla, androecium and
gynoecium, as well as of the individual whorls of corolla and androecium,
is in centripetal succession.

1949]: SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 15

As a result of the confinement of meristematic activity to the rim of
the carpel primordium, an embossed periphery is created. At this stage,
the primordium appears as a shallow cup, having an outline similar to
that of a horse-shoe, the free ends of which are fused, Fig. 5. A few cells
on this ventral side fail to divide but the surrounding cells continue divi-
sions; as a result, a notch appears in the wall of the cup, Fig. 6. Further-
more, the cells of the rim on the dorsal side maintain a faster rate of
division and the tissue on this side grows rapidly in height. This results
in a conduplicate structure enclosing a cavity between the folds. The
structure itself is abaxially deformed, Fig. 7. The free edges grow out
externally in the form of flanges, whose margins flare apart, Figs. 8,9. A
transverse section of the carpel at this stage is represented in Fig. 89.
The internal surfaces of the flanges later become stigmatic, as will be
described on a subsequent page.

VASCULARIZATION OF THE FLOWER

An ontogenetic outlook in studies dealing with the vascular anatomy
of flowers is an essential factor for an understanding of the real significance
of the phenomena involved. The procambial system laid down during
the early development is subject to a high degree of modification as the
flower attains maturity. The modifications involve either an elaboration
of the system which brings greater complexities into the adult vascular
structure, or a reduction whereby the mature vasculature becomes very
much simplified. In order to emphasize the degree and nature of the
complications that attend the procambial pattern of the Degencria flower,
the topic will be considered at present under two heads — the procambial
pattern and the vascular pattern at anthesis. The adaptation of this
method should not be interpreted as implying that all the organs of the
flower have their vasculature represented only by the procambial strands
at any one single step and that the strands become differentiated into
xylem and phloem at another single step. In fact, hand in hand with the
extension of the procambial strands into the newly formed structures on
the floral axis, the older portions of the same strands become transformed
into xylem and phloem. This being the case, it would be misleading to
look at the phenomenon as involving two definite steps. The distinction
made at present is only for the sake of description.

Procambial pattern:

Figure 10 represents what may be called the “procambial diagram” of
a sector of the flower. This diagram is synthesized after following the
course of differentiation and extension of the procambium into the respec-
tive structures of the flower as they are being formed on the thalamus;
the secondary modifications of these strands have been deliberately omitted
in this figure, but have been incorporated in the diagram representing the
vascular pattern at anthesis, Fig. 11.

In general, the method of supply of the procambial strands to the sepals
and petals is very similar. Each member receives a median and two mar-

16 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

from the side with one of the calyx lobes facing the reader. s— vascular traces supply-
ing the sepal members; c—of petal members; a—of androecial members; g—of
gynoecium. Fig. 11. Wascular diagram of the same sector at anthesis. Lettering as
in Fig. 10

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 17

ginal traces, Fig. 10 (traces of the sepals marked s, of the corolla, c).
The ma rginal traces of the adjacently placed members are usually derived
from a common stelar bundle. However, each of the petals belonging to
the innermost whorl receives a single bundle, which trifurcates into median
and marginal traces. Each member of the androecium receives a single
trace. It may be noted that the supply of a group of four or five stamens
originates from a single stelar bundle (marked a). The remainder of the
procambial bundles (marked g) fuse with each other into seven to nine
larger units before entering the carpe

Vascular pattern at anthesis:

By the time the flower bud attains the stage of anthesis, the simple
pattern of vascularization just described becomes highly modified, of
course, retaining the original ground plan in all essentials (compare jae
10 otal 11; the latter diagram represents the same sector of the flower as
that depicted in the former). The important modifications that are in-
volved may be briefly stated as follows.

PERIANTH. — The vasculature of the perianth members is subject to a
high degree of modification. The median and marginal traces of each
member branch, and the branches anastomose not only with one another
but also with similar branches belonging to the members of the adjacent
whorls. Numerous new traces originate from the points of anastomoses
and occupy positions on either side of the median and marginal traces of
each perianth lobe.

ANDROECIUM. —In the androecial members, the modification is less
pronounced. The single vein trifurcates into the median and marginal
traces at the very base of the microsporophyll. In the staminodes, the
marginal traces undergo a further bifurcation at the base and are gen-
an more highly developed than in the stamens.

In this connection some points in the ontogeny of the stamens and
staminodes may be considered. The primordia of all the individual
members of the androecium are alike in shape, size and histological char-
acteristics. The cells are richly protoplasmic and show a great and uniform
avidity for stains. With subsequent development, parenchymatization
starts in the adaxially situated cells and proceeds in an abaxial direction.
When this phenomenon advances to about half the distance between the
abaxial and adaxial surfaces, a nest of small cells in the center becomes
differentiated as the procambial strand. It is at about this time that the
archesporial cells begin to differentiate in the hypodermal layer on the
abaxial side. The archesporial cells undergo the first periclinal division,
leading to the formation of the primary parietal and sporogenous layers.
The dévelopment until this stage is characteristically seen to take place
in all the primordia of the androecium, irrespective of their destination,
whether staminal or staminodal.

From this stage onwards, in those primordia that belong to the inner-
most whorl and in a varying number of primordia that belong to the next
outer whorl, the subsequent development of the sporogenous tissue be-

18 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xxx

comes arrested and all the cells excepting the procambial strand become
rapidly parenchymatized. By further differentiation of the component
cells and by a modification of the external form, the appendages become
staminodes. On the other hand, the development of the sporogenous
tissue continues in thé remaining primordia and the rate of parenchymati-
zation of the cells becomes somewhat retarded, so that an abaxial patch
of darkly staining cell layers still persists (compare the staminodes and
the stamens in Fig. 85). These appendages mature into stamens.

Thus it will be seen that both the staminodes and stamens show a funda-
mentally similar origin and development until a certain stage, and the
factors that make them develop into one or the other of these structures
operate later; both structures are similar in their method of vasculariza-
tion, although .the marginal traces in the staminode divide once again.
Furthermore, the stamens may often bear degenerate sporangia (1). All
of these facts indicate that in Degeneria, the staminodes are best inter-
preted as sterile microsporophylls.

GyNoEcIuM. — The gynoecium of Degeneria is represented by a soli-
tary carpel occupying a terminal position on the floral axis. After sup-
plying the members of the preceding whorls, a fairly large number of
cauline bundles (varying usually between 18 and 25) are left over in the
floral axis, Fig. 12. A few of these bundles, however, disappear at the
base of the carpel; others recombine into seven to nine larger bundles,
Figs. 13, 14. The three adjacently placed bundles (marked m and-v in
the figures) situated below the midrib region of the folded megasporophyll
run into it as median and ventral traces; the other bundles (marked s)
also enter the carpel and occupy positions in its wall as indicated in the
illustrations, Fig. 15. The median trace gives out extensive side branches,
which spread in the carpellary walls in a pinnate manner, Figs. 24-27.
The ends of the branches finally anastomose with the ventral trace of the
corresponding side. The branches of the ventral traces and to a small
extent those of the dorsal trace take part in the vascularization of the
ovules, as will be explained in a later paragraph. In contrast to the
behavior of the ventral and median traces, the extra traces (marked s)
in the carpel remain distinct and isolated — in location as well as in the
degree of branching — from the system of the median and ventral traces.
Even in the mature fruit, they persist as robust cords, maintaining the
same configuration, Fig. 29. In other words, in spite of the fact that
several traces enter the carpel, only three of them (the median and ven-
trals) behave as normal carpellary traces, whereas the others do not; the
presence and behavior of the median and ventral traces in the carpel is
typical and consistent, whereas the number of the extra traces is subject to
fluctuation. This situation leads one to suspect strongly that the extra
traces do not belong to the integral vascular system of the carpel and to
conclude that the carpel of Degeneria is essentially a three-trace mega-
sporophyll as in the Winteraceae(2).

How then can the extra traces of the carpel be explained? What may
be their possible significance? In this connection it is worth while to

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 19

recall the general behavior of floral apices in relation to vasculature.
Unfortunately such studies have not been extended to a wide range of
ORGS, | PE OWEVER, the information provided by Arber and other work-

7 Ts) 20
Fics. 12-23. Figs. 12-15. Transverse sections at successive levels starting from the
base of the carpel. Explanation in text. Fig. 16. Semidiagrammatized pattern of

vascularization in a two-carpelled gynoecium as seen from a side; reconstructed after
a study of serial transverse sections. Figs. 17-23. Transverse sections of the two-
Sone gynoecium at levels indicated in Fig. 16 by corresponding numbers. All
figs., 10.

ers 3 is indicative of the following salient generalizations: In multi-
carpellate flowers the vascular bundles of the axis that are left over after
supplying the perianth and androecium take part in the vascularization
of the carpels in a normal and uniform manner. Thus all vascular
bundles are “‘used up” by the carpels. This seems to be the usual be-
havior in the great majority of instances. But, particularly in genera
3 See , A. “The pietaaer ee of the flower: A study of some aspects of mor-
Reon Rodent ” Biol. . 12: 157-184. 1937, and literature cited therein.

20 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

and species that exhibit a series in the reduction of many carpels to one,
the following modifications are frequently seen, either singly or in various
combinations. (i) The vascular bundles (residual vascular tissue) that
would have supplied the missing carpel or carpels may still persist in the
floral axis either as such or after anastomosing with one another in various
ways. (ii) The residual bundles may disappear at a considerable distance
below the persisting carpel or carpels. (iii) They may fuse with the
bundles that are concerned in supplying the persisting carpel or carpels.
(iv) They may enter the persisting carpel itself as supernumergry traces.

The situation in Degeneria seems to be a combination of the factors
involved in items (i) and (iv). It may now be recalled that in Degeneria
a large number of the stelar bundles are left over in the floral axis at the
base of the carpel; that these fuse with one another into seven to nine
larger bundles; that three of them form the median and ventral traces of
the carpel; that the other extra bundles also enter the carpel and travel
throughout its entire length; and that the behavior of the extra bundles
within the carpel is markedly different from that of the median and
ventral traces. The extra bundles therefore probably represent those that
formerly supplied other carpels. In other words, the ancestral flower of
Degeneria was in all probability multicarpellate. That this was most
likely the actual condition is further supported by the method of vascula-
ture of a bicarpellate flower.

Very rarely, flowers of Degeneria bear two carpels, attached at slightly
different levels on the thalamus, Figs. 16-23. transverse section just
below the level of attachment of the carpels, Fig. 17, shows the bundles of
the axis arranged in a ring in the same manner as in a monocarpellate
flower, compare Fig. 12. The bundles unite in various ways with one
another resulting in eight larger bundles, Fig. 18. The three bundles
disposed towards the right hand side (marked m, v) supply the median
and ventral traces to the first carpel. Of the remaining five bundles, the
middle three (marked m’, v’) form the median and ventral traces of the
second carpel; the other two bundles (marked s) also enter the upper
carpel as supernumerary traces and behave in the same manner as in the
case of the normal monocarpellate flower, that is, remain distinct from
the integral vascular system of the carpel.

Two significant points emerge from this data. (i) The first carpel on

comes reduced and probably eliminated. . Conversely, the presence of
supernumerary traces in the normal monocarpellate flowers of Degeneria
is closely associated with the phylogenetic reduction in the number of
carpels.

VASCULARIZATION OF THE OVULES.— At the very base of the carpel,
the ventral trace gives off two or three side branches which further ramify

1949 ] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 21

and anastomose with the system of the median trace, Figs. 24-27. The
remainder of the trace continues to run along the ovule-bearing region.
In general, this trace remains relatively weakly developed and on this
account there is every possibility of overlooking its presence and at the

aa

WN

J

: \ \\

vy
vi i.

Same, at anthesis, x 25. Figs. 26

of origin of ovular primordia, x 25. 5 fle :
and 27. Variations in the vascularization of the ovules at anthesis. Small circles de-

note the funicles with the ovules removed. Semidiagrammatic,

are removed and their respective places of attachment are denoted as empty circles,
x 20. Also compare figs. 52 and 87. Fig. 29. Vasculature of mature fruit, x 2.5.

22 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. Xxx

same time mistaking the otherwise well developed supernumerary trace,
s in Figs. 24-27, 41-44, for the ventral.

The ventral trace is concerned with the vascularization of the majority
of the ovules. It sends out a number of slender branches all along its
length and each branch enters an ovule. This pattern is stabilized in
the ovules that are situated towards the basal region of the carpel. On the
other hand, some of the ovules that are situated especially towards the
distal end exhibit an altogether different situation. It has been pointed
out that the median trace of the carpel builds up an extensive branch
system and that the branches reach towards the ventral traces with which
they ultimately fuse. Frequently, some of the branches of the median
system fuse with the ovular traces given out by the ventrals (see the
penultimate ovules towards the distal end in Fig, 25), whereby such ovules
should be considered as having been vascularized by both the median and
ventral systems. Less frequently the ventral trace is precociously used
up and the distally situated ovules derive their vascular supply only from
the median system, Fig. 27. In very rare instances, the supernumerary
traces may also send out one or two slender branches that fuse with those
of the distally situated ovular traces, Fig. 26. These various methods of
vascularization often occur in the same carpel in varying combinations
and fluctuate from carpel to carpel.

MICROSPORANGIUM AND MICROSPORES

As pointed out by Bailey and Smith (1), the microsporangia are im-
mersed beneath the abaxial surface of the sporophyll. As already men-
tioned on a previous page of the present contribution, the primary
archesporium' differentiates in the hypodermal layer of the abaxial surface.
This fact negates any suggestions that the extrorse dehiscence in Degeneria
is due to an ontogenetic shift in the position of the sporangia.

Although the primary archesporial cells are strictly hypodermal in origin,
occasionally one or two cells belonging to the sub-hypodermal layer also
show the characteristics of archesporial cells (group of three cells in the
right side of Fig. 30). It is quite possible that the cells of the sub-
hypodermal layer may infrequently become a part of the archesporium
and share its subsequent development. Two adjacent groups of arche-
sporial cells, each group consisting of two or three cells as seen in
transverse sections, are differentiated between the median and marginal
traces of the sporophyll. The first division in the archesporium results
in the formation of parietal and sporogenous layers, Fig. 30. The parietal
layer by further periclinal divisions builds up four or five wall layers
and the tapetum, whereby the sporogenous cells become deep-seated in
the tissue of the sporophyll, Fig. 31. The outermost of the wall layers
later transforms into the endothecium. The sporogenous cells also may
divide once and thus increase in number before differentiating as micro-
spore mother cells.

Simultaneously with their enlargement, the microspore mother cells
become jacketed by a continuous layer of tapetum, Fig. 32. When the

ose
esnaeen
aS ne
Uy Way
; Pa
O26 a
Yipat=

)
1

:
58

> ey
C)
ae

oe

Sx

has divided into the parietal and sporogenous layers, x 200. Fig. 31. Same, show-
ing the differentiation of wali layers; the primary sporogenous cells have also divided,

x 200. Fig. 32. Same, showing the organization of tapetum, x 200. Fig. 33.

single sporogenous group with its tapetum, some of the tapetal cells are binucleate and
the nuclei in others are in various stages of division or reunion, x 200. Fig. 34. Two
adjacent sporangia showing the disappearance of the tapetum, uninucleate micro-
spores and the initiation of the endothecial thickening, x 50. Fig. 35. Same, at
the shedding stage; the pollen grains are two-celled, x 50. Fig. 36. Division of the
microspore. The broken line indicates the position of the germinal furrow, x 200.
Fig. 37. Pollen grain at the shedding stage showing the faintly-stained vegetative
nucleus and the darkly stained generative cell, x 200. Fig. 38. A young tetrad
showing the initiation of the furrow on the distal face of each spore, x 135. g.
39, Pollen grains of a tetrad immediately after separation, showing the well defined
furrow, x 135. Fig. 40. A germinated pollen grain removed from the papillate sur-
face of a carpel; note the emergence of the tube from the broadened end of the
furrow, xX 200.

24. JOURNAL OF THE ARNOLD ARBORETUM [VoL, Xxx

nucleus of the microspore mother cell is in the prophasic stages of the first
meiosis, the tapetal cells become binucleate, Fig. 33. Frequently some
of the nuclei fuse again and occasionally redivide. However, they do not
migrate out of their cells and their general behavior is in conformity with
the secretory type of tapetal organization. The cells of the tapetum
function actively until after the first division of the microspore mother
cell and then degenerate.

The microspore mother cell undergoes the meiotic divisions in a simul-
taneous manner and produces a tetrad of microspores. Usually, the
configuration of the latter is tetragonal, Figs. 38, 39, rather than tetra-
hedral. The young microspore, while it is still lodged within the original
wall of the mother cell, shows a conspicuous fold on its distal face, Fig.
38. At the time of separation of microspores from the tetrad, Fig. 39,
a narrow groove with slightly broadened extremities takes the place of
the distal fold. It is this groove that differentiates as the germinal furrow
(colpa) of the mature pollen grain. Thus it is clear that the position of
the germinal furrow in the pollen of Degeneria is distal.

The uninucleate microspores, Fig. 34,.soon become scattered in the
sporangial cavity. The nucleus of the microspore migrates towards the
distal side of the spore and a vacuole develops towards the proximal side.
The first division of the spore nucleus is accomplished in this position.
The orientation of the spindle during this division is such that the gen-
erative cell is always cut off towards the furrow-end of the grain, Fig. 36,
that is, towards the exterior end of the tetrad. The generative cell, after
a time, shifts its position and comes to lie nearer the interior of the
grain, Fig. 37. The pollen grains are shed in this two-celled condition.

At the time of the division of the microspore nucleus, the cell layers
separating the adjacent sporangia break down and their cavities become
continuous. The endothecium is characteristically localized in disposi-
tion, not extending beyond the outer faces of the sporangia, Fig. 35.

MEGASPORANGIUM AND FEMALE GAMETOPHYTE

The ovule-bearing region of the carpel is situated far back of the mar-
gins of the conduplicately folded carpel, a feature clearly demonstrated in
transverse sections, Figs. 41-44, 88-91. Longitudinal sections passing
through the plane of conduplication also reveal the same feature, Fig. 86,
and in addition, reveal the slanting orientation of the carpellary cavity
and the ovule-bearing region in conformity with the abaxial deformation
of the carpel. A single row of ovular primordia arises opposite each of
the ventral bundles, Figs. 41, 88. Ten to thirteen primordia constitute a
row and the ovule-bearing region itself is confined to the upper end of the
carpellary cavity.

The ovular primordia grow into the cavity at first in a vertically down-
ward direction, Figs. 41, 88. With the differentiation of the nucellus
and integuments, the apex of the ovule undergoes a curvature of 90°,
Figs. 42, 43, and by the time of sporogenesis, becomes bent on itself,
thereby assuming a completely anatropous position, Figs. 44, 91.

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 25

Soon after the differentiation of the archesporial cell in the nucellus,
the two integuments originate more or less simultaneously, Fig. 47.
When the megaspore mother cell is fully differentiated, the inner integu-
ment comes to have three layers of cells and the outer, four to five layers.
The number of cell layers in the integuments continue to increase and
at the time of fertilization the inner integument consists of four cell layers
at the region of the micropyle, but remains three layered in the part sur-

44

Fics. 41-44. Transverse sections of carpels at different stages of development to
show the curvature undergone by ovular primordia in soyaee the ae posi-
tion, X 12. m-—median trace; v—ventral trace; s—supernumerary trace

rounding the nucellus; the outer integument comes to have seven or eight
layers. As will be shown later, the thickness of the outer integument
becomes further increased during post-fertilization development. Both
integuments of the mature ovule envelop the nucellus completely and the
inner integument organizes the micropyle, Figs. 49-51.

The subepidermal archesporial cell, Fig. 45, divides into a parietal and
a sporogenous cell. The former gives rise to seven to ten layers of parietal
tissue that cap the gametophyte; some of the layers persist in the develop-
ing seed for a long time, Fig. 101. The sporogenous cell functions as the
megaspore mother cell, Figs. 46-48 and after undergoing the reduction
divisions gives rise to a linear tetrad of megaspores, the chalazal one of
which develops into the eight-nucleate gametophyte, Figs. 49-51. The
fusion of the polar nuclei takes place usually before fertilization. How-
ever, in a few cases it may be postponed until the intervention of the sperm.
The antipodals organize into cells and show signs of degeneration before
the ovules are fertilized.

STIGMA, POLLEN TUBE, FERTILIZATION
It should be emphasized again that the ovule-bearing region of a De-
generia carpel does not represent the margin of the carpel and that the
ends of the flared-out region of the sporophyll are the true margins.
During earlier stages of development, Fig. 88, the closely approximated
ventral surfaces of the conduplicate carpel form an open cleft extending

26 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

from the locule to the exterior. Later, when the ovular primordia are
differentiating into nucellus and integuments, the ventral halves of the
conduplicate carpel tend to flare apart externally, Fig. 89. The epidermal
cells of the adaxial surfaces in the region of this flaring develop into pro-

LL?

biup

l

nn

LT 7
ei
aN
eae
HH

a
why +
KS
AY See
FRU TLS
Pa
OO)
esd
Ss
UC)

9
Y»
a
LS
(Y
SY

SS rr TORT pene ichiss eg col

ce

yy

t) iy
oe s
Seta By
OOS
SES
a

Mere
LU

Fics. 45-51. Figs. 45, 47 and 48 are enlarged from
mounted with their corresponding orientation. Fig.

Figs. 41-43 respectively and are
Archesporial cell in the

Fig. 49. Linear tetrad of megaspores, the chalazal one enlarging. Fig. 50.
nucleate embryo sac. Fig. 51. Mature eight-nucleate embryo sac. Figs. 45-48,
< 182; Figs. 49-51, x 40.

tuberant hairs, Figs. 89, 93. This wave of glandular differentiation
spreads outward toward the margins of the carpel and inward into the
locule, Fig. 90, ultimately extending internally beyond the region of
attachment of the ovules, Fig. 87. As a result, the narrow cleft between
the adaxial surfaces that lie outside the ovule-bearing region becomes
occluded by a closely interlocking system of papillate hairs, Figs. 90, 91,
94. At anthesis, the hairs on the flaring region grow into a dense felt and
many of the individual hairs become two- or three-celled, Fig. 95, whereas

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 27
in the locule they do not become as long or multicellular. Thus, a
greater portion of the adaxial surface of the sporophyli — from the margin
to a considerable distance towards the interior of the ovule-bearing region
— becomes evenly papillate, Fig. 28. As will be shown presently, this
entire surface is concerned in the penetration of the pollen tube and hence
Atk :

xamination of alcohol-preserved specimens reveals that during
pollination the pollen grains become deposited anywhere on the adaxial
surfaces of the outwardly flaring parts of the carpel. A few grains were
also found attached on the abaxial surface; however, none of these grains
show any signs of germination and appear to have lost their cell contents
through degeneration. This observation indirectly emphasizes the im-
portance of the papillate surface as an essential factor in the germination

ee YK REN ESE
SEA OT OO eee Sa
BAS ah Wat es Wie

as ac}

AT LY ee.
SK een
ee. TT
ms PaO.
peal sGeiatawe
AN \eseOngee
ren NGinke
rl Saltersee
C3 AE
30) Nera syreOen
Be enaate

Fic. 52. Micropylar half of a mature ovule with a part of its funicle and associated
carpellary tissue showing the papillate epidermis, the course of the pollen tube,
10.

etc.,

of the pollen grains in Degeneria. The grains that are lodged on the hairy
surface put forth pollen tubes. The tube arises not from the narrow
region of the germinal furrow but rather from one of its broadened ends,
Fig. 40. The course of the tube into the carpellary cavity is strictly along
the papillate surface and at no time does the tube penetrate the carpellary
tissue. Upon reaching the ovule-bearing region, it continues to grow
between the funicles and reaches the extension of the papillate surface on
the inside of the ovule-bearing region. Here the tube wanders about a little
and then characteristically curves back and enters the micropyle, Fig. 52.

28 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

After reaching the lower end of the micropylar canal, the tube bores
through the apex of the nucellus until the upper end of the embryo sac
is reached. During the discharge of the pollen tube contents into the
embryo sac, one of the synergids is usually destroyed. Double fertiliza-
tion takes place in a typical manner. Triple fusion precedes syngamy.

ENDOSPERM, EMBRYO

The first division of the primary endosperm nucleus results in a trans-
verse chambering of the embryo sac, Fig. 53. The separating membrane
is very thin and delicate. The nuclei of both chambers divide almost
simultaneously and in quick succession, each division being followed by
wall deposition until the embryo sac cavity becomes filled with a mass
of extremely thin-walled cells, Figs. 54-58. The sequence of the divi-
sions, however, does not follow any prescribed method. The tissue keeps
on steadily increasing at the expense of the nucellus and finally occupies

Fics. 53-58. Fig. 53. An ovule showing two-celled endosperm and zygote, x 40.
Figs. 54-56. Early stages in the development of endosperm, x 100. Figs. 57, 58.
a, later stages; note the differentiation of oil-bearing cells in the outer integu-
ment, e pad of degenerated nucellar ceils (shown in black) at the chalaza
and = extension of the vascular bundle on to the opposite side in Fig. 58, x 30.

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 29

the entire space enclosed within the integuments. The development of
ruminations in this tissue will be considered in connection with the seed.

o@.2 @8

Fics. 59-76. Stages in the development of embryo; explanation in text. The
heavy transverse line in the figures denotes the boundary of the cells derived from
the basal and terminal cells of the two-celled proembryo, X 283. Fics. 77, 78. Em-
bryos dissected from mature seeds; note the three cotyledons, swollen hypocotyl and
stalk-like suspensor, x 90.

30 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

The zygote divides only after about 300 endosperm cells are formed.
The first division is in a transverse plane and results in the formation of
basal and terminal cells. The subsequent method of segmentation does
not follow a strict pattern, as may be seen from Figs. 59-69. However,
when a fairly large amount of material is studied, certain broad trends
of developmental aaa suggest themselves. The second cell generation
commences first in the terminal cell and this is invariably by a vertical
wall, whereas the Renee ene of the basal cell is generally delayed, Figs.
59, 60, 65. The division in the latter cell also is usually by a vertical wall,
Figs. 62-65, thus resulting in two superposed tiers of two cells, Fig. 64
less frequently, the wall laid down is oblique or transverse. The division
of the basal and terminal cells may be more or less simultaneous or in rare
instances the development of the terminal cell lags behind, Figs. 66, 67.
However, subsequent divisions follow in rapid succession, resulting in an
ovoid mass of cells. The derivatives of the basal and terminal cells are
distinguishable for a fairly long time, Figs. 73-75.

In spite of the highly plastic nature of the early segmentation, the
method of tissue differentiation in the undifferentiated mass of cells is
quite stabilized. All the derivatives of the terminal cell and about half
of the derivatives of the basal cell enter into the construction of the
embryo proper. The other half takes part in the construction of a massive
suspensor. With the differentiation of the suspensor, the identity of the
derivatives of the basal and terminal cells becomes obliterated, Fig. 76.
In the mature embryo, the suspensor persists as a short stalk, Figs. 77,
78, 106; the hypocotyl region is bulbous; the cotyledons are well dif-
ferentiated and three or rarely four in number. Internally, the stem
and root apices, an incipient root cap and the procambial system are well
organized. However, the size of the embryo is significantly small in
relation to the large amount of endosperm, Fg. 79.

The cotyledonary number in Degeneria is remarkable. In Magnolia
grandiflora (4) and possibly in other species of this family, tricotyledonous
embryos are encountered occasionally as abnormalities. On the other
hand, a tricotyledonous development in Degeneria seems to be the rule,
as not even a single instance of an embryo with two cotyledons was seen
among the large number of seeds examined. The following data give an

COLLECTION NUMBER OF TRICOTYLEDONOUS TETRACOTYLEDONOUS
NUMBER SEEDS EXAMINED EMBRYOS BRYO
5880 a. 42 from 3 fruits 34 8
b. 50 free seeds 44 6
6190 a. 46 from 4 fruits 39 7
b. 50 free seeds 40 10
6318 a. 72 a 6 fruits 63 9
b. 50 free seeds 42 8
310°

262 48
or, nearly 87% or, nearly 13%

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 31

estimate of the frequency of the occurrence of the tricotyledonous and
tetracotyledonous embryos in Degeneri

The embryogeny of magnoliaceous ae such as Magnolia virginiana
(5), M. grandiflora (4), M. Soulangeana, Michelia fuscata, M. Cham-
paca and Liriodendron tulipifera (unpublished observations of the author)
presents several points of similarity with Degeneria. The plasticity in
the developmental pattern of the two-celled proembryo, the organization
of an oval mass of undifferentiated cells before tissue differentiation, the
development of a massive suspensor and its persistence in the mature
embryo, the swollen nature of the hypocotyl, the nearly triangular shape
of the cotyledon, and finally the size relationships of the mature embryo
and endosperm, are the more significant resemblances. Although the em-
bryos of Degeneria are prevailingly tricotyledonous, such tricotyledonous
nodes are of not infrequent occurrence in large populations of seedlings of
magnoliaceous plants.

SEED

Fertilization affects not only the structures within the embryo sac but
also other parts of the ovule. The fertilized ovule undergoes enormous
increase in size and the unfertilized ones soon degenerate, Fig. 97. In
the former, the nucellar cells abutting upon the antipodal end of the
embryo sac soon begin to degenerate and stain very deeply, Fig. 96. With
subsequent growth of the ovule, the mass of degenerating cells becomes
compressed in the form of a pad, Figs. 57, 98. More and more nucellar
cells in the chalazal region become involved and cause an expansion in
the width of the pad, Fig. 58. During later stages, this structure becomes
very prominent and is contacted by the ramifications of the vascular
bundle of the ovule, Figs. 99, 100. Some of the magnoliaceous genera also
exhibit a similar feature in varying degrees.

At the time of fertilization, the outer integument consists of about eight
layers of cells, all the cells appearing homogeneous. After fertilization,
clusters of two to ten cells in the outer layers develop into oil- bearing
cells, Figs. 57, 58, 99. Some of the walls of the oil cells break down
during later stages to result in cyst-like cavities. In the mature seed
they become very conspicuous and occupy a more superficial position, Figs.
79, 100, 102.

Hand in hand with the early segmentation of the zygote, localized
patches of cells from the innermost layers of the outer integument begin
rapid divisions largely by tangential walls and grow out in the form of
wedges, Fig. 102. Thus the entire inner surface of the outer integument

mature endosperm thus becomes irregularly grooved and cleft, Figs. 79,
105, and presents all the essential characteristics of a ruminated type.
It must be emphasized that the nature of the rumination in Degeneria
is less exaggerated than in Annonaceae, Myristicaceae, Palmae, etc. In
these families, the ingrowths of the integument penetrate the endosperm

32 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

far deeper in the form of thin transverse plates, whereas in Degeneria,
they are more massive and wedge-shaped than plate-like; also, in the
families cited above, the partitions are more or less transverse and parallel
to one another, or extend in a converging manner from the periphery,
thus exhibiting a specialized manifestation, whereas in Degeneria the
wedges are scattered indiscriminately and the form of the individual wedges
is somewhat fluctuating. In some members of the Myristicaceae (6), it
is the inner integument that takes part in the organization of the rumina-
tion, but in Degeneria it is clearly the outer.

The epidermis of the mature seed becomes covered by a thick cuticle
and the epidermal cells undergo conspicuous elongation in a radial direc-
tion, Figs. 79, 80. That part of the outer integument wherein the oil
cells are located continues to persist as a succulent outer coat. The inner

ap

sca

av ty

yes
ne
ie

fey

ee
he
y

bee ote ae
1 e
Ser oN
ne a)
Pee

os
nN

(

a

80
79

Fics. 79, 80. Fig. 79. Diagrammatic representation of a median longitudinal
section of a mature seed, the plane of section passing through the vascular bundle,
x 20. Fig. 80. A portion of the seed coat and endosperm of Fig. 79 enlarged to
show the histological details, x 40. a@—cuticle; b— epidermis; c — oil-bearing
cells; d— fleshy coat; e—stony coat; f—dinner integument; g—endosperm; h —
embryo; i— vascular bundle.

part together with its ruminated outgrowths undergoes considerable hard-
ening and transforms into an inner stony coat, Figs. 79, 80, The
process of hardening first commences in the cells of the wedges and gradu-
ally works outward. Due to the uneven pattern of hardening, the stony
coat becomes irregularly ruminate on its outer surface also, Figs. 79, 103.

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 33

The inner integument becomes crushed into a membranous covering be-
tween the endosperm and the stony coat. The vascular bundle extends
to the opposite side of the funicle as far as the micropylar end, Fig. 79.

At this point, a reference to the nature of the hardening material in the
cells of the stony coat deserves special mention. In Degeneria, as also
in magnoliaceous genera, the cells become filled with vacuolar substances
during development and finally the entire cell contents are rendered ex-
cessively hard; the cell walls, however, remain thin and the cells do not
increase in size, Fig. 80. Thus the hardness in these cases is due to the
transformation of the cell contents. On the other hand, in the corre-
sponding tissue of annonaceous genera, the cell walls become heavily ligni-
fied and the individual cells elongate in the form of sclerotic fibres; the
secondary thickening frequently occludes the cell lumen. Thus the hard-
ening is due to the lignification of the cell walls. This important feature
serves to distinguish the seeds of the Magnoliaceae and Degeneriaceae
from those of the Annonaceae.

The mature seeds of the families Degeneriaceae, Magnoliaceae and
Himantandraceae share several features in common, at the same time
showing significant differences in histological details: In all the three
families, the inner integument is reduced to a membranous layer and the
outer integument increases in thickness during post-fertilization develop-
ment and takes part in the construction of the seed coat; the vascular
bundle becomes extended from the chalazal side toward the micropyle;
and the minute embryo is embedded in an extensive endosperm. ‘The
seeds of the Degeneriaceae and Magnoliaceae are highly opaque, oval or
round with a fleshy exterior, whereas those of the Himantandraceae are
decidedly less opaque, suborbicular, greatly compressed and submem-
branous. The outer integument differentiates into fleshy and stony coats
in the former families, whereas in the latter, such a differentiation is want-
ing; instead, the outer cell layers attain a cartilaginous texture. The
fleshy layer in the Degeneriaceae and Magnoliaceae contains the oil-
bearing cells. In the former family, their distribution is characteristically
in clusters and in the latter, isolated and diffuse. Whether the seed coat
of Himantandra contains oil cells could not be determined from herbarium
specimens. The nature of hardening of the stony coat is essentially
similar both in the Degeneriaceae and Magnoliaceae. The epidermal cells
lining the exterior of the seed in the Magnoliaceae and Himantandraceae
are isodiametric in shape, whereas in Degeneria they are conspicuously
elongated radially.

FRUIT

Long before anthesis, the carpellary wall differentiates into two regions.
The cells of the outer region stand out more prominently with their
darkly staining walls and protoplasts in contrast to the faintly staining
inner region, Fig. 81. After fertilization, important changes occur within
each of them. The cells of the outer region appear to be retarded in

34 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XxX

meristematic activity, numerous spicular cells develop, and the vascular
system built by the median trace becomes excessively ramified. On the
other hand, the cells of the inner region suddenly step up their meri-
stematic activity and grow into the carpellary cavity in the form of lobes
of a spongy consistency. The lobes intrude between the developing ovules,
at the same time overtaking the degenerate ones and pushing them to one
side, Fig. 97. Ultimately the entire ee cavity is densely packed
with the spongy ingrowths, Fi : and the interior of the fruit
becomes fleshy. The outer region of < eed eee wall, in which the
spicular cells and vascular system ramify, becomes increasingly tough and
coriaceous.

According to field observations of Dr. Smith, the fruits dehisce along
the ventral suture after falling from the tree. The seeds become dis-
seminated through this split, also perhaps facilitated by the decay of the
coriaceous wall. In such fruits, the vascular skeleton built by the median
and ventral traces appears as a coarse and closely woven mesh, and the
supernumerary vascular traces, which have now assumed the form of large
cords, clasp the outer surface of the mesh, Fig. 2

The observation of Bailey and Smith (1) that the seeds of one row
are strictly sessile, while those of the other row are borne on slender
elongated funicles, needs to be amended in view of the large amount of
material now available. Although such a condition occurs in a few in-
stances, it is subject to numerous exceptions. Sessile as well as long-
funicled seeds are indiscriminately distributed and invariably the former
outnumber the latter. Furthermore, the distinction between the two
categories is one of degree. The manifestation of these conditions seems
to be largely dependent upon the degree and direction of pressure exerted
by the invading spongy lobes of the carpellary wall on the developing
seeds. The rate of ingrowth is by no means uniform and simultaneous
throughout the surface, especially in the early stages. If the activity of
the spongy ingrowth in the immediate vicinity of the micropylar end of the
seed dominates, it is likely that the ovule in question is carried more
towards the interior of the carpel, whereby the funicle also becomes cor-
respondingly stretched. On the other hand, if the activity of the lobes
from the opposite direction dominates and tells upon the chalazal region
of the ovule, it is possible that the funicle is not subjected to extensive
stretching. It may be further noted that such a differential method of
growth of the carpellary wall is also responsible for the frequent readjust-
ment or reorientation of the seeds in the mature fruit.

In dried fruits, the fleshy region shrinks and is pressed back against
the outer coriaceous wall. The latter becomes very hard, as might be
anticipated in view of its high ratio of spicular cells and vascular ele-
ments. When the seeds are broken from their points of attachment with
the carpellary wall,” some of the shrunken spongy tissue also becomes de-
tached and forms a cupule-like appendage on the micropylar end of the
seed, as pointed out by Bailey and Smith (1). However, it must be borne
in mind that the seed in reality has no true appendages and that what

1949 ] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 35

appears to be a cupular appendage is clearly the dried remnants of the
spongy ingrowth of the carpellary wall.

RELATIONSHIPS OF THE DEGENERIACEAE

A critical evaluation of the various vegetative and reproductive char-
acteristics of Degeneria, Himantandra and the Magnoliaceae (sensu
strictu) with regard to their mutual affinities and systematic position has
already been published by Bailey and his coworkers (1, 3). Only a few
points that are not covered by them will be dealt with here.

The vascularization pattern of the cotyledonary node of Degeneria
differs from that in most of the investigated species of Magnoliaceae. In
Degeneria, each of the three or four cotyledons has a trilacunar attach-

ment, whereas in most Magnoliaceae the four lateral strands of the
paired cotyledons arise by the bifurcation of two traces, each related to
an independent gap. When three cotyledons are formed, the six lateral
strands arise by the bifurcation of three independent traces. However, a
nodal anatomy resembling that of Degeneria has been reported in Mag-
nolia Soulangeana (7) and may ultimately be found to occur in seedlings
of other representatives of the Magnoliaceae.

he nodal anatomy and the vascularization of the adult petiole follow
a characteristic and basically similar plan in the Degeneriaceae, Mag-
noliaceae and Himantandraceae, a fact that has already been recorded
(3). Furthermore, in the case of the Degeneriaceae and Magnoliaceae,
there are similar transitions between the trilacunar nodes of the first seed-
ling leaves and the multilacunar nodes of leaves from older plants. The
emarginate nature of the first one or two leaves succeeding the cotyledons
characterizes Degeneria and is of not uncommon occurrence in Mag-
noliaceae.

During microsporogenesis, the method of differentiation of the parietal
and sporogenous tissues, the organization of a binucleate secretory tapetum
from the innermost layer of the parietal cells,* the simultaneous method
of meiosis, the prevalence of a tetragonal arrangement of the microspores
at the tetrad stage, the cutting off of the generative cell towards the
exterior side of the tetrad, and the two-celled shedding condition of the
pollen grains are features shared by both Degeneria and the Magnoliaceae.
The broad microsporophylls of Degeneria and Himantandra, bearing long,
slender, deeply embedded sporangia have similar counterparts in certain
representatives of the Magnoliaceae. However, the stamens of this family
differ from those of Degeneria and Himantandra in having a less typically
parenchymatous hypodermal layer.

The origin and differentiation of the nucellus, the thin inner and thick
outer integuments, the formation of seven to ten parietal layers in the
ovule, the monosporic eight-nucleate embryo sac and the ephemeral nature

4 Maneval (5) reports that the tapetum in Magnolia ata is cut off from the
sporogenous cells. This observation seems to be erroneous. M. Soulangeana, M.
stellata, as well as in Michelia fuscata and Liriodendron plore I have seen only

a parietal origin of this layer

36 JOURNAL OF THE ARNOLD ARBORETUM [Von see

of the antipodals of Degeneria are present point by point in the Mag-
noliaceae. The ab initio cellular endosperm,® the excessive broadening
of the chalazal region of the ovule, and the development of the darkly-
staining pad of degenerated nucellar cells at the chalaza are again a
combination of characters that are common to Degeneria and the Mag-
noliaceae; the presence of the last two characters in Himantandra appears
to be almost certain, as can be judged from herbarium material. In
Degeneria the endosperm becomes ruminated, and this feature has no
counterpart either in the Magnoliaceae or in Himantandra.

The plasticity in the sequence of early cell divisions in the two-celled

ge

portion of tricotyledonous embryos appears to be confined to Degeneria
alone.

The outer integument takes part in the construction of the seed coat

l

layer. In Degeneria and the Magnoliaceae, the outer integument differ-
entiates into fleshy outer and stony inner coats; such a distinction is
absent in Himantandra, and instead, some of the outer cell layers become
modified so as to render the seed coat cartilaginous. The stony coat of
Degeneria is ruminate internally as well as externally, whereas that of
the Magnoliaceae is smooth; however, the histological nature of hardening
in both families is identical. Although the fleshy coat of Degeneria and
the magnoliaceous genera contains oil cells, the pattern of distribution is
different in the two families.

The fruit of Degeneria is relatively large as compared with that of
Himantandra or of the Magnoliaceae. Its interior is packed with fleshy
outgrowths of the carpellary tissue and the seeds become embedded in it.
In the Magnoliaceae, the fruit is either dry or succulent and in Himan-
tandra, fleshy. However, in the fleshy fruits of the latter families, the
aoe) wall does not form spongy lobes that embed the seeds as in
Degeneri

It will, = evident from the brief review presented above that a large
number of embryological characters are common to all three families.
Nevertheless, the dissimilar characters are seen in specific combinations
in each of them. Thus the summation of evidence again points to the

5 Earle (4) states that in Magnolia eae “The early divisions of the endo-
sperm nucleus are very rapid and the number of free nuclei formed is aa

Il.

small. .. Wall formation occurs shortly after the appearance of these free

i a t in M. virginiana (5), M. Soulangeana, M. stellata, ee ne fuscata
and Ke adenaron tulipifera (unpublished observations of the author), endosperm
is cellular from i T alls of this tissue in the early — are extremely

delicate and hard to see as also in Degeneria, unless the sections are counterstained
with eosin or fast green. It is very likely that the presence of walls in Magnolia
grandiflora escaped Earle’ s notice

1949] SWAMY, MORPHOLOGY OF THE DEGENERIACEAE 37

same conclusion already reached by Bailey, Nast and Smith (3) that “
the Degeneriaceae, Himantandraceae and Magnoliaceae we are concerned
with three distinct but closely related families.”

SUMMARY

Additional information on the anatomy of the secondary xylem, node,
petiole, seedling, flower, fruit and seed of Degeneria vitiensis is presented.

Ontogenetic and anatomical evidence indicate that the staminodes are
sterile microsporophylls in which the development of the sporogenous
tissue becomes arrested.

After supplying the vascular traces to the perianth and androecial
whorls, a large number of bundles are left over in the floral axis. They
reunite into seven to nine larger bundles and enter the carpel. Only three
of these behave as true carpellary traces. The remaining traces represent
the residual vascular traces that once supplied the now missing carpels.
The anatomy of bicarpellate flowers also supports the hypothesis that the
ancestral flower of Degeneria was multicarpellate.

The ovules derive their vascular supply in part from the branches of
the ventral veins, in part from those of the median vein, and in part from
the branches of both sets of veins. The stigmatic papillae extend to a
considerable distance towards the interior of the ovule-bearing region.
The ontogenetic occlusion of the cleft of the carpel is accomplished by the
interlocking arrangement of the papillae. The germination of the pollen
grain and the path of the pollen tube is largely determined by this papil-
late surface.

The germinal furrow of the pollen grain develops on its distal face. A
summation of evidence from the development of the gametophytes, endo-
sperm and embryo, and the structure of the node, petiole, seed and fruit,
confirms the earlier conclusion that in the Degeneriaceae, Himantandraceae
and Magnoliaceae we are concerned with three distinct but closely related
families.

LITERATURE CITED

. Bartey, I. W. C. SmirH. Degene pees a new family of flowering plants
from Fiji. ee Atnola Arb. 23: 356-365.

. Battey, I. W. & C. G. Nast. The co parative ee ioe of the Winteraceae.
IJ. Carpels. Jour. Arnold Arb. 24: 472-481. 1943.

. Battey, I. W., C. G. Nast & A. C. Smira. =o family Himantandraceae. Jour.

Arnold Arb. 24: 190-206, 1943.

Earte, T. T. Embryology of certain Ranales. Bot. Gaz. 100: 257-275. 1938.

. Manevat, W. E. The development of Magnolia and Liriodendron, including a
Ree of the primitiveness of the Magnoliaceae. Bot. Gaz. 57: 1-31. 1914.

. Maurirzon, i Contributions to the ie of the orders Rosales and Myr-
ales s Univ. Arsskr. N.F. Avd. 2, 35(2): 1-120, 1939.

. Tuomas, E. ie Seedling anatomy of erm Rhoeadales and Rosales. Ann. Bot.
28: 695-733. 1914.

ra

ise)

Ww

a

un

an

~I

38 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

EXPLANATION OF PLATES
PLATE I. Fics. 81-87. fig. 81. Transverse section of a flower bud long before
S. 82-85 ower

buds at different stages of SeeoaSny x see Fig. 86. pas section of a

portion of a eee at the time of fertilization to show the extension of the cay
epidermis on the inner side of the ovule- i region (indicated by arrow), §
m — micropylar tees occluded by papillae

PLATE II. Fics. 88-95. Figs. 88-91. Transverse sections of Re oe at different
stages of development of the papillate epidermis. Fig. 88. Very young stage; the
adaxial surfaces do not show any papillate differentiation at this stage (detail in
Fig. 92). Fig. 89. The epidermal cells of the flaring ventral surfaces are beginning

88-91, X 21; Figs. 92-95,

PLATE III. Fics. 96-102. Fig. 96. An ovule after fertilization. Note the degen-
erated and darkly-staining nucellar cells at the chalazal end of the ovule, x 38.
ig. 97. Transverse section of a young fruit to show the ingrowth of t he carpelary

a slightly older than in Fig. 97. The spongy tissue is invading between the ovules.
e darkly-staining chalazal pad is increasing in breadth (compare the seed in Fig.

a oon is from a corresponding stage in development), x 20. Fig. 99, Same, ata

still later stage. The ovule is cut somewhat obliquely. The characteristic distribution

effect of the ruminate outgrowth of the inner surface of the outer integument on the
endosperm, are clearly seen, x 18. — 100, Detail of the eee region of a young
seed at a stage similar to Fig. 99, showing the pad and its relation to the surrounding
tissue, x 20. ig. 101. malas region of a young seed (outer integument
removed) showing ‘the endosperm and embryo. Note the persistence of parietal
cells in contact with the oe a of the embryo, x 40. fig. 102. A portion
of a young seed in section showing the wedge-shaped ruminate outgrowths from
the inner surface of the outer integument, x 30.

PLATE IV. Fics. 103-106. Fig. 103. External surface of the mature stony coat
of the seed, x 10. Fig. 104. Internal surface of the same, x 10. Fig. 105. Endo-
sperm dissected from an immature seed to show the early stages of rumination, x 10.
Fig. 106. A group of mature embryos, 0

INSTITUTE FOR RESEARCH IN GENERAL PLANT MORPHOLOGY,
.Harvarp UNIVERSITY.

Jour. ArNotD Ars. VoL. XXX PLaTE |

MorPHOLOGY OF THE DEGENERIACEAE

Jour. Arnotp Ars. VoL. XXX Prate II

HES

Be
Len¢*

ote

2
¥ Sn .
Pe

h
A 7% 2

NS DS
ne
2
ate
srt >

ren

a eomeeraar lr

MorPHOLOGY OF THE DEGENERIACEAE

Jour. ARNOLD ArB. VoL. XXX Pirate III

fo a

et

MorPHOLOGY OF THE DEGENERIACEAE

Jour. ArNotp ArB. VoL. XXX PLATE IV

MORPHOLOGY OF THE DEGENERIACEAE

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 39

PLANTAE PAPUANAE ARCHBOLDIANAE, XVIII *

E. D. Merritt and L. M. Perry

THIS ARTICLE, like the others of this series, is a miscellany. In it we
have tried to bring together the misplaced odds and ends of families

lanaceae, and Cucurbitaceae. Owing to the official retirement of the
senior author in June, 1948, this will be the last of these joint papers.
For the convenience of those who are particularly interested in New Guinea
flora we are adding an index of the families and genera included in this
series.
TAXACEAE

Dacrydium xanthandrum Pilger, Bot. Jahrb. 69: 252. 1938.

In eat to the specimens cited by Pilger in oe original description, we have the
follow

NORTHEAST NEW GUINEA: Ogeramnang, eee 6408, May 1937, alt. 1750
m.; Yunzaing, Clemens 6488, June 1937, alt. 1350

SOLOMON I ISLANDS: Biota eae ilice Kupei Gold Field, Kajewski 1694,
1709, April 1930, alt. 950 m., rare in rain-forest (tree up to 25 m. high). Closely allied
but perhaps not Eonepecific with these are the following two specimens from the
Solomon Islands: Ysabel: Mt. Sasari, Brass 3264 (3), December 1932, alt.
100 m., found only on summits of highest peaks (thick boled tree seldom
more than 15 m. tall; branches spreading, upturned). Guadalcanal: Uulolo,
Tutuve Mt., Kajewski 2652, May 1931, alt. 1200 m., rare, in gold-mining country.

CARYOPHYLLACEAE

Sagina monticola sp. nov.

Herba perennis; caulibus ea anys glabris, novellis puberulo-glandu-
loso-pilosulis (pilis tenuissimis), breviter sarmentosis vel decumbentibus

nodos saepe radicantibus et yeaa plantas novellis evolventibus apice
floriferis; foliis inferioribus subcarnosulis subrosulatis, linearibus,
cronatis, usque 3.5 cm. longis, vix 5 mm. latis, basi (2.5 mm. latis) late
vaginantibus, margine et subtus costa distanter ciliolatis, supra planis,
subtus convexiusculis, utrinque minute pustulatis; foliis superloribus
similibus sed minoribus, 3-5 mm. longis; pedicellis tenuibus, 0.6-2 c
longis, consperse glanduloso- puberulis; floribus pentameris; sepalis
glanduloso-puberulis, ellipticis, 3-4 mm. longis, 2 mm. latis, apice late
obtusis, e basi 3-nerviis, nervis ramosis; petalis albis, 4-5 mm. longis,
oblongis, apice rotundatis; staminibus 10, filamentis exterioribus 2.5 mm
longis, basi glandula insidentibus, interioribus paullo brevioribus eglandu-
losis; ovario ovoideo 3 mm. longo, 2 mm. lato; stylis 5, ex apice ad
medium papilliferis; capsulis 4 mm. longis, seminibus 0.6-0.8 mm. longis,
oblique subrotundatis, extus minute muriculatis.

* Botanical results of the Richard Archbold Expeditions. See Jour. Arnold Arb.
27: 193-233. 1946

40 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

NETHERLANDS NEW GUINEA: Bele River, 18 km. northeast of Lake Hab-
bema, Brass 11569 (type), November 1938, alt. 2350 m., rooting in earthy niches on
a sparsely vegetated limestone precipice (flowers white).

This species was published, Jour. Arnold Arb. 23: 388. 1942, as Sagina
echinosperma Hayata. Since the war, Dr. Hiroshi Hara, who has done
some work on this group, called our attention to the fact that the New
Guinean plant is clearly distinct from the Formosan plant which has been
proved on further study to be S. japonica (Sw.) Ohwi, common in Japan.
The latter has short (less than 2 mm. long) petals, smooth leaves, and
5-8 stamens. The habit of this plant is rather distinctive. It suggests a
slender short caudex-like stem covered by relatively long leaves with
overlapping very much broadened bases, from this stem or stem-like base
slender decumbent branches appear to grow, and these eventually develop
new plants at the nodes, but may also terminate in an inflorescence. Noth-
ing in the duplicates indicates a tangled mass of branches such as one
expects to see in Sagina.

AQUIFOLIACEAE
Tlex een caer Loesener, Nov. Act. Abh. een Akad. Naturf. (Monog.
Aquifol.) 78: 432. 1901, var. stenura var.

A specie recedit foliis oblongis, 10— fs cm. longis, 2.4—5 cm. latis (uno
8.2 >< 2.5 cm.), basi obtusis vel cuneatis, apice caudato-acuminatis,
acumine 1.5—2 cm. longo, basi 4 mm. lato, apicem versus + 1.5 mm. lato,
obtuso.

BRITISH NEW GUINEA: Central Division, Mafulu, Brass 5175 (type of var.),
Sept.-Nov. 1933, alt. 1250 m., oak forest undergrowth, common (slender tree or
; small, fleshy purple-black fruit)

Although we have no material with which to compare this collection,
according to the description in Loesener’s monograph of the Aquifoliaceae
it seems to be very close to /lex malaccensis Loes. The fruits are young,
but in the one which was cut open there were apparently 14 locules. It
differs from the description of the species chiefly in the narrower leaves
with longer acumen.

log
c
mn
=a
t
a
3
5.
gQ
i=

STERCULIACEAE

In the Sterculiaceae of these collections we have the usual widely dis-
tributed species of Abroma, Kleinhovia, Commersonia, Melochia, Helic-
teres and Heritiera. In addition to the common Heritiera littoralis, there
is a sterile specimen collected near Bernhard Camp, at 400 m. altitude,
a specimen of Pterygota Forbesii, a sterile collection of Pterocymbium,
and one of Tarrietia. The last genus does not appear in our Index for New
Guinean plants, and as far as we know this is the first record of its occur-
rence in New Guinea. The specimen was determined at Buitenzorg as
Tarrietia Riedeliana Oliv. Other species of interest are recorded below.
Sterculia Clemensiae sp. nov.

Arbor magna, trunco 37-45 cm. diametro (fide M. S. Clemens) ; ramulis
10-15 cm. diam., rugosis, novellis dense pubescentibus; foliorum petiolis
4—7 cm. longis, circiter 1/3 longitudinem laminae aequantibus dense ad-

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 41

NORTHEAST NEW GUINEA: Morobe District, Lae, Clemens 10452 (TYPE),
10466, July 1939, in brush near Busu River (big tree, trunk 15 to 18 inches in di-
ameter; inflorescence brick-red-purple; fruits pale yellowish green).

Sterculia Clemensiae appears to belong in the same group with S. comosa
Wall. Both have very small flowers. The latter however has a shorter
smoother pubescence on the lower surface of the leaves giving them a
grayish color. It has been reported from Amboina, the Celebes, and Key
Islands.

Sterculia oncinocarpa F. v. Muell. & Forbes, Victoria Nat. 3: 48. 1886; Mildbr. Bot.
Jahrb. 62: 355. 1929; vel aff.

NETHERLANDS NEW GUINEA: Bernhard Camp, Idenburg River, Brass &
Versteegh 14102, April 1939, alt. 70 m., occasional in primary forest on edge of flood
plain (tree about 30 m. high, 58 cm. diameter; bark 12 mm, thick, black-brown, fairly
smooth; wood light brown; flowers yellow).

BRITISH NEW GUINEA: Palmer River, 2 miles below junction Black River,
Brass 7348, July 1936, alt. 100 m., common canopy tree of lower clayey ridges
(+ 30 m. high; trunk buttressed and covered with gray lenticellate bark; leaves
crowded at the apex of the thickened branchlets; fruit close below the leaves on
peduncles 20-25 cm. long).

Among the Papuasian species these two collections are closest to this
species and possibly identical with it. In any case it seems best to point
out the differences or variations. Mr. A. W. Jessep, Director of the Mel-
bourne Botanic Gardens and National Herbarium, very kindly sent us a
photograph of the type and a fragment of the leaf for comparison. The
texture of the leaf is very similar to that of Brass 7348 but the pubescence
is much finer and sparser. The stellations are minute. The rays of those
on the Brass specimen are at least five times as long as those in the Forbes
collection. Again, in the original diagnosis the fruitlet is described as
being about four times longer than broad; in the photograph, however,
there is one scarcely twice as long as broad. The fruitlet of the Brass
collection is approximately 8 x x 4.5 cm., the seeds are olive brown
and the surface is reticulate (the other seeds of Sterculia species in the
herbarium are black; possibly these are not ripe). The collection, Brass

42 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

& Versteegh 14102, has only very young leaves and flowers. Most of the
inflorescences (up to 20 cm. long) appear to be sessile or almost so at the
apices of short shoots 3—6.5 cm. long; rhachis, branches, branchlets, pedi-
cels, and flowers are all spreading-hirtellous with hairs -+ 1 mm. long;
6 FLOWER: calyx-tube 3 mm. long, puberulous inside, lobes 1.3 mm.
long, ovate, acute, finely pubescent on the upper surface; staminal column
2 mm. long, anthers 5 (4-6), annular; @ FLOWER: larger than the stami-
nate and terminal on the branches; calyx-tube 4 mm. and lobes 1—-1.5 mm.
long, pubescent as in the ¢ flower; ovary 3-lobed, + 1.5 mm. long,
densely hairy, style 1 mm. long, pubescent, stigma capitate. This collec-
tion is myrmecophilous.

Sterculia Shillinglawii v. Muell. Australas. Jour. Pharm. ee 1887.

Sterculia Conwentzit Schum. Bot. Jahrb. . oa Pes 18

Sterculia multinervia aa Rep. Sp. Nov. 9:

BRITISH NEW es EA: Lower Fly ee east oul opposite Sturt Island,
Brass 8089, Oct. 1936, n forest, on es ridges (common canopy tree; stem wi
well developed plank- reer bark gray-brown; i wers yellow- brow galled
flowers pink; fruit red, mericarps narrow and recurved).

NORTHEAST NEW GUINEA: vicinity of Kajabit, Markham Valley, Clemens
10557, 10857bis, August 1939, alt. + 240 m

HIPELAGO: New Brit ain: near Blanche Bay, Parkinson
(ryPE of S. Shillinglawii) ; vicinity of Rabaul, Kanehira 3994, Waterhouse 424 (large
spreading tree, excellent for shade; white flowers; striking red seeds in open fruit).

SOLOMON ISLANDS: Gua da leanal: Rerande River, Kajewski 2411, sea
level, rain forest (a medium-sized buttressed tree; pod red outside, salmon-colored
inside, seeds green. A person recovering from a long illness, such as fever, is given

Cristoval: Wainamura, Brass 2843, Sept. 1932, lowland rain forests, not com-
mon (small tree with thick, gray, fairly smooth bark; leaves stiff, glaucous below,
midrib and main nerves brown on lower surface; flower yellowish green with red
throat, the narrow lobes arched and inflexed).

Mildbraed, in his work on the Sterculiaceae of New Guinea, Bot. Jahrb.
62: 347-367, 1929, hesitated to reduce S. Conwentzii K. Schum to S.
Shillinglawii F. v. Muell., although admitting S. muliinervia Rech. as a
synonym. We are indebted to Mr. Jessep for a photograph and fragment
of the type of Mueller’s species, the isotype of K. Schumann’s species at
the Kew Herbarium has been examined by the senior author. There
seems to be no doubt that S. Shillinglawii is the oldest name for this
apparently fairly widespread species. Needless to say, there is some
variation in the specimens, but not enough to be of specific value.
Sterculia porphyroclada sp. nov.

Arbor + 20 m. alta; trunco + 20 cm. diam.; ramulis novellis 2—2.5
mm. diam., , purpureis in sicco jongitudinaliter ruguloso- sulcatis, annotinis
4.5 mm. diam., lenticellatis rimosis, sparsim patenti-hirsutis; foliorum
petiolis 1.5—6.2 cm. longis, circiter 1/ 4 longitudinem laminae aequantibus
patenti-hirsutis; laminis oblongo-ellipticis, 10-24 cm. longis, 5-11.5 cm
latis, Ae paululo retusis vel rotundatis vel obtusis, apice acuminatis,
acumine 1—2 cm. longo, obtusiusculo, supra costa et nervis -+ dense
Sais pilosis ceterum fere glabris, subtus stellato- -pilosis, ima basi tri-

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 43

nerviis (in foliis majoribus subquinquenerviis) , nervis lateralibus utrin-

spicua: inflorescentiis non visis; axi in fructu + . longo, 4 m

diametro; mericarpiis + 9 cm. longis, 2 cm. diametro, subfalcatis, apicem
versus paulo angustatis, apice acutiusculis, extus brevissime et densissime
ferrugineo- stellato-tomentellis, intus stellato -pilosulis; seminibus pluribus

8 in uno maricarpio ) in sicco atrobrunneis, circiter 15 mm. longis et 7—9
mm. crassis.

NETHERLANDS NEW GUINEA: 6 km. southwest of Bernhard Camp, Iden-
burg River, Brass ae ire ane ioe alt. 1200 m., occasional in rain-
forest (subsidiary - tree and 20 cm. diameter; fruit orange-red, seeds
white).

The species appears to be related to S. malacophylla K. Schum. It
has longish somewhat stiff hairs on the branchlets and petioles, and 4-
pronged stellate hairs on the lower surface of the leaves; however, the

on the younger branchlets is dark purple, not yellowish gray; and the
leaves have 10-13 pairs of veins as compared with 7-8 in S. malaco-
phylla K. Schum

Sterculia ee Baker f. Jour. Bot. 61. Suppl.: 5. 1923; White, Jour. Arnold Arb.

929.

Siercuta ee aed Bot. Jahrb. 62: te 1929.

BRITISH NEW GUIN Western Division: Oriomo River, Wuroi,
Brass 5785, Jan.-Mar. ae nivenne rain-forest ee mon erect sparsely branched
tree up to about 15 m. high; leaves clustered at branch tips; large red fruit, seeds
black) ; Lake Daviumbu, Middle Fly River, Brass 7711, 7785, common in rain-forest
(substage or lesser canopy tree; flowers brown-green).

In our herbarium is an isotype of S. coggygria Mildbr. which we are
unable to distinguish from the earlier described S. ampla Baker f. Again,
judging by the specimens Kanehira & Hatusima 11884, 12395, we are in-
clined to believe that S. gigantifolia sensu Kaneh. & Hatus. in Bot. Mag.
Tokyo 55: 389. 1941, also belongs here. The stipules are somewhat larger
and coarser than in the Brass collections, and the pubescence varies. It
might be worth noting that the inflorescence of Kanehira & Hatusima
11884 is 60 cm. long.
erat quadrifida R. Brown in Benn. Pl. Jav. Rar. 233. 1844; F. M. Bailey,

ns]. Fl. 1: 136. ee Ser Agric. Jour. 22: 147. 1909; C. T. White, Proc.
Rae "Soc, Queensl. 33:

BRITISH NEW ee sees Division: Mabaduan, Brass 6480,
6503, common in monsoon forests on granite pee (tree to 10 m. high; branches
flatly spreading; gh, brown; fruit generally 4-lobed, red, somewhat rugose;
seeds a beautiful velvety-black) ; Wassi Kussa R ; ara, Br 4, common i

dry brushy rain-forests (tree 14-15 m. high; bark gray, lenticellate, somewhat scaly ;
leaves gray underneath; fruit orange-red, seeds velvety-black).

Previously reported by F. M. Bailey from Boku, Papua. These speci-
mens are a reasonably good match for Brass 2378 collected on the Mow-

44 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

bray River, Queensland. All are finely and closely velvety-tomentose

on the under surface of the leaves. In the Queensland collection the fruit

is less pointed at the base, a minor difference. There is considerable

variation in the length and breadth of the leaves, those of Brass 8414

are oblong-lanceolate, 7.5-16 cm. long, 3.2-5.3 cm. wide; in the other

oe the leaves are more nearly elliptic or slightly obovate, 5.5-10
m. long, and 3.5—6 cm. wide.

Keraudrenia J. Gay

Keraudrenia corollata (Steetz) Domin, Bibl. Bot. ans 895): 974, 1928.

Seringia corollata Steetz in Lehm. Pl. Preiss. 2: 350.

NETHERLANDS NEW GUINEA: Waren, 60 miles south of Manokwari, Kane-
hira . ie ene RG 13205, March, 1940, alt. 200-300 m., on grassy hill (plant
1m. ; flower pink).

ae slant nen very well with a number of Queensland collections
placed in this species. Keraudrenia lanceolata has been reported once
from northern New Guinea, but this seems to be the first oa of this
species from hes sland.

PASSIFLORACEAE
Passiflora Linnaeus
Passiflora moluccana Blume Bijdr. Fl. Nederl. Ind. 938. 1826; DC. Prodr. 3: 323
1828; Rumphia 1: 169. ¢. 32. 1835; Koorders Seen Java 2: 638. 1912;
vel aff.
SOLOMON ISLANDS: Bougainville: Kugumaru, Buin, Kajewski 1913,
July 1930, alt. 150 m., rain forest (common vine on rain forest trees; fruit green,

4.2 cm. long, 4 cm. dia ); Koniguru, Buin, Kajewski 2076, October 1930, alt. 950
m., rain forest (vine; fruit globular, 3.9 cm. diameter, slightly flattened top and
bottom). Ysabel: Tasia, Brass 3283, Dec. 1932, coastal rain forests (large climber ;

leaves stiff, with pale nerves, glands pale brown; immature fruit about 2.5 cm
diameter, een):

These specimens differ from the description and plate of this species in
having somewhat larger leaves (12-17 cm. long, 5.5-8.5 cm. wide) and
globose fruits. Since we have no flowering material it seems best to place
it here for the present. Passiflora moluccana Bl. is a native of the
Moluccas.

Hollrungia K. Schumann
Hollrungia aurantioides K. Schum, Bot. pena 7 — 1888; Merr. & Perry, Jour.
Arnold Arb. 24: 210. 1943, op. cit. 29:

BRITISH NEW GUINEA: Central ace ae ulu, Brass 5239, Sept.-Nov. 1933,
alt. 1250 m., on edge of forest ae only one example seen (large woody climber ;
leaves smooth and shining; flowers yellow-green; fruit immature).

Our duplicate specimen ca lost all its flowers, but it was finally named
with the aid of some flowers from the first set in New York. Now we
have one or two collections from each political division of New Guinea,
indicating the plant is probably wide-spread over the island, but appar-
ently not plentiful anywhere.

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVII 45

GENTIANACEAE
Nymphoides Hill.
pee hanes hydrocharoides (F. v. Muell.) O. Ktze. Rev. Gen. 2: 429. 1891 (as
hodes hydrocharodes).
Villarsia hydrocharoides F. v. Muell. Fragm. Phytogr. Austr. 6: 139. 1868.
seals hydrocharoides F. v. Muell. ex Bentham Fl, Austr. 4: 380. 1868 ;
. M. Bailey, Queensl. Fl. 3: 1030. 1900

BRITISH NEW GUINEA: Wuroi, Ori riomo , Brass 5871, Jan.-Mar. 1934,
alt. 20 m., plentiful in small pools of a little are on Men Geaces reddish oe
neath; Rewer orange-colored).

Through the co-operation of Mr. C. T. White of Brisbane, a duplicate
of Brass 5871 was sent to Melbourne, where it was checked with the type
of Mueller’s species by Mr. J. H. Willis, who found it to match very well
indeed. This appears to be the first record of this Australian species. from
New Guinea. Its most striking character is found in the comparatively
large and minutely pubescent seeds.

Nymphoides parvifolium (Griseb.) O. Ktze. Rev. Gen. 2: 429, 1891 (as Nymphodes
parvifolium).

Linmanthem ps aie Griseb. in DC. Prodr. 9: 141. 1845; F. M. Bailey,

Queens. Fl. 3: . 1900.

pn ee NEW ee Mabaduan, Brass 6534, April 1936, common in shallow
grassy rain pools in savannah forests (very small white flowers). Previously reported
from India, Siam, Ceylon, the Malay Peninsula, and Queensland.

SOLANACEAE

In the Solanaceae, as in so many other instances, we have very little
authentic material for comparison. Many of our collections differed so
widely from the descriptions of the species already reported for that
island that we have described them as new. Following the latest work
on the Solanaceae by Mr. C. V. Morton, Contrib. U. S. National Her-
barium, 29: 54, 55. 1944, we have treated Lycianthes as a section of the
genus Solanum.

Solanum Re sp. nov,

Frutex + 1 m. altus; ramulis teretibus, 2—2.5 mm. diametro, eee
Silene Be eatini. sparsim aculeatis, aculeis tenuissimis, usque :
longis basim versus 0.5 mm. diametro, internodiis 2.5—3 cm. longis; foliig
alternis membranaceis, lanceolatis, 9— 15 cm. longis, 2.5—4.5 cm. latis, apice
acuminatis, acumine + 2 cm. longo, basi interdum inconspicue inaequilat-
eralibus interdum valde obliquis, anguste cuneatis, margine subintegris vel
undulatis vel leviter angulatis, nonnumquam non profunde lobatis, utrinque

atis) mm. longis; calyce cupulari, 1 mm. longo, 5-dentato, dentibus

46 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

minutis; corolla 6 mm. longa, tubo 1 mm. longo, lobis lanceolatis, intus
glabris; filamentis brevissimis, antheris 4 mm. longis, anguste lanceolatis,
apicem versus angustatis, poris apicalibus minimis; ovario ovoideo, 1 mm
longo, stylo gracili, 4 mm. longo; fructibus ovoideis, 1.8 cm. longis, 1.5
cm. diametro, seminibus + 20, oblique reniformibus, 6 x 4.5 0.6 mm.,
minute reticulatis.

BRITISH NEW GUINEA: Central Division, Ononge Road, eek Brass 3814
(type), April 1933, alt. 500 m., in rain forest (one plant seen; shrub 1 with ter-
ius al flat- spreading branches; flowers pink; fruit broadly ovoid, evanke: -yellow,

1 . long, 1.5 cm. diameter).

“he addition to the above cited collection, we have a specimen from Thu,
Vailala River, Brass 972, which agrees very closely with the type except
that the calyx has very definite lobes and is about 3 mm. long, the fruit
is very much like that described above in color and size of seeds, the
plant is described as a weed in rain forest clearings. Unfortunately there
are not very many flowers on any of the specimens, in fact only about a
single open one on each, and several buds on the type, hence it is diffi-
cult to estimate the amount of variation in the character of the calyx.
Solanum oligolobum sp. nov.

Frutex magnus; ramulis teretibus 3-6 mm. diametro, aculeatis, primum
stellato-tomentosis demum paullo glabratis, pilis sessilibus vel stipitatis
apice stellatis grossis, flavidis, aculeis + crebris, rectis, 5-10 mm. longis
basi 1-2 mm. latis, internodiis 2.5-5.5 cm. longis; foliis alternis vel
superioribus plerumque geminatis vel parum inter se distantibus, paullo
inaequalibus, tenuiter chartaceis fragilibus, ambitu lanceolatis, apice acutis
vel sensim acuminatis, basi inaequilateralibus, obtusis, margine grosse
3—4-lobatis (lobis obtusis vel acutis interdum mucronatis, sinubus
plerumque late subrotundatis in foliis minorum fere planis), utrinque costa
et venis primariis (utrinsecus - 6 inconspicuis) aculeatis (aculeis paucis
usque 1.5 cm. longis), supra + dense tomentosis (pilis stellatis stipitatis
sessilibusque) interdum consperse glandularibus, atro-olivaceis, subtus
densé tomentosis et glandularibus paullo pallidioribus, lamina majorum
11-21 cm. longa, 4-13 cm. lata, petiolo 3-5 cm. longo, tomentoso et
aculeato (aculeis usque 2.2 cm. longis), minorum 5-13 cm. longa,
2-5.5 cm. lata, petiolo 1.5-2 cm. longo; inflorescentiis a tomentosis,
primum fere terminalibus, serius in n latus coactis, + 12-flor - pedunculo
2-10 mm. longo, rhachibus simplicibus vel furcatis, pedicel 1 (-1.5
in fructu) cm. longis interdum aculeatis, aculeis paucis usque 4

in dentes lineari-subulatos vix 1.5 mm. longos abrupte angustatis; corolla
violacea, rostrata, tubo + 2 mm. longo, lobis 8-9 mm. longis, extus
margine lato excepto stellato-tomentosis; staminibus 5 (interdum 6),
circiter 1 mm. corollae basim supra insertis, filamentis vix 2 mm. longis,
antheris 3-3.5 mm. longis, apicem versus angustatis; ovario ovoideo, 1.5
mm. longo, apicem versus glandulis paucis breviter stipitatis instructis,
stylo + 4 mm. longo; fructibus + 1 cm. diametro, glabris, seminibus
numerosis, 2.5 xX 2 * 0.4 mm., minute reticulatis.

NETHERLANDS NEW GUINE i. 9 km. NE. of Lake Hab a, Brass 10876
(TYPE), Oct. 1938, alt. 2650 m., sunny situation at base of aa Gaie shrub with
violet flowers).

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 47

In the outline of the leaves and the straight prickles this species sug-
gests S. Gibbsiae J. R. Drummond; in S. oligolobum, however, the leaves
are very much larger, the prickles fewer, the tomentum denser and; particu-
larly on the lower leaf-surface, mixed with stipitate glands, the inflo-
rescence compound, and the fruit has many more seeds. The stalked
coarse hairs forming a thick and loose tomentum give the impression of a
thicker leaf than where the hairs are sessile (and usually finer); at the
same time both leaves may be similar in texture and thickness.

Solanum trichostylum sp. nov.

Frutex usque 2 m. altus; ramis sensim glabratis, + 4 mm. diametro;
ramulis novellis dense tomentosis (pilis stellatis, ate sessilibus), aculeo-
ee (aculeolis vix 2 mm. longis, basi 2 mm. latis, lateraliter compressis) ,

3 : :

ar)
-
io’)
ao!
=“
ar
= 5
=
a)
Lay 3
=
jo)
S
7
i
wn
E
—
)
=)
gg
—
1¢2]
=r
o)
=
—_
wn
=
ct
oO
+t
=
n
ch
5
oO

ae ome cm. _longis, 1.5-4 cm. eae ovatis, apice acuminatis vel

leatis (aculeis remotis paucisque 2-6 mm. longis, rectis) vel fere inermibus,
supra olivaceis et sparsim, subtus subflavidis et dense stellato-tomentosis,

rete supra impresso subtus prominulente; petiolo 1-2 cm. longo, dense
stellato-tomentoso interdum 1—3-aculeato; inflorescentiis 6—10-floris, dense
stellato-tomentosis in sicco subflavidis, primum fere terminalibus serius in
latus coactis; pedunculo 5-10 mm. longo, rhachi brevi, pedicellis + 1 cm.
longis; calyce 4 mm. longo, subcampanulato, 5-lobato, lobis vix 2 m
longis, subtruncatis, apiculatis, apiculo 0.5 mm. longo; corolla 5- eto:
circiter 1 cm. longo, lobis late ovatis apice leviter cucullatis, utrinque
stellato-pubescentibus margine tantum glabris; staminibus supra corollae
basim 1.5 mm. insertis, filamentis 2 mm. longis, antheris lanceolatis, basi
2 mm. apice 0.6 mm. latis, poris apicalibus parvis extrorsis; stylo 7 mm.
longo, in dimidio inferiore stellato- pubescente; ovario subgloboso 1.5 mm.
diametro, glabro; fructibus + 1 cm. diametro, apiculatis, seminibus
oblique subreniformibus, 4 xX 3.5 & 0.5 mm., margine minute reticulatis.
BRITISH NEW GUINEA: Central Division, Mount Tafa, Brass 4934 oe

damaged by fire (sparsely branched shrub 1-1.5 m. high; leaves pale; corolla purple,
anthers bright yellow; soft, globose, black fruit about 1 cm. diameter).
Solanum torvoideum sp. nov.

Frutex altus vel arbor parva; ramulis teretibus + 5 mm. diametro
novellis densissime stellato-tomentosis deinde glabratis, pilis stipitatis vel
sessilibus grossis fale aculeis paucis, vix 1 mm. longis, inconspicuis,
internodiis 2.5—5.5 . longis; foliis inermibus alternis vel interdum
superioribus gemin ae ne minore), fragilibus, tenuiter chartaceis ambitu
lanceolatis vel lanceolato-ellipticis (5.5 —)7—16.5 cm. longis, (1.5 —)3-9

tosis, pilis saepissime stipitatis, venis primariis utrinsecus + 5, supra

48 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XXX

aaron subtus latioribus et prominulis, oblique adscendentibus mar-
ginem versus arcuatis, rete laxo tantum subtus in foliis soliiae es
een petiolo 1.2-2.5 cm. longo densissime stellato-tomentos
florescentiis (in fructu 7-8 cm. longis) primum fere ee a Satie serius
in latus coactis, + 20-floris, dense stellato-tomentosis, serius glabratis;
pedunculo 7-15 mm. longo, furcato, rhachibus et pedicellis (1-2 cm.
longis) etiam -+ glandulosis; calyce campanulato, tubo 2 mm. longo,
lobis 4 mm. longis, lanceolatis apice 2 mm. subulatis, corolla + 10 mm.
longa, lobis intus apice tantum pubescentibus leviter cucullatis vel apicu-
latis; staminibus 1 mm. supra corollae basim insertis, filamentis 1.5 mm.
longis, antheris 4 mm. longis, lanceolatis sursum angustatis; ovario sub-
globoso minute glanduloso, stylo 6 mm. longo; fructibus globosis, + 1 cm.
diametro glabris; seminibus numerosis, 2 * 1.8 0.5 mm., ovalibus,
fere levibus.

BRITISH NEW GUINEA: Central Division, Mafulu, Brass 5411 (typE), Oct.
1933, alt. 1250 m., forest regrowths (tall bush or small tree; very few small prickles
n branches; flowers white; fruit orange-brown

This collection differs from what we take to be typical S. toruum Sw.
in the coarse thick tomentum on both surfaces of the leaves which gives
the plant a yellowish brown appearance when dry rather than pale or
cinereous green which is characteristic of typical S. torvuwm Sw.; the leaves
also are only slightly inequilateral at the base, and the pedicels are more
robust; in the specimens at hand there are also many less flowers in the
flower-cluster.
Solanum inaequilaterale Merr. Philip. Jour. Sci. 1, Suppl.: 236. 1906; Enum. Philip.

]. 3: 426, 1923,

NETHERLANDS NEW GUINEA: Bele River, 18 km. NE. of Lake Habbema,
Brass 11590, Nov. 1938, alt. 2200 m., occasional in open grassy second growths
(spreading tree 1.5-2 m. high; flowers levenders fruit yellow).

This specimen is a very good match for the Philippine material of this
species, and both come from the same altitude.
Solanum Schefferi F. v. Mueller, Descr. Notes Papuan Pl. 1: 44.

Solanum incanum Scheffer, Ane. Jard. Bot. Buitenz. 1: 39. nS non Linn.

Solanum smilacocladum Bitter, Bot. Jahrb. 55: 79. 1917.

In citing the Buitenzorg Herbarium specimen no. 9853! (sub. nom. S.
incanum Scheff.) we believe that Bitter was citing the type of Scheffer’s
Solanum incanum, hence the first valid name for the species is S. Scheffert
F. v. Muell., rather than S. smilacocladum Bitter, even though the de-
scription is emended.

Solanum heteracanthum sp. ncv.

Planta suffruticosa -- 1.5 m. alta; ramis aculeatis et fere glabris, ramulis
primum pilis stellatis dense tomentosis demum glabratis, aculeati is, aculeis
5-7 mm. longis, ad basin 5 mm. attingentibus, recurvis, a latere com-
pressis; foliis alternis vel saepe ad nodos binis et inaequalibus, 5-12.5 cm
longis, 1.8—5 cm. latis, margine angulato- -sinuatis vel interdum subintegris,

acuminatis, supra maturis in sicco atro-olivaceis glabris, novellis costa
et venisque stellato-pubescentibus, subtus fulvo-cinerascentibus dense
stellato-tomentosis, costa venisque consperse aculeatis, aculeis paucis rectis

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 49

vel leviter recurvis usque 9 mm. longis ad basin 1-1.5 mm. latis; petiolo
7-9 mm. longo, tomentoso; inflorescentiis extra-axillaribus usque 14-floris,
pedunculo + 5 mm. longo, dean thachibus 2-5 mm. longis, pedicellis
confertis, 7-10(—17 in fructu) mm. longis apicem versus incrassatis et
angulatis, sicut pedunculo et eae stellato- -pubescentibus; alabastris
dense stellato-tomentosis; calyce campanulato, tubo vix 2 mm. longo, lobis
mm. longis, abrupte obtusis apiculatis; corolla rotata, + 11 mm. longa,
5-lobata, lobis lanceolatis extus dense stellato-tomentosis in dimidio in-
feriore membranis glabris conjunctis et marginatis, intus in parte ee
venae mediae interdum pilis stellatis obsitis; staminibus 5, + 1
supra corollae basin insertis, filamentis 2 mm. longis glabris, Ae
anguste lanceolatis 6 mm. longis basi 1 mm. latis, fere a basi subcordata

cm. diametro, seminibus + 16 reniformibus rca compressis, 4 x 3
mm., margine 0.6 mm. circumdatis, minute reticulatis.

NETHERLANDS NEW GUINEA: 9 km. NE. of La a Habbema, Brass mone
(TYPE), Oct. 1938, alt. 2700 m,, rain ae of valley bottom (weak shru
high, in a native clearing; flowers purple; fruit unripe); Bele River, 18 km. NE. ‘of
Lake Habbema, Brass 11505, Nov. 1938, alt. 2200 m., scrambling on an open land-
slip (flowers purple; fruit immature).

Among the descriptions of the New Guinean species of Solanum this is
perhaps related to S. Gibbsiae J. R. Drummond; but in the latter the
spines are straight, and the inflorescence 1—3-flowered. Among the species
which we have seen, it is most like S. dimorphispinum C. T. White of
Queensland. The latter, however, is larger in most dimensions, the leaves
are only repand, the inflorescence is much more open, and the pubescence
more compact. Our species seems best characterized by the strongly
recurved prickles with broad bases on the branchlets, the angled sinuate
margins of the leaves, the longer and practically straight prickles with
narrow bases on the leaves (on the midrib and main primary veins of the
upper surface, usually confined to the midrib on the lower surface), and
the rather compact inflorescence; whether the thin margin of the seed is
a good diagnostic character or only on account of its immaturity we cannot
say.

Solanum acuminatissimum sp. no § LYCIANTHES

Frutex 60 cm. alta; ent sparsim ee vel puberulis, internodiis
2-3.5 cm. longis; oliis gem ie inaequalibus, tenuiter chartaceis;
majoribus 6.5-16 cm. longis, 2.8-7 cm. mice ovato- ce prea te
angustatis, basi oblique uaa apice longe acuminatis, acumine 1.2-3

lato, lineari-lanceolato + falcato, margine ere
vel subintegris, utrinque abe supra olivaceis, subtus pallidioribus,
venis lateralibus utrinsecus + 6 oblique adscendentibus prope marginem
arcuatis, utrinque perspicuis, rete laxo + manifesto; petiolo sparsim
pilosulo vel puberulo, 1-1.3 cm. longo; foliis minoribus 2—5.5 cm. longis,
1.3-3.5 cm. latis, petiolo pach mm. longo; inflorescentiis axillaribus 2—3-
floris, glabris, pedicellis 2—2.5 cm. longis; calyce (cupulari) in fructu
patenti, truncato, infra marginem dentibus 3—5 obtusis, 0.5—-1 mm. longis,
oriundis; baccis globosis, circiter 1 cm. diametro; seminibus s atis nu-
merosis, reniformibus, in sicco vix 3 K 3 X 1 mm., distincte reticulatis.

50 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

NETHERLANDS NEW GUINEA: 15 km. SW. of Bernhard Camp, Idenburg
River, Brass 12290 (type), Jan. 1939, alt. 1800 m., one example on a small clearing
‘in mossy forest (shrub 60 cm. high; fruit green; pedicels: calyx, and lower surface
of leaves tinged with ao.

This species is perhaps best distinguished by the thinly chartaceous +
sinuate-margined larger leaves with a long slender somewhat falcate
acumen, the few-flowered inflorescence, and the thin calyx. It is very
much like Solanum banahaense Elmer in that both have calyces with
minute teeth (or lobes) developing just below the margin. In the Philip-
pine material (which also has entire leaves of firmer texture) this charac-
ter appears to be fairly constant, the teeth originating from 4 angles or
ribs; in the New Guinean collection the character is variable, sometimes
only 3 teeth are developed, and sometimes 5
Solanum multifolium sp. nov. § LYCIANTHES

Frutex gracilis, 2-3 m. altus; ramulis 1-2 mm. diametro, oe
ee aaaman es pilis brevibus apice incurvis, internodiis 0.7—3 cm. longis;
foliis membranaceis, plerumque geminatis interdum solitariis; permite
(1. i )4—9 cm. longis, 1.5-3.7 cm. latis, lanceolatis, utrinque angustatis,
basi cuneatis inconspicue obliquis, apice acutis interdum obtusiusculis vel
obtusis, margine integris, supra glabris, subtus costa et venis tantum pilis

sicut ramulis adscendenti-pubescente; foliis minoribus a “a 5 cm. longis,
0.4-0.8 cm. latis, lanceolato-ovatis vel suborbicularibus, obtusis; petiolo
brevissimo vel usque 2.5 mm. longo; floribus solitariis vel seminatis,
glabris, pedicellis 0.7—-1.8 cm. longis, alabastro uno tantum viso, 5 mm.
longo; calyce cupulari, 2.5 mm. longo, apice 3 mm. diametro; corollae
lobis 5, apice leviter ee et puberulis: filamentis brevibus, antheris
3 mm. ‘longis, lineari- mee pistillo minuto; fructibus globosis, 7]
mm. diametro, seminibus + 15, 2.5-3 mm. longis latisque, 1.5 mm. crassis.
NETHERLANDS NEW GUINEA: 6 km. SW. of Bernhard Camp, Idenburg
River, Brass 12907 (TYPE), Feb. 1939, alt. 1150 m., common in a rain forest gully
(very slender tree 2-3 m. high; fruit red; one white. flower-bud).
A species easily recognized by its relatively small and very thin leaves,
short internodes, and mostly solitary flowers.
Solanum belense sp. nov. § LYCIANTHES
Frutex parvus; ramis cinereis, glabris, ramulis novellis dense puberulis
demum glabratis, 1.5—-3 mm. diametro, internodiis 1.5—3 cm. longis; foliis
tenuiter chartaceis, plerumque pee Sa interdum solitariis,
majoribus 45-155 cm. longis, 2.5—5.5 cm. latis, late lanceolatis, apice
tuse vel acute acuminatis, basi oblique cuneatis, margine integris,
utrinque glabris, novellis subtus costa venisque puberulis mox glabratis,
venis primariis utrinsecus 6-8 supra manifestis, subtus prominulis, ad-
scendentibus, marginem versus arcuatis, - laxo supra inconspicuo sub-
tus distincte manifesto; petiolo 0 051.5 m. longo, + puberulo; foliis
minoribus 1—3.5 cm. longis, 0.5—-1.8 cm. "lati lanceolatis vel elliptico-
lanceolatis, basi plerumque obliquis, apice obtusis vel acutis, glabris, venis
primariis + 4; petiolo 3-5 mm. longo; inflorescentiis inter bases petiolorum
inse: tis, glabris; floribus solitariis vel 2—6-fasciculatis, fasciculis interdum

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII ol

pedunculo Peles vel pulvillo insertis, pedicellis usque 2 cm. longis;
calyce cupulari, 3 mm. longo, apice 3 mm. diametro; corollae tubo 5 mm.
longo, lobis s ‘1 em. lon ngis, 4 mm. latis, acutis, extus apice cucullato et
margine dense puerta staminibus tubi corollae apicem prope insertis,
filamentis 1.5-2 mm. longis, antheris 3 mm. longis, oblongis apice tantum
paullo ennoas ovario ovoideo, 1.5 mm. longo, stylo 7 mm. longo;
fructu non v

Se NEW GUINEA: Bele River, 18 km. NE. of Lake Habbema,
Brass 11223 (tyre), Nov. 1938, alt. 2300 m., Fagaceae forest, common in moist
semi-shade (small shrub; pedicel and calyx vines corolla white).

Only one branch seems to have any perfect flowers, the rest are all
staminate; in the latter the filaments are a little longer than in the perfect
ones.

Solanum acuminatissimum, S. belense, and S. multifolium are all closely
related. The first appears to be easily separable on its somewhat undulate,
long-acuminate leaves, and slightly dentate calyx, the latter character is
not present in either of the other species. For the most part, S. belense
has more pubescent branchlets, entire, narrower and more shortly acumi-
nate leaves, and more flowers in a fascicle. These are fastened to a slight
swelling or a very short peduncle situated between the bases of the petioles
of each pair of leaves. Solanum multifolium has very short internodes,
and on the average much smaller leaves than the other two. All three
species have thin leaves and calyces, but in the last species the leaves are
seemingly more translucent than in the other two.

Solanum rostellatum sp. nov. § LYCIANTHES

Planta suffruticosa, + 1 m. alta; ramulis teretibus, glabratis, 1-3 m
diametro, novellis dense hirtellis, pilis simplicibus pluri-cellularibus, ee
nodiis t= 3 cm. longis; foliis saepe majoribus cum foliis minutis geminatis,
mee 4— 0 cm. longis, 1-3.5 cm. latis, chartaceis, late oblongis vel
lanceolatis, basi -+ oblique eee ae obtusis, ee subabrupte acumi-
natis vel rostellatis, acumine 8-1 m. longo i 4-5 mm. lato, integris,
in sicco atro- olivaceis, supra ares subtus et, et venis primariis -
hirtellis, venis primariis utrinsecus 5—6 curvatim adscendentibus prope
marginem arcuatim conjunctis, rete utrinque manifesto; petiolo 5-9 mm.
longo, + hirtello; foliis minutis 4-8 mm. longis, 2—4 mm. latis, suborbicu-
laribus, subsessilibus; inflorescentiis axillaribus, 1—2-floris, pedicellis usque
2 cm. longis, sparsim hirtellis vel fere glabris; calyce cupulari, truncato,
2 mm. longo, apice 3 mm. diametro, consperse pilosulo vel glabro; corolla
glabra, tubo 2 mm. longo, lobis 5, + 5 mm. longis, 1.2—-1.5 mm. latis, apice
subacutis, margine (in ae minutissime papillatis: filamentis latis
et brevissimis, sub apice tubi corollae insertis, antheris lineari-lanceolatis,
4 mm. longis, basi 1 mm., apice 0.5 mm. latis, poris apicalibus parvis,
ovario ovoideo, stylo 5 mm. longo; baccis non visis.

BRITISH NEW GUINEA: Central Division, East Mt. Tafa, Fie i 35 (TYPE),
May 1933, alt. 2100 m., common on roadside in tall foothill forest (wea rub about
1 m. high; leaves dull aon on purple petioles; corolla purple, anthers act yellow).

This species might be compared to Solanum cladotrichum Bitt., but the
latter is a high-climbing liane with gradually acuminate leaves, consider-
able pubescence and small flowers; our species is a small shrub with some-

52 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

what abruptly acuminate leaves, the pubescence (unbranched hairs) is

scanty occurring chiefly on the younger branchlets, the midrib and veins

of the lower surface of the leaves, and very few hairs on the pedicels and

calyx. The flowers are larger than those described for Bitter’s species.

Solanum Rechingeri Witasek, Rep. Nov. Sp. 5: 165. 1908, ee eae Math.-Nat.
Kl. Akad. Wiss. Wien 89: 602. 1913; Bitter, Bot. Jahrb. 102. 1917.

Lycianthes Rechingeri (Witasek) Bitter, Abhandl. Nat. Ver. eae on 504. 1920.

SOLOMON ISLANDS: Bougainville: Kugimaru, Buin, Kajewski 1800,
June 1930, alt. 150 m., rain forest, common (small tree with white trunk; petioles
of leaves blue, darker near the base; petals blue, stamens yellow); same locality,
Kajewski 1863 (fruit 1.4 cm. long, 1.2 cm. ene: ue green-blue when ripe) ;
Siwai, Waterhouse 66 (small tree). Guadalcanal: Berande River, Kajewski
2388, June 1930, rain forest (small tree up to 10 m. high: purple petals, yellow sta-
mens). Ysabel: Tiratona, Brass 3314, Dec. 1932, alt. 600 m., rain forests, common
(slender tree up to 10 m. tall; pale brown bark; small bright red fruits).

This species is closely related to S. vitiense Seem. Just about the time
of anthesis the flower-bud is 8 mm. long, and 3.5 mm. diameter (in the
dried specimen); the calyx is cupular, 2 mm. high, the margin + tufted
puberulous-ciliolate, as is also the margin of the corolla-lobes; in the full-
blown flower the corolla- lobes are about 8 mm. long, the tube 4 mm.; the
filaments are 2.5 mm. long, the anthers 3 mm. long; the style 2 mm. long.

n S. vitiense Seem. the flower-buds are broader in proportion to the
length, and the inflorescence does not show any tendency to become
racemose in fruit, as is the case in S. Rechingeri Witasek. In the latter
species the leaf-bases are both symmetrical and unsymmetrical in the
specimens cited.

Solanum impar Warburg, Bot. Jahrb. 13: 415. 1891; Bitter, Bot. Jahrb. 55: 109.
1917

Lycianthes impar (Warb.) Bitter, Abhandl. Nat. Ver. Bremen 24: 504. 1920.

BRITISH NEW GUINEA: Fly River, 528 mile Camp, Brass 6796, May 1936,
alt. 80 m., climbing on undergrowth trees in the river flood-bank forests (attractive
purple ae and blue fruit in axillary fascicles, fruiting calyx white

This specimen seems to fit the description of Warburg’s re reason-
ably well; the leaves are chartaceous and vary a little more in size (15—
32.5 cm. X 5-8 cm.), the peduncle does not exceed the length of a
centimeter, and the berries are oblong or slightly ovoid (7-8 mm. x 5
mm he c calyx is cup-shaped, truncate or slightly undulate, 2-3 mm.
high: the corolla tube is about the same length, the lobes 5(—6), about 3.5

mm. long, 1 mm. broad, apex slightly cucullate and very slightly puberu-
lous; filaments inserted near the top of the corolla-tube, 1 mm. long, gla-
brous, anthers oblong, 2.5 mm. long; pistil sometimes minute (as described
in S. Ledermannii Bitter) or aborted, or again normal with the style pro-
truding beyond the stamens about 1.5 mm

PLANTAGINACEAE

Plantago Linnaeus
In van Steenis’s article, ‘On the origin of the Malaysian mountain flora,”
Bull. Jard. Bot. Buitenz. III. 13: 235. 1934, we find concerning Plantago

1949 | MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 53

that ‘no species seems to be indigenous in Malaysia.” Since then four
species have been collected in the high mountains of New Guinea during
the Archbold Expeditions. The first, from British New Guinea, reported
by Pilger in 1935, was originally described from New Zealand. Now this
same species (at least i in a broad interpretation) and three others also with
southern affinities are reported here.
ae eee sp. nov. § MESEMBRYNIA
t videtur perennis, acaulis vel caudex brevis crassiusculus;
tollis ee usque 13 cm. longis (parte superiore 3— ui cm
cm. lata), deorsum sensim Bree is in petiolum 2—5 c ae ongum, 1.5—2
mm. latum, crispe pilosum, transeuntibus, basi paulo dilatatis ibique lana
longa fulva involutis, apicem versus angustatis, apice ipso obtusiusculis,

obtusis, obtuse carinatis, glabris, margine sparsim ciliatis, intus basi +
pilosis; sepalis 3 mm. lon aoe posticis ellipticis, valde concavis, apice vix

angustatis, carina crassiuscula prominula, anticis paulo angustioribus,
parum inaequilateralibus, ee vel er m oar: et consperse cilio-
latis; corollae tubo 2 mm. longo, lobis 1.5 mm. longis, rotundato-ovatis;

capsulis ellipsoideis, apice ad basim styli cere circiter 4 mm. longis
(parte superiore 2.5-3 mm., inferiore 1.5 mm.), seminibus 4—5, fuscis,
ambitu ovato-oblongis, crassis, fere 2 mm. longis, subtiliter punctatis,
facie hili convexis.

BRITISH NEW GUINEA: Mount Albert Edward, Brass 4352 (TYPE), June
1933, alt. 3680 m., common; wet slopes of alpine grasslands (leaves flat, spreading,
concave, margins red).

This plant belongs to the section MESEMBRYNIA Decaisne. From the
descriptions at hand we are unable to suggest a closely related species.
Plantago ate Hook. f. Handb. New Zeal. Fl. 227. 1864; Pilger, Bot. Jahrb.

67: 236. 1935; Pflanzenr. eG 269): 120. 1937.

NETHERLANDS NEW GUINEA: Lake Habbema, Brass 9457, Aug. 1938, alt.

3225 m., filling the spaces between scattered grass-tussocks on marshy slopes; 11 km

east of Wilhelmina-top, Brass & ees ee 10220, Sept. 1938, alt. 3800 m., alpine
grassland, scattered along banks of a
In reporting Brass 4646, Pilger tea out that it varied somewhat from
the New Zealand material. These collections vary even more, particularly
in size. The plants are 4.5—-16.5 cm. in diameter with leaves 1.5—8 cm.
long, and 0.4-1.8 cm. broad; the pubescence of the leaves is variable, in
some ashe appearing only on the lower surface. We have considered
nigera var. Petriet Cheeseman for these collections, but the leaves
are even larger than described there, and numerous. In the collection
from Lake Habbema the peduncles are glabrescent, sometimes having only
a few scattered hairs.

54 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Plantago stenophylla sp. nov. § OLIGANTHOS

Planta parva, acaulis; rhizomate valde verticali, crasso vel crassiusculo,
radice primaria non visa (probabiliter demum nulla), radicellis robustis,
elongatis (usque ad 8 cm. longis); foliis numerosis, rosulatis, sub- erectis
vel adscendentibus, anguste linearibus, 1.5—4 cm. lon ngis, 0.7— 2 mm. latis,
integris vel remote et minute dentatis, margine + revolutis, apice obtusis,
glabris, crassiusculis, supra costa impressa, subtus prominula, basi paullo
dilatatis ibique lana longa flavescente instructis; inflorescentiis axillaribus,
primum breviter (2 mm.), demum longe (2 cm.) pedunculatis, pedunculo
glabro; spicis plerumque 1-floris; bracteis + 1 mm. longis, rotundatis,
concavis, basi intus longiuscule pilosis: sepalis oblongo-lanceolatis, 2—2. 5

lobis lanceolatis, 1.2 mm. longis; staminibus 4, antheris non visis; capsulis
ellipsoideis, 8—10-spermis, circiter medio circumscissis; seminibus fuscis,
3—4-angulatis

NETHERLANDS NEW GUINEA: Lake Habbema, Brass 9456 oe Aug.
1938, alt. 3225 m., alpine grassland (common bog herb); 7 km. northeast of Wil-
eee ig Brass & Mevyer-Drees 9920, Sept. 1938, alt. 3720 m., abundant in alpine
bog t

a eeeeren NEW GUINEA: Mt. Sarawaket, Clemens, sn. May 1939, alt.
about 3300 m., forming sod around pond margins with Potentilla, Gentiana and
Carex (flower wine-purple, minute).

Plantago depauperata sp. nov. § OLIGANTHOS
Planta pulvinata; rosulis parvis, 1-3 cm. diametro, be ae aggre-
gatis ; foliis crassiusculis, 6-15 mm. longis, 1-1.5 mm. latis, linearibus,

apicem versus angustatis obtusiusculis, basi sensim dilatatis ibique lana
fulva involutis, ceterum glabris, integris, margine angustissimo

latis, 1 ngis, + reflexis; antheris longe exsertis; stigmate longe
exserto; capsulae parte . inferiore valde elongata, + 5 mm. longa, anguste
obconica, arte superiore conica, -- 2 mm. longa; seminibus 4-8, parvis,

angulatis, capuliter impresso- punctatis.

NETHERLANDS NEW GUINEA: Lake Habbema, Brass 9199 (typPE), Aug.
1938, alt. 3225 m., plentiful in open boggy ground (gregarious, each plant forming a
separate tuft + 1 cm. high; leaves blue-gray); 7 km. northeast of Wilhelmina-top,
Brass & Meyer-Drees 9921, Sept. 1938, alt. 3800 m., gray-green rosettes forming the
bulk of the sparse herbaceous cover on shallow soil of sandstone summits; northern
slopes of Mount Wilhelmina, Brass & Meyer-Drees 10135, Sept. 1938, a 4150 m.,
often the chief plant on black sandstone heights.

Both this and the preceding species belong to section OLIGANTHOS
Barneoud. In P. depauperata the capsule seems to answer the description
of that of P. barbata var. monanthos (D’Urv.) Pilger; however, on com-
parison with scanty herbarium material from the Falkland Islands, we
find the plants of the latter much coarser and of much more open habit
than the New Guinean material. Whether the Falkland Islands material

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIIE 55

is properly named we cannot say, but certainly the Archbold collections are
not conspecific with it.

RUBIACEAE
Maschalodesme simplex Merr. & Perry, Jour. Arnold Arb. 25: 197. 1944.

BRITISH NEW GUINEA: Vailala River, Thu, Brass 914, Feb. 1926, rain forest
(erect unbranched bush 5 feet high; leaves shining above; large ae stipules; inflo-
rescence in supra-axillary clusters; fruit red, eight-angled).

At the time this species was described this specimen was unavailable for
examination. Since the genus has not been reported in fruit before, it
seems worth while to add a brief note. Two fruits are available, one of
which has been opened. It contains 16 seeds, eight to a locule, biseriately
arranged and imbricating upward. Dry fruit elliptic, 8-costate, 3-3.5 cm.
long, 1.5-2 cm. diameter, minutely puberulous. Seed after being soaked
irregularly subobtrigonous or subrhomboid, more or less angled, somewhat
flattened and curved, hardly 1 cm. long, about 5 mm. wide at its greatest
width, attached at the inner base and here somewhat pointed; radicle
superior scarcely 3 mm. long, cotyledons ovate, about 2 x 1.5 mm., em-
bryo straight; albumen thick, somewhat cartilaginous.

VALERIANACEAE

Duce repens Hemsl. Kew Bull. 1899: 101. 1899; ae os Jahrb. 62: 493.
929; van Steenis, Bull. Jard. Bot. Buitenz. III. 13:

Medea NEW GUINEA: Lake Habbema, coe ee Aug. 1938, alt.
3225 m., locally common on open boggy ground (flowers white).

NORTHEAST NEW GUINEA: Sarawaket, Clemens 6325, April 1937, alt. 1800—
2400 m.; Morobe District, Upper Camp A, Clemens 10085bis, Mar. 1939; Ulap Trail,
Clemens 41133, April 1940 (flowers white).

The two previous records of the New Guinean species of Triplostegia
were based on fragmentary material. Hemsley described the plant as
having repent stems, but the radical leaves were not seen. Diels indi-
cated that, although in the plant from Sarawaket the stem was upright and
much shorter than in the original which he had examined, he was of the
opinion that it belonged to the same species. In the light of the material
cited above we believe it is clear that the habit of the plant is closely
adapted to its environmental conditions. If the roots are fairly close
to the surface of the substratum, the basal part of the shoot is very short,
and a pseudo-rosette of leaves develops at the surface. On the other hand,
if the roots are deeply buried in the substratum the basal part of the shoot
is longer depending on the distance it has to travel before reaching light
and developing leaves. Possibly if the plant is in the shade longer inter-
nodes develop, the longest in the specimens at hand is 1.5 cm., but mostly
they are very short (2-5 mm.); above the cluster of leaves the stem be-
comes scape-like with a pair of reduced leaves half way between the
substratum and the inflorescence. There is considerable variation in the
number of glands on the inflorescence, perhaps depending on the stage
of development. The genus is known only from the Himalayas, Yunnan
and New Guinea. We have reported similar ranges for Stellaria saxatilis

d

56 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Buch.-Ham. in Don, and Potentilla of the P. leuconota complex, but the
last two are also found in intermediate localities.

CUCURBITACEAE
Neoalsomitra Hutchinson

In a brief article published in the Annals of Botany, n. ser. 6: 95-102.
1942, Dr. J. Hutchinson has clearly shown that Macrozanonia Cogniaux
is only a synonym of Alsomitra Roemer, both having been established on
the same type-species. He proposed the new genus Neoalsomitra Hutchin-
son to take care of the rest of the species later included in Alsomuitra
Roemer.

Neoalsomitra Schultzei (Cogn.) Hutchinson in Ann. Bot. n. hae - 98. 1942.

Alsomitra Schultzei aan Pflanzenr. 66(1V. 275. I):

BRITISH NEW GUINE Lower Fly River, east a ae Sturt Island,
Brass 8149, Oct. 1936, ee in reed swamps (flowers yellow; fruit green, in-

This appears to be the first record of this species since it was originally
described on a specimen collected on the Augusta River, Northeast New
Guinea.

Neoalsomitra integrifoliola (Cogn.) Sls in Ann. Bot. n. ser. 6: 99, 1942.

Gynostemma integrifoliola Cogn. in DC. Monog. Phan. 3: 916.

Alsomitra shy aes eae Jour. Coll. Sei, Tokyo 30: i st, Ic. Pl. For-

s t. 38, 39. 1911; Cogn. Pflanzenr. 66(1V. 275. I): 1916

BRITISH NEW GUINEA: nae ee Division: ee R., Rona, Brass
3606, Mar. 1933, alt. 450 m., edge of rain forest, common (slender spieaing climber ;
upper surface of leaves Aull lower shining; flowers very pale yellow). Western
rationed Lower ie River, east bank opposite Sturt Island, Brass 8116, rain
forest, comm trees bordering swamps (liane; branchlets glaucous; leaves som
what ia ere Reported by Hutchinson from Formosa, the Philiouines:
and Fij

Melothria Linnaeus
Melothria scabridula sp. nov.

Monoica, scandens; caulibus gracilibus, sulcatis, nodis sparsim pube-
scentibus exceptis, glabris: foliorum petiolo 1.5-3 cm. longo, supra canali-
culato et marginibus minute pubescentibus, ceterum glabro; lamina
tenuiter chartacea, 3—-5-partita, margine subinciso-dentata, dentibus apicu-
latis, supra in sicco fusco-olivacea, consperse albido-scabridula et venis
minute pubescente, subtus paullo pallidiora, levi, glabra vel novella venis
minute pubescente, lobo terminali lanceolato, apice longiuscule acuminato,
48.5 cm. longo, ad basin 0.5-1 cm., in medio 1-2.3 cm. ato, lobis

mm. longo, 15 mm. lato) vel laminarum basi fere truncata; cirrhis
simplicibus, gracilibus minute pubescentibus; ¢ inflorescentibus racemosis
vel subcorymbosis, paucifloris (2—5-floris), pedunculo 1—1.5 cm. longo,
minute pubescente; pedicellis usque 5 mm. longis; calyce campanulato,

mentis 2.5 mm. longis, puberulis, apice curvatis; antheris ellipticis,

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 57

connectivo apice leviter producto; @ solitariis in eadem axilla cum mas-
culis dispositis ; pedunculo 1.5-4 cm. longo; ovario subgloboso; calyce
et corolla maris; staminodiis consperse et breviter pilosulis: stylo 3 mm.
longo, basi disco cupulari; stigmatibus 3, 1 mm. longis, apice expansis
complanatis, bilobatis; fructibus + globosis, 1 cm. diametro; seminibus
stramineis, obovoideis, circiter 4 mm. longis, 3 mm. latis, basi angustatis,
ae complanatis, immarginatis, scrobiculatis.

NETHERLANDS NEW GUINEA: 9 km. NE. of Lake Habbema, Brass 10621
(TYPE), Oct. 1938, alt. 2800 m., mossy fen climbing over a stump on forest path;
Bele River, 18 at NE. of Lake Habbema, Brass 11046, Nov. 1938, alt. 2200 m.,
small anal in forest clearings (flowers alu fruit red).

This species is easily recognized by the deeply cut ei and the
scrobiculate seeds, a character obvious even in the dried fru
Melothria belensis sp. nov. § EUMELOTHRIA

Monoica scandens; caulibus gracilibus, ey scabridulis, breviter et
sparsim pilosulis vel puberulis: foliorum petiolo 1 cm. longo, gracili, den-
siuscule adpresse pubescente; lamina tenuiter ae 4-10 cm. longa,

3.5—5 cm, lata, integra, cordata vel subhastata, apice acuminata vel interdum
ae margine denticulata (dentibus inter se 4-15 mm. remotis) in SICCO

sup

scabridula vel fere levi, sinu basilari latissimo concavo, 3-6 mm. profundo
vel fere nullo; cirrhis filiformibus basin versus sparsim et breviter pilosulis:
3 racemis simplicibus vel interdum ramosis, 7—23- floris, pedunculo com-

dentibus triangularibus, acutis, 1 mm. longis, basi paullo latioribus, +
nervatis; corolla 5-6 mm. longa, utrinque minute pubescente vel puberula,
S-lobata, lobis late ovatis; staminum filamentis 3 mm. longis, antheris
subquadratis, bilocularibus, 2 mm. longis, vix 2 mm. latis, loculis rectis,
connectivo latiusculo, obtuso, + ciliato; @ floribus solitariis, peduncu-
latis, pedunculo usque 4 cm. longo, puberulo, filiformi ; calyce et corolla

longo, basi disco cupulari (non annulariformi) ; stigmatibus Sie o} iaeten'
longo, apice See ees! fructibus tantum immaturis visis

LANDS NEW GUINEA: Bele River, 18 km. NE. of Lake Habbema,
Brass 11082 (rype), Nov. 1938, alt. 2200 m., common in open eng second growth
forest (flowers yellow).

Melothria idenburgensis sp. nov. § EUMELOTHRIA
Monoica, scandens; caulibus glabris vel novellis oa peu
sulcatis, nodis + puberulis; foliorum petiolo 1-2.5 OF iste

1-2 cm. profundo, plerumque late triangulari; ree simplicibus glabris,
tenuibus: $ floribus solitariis vel geminatis, 5-10 m. pedicellatis; calyce
campanulato, 2 mm. longo, lobis subulatis, 1 mm. Sere corolla fauce
pilosula, lobis extus + puberulis, ovatis, acutis, 3 mm. longis; staminum
filamentis brevibus; antheris subquadratis, loculis rectis, margine dorsali

58 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

minute ciliatis, connectivo latiusculo, apice ela producto; @ flori-
bus solitariis vel interdum geminatis, pedunculo + 1 cm. longo; ovario
fusiformi; fructu ovoideo, apice breviter ae creer 1.7 cm. longo,
1 cm. iametro, in sicco longitudinaliter et irregulariter rugoso; seminibus
oblongis basi paullo angustatis, 5 mm. longis, incl. 1 mm. alato- appendicu-
latis, 2 mm. latis

NETHERLANDS NEW GUINEA: Bernhard Camp, Idenburg River, Brass 14100
(type), April 1939, alt. 50 m., rain forest, occasional in fringe vegetation along streams
(flowers white; fruit unripe).

A species showing some similarity to the description cf M. Thwaitesi
Schweinf. (IM. zeylanica C. B. Clarke) but fruits not angled-costate.
Melothria i parece (Bl.) Cogn. in DC. Monog. Phan. 3: 601. 1881; Cogn.

flanzenr. 66(IV . 1): 101. 1916.

Bryonia oe ee Bijdr. 924. 1826.

BRITISH NEW GUINEA: Lower Fly River, east bank opposite Sturt Island,
a ie Oct. 1936, in low second growths on riverbank (flowers greenish; fruit
soft,

This calecion looks very much like the scanty material we have at hand
of Melothria leucocarpa (Bl.) Cogn., but the fruits are red, not white.

Trichosanthes Linnaeus

Trichosanthes Pulleana Cogniaux ex Harms, Bot. oo | 160, 1925, vel aff.
les papuana Pulle, Nova Guin. Bot. 8(2): . 1910, non F. M. Bailey
(1900).

BRITISH NEW GUINEA: Central Division, Mafulu, Brass 5347, Oct. 1933, alt.
1250 m., climbing in low regrowth forest (numerous pendent large red fruits about
20 cm. 16 ong, 8 cm. diameter, yellow inside, seeds black).

The above cited collection (in fruit only) agrees with the original
description of this species in the general shape of the leaf and the glands
near the insertion of the petiole. The petiole is 4-12 cm. long, the upper
surface of the leaf is scabrous; the fruit is somewhat oblong-obovoid with
a smooth surface; the seeds are a little more than 1 cm. long, 5 mm. wide,
flattened, slightly narrowed at one end and shallowly emarginate at the
other
Trichosanthes mafuluensis sp. nov.

Caules graciles angulato-sulcati, + puberuli demum ape Lt
membranaceis, supra parce e puberulis et scabriusculis, subtus mollit
berulis, interdum hic illic glandulis parvis conspersis; petiolo 3— 5 cm.
longo, pubescente: lamina ovata, profunde cordata, 9-20 cm. longa, 6—14.5
cm. lata, ae breviter acuminata, margine + remote serrato-denticulata,
sinu basali 2 —5 cm. profundo, 1.5-2 cm. lato, truncato-obtuso, medio ob
lobos basiles paullo angustato; cirrhis sulcatis = pubescentibus, 2—3-fidis ;
racemis ¢ 4-8 cm. pedunculatis, plurifloris (+ 15), pubescentibus et +
minute glandulosis; bracteis sub pedicellis lineari-lanceolatis, circiter 8

mm. eas integris, pedicellis gracilibus sub anthesi erectis, ad articula-
tionem -- 5-7 mm. longis; calycis tubo anguste infundibu lar a 2 cm
longo, ie lobis 4 mm. longis, lanceolatis, integris, ie corollae
lobis oblongis, 7-8 mm. longis, obtusiusculis, granulo-puberulis, ciliis fim-
briatis; filamentis eee glabris; antherarum capitulo 3 mm. longo,
faucem attingente non exserto; “floribus ? non visis.

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII 59

BRITISH NEW GUINEA: Central Division, Mafulu, Brass 5257 (TYPE), Oct.
1933, alt. 1250 m., climbing in forest regrowth brush (corolla white).
Trichosanthes dieniensis sp. nov.

Caules gracillimi, novelli puberuli demum glabrati, angulato-sulcati;
foliis membranaceis, supra scabriusculis, subtus glabris; petiolo 2-4.
cm. longo, glabro; lamina ovata, cordata 4. 5-12 cm. longa, 2.5—8 cm. lata,
apice breviter acuminata, margine inconspicue et remote glanduloso-
denticulata, sinu basali Oy 2 cm. profundo, 1.3-3 cm. lato, obtuso vel
subrotundato; cirrhis gracilibus bili, glabris; racemis 8 3-5.5 cm

obovato-oblongis, ue 5-lobatis ‘vel incisis: medic erectis, ad articula-
tionem vix 3 mm. longis; calycis tubo anguste infundibulari, 4.5—7.5 cm.
longo, glabro, lobis circiter 1 cm. longis, lineari-subulatis, basim versus
1-3- denticulatis: eee lobis -- 1 cm. longis, ramoso- fimbriatis, fimbriis
circiter 2 cm. longis, utrinque puberulis, fauce ++ puberulo- -pilosulo;
filamentis brevibus gracilibus: antherarum ‘capitulo 6 mm. longo, novello
glabro, sub anthesi puberulo: floribus @ non visis.

BRITISH NEW GUINEA: Ononge Road, Dieni, Brass 3898 (type), April 1933,
alt. 500 m., massed on a dead tree trunk (flowers white).

CAMPANULACEAE

Pratia Archboldiana sp. nov.

Planta in sicco habitu rosulata, acaulescens vel caulibus brevissimis
2-4 mm. longis); foliis approximatis petiolatis, petiolo 3-7 mm. oe
piloso vel glabrato, lamina ovato-rotundata vel subrotundata, 0.5-1.5 ¢
longa lataque, membranacea vel chartacea, margine undu lata vel ae
dentata, supra consperse pilosa, subtus glabra; floribus solitariis, axillari-
bus, pedunculatis, pedunculo 2-4 mm. longo, pilosulo; ovario piloso, cir-
citer 3 mm. diametro; calycis lobis lineari- lanceolati tis, mm. longis,
pilosulis; corollae tubo 2.5 mm. longo, lobis 3 mm. longis, lanceolato-
linearibus; antheris in tubum connatis, anticis breviter aristulatis: baccis
globosis, -— 3 mm. diametro, calyce et ‘interdum corolla marcescente coro-
eae seminibus brunneis, nitidis, levibus, obovoideis, circiter 0.8 mm.
longi

aati NEW GUINEA: Murray Pass, Wharton Range, Brass 4943 (TYPE),
Aug. 1933, alt. 2840 m., under a rock wall on grassy bank of a creek, very rare
(flowers red).

This specimen consists of a half dozen small plants, the largest of which
is between 3 and 4 cm. diameter. Most of them are in fruit with the
withered corolla still adhering to the apex, one flower-bud was found, and
in this we found the anthers united in an oblique ring around the style,
which character gave us the clue to the family. The species may possibly
be related to P. irrigua (R. Br.) Benth., a native of Tasmania.

Lobelia conferta sp. nov.

Planta prostrata, parva, carnosula, glabra; ramis brevibus; foliis alter-.
nis, integris, oblongo-lanceolatis, 4-5 mm. longis, 1-2 mm. latis, apice
rotundatis vel obtusis, sessilibus, basi decurrentibus, confertis; floribus in
axillis foliorum superiorum solitariis; pedicellis 3-5 mm. longis, ebracteo-
latis; hypanthio brevi, calycis lobis lineari-lanceolatis, 1-1.4 mm. longis,

60 JOURNAL OF THE ARNOLD ARBORETUM

acutiusculis;

longis, inferioribus 3 vix 2.5 mm. longis, lanceolatis; fila

[VOL. Xxx

corollae tubo 2 mm. longo, lobis 2 superioribus vix 2 m

mentis in part
m

inferioribus liberis, superiore connatis, glabris; antherarum tubo 1
longo, antheris 2 inferioribus apice sparsim et minute barbatis et ee

diculato minuto instructis;
BRITISH NEW GUINEA: Mt.

capsula non visa
Albert Edward,

southwest slope, Brass 4417

(TYPE), July 1933, alt. 3680 m., on wet grassland (prostrate plant, very rare; leaves

mooth, shining; flowers very pale purple

This species, if it is a Lobelia, Belen to section Hrmipocon, but we

are unable to suggest any near relationships.

It may possibly belong to

Pratia, but without fruit it is difficult to say.

GOODENIACEAE

a Serre R. Br. Prodr. 5
1868

a Bailey, oe FI. 3: 892.

ae 7) 108. Fe
H NEW ae

northern Australia

Tarara, Wassi Kussa River, Brass 8387, Dec.
nea forest (common ee on acid gray soil;

810; F. v. Muell. Fragm. Phytogr. Austr.
1900; Krause, Pflanzenr.

1936,

flowers pale blue). Queensland and

INDEX TO GENERA AND FAMILIES CONTAINED IN PLANTAE
PAPUANAE ARCHBOLDIANAE, I-XVIII, AND FOUR SMALL
PAPERS WITH SOME NEW GUINEAN SPECIES

In each instance the number preceding the colon represents the volume
number of the Journal of the Arnold Arboretum and the one following
the colon is the page on which the name first appears.

Acacia, 23: 395

Acronychia, 22: 55
Acenonticr, ae 329; 23: 396
Agatea, 2 8

Aglaia, 21: 316; 29: 158

Amaracarpus, 27: 221
Amar ranthaceae, 23: 385
mpelocissus, 22: 375
Fase a 22: 529; 29: 158
Antirhea, 26: 233
fee ees 21: 314
5 : 199; 24: 210; 29: 163
Aquifoliaceae, 20: 333; 22: 258; 30: 40

A 1
Aristolochia, 23: 383; 29: 152
Aristolochiaceae, 23: 383; 29: 152

Arytera, 21: 522
Astronia, 24:
Astronidium, 24: 435

B
w
rs

Badusa, 25: 187

Balanophora, 23: 383; 29: 152

Balanophoraceae, 23: 383; 29: 152
390

Begonia, 24: 41; 29: 160
Begoniaceae, 24: 41; 29: 160
9

6: 34
Buchanania, 22: 529

Callitrichaceae, 22: 258; 29: 158
Callitriche, 22: 258; 29: 158
Calogyne, 22:

Calycosia, 27:
Campanulaceae, 22: 383;
Campnosperma, 22: 533
Canavalia, 23: 406

30: 59

1949]

Canthium, 26: 230
Cardamine, 24: 207
Carruthersia, 24: 215
Caryophyllaceae, 23:
Cassia, 29: 155
Cayratia, 22: 378
Celastraceae, 20: 335;
53

386; 30: 39

22: 260

Chisocheton, 21: 311; 29:
Cissus,

Cleistocaly 23: 233
4: 34; 29:

#163
Gants 23: 397

Dacry: dium, 30: 39

Dentella, 1

Derris, 23: 403

Des an 23: 401; 29: 156
Dodonaea, 21: 525

Dolicholobium, 25: 183

Duabanga, 22: 269

Dysoxylum, 21: 299; 22: 256; 29: 157

Elaeagnaceae, 22: 267
Elaeagnus, 22: 267
nieeoden tron: 20: 336

MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII

Eugenia, 23:
haa cs yes
Evodia, 22: 34
Excavatia, 21: 199
Fagraea, 23: 412
Flindersia, 20: 330;
Freycinetia, 20: 140; 21:

22: 58
163

Galium, 26: 265

Glinus,

Glycine, 29

Gompholobium, 23: 400
Goodeniaceae, 22: 387: 30: 60
Guioa, 21: 512

Gymnosporia, 20: 335
Gynostemma, 29: 168
Gyrinops, 22: 264
Haloragaceae, 23: 407; 29: 161
Se ee 29: 161

iivppocrsteaceac! 20: 336
Hollrungia, 24: 210; 29:
Hybanthus, 24: 208
Hydnophytum, 26: 14

Tlex, 20: a 22: 258; 30: 40
Ixora, 26:

Jagera, 21: 522

Kelleria, 22: 267
Keraudrenia, 30: 44
Kingi uae 23: 399
Kurrimia, 20: 335
Lagerstroemia, 22: 268
Lasianthus, 27: 219
Le Bp waren 29:
Leea, 22: 380
Teena YF 329;
521

161

Loganiaceae, oe 408; 29: 163
Lucinaea, 25:

iLgastnaeveitiay 55 a
Lythraceae, 22: 268

Mangifera, 22: 532

23: 391;

160; 30: 44

29:

61

155

62 JOURNAL OF THE ARNOLD ARBORETUM

Maniltoa, 23: 397; 155
Maschalodisns, 25% a 30: 55
Mastixia, 21

3
Mastixiodendron: 23: 416; 25: 205; 26:
254

Medinilla, 21: 198; 24: 4

Melastomataceae, 21: i 24: 422

Meliaceae, 21: 296; 22: 256; 29: 157
5

Myrtus, 23: 237

Nauclea, 25: 187

Neuburgia, 15
Nyctaginaceae, 20: 327
Nymphaea, 23

Nymphaeaceae, 23: 388
Nymphoides, 30: 45

Ochrosia, 24: 213
Colton, A
Ormosia, 23:

Pachystylus, 26: 229

Parasponia, 20: 324; 22: 254
Parastemon, 21: 197; 22: 256
Parkia, 23: 396

Parsonsia, 24: 29: 164
Passiflora, 30:

Passifloraceae, ae 210; 29: 160; 30: 44
384

Pentaphragma, 22:
Perrottetia, 20: 336; 22: 261
Phaleria, 22: 265

Phyllocharis, 22: 387
Piptadenia, 23:
Pithecellobium, 23: 393

Polygonaceae, 23: 384; 29: 154
Polygonum, 23: 384; 29: 154
386

Po aero 23: 386
Potentilla, 184

Randia, 25: 198; 29: 165

Rhus, 22: 536; 29: 159

eee 2: — 22: 256
Rotala, 22:
Rourea, fs

Rubiaceae, i. me 25: 183; 26: 1, 229;
30: ‘$8

193; 29: 164;
Rubus, 21: 179
R

mex, 23: 384
Rutaceae, 20: 330; 22: 32
Sabiaceae, 20: 356; 22: 262
Sagina, 23: 387; 30: 39
336

Syzygium, 23: 248

Tarrietia, 30: 40

[VOL. XxX

1949] MERRILL & PERRY, PLANTAE ARCHBOLDIANAE, XVIII

Taxaceae, 30: 39

Turpinia, 22: 543

Ulmaceae, ZO. 324; 22: 253
189

Uncaria,
Urophyllum, 26: 229

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

Medes 30: 55
97

Vitaceae, 22: 375

Wahlenbergia, 22: 383
Wikstroemia, 22: 266
Wrightia, 24: 215

Xanthomyrtus, 23: 240
See 203 13305.-22:. 32
337

Zizypous, 2
Zygophyllaceae, 21: 511

64 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

ORIGIN OF THE ANGIOSPERMS: NEED FOR A
BROADENED OUTLOOK 1!

I. W. BAILtey

WE ARE HEARING more and more of late, in various fields of human
endeavor, regarding the significance of a “New Look.” Much is being
spoken and written regarding the necessity for a “New Systematics”
and for a ‘New Morphology.” All of us will, I believe, readily admit the
need for new methodologies and for new points of view in the develop-
ment of botanical science. However, if real progress is to be made, it is
essential to analyze trends of putative originality to determine whether
they are soundly conceived and of productive potentialities. This is
particularly true in dealing with such complex and difficult problems as
to how, when, and where the angiosperms originated, how they became so
highly diversified and so widely distributed, and how they should be classi-
fied in a system of truly natural relationships.

The “New Morphology” commonly involves the basic assumption —
to quote a recent exponent ~——that a comparison of the organs of
angiosperms and of lower plants only makes sense if the former are
abandoned as a starting point and if we allow ourselves to be guided only
by paleobotanical evidence in the direction of evolutionary development.
Is the sweeping generalization that we should be guided in the study of
angiospermic phylogeny and classification solely by paleobotanical evi-
dence a sound one? When shorn of its psilophytalian halo, the “New
Morphology” frequently consists in practice of highly speculative attempts
to homologize specific parts—even teratological ones—of a few
selected angiosperms with those of ancient land plants. In so doing,
scant attention is given to long stretches of geological time during which
profound morphological changes are likely to have occurred. Further-
more, the fact that many salient trends of morphological specialization
are adequately preserved in the huge assemblage of surviving ei
is overlooked or ignored. This procedure of the ‘““New Morphology”
not new, since it is essentially similar to that utilized in the past, eS
example, in interpreting the Magnoliaceae in terms of the Bennettitales
and the Amentiferae in homology with the Gnetales.

It is not my intention to belittle the great contributions that the study
of the Psilophytales and of other ancient land plants have made to a
clearer understanding of the stem and of its appendages, but rather to
1 Invitation paper read before the Paleobotanical Section of the Botanical Society
of America, Sept. 13
2Lam, H. J. Classification and the New Morphology. Acta Biotheoretica 8: 109.
1948.

1949] BAILEY, ORIGIN OF THE ANGIOSPERMS 65

emphasize the necessity for much broadened outlooks in discussion of
the origin and the relationships of the angiosperms. We need “new
looks,” but they should be comprehensive, rather than narrow and ex-
cessively speculative ones. There are as grave dangers in indiscriminately
interpreting selected structures of living angiosperms solely in terms of
ancient Land Plants as there formerly were in studying fossil plants with
excessive dependence upon the classical concept of the stem and leaf.
For example, recent attempts to establish the Sarcopodaceae as a new
family of gymnosperms, and to utilize it as support of a stachyosporous
line of angiosperms provide cogent evidence of inherent dangers in the
methodologies of the “New Morphology.” The authors? of the new
family have been forced to admit that Sarcopus aberrans is a species of the
santalaceous genus Exocarpus and to change the name of the family to
Exocarpaceae. If the Exocarpaceae are to be removed from the Santa-
laceae and placed in closer relationship to the gymnosperms such a transfer
should be based, not solely upon superficial examinations of the flower
of a single species, but upon comprehensive cytological, embryological,
anatomical and general morphological studies of Execarpus in comparison
with other genera of the Santalaceae and of other families of the Olacales.
In addition, phylogenetic conclusions should be in harmony with salient
trends of morphological specialization in the angiosperms as a whole. In
the case of Exocarpus, even an examination of the pollen and the xylem

generalizations. Similar objections apply in the use of selected parts of
such genera as Casuarina, Salix, Ricinus, Portulaca, Calothamnus, etc.
when divorced from their context in the dicotyledons as a whole.

It should be emphasized in these connections that diversified investi-
gations of surviving angiosperms provide the only available means at
present of morphologically characterizing this great group of the vascular
plants, and for recognizing and of accurately identifying ancestral forms
when found. Each botanical discipline, Taxonomy, Paleobotany, Phyto-
geography, Cytolog if Embryology, Anatomy, Developmental Morphology,
Genetics, etc., has important contributions to make in the ultimate solu-
tion of various sae of the great central mysteries, but the limitations
of each field of research in the solution of specific problems of phylogeny
and relationship should be clearly visualized and freely admitted.

The Land Plants from the Psilophytales onward are aptly designated
by some the Tracheophyta, since they are characterized throughout by
the formation of tracheary tissue. The constituent cells of this tissue
obviously perform two fundamentally important functions, 1) the con-
duction of water and solutes from the absorbing to the transpiring parts
of the plant, and 2) the provision of mechanical strength in parts where
they occur. The former function is facilitated by peculiarities in the

3 GacNEPAIN, F. & Ep. Bourreau. Une nouvelle era de Gymnospermes: les Sar-
copodacées. Bull. ae Bot. France. 93: 313-320.

———— & ———.. Nouvelles considérations are oe a propos du Sarcopus
aberrans Gagnepain. Bull. Soc. Bot. France. 94: 182-185. 1947.

66 JOURNAL OF THE ARNOLD ARBORETUM [VOL, Xxx

form of the cells and by the loss of their cytoplasmic contents at maturity,
the latter function by the remarkable physical structure and the chemical
composition of their cell walls. The major trends of phylogenetic modi-
fication of the tracheary tissue of the Land Plants are associated with
changes of equilibrium between these two fundamentally important physio-
logical functions. In the case of most dicotyledons, there is a striking
division of labor within the vascular tissues, certain linear series of
tracheids becoming modified to form vessel members, whereas others
assume an increasingly fiberlike form and may at times take on a storage
function in addition to an exaggerated mechanical one. These salient
trends of evolutionary specialization of the tracheary tissues are largely
unidirectional and irreversible, and are fully and adequately preserved in
surviving angiosperms. ‘There fortunately are no serious missing links
in these phylogenetic chains and it is not essential, for example, to search
geological strata for vesselless pro-angiosperms since ancestral types of
primitive xylem occur in living representatives of both the dicotyledons
and the monocotyledons.

Before discussing the significance and the limitations of these phy-
logenies in the study of plant relationships, it is essential to comment
briefly upon the character of the evidence upon which they are based.
Large collections of materials of woody dicotyledons have gradually been
accumulated at Yale, Oxford, Harvard and many other institutions.
These collections when supplemented by herbarium specimens have
afforded material for the study of many genera of all of the principal
families of the angiosperms. In addition, sufficient developmental investi-
gations have been made so that it now is possible to visualize salient trends
of specialization of the xylem in terms of a phylogeny of successively
modified ontogenies. In other words, the evolutionary picture has emerged,
and has become increasingly distinct, as larger and larger volumes of data
have been analyzed from the angiosperms and the gymnosperms as a
whole. It has crystallized solely from the study of the vascular tissues
themselves, and entirely independently of assumptions regarding the puta-
tive primitiveness of specific representatives of the dicotyledons or mono-
cotyledons. Furthermore, these particular trends of evolutionary speciali-
zation cannot be read in a reverse direction, since no one is likely to argue,
in the light of our present knowledge of the Tracheophyta, that tracheids
originated from the dissociated members of vessels.

What then are some of the more important contributions that a knowl-
edge of these unusually extensive and reliable vascular phylogenies can
make in discussions of the origin and the relationships of the angiosperms?
Obviously it would be possible to arrange the angiosperms according to a
system of increasing specialization of the vascular tissues, but such a
system of classification would prove to be as unsatisfactory as others that
have been based primarily upon the study of restricted parts of plants. If
a truly natural classification is to be attained, it must be based upon the
analysis and the harmonization of evidence from all organs, tissues and
parts. This is due to the fact that the morphological specializations of

1949] BAILEY, ORIGIN OF THE ANGIOSPERMS 67

different structures commonly are not synchronized. The flower or one
of its parts may be changing rapidly at the time when the vascular tissues
exhibit morphological stability or vice versa. In studying plant rela-
tionships, evidence from the vascular tissues is in general more significant
in negations than in affirmative conclusions. This is due to the frequent
occurrence of parallel trends of evolution in the various organs and tissues
of the Land Plants. Thus, although it is impossible to derive truly primi-
tive forms of xylem from highly specialized ones, close structural
similarities may be due to parallel or convergent development rather than
necessarily to close genetic relationship.

In illustration of these basic considerations, let us turn to a specific dis-
cussion of the present revival of assumptions regarding the possibility of
deriving the angiosperms from the Coniferales, Gnetales, Bennettitales or
Williamsoniales.

The Lycopsida, Sphenopsida, and most of the Pteropsida are charac-
terized by having essentially similar ontogenetic sequences in the develop-
ment of their tracheary tissue. In a fully elongated axis, as seen in
radial longitudinal sections, the earlier formed tracheary cells have loosely
coiled helical thickenings, although they may be preceded at times by
tracheids with annular thickenings. These cells are succeeded by tracheary
cells having compactly coiled helical thickenings of very low pitch, and
these in turn by elements exhibiting transitions between scalariformly
reticulate and scalariformly bordered pitted walls. In certain groups of
Land Plants, e.g., many Lycopsida, Sphenopsida, and ferns, the subse-
quently formed tracheids have dominantly scalariform pitting, whereas
in other groups, e.g., many seed ferns, Cordaitales and Cycadales, there
are subsequent transitions between scalariform bordered pits and approxi-
mately circular ones of varying number and distributional patterns.

In the Ginkgoales, Coniferales, and Gnetales, this normal ontogenetic
sequence, characteristic of a vast majority of the Land Plants, is profoundly
modified. True scalariform bordered pitting is eliminated and circular
bordered pits are formed in the earlier stages of the ontogenetic sequence,
even in the first-formed tracheary cells with loosely coiled helical thicken-
ings. This modification of the primary vascular tissues is firmly established
in the Coniferales and Gnetales and is indicative of an ancient trend of
drastic anatomical specialization. In the latter order, vessels developed
from tracheids having circular bordered pits. On the contrary, in the stems
of angiosperms, which have retained a primitive type of vesselless xylem,
the ontogenetic sequence is of the normal type, and vessels have developed
in both dicotyledons and monocotyledons, as in Pteridium, by the modi-
fication of tracheids with scalariform bordered pitting. Such funda-
mentally significant anatomical differences form an insuperable barrier
to a derivation of the angiosperms from the Coniferales or the Gnetales.
Thus, the presence of vessels in both the Gnetales and the angiosperms,
which has so frequently been cited as evidence of relationship, actually
negates such relationship. There are similarities between the end prod-
ucts of tracheary specialization in Gnetum and certain of the dicotyledons,

68 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

but they have arisen by entirely different developmental changes.

hat then are the possibilities of deriving the angiosperms from the
Bennettitales or Williamsoniales? Here the evidence from the vascular
tissues alone is inconclusive. Although similar transitions between
tracheids with scalariform and circular bordered pitting occur in the
secondary xylem of vesselless dicotyledons and of certain cycadeoideas, the
similarity is not indicative necessarily of close relationship, since it may
be due to parallel evolutionary development. Furthermore, the possibility
of deriving the angiosperms from the Bennettitales or Williamsoniales, as
also from the Corystospermaceae or Caytoniales, appears to be negated
by evidence from other parts of the plants, particularly the reproductive
ones.

It is essential at this point to digress briefly for comment upon a recent
attempt to derive certain ranalian families from the Bennettitales, and
to arrange them in a linear evolutionary series. This effort is based
largely upon the reported discovery of a new type oi tracheary element in
the Magnolicaeae and supposedly related families, and affords a pertinent
illustration of inherent dangers in a narrow anatomical look. The cells
in question are of a well known tracheary type which was fully discussed
at meetings of the Committee on Nomenclature of the International
Association of Wood Anatomists. They are new only through changes
in terminology.® A broad anatomical look at the vascular plants as a
whole reveals the fact that tracheids with a similar configuration of
bordered pitting occur, not only in the Bennettitales but also in other
gymnosperms, and in many orders of the dicotyledons as well as in certain
families of the Ranales. Furthermore, the true significance of the con-
figurations cannot be fully understood without reference to accumulated
data regarding the physical structure of the walls of tracheary cells. It
should be emphasized in passing that it is such sweeping generalizations,
based upon limited and inadequate data, that have raised uncertainties
and doubts in the minds of many botanists regarding the value and the
reliability of anatomical evidence in the study of phylogeny.

Up to this point, I may have given the impression of an excessively
pessimistic look at the problem of the origin of the angiosperms, in which
the only certainties are negations and in which potential ancestors of the
angiosperms are successively eliminated. However, if real progress is to
be made in the solution of the great mystery, it is essential periodically to
differentiate the wheat of reliable cumulative evidence from the chaff of
excessive speculation. As a matter of fact, I am quite optimistic regarding
advances that can be made in the near future, provided investigators in
various botanical disciplines cooperate more actively in providing essential
data for a well coordinated attack along a broad morphological front.
+LeMESLE, Ropert. Les divers types de fibres a ponctuations aréolées chez les
dicotylédones apocarpiques les plus archaiques et leur réle dans la phylogénie. Ann.
Sci. Nat. Bot. et Biol. Végétale 7: 19-40. 1946.

5 Bar I. W. The problem of differentiating and classifying tracheids, fiber-
ee and libriform wood fibers. Tropical Woods 45: 18-23. 1936

1949 ] BAILEY, ORIGIN OF THE ANGIOSPERMS 69

Our knowledge of the vascular tissues of Land Plants has advanced
to a position where it can aid materially in planning such a campaign, in
avoiding serious pitfalls, and in restraining fruitless sorties in haphazard
directions. Thus, the tracheary phylogenies, that are so clearly and fully
preserved in surviving angiosperms, provide us with significant clues
regarding the general habit of growth of ancestral dicotyledons. This is
due to the fact that, in the various dicotyledonous families, herbs, vines,
lianas, succulents, aquatics and extreme xerophytes exhibit a high degree
of specialization of the vascular tissues, whereas vesselless xylem and the
less modified forms of vessel-bearing xylem occur in large woody peren-
nials, viz. shrubs or trees. Such facts negate the possibility not only of
deriving structurally primitive types of arboreal dicotyledons from her-
baceous ancestors, but also of deriving the monocotyledons from herbaceous
dicotyledons, e.g., the Ranunculaceae. If the angiosperms are mono-
phyletic, the monocotyledons and the dicotyledons must have diverged at
an early date, since vessels have developed independently in the two
groups.

Of the nearly 100 species of living dicotyledons, which have a primitive
vesselless type of both primary and secondary xylem, all occur within
the order Ranales, as broadly conceived by Engler and Prantl. Further-
more, an unusually complete record of the origin of vessels and of salient
trends of specialization of the xylem is preserved within this order. That
these occurrences are not purely fortuitous, and may have significant impli-
cations, is suggested by accumulating data regarding the pollen morphology
of seed-bearing plants. Most of the families and orders of the dicotyle-
dons are characterized by having tricolpate pollen or types of pollen that
appear to have been derived phylogenetically from such grains. Tricolpate
pollen is not known to occur in any other group of seed-bearing plants and
therefore is indicative of a highly distinctive trend of specialization in
pollen morphology. Monocolpate pollen, which is characteristic of so
many seed ferns, Bennettitales, Cycadales and Ginkgoales, is of common
occurrence in monocotyledons, but is largely confined in the dicotyledons
to a number of the more woody families of the Ranales. In other words,
insofar as I have been able to determine, the Ranales are the only order
of the angiosperms in which both monocolpate and tricolpate pollen occur.

However, it was the remarkable carpels and stamens of a previously
undescribed family of the Ranales, viz. the Degeneriaceae, which induced
Dr. A. C. Smith and myself to initiate a series of comprehensive coordi-
nated investigations of ranalian families, in which others are now actively
cooperating. I have already reported, earlier in these meetings, upon the
primitive, unsealed, conduplicate ranalian carpel and concerning its
salient trends of specialization in different ranalian families. There are
equally significant transitions in the Ranales between broad microsporo-
phylls and stamens of a more conventional form. These investigations
at least justify a “New Look” at the flower of the angiosperms as a whole.
Comprehensive studies of the various orders and families of living angio-
sperms should eventually provide adequate accumulations of evidence

70 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

for determining whether the putative lines of stachyosporous dicotyledons
actually are such or have been derived from a phyllosporous one such as
characterizes the Ranales. It should be noted in this connection that
detailed and extensive investigations of other parts of living angiosperms,
viz. the leaf, fruit, seed, embryo-sac, seedling, etc. are essential in synthe-
sizing a picture of a primitive angiosperm or pice to serve as a
guide to paleobotanists in their search for pro-angiosper

Allow me to conclude my remarks with a few brief ane upon those
parts of the Earth’s surface which appear to me to offer the most promising
fields for future explorations. It is the living floras of northern Australia,
New Guinea, New Caledonia, and Fiji and adjacent regions northward to
southern China that have yielded and are continuing to yield the richest
crop of missing links in the chains of angiospermic phylogenies. For
example, of the nine known genera of primitive vesselless dicotyledons,
five occur on New Caledonia and three are endemic on that island. Only
a beginning has been made in the exploration of many of these floras and
a continuous flow of significant new plants may be anticipated for some
time to come as more and more complete collections are assembled. Fur-
thermore, these living floras of southern latitudes have already yielded
more structurally primitive dicotyledons than have all of the known fossil

Tertiary and Mesozoic rocks of these southern lands have been adequately
explored. It should be emphasized, in addition, that comparatively little
is known regarding the seed-ferns of the Glossopteris flora. Here again
is a field in need of detailed exploration.

In conclusion, I may be permitted, as one who is approaching the ter-
minal stages of his career, to offer a word of friendly advice to those of
the younger generation of taxonomists, paleobotanists and morphologists
who may become interested in the origin of the angiosperms: Look West
young man toward the remnants of Gondwana Land!

INSTITUTE FOR RESEARCH IN GENERAL PLANT MorpHo.ocy,
Harvarp UNIVERSITY.

1949] CHATTERJEE, VALIDATING BOTANICAL NAMES 71

VALIDATING BOTANICAL NAMES BY REFERENCE
TO PRE-1753 LITERATURE

D. CHATTERJEE
IN THE PAST, botanical names were considered validly published when

lished literature. In many cases such literature was published before
1753. Furtado (Gard. Bull. Str. Settlm. 10: 170-71. 1939) has suggested
that this method is wrong, and that such publication does not validate the
names. This view appears to be contradictory to the existing practice and
it seems necessary to discuss the subject with a view to clarifying this

oint.

Furtado’s views summarised by himself are as follows (/. c.):—

“iv. No see 1 Name can be admitted as ae unless (7) it satisfies the condi-
tions laid down as ee for the word o ords constituting the name itself, and

s
(2) it is accompanied by either (a) a valid descupubel or (b) a valid reference (see
below v) .

mV REFERENCE is valid when it refers to a description published both in valid
literature, and in come with the rules laid down as essential for the words
ae the descriptio
i. Consequently ALL REFERENCES TO INVALID LITERATURE, even when they include
ae should be ER OEE under the Rules.
“vij. SevERAL Names (formerly ignored but recently resurrected and made the bases
ny nomenclatural changes) whose sole claim for validity is that they were

eee to a plate or description in invalid literature should be invalid: e.g.

“(1) In O. SrickMan’s Dissertation on RumMpuivs’s Herbarium Amboinense eds. 1754
and 1759, accepted as valid by Dr. E. D. MERRILL (Interpr. RUMPH. Herb.
Amboin, 1917).

“(2) In BurMann’s Index to RUMPHIUS’S eS Amboinense eds. 1755 and 1769,

accepted as valid by Dr. Merrit (cf. op. cit.).

“(3) In Roxsurcu’s Hortus Bengalensis oo accepted as valid by the Jate Dr. C.
B. Rosinson (Philipp. Journ. Sci. VII, 1912, pp. 411-419) because RoxBuRGH
had referred under the names to plates in Rumputius’s Herb. Amb.

““(4) In Moon’s ge of te and ee Plants of Ceylon (1824) ae
as valid by . A. G. ALsTon (Ann. . Bot. Gard. Peradeniya XI, 1929
pp. 203-205) because the ey contain eee a the ee or Saas either
in RHEEDE’s Hortus eee or in Rumpuius’s Herbarium Amboinense.

(5) Several generic and specific names listed in nomenclators and In dex ie wensis

with arena to invalid descriptions, whether or not published in valid
Sante

It eae be noted that Furtado has introduced a few unfamiliar terms

in the above summary such as “valid and invalid literature,” “invalid

description,” etc. By “invalid literature” Furtado means all botanical

literature published before 1753. This is evident from his statement

(1. c. 165), “Epipogium and Pterocarpus are said to have been validated

72 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

by a reference to pre-1753 literature; but Art. 20 invalidates this litera-
ture.” It should be pointed out that Art. 20 reads as follows: “Legitimate
botanical nomenclature begins for the different groups of plants at the
following dates (a) Phanerogramae and Pteridophyta, 1753 (Linnaeus,
Species Plantarum ed. 1).” This article essentially refers to botanical
nomenclature and not to botanical literature. The suggestion to describe
all botanical literature published before 1753 as “invalid” appears to be
due to a misinterpretation of Art. 20. Botanical literature containing
descriptions or plates of plants is effectively published when it is printed
and distributed to botanical institutions. So far as the descriptions and
plates are concerned they should be regarded as validly published irrespec-
tive of the year 1753. But the validity of binomial names used in these
publications will be considered only from 1753.

Unless this interpretation is accepted we have to reject a large number
of names from Linnaeus’s Species Plantarum. For example, the name
Plumbago zeylanica L., is published in Species Plantarum 151 (1753),
with references to Hortus Cliffortianus 53 (1737), Hortus Upsaliensis 43
(1748), Flora Zeylanica 73 (1747), etc. The only description available
in the Species Plantarum is “foliis petiolatis,” which is an exact repetition
of what is available in Hortus Cliffortianus. It is evident that the two
words constituting the description are indeed insufficient and an adequate
idea of the plant can only be obtained by reference to other effectively
published descriptions in earlier literature. In a large measure, therefore,
Linnaeus validated the name Plumbago zevlanica L. by reference to pre-
1753 literature. Similar examples are also available from other names
like Jasminum officinale L., Justicia echioides L., Nyctanthes Sambac L.
and Salvia pinnata L. In the case of Plumbago zeylanica L., the repetition
of the two words from Hortus Cliffortianus conveyed no additional infor-
mation, as the reference itself was enough. What Linnaeus did, in this
case, was to publish a binomial and validate it by reference to a previously
and effectively published description in pre-1753 literature. This impli-
cation should be clearly understood.

In the above cases Linnaeus had actual specimens as well as previously
published descriptions of these plants. There are, however, quite a number
of cases where binomials were published and validated by Linnaeus solely
on the basis of descriptions or plates in pre-1753 literature. Svenson’s
interesting paper on the descriptive method of Linnaeus (Rhodora 47:
278. 1945) shows that a large number of species described in his Species
Plantarum were based on “Figures such as those of Plukenet and of
Cornut’s Canadensium Plantarum Historia 1635, and Sloane’s History
of Jamaica, of which Linnaeus had seen no herbarium specimens.” A
random examination of one hundred specific names from Linnaeus’s Species
Plantarum shows that about fifteen of these were based solely on such
descriptions or plates in pre-1753 literature, and no herbarium specimens
of these species are to be found in the Linnaean Herbarium. This very
rough estimate may give some idea of the total number of species described
by Linnaeus on this principle. The name Cyclamen indicum L. may serve

1949] CHATTERJEE, VALIDATING BOTANICAL NAMES 73

as an example in this connection. The name was validated in the Species
Plantarum by a reference to Flora Zeylanica 401, 1747, which was based
on two drawings by Hermann (Zeyl. 50-57). It is almost certain that
Linnaeus did not see any specimen of this species and no specimen is
available in the Linnaean Herbarium. It may be pointed out, in passing,
that as a result, the correct identity of this plant is still very obscure.*
Even if some of these names are still, and probably will always remain
obscure, because they were based on such pre-1753 literature, that is no
reason for rejecting this type of publication. The existing rules of nomen-
clature do not prevent others from following the practice initiated by Lin-
naeus. The first author on East Asiatic Botany who followed this principle
and validated a binomial by reference to pre-1753 literature appears to
have been Prain. He published the name Sindora galedupa Prain (Jour.
As. Soc. Beng. 66: 483. 1897) by reference to Rhumphius’s Herb. Am-
boinense (2: 59, tab. 13. 1750). Furtado appears to have missed Prain’s
name in his list of Robinson, Merrill and Alston (J. c.).

From a consideration of these cases it will be understood that this
practice of validating names by reference to pre-1753 literature is indeed
an old one, starting in fact in 1753 with Species Plantarum itself. These
names cannot be regarded as invalid, as argued by Furtado. The fallacy
in Furtado’s discussions lies in his assumption of all pre-1753 literature
as “invalid literature.” It appears to me that he has also failed to inter-
pret correctly Articles 20, 36 and 37 of the rules of botanical nomenclature.
Further, he has unfortunately confused the two phrases “effective pub-
lication” and “valid publication,” and in all probability considered them
as synonymous. It is admitted that “valid publication” of botanical names
started from 1753, but the “effective publication” of botanical literature
started much earlier, perhaps with the art of printing. A proper and
logical interpretation of the rules would only lead us to accept these works
(botanical literature) as effectively published. It is unfortunate that by
misinterpreting the rules, Furtado has coined such terms as “invalid
literature” and “invalid description.” + In my opinion, this practice of
validating new binomials by reference to pre-1753 literature is perfectly
legitimate and in accordance with the existing rules. Such names are
therefore to be regarded as valid names

This subject was discussed at the Symposium of botanical nomenclature
held at Utrecht (June, 1948) and there was general agreement to the view
expressed by the writer in this paper.

*In Pflanzenreich — Primulaceae 4, 237: 248. 1905, Pax and Knuth wrongly con-
sidered Cyclamen indicum L. as a synonym of Gruner persicum Mill. Bailey, appar-
ently following Pax and Knuth, has adopted Cyclamen indicum L. as the valid name

plant | is very different from Cyclamen persicum Mill., and the difference is so pro-
nounced that Cyclamen indicum L. may ultimately prove not to be a Cyclamen at all.
{ In Gard. Bull. Str. SettIm. 11: 1. 1939, Furtado has introduced such other unfamiliar
terms as “priorable”’ and “impriorable” names, “typonymous formal homonyms”
(I. c. 27), etc.!!

74 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

A few individual members also suggested that the practice referred to
above, although perfectly legitimate, should be discouraged in future.
There are two grounds for abandoning this practice, viz. (i) our present
conception of a species has greatly changed from that of Linnaeus and
other earlier authors. By supplying a binomial and referring to a descrip-
tion in pre-1753 literature, we indirectly admit that our conception of
the species is the same as that of these early authors; (ii) the main object
of publication is to clarify and amplify existing knowledge. The descrip-
tion of a plant should present a clearer idea of the species than could be
had by reference to some of these early works. Furthermore, these works
are becoming very rare; hence many herbariums do not have them. It is
therefore desirable to give a description to validate new binomials. On
these two considerations, I suggest the following recommendation to
Art. 36:—

From 1950 onward, botanists are recommended to discontinue the prac-
tice of validating new binomials solely by reference to descriptions or
plates in pre-1753 literature. This modification shall not affect binomials
which are already published and validated by such reference.

Roya Botanic GARDENS,
Kew, ENGLAND.

1949] GREENWOOD, FIJIAN WEEDS 75

NOTES ON SOME FIJIAN WEEDS AND
INTRODUCED PLANTS

WILLIAM GREENWOOD

A NUMBER of interesting weeds and introduced plants have been found
in Fiji in regions other than the leeward coasts and so were mentioned
only briefly in my previous articles on this general subject (Proc. Linn.
Soc. 154: 92-106. 1943; Jour. Arnold Arb. 25: 397-405. 1944). In the
present article notes on some of these plants are given. Some records: of
weeds common to both the dry and wet zones in Fiji are also given, but
no attempt has been made to list all the weeds noticed or collected during
various trips outside the dry zones. Some hitherto unpublished records
for Fiji are also included. In most such cases the records are based upon
specimens which are deposited in the herbarium of either the Royal Botanic
Gardens, Kew, the Arnold Arboretum, or the Gray Herbarium.

In the articles mentioned above, published in 1943 and 1944, about 80
records new to Fiji were published; A. C. Smith has also listed new records
of weeds and introduced plants in Sargentia 1: 1-148. 1942, and in Bull.
Torrey Bot. Club 70: 533-549. 1943. From these records one might con-
clude that the number of weeds and introduced plants in Fiji is increasing
rapidly, but I do not believe this to be the case. Many of the new records
of recent years are from southeastern Viti Levu in the coastal portions of
Serua, Namosi, Rewa, Naitasiri, and Tailevu Provinces, especially near the
towns of Navua, Nasinu, and Nausori. Although the plants discussed
may have been established for years, these districts have not previously
been visited by anyone sufficiently interested in the weeds to collect and
identify them and to publish the records. 3

In this paper families are discussed in the order of Bentham & Hooker’s
Genera Plantarum. Collection numbers found in the text italicized and
in parentheses refer to the writer’s specimens; these numbers are inserted
only when the species has not previously been reported from Fiji. Some
of the plants discussed were identified by members of the staff of the Royal
Botanic Gardens, Kew, to whom I.am grateful. Dr. E. D. Merrill and
Dr. A. C. Smith, of the Arnold Arboretum, have identified most of the
new records and have assisted me in the preparation of this article.

PORTULACACEAE
Talinum paniculatum (Jacq.) Gaertn.

In Bull. Torrey Bot. Club 70: 537. 1943, Smith states that this species
has been recorded, in Fiji, only from the island of Ovalau. During 30
years’ collecting in Fiji I have never seen it except in this one locality,
near Levuka. Why this weed should have failed to spread, after its intro-
duction into Fiji at least 80 years ago, and why it should have persisted
around Levuka are interesting questions.

76 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

MALVACEAE
Sida acuta Burm. f.

This species has increased rapidly in recent years in Mba and Ra Prov-
inces, Viti Levu, especially in open hilly country and sometimes to the
exclusion of everything else. What would appear to be the earliest collec-
tion of this plant in Fiji is Greenwood 99, collected at Lautoka, Viti Levu,
and identified and deposited at the Royal Botanic Gardens, Kew.

LEGUMINOSAE

Albizzia Lebbeck (L.) Benth.

Vanua Mbalavu; collected by Mrs. J. D. Tothill.
Acacia sp.

Near Varoko, Mba Province, Viti Levu (7782). This is apparently a
recent arrival in the Colony and, as far as I know, occurs only at this one
locality, where efforts are being made to eradicate it before it spreads. It
grows to 12 feet high and is armed with strong spines.

Desmanthus virgatus (L.) Willd.

Lautoka, Mba Province, Viti Levu, near sea-level.
Cassia Leschenaultiana DC.

Near Tawarau, road from Lautoka to Rarawai, Mba Province, Viti Levu
(1183); also near Lautoka, alt. about 180 m. (1183A). This species
attains a height of 5 feet and appears to prefer well-drained, sloping ground.
Desmodium triflorum (L.) DC.

Near Nasinu, Naitasiri Province, Viti Levu. Also seen in several places
in Tailevu Province but not collected there.

Alysicarpus vaginalis (L.) DC.

Near Nasinu, Naitasiri Province, Viti Levu.
Lourea vespertilionis (L.f.) Desv.

Mango and Lakemba Islands. The only previous Fijian record (Green-
wood in Proc. Linn. Soc. 154: 96. 1943) is from Penang, Ra Province, Viti
Levu. The specimens from Mango and Lakemba were collected by Mrs.
J. D. Tothill and identified at Kew.

ROSACEAE
Chrysobalanus Icaco L.

Fairly plentiful in parts of Naitasiri Province, Viti Levu, especially
between Nasinu and Nausori; also just above high- water mark in sand,
Naitonitoni Beach, Serua Province, Viti Levu. This is one of the plants
which A. de Candolle considered as probably being spread by ocean cur-
rents (Hemsley, Bot. Voy. Challenger 1 (4): 279. 1885).

CRASSULACEAE

Bryophyllum pinnatum (Lam.) Kurz.

On flat top of large rock near creek, alt. about 1800 ft., western slopes of
Mt. Evans Range, Mba [formerly Lautoka] Province, Viti Levu; no
specimen collected.

1949] GREENWOOD, FIJIAN WEEDS Cue

MYRTACEAE
Psidium littorale Raddi.

Near sea-level on flat land between the Navua River and the Serua hills,
Serua Province, Viti Levu. Since first recording this species, in Jour.
Arnold Arb. 25: 397. 1944, I have heard that it was originally planted
in the mentioned locality. However, it was quite naturalized when col-
lected there in 1943.

TURNERACEAE
Turnera ulmifolia L.

Mango Island: R. J. Lever. Mr. Lever, Government Entomologist,
ee this material in 1945; he Seed me that he had also seen this
species at Levuka, Ovalau. The only other See ues record is from
Thakaundrove on Vanua Levu (Sargentia 1: 64. 1942)

CUCURBITACEAE
Momordica Charantia L.

Waste places, Navua, Serua Province, Viti Levu.
Coccinea cordifolia (L.) Cogn.

First recorded from Fiji in Jour. Arnold Arb. 25: 400. 1944, from the
Lautoka district. Although it has not yet been noticed elsewhere, this
species is starting to spread in the vicinity of Lautoka, Mba Province, Viti
Levu. It prefers open sunny places, where it covers shrubs and small
trees. It may become a pest in the cane-fields by smothering the cane or
causing it to fall, as do Quamoclit coccinea (L.) Moench and Q. pennata
(Desr.) Boj.

CACTACEAE
Opuntia vulgaris Mill.

Seashore near Tavua, Mba Province, Viti Levu. Recorded by Mr. B.
E. V. Parham in Agric. Jour. Fiji 18: 56. 1947, where it is stated that all
existing stands have been eradicated.

RUBIACEAE
Borreria laevis (Lam.) Griseb.

In damp open grassland at Nandarivatu, Mba [formerly Tholo North]
Province, Viti Levu, alt. about 2700 ft.; also at Nasinu, Naitasiri Province,
Viti Levu.

COMPOSITAE
Synedrella nodiflora (L.) Gaertn.
Navua, Serua Province, and also near Nasinu, Naitasiri Province, Viti
evu
Cosmos caudatus H. B. K.

Near Nasinu, Naitasiri Province, Viti Levu.
Tithonia diversifolia (Hemsl.) A. Gray

Waste places, Navua, Serua Province, Viti Levu.

78 JOURNAL OF THE ARNOLD ARBORETUM [VoL. xxx

Erigeron pusillus Nutt.

In coral sand in depressions just behind the coast near Saweni, near
Lautoka, Mba Province, Viti Levu; on open steep grassy western slopes
of Mt. Evans Range, alt. about 2500 ft., Mba [formerly Lautoka] Prov-
ince, Viti Levu.

Erigeron floribundus (H. B. K.) Sch.-Bip.

In Proc. Linn. Soc. 154: 99. 1943, I recorded this plant as growing up
to 1000 ft. elevation. Actually it occurs up to 2700 ft. and is very common
at about 2000 ft. on ground used by the Fijians for crops in the Lautoka
section of Mba Province, Viti Levu.

Elephantopus mollis H. B. K.

Nandarivatu, Mba [formerly Tholo North] Province, Viti Levu, alt.
about 3000 ft. In open sunny places and also in thick forest wherever a
clearing occurs or the track is wide enough to admit a certain amount of
sunlight.

Distreptus spicatus (Aubl.) Cass.

Sandy soils near coast, Lautoka, Mba Province, Viti Levu (1132).
Spilanthes acmella (L.) Murr.

Along roadside on lower slopes, Serua hills, Serua Province, Viti Levu;
waste land near Nausori, Tailevu Province, Viti Levu. The only previous
record for Fiji (Sargentia 1: 141. 1942) is from Thakaundrove Province,
Vanua Levu.

Taraxacum officinale Weber.

Open grassland, Nandarivatu, Mba [formerly Tholo North] Province,
Viti Levu (1172).

Struchium sparganophorum (L.) Kuntze.

In low-lying wet places near Nasinu, alt. about 100 ft., Naitasiri Prov-
ince, Viti Levu (1107).

LOBELIACEAE

Isotoma longiflora (L.) Presl.

Damp places in Rewa, Tailevu, and Naitasiri Provinces, Viti Levu; also
at Nandarivatu, alt. about 2700 ft., Mba [formerly Tholo North] Province,
Viti Levu.

Lobelia zeylanica L.

Damp shady places, near road hills between Navua and Suva in Namosi
Province, Viti Levu; shady places, Serua hills, Serua Province, Viti Levu;
in wet places, sometimes in standing water, near Nasinu, Naitasiri Province,
Viti Levu. When not in flower, this plant somewhat resembles Cardamine
sarmentosa Forst., which also grows in water and other very wet places and
is gathered and eaten as a water-cress in Fiji.

APOCYNACEAE
Allamanda cathartica L.
Open sunny places near Nasinu, alt. about 100 ft., Naitasiri Province,

1949] GREENWOOD, FIJIAN WEEDS 79

Viti Levu (1195). This commonly cultivated garden plant was found
covering several acres near Nasinu, where there had evidently been a
house at one time. It was covering the clearing and climbing over small
trees near by, appearing quite naturalized.

CONVOLVULACEAE
Ipomoea obscura (L.) Ker.

In shady places near sea-level, Lautoka, Mba Province, Viti Levu (821).
First noticed by me about 1935, this species thus far shows no tendency to
spread or to become a bad weed.

Cuscuta australis R. Br.

The Director, Royal Botanic Gardens, Kew, informs me (in litt.) that
the record of C. densiflora Hook. f. given by Miss Gibbs from Nandarivatu
refers to this species.

SOLANACEAE
Browallia americana L.

Near Nasinu, Naitasiri Province, Viti Levu (1116). A garden escape,
found semi-naturalized about European houses in several places in Naita-
siri and Tailevu Provinces. It was also seen as a weed in gardens at
See alt. about 2700 ft., Mba [formerly Tholo North] Province,
Viti L

SCROPHULARIACEAE1

Scoparia dulcis L.

Near Nasinu, Naitasiri Province, and near Nausori, Tailevu Province,
Viti Levu.

Bonnaya veronicaefolia (Retz.) Spreng.

Nausori, in damp places, Tailevu Province, Viti Levu; collected by Mr.
R. Veitch (330). I am indebted to the Director, Royal Botanic Gardens,
Kew, for the identification of this plant, which was collected in May, 1921.
I did not see it during a visit to Nausori a few years ago, and the record
was overlooked when I prepared the previous weed articles.

Lindernia anagallis (Burm.f.) Pennell.

Near Nasinu, Naitasiri Province, Viti Levu, in wet places. The only
other record for a (Jour. Arnold Arb. 25: 397. 1944) is from Serua
Province, Viti Lev
Lindernia diffusa (L.) Wettst.

Covering wet banks at side of road, hills between Navua and Suva in
Namosi Province, Viti Levu.

Lindernia microcalyx Pennell & Stehlé.

In wet places and sometimes in standing water, near Nasinu, Naitasiri
Province, Viti Levu (1110).

1 Dr. F. W. Pennell, of the Academy of Natural Sciences, Philadelphia, has kindly
anaes several of the identifications listed in this family.

80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Torenia polygonoides Benth.

Damp shady places near Nasinu, Naitasiri Province, Viti Levu (1099).
Mazus japonicus (Thunb.) Kuntze.

Damp shady banks near Nausori, Tailevu Province, Viti Levu (1105).
Angelonia angustifolia Benth.

Damp open sunny places between Mulamula and Yavuna, alt. about
500 ft., Mba [formerly Nandi] Province, Viti Levu; open hillsides between
Lautoka and Mt. Evans, alt. about 1800 ft., Mba [formerly Lautoka]
Province, Viti Levu (25). This plant is commonly grown in gardens in
Fiji by Europeans, Fijians, and Indians. Greenwood 25 was collected
in 1920 and identified at Kew, but I did not see the plant again in this
locality and therefore did not include it in my articles on the weeds and
adventive flora of the leeward coasts. However, in June, 1947, I observed
the species flowering well and apparently naturalized at two places several
miles apart inland from Nandi, and therefore it should doubtless be re-
corded here.

ACANTHACEAE
Asystasia gangetica (L.) T. Anders.

Near Nasinu, Naitasiri Province, Viti Levu. First noticed in Fiji at
Levuka, Ovalau, in 1927.

Thunbergia alata Boj.

Near Thuvu, Nandronga & Navosa Province, Viti Levu. Commonly
called Black-eyed Susan, this creeper has a corolla of which the throat is
sometimes black and sometimes not.

Blechum pyramidatum (Lam.) Urb. [B. Brownei (Sw.) Juss.]

Near Nasinu, Naitasiri Province, Viti Levu.

LABIATAE
Hyptis pectinata (L.) Poit.

Open places on hills, road from Navua to Suva in Namosi Province, Viti
Levu. During a trip around Viti Levu by car this species was not noticed
in Serua Province, but it occurs from the cited locality in Namosi Province
through Rewa, Naitasiri, and Tailevu Provinces where the road passes.
During such a trip in 1945 the northernmost plants along the road were
seen about two miles south of the turn-off to Ellington Wharf, Ra Province.
Salvia coccinea Juss. ex Murr.

Listed by Gibbs (in Jour. Linn. Soc. Bot. 39: 161. 1909) from the
Nandrau Valley, alt. about 1000 ft., Nandronga & Navosa [formerly Tholo
North] Province, Viti Levu. The species does not otherwise appear to
have been recorded from Fiji.

PLANTAGINACEAE
Plantago major L.
Nandarivatu, alt. about 2700 ft., Mba [formerly Tholo North] Prov-
ince, Viti Levu. Seemann, in Flora Vitiensis, notes this species as common,

1949] GREENWOOD, FIJIAN WEEDS 81

but during the past 30 years I have seen it only at Nandarivatu and Suva.
It has not been noticed on the leeward coasts of Viti Levu or Vanua Levu.

AMARANTHACEAE
Iresine Herbstii Hook.
This is another species listed by Gibbs (in Jour. Linn. Soc. Bot. 39: 161.
1909) from the general vicinity of Nandarivatu, Viti Levu, which appears
not to have been collected — or at least not recorded — since her report.

CHENOPODIACEAE
Chenopodium ambrosioides L.

Nandi, Mba, [formerly Nandi] Province, Viti Levu. Mr. B. E. V.
Parham has published an interesting note on this species in Agric. Jour.
Fiji 17: 24. 1946. Here the plant is noted as a new record for Fiji, but
it was earlier listed by Horne (A Year in Fiji, 259. 1881) without dis-
cussion. At present it appears to be confined to the Nandi district.

PIPERACEAE
Peperomia: pellucida (L.) H. B. K.
In shady places near Nasinu, Naitasiri Province, Viti Levu.

EUPHORBIACEAE

Phyllanthus urinaria L.

Near Navua, Serua Province, Nasinu, Naitasiri Province, and Nausori,
Tailevu Province, Viti Levu. Formerly recorded from Nandarivatu, in
Sargentia 1: 46. 1942.

Phyllanthus Niruri L.

Near Navua, Serua Province, Viti Levu.
Euphorbia Peplus L.

Nandarivatu, alt. about 2700 ft., Mba [formerly Tholo North] Prov-
ince, Viti Levu (1171). A weed in gardens and on cultivated ground.

URTICACEAE
Pilea microphylla (L.) Liebm.
Along roadside in shade, hills between Navua and Suva, Namosi Proy-
ince, Viti Levu; Nandarivatu, alt. about 2700 ft., Mba [formerly Tholo
North] Province, Viti Levu; Lautoka, near sea-level, Mba Province, Viti

SALICACEAE
Salix sp.
On bank of sand and stones at edge of river a few miles below Tum-
benasolo, Mba [formerly Nandi] Province, Viti Levu (1191). The family
has apparently not previously been recorded from Fiji.

IRIDACEAE
Sisyrinchium micranthum Cav.

F. M. Bailey (Weeds and suspected poisonous plants of Queensland,

82 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

192. 1907) states that this species is believed to be poisonous and that it
causes violent scouring in stock. The plant is very small and, when not in
flower, is difficult to discern among grasses. It is recorded thus far in Fiji
only from the vicinity of Nandarivatu (in Bull. Torrey Bot. Club 70: 535.
1943), but it may be expected to spread over some coastal pastures by
means of flood waters.
AMARYLLIDACEAE

Agave sisalina Perr.

Open hillsides at about 500 ft., near Thuvu, Nandronga & Navosa Prov-
ince, Viti Levu.

PONTEDERIACEAE
Monochoria hastata (L.) Solms.

Mr. B. E. V. Parham records this plant (as M. hastaefolia) for the first
time from Fiji, in Agric. Jour. Fiji 18: 39. 1947. He states that it occurs
throughout the Navua district, in all parts of the lower Rewa, and in rice
fields between Samabula and Suva Point. It was pointed out to me by
Mr. Parham at Nanduruloulou and I also saw it near Nasinu (both in
Naitasiri Province, Viti Levu), but specimens were not collected.

CYPERACEAE
Cyperus Iria L.

Lautoka, Mba [formerly Lautoka] Province (806), and Wangandra,
Mba [formerly Nandi] Province (806A), Viti Levu. This species was
first noticed by me at Wangandra in 1932 and near Lautoka in 1940. It is
a bad weed in rice fields in the Lautoka and Nandi districts, springing up
and dying down during the wet season. It has been seen at about 1000 it.
elevation on hills near Lautoka.

Cyperus difformis L.

In standing water near Nasinu, Naitasiri Province, Viti Levu.
Cyperus compressus L.

The annual form of this, mentioned in Proc. Linn. Soc. 154: 105. 1943,
appears to be confined in Fiji to the vicinity of Lautoka, in my observa-
tion. A perennial form, up to 10 or 12 inches high, is found in Mba
[Lautoka], Nandronga & Navosa, Naitasiri, and Tailevu Provinces, Viti
Levu, usually in damp or wet grassland.

Cyperus alternifolius subsp. flabelliformis (Rottb.) Kiikenth.

Banks of Navua River near Navua, Serua Province, Viti Levu.
Cyperus Haspan L.

Common in drains and other wet places in Naitasiri and Tailevu Prov-
inces, Viti Levu.

Cyperus distans L. f.
Near Nausori, Tailevu Province, Viti Levu.
Scirpus Purshianus Fernald [S. debilis Pursh, non Lam.]
In standing water in drains near Navua, Serua Province, Viti Levu

1949] GREENWOOD, FIJIAN WEEDS 83

(980). During a month’s visit at Navua this species was observed at only
one place and was not plentiful there.
Eleocharis geniculata (L.) R. & S.

In coral sand in depressions ies: behind the coast near Saweni, near
Lautoka, Mba Province, Viti L
Fimbristylis miliacea (L.) Vahl.

In standing water in drains, near Nasinu, Naitasiri Province, Viti Levu.

GRAMINEAE2
Ischaemum aristatum L. [J. ciliare Retz.].

Near Nasinu, Naitasiri Province, Viti Levu.
Ischaemum timorense Kunth.

The record of this species in Fiji, noted in Bull. Torrey Bot. Club 70:
534. 1943, was based on Greenwood 817, collected in 1939 from roadside
on the Navua flats between Navua and Suva in Serua [not Rewa as pre-
viously stated] Province, Viti Levu. In May, 1943, the species was col-
lected all along these flats, on the banks of the Navua River (Serua
Province), and also on the hills along the road from Navua to Suva in
Namosi Province.

Ischaemum rugosum var. distachyum (Cav.) Merr.

In soak area near creek in open, western slopes of Mt. Evans Range,
Mba [formerly Lautoka] Province, Viti Levu (1038B).

Sorghum halepense (L.) Pers.

Lambasa district, Mathuata Province, Vanua Levu. This record should
have been included in my earlier articles but it was overlooked. The
species has been present near Lambasa for over 30 years.

Themeda quadrivalvis (L.) Kuntze.

Roadside between Lautoka and Rarawai, Mba Province; hills at about
1200 ft. alt., between Lautoka and Mt. Evans Range, Mba [formerly
Lautoka] Province, Viti Levu.

Brachiaria subquadriparia (Trin.) Hitchc.

Damp places near sea-level, Lautoka, Mba. Province (934), and near
Nausori, Tailevu Province (93B), Viti Levu. This species has been
combined by some botanists with B. distachya Stapf; I am indebted to
Mrs. Chase for pointing out the differences. She writes: “It is related to
B. distachya Stapf but has more numerous racemes and slightly longer
spikelets. It is more frequent in the islands than is B. distachya, to which
it has commonly been referred.”” Much of the material listed by Summer-
hayes & Hubbard (Kew Bull. 1927: 32; 1930: 256) under B. distachya
will prove to represent B. subquadriparia. Apparently B. distachya prefers
damper places than B. subquadriparia and has a more compact growth of
a lighter green color.

2 For some of the listed identifications of grasses I am indebted to Mrs. Agnes
Chase and Mr. J. R. Swallen, of the U. S. National Herbarium.

84. JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Brachiaria erucaeformis (J. E. Sm.) Griseb.

On banks of sand and stones along creek in open, alt. about 200 ft.,
Lautoka, Mba Province, Viti Levu (1081, depauperate). Previously
known, in the Pacific, from only the Philippines and Guam, according to
Mrs. Chase.

Axonopus compressus (Sw.) Beauv.

Fairly common near Navua, Serua Province, near Nasinu, Naitasiri
Province, and near Nausori, Tailevu Province, Viti Levu.

Echinochloa stagnina (Retz.) Beauv.

In drains, near Nausori, Tailevu Province, Viti Levu. The only previous
record from Fiji (Jour. Arnold Arb. 25: 397. 1944) was from drains near
Navua, Serua Province, Viti Levu.

Cyrtococcum oxyphyllum (Hochst. ex Steud.) Stapf.

In open places in forest along road between Navua and Suva, Namosi
Province, and in shady places near Nasinu, Naitasiri Province, Viti Levu.
Setaria barbata Kunth.

Levuka, Ovalau, in shady places. Although the species was apparently
quite common at Levuka when I was there in 1932, it has been previously
recorded, in Fiji, only from Koro (Kew Bull. 1930: 260).

Setaria geniculata (Lam.) Beauv.

On open flat rock at extreme western end of Mt. Evans Range, alt.
about 3400 ft., Mba Province, Viti Levu (962, 962A). The only previous
record of this species in Fiji (in Sargentia 1: 6. 1942) is based on
Degener & Ordonez 13514, which I believe to represent S. pallidifusca
(Schumacher) Stapf & Hubbard, a tall annual grass; S. geniculata is
perennial.

Setaria pallidifusca (Schumacher) Stapf & Hubbard.

Near sea-level near Navua, Serua Province, Viti Levu.
Isachne dispar Trin.

In wet ground in open beside creek, alt. about 2700 ft., Nandarivatu,
Mba [formerly Tholo North] Province, Viti Levu (1178).

Eragrostis pilosa (L.) Beauv.

Navai, near Nandarivatu, alt. about 2500 ft., Mba [formerly Tholo
North] Province, Viti Levu.

Arundo Donax L.

Along road, hills between Navua and Suva, Namosi Province, Viti Levu;
along creek, lower western slopes of Mt. Evans Range, alt. about 1800
ft., Mba [formerly Lautoka] Province, Viti Levu.

CoLontaL SuGAR REFINING Co.,
Lautoka, Viti Levu, Fiji.

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 85

STUDIES IN THE BORAGINACEAE, XVII
Ivan M. JoHNSTON

A. CORDIA SECTION VARRONIA IN MEXICO AND
CENTRAL AMERICA

WHILE PREPARING AN ACCOUNT of the Boraginaceae for the southern
West Indies it became necessary to typify the Cordias proposed by Lin-
naeus and also to re-examine the definition of certain widely ranging
species. The results of these investigations came to involve a goodly pro-
portion of the Varronias growing in Mexico and Central America. Since
a discussion of them would be out of place in a paper dealing with the
Antillean flora, they have been brought together and, with other observa-
tions, presented as a preliminary synopsis of the group for the region
mentioned. Some of the species, e.g., Cordia oaxacana, C. Pringlei, C.
coyucana, C. curassavica, and C. spinescens, have been treated cursorily
and conservatively. They need much further study.

A few words are needed regarding the history, definition, and typifica-
tion of Varronia. The name was coined and first used by Browne, Hist.
Jam. 172 (1756), in describing Jamaican plants later called C. globosa
and C. curassavica. Linnaeus was the first to use it in a strictly binomial
classification, Syst. ed. 10, 916 (1759). He referred two species to it,
Varronia lineata L. and V. bullata L. These two, both treated in the
present paper, obviously belong to Varronia as usually defined. The group
is, however, only a well-marked section of Cordia. It is confined to Amer-
ica and is best distinguished by fruiting structures. The fruit is usually
matured in a cup-shaped calyx and is drupaceous. When ripe it has a
conspicuous, thin, fleshy, usually cherry-red mesocarp which shrinks and
fades and seems to disappear when the fruit is dried. The bony endocarp
at maturity is usually single-seeded, irregularly ellipsoidal, and generally
coarsely and irregularly tuberculate. The corolla in most Varronias has
lobes much broader than long. The inflorescence is an open cyme in some
species, but in most is either a dense head or a spike that matures its
terminal flower first. In habit the plants are shrubs, woody scrambling
climbers, or small trees. Many of them produce tiny aromatic granules
on the herbage that give them a characteristic sagy odor. Cordia parvi-
folia DC. (C. Greggit Torr.; C. Watsonit Rose), although placed in Var-
ronia by some authors, is accordingly here excluded. Among Mexican
species C. parvifolia is actually most closely related to C. elaeagnoides
DC. Note should be made of the treatment of Varronia by Friesen, Bull.
Soc. Bot. Genéve sér. 2, 24: 115-201 (1933). That author not only treats
Varronia as a genus, but has even broken it up into a number of segregate

86 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

genera. Personally I am entirely unable to find any justification for
treating the Mexican and Central American species under four distinct
genera, Varronia, Montjolya, Ulmarronia, and Cordiopsis, as Friesen
proposes.
KEY TO SPECIES

_ ce 1-3 cm. long, pure white, salverform or funnelfor

clothed with abundant spreading hairs; tip of ae ee ae
ae filiform, 2-5 mm. long.......... 02. cece eee ees C. oaxa
Calyx clothed with ae appressed hairs, tip of lobes a or not “1 “al
. lon

prolonged (less than 2 m ng).
Corolla 2-3 cm. long, ae a slender elongate tube much exserted from the
calyx; leaves ovate to ea 2-3 times as long as puts calyx-lobes
prolonged into free tips 1- ONS wes ladia ated foe C. limicola.

Corolla 1-2 cm. long, its oe tube short and barely ee from the
calyx; leaves lanceolate, 3-5 times as long as broad; ae lobes Bie tips
carcely: DYOlONLEd b.4orhs ia seid awd epee eis C. podocephala.

Corolla much smaller
Foliage clothed nah stellate hairs
Calyx-lobes with well- developed very elongate linear tips 2-5 mm. in length ;
Cc.

heads in bud 2-3 cm. in diameter...........--2+-+ ee eeeee 4, Pringlet.
Calyx-lobes with eee or scarcely if at all developed - ps iar
in bud 10-18 mm. in diameter......... 0... eee eee ee eee eee C. stellata.

Foliage bearing simple hairs rs
Flowers borne in a forking cyme
Leaves firm, usually broadest at or above middle, the upper surface dis-
tinctly scabrous from an abundance of short stiff appressed hairs;
calyx-lobes abruptly short-acuminate, a apices free and forming
a short but ar tip on the unopened calyx.......... 6. C. foliosa.
Leaves not very firm, usually broadest < or below middle, the upper
surface ae only widely scattered strongly appressed hairs; calyx-
lobes acute; unopened calyx with unappendaged apex.............
7

Laan hone une ace ane asthe wastegate a ao ene oe aes C. bifurcata
Flowers borne in heads or spikes.
Flower-clusters capitate.
Peduncles aa ANALY 528s ood ae lee eee 8. C. lineata.

Pedande: never axillary, but terminal or internodal.
Calyx-lobes with apex acute or with prolonged tips only 1 mm.
long or less
Heads ellipsoidal, somewhat longer than broad, 10-15 mm.
ick; mature fruit enclosed by a loose papery —
calyx and ae freed by the break- t

LALCOR cores Sad eon anaes Se eee toe C. ee
Heads globose, not longer than broad, mostly 10 mm. or
less thick

Calyx at anthesis 4-6 mm. long, glabrous or with only

scattered relatively coarse appress sed hairs; calyx-
lobes distinctly attenuate, the tips 0.5- ong,
free and usually spreading in the bud; corolla flaring,
about 5 - leaves lance-ovate, ovate, or ellip-
tic, 2-6 broad, upper surface with branches of
veins ee evident...........04- bigua
Calyx at anthesis 1.5 . long, clothed with abun-

t
mm. long; leaves usually lanceolate or ob-

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 87

lanceolate and 1-2 cm. broad but ee be-

ame lance-ovate to elliptic and up to 5 cm. broad,
upper surface with branches of veins ee if at
alleviden Gis... veaoun ia wanna ll. C. inermis.

Calyx- eae with prolonged free linear tips 2-4 mm. long.

Leaves firm, thickish, with sharp teeth and a sharply acute
pet lower surface with a prominent reticulum of re-
peatedly branched and anastomosing veins, and with one
or more small, well-defined concave or obconic depres-

elongating, usually 2c ong; corolla 4—
long, usually hardly St DASeIne the ge the tube ex-
anding into a subcampanulate throat...12. C. bullata.

pee thinner, usually with rounded teeth and obtusish at

Bee 1 a.nd ces ep st, RS ae
Flower-clusters elongating, spikes.

f petiole not uae on the subtended axillary shoot, at

Bere only the basal 1-2 m. hardening and persisting on defoli-

ge biancies: plant | an erect shrub bearing no axillary eee

internodes along leafy stems 5) culls Senate enone Cc CUYASSAVICA.
Base of petiole decurrent for 5-10 mm. on the subtended axillary

along leafy stems, only those first produced by the stem being
terminal.

Calyx-lobes with short free tips forming a short but distinct
apical prolongation on the unopened calyx; spikes borne in
terminal and axillary panicles.......... 15. C. costaricensis.

Calyx-lobes without free tips, the unopened calyx with a
rounded unappendaged apex; spikes rarely paniculate, most
of them springing singly from the leaf- 2a aoe the stems.

F Ruinescens:

1. Cordia oaxacana DC. Prodr. 9: 497 (1845).—near Tololapa, Oaxaca, Andrieux
203.
Varronia oaxacana (DC.) Friesen, Bull. Soc. Bot. Genéve sér. 2, 24: 175 (1933).
pikes peas ata Shee Field Mus. Pub. Bot. 2: 338 Gigi) s—=Garon de
, Oaxaca, Conzatti 2218.
Varrenis Aiea var. sppendicuat (Greenm.) Friesen, Bull. Soc. Bot. Genéve
, LO Mt lef:
Cordja Realapye Fernald, ae oc. ne Acad. 33: 90 (1897).— Acapulco, Palmer 70.
Cordia urticacea Standley, Contr, U. S. Nat. Herb. 23: 1222 (1924).— Real de
uad

Tey urticacea (Standl.) Friesen, Bull, Soc. Bot. Geneve sér. 2, 24: 174 (1933).

88 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

MEXICO: Aguas Calientes, 1837, Hartweg 181 (G); Zacuapan, Vera Cruz, Pur-
pus 8477 (G, US); Lake Chapala, Jalisco, Hitchcock & Stanford 7174 (G, US);
Zitacuaro to Coyota, Michoacan, ae 3 m., Hinton 13153 (G, US); Acapulco,
Guerrero, 1895, Palmer 70 (G, Type of C. perlonga; US); San Antonio, Guerrero,
shrub 15-2 m., Hinton 10256 (G, US) and 10542 (G, US); Real de Guadalupe,
Guerrero, shrub 2 m., Langlassé 355 (US, Type); Picacho to San Geronimo, Oaxaca,
Purpus 6692 (G, US); Rio de los Vueltos, Oaxaca, Liebmann 15158 (G); Cuesta de
Quiotepec, Oaxaca, Smith 700 (G) and Conszatti 4027% (G, US); Tomellin Canyon,
Oaxaca, 3-4.5 m. tall, Pringle 4630 (G, US); between San Geronimo and La Venta,
Oaxaca, Nelson 2785 (G, US); San Geronimo, Oaxaca, Mell 2133 (US); Ixtepec,
Oaxaca, Fisher 35250 i Salina Cruz, Oaxaca, Fisher 35492 (US) ; Tehuantepec,
Oaxaca, Matuda 244 (U

The species is ae to southern Mexico and appears to be very
variable. Specimens referred to it are rather diverse in appearance, but
until more collections of it accumulate and more is learned of its behavior
in the wild, I believe it had best be given the broad definition here
accepted. From plant to plant the corolla may be moderate-sized or large
and differ greatly in the length of corolla-tube protruded from the calyx.
These differences, however, may possibly be associated with heterostyly.
Leaf-form is variable, and so also is the type of hairs on twigs and foliage.
The leaves are usually ovate or elliptic and rounded or obtuse at the base.
The plant described as C. perlonga, however, has very elongate leaves that
are acute or attenuate at the base and very scantily pubescent beneath.
Its peduncles, also, are unusually elongate. The plant described as C.
urticacea has the twigs and peduncles bristly with spreading brownish
hairs. Hinton’s two collections from Guerrero (nos. 10256 and 10542)
show that it is not sharply distinct from other forms of the species.

2. Cordia limicola Brandeg. Univ. Cal. Pub. Bot. 6: 502 (1919).— Corral de
Piedras, Vera Cruz, Purpus 8017
Varronia limicola (Brandeg.) Friesen, Bull. Soc. Bot. Geneve sér. 2, 24: 148 (1933).
Cordia passa Johnston, Jour. Arnold Arb. 29: 227 (1948).— Jacala, Hidalgo,
Chase 7461.
MEXICO: near San Lucas, Jaumay ve, Tamaulipas, Rozyuski 563 (G); Jacalay

Zacuapan, Vera Cruz, Pupils 8017 (G, US, tsotype of C. limicola) and Purpus 7491
G, US); Mirador, Vera Cruz, 1857, Mohr (US); Rio Talva, Aug. 1842, Liebmann
12718 ( ;

A well-marked species known only from middle eastern Mexico. Its
closest relatives are C. podocephala Torr. of northeastern Mexico and C.
mirabiloides (Jacq.) R. & S. of Cuba and Haiti. Typical C. passa
differs from typical C. limicola only in having smaller, more hairy leaves,
and is obviously only an ecological form of the latter.

3. Cordia podocephala Torr. Bot. Report U. S. & Mex. Bound. Survey 135 (1859).

Varronia longifolia ee & Moc. FI. vee 48 (1893), ed. [ 44 (1894) — Type from

“Praedio de la Pu
TEXAS: Eagle ate — Co., Havard 10 (G); Montell Creek, Uvalde Co.,

9 G);
Falfurrias, Duval Co., Cory 14751 (G); inianco: Bien rich bottom land,” Rever-
chon 1562 (US).

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 89

MEXICO: 11 mi. south of Allende, Coahuila, weak shrub 1-1.5 m. tall, Johnston
7023 (G); Hacienda Mariposa, Coanuila: Wynd & Mueller 240 (G); Musquiz,
Coahuila, Marsh 140 (G); Soledad, southwest of Monclova, Coahuila, 1880, Palmer

4 (G); Cafion Fe Coahuila, Muller 3112 (G); Meee Nuevo Leon,
Phi 1877 (G, US), ee (G, US), and Mueller 322 (G); ctoria, Tamaulipas,

1907, Palmer 453 (G, US); Cerro Tamaulipeca, Tamaulipas, fee 10631 (G, US);
acces Tamaulipas: Viereck 319 (US).

A very well marked species which is closely related only to C. limicola.
It ranges just north of the latter species in Tamaulipas, Nuevo Leon,
Coahuila, and in adjoining Texas.

4. Cordia Pringlei igor Proc. Am. Acad. 26: 169 (1891).—Las Palmas, San
Luis Potosi, Pringle 3
Varronia Bp aes Friesen, Bull. Soc. Bot. Genéve sér. 2, 24: 168 (1933).
Cordia Pringlei var. altatensis Brandeg. Zoe a: 219 (1904).— Yerba Buena near
Altata, Sinaloa, Brandegee
aad asterothrix Killip, Tou Wash. Acad. 17: 330 (1927). — Rio Limon, Vene-
zuela, Curran & Haman 808.
Cordia Storkii Standley, oe pe Mus. Bot. 18: 984 (1938).— Catalina, Guana-
caste, Costa Rica, Stork
The species ranges in i Aa: parts of Mexico, Central America and
northern Colombia and Venezuela.
MEXICO: Victoria, Tamaulipas, pe Fisher (G); Las Palmas, San Luis Potosi,

shrub 3-4.5 m. tall, anae 3091 (G, TYPE of C. Pringlei; US) and 4058 (US); 100
mi. north of Tamazunchale, San Luis ee 1-1.5 m. tall, Hitchcock & Stanford
689% ); State of San Luis ee peer ie (G); eee June 1841,

Redondo west of Caymanero, Sinaloa, 3. eee . tall, (es 7091 (G) and 7091A (G);
Buena Vista, Apatzingan, Michoacan, shrub 4 m., Hinton ae es US)

GUATEMALA: near Estanzuela, dept. Zacapa, 7-9 dm. , Steyermark 29102
(G); between Zacapa and Chiquimula, 2-3 m. tall, Standley ee (G) and 74448
(G); near Jutiapa, 2 m. tall, Standley 74969 (G).

NICARAGUA: Managua, Garnier 1084 (US) ; indefinite, Wright (G, US).

COSTA RICA: Bebedero, Guanacaste, Brenes 12546 os FM); Catalina, Guana-
caste, 5 m. tall, Stork 2758 (FM, Type of C. Storkii); Puerto Jesus, Nicoya, 1903,
(Gita & Doyle 747 (US); Estero de Puerto Jesus, 1903, ae 16690 (US).

BIA: eevee Quebrada de Angeles and Rio Cabrera, dept. Huila, Rusby

& Pennell 333 (G).

VENEZUELA: Rio Limon, Curran & Haman 808 (US, Type of C. asterothrix; G).

The species is most closely related to C. stellata of Mexico and to C.
macrocephala (Desv.) HBK. of Peru and C, polyantha Benth. of Ecuador
and northern Peru.

5. Cordia stellata Greenm. Proc. Am. Acad. 39: 86 (1903).— Cuicatlan, Oaxaca,
mith 147.

MEXICO: Cameron, Vera Cruz, Purpus 8715 (G, US); near San Luis Tultitlanapa,
Puebla, Purpus 3242 (G, US); Cire tlan, Oaxaca, Smith 147 (G, type) and Gonzales
983 (G): Cuesta de Quiotepec, dist. Cuicatlan, Goneait 2461 (G); Cuesta de Coyula,
dist. Cuicatlan, Conzatti 4136 (US); 6 mi. above Dominguillo, Oaxa aca, Nelson 1646
(G, US); Cerros Teotitlan, Oaxaca, Conzatti 3456 (G); Rio Vuelto, Oaxaca, July
1842, Liebmann 12715 (US).

The plant is very closely related to C. Pringlei, but it is always dis-

90 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

tinguishable by its smaller heads and very short or inconspicuous tips on
the calyx lobes.
6. Cordia foliosa oe & Gal. Bull. Acad. Brux. 112: 330 (1844).— Zacuapan, Vera
Cruz, Galeotti 7
Cordia sp a a Proc. Am. Acad. 40: 52 (1904).— Ocuilapa, Chiapas,
Nelson
MEXI ae ae de Piedras, Vera Cruz, Purpus 8062 (G); Zacuapan, Vera Cruz,
Purpus 10785 (G, US) and 10811 (G, US); La Palmilla, Purpus 16410 (G); Mirador,
Vera Cruz, Mobmaie 12703 (G, US); Hacienda de la Laguna, Vera Cruz, Schiede
146 (NY); Ocuilapa, oo Nelson 3003 (G, TYPE of C. chiapensis; US); Monser-
rate, Chiapas, Purpus 14 (U
Puebla viejo Quen Santo, Huehuetenango, Seler 3011 (G); north-
west of Cuilco, Huehuetenango, Steyermark 50806 (G)
A very well marked species and the only Varronia with loose, distinctly
forked cymes known from Mexico and northernmost Central America.

7. Cordia bifurcata R. & S. Syst. 4: 466 (1819).— Based on V. dichotoma R. & P.;
not Cordia dichotoma — (1786
leg dichotoma R. & P. Bee 2: 23, t. 146 (1799).— Type from Chacha-
huasi, Peru.
rent slopes of Costa Rica and Panama and from there south along
the Andes of South America.
COSTA RICA: La Colombiana Farm, prov. Limon, 2-3 m. tall, Standley 36641

(US); Juan Vifias, prov. Cartago, Cook & Doyle 302 (US); Valle Tuis, basin of
Rio Reventazon, prov. Cartago, Tonduz 11353 (G, US); Aguacaliente, prov. Cartago,
Stevens 324 (US); Rio Hondo, plains of Santa Clara, Cook & Doyle 530 (S); San
Jose, Tonduz 11353 (US); Villa Quesada, prov. Alajuela, Austin Smith 2542 (G);
Las Vueltas, Tucurrique, Tonduz 12759 (US).

PANAMA: Almirante, Bocas del Toro, Cooper (G); psn eae Valley,
Bocas del Toro, Dunlap 90 (G); western Panama, ee Stork 16 and 19 (US).

This is one of the previously unrecognized species that has been con-
tained in the complex, passing under the names C. ulmifolia, C. corymbosa,
and C. polycephala. It is the common representative of the complex in
Colombia and southward along the Andes into northern ieee and is
readily delimited except where it approaches the area of its Brazilian
relatives in Paraguay and northeastern Argentina. Though formerly not
distinguished from the other Central American member of the complex,
C. lineata, it is decisively separable and utterly different in appearance.
In C. bifurcata the cymes are all terminal or internodal, and usually very
evidently so. Though at times the flower clusters may be rather dense,
they always show some evidence of being two-lobed and in the very
mature fruiting condition always reveal a forked rhachis to some extent.
Usually, however, the cyme is evidently forked even in bud and is gen-
erally very conspicuously so when in fruiting state. This is all very
different from the condition in C. lineata, where the small, distinctly capi-
tate inflorescences spring from axils along the leafy twigs. The foliage
of the two species also differs. In our present plant the leaves usually

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 91

have a thinner texture than in C. dineata and are conspicuously very much
less hairy, especially on the upper surface.

8. Cordia lineata (L.) R. & S. Syst. 4: 464 (1819).
Varronia lineata L. Syst. ed. 10, 916 (1759); Amoen. Acad. 5: 394 (1759); Sp.
Pl. ed. 2, 275 (1762); Syst. ed. 12, 176 (1767). — Jamaica.
Cordia ulmifolia var. lineata (L.) DC. Prodr. 9: 495 (1845).
Lantana corymbosa L. Sp. Pl. 628 (1753); not eae corymbosa Desv. (1809),
nor Cordia corymbosa Willd. (1819), nor Don (18 — Jamaica.
Ulmarronia corymbosa (L.) eae a Soc. Bot. ee sér. 2, 24: 143 (1933),
as to name- canes synonym
? ? Cordia adnata DC. Prodr. ro ah (1845). — Cuba.
Ranging from southern Mexico to Panama, in the northern West Indies
(Jamaica, Cuba, and Haiti), and possibly also in northern South America.
MEXICO: Zacuapan, Vera Cruz, Purpus 6272 (G, US) and 10825 (US); Jical-
tepec, Vera Cruz, March 1841, Liebmann 12704 (US); Rio Cascabel, Oaxaca, Mell
2303 (US).
BRITISH HONDURAS: San Agustin, E] Cayo Dist., Lundell ee (G, US).
GUATEMALA: Quirigua, dept. Izabel, Siandley 23698 (G,
COSTA RICA: El General, Skutch 3831 (G, US), 4314 (G, ton and 4769 (US).
PANAMA: San Blas Dist., Cooper 276 (G US): Balboa, Standley 25474 (G, US),
26068 (US), and 29245 (G, US); Punta Paitilla, Heriberto 212 (G, US); between
Pacora and Chepo, Woodson, Allen & Seibert 1671 (G); Rio Tapia, Standley 28153
(US)

JAMAICA: St. Margarets Bay, Fredholm 3271 (US); Oxford, near Troy, Harris
9434 (Us): indefinite, Macfadyen (G).

CU Heme Pinar del Rio, Van Hermann 858 (G, US); between Herradura

and ae Real, Shafer 11769 (US); Vinales, Pinar del Rio, Killip 13543 (US); Los
Palacios, Pinar del Rio, Shafer 11652 (US); Pitajones, ee Clara, Shafer 12276
US) ; Sancti Spiritus, sane Clara, Shafer 12111 (US); Cieneguita, Santa Clara,
Combs 420 (G); vicinity of Soledad, Santa Clara, Jack 5483 and 6629 (G, US);
Soledad, Howard 6619 (G); La Gloria, Camaguey, Shafer 119 (G, US); Saltadero,
Oriente, Wright 423 (G); between Lebisa Bay and E] Purio, Oriente, Shafer 3429
(US).

HAITI: St. Michel de l’Ayalaye, Leonard 7519 os hae St. Louis du Nord
Leonard 14289 (US); Terre Neuve, Buch 542 (US); Pilate, iets 9651 (US) :
Plaisance, pes 9173 (US); Gros Morne, |’Artibonite, ee 9852 (US); Massif
du Nord, Port Margot, Bayeux, Ekman 2560 (US); noni of Morne Aaa south-
ern Reais Bartlett 17559 (US

DOMINICAN REPUBLIC: San Jose de las Matas, prov. Santiago, Jimenez 959
(US); banks of Rio Mao to La Ceiba, prov. Monte ane Valeur 474 (US).

VENEZUELA: Perija, Zulia, Tejera 8 (US); between Guamitas and Rancho
Grande, P. N. Aragua, Williams 10368 (US); La Paragua, Bolivar, Killip 37614 (G);
Avila, savana del monte, pee 710 (G).

Co: OMB Boyaca, Rio Meta, Orocue, Los Llanos, Cuatrecasas 4401 (US).

A very well te but previously unrecognized species here associated
with a neglected specific name long ago proposed by Linnaeus. The plant
is known only from Mexico, Central America, and the northern West Indies
and is to be found in herbaria mistakenly identified as C. ulmifolia, C.
corymbosa, or C. polycephala. ‘These latter names have, at one time or
another, been used to cover a complex of species distributed from Mexico
and Cuba south to Argentina. A study has shown that the aggregate can
be broken up into a number of well-defined species. In the West Indies,
Mexico, Central America, and coastal northern South America three species

92 JOURNAL OF THE ARNOLD ARBORETUM [VOL, Xxx

should be segregated from the complex, our present species being one of
them. Cordia polycephala (Lam.) Johnston, sensu str., is another. It
ranges in the West Indies from Santo Domingo through Porto Rico, the
Virgin Islands, the Lesser Antilles, and Trinidad into northern South
America, where it occurs in the Guianas, Venezuela, and coastal northern
Colombia. The third segregate is C. bifurcata R. & S., which ranges
north along the cordillera into Colombia and has outlying stations on the
Atlantic slopes of Panama and Costa Rica. These three species, C. lineata,
C. polycephala, and C. bifurcata have distinct patterns of geographic dis-
tribution and are easily distinguished by differences in inflorescence.

Most of the species of the section Varronia have peduncles that are either
terminal or internodal and never axillary. This normal condition prevails
in C. bifurcata. In C. lineata and C. polycephala s. str., however, while
the first few inflorescences on a long shoot may be terminal or internodal,
the very abundantly produced later ones are all axillary. Their peduncles
arise directly from the leaf-axils along the leafy shoot and have their foot
confluent with the indurated basal portion of the petiole of the subtending
leaf. In C. lineata the peduncles are very slender, 2-3 cm. long, and
terminated by a small dense capitate flower-cluster. Only exceptionally
do they branch or bear a reduced leaf. They are never clustered. In C.
polycephala s. str. the cymes are commonly glomerate and are only rarely
markedly spherical in form. The peduncles are coarser than in C. lineata
and frequently branched. They tend to be especially numerous and even
paniculately arranged towards the ends of the shoot. The corolla of C.
polycephala s. str. averages larger than that in C. lineata and is more
abundantly glandular-hairy inside. The upper surface of its leaves is
minutely tuberculate and only sparingly strigose. In C. lineata the dis-
tinctly lanceolate leaves are less variable in form, dry a darker color, and
average more elongate than in C. polycephala s. str. Their upper surface
is evidently more hairy, since in addition to a coarse strigosity it bears
minute tuberculations, comparable to those already noted in C. polycephala,
which are here prolonged upwards into minute appressed hairs.

In geographic distribution C. lineata and C. polycephala approach each
other only in the Dominican Republic and in western northern South
America. The latter species I have seen from only three general localities
in the Dominican Republic (Paradis, Barahona, Fuertes 998; vicinity of
Ciudad Trujillo, Ekman 12336, Allard 13226 & 15742; and San Pedro
de Macoris, Rose, Fitch & Russell 4162), all near the southern coast of
the country and apart from the known localities in northern and western
parts of the republic where C. lineata has been collected. Judging from
material examined, the two species remain distinct and readily separable
in the West Indies. Only in Venezuela is there any difficulty in distin-
guishing them. In northern South America C. polycephala ranges as far
west as Santa Marta, Colombia (Smith 2737) and is particularly common
and well distributed in Venezuela. In the latter country, unlike its be-
havior in other parts of its range, however, it tends to lose its stability
and is accompanied by a goodly number of variants, some of which are

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 93

separable from C. lineata only with difficulty. The proper interpretation
of these forms must await a detailed study of the Venezuelan Varronias.
The specimens from Colombia and Venezuela cited above can, accordingly,
be now only tentatively identified with typical C. lineata.

The oldest name applied to the present plant is Lantana corymbosa L.
Sp. Pl. 627 (1753). Linnaeus applied the name to Jamaican plants treated
by Sloane, Hist. 2: 83, t. 194, f. 3, and Plukenet, Alm. t. 328, f. 5.
Sloane’s plant most suggests true Cordia polycephala, a plant not known
from Jamaica, but in any case it is probably not the same as the plant
illustrated by Plukenet. Plukenet’s plant is clearly representative of C.
lineata as here accepted. The name Lantana corymbosa L. cannot be
transferred to Cordia, since the resulting combination is already preoccu-
pied by Cordia corymbosa Don, 1838 (based upon Varronia corymbosa
Desv., 1809, a renaming of V. monosperma Jacq.), as well as by Cordia
corymbosa Willd. ex R. & S., 1819. The name “Varronia corymbosa L.,”
Desf. Tableau 71 (1804), possibly may have been based on Lantana
corymbosa L., but it is published merely as a bare name in a list for a
botanic garden and cannot be accepted as a nomenclatorial transfer since
its association with L. corymbosa L. is merely conjecture.

The name Varronia lineata L. Syst. ed. 10, 916 (early 1759) is the
second one applied to our present plant. It was published thus: “lineata.
A. VV. fol. lanceolatis linearis, spicis oblongis. Lantana corymbosa.
Spec. pl. 628. Brown, jam. t. 13. f. 2.’ The concept is a confused one.
The phrase “spicis oblongis” in the diagnosis best describes Cordia
curassavica. Browne’s plate represents Cordia globosa. The reference
to Lantana corymbosa involves two other species, one of which is the
species we now have under consideration.

In Nov. 1859, Linnaeus’s student, Elmgren, Amoen. Acad. 5: 394, gave
under the name Varronia lineata a very good detailed description of our
present species. It was based on a specimen collected in Jamaica by
Browne. , The specimen is now catalogued in the Linnaean Herbarium
as no. 255.1 and has been discussed by Smith, Rees Cyclop. 36: sub
V. lineata (1817)

In Dec. 1859, another student of Linnaeus, Sandmark, Amoen. Acad.
5: 376, applied the name V. lineata to a plant figured in Browne’s History
of Jamaica. Sandmark, by his reference to the second species treated by
Browne on page 172, associated the name with what is probably Cordia
curassavica.

The application of the name Varronia lineata, previously confused, was
finally clarified by Linnaeus in the second edition of his Species Plantarum,
275 (1762), where he treated it as follows: “lineata. 1. VARRONIA
foliis lanceolatis lineatis, Deru lateralibus petiolo adnatis, spicis
globosis. Amoen. acad. 5. p. 394.* Lantana corymbosa, foliis alternis,
floribus corymbosis. Sp. pl. 628. Ulmi angustifoliae facie baccifera
jamaicensis, foliis superne scabris, subtus villosis, floribus flavis perpusillis,
fructu botryoide monospermo. Pluk. alm. 393, t. 328, f. 5. Habitat in
America.” The diagnosis is a new one and applies perfectly to our present

94, JOURNAL OF THE ARNOLD ARBORETUM [voL. XXX

plant. The asterisk calls attention to the full description given by Elm-
gren. Sloane’s plant, cited when Lantana corymbosa was published, is
excluded by both the diagnosis and description. The Plukenet plant is
the Jamaican plant here treated and the one with which Linnaeus is almost
exclusively concerned in the second edition of the Species Plantarum.
In the twelfth edition of the System, p. 176 (1767), the species appears as
follows: “lineata. 1. V. fol. lanceolatis lineatis, pedunc. lateralibus
petiolo adnatis, spicis globosis.” This is the same diagnosis as that which
first appeared in the second edition of the Species Plantarum, five years
earlier.

In disposing of the two Linnaean names, Lantana corymbosa and Var-
ronia lineata, there are only two choices that can be justified logically.
They can be discarded as nomina confusa, or they can be typified by the
Jamaican plant illustrated by Plukenet. The names were applied by
Linnaeus to an aggregate, but to no more confused an aggregate than
many other names of that author which have been subsequently restricted
and are now generally accepted. In five out of the six times which Lin-
naeus or his students used the names between 1753 and 1767, the Jamaican
plant illustrated by Plukenet was included in the species treated. It was
not only a recurring element in the Linnaean concept, but also the one
which Linnaeus eventually emphasized and allowed to dominate in his
concept. I have accepted Plukenet’s plate as typifying both Lantana
corymbosa and Varronia lineata. ‘This seems logical and furthermore is
of some practical moment, since it provides a name for a species otherwise
without one.

9. Cordia coyucana Johnston, Jour. Arnold Arb. 29: 227 (1948).— Coyuca, Guerrero,
Hinton 8156

MEXICO: Coyuca, Guerrero, 3 m. tall, Hinton 8156 (G, TYPE).

GUATEMALA: Jalapa, shrub 2 m. tall, Standley 76595 (G).

HONDURAS: (dept. Morazan): Zamorano, Valerio 139 (G); El Pedregal,
Valerio ve (G); road to Tatumbla, Valerio 532 @.

ICA: Hacienda Santa Maria, Guanacaste, Dodge & Thomas 6294 (G);
La Co a Guanacaste, 1890, Pittier 2751 (G).

The species is known only from the collections cited above. Although
the plant has a large, plump, ellipsoidal or capitate inflorescence which is
rounded at the base and hence very different from the elongating, more
slender and narrower, basally attenuate spikes of C. curassavica, I believe
it is most closely related to the latter species. In foliage, in bristly twigs,
and in form and indument of the inflated fruiting calyx it is especially
suggestive of those forms of C. curassavica found in-Guerrero which have
been described as C. imparilis Macbr. It can be noted here, as an inter-
esting coincidence, that among the Mexican Varronias three very different
species have developed very bristly forms in the state of Guerrero, i.e.,
C. oaxacana (C. urticacea), C. curassavica (C.imparilis) and C. coyucana.
10. Cordia ambigua Schl. & Cham. Linnaea 5: 115 (1830).— Jalapa, Schiede 210.

MEXICO: Jalapa, Vera Cruz, a 216 (G, photo of TypE); Jalapa, Rose &
Hay 6139 (US); Jalapa, shrub 3-6 m., Pringle 8193 (G, US) and 9407 (G, US);
Izuatlanchillo, near Orizaba, Vera aie Renews 2625 bis (G); near Orizaba, Boureau

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 95

2625 (G, US); Orizaba, Botteri 169 (G) and 181 (G); Tenango, near Orizaba, Botteri
482 (US); Dos Puentos, Oaxaca, Aug. 1842, Liebmann 12733 (US); Mt. Orando,
Chiapas, Matuda 771 (G).

A very well marked species known only from southeastern Mexico.

11. Cordia inermis (Mill.), comb. nov
Lantana inerma Miller, Dict. (1768 ¥
Cordia cana Mart. & Gal. Bull. Acad. Brux. 117: 331 (1844).— Pacific slope of

ee tiie Greetm Proc. Am. Acad. 33: 482 (1898).— Maria Madre Island,
Nels
rear a northwestern Mexico south through Central America to
Panama and the dry north coast of Colombia. In Mexico it is confined to
the seasonally dry Pacific slope and is found as far north as Sinaloa.
MEX : Ymala, Sinaloa, 1891, Palmer 1419 (G, US); San Juan, Sinaloa,
Ortega 4030 (US) ; Rosario, Sinaloa, Rose 1825 (G, US) ; San Ignacio, Sinaloa, Montes
& Salazar 550 (US); See ican, Sinaloa, 1904, Brandegee (G); Yervacito, Sinaloa,
1904, Brandegee (US); Maria Madre icand. Nelson 4296 (G, TYPE; US), Solis 68,
Ferris 53629 (US), Bae 10518 (G) and Mason 1779 (G, US); Zapotlan, Jalisco,
3 m. tall, Pringle 4389 (G); Sacoalco, tome Jones 363 (US); barranca near Guadal-
ajara, Talicees 1886, Palmer 84 (G, US); La Palma, ae Jones 364 (US); Barranca
de Tequila, Jalisco, 1.5-4.5 m. tall, ee 4436 (G, US); Zitacuaro to EI Souse,
Michoacan, tree 2 m. tall, Hinton 13248 (G, US); Acapulco, Guerrero, 1895, Palmer 2
(G, US); Acatitlan, Mexico, 1 m. tall, inion 4337 (G, US); Tenatac, Mexico,
3.5 m. tall, Hinton 4213 (G); Salitre, Mexico, 1.5 m. tall, Hinton 6510 (G, US);
Tehuantepec, Oaxaca, Matuda 584 (G, US); Jalisco, ae Purpus 9211 (G, US);
i S).

GUATEMALA: divide between Zacapa and aa shrub 2-3 m. tall,
Standley 73704 (G); Gualan, Zacapa, shrub 3 m., Deam 6319 (G, US) and pp
(G, US); Coban, Johnson 39 and 567 (US); Giatemals ou pe 24 and 147
(US).

SALVADOR: Santa Ana, shrub 1.5-2.5 m., Standley 19689 (G, US); La Union,
shrub 9-12 dm. tall, Standley 20668 (G, US); La Union, Standley 20843 (US) ; Izalco,
dept. Sonsonate, Standley 22182 (G, US); San Salvador, Velasco 8926 (US) and
Calderon 1068 (US); dept. Ahuachapan, Sisto Alberto Padillo 84 and 377 (US

HONDURAS: Isla de Disposicion, 1.5 m. tall, Stork & Worth 8864 (G); Bella
Vista, dept. Choluteca, shrub 2-3 m., Williams & Molina 10873 (G); Yeguare,
Morazan, Valerio 314 (G); Zamorano, Morazan, Valerio 174 (G).

NICARAGUA: Granada, Bere 12799 (US) ; near Granada, Maxon, Harvey &
Valentine 7624 (G, US); Managua, Garnier 1093 (US) and Chaves 9 (US); Momo-
tombo, 1.8-2.1 m. tall, C. L. Smith 123 (G, US); Pueblo Nuevo, dept. Esteli,
Williams oe Molina 10808 (G).

COSTA RICA: road to Nicoya, Tonduz 13789 (G, US); Puntarenas, Rowlee 19,
65 and 141 (US); Nicoya, Cook & Doyle 645 (US).

NAMA: Isla Taboga, 2 m. tall, Woodson, Allen & Seibert 1478 (G); Isla
Taboga, 2-3 m. tall, Standley 27039 ( 8).

COLOMBIA: Santa Marta, Smith 582 (G, US).

For a widely ranging member of the section Varronia this species shows
very little variation. The species usually has an indument of fine, closely
appressed hairs. At the northern end of its area of dispersal the hairs, and
especially those on the calyx, tend to become spreading. Cordia insularis
was distinguished on this character.

The plant has for a century been identified, and correctly so, with C.
cana Mart. & Gal., a species based on material from southern Oaxaca. It

96 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

was, however, named very much earlier by Miller in his Gardeners Dic-
tionary as Lantana inerma. ‘The type of Miller’s species, preserved at the
British Museum, is an unmistakable, very characteristic specimen of the
present plant, labeled, subsequent to Miller’s time, as having been collected
by Houston. According to Miller, “The seeds of this sort were first sent
me by the late Dr. Houstoun, from La Vera Cruz, but I have since received
them from Jamaica.” The plant, however, is known neither from Jamaica
nor from Vera Cruz, Mexico, or their general vicinity. Except for mistakes
as to the color of flowers and fruit, Miller’s description agrees reasonably
well with his type. Even the ambiguous polynomial quoted from Sloane,
really applying to a very different plant of Jamaica, might fit our species
also

12. Cordia bullata (L.) R. & S. Syst. 4: 462 (1819

Varronia bullata L. Syst. ed. 10, 916 (1759); Amoee: Acad. 5: 394 (1759); Sp.
Pl. ed. 2, 276 (1762).— Jamaica

Cordia asperrima DC. Prodr. 9: 498 (1845) ; Urban, Symb. Ant. 3: 360 (1903).—
Jamaica, Bertero. Not Cordia asperrima Spreng. Syst. 1: 649 (1825), which is

tis.

at asperrima (DC.) Friesen, Bull. Soc. Bot. Geneve, sér. 2, 24: 155, t. 1,
f. 5 (1933
Va me ar Britton, Bull. Torr. Bot. Cl. 16 (1914). — Jamaica,
Harris 10995.

The species is ake only from scattered localities in Jamaica, Mexico,
and Central Amer

MEXICO: ssa a Gaumer 23479 (G, US); Dzitas, Yucatan, Stewart
372 (G

eee ALA: La Libertad, dept. Peten, Aguilar 171 (G); near Chinana, shrub
1-1.5 m., July 1860, Hayes (G); Fiscal, dept. Guatemala, shrub 18 dm., Standley
59552 (G); near Guatemala City, Tonduz 630 (G, US); barranca near Guatemala
City, 2.5-3 m., 1860, Hayes (G); indefinite, Heyde 171 he

SALVA DOR: near Chalchuapa, Calderon 1002 (G, U

ence near Granada, Maxon, Harvey & a e 7612 (G, US).

HONDURAS: Zamorano, dept. Morazan, Valerio 140, 175, o 1157 (G); Medina
near Coyoles, dept. Yoro, Vanienen. Kouta & Wagner 8637 (G).

JAMAICA: Farm Pen, Sanit Town, Campbell 5857 ae Yardly —
Harris 9671 (NY); upper Clarendon, Fas 10995 (TYPE of V. clarendonensis, NY) ;
Indefinite, Wolle (G); Union Hill, north slope of Mt. Diable, Maxon 10417 ey

The species has a close relative in the widely distributed C. globosa
(Jacq.) HBK. and in C. Bonplandii (Desv.) R. & S. of northern Vene-
zuela. It differs from the Venezuelan plant by having much smaller leaves
and much smaller heads with less thickened calyx-tips and more elongate
peduncles. Although C. dzllata practically always can be separated at a
glance from the closely related C. globosa, no single character has been
found that decisively separates them. There are numerous differences,
but these do not vary together and are not always positive. Although an
honest statement of differences in the key may not be very impressive, I
am of the opinion that C. bullata and C. globosa are two species worthy
of recognition. Of the two, C. bullata is more loosely branched and has
more slender and more woody twigs. Its leaves are usually ovate, sharply
and frequently doubly serrate, and are sharp-pointed at the apex. They

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 97

are usually a darker green than the lanceolate or elliptic, usually bluntly
toothed thinner leaves of C. globosa. The leaf-veins are usually much
branched and anastomosing to form an evident reticulum that is usually
prominent on the lower face and frequently evident as impressed lines
on the upper face. The lower surface of the leaf, and generally the upper
also, is therefore broken up into small areoles a few millimeters in breadth.
This is generally most evident on old, very mature foliage. The upper
surface has short stiff ascending bristles, usually arising from bulbose
mineralized bases. On hardened old leaves, on which the veins are well
impressed on the upper surface, usually only one to several bristles arise
from within each areole. On the lower leaf-surface the vein-reticulum
may be moderately prominent on lush new foliage or, in some plants, par-
tially hidden by an abundance of appressed hairs. Usually, however, and
especially on hardened old leaves, the reticulum on the lower leaf-face is
prominent and conspicuous. The areoles on the lower face in herbarium
material, when examined under a lens, usually reveal a rather distinctive
development of this species. In drying there is shrinkage of tissue beneath
many of the thick bases of the major bristles on the upper surface. Accord-
ingly, within the areoles on the lower surface there are usually present tiny
concave depressions. A suggestion of this development may be detected
very rarely in some specimens of C. globosa. In C. bullata, however, it
is practically always present, at least in some degree, and in most speci-
mens is very evident when looked for. The flowers of C. bullata are borne
in heads averaging perceptibly smaller than those in C. globosa. The
peduncles of C. dbzdlata, furthermore, usually become distinctly more elon-
gate than those of its relative. The tips of its calyx-lobes tend also to be
thicker and somewhat shorter than in C. globosa. The corolla of C. bullata
seems to be always small. Its short tube is very hairy within and swells
only moderately to form a subcampanulate throat. The usually larger
corolla of C. globosa is funnelform with the tube flaring into a widely
dilating throat. It has a well-defined hairy band inside the tube. In
C. bullata the band is less well defined and hairs frequently occur pro-
portionately much lower down on the tube surface.

The name Cordia bullata (L.) R. & S. is clearly the oldest name for the
plant of Jamaica later described as C. asperrima DC. It was originally
published as Varronia bullata L. Syst. ed. 10, 916 (early 1759) and treated
as follows: “bullata. B. V. fol. ovatis venoso-rugosis, spicis globosis.
Sloan. jam. t. 195, f. 1.’ The descriptive details and the cited plate both
apply to the Jamaican plant later called C. asperrima. The name V. dul-
lata next appears in the literature associated with the short original de-
scription of a Jamaican specimen collected by Browne. This was pub-
lished by Elmgren, one of Linnaeus’s students, Amoen. Acad. 5: 394
(Nov. 1759). The plant described is that now catalogued in the Linnaean
Herbarium as no. 255.2 and is also the one discussed at length by Smith
in Rees Cyclopedia, 36: sub Varronia bullata. Jamaican Cordia asperrima
is Clearly represented. A month after Elmgren published his dissertation,
another was published by Sandmark, Amoen. Acad. 5: 376 (Dec. 1759),

98 JOURNAL OF THE ARNOLD ARBORETUM [voL, xxx

in which the name V. dullata was again mentioned. In this case it was
associated with the first of the Varronias treated on page 172 of Browne’s
History of Jamaica. Browne’s descriptive matter is ambiguous, but the
plate cited, t. 13, f. 2, evidently represents the northern form of C. globosa.
The fourth appearance of the name V. bullata is in the second edition of
the Species Plantarum, 276 (1762). Linnaeus gave only two references
as basis for the name, the first being Elmgren’s description of C. asperrima
and the second Jacquin’s account of the utterly different V. mirabiloides.
It is to be noted that Lantana bullata L. Sp. Pl. 627 (1753), although a
species of Cordia, is not synonymous with Varronia bullata L. Syst. ed.
10, 916 (1759). In the Species Plantarum, ed. 2, 276 (1762), Linnaeus
correctly cites Lantana bullata as a synonym of Varronia curassavica.

13. Cordia globosa (Jacq.) HBK. var. humilis (Jacq.), comb. no

Varronia humilis Jacq. Enum. 14 (1760); Sel. peek 41 (1763). aie,

Cordia humilis (Jacq.) Don, Gen. Syst. 4: 383 (18

Lithocardium corymbosum var. humile (Jacq.) eee Rev. Gen. 2: 438 (1891).

Cordia jacmeliana Krause, Beih. Bot. ea 32: 344 (1914); Urban, Symb.
Ant. 8: 579 (1 G1) cx aac Haiti, Ara

Varronia jacmeliana (Krause) Friesen, Bull Soc. Bot. Genéve, sér. 2, 24: 177
(1933),

Varronia humilis var. mexicana Friesen, Bull. Soc. Bot. Genéve, sér. 2, 24: 162,
1, f. 4 (1933).— Morelos, Mexico, habia 6346.

Varronia mexicana Friesen, Bull. Soc. . Genéve, sér. 2, 24: 162 (1933).

MEXICO: Mazatlan, Sinaloa, Rose, ae & Russell 13719, 14136 (US), Howell
re (US), and Rose 3108 (US); La Constancia, Sinaloa, shrub 4 m., Ortega 5518

US); between Rosario and Acaponeta, Sinaloa, Rose 1871 (G); Topolobampo,
ead 1897, Palmer 180 (US); Culican, Sinaloa, shrub 2-3 m. tall, Gentry 7081
(G) and 7087 (G); Colima, 1897, Palmer 52 (G, US); Zitacuaro to Laurelles,
Michoacan, shrub 4 m., Hinton 13209 (G, US); Ixtapan, Mexico, tree, Hinton 1164
(G, US); mear Cuernavaca, Morelos, 3-4.5 m. tall, Pringle 6346 (G, US); Yautepec,
Morelos, Rose, Painter & Rose 8547 (G, ee ; San Luis Tultitlanapa, Puebla, Purpus
2472A (G, US); Mt. Male, near Porvenir, Chiapas, Matuda 4661 (G); Cosumel
Island, Gaumer 126 (G); ame, Yucatan, ee 801 (G, US); Progreso, Yucatan,
Lundell 7961 (G, US).

MALA: between Rio Hondo and Santa Cruz, dept. Zacapa, shrub 2
Standley 74084 (G); Zacapa, shrub 2-3 m. tall, Standley 73627 (G) and 73964 (G);
Chiquimula, shrub 2 m., Standley 73967 (G); between San Ildefonso and Cuilco,
dept. Huehuetenango, Stevennark 50751

SALVADOR: La Libertad, shrub 1-3 be Standley 23218 (G, US); Acajutla,
shrub 115 m., Eyerdam & Beetle 8735 (G

NICARAGUA: base of Coseguina Velen , Howell 10256 (G); Santiago Volcano
near Masaya, Maxon 7678 (US); Managua, Chat 386 (

MA: Punta Paitilla, shrub 1-2.5 m. tall, Standley 26268 (G, US).

The widely ranging Cordia globosa breaks up into two geographic vari-
eties. The northern and western plants, those found from Florida and
northwestern Mexico south through the Greater Antilles and Central
America, are here distinguished as the var. humilis. Typical Cordia
globosa, a plant of the Lesser Antilles, Trinidad, Venezuela, and eastern
Brazil, has larger, more pointed (acute or acuminate) leaves sharply

serrate on the margins. Its flower-heads are also larger. hese tw
eens of the species are to be detected in any large and representative

cot

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 99

suite of specimens. The differences are not alway positive, especially in the
Lesser Antilles, but serve reasonably well for the separation of northern
and southern plants.

The variety is typified by the Jamaican plant described and illustrated
as the “Round spiked Varronia” by Browne, Hist. Jam. 172, t. 13, f. 2,
since Jacquin specially mentioned it when he described Vernon humilis.
The plant described as V. globosa by Jacquin is evidently the southern
plant. He stated that it resembled his V. martinicensis, but differed in
flowers and inflorescence. In size and form of leaves the southern plant
is indeed very similar to C. martinicensis as illustrated by Jacquin, Sel.
Stirp. 41, t. 32 (1763).

14. Cordia curassavica ee R. & S. Syst. 4: 460 (1819).
Varronia curassavica Jacq. Enum. 14 (1760) and Sel. Stirp. 40 (1763).— Curacao.
Lantana bullata L. Sp. PL ou (1753). Not V. bullata L. (1759) nor C. bullata
R.& S. (1819). — Jamaic

Varronia macrostachya =e Enum. 14 (1760) and Sel. Stirp. 41 (1763).— Car-

tegena, Colombia.

Cordia Oe ao (Jacq.) R. & S. Syst. 4: 461 (1819).

Cordia brevispicata Mart. & Gal. Bull. Acad. Brux. 11?: 331 (1844).—- Tehuacan,

Puebla, Galeotti 7192.
fg oe rue peruviana var. mexicana DC. Prodr. 9: 491 oe i

Cordia linearis DC. Prodr. 9: 493 (1845). — Mexico, herb.

Cordia hispida Benth. Bot. Sulphur 139 (1845).— Gulf es ses Sinclair

es Palmeri Wats. Proc. Am. Acad. 24: 62 (1889).— Guaymas, Sonora, Palmer

ah socorrensis Brandeg. Erythea 7: 5 (1899).— Socorro Island, Anthony 384.

Cordia brevispicata var. hypomalaca Greenm. Field Mus. Pub. Bot. 2: 338 (1912).
—Cerro San Felipe, Oaxaca, Conzatti 1831.

Cordia imparilis eye Contr. Gray Herb. 49: 16 (1917).— Banio nuevo,
Michoacan or Guerrero, Langlassé 265.

Cordia chepensis Bite Contr. U. S. Nat. Herb. 18: 253 (1917).— Chepo, Panama,
Pittier 4511.

Cordia littoralis Pittier, Contr. U. S. Nat. Herb. 18: 253 (1917).— Porto Limon,

Rica, Pittier 3641.
Cordia mollis Pittier, Cont _S. Nat. Herb. 18: 294 (1917).— between Guate-
mala City and Salama, pe 134.

Ranging from Mexico and Cuba south through Central America and
the West Indies into northern South America. In Mexico the plant grows
along the Pacific slope in the western and southern parts of the country,
extending as far north as middle Baja California and middle-western
Sonora. Only in Yucatan is it found on the Caribbean slopes of Mexico.
In Central America it occurs on both slopes but apparently with greater
frequency on the Pacific.

The.species is very widely ranging and includes very diverse forms. It
has accumulated a formidable list of synonyms. The names given above
include only the very oldest few and those based upon plants originating
in Mexico and Central America. Although readily distinguishable by its
inflorescence, our plant has been frequently confused with C. cylindro-
stachya (R. & P.) R. & S. and its allies, of northwestern South America.
The latter range along the Andes from Colombia to Bolivia and produce
most of their spikes directly from leaf-axils along leafy twigs, only the

100 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

very first produced on the shoot being terminal. The spikes of C. curas-
savica, on the other hand, are all terminal. By subsequent elongation of
the twig, however, the spikes may appear to arise from stem-forks or,
still on naked peduncles and without subtending leaves, from internodes
along leafy stems. None of the spikes in C. curassavica ever arise from
the leaf-axils.

The oldest name for this widely distributed plant is Lantana bullata
L. Sp. Pl. 627 (1753). The description of the species records it as “spicis
oblongis” and as originating in Jamaica. Linnaeus’s knowledge of it was
apparently derived entirely from the writings of Plukenet and Sloane. Of
the three references given, one is to the works of the former and two
to the latter. The first given is to Plukenet, Alm. 329, t. 221, f. 3, and
clearly applies to a form of C. curassavica. The second reference, Peri-

clymenum ... folio majore oblongo bullato .. ., Sloane, Hist. 2: 81, applies
to a close relative of C. curassavica. The third reference, however, Peri-
clymenum ... folio majore subrotundo bullato, Sloane, Hist. 2: 81, t. 195,

. 1, involves a very different species, one bearing flowers in heads, which
Linnaeus, Syst. ed. 10, 916 (1759), later distinguished as Varronia bullata.
In the second edition of the Species Plantarum 2: 276 (1762), Linnaeus
cited Lantana bullata (without name) as a synonym of Varronia curas-
savica Jacq. Of the three references originally given for Lantana bullata
only the first of the two Sloanean ones is repeated in his synonymy of
Jacquin’s species. It is to be particularly noted that the significant phrase,
‘‘spicis oblongis,” first used in 1753, again appears in his diagnosis of C.
curassavica. Although the name Lantana bullata L. (1753) much ante-
dates Verronia curassavica Jacq. (1760), it cannot be transferred to Cordia
or Varronia, since the resulting combination would be a homonym of C.
bullata R. & S. (1819) or V. bullata L. (1759).

The next available names for our plant are Varronia curassavica Jacq.
and V. macrostachya Jacq. These were first published in Jacquin’s Enu-
meratio, p. 14 (1760) as follows:

“curassavica 1. VARRONIA spicis oblongis; foliis lanceolatis. kh Brown.
Jam. 2. p. 172

“macrostachya 2, ———— -——— spicis oblongis; foliis lanceolato-
oblongis. .” Three years later they were again treated, with more
detail, by Jacquin, Sel. Stirp. p. 40-41 (1763), as follows:

“1 VARRONIA spicis oblongis; foliis lanceolatis. WVarronia assurgens
sarmentosa, foliis & capitulis oblongis. Brown. jam. 2. p. 172. Frutex
quindecimpedalis; vamis teretibus, scabris, senioribus ferrugineis. Folia
lanceolata, acuta, serrata, rugosa, venosa, facie scabra, petiolata, alterna.
Spicae densae, terminales, duos tresve pollices longae, insidentes pedun-
culo bipollicari. Flores parvi, inodori. Calycis inflati laciniae subovatae
desinunt in denticulos setaceos ejusdem secum longitudinis. Corollae
albae tubus longitudine calycem aequat; limbus brevis est & emarginatus.
Stamina petalo sunt paulo breviora. Stylus etiam brevis stigmate capitato
simplicique instruitus. Drupa rubra & parva est. Habitat in Curacao in
sepibus arboreis.

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 101

“2. VARRONIA (macrostachia) spicis oblongis; foliis lanceolato-ob-
longis. Arbuscula erecta, duodecimpedalis. Spicae semipedalis. Folia
angusta, semipedalis. Reliqua, ut in praecedente. Habitat Carthagenae
in fruticosis & silvaticis.”

t will be noted that the trivial name curassavica was omitted in the
second publication. The lengthy discussion published in 1763, however,
evidently applies to V. curassavica.

I have come to the conclusion that Jacquin’s two binomials apply to
forms of a single species. Varronia curassavica is the common xerophytic
phase, and V. macrostachya is the more luxuriant mesophytic phase of the
most common of the spicate Varronias in northern South America. This
is a departure from some recent usage in which V. eh a has been
applied to the plant distinguishable as C. divaricata HB

The identity of V. macrostachya is readily established since there is
only one species to which it is possibly applicable that has been found at
its type locality (Cartagena) and in ecologically similar stations (Barran-
quilla, Santa Marta) on the dry north coast of Colombia. Jacquin says
that V. macrostachya differs from V. curassavica in being a shrub 3.6 m.
(rather than 4.5 m.) tall, in having spikes 15 cm. (rather than 5—7.5 cm.)
long, and in having lance-oblong (rather than lanceolate) leaves reaching
15 cm. in length. The elongate spikes and large leaves indicate that he
had one of the luxuriant mesophytic phases of the species repeatedly col-
lected near the ports of northern Colombia.

The specific name selected by Jacquin when he named V. curassavica
makes it desirable that the binomial should apply to some plant of
Curacao. Only two spicate Varronias occur on the island. Judging from
the relative number of collections made there, C. divaricata must be rare
or local on the island, and the xerophytic phase of the plant named V.
macrostachya must be common and widely distributed there. The latter
seems to be the plant to which the name V. curassavica belongs. Jacqut S
description of the leaves of the island plant, “lanceolate” and “acute,”
certainly apply best to it. Had Jacquin possessed material of C. divaricata,
he would certainly have mentioned the hairy upper leaf-faces in distin-
es it from the plant of Cartagena. He does speak of the scabrous
leaves of V. curassavica, but this almost certainly refers to the harshness
resulting from the great numbers of minute siliceous tubercules usually
present on the glabrous upper surface of the xerophytic form of V. macro-
stachya. If V. curassavica is taken as one of two possible species of
Cue and is judged by comparisons with V. macrostachya, its identity

* A species widely distributed in northern Venezuela and best known from the region

lobes tend to be shorter and proportionately broader, and the corolla-lobes are Be
more erose-denticulate and crisped than in C. curassavica. C. cuneiformis DC.
synonym.

102 JOURNAL OF THE ARNOLD ARBORETUM (VOL. Xxx

seems clear. Varronia curassavica and V. macrostachya are forms of one
species.

Some of the details in Jacquin’s later discussion of V. curassavica, how-
ever, are puzzling. They agree neither with C. divaricata nor with forms
of V. macrostachya. They suggest that Jacquin’s notes may have been
faulty, that he may have been misled by extraneous material associated
with his specimens, or that there was some confusion in his manuscript. In
the first place he states that V. curassavica grew 4.5 m. tall, a height which
is certainly unusual if not greatly excessive for the species. Most im-
portant, however, he gives the calyx as inflated and having subovate lobes
prolonged into setaceous tips. This agrees with the calyx of Varronia
globosa Jacq., a species with capitate inflorescences common on Curacao,
but certainly not with any of the spicate Varronias known from the island
or from Venezuela. Indeed almost all of the details given for the flowers
of V. curassavica could apply better to V. globosa. It is surprising that
Jacquin, in his description of V. globosa, made no mention of the setaceous
tips of the calyx-lobes, one of the most conspicuous and distinctive traits
of that species. Finding reference to such structures in Jacquin’s descrip-
tion of V. curassavica, where they certainly do not belong, I am willing to
believe that the notes on flowers and fruit appearing there may have been
intended for his account of V. globosa and were somehow misplaced. For-
tunately, the confusion described can be overlooked, since it does not
appear in the Enumeratio (1760), where the species were established. It
appears only in the Stirpium (1763), his second amplified treatment of the
pierre where by oversight even the specific name “curassavica” was

mitted. Since we are able to associate it with one of the two spicate
Varronias of Curacao, the name V. curassavica can be used. Moreover,
it is preferred to macrostachya in combining Jacquin’s two species, since
it belongs to the species originally given priority of position and a fuller
discussion. Also it has been more frequently used by subsequent authors
and is hence the more familiar name

In the present paper I have defined C. curassavica very broadly, for the
simple reason that I have not had time to give the Mexican and Central
American plants of this complex group the study they need. I have in-
cluded under it some very diverse forms which will almost certainly be
treated eventually as separate species. The diverse forms of Mexico and
certain hairy plants of Central America are in many ways more similar
to plants of the Greater Antilles than to typical C. curassavica, and must be
studied with the former

ical C. curassavica is known only from northern South America from
fo)

Cc
risers Pittier. It is the only form of the present aggregate found in
Panama. Specimens in the Gray pee representing typical C.
curassavica from Central America are as follow

Chepo, Pittier 4511 (C. chepensis es. ee Standley 25552 and
27152; Ancon Hill, Killip 12059; Race Track near Panama, Standley 27795; Bella

1949 ] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 103

Vista, Maxon & Valentine 6945; Miraflores Lake, White 244; Isla Taboga, Woodson
Allen & Seibert 1485; Aguadulce, Coclé, Pittier 4860; between Aguadulce and Anton

Island, Hayes 5; Chagres, Fendler ie near Chiriqui Lagoon, Bocas del Toro, Wedel
562, 2478, 2608, and 2923.
COSTA RICA: Port Limon, Pittier 3641 (C. littoralis Pittier).
3

NICARAGUA: San Juan del Norte, C. oa Smith 91.

GUATEMALA: ee Standley 7526

BRITISH HONDURAS: Manatee ee Peck 65; Lower Belize River, Record;
New Town, Schipp 821; Bakers Pine bee Lundell 7002; Honey Camp, Lundell
508; San Antonio, Corozal, a 498

Yucatan, Chank i ae nt 84; indefinite, Gaumer 24027 es 24234;

Quintana Root Lake Coba, Lundell 7686; Campeche, Hecelchakan, Stewart 11.
15. Cordia costaricensis, sp. nov.

Frutex scandens, ramulis minute brunneo- pubescentibus, internodis 1-7
cm. longis; foliis ‘alternis late lanceolatis penninervis reticulato-venosis,

ratis, facie superiori minute papillatis pilis rigidis ad 0.5 mm. longis plus
minusve donatis rare subglabris, facie inferiori pilis minutis mollibus 0.2—
0.5 mm. longis plus minusve donatis; petiolo 1-3 cm. longo cum pedunculo
vel ramo axillari basi connato; spicis densis vel laxis 2-8 cm. longis ad
8 mm. crassis 1-3 cm. longe pedunculatis in paniculis laxis 15-25 cm.
longis terminalis vel axillaribus paucissime Wee gestis; calyce sessile
fere ad medium lobato, in alabastro late obovato 2-3 mm. crasso summum
ad apicem rotundatum apices liberes een 0. 2—0.5 mm. longas prore-
renti, extus subglabro vel supra medium pilis minutis Seder sparse
obsito, intus glaberrimo, sub anthesi 4-5 mm. longo basi et paulo ae
basem 1—1.5 mm. crasso deinde sursum gradatim ampliato apice 4-5 m
diametro; lobis calycis ascendentibus deltoideis vel ovato-deltoideis, apice
apiculatis et subincrassatis, sinibus acutis; corolla campanulat a 5-6 m
longa, basi 1.5-2 mm. crassa ceinde sursum ampliata, apice 5-6 mm.
diametro, margine erosa haud lobata, extus glabra, intus sub insertione
staminum villosa alibi glabra; staminibus inclusis, antheris 1 mm. longis,
filamentis 2.5 mm. longis; ovario cum stylo glabro; fructu ignoto.
COSTA RICA: Vicinity of El General, prov. San Jose, 1190 m. alt., scandent

Vara Blanca de Sarapiqui, north slope of Central’ Cordillera, 1500-1750 m. ait., in
thicket, fl. white, July-Sept. 1937, Skutch 3175 (G).

A very well marked species with its closest relations in C. multispicata
Cham. of eastern Brazil. From that species it differs in its somewhat
larger and more elongate leaves, axillary panicles of spikes, and tubular
erose-margined corollas

16. Cordia spinescens L. Mant. 2: 206 (1771).— “India orientali.”

Varronia ferruginea Lam. Tab. Encyc. 1: 418 (1791); Poir., Encyc. 4: 263 (1797) ;
Desv. Jour. de Bot. 1: 266, t. 9 (1809). Freie on plants cultivated at Paris.

Cordia ferruginea Geen R. & S. Syst, 4: 458 (18

sae dia riparia HBK. Nov. Gen. et Sp. 3: 71, t. aah (1818). — Type from Mom-

ie pease Val, ber same

ae laxiflora Gen. et a 3: 72 (1818).— Between Mompox and

Morales, eae ae Colombia

104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Cordia pvrifolia Willd. ex R. & S. 4: 802 (1819).— South America,

Cordia Thibaudiana DC. Prodr. 9: 489 (1845).— America

Cordia crenulata A.DC. Prodr. 9: 492 (1845).— Mexico, herb.

Varronia Nicaea Sessé & Moc. FI. Mex. 48 (1893), ed. 2, yi ee — From
Orizab exico.

Cordia seria gs Mem. Torr. Bot. Cl. 6: 83 (1896).— Near Cochabamba,
Bolivia, Bang

Ranging from ae southward through Central America into Colom-
bia and from there south along the Andes to Bolivia. In Mexico it is
practically confined to lower altitudes towards the coasts and is found as
far north as Nayarit and middle Vera Cruz. Most of the many speci-
mens examined come from Mexico, Guatemala, and Panama.

This species has been called C. ferruginea (Lam.) R. & S. It has, how-
ever, an older name in Cordia spinescens L. This latter was based upon
a specimen still preserved in the Linnaean Herbarium. It was launched
with a detailed description, but, unfortunately, was mistakenly said to be
an East Indian plant. Unrecognized by botanists in the Old World and
unstudied by botanists in America the species has remained neglected and
unplaced. No Cordia remotely suggesting it is known from the Old
World. It is, however, indistinguishable from the well-known American
species usually called C. ferruginea. The type of C. spinescens has a few
spikes and numerous developing axillary shoots and the general appear-
ance of growing in a botanic garden away from its natural environment.
I suspect that it may have been grown from seeds sent from Colombia by
Mutis or have been derived from plants of other sources grown at Madrid
or Paris. The species is known to have been in cultivation in Europe in
the late eighteenth century. The synonymous C. ferruginea, indeed, was
based on plants growing at Paris sometime previous to 1791. The specific
name “spinescens” adopted by Linnaeus refers to the decurved spurs,
persisting indurated basal portions of petioles, that are very well devel-
oped along the clambering shoots of this shrub.

B. THE IDENTITY OF SPECIES PROPOSED BY
SESSE AND MOCINO

A very large number of puzzling species were saa about fifty
years ago, when the manuscripts of Sessé & Mocino were dug out of the
archives at Madrid and printed in eee as the Flora Mexicana” and
“Plantae Novae Hispanicae.” The Boraginaceae described in these works
mostly defied identification. On various trips to Europe, accordingly, I
methodically recorded the data on all specimens believed to be collections
of Sessé & Mocifio. Notes were also made on the set of Mocino drawings
at Geneva. In 1936, when the Sessé & Mocifio herbarium at Madrid was
loaned to the Chicago Natural History Museum for study, I was allowed
to examine all the Boraginaceae it contained. With data assembled from
the sources mentioned it is now possible to identify with reasonable ac-
curacy the species first published in the posthumous writings of Sessé &
Mocifio. Only those species which the authors themselves obviously in-

1949 ] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 105

tended as new proposals are here discussed, since only these have any
nomenclatorial importance. The many misidentifications contained in the
work are properly ignored.

The specimens from Madrid commonly bear the identification of Sessé
& Mocino and frequently also their collection number, but never any
geographical data. To facilitate reference, a new set of herbarium num-
bers was applied to the sheets in the collection after it arrived at Chicago.
In referring to the Sessé & Mocifio material belonging to Madrid, this new
set of numbers has been used. For each collection mentioned I have given
its recently assigned number, and, when it is not an unpublished binomial,
have given the Sessé & Mocifio identification also.

The study of the Sessé & Mocifio borages has made it clear that the
authors either collected in northern Mexico or had collaborators there.
In the past it has been generally stated that their material came only
from central and southern Mexico, Guatemala, Cuba, and Porto Rico.
Rickett has recently found proof that northwestern Mexico was another
source. Northeastern parts of the country, very likely the vicinity of
Monterrey, are also to be included. This is evidenced by the presence
in the Sessé & Mocifio herbarium of material of Cordia podocephala Torr.,
Cordia Boissiert DC., Ehretia Anacuna (Berl.) Johnston, and especially
Omphalodes aliena Gray.

Cordia alliodora (R. & P.) Cha
Varronia tuberosa S. & M. Pl. N. isp. 30 (1888), ed. 2, 28 (1893). — From Metepec,
Vera Cruz

The pte of this species assembled by Sessé & Mocifio and now at
Madrid is numbered and bears original identifications as follows: 752
(Cerdana), 1352 (Varronia tuberosa), 1366 (Cordia nodosa), 1452 (with-
out name), 1568 (Cordia Gerascanthus), 1569 (Cerdana), 5275 (Cordia
globosa). A specimen at Oxford is labeled “Varronia tuberosa, Mexico
ex herb Ruiz.”

When Varronia tuberosa was described reference was-given to “Fl. Mex.
Ic. 193.” This plate number appears on no. 882 of DeCandolle’s copies of
the Mocino plates at Geneva. The plate is labeled “Cordia nodosa,” a
name apparently originating with Sessé & Mocijfio. It is incorrectly cited
under Cordia nodosa Lam. in the Prodromus 9: 475 (1845).

Cordia diversifolia Pavon ex DC.

Cordia paniculata 8S. & M. FI. es 49 (1894), ed. 2, 45 (1894).— Habitat in Op-

pido de la Punta, ? Vera Cru

An unidentified specimen at nae bears the number 760. It is con-
specific and probably a duplicate of at least the Mexican specimen from
Pavon’s herbarium cited by DeCandolle, Prodr. 9: 474 (1845), when he
described C. diversifolia.

Cordia dodecandra DC.
Cordia dodecandria S. & M. FI]. Mex. $0 (1894), ed. 2, 46 (1894). — From coast at
Alvarado, Vera Cruz.

This species is represented by an unidentified specimen, no. 758 bis,

at Madrid, and also by a fine colored plate, tab. 874, among the copies of

106 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

the Mocifio drawings at Geneva. The plant, bearing the name “Cordia
dodecandra,” is the basis upon which DeCandolle, Prodr. 9: 478 (1845),
described the species.
Cordia elaeagnoides DC.
? Cordia exsucca S. & M. FI. Mex. 49 (1894), ed. 2, 45 (1894).— Habitat in
Apataingani [Michoacan] aliisque calidissimis Novae Hispaniae locis

I have seen nO material collected by Sessé & Mocino and am able to
judge their C. exsucca only from the description. This fits C. elaeagnoides
DC. reasonably well.

Cordia pauciflora DC. ex Ramirez, Estud. Hist. Nat. 295 (1904), not Rusby (1896).

Ramirez identifies plate 877 of the set of Sessé & Mocino plates at
Geneva with plate 25 of the set at Madrid and associates them with
“Cordia Sebestena, Jacq. — Plantae Novae Hispaniae pag. 28. Ic. 36.”
The plate at Geneva is very sketchy, being only a poor pencil outline. It
bears the name “Cordia pauciflora,’ which probably originated with Sessé
& Mocifio. In general habit the sketch most suggests a species of Bour-
reria. 1 do not believe it is the same as Cordia Sebestena sensu Sessé &
Mocifio, Pl. N. Hisp. 30 (1888), ed. 2, 28 (1893), which is given as
originating in dry fields near Cuernavaca. Ramirez’s publication of the
name “Cordia pauciflora” consists of mere casual mention. No description
is given.

Cordia podocephala Torr.
Varronia longifolia S. & M. Fl. Mex. 48 (1893), ed. 2, 44 (1894).— Habitat in
Praedio de la Punta, ? Vera Cruz.

A specimen at Madrid, no. 5269, bears the name “Varronia longifolia
N.” The plant is unquestionably conspecific with C. podocephala, known
only from northeastern Mexico and adjacent Texas. The published de-
scription of V. longifolia agrees with Torrey’s species, also.

Cordia spinescens L.

Varronia eric S. & M. Fl. Mex. 48 (1893) ed. 2, 44 (1894).— From the vicin-

ity of Orizaba, Vera Cruz.

Material at Madrid representative of the species is as follows: 739
(Varronia crenulata), 740 (without name), 741 (without name), 742
(without name), 5262 (new sp. aff. spinescens), 5264 (Varronia spines-
cens), 5273 (Varronia crenulata). Judging from similarities in material
I would say that parts of only three different collections = represented
in the suite cited. The type of Cordia crenulata DC., Prodr. 9: 492
(1845), is almost certainly a duplicate of one of these ae

hretia Anacuna (Berl.) Johnston

Ehretia lancifolia S. & M. Fl. Mex. 51 (1894), ed. 2, 44 (1894).— Habitat in Tex-

CUuco.

Representative specimens of the species at Madrid are as follows: 773
(E. lancifolia), 774 (Ehretia sp. nov.), 775 (without name), 5247 (Ehretia
sp. nov.), 5250 (E. lancifolia, communicavit D. Ignatius de Leon, Horti
Reg. Mex. Alumnus). Judging from general appearance, the material
cited probably consists of duplicates of no more than two different collec-

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 107

tions, one bearing the name E&. lancifolia and the other an unpublished
binomial. The species is known as a wild plant only in northeastern
Mexico and adjacent Texas, but I have seen cultivated specimens of it
from as far south in Mexico as Guanajuato.

Bourreria rotata (DC.) comb. nov

Cordia rotata Mocino ex DC., Prode 9: 483 (184

Ehretia calophylla Richard in Sian Fl. Cubana Ae t. 61 (185

Sr eileog calophylla (Rich.) Wright in Sauvalle, Fl. Cubana 110 oe Schulz

an, Symb. Ant (1911).

The species described by DeCandolle is based upon a handsome colored
plate, no. 880, in the set of Mocifio drawings at Geneva. The plate shows
a plant with the characteristic forking style of a Cordia, but in all other
details it is a remarkably good representation of the Cuban, plant now
current as Bourreria calophylla. The forked style is evidently a mistake
of the artist.

Bourreria spathulata (Miers) Hemsle

Ehretia cuneifolia 5S. & M. Fl. Mex. "51 (1894), ed. 2, 47 (1894).— From Chilpan-

zingo, Guerrero.

Two collections of the species are at Madrid: 1354 (Ehretia exsucca)
and 5252 (Ehretia cuneifolia, olim exsucca). A fine original colored plate
of the species occurs among the Mocino plates at Geneva, no. 887. It
bears an unpublished binomial, under EAretia, and also Mocifio’ s number,
292. This latter was cited when Ekretia cuneifolia was described. Miers
based his species upon a plant from Pavon’s herbarium. The type is
almost certainly a duplicate of one of the two Madrid specimens cited

bove

Heliotropium angiospermum Mur
peel lancifolium S. & M. Fl. Mex. 31 (1893), ed. 2, 29 (1894). — Cuerna-
vaca, Morelos
pee lanceolate S. & M. Pl. N. Hisp. 20 (1888), ed. 2, 19 (1893). — Cuernavaca,
orelos.

The herbarium at Madrid contains a number of specimens of this
species: 858 (Heliotropium sp., Cerinthe lanceolata), 1384 (Heliotropium
indicum), 1718 (Heliotropium lancifolium), 1719 (Heliotropium sp. nov.),
1722 (Heliotropium parviflorum), 5254 (Heliotropium parviflorum).
Some of the material probably came from Havana and is that treated as
H. parviflorum in Flora Mexicana 31 (1893), ed. 2, 28 (1894).
Heliotropium assurgens ee

Anchusa incana S. & M. Fl. Mex. 33 (1893), ed. 2, 30 (1894).— Cuernavaca,

Morelos

Two specimens at Madrid clearly belong to this species: 931 (Litho-
spermum olim Anchusa incana) and 1435 (Anchusa incana).
Heliotropium calcicola Fernald

Symphitum fruticosum S. & M. Pl. N. Hisp. 21 (1888), ed. 2, 20 (1893). — Chil-

pancingo, Guerrero.

There are three specimens at Madrid, 861 (Heliotropium frutescens),
1716 (Heliotropium sp. nov.), and 5256 (Heliotropium sp. nov.).

108 JOURNAL OF THE ARNOLD ARBORETUM [VOL. xxx

are so similar in appearance that they may well be only duplicates of a
single collection. At Geneva there is a beautiful plate of the species, tab.
901, labeled Symphitum fruticosum and bearing the number 288 of Sessé
& Mocifio’s series of plates. This latter number is"cited in Sessé &
Mocifio’s account of S. fruticosum. Antiphytum mexicanum DC., Prodr.
10: 121 (1846), was based entirely upon the plate at Geneva.
Heliotropium oaxacan
Lithospermum sna oe S. & M. Fl. Mex. 33 (1893), ed. 2, 30 (1894).—
From Uruapam, Michoacan.
Anchusa es S. & M. Pl. N. Hisp. 21 (1888), ed. 2, 20 (1893). — From Uruapam,
Michoa

ani of the species at Madrid is one specimen, no. 5238. This
is associated with an unpublished binomial of which “rosmarinifolium’’ is
the specific adjective. The descriptions of Sessé & Mocifio apply reason-
ably well to H. oaxacanum DC., and I am content to refer this proposed
species to the synonymy of it.

Heliotropium Sessei Johnston
Myosotis mexicana S. & M. Fl. Mex. 33 (1893), ed. 2, 31 (1894).—In temperatis
N. Hispan. montibus

Two specimens are - piesa at Madrid, 1725 (Mvyosotis mexicana)
and 5229 (Myosotis mexicana). They are probably duplicates of a single
collection. The description of Myosotis mexicana fits them reasonably
well. The species is known only from Sierra de la Mesa, Hidalgo (Purpus
1402 and Rose, Painter & Rose 9122) and from Sessé & Mocifio’s collec-
tions cited above.

Tournefortia ee Sw.
Tournefortia laevis S. & M. Fl. Mex. 52 (1894), ed. 2, 48 (1894).— Habitat in
Novae Hi oe regionibus.
A specimen at Madrid represents Sessé & Mocifio’s species, 1712 (T.
laevis).
Tournefortia hirsutissima L.
Tournefortia odorata S. & M. Fl. Mex. 52 (1894), ed. 2, 48 (1894). — Habitat in
Apatzingani [Michoacan].

Several collections of this species are at Madrid, most of them under
unpublished names: 1705 (J. suffruticosa), 1706 (Tournefortia n. sp.),
1707 (Tournefortia sp. nov.), 1715 (Tournefortia sp. nov.), 5261 (Tourne-
fortia sp. nov.). I suspect that no. 1707 may be the type of 7. odorata.
The other collections are perhaps those described as 7. suffruticosa sensu
S. & M. Fl. Mex. 52 (1894), ed. 2, 48 (1894), and T. cymosa sensu S. &
M. PI. N. Hisp. 31 (1888), ed. 2, 29 (1893). These latter two names apply
to plants from San Juan de los Platanos near Apatzingan, Michoacan.
Tournefortia syringaefolia Vahl

peel ce lanceolata S. & M. Fl. Mex. 52 (1894), ed. 2, 48 (1894).— From

ains near Tuxtla, Vera Cru

Two specimens representative a T. syringaefolia (T. peruviana) are
found at Madrid: 1711 (7. lanceolata) and 5260 (Tournefortia n. sp.).
One is probably the type of T. lanceolata S. & M

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 109

Lithospermum distichum Ort
Lithospermum laevigatum 5. & M. Fl. Mex. 32 (1893), ed. 2, 30 (1894).— Habitat
in montibus Predii S. Nicolai [state of Mexico]
? Anchusa mexicana S. & M. Pl. N. Hisp. 21 “ae ed. 2, 20 (1893).— Moun-
tains of Patzquaro, Michoacan.

I have seen two specimens from Madrid representative of L. distichum,

z., 1732 (Lithospermum laevigatum) and 5226 (Lithospermum n. sp.).
No specimen identified as “Anchusa mexicana’ has been encountered.
The original description of that species, however, agrees reasonably well
with L. distichum, and I am content to refer it to the synonymy of the
latter
Lithospermum discolor Mart. & Gal

Lit age eb al eae S.& M. FI. Mex. 32 (1893), ed. 2, 29 (1894). — Habitat

in mon s Oppido de El Valle.

Two specimens at Madrid belong here: 1737 (L. obtusiflorum) and
5232 (L. obtusiflorum). Specimens in the Boissier Herbarium labeled
“Nueva Espana, herb. Pavon, Lithospermum obtusiflorum,” are also ref-
erable to L. discolor.

Lithospermum strictum
Lithospermum omeioliom S & M. FI. Mex. 32 (1893), ed. 2, 29 (1894). — From
Ario, Michoac

Heliotropium mexicanum S. & M. PI. N. fas 20 (1888), ed. 2, 19 (1893). — Habi-

tat in Sancti Angeli hortis, Valley of Mex

meee ones rosmarinifolium S. i a Pl. N. "Hisp. 20 (1888), ed. 2, 19 (1893).—

in O rio, Michoacan. Not L. rosmarinifolium S. & M. Fl. Mex.
ae een ed. 2, 30 (1894), eae cone Heliotropium oaxacanum DC.

Three specimens from Madrid have been seen, viz., 1389 (Heliotropium
mexicanum), 1734 (Lithospermum angustifolium), and 5234 (Litho-
spermum angustifolium). Among the drawings at Geneva there is a poor
pencil sketch, no. 1174, labeled “Heliotropium mexicanum.” In both of
the large herbaria at Geneva there are specimens from the Pavon herbarium
bearing the name “Lithospermum angustifolium.” The description accom-
panying the name L. rosmarinifolium in the Plantae Novae Hispaniae is
a repetition of that associated with Lithospermum angustifolium in the
Flora Mexicana. The names, accordingly, must be synonymous.
Macromeria exserta Don

Echium longiflorum S. & M. Pl. N. Hisp. 20 ee ed. 2, 19 (1893).— Habitat in

Mazatlani et Chilpanzingi montibus, Guerr

One specimen from Madrid has been seen, “359 (Echium longifolium). In
the library at Geneva there is a fine original plate (no. 903 of the DeCan-
dolle series) which bears Sessé & Mocino’s original number, no. 293, and
their original name, Echium longiflorum. This latter plant-number is
cited with the published description of Echium longiflorum.

Macromeria longiflora Don

Lithospermum flavum S. & M. FI. Mex. 32 hae ed. 2, 30 (1894). — Mountains

between Zitacuaro and Malucatepec, Michoa

One specimen from Madrid has been seen, 1738 (Lithospermum longi-
florum). There is a good plate at Geneva, no. 905, determined by Sessé

110 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

& Mocifio as a species of Philonomia. DeCandolle has identified it with
Macromeria discolor Benth. A specimen in the Boissier herbarium from
Pavon was determined as Lithospermum longiflorum as was also Don’s
type of Macromeria longiflora at the British Museum. It should be noted
that M. longiflora Don was formerly misunderstood and that, as has been
shown, Contr. Gray Herb. 92: 93 (1930), it properly applies to the well-
marked species later described as M. discolor Benth
Macromeria viridiflor

? Lithospermum ioniorim S. & M. Fl. Mex. 32 (1893), ed. 2, 29 (1894).— Santa

Rosa near Guana

Only one specimen ton Madrid has .been seen, 5231 (Lithospermum
longiflorum). There are two plates of the species at Geneva. Upon them
DeCandolle founded his M. viridiflora. One is plate no. 904 of the folio
series, and the other is plate 12, f. 1 of the volume of Sessé & Mocifio’s
generic analyses. They were treated as members of the new genus Philo-
nomia and associatd with the specific name adopted by DeCandolle. The
two plates and the specimen from Madrid are evidently conspecific with
the plants subsequently treated as Macromeria (or Onosmodium) Thurberi.
This latter species, however, is known only from southern Arizona and New
Mexico and south along the northern Sierra Madre Occidental to southern-
most Chihuahua. In the absence of modern collections from more southern
localities it seems well to question old records attributing the species to
central Mexico. Rickett, Chron. Bot. 11 !: 29 (1947), has given evidence
that Mocifio and his artist traversed the area in Chihuahua in which the
species is well known. I am inclined to believe that they may have ob-
tained it there.

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXX APRIL 1949 NuMBER 2

STUDIES IN THE BORAGINACEAE, XVIII
BORAGINACEAE OF THE SOUTHERN WEST INDIES

IvAN M. JoHNSTON

PRELIMINARY TO THE PREPARATION of an account of the Boraginaceae
of Trinidad and Tobago it has been found desirable to survey the rep-
resentation of the family in the West Indies. Some of the results are here
presented in a synopsis of the species known from the Lesser Antilles,
Tobago, and Trinidad. Special attention has been given to the details of
distribution in the southern West Indies and to the taxonomic history of
the species concerned. The study is primarily based on the collections of
the Gray Herbarium (G), but numerous specimens from the New York
Botanical Garden (NY), the United States National Herbarium (US),
and the Trinidad Botanic Garden Herbarium (T), have also been examined.

KEY TO GENERA
Style twice cleft, with the four branches each bearing a stigma...1. Cordia.
Style cleft or simple.
Style evidently cleft, bearing two stigmas.
Corolla subrotate; flowers and leaves produced on dense axillary
short-shoots; plant frequently spiny........ : ochefortia.
Corolla salverform; plant unarmed................ 3. Bourreria.
Style simple or very short or none; stigma annular, usually surmounted
by a sterile, conic or cylindric, occasionally lobed appendage.
Fruit with fleshy mesocarp; large shrubs........ 4. Tournefortia.
Fruit dry, without any fleshy mesocarp; herbaceous............
Ec cna CRE aa SRO re IO es 0 ti Ware pity te ee . Heliotropium.

LesCORDIEA

KEY TO THE SPECIES
Corolla vermilion, about 4 mm. long, frequently 6-lobed; mature fruit hard,
ry, ovoid, 20-25 mm. long, completely and tightly invested by the
calyx that becomes white and fleshy............... 1. C. Sebestena.

LZ JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Corolla white or cream-colored, 5-lobed; cae not fleshy at leer

Pubescence stellate; corolla marcescent, persistent; fruit slender, elon-
gate, thin-walled, dry, developing re by aie tube and
10-ribbed cylindrical calyx, when fully mature the whole flower
falling with the enclosed fruit, oe stil spreading corolla-lobes
TUNELIONING Be A PATaCNUtes ato ss eee ereu es ee a. . alliodora.

Pubescence simple or none; corolla withering, ‘falling after anthesis ;
ruit drupaceous.

Corolla very large, 35-50 mm. long, the slender exserted tube
abruptly and broadly expanded into a well-developed cylin-
drical throat; inflorescence capitate; tips of calyx-lobes con-
spictously lone-attentidtes si.avcescus bes 3. C. grandiflora.

Corolla less than 15 mm. rae besa much shorter, without a
well-differentiated tube ©)

Fruiting calyx half to Kee ed investing the drupe,
the latter red, 4-6 mm. long, with a thin quickly dessi-
cated flesh; stone irregularly warted; corolla-limb usu-
ally only shallowly and very ‘broadly lobed: plant usually
with as

Inflorescence not distelly spicate.
Calyx-lobes with prolonged hairy setaceous tips about
5 mm. long; atone ene a globose head 1.5-3
Gil. Ti Wa metEy 4s k eee necaes 4. C. globosa.
Calyx-lobes acute, without prolonged tips.
Upper leaf-surface bearing short stiff hairs in
great abundance; inflorescence terminal or

Ame

maturity; corolla with reflexed, a
lacerate-dentate lobes about as lon
Breil oc uraravecweet anes C. salv ifolia,
Upper leaf-surface glabrous or nee so, stiff
hairs few, scattered and inconspicuous; in-
florescences mostly axillary, simple or spar-
ingly branched, the cymes capitate or glom-
erate and loosening little if at all in most
advanced maturity; corolla with obscure
lobes several times as broad as long, mar-
gins entire or nearly so..6. C. polycephala.
Inflorescence spicate.
Spikes mostly axillary, with the peduncle and the peti-
e of the subtending leaf united for 3-9 mm.
at their base; upper surface of leaves more or
ney aaah with scattered tightly appre
ren ere ee i oe i
Spikes : all teri or internodal, never axillary ; upper
surface of leaves glabrous or bearing ascending
or eee hairs.

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE. XVIII 113

Corolla-limb subentire, its very short inconspicu-
ous lobes several times broader than long;
calyx-lobes usually with short but distinct
tips that are free in the bud; spike dense,
in fruit 10-15 mm. thick; fruit matured in

inflated calyx....8. C. martinicensis.

Corolla aes the limb with evident reflexed
lobes about as broad as long; calyx-lobes
without prolonged tips that are free in the
bud; spikes frequently becoming loose and
interrupted, in fruiting state usually. 8 mm.
or less thick: fruit partially protruding
from the cup-shaped calyx

Leaves lanceolate to lance- EI, broadest at

ec

nate, upper surface usually glabrous or
rarely with short stiff hairs; corolla-

lobes entire or obscurely erose-dentate,
weakly ai “at ‘all icrisped:s. ss. 3

Bead ed eee C. curassavica.
Leaves broadest at or above riddle apex usu-
ally rounded, margin usually evidently
crenate or dentate, upper surface at

dent soft hairs; corolla-lobes usually

erose-dentate and crisped............

a ene seyrre ee 320% 2 0. C. divaricata.

Fruiting calyx usually explanate or saucer-shape, covering only
e very base of the drupe, the latter 8-16 mm. long and

with a very well-developed juicy mesocarp; stone ribbed

Calyx 10-ribbed, in the bud appendaged at apex, usually
irregularly circumscissile below summit: corolla
ee broadly funnelform, its lobes emarginate and
about as broad as long............. 11. C. dentata.

Calyx not eae ribbed, not appendaged at apex, usually
opening regularly and 5-lobed; corolla with cylin-
drical tube and spreading elongate lobes.

Petioles of well-developed leaves 1.5-4 cm. (usually

—3 cm

Leaves 5-11 cm. long, lower surface (as well as
petiole and branches of inflorescence) prac-

tically glabrous at maturity; fruit a very

plump and juicy drupe, 1.5 cm. me pink

or salmon in color......... ] iqua.

Leaves 10-20 cm. long, with brownish eee

branches of inflorescence: fruit a whitish
drupe, 8-10 mm. long... .13. C. tetrandra.

114 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xxx

Petioles of well-developed leaves 0.2-1.5 cm. (usually
0.5—
Leaves usually broadest above middle, apex acute
a coarse, generally obtusish short
prolongation; tree deciduous, producing its
inflorescence on bare twigs before or simul-
taneously with the appearance of the new
leaves of the season; fruit red; flowers
markedly heieremen one those on each tree
functionally either male or female.......
iaatGeeaed 4aaeeeeens 14. C. Collococca.
Leaves usually broadest below the middle, at least
the larger ones with a sharp prolonged at-
tenuate tip; plant bearing its inflorescence
on leafy shoots; fruit white or yellowish.
Lower leaf-surface with an evident pallid in-
dument of abundant minute whitish
hairs, the hairs confined to the veins
and veinlets and appressed and directed
concentrically over the numberless
areoles in the veinlet-reticulum; fruit
strigose; flowers homomorphic, per-
fect, functionally bisexual sues saa44

Lower leaf-surface velvety and ne

areoles in the veinlet-reticulum not

covered by appressed, ccna

directed hairs; fruit glabrou
Flower-buds large, 7-9 mm. ae broad-

fragmenting and erose; flowers
momorphic, all perfect and
functionally bisexual............

eA eee new ue es 16. C. elliptica.
Flower ee smaller, 2.5-4 mm. long;
with 5 small subequal teeth ;

iene markedly dimorphic, those

flowers with a glayaie erudite
bud, at anthesis with a cylindric
calyx, an inconspicuous corolla,
small stamens and evidently pro-

anthesis with a cup-shaped calyx,
evident corolla-limb, conspicuous-
ly exserted stamens and reduced
included female organs.

1949] JOHNSTON, STUDIES IN THE.BORAGINACEAE, XVIII 115

Larger leaves usually subcordate at

pre
Larger leaves acute, obtuse or
rounded at oe the tip at-

te and prolonged.

Leaf decidedly aes above,
rigid hairs only

moderately appressed ;
blade up to 15 cm. broad;
twigs, petioles, and inflo-
rescence bristly with stiff
hairs 1-2 mm. long, if
also (eae other hairs
these minute appressed

and (especially on the

twigs ) ee retrorse.
aden anamensis.
Leaf with very aie closely
appressed hairs and rath-
er smooth; blade up to 11
cm. broad; twigs, petioles
and inflorescence not at

appressed and
trorse.19. C. sericicalyx.
1. Cordia Sebestena L. Sp. Pl. 190 (1753).— Type West Indian.
Native of the West Indies and ranging from the Bahamas to northern
Venezuela; now widely cultivated in the tropics.
GUADELOUPE: Duss 2754 (G). MartTINIQUE: Hahn 763 (G). BARBADOS:
Eggers 7183 (G).
2. Cordia alliodora (R. & P.) Cham. ex DC. Prodr. 9: 472 (1845);
Urban, Symb. Ant. 8: 574 (1921); Johnston, Contr. Gray Herb. 73:
77 (1924) and lc. 92: 13 (1930).
Cerdana alliodora R. & P. FI. Peruv. 2: 47, t. 184 (1799). — Type Peruvian.
Cordia pene at var. subcanescens DC. Prodr. 9: 472 (1845). — Type
West India

116 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXX

Cordta Gerascanthus, var. domingensis Cham. ex DC. Prodr. 9: 472 (1845),

in synonymy.

Cordia Gerascanthus, forma inartinicensis Chodat, Bull. Soc. Bot. Geneve,

sér. 2, 12: 210 (1920). — — Type from Martinique, Hahn 626.

Cordia Gerascanthius, forma micrantha i at, Bull. Soc. Bot. Genéve,

sér. 2, 12: 210 (1920). — Type West India

Cordia Coviccaniiias of many authors, not a aie

Widely distributed in tropical America.

Montserrat: Shafer 557 (NY, US). Anticua: Box 1316 and 1342
(US); Rose, Fitch & Russell 3320 and 3456 (US). GuapeLoure: Duss
2580 (G, US, NY); Questel 694 (US); Stehlé 85 (NY), 126 (US), 559
(US), and 2684 (G). Dominica: Imray (G); Hodge 1627 (G). Mar-
TINIQUE: Duss 1425 (NY, US); Stehlé 6031 (G). St. Vincent: Smith
1249 (G, NY). Trinipap: Sieber 121 (G, NY); Broadway 5547 and 9845
(G); Fairchild 2863 (G

North of our area the species is known from the Virgins, Porto Rico,
Haiti and Cuba. Although reported from Jamaica it is almost certainly
not native there. The West Indian material of C. alliodora has its twigs
much less deformed by ant-domatia than that of Central America and
northwestern South America. The older writers identified the plant with
C. Gerascanthus, but that is a very different species native to Cuba,
Jamaica, and adjoining Central America. For an excellent very detailed
account of our tree see R. C. Marshall, Silviculture of trees of Trinidad
and Tobago, pp. 167-172 (1939). He reports the plant from Tobago as
well as from Trinidad.

3. Cordia grandiflora (Desv.) HBK. Nov. Gen. et Sp. 3: 77 (1818);
Johnston, Jour. Arnold Arb. 16: 32 (1935).
Varronia grandiflora Desv. Jour. de Bot. 1: 273 (1809).— Type from
Venezuela.

Ranging from British Guiana to central Venezuela and southward into
the Amazon Basin; Trinidad.

TRINIDA Caparo, June 1918, Broadway (T); “Trinidad,” Aug. 1928,
Abrahams, "Herb. Trin 12215 (TY;

According to Williams & Williams, Useful and Ornamental Pl. Trinidad,
ed. 3, p. 112 (1941), the plant is frequently found in peasant gardens in
southern Trinidad.

4. Cordia globosa (Jacq.) HBK. Nov. Gen. et Sp. 3: 76 (1818).

Varroma globosa Jacq. Enum. 14 (1760) and Sel. Stirp. 41 (1763). —In

Caribaearum maritimis
Varronia dasycephala Des, Jour. de Bot. 1: 274 (1808).— Habitat in
Cumana, Antigua, arenosis humidis.

Cordia dasveephaln (Desv.) HBK. Nov. coy et Sp. 3: 76 (1818

Cordia Ss a Humb. ex R. & S. Syst. 4: 801 (1820). tg: areno-

sis Cum :

Varronia ieee ephala Willd. ex HBK. Nov. Gen. et Sp. 3: 453 (1820).

Cordia bullata, var, angustata DC. Prodr, 9: 496 (1845).— Type from

Guadeloupe.

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 117

Ranging south through the Lesser Antilles into South America.

ST. BARTHELEMY: Questel 578 (US). St. Kitts: Britton & Cowell 257
(US); Rose, Fitch & Russell 3240 (US). Anticua: Box 1160 (US);
Rose, Fitch & Russell 3330 (G, US). Montserrat: Shafer 232 and 510

US). GuapELouPE: Duss 2575 (G, US); Stehlé 2681 (G). Dominica:
Hodge 2769 and 3861 (G). Martinique: Duss 652 (US).

The species as here accepted includes plants from the Lesser Antilles,
Trinidad, Venezuela, and eastern Brazil. Related plants of the Greater
Antilles, Mexico, and Central America are separated as C. globosa var.
humilis (Jacq.) Johnston, Jour. Arnold Arb. 30: 98 (1949). They are
distinguished by their leaves, which not only average much smaller than in
typical C. globosa, but also have the marginal teeth and apex obtuse rather
than acute. Also related to C. globosa are C. caput-medusae Taub., a
very glandular plant of Brazil, and C. swbtruncata, a Colombian plant with
pallid strigose herbage. There are some other closely related but sep-
arable plants in Colombia and Venezuela, but these are as yet unnamed.

No type of Varronia globosa is preserved. It is necessary to interpret
the species entirely from Jacquin’s short diagnosis and vague remarks.
They are as follows: ‘‘vARRONIA (globosa) spicis globosis, aequalibus.
Frutex hic perfecte congruit cum Varronia martinicensi. Discrimen solum
est in constanti globositate spicarum, in corollarum laciniis emarginatis
ipsis, & in stigmate obtuso quadruplici. Habitat in Caribaearum mari-
timis.” Curiously, there is no mention of the setaceous calyx-lobes, prob-
ably the most distinctive feature of our species, cf. Johnston, Jour. Arnold
Arb. 30: 102 (1949). Jacquin’s statement that his V. globosa agrees
closely with V. martinicensis, except as to inflorescence and floral struc-
tures, offers the most help in placing it. The foliage of typical C. globosa,
as I have delimited it, agrees well in size, shape, and dentation with that
shown in Jacquin’s illustration of the type of V. martinicensis. Since V.
globosa is given by Jacquin as Caribbean (i.e., West Indian and not conti-
nental), it very likely came from the Lesser Antilles, where he did much
collecting and could have compared growing plants with C. martinicensis.

5. Cordia salvifolia Juss. ex Poir. Encyc. 7: 46 (1806). — Source of type
not given

Endemic to the Lesser Antilles.

BUDA: Box ae (US). Antigua s Bow 1054 (US); Rese Fitch &
Ren 3339 (G, US). Montserrat: Shafer 495 (US). GUADELOUPE:
Duss 3959 (US) ; Stehlé 2025 (US), 2689 (G), 2698 (G), 6161 (G). Marie
GALANTE: Stehlé 165 (US), 2678 (G). DeésrtrapvE: Stehlé & Quentin 5340
(US). Martinique: Duss 287 (US). St. Lucta: Beard 1080 (G).

A very distinct member of the section Varronia which is endemic to our
area. It is readily distinguished from all immediate relatives by its large,
loose, repeatedly forked cymes. Many of the specimens have parasitized
fruit, which eventually contains smooth ellipsoidal pupae suggestive of
seeds. Normal fruit, however, has the irregular tuberculate single-seeded
stone characteristic of all Varronias.

118 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx
6. Cordia polycephala (Lam.) Johnston, Jour. Arnold Arb. 16: 33
(1935).

Varronia eae Lam. Tab. ee 1: 418 (1791); Poir. Encyc. 4:
6 98).— Type from “Amer

Varronia paniculata Wikstrom Oty, Guadeloupe. Fl.], K. Vet. Akad.
Handl. 18271: 59 (1828).— Type from Poe ee Forsstrom,

Cordia Wickstroemii Steud. Nom. ed. 2, 1: 419 (1841). — Based on Var-
ronia paniculata Wikst. not Cordia a Roth.

Cordia Wikstromii Steud. ex DC. Prodr. 9: 495 (1845).

Cordia sulfurata Krause, Beih. Centralb. 322: 341. (1914).— Type from
St. Vincent, Krause 11821.

ar wlmifolia, var, ovata DC. Prodr. 9: 495 (1845).— Type West

Indian

Cordia winapona: var. ovalis DC. Prodr. 9: 495 (1845).— Type West

nN ian

Ranging from the southern parts of the Dominican Republic south
through Porto Rico, the Virgins and Lesser Antilles into northern South
America.

ST. BARTHELEMY: Stehlé 7076 (G). Sr. Kitts: Britton & Cowell 203
(NY). GuapeLoupe: Stehlé 3 (NY), 1100 (US), 2614 (US), 2680 (G),
2695 (G); Duss 2576 (G, NY, US). Dominica: Cooper 22 and 141 (G
NY, US); Lloyd 420 and 688 (NY); Hodge 792 (US), 793 (NY, US),
1451 (G), 2217 (G), 2319 (G), 2905 (G), 3900 (G). Martinigur: a
873 (G, US); Duss 1421 (G, NY, US); Stehlé 3646 (G). St. Vincen
Eggers 6551 (G); Smith 781 (G); Beard 1347 (G). GRENADA: ae
(G,. US, NY),

The name C. polycephala is here restricted to the more southerly and
easterly of the two West Indian species formerly included in the aggregate
: tropical American shrubs assembled under the names “C. corymbosa”

r“C. ulmifolia.” It ranges from the Dominican Republic south through
our area into northern South America. On the southern continent it is
most abundant in Venezuela, but it also occurs west to northern Colombia
and in Brazil at least as far south as Ceara. The species is reported from
Trinidad, but I have seen no material from the island and doubt that it is
native there. It has recently been reported from St. Lucia, Stehlé, Carib.
Forester 8: 106 (1947).

The former aggregate species of which C. polycephala, sensu stricto, is a
small part, breaks up into two well-marked groups distinguishable by
differences in inflorescence. In one the cymes are chiefly axillary, only
the first produced by the shoot being terminal. In the other the cymes are
all terminal, but, with the elongation of the shoot, come to be borne on
stem internodes without any subtending leaf. Cordia polycephala, as
here narrowly delimited, has an axillary inflorescence as does also C. lineata
(L.) Don and C. patens HBK. These three segregates of the old complex
may be at times difficult to distinguish in Venezuela but elsewhere are rec-
ognizable at a glance. They have very different patterns of geographical
distribution. Cordia polycephala extends north via the Lesser Antilles to

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 119

Hispaniola. Cordia lineata appears in the northern and western parts of
Hispaniola and extends to Cuba, Jamaica, Central America, and, appar-
ently, Venezuela, cf. Johnston, Jour. Arnold Arb. 30: 92 (1949). Cordia
patens, in contrast, is a shrub of the Amazonian headwaters and extends
from the Guianas to eastern Peru.

The segregates of the old complex having extra-axillary inflorescences
have their center of distribution south of the Equator. Cordia bifurcata
R. & S. ranges from Costa Rica and Colombia south along the Andes into
Argentina, cf. Johnston, Jour. Arnold Arb. 30: 90 (1949). Cordia bolivi-
ana Gandog. is Bolivian. Cordia urticifolia Cham. is a coarsely strigose
representative of the group in Brazil and Paraguay. Another member is
C. discolor Cham. of Brazil. Of all these southern representatives only
C. discolor shows instability as to type of inflorescence. It is the only one
of the southern group which now and again makes embarrassing approaches
towards C. polycephala.

The first name given to our plant appears to be Varronia polycephala
Lam., which was described as follows: “1887, Varronia polycephala. V.
foliis ovato-lanceolatis, serrati; pedunculis lateralibus; spicis globosis.
Ex America. kh Pluk. t. 328 f. 52 Il varie & pédoncules rameaux.”
The brief description certainly applies to some member of the former aggre-
gate of which our plant is a part. Poiret, Encyc. 4: 263 (1797), so
accepted it and so have all subsequent authors. I am assuming that La-
marck’s type came from the French West Indies. If investigations at
Paris prove this to be incorrect, then the next available name for our species
is C. Wickstroemti Steud.

A few comments should be made regarding the name Varronia mono-
sperma Jacq. Pl. Rar. Hort. Schoenbr. 1: 18, ¢. 39 (1797). The material
upon which that species is described is said to have come from Caracas,
Venezuela, a locality at which C. polycephala has been repeatedly collected.
Jacquin’s beautiful and very detailed plant of V. monosperma is unlike
any plant I have seen from Central America, the West Indies, or northern
South America. In fact it seems to represent the Brazilian and Para-
guayan plant described as C. urticifolia Cham. (1829). The illustration
shows very clearly the extra-axillary, internodal or oppositiflorous in-
florescences that characterize the relatives of true C. polycephala south of
the Equator. Jacquin’s name is, in fact, the oldest name applied to that
group. It has exact synonyms in C. monosperma (Jacq.) R. & S., 1819,
Varronia corymbosa Desv. Jour. de Bot. 1: 275 (1809), and Cordia
corymbosa (Desv.) Don, 1838.

7. Cordia Schomburgkii DC. Prodr. 9: 490 (1845); Johnston, Jour.
Arnold Arb. 16: 39 (1935).— Type from British Guiana.
Cordia tobagensis Urban in Fedde, Repert. 16: 39 (1919).— Type from
Tobago, Broadway 3072.
Cordia tobagensis, var. Broadwayi Urban in Fedde, Repert. 16: 40 (1919),
— Type from Tobago, Broadway 4235.
Known only from the Guianas and from Tobago.

120 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

aco: Dennet Estate, Broadway 4235 (G, pt. of Tyre); Forest Reserve
beyond Caledonia, Broadway 3072 (G, pt. of Type); Menna Road near Mason
Hall, low shrub, fl. white, June 6, 1925, Williams, Herb. Trin. 11140 (T);
Mt. St. George-Castara road, Oct. 18, 1937, Cheesman, Herb. Trin. 13208
(fT)
8. Cordia martinicensis (Jacq.) R. & S. Syst. 4: 461 (1819).
Varronia martinicensis Jacq. Enum, 14 (1760) and Sel. Stirp. 41, ft. 32
(1763).— Type from Martinique.

Known only He Dominica, Martinique and St. Lucia.

Dominica: Hodge 2543 and 2663 (G). Martinigue: Duss 1423 (G,
US); Hahn 286 (G, US); Sieber 60 (G); Batley 252 (US); Stehlé 1040
(G,.US), 2125 (NY), 3470 1G), 3042 0G), and 0925 (G), St. Lica:
Beard 1011 (G).

9. Cordia curassavica (Jacq.) R. & S. Syst. 4: 460 (1819); Johnston,

Jour. Arnold Arb. 30: 99 (1949).

Varronia curassavica Jacq. Enum. 14 (1760) and Sel. Stirp. 40 (1760). —
Type trom Curacao.

Varronia macrostachya Jacq. oe 14 (1760) and Sel. Stirp. 41 (1763).
— Type from Cartagena, Colom

Cordia macrostachya (Jacq.) R o S. ‘Syse 4: 461 (1819) ; Johnston, Jour.
Arnold Arb. 16: 36 (1935).

Varronia guianensis Desv. Jour. de Bot. 1: 270 (1809).— Type from
French Guiana.

Cordia canescens HBK. Noy. Gen. et Sp. 3: 73.(1818).— Type from
Ibaque, Colombia.

Cordia graveolens HBK. Nov. Gen. et Sp. 3: 74 (1818).— Type from

between Ferreras and Angostura, Venezuela.

Cordia cylindrostachya, var. graveolens (HBK.) Griseb. Fl. Brit. W. I.

480 61).

Cordia spicata Willd. ex R. & S. Syst. 4: 799 (1819).— Type from Angos-

tura, Venez

uela.
Cordia rugosa Willd. ex R. & S. Syst. 4: 801 (1819). — Type South Amer-

Cordia interrupta DC. Prodr. 9: 491 (1845).— Type from French Guiana.
Cordia cylindrostachya, var. interrupta (DC) Griseb. Fl. Brit. W. I. 480
1861

(
Cordia sauna DC, Prodr, Os pe (1845).— Type from British Guiana.
Lithocardium cylindrc n, var. platyphyllum Kuntze, Rev. Gen. 2

438 (1891). ape from Trinidad.
— ial Pittier, Contr. U. S. Nat. Herb. 18: 253 (1917). — Type
m Pana
Cordia littoral's Pittier, Contr. U. S. Nat. Herb. 18: 253 (1917).— Type
from Costa Rica.
Ranging in northern South America (French Guiana to Colombia) and
north into Central America and the Lesser Antilles.
MarTINIQUE: Duss 289 (NY). St. Vincent: Smith 470 (G, NY).
Grenava: Eggers 6139 (G); Broadway 141 (G, NY). Barsapos: Dash
14 (NY). Topaco: Eggers 5458 (G); Elmore (G); Broadway 4254 (G,

1949 ] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 121

NY). Trrnipap: Johnston 78 (G); Broadway 11 (G); Britton & Hazen
18 (G); Riley 68 (NY).

The plant has sharp-pointed, lanceolate or lance-ovate leaves which are
green and usually glabrous above. The upper leaf-surface may be smooth
but commonly is more or less roughened by minute siliceous tuberculations
present there in varying abundance and development. Occasionally the
tuberculations become conic or even prolonged into short stiff bristle-tips.
The surface of the leaf, however, is never soft hairy nor more or less
velvety. The xerophytic form of the species has thickish ianceolate
leaves 1.5—4 cm. broad, roughened on the upper face by an abundance of
tuberculations. It is the form represented by the type of Varronia curas-
savica. In more sheltered places and especially those with much rainfall,
the plants produce larger, proportionately broader and thinner leaves, 4—9
cm. wide, on which the upper surface is nearly bare or has tuberculations
much reduced in size and number. Its spikes also tend to become very
elongate. This is the form described as Varronia macrostachya Jacq. An
extreme phase of this broad-leaved form was described from Trinidad as
Lithocardium cylindrostachyum, var. platyphyllum Kuntze. For our pur-
poses it may be called Cordia curassavica, var. platyphylla (Kuntze),
comb. nov. As the name for the mesophytic phases with broad leaves it
is applicable in our region only to plants collected on Trinidad and Tobago.

The present species, though not distinguished from C. cylindrostachya
by many West Indian botanists, is readily separable from that more south-
erly ranging South American species by its complete lack of axillary spikes.
True C. curassavica is very different in appearance and readily separable
at a glance from the various closely related but as yet imperfectly defined
species that occur north of it in the Greater Antilles and in Mexico. These
latter are excluded from C. curassavica as here accepted. Our species is a
natural’and practicable one that occurs in northern South America from
the Guianas to Colombia and extends north into Central America and the
Lesser Antilles. A-complete listing of its synonyms has been given above.

10. Cordia divaricata HBK. Nov. Gen. et Sp. 3: 74 (1818); R. & S.
Syst. 4: 802 (1819); Johnston, Jour. Arnold Arb. 30: 101 (1949).
— Type from Cumana, Venezuela.

Cordia cunetformis DC. Prodr. 9: 492 (1845).— Type from Caracas,

Venezuela.

Northern Venezuela and adjacent Colombia, Curacao, Martinique, Do-
minica.

Dominica: Hodge 3808 (G); Lloyd 839 (US). Martinique: Duss
288 and 289 (NY); Stehlé 3471 (G).

A relative of C. curassavica distinguishable by its usually smaller, gen-
erally more or less oblanceolate or obovate leaves, usually soft hairy on
the upper surface. Its calyx tends to have lobes less sharply triangular
and proportionately broader than in its relative.

122 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

11. Cordia dentata Poir. Encyc. 7: 48 digae Johnston, Jour. Arnold
Arb. 21: 347 (1940). — Type from Curaca

“Cordia alba’ of most authors.

West Indies, Mexico, Central America, and northern South America;
frequently cultivated.

T. BARTHELEMY: Questel 682 (US). GuapbELouPE: Duss 2757 (US);
Questel 824 (US); Stehlé 950 (US), 2691 (G), 2843 (G). MartTINIQUE:
Duss 285 (US). Antigua: Box 1188 (G, US). Barsapos: Waby 68
(US); Warming 98 (US). Tosaco: Broadway (G). Trinmap: Chaca-
chacare Isl., Finlay, Herb. Trin. 1842 (T).

Readily recognized because of its toothed, slenderly petiolate leaves,
broad funnelform shallowly lobed white corollas, ribbed calyx, and large
watery white fruits. In West Indian botany the plant is well known under
the name Cordia alba or Calyptracordia alba, names unhappily not prop-
erly applicable to it.

12. Cordia obliqua Willd. Phytogr. 4, ¢. 4 (1794) and Sp. Pl. 1: 1072
(1797). — Type from western Tridis,

Cordia tr ja Griseb. Fl. Brit. W. I. 479 (1861).— Type from Bar-

bados, Lar

An Indian tree introduced into the West Indies, where it is now widely
distributed.

SP. iris: Hee 2635 (G, US); Britton & Cowell 135 (NY). AN-
TIGUA: Box 1428 (US); Rose, Fitch & Russell 3370 (US). MONTSERRAT:
Shafer 124 bei 209 (NY, US). GuapELoure: Hahn 953 (US); Stehlé
260 (US), 2683 (G). Marie Garante: Stehlé 195 (US). St. VINCENT:
Smith 462 (G). Grenapa: Beard 34 (G); Broadway 1810 (G, US).
Barpapdos: Bovell 466 (NY); Gooding (G); Eggers 7179 (G, US).

For the most of a century this tree has been known in the West Indies as
C. tremula Griseb. and accepted as a native species, endemic to the Lesser
Antilles. It is, however, a native of India and was introduced into the
West Indies, probably during the eighteenth century, and most likely at
Barbados. Mr. E E. G. B. Gooding (in lit., June 1942) has some significant
observations concerning the plant on Barbados: He notes that the species
is absent from the few remnants of original forest on the island. It grows
“wild” only on open waste land, seldom far from habitations, and is fre-
quently cultivated. This ability to go wild is also demonstrated in Cuba.
The trees in the Botanic Garden at Soledad, originating from Lesser An-
tillean seed, now have progeny in waste land outside the garden.

The name Cordia obliqua Willd. here applied to our plant is subject to
future revision, although it certainly belongs to the particular form of the
Indian plant represented in the Caribbean area. Unfortunately, it is only
one of a confusing complex of many closely related forms represented in
the tropics of Asia and Malaysia. Until the complex is thoroughly studied
and a modern classification of it available, names applied to its forms can
be only tentative. Nevertheless I am inclined to believe that the name
selected may stand. The Indian tree, along with its other relatives in

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 123

southeastern Asia and the South Pacific, are readily distinguishable by
form of style from the African and Levantine C. Myxa L., with which they
were formerly associated. For the eastern complex there is only one name
older than C. obliqua Willd. (1794), namely C. dichotoma Forst. (1786).
This latter was applied to a plant of the wetter tropics and is distinguish-
able from C. obliqua by its thinner elliptic leaves and smaller flowers.
The eventual monographer of the group will, I am confident, distinguish
these two extreme forms, if not as species, then certainly as varieties.

13. Cordia tetrandra Aublet, Hist. Pl. Guian. 1: 222, ¢. 87 (Chins
Johnston, Jour. Arnold Arb. 16: 11 (1935).— Type from French
Guiana.

Widely distributed at low altitudes in tropical South America, but not
common.

Barpapos: Dodd’s, St. Philip, 1902, Bowell 437 (NY).

The specimen cited is probably from a cultivated tree. The species,
however, has such a hit-and-miss distribution over so wide an area in trop-
ical South America that its natural occurrence a little further north, as a
rare tree in the southern West Indies, would not be surprising.

14. Cordia Collococca [Sandmark] L. Amoen. Acad. 5: 377 (1759), as
“C, Callococca.” — Type from Jamaica.
Cordia Collococca L. Sp. Pl. ed. 2, 274 (1762), excluding “Cordia glabra
Spe eicprdep Ole

Cordia Callogocca L. Syst. ed. 12, 176 (1767), excluding phrase “fol.

cordato-ovatis.”
Cordia micranthus Sw. Prodr. 47 (1788) and Fl. Ind. Occ. 1: 460 (1797).
— Type Jamaican.

Cordia ehretioides Lam. Tab. Encyc. 1: 421 (1791); Urban, Symb. Ant.
8: 576 (1921).— Type from Santo Domingo.

Cordia Collococca, var. ehretioides (Lam.) Poir. Encyc. 7: 42 (1806).

Cordia glabra of authors, not Linnaeus, cf. Johnston, Jour. Arnold Arb.
21: 345 (1940).

Cuba and Mexico south to northern South America.

GUADELOUPE: Duss 2577 (NY); Stehlé 91 (NY), 66 (NY), 2682 (G).
Dominica: Hodge 3890 (G). Martinique: Duss 282 (NY). ;
CENT: Smith 1546 (NY), 1836 (G, NY). Grenapa: Broadway (G, NY).
Barpapos: Gooding 609 (NY). Toraco: Montserin, Herb. Trin. 13647

T). Trinipap: Britton, Freeman & Watts 2692 (G).

The plant is deciduous and develops its inflorescences late in the dry
season on leafless or nearly leafless twigs, at least before the new growth
of leaves is expanded. In having precocious inflorescences and bright red
fruit it is readily separable from all other congeners in our area. In addi-
tion to those islands from which specimens are cited, the species is also
reported from St. Barthélemy, Antigua, Bequia, and Mustique. A good
account of the tree, with interesting field observations, has been published
by R. C. Marshall, Silviculture of trees of Trinidad and Tobago, pp. 172-3
(1939).

124 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXX

The name of this well-marked species merits a few words of discussion.
The first name applied to it was Cordia Callococca [Sandmark] Linnaeus
(1759). It is one of the many binomials applied by Sandmark to Ja-
maican species described and named under polynomials in Browne’s History
of Jamaica. In the present instance Sandmark’s page-reference to the
History of Jamaica is incorrect, being “166” and not 167. In such a very
abbreviated synopsis as that published by Sandmark this error might be
very confusing indeed. However, a perusal of his dissertation makes it
obvious that he intended his binomial, C. Callococca, to apply to the first
of the two species of trees for which Browne had erected the genus Col-
lococca. Sandmark’s specific epithet was derived from Browne’s generic
name. The latter was given in allusion to the glutinous pulp in the fruit
of the plant, a meaning lost by the vowel change in the specific epithet as
published by Sandmark. This vowel change, I believe, is another example
of typographical or clerical error in the dissertation. This is consistent
with the fact that when Linnaeus, adopting the binomial proposed by his
student, first treated the species in his own publications, he used the orig-
inal Brownean spelling in the specific name. This he maintained in the
third edition of the Species Plantarum, 274 (1764). Only in the last of
his major works, Syst. ed. 12, 176 (1769), is the specific epithet Callococca
used. Subsequent authors have spelled the name both ways. I have
adopted “Collococca” as the correct spelling for the species, since I believe
“Callococca” originated in error. The epithet has usually been capi-
talized and treated as a generic noun in apposition. As such, it is mean-
ingless unless it conforms to the original Brownean spelling.

Though there may be some doubt as to the correct spelling of Sand-
mark’s binomial, its correct application seems clear. The Jamaican tree
discussed by Browne, History 167 (1756), as “Collococcus. 1. Foliis rugosis
venosis oblongo-ovatis, floribus laxe racemosis,” is evidently the same as
the plant now under discussion.

In the past Sandmark’s publication (Dec. 1759) has been overlooked,
and the name Cordia Collococca has been accepted as established in the
second edition of the Species Plantarum (1762). This publication of 1762
also has its complications. As has been discussed elsewhere, Jour. Arnold
Arb. 21: 345 (1940), by some error the name Cordia glabra L., Sp. PI.
191 (1753), was listed as a synonym of Cordia Collococca rather than
under Ekretia Bourreria L., where it properly belongs. With this incorrect
synonym excluded, Cordia Collococca L. (1762) is based partially on the
reference to Browne’s History already used by Sandmark and partially
on a reference to Sloane’s great book on Jamaica, 2: 95, ¢. 203, f. 2
(1725). Sloane gives a good illustration and discussion of our plant.

15. Cordia bicolor A. DC. Prodr. 9: 485 (1845); Johnston, Jour. Arnold
Arb. 16: 23 (1935) and 21: 349 (1940). — Type from Dutch Guiana.
Lithocardium Lockartii Kuntze, Rev. Gen. 2: 438 (1891).— Type from
Trinidad, Lockart.
Cordia Lockartiu Kuntze, Rev. Gen. 2: 438 (1891), in synonymy.

1949 ] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 125

Northern South America and northward in Central America; Trinidad.

St. VINCENT: Anderson (Kew). Trintpap: Britton & Broadway 2797
(G) ; Britton 2500 (G, NY); Britton & Hazen 1926 (G, NY); Britton 549
(G, NY); Eggers 1431 (Kew); Eggers 1178B (Berlin); Lockart (Kew,
TYPE).

The specimens cited from St. Vincent probably came from a cultivated
tree. Alexander Anderson was director of the botanical garden on St.
Vincent from 1785 to 1811. I doubt if C. bicolor occurs naturally in the
West Indies north of Trinidad. Some excellent field observations concern-
ing the Trinidad tree are given (sub “Cordia Lockhartit”) by R. C. Mar-
shall, Silviculture of trees of Trinidad and Tobago, pp. 173-4 (1939).
Marshall states that the ripe fruit is a greenish yellow drupe, ovoid-rounded,
12 mm. long, and covered with minute hairs. Its stone is said to be 10
mm. long, pointed at the apex, and irregularly furrowed.

16. Cordia elliptica Sw. Prodr. 47 (1788) and Fl. Ind. Occ. 1: 461
(1797); Urban, Symb. Ant. 3: 357 (1903).— Type from “Jamaica.”’
Cordia reticulata Vahl, Ecol. 3: 5 (1807). — Type from Montserrat, Rvan.

Cordia laevigata of authors.

A species endemic to the Lesser Antilles.

MONTSERRAT: Shafer 345 (NY). GuapeLoupre: Fairchild (G): Duss
2584; Stehlé 2679, 2692, 2693, and 2844 (G). Dominica: Cooper 58 and
142 (G, NY); Beard 1464 (G); Lloyd 244 (NY); Hodge 1126, 1809, 2093,
2247, 2349, 2612, 3012, 3359, and 3458 (G). Martinique: Duss 242 (NY),
1422 (G, NY). Sr. Lucia: Beard 495 (G).

I am following Urban in identifying our present plant with C. elliptica
Sw. The specimen from the Swartz herbarium at Stockholm that appears
to be the type of C. elliptica is labeled as from “Jamaica.” It has an old
inflorescence with some persisting, apparently blighted calyces. There
are no corollas nor fruit. The indument on the inflorescence and calyces
is sparse, but if these are persisting old ones found on a tree out of flower,
that is not surprising. Although our species does have a dense strigose
indument on new flowering inflorescences, the hairy covering thins out
considerably on old inflorescences and even on fruit-bearing calyces. The
herbage and stems of the Swartz type are indistinguishable from those of
the plant of the Lesser Antilles. The type agrees well with plants from
our area but is very different and readily separable from all members of
its genus known from Jamaica. I am willing to believe that the type of
C. reticulata was attributed to Jamaica through some error. It was more
likely collected on Dominica, probably by Ponthieu.

The name Cordia laevigata Lam. Encyc. 1: 422 (1791); Poir. Encyc.
7: 46 (1806), has been applied to our plant. Judging from the original
descriptions, however, Lamarck’s plant must be very different from C.
elliptica, particularly as to flowers. Indeed, it appears to be the same as
the more northern C. nitida Vahl (1793), and being older, probably should
be taken up in place of the latter.

126 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Among West Indian species our plant is notable for its firm, glabrous,
lustrous, strongly acuminate leaves, its elongate flower-buds that are
clothed with abundant short appressed sericeous uairs, and its large gla-
brous drupe containing an obliquely ascending apically pointed stone. It
is a very distinct species, perhaps most closely allied to plants of Vene-
zuela and Colombia. Its drupe must be very large and conspicuous, but
little has been recorded by collectors or authors concérning the color, size,
configuration, and other features in the fresh state.

17. Cordia suleata DC. Prodr. 9: 488 (1845).— Type from Guadeloupe,
Bertero.

Ranging from the Barbados and St. Vincent north to Cuba.

Sasa: Boldingh 1477 and 1495 (NY). Sv. Kitts: Britton & Cowell 302
(NY). Antigua: Box 1491 (US). Montserrat: Shafer 475 (NY, US);
Ma'oney 420 (G). GuapDELouPE: Bertero (G, photo of TyPE); Duss 3079
and 3773 (NY, US); Stehlé 295 (US), 1078 (US); 265 (US ).2220-( 0s),
2677 (G). Dominica: Imray (G); Beard 661 (G); Narodny 1 (G);
Eggers 529 (G); Hodge 3154 and 3745 (G). Martinique: Duss 1426
(NY, US), 4502 (NY), 4703 (G, NY, US); Hahn 752 (NY); Stehlé 4468
(US). Sr. Vincent: Smith 771 (G, NY). Barpapos: Eggers 7294 (G,
WS).

This is a species endemic to the West Indies which is found in the Lesser
Antilles and in all the Greater Antilles except Jamaica. Its closest relative
is C. macrophylla L., an endemic of Jamaica, which differs only in having
much larger, more elongate leaves, and a shaggy-velvety indument on
twigs, petioles, and inflorescences. The only other close relative of C.
sulcata is C. panamensis of Trinidad, Tobago, and Central America. This
latter differs in its bristly twigs, very scabrous eae leaf-surface, and non-
cordate leaf-bases. These three species, C. sulcata, C. macrophylla, and
C. panamensis, are immediately related and are nearer by neither
flowers nor fruit. They have, however, distinct geographic ranges and are
separable by vegetative characteis, Tho ugh not “strong” species, they
are useful units worthy of continued recognition.

The type of C. sulcata DC. is an immature specimen with new, partially
expanded foliage and flower-buds. It was collected on Guadeloupe by
Bertero. In addition to the islands from which I have cited specimens,
the species is reported from St. Eustatius and St. Lucia.

18. Cordia panamensis Riley, Kew Bull. 1927: 125 (1927); Johnston,
Jour. Arnold Arb. 21: 347 (1940).— Type from Panama.
Known only from Central America, Trinidad and Tobago.
Trinipap: Southern Watershed Reserve, Marshall, Herb. Trin. 12433
(T); Southern Watershed as main ridge, tree 25 ft. tall, Marshall,
Herb. Trin. 12410 (T); S. W. erve, Russell, Herb. Trin. 12262 (T);
Quinam Road, Southern Range, a tree, Wilhams, si Trin. 12186 (T);
Maracas Bay, Dannouse, Herb. Trin. 6749 (T); Maracas, Herb. Trin. 1843
(T); Arima, Dannouse, Herb. Trin. 8986 (T); St. Patrick, Dardaine, Herb.
Trin. 11588 (T); Monos, Herb. Trin. 1845 (T); Botanic Carden near

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 127

Carpenter Shop, Broadway, Herb. Trin. 3833 (T); Botanic Garden, Herb.
Trin. 1050 and 5942 (T). Toxpaco: tree 40 ft. in woods, 800 ft. alt., Eggers
5590 (G

In the past this tree of Trinidad and Tobago usually has been identified

s “Cordia sulcata.” I am unable to distinguish it from C. panamensis,
a species previously recognized only in Central America. No species na-
tive to northern South America is readily confused with it. The relations
of the plant are with Cordia sulcata DC., which ranges in the West Indies
from Barbados to Cuba. Among its distinctive features are its bristly
twigs and inflorescence and its harsh, very scabrous upper leaf-surfaces.
Excellent field observations concerning the tree in Trinidad are given (sub
Cordia sulcata) by R. C. Marshall, Silviculture of trees in Trinidad and
Tobago, pp. 174-5 (1939). He reports the ripe fruit as a round, whitish,
translucent drupe that is smooth, about 1 cm. in diameter, and very sug-
gestive of a rather large English white currant. It contains a hard, woody,
irregularly shaped stone, about 6 mm. in diameter, embedded in muci-
laginous pulp.

19. Cordia sericicalyx A. DC. Prodr. 9: 485 (1845); ee Jour.
Arnold Arb. 16: 25 (1935).— Type from British Gui

Cordia ierensis Britton, Bull. Torr. Bot. Club 50: 54 (1923). ae from
Morne Bleu, Trinidad, Britton, Freeman & Bailey 2277.

Ranging from Dutch Guiana to western Colombia and north into
Trinidad.

TrinipAD: Morne Bleu, tree 10 m. tall, Britton, Freeman & Bailey 2277
(NY, type; G, T) ; Mount Tocuche, tree 15 m., corolla white, Britton, Hazen
& Mendelson 1344 (G, NY); Blanchisseuse Road near 9-mile post, small
tree, Broadway 6000 (G); St. Anne, Herb. Trin. 1848 (T); Maracas, Herb.
Trin. 588 (T); Maraval, Herb. Trin, 4533 (T); Arima-Blanchisseuse road,
14th-15th mile, Marshall, Herb. Trin. 12882 (T); Tucuche, Britton & Free-
man, Herb. Trin. 9094 (T); Tucuche, Williams, Herb. Trin, 11016 (T).

This species commonly has relatively thin, rather smooth, parchment-
like leaves that are usually markedly heteromorphic. The leaf-surfaces,
twigs, and inflorescences are very finely and minutely short-strigose. The
indument, though readily detected with a hand lens, is inconspicuous. At
a casual glance the mature twigs and leaves might pass as glabrous. Some
field observations concerning this species (sub Cordia ierensis) are given
by R. C. Marshall, Silviculture of trees of Trinidad and Tobago, p. 175
(1939). Marshall states that the drupe is green, about 1 cm. long, and
contains an irregular stone embedded in a mucilaginous pulp.

2. PROCHERORTIA

KEY TO THE SPECIES
Twigs very thorny; flowers nearly sessile, borne one or two together directly
On axillary ShOrt=shiGOts saree meee eee 1. R. acanthophora.
Twigs with few or no thorns; flowers borne several to many in a aie
PECUNCUIAe Cyan eee: epee ne ner rene nese ie clo) stele oie 2. R. cuneata,

128 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

1. Rochefortia acanthophora (DC.) Griseb. Fl. Br. W. I. 482 (1861).
Ehretia acanthophora DC. Prodr. 9: 510 (1845).— Type from Santo
Domingo.

AnTIcua: Box 858 (G).
The species is reported from St. Martin and St. Eustatius.

2. Rochefortia cuneata Sw. Prodr. 54 (1788) and Fl. Ind. Occ. 1: 552
(1897).— Type from Jamaica.
GUADELOUPE: Duss 2753 (G).
The species is also reported from Dominica and Martinique.

3. BOURRERIA

Bourreria succulenta Jacq. Enum. 14 (1760) and Sel. Stirp. 44 (1763).
— Type from Curacao.

Bourreria recurva Miers, Contr. 2: 234 (1869).— Type from Dominica.

Ranging from Florida to Venezuela and Panama, generally distributed
in the West Indies.

ANGUILLA: Boldingh 3518 (NY). St. Marrin: age 2602 (NY).
ST. BartHELeMy: Forsstrém (NY); Questel 90 (NY). Sr. Kitts: Brit-
ton & Cowell 366 (NY). Montserrat: Shafer 310 and 497 (NY). Guapr-
LOUPE: Duss 2752 (G, NY). Marie GaLtante: Stehlé 2837 (G). Domin-
Ica: fishlock 53 (NY); Lloyd 616, 661, and 664 (NY). Martrnigue:
Stehlé 6051 (G); Hahn 920 and 1095 (G); Bailey 203 (G). St. VincENT:
Smith 551 (G, NY) and 569 (G). Grenapva: Broadway (G). Topaco:
Broadway 9166 (G); Eggers 5517 (G); Williams, Hare, Trin, 11444 (T).
Trinipap: Maraval, Baptiste, Herb. Trin. 5860 (T).

Schulz in his motion ao of the genus, Urban Symb. Ant. 7: 58 (1911),
reports collections from St. Eustatius, Antigua, St. Lucia, Bequia, and
Mustique

The name of the present genus has variant spellings in “Bourreria,”
“Beurreria,” and “Beureria.”’ The first, however, is correct. The name
was proposed and first applied to our genus by Patrick Browne, Nat. Hist.
Jamaica 168 (1756), who spelled it “Bourreria.” This spelling was
adopted by Jacquin, Enum. 2 and 14 (1760), who was the first author to
use the concept and name in a publication employing binomial nomen-
clature. This act established the correct orthography. To be sure Jac-
quin, Sel. Stirp. 44 (1763), later adopted the spelling “Beurreria,” and sub-
sequent authors, notably O. E. Schulz, the spelling “Beureria,”’ but these
are not corrections of error but expressions of personal taste, and are not
sanctioned by rules of nomenclature. Cf. Kew Bull. 1935: 385 (1935).
BOURRERIA CUMANENSIS (Loefl.) O. E. Schulz.

A form of this species was described as Crematomia Guildingiana Miers,
Contr. 2: 246, t. 87 (1869), and given as having been collected on St.
Vincent by Guilding. If the plant is not mislabeled it almost certainly
represents, not a wild, but a cultivated plant, probably one from the botanic
garden that flourished on St. Vincent early in the nineteenth century. The

1949 ] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 129

species is known with certainty only from the dry north coast of Colombia
and Venezuela, and though it might be expected on some of the small
dry islands near Trinidad, it is certainly not to be expected on St. Vincent.

4. TOURNEFORTIA
KEY TO THE SPECIES
Leaves linear-spathulate, broadest below the apex and gradually contracted
towards the attachment, 3-10 cm. long, silvery silky from an abundance
of appressed hairs; fruit with a very thin flesh, having a basal central
cavity, ae tardily breaking up into a pair of two-seeded nutlets;
Sttandieplantansc cs Fk bat ny eda eee ous 1. TV. gnaphalodes.
Leaves with a Hoes blade and petiole; fruit with a very fleshy mesocarp,
with a lobed or unlobed stone not hollow at the base.

Fruit with a conspicuously 2-4-lobed stone; embryo curved; corolla-lobes
subulate or cuneate, greenish or yellowish; style elongate, falling
with the corolla.

Corolla-tube 3-8 mm. long, with a moderately expanded throat, lobes
not more than half the length of the tube, cuneate; ripe fruit

ange or yellow; leaves nearly glabrous....2. 7. maculata.
eee fine 1.5- 3 mm. One) peng at summit, Ho bes subulate,
nearly as long as tube; ripe fruit usually whi

Leaves 4-10 cm. broad, g ie m. long, leathery ; Bforeccai nce
narrow, stiff, the branches short, densely flowered and
divaricate, borne on an elongate axis...3. 7. subsessilis.

ae 0.54 ( 6) cm. broad, 3-10(-12) cm. long, firm but not

leathery ; inflorescence loosely branched. .4. JT. volubilis.
Fruit with ovoid stone, at most merely sulcate on the sides; embryo
straight; corolla-lobes elliptic to triangular-ovate, white; ripe
fruit white.
Style Somer 1-3 mm. long, usually falling with corolla; leaves
oe large, Suede 15-40 cm. long, and having oe 15 pairs
) jor veins; tree, up to 9 m. tall....4 4: us T. filiflora.
Style very Ane sessile and ieee on the mature fruit: leaves
5-20 cm. long, with 4-9 pairs of veins; plant usually scandent
or climbing

Foliage and fae glabrous or practically so: calyx-lobes 1-2
IT oe UT Op ees Ney Se a a ie Ne ee Gr er erbi color,

Foliage and twigs evidently hairy.

Calyx-lobes lanceolate, 2-4 mm. long: stems bearing short

curved grayish hairs Aboit 1 mm. long; corolla-tube

PTET asl O11 ss dae ave eB cetteneehes 7. T. hirsutissima.

Calyx-lobes subulate, 5-9 mm. long, slender and long-at-
nuate; stems shaggy, bearing abundant slender

brown hairs 2-4 mm. long; corolla-tube 5-8 mm.

LOM Opts ke Sins) Harpe ee oe eee 8. T. cuspidata.

1. Tournefortia gnaphalodes (L.) R. Br. ex R. & S. Syst. 4: 538 (1819).

Heliotropium gnaphalodes L. Syst. ed. 10, 913 (1759) ; Amoen. Acad. 5:
376 and 394 (1759).— Type West Indian.

130 JOURNAL OF THE ARNOLD ARBORETUM [VoL. xxx

Mallotonia gnaphalodes (L.) Britton, Ann, Missouri Bot. Gard. 2: 47
(1915).

Messerschnudia gnaphalodes (L.) Johnston, Jour. Arnold Arb. 16: 165
(1935).

A strand plant widely distributed in the West Indies.

GUADELOUPE: Stellé 2687 and 2842 (G). Marrinigue: Duss 246 (G).
MusTIQuE: Smith G26 (G).

In addition to the islands mentioned above it has been reported in the
Lesser Antilles from Anguilla, St. Martin, St. Barthélemy, Saba, St. Eusta-
tius, Antigua, St. Vincent, Bequia, Carriacou, and Barbados. It is not
reported from Trinidad nor Tobago, but is well known to the westward on
the islands off the Venezuela coast. Guppy, Plants, Seeds and Currents,
p. 247 (1917), has given observations regarding the behavior of the fruits
of T. gnaphalodes and its modes of distribution in the West Indian drift.

Considering the present status of classification in the subfamily Helio-
tropioideae it has seemed best to assign this species to Tournefortia. Tra-
ditionally in this subfamily those species with dry fruits have been placed
in Heliotropium and those with a fleshy mesocarp in Tournefortia. There
are reasons for believing that this may not be a natural division. However,
until the subfamily is thoroughly studied and a convincingly natural new
classification has been worked out, it seems best to continue the separation
of Tournefortia and Heliotropium in the traditional manner. To assign
the present species to the monotypic Mallotonia or to group it with two
Old World species in Messerschmidia is inconsistent when other equally
distinct groups of species are still submerged in the older traditional genus.

This West Indian species has a fleshy mesocarp, although a very scanty
one that soon dries. Otherwise the fruit is very similar to that found in
certain groups of Heliotropium. Its only distinctive feature is the presence
of corky tissue, a feature which I emphasized in assigning the plant to
Messerschmidia. This, incidentally, has its weakest development in our
West Indian plant

2. Tournefortia maculata Jacq. Enum. 14 (1760) and Sel. Stirp. 47
(1763).— Type from Cartagena, Culombia.
Tournefortia syringaefolia Vahl, Symb. 3: 23 (1794); Johnston, Jour.
Arnold Arb. 16: 48 (1935).— Type from French Guiana.
Tournefortia peruviana Poir. Encyc. Suppl. 4: 425 (1816); Urban, Symb.
Ant. 8: 586 (1921).— Type from Per
Tournefortia Sagraeana DC. Prodr. 9: 522 (1845).— Type from Cuba.
Tournefortia guadelupensis Urban, in Fedde Repert. 17: 169 (1921).—
Type from Guadeloupe, Duss 39092
Ranging from Cuba and southern Mexico south through the West Indies
and Central America into northern and western South America.
GUADELOUPE: Duss 3992 (G). Dominica: Hodge 2385 (G). Toxsaco:
Broadway 4606 (G). Trinipap: Broadway 6389 (G).
The long-neglected name 7. maculata Jacq. is evidently the oldest and
proper name for this widely distributed plant. Jacquin (1763) described

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVII 131

fruiting specimens from Cartagena, Colombia, as follows: “TOURNE-
FORTIA (maculata) foliis ovatis, utrinque glabris; spicis ramosissimis,
pendulis. Frutex ramis debilibus; foliis integerrimis, acuminatis, petio-
latis, inodoris; fructibus luteis basi quatuor maculis subrotundis atrisque
notatis. Habitat Carthagenae in arbustis & sepibus.” The glabrous
leaves and the yellow drupes splotched with black are distinctive features
of the present plant.

In the West Indies and Central America T. maculata presents problems
in delimitation only in Jamaica, Haiti, and Porto Rico. Plants from these
islands, described as 7. laurifolia Vent. and T. jamaicensis Urban, differ
from ordinary 7. maculata in their larger, long-attenuate calyx- and
corolla-lobes. In Jamaica and Haiti plants with short or elongate lobes
both occur, but in Porto Rico only the latter. The plants with elongate
attenuate lobes seem to deserve some nomenclatural recognition, but only
field work will determine whether this should be as species, variety, or
form.

3. Tournefortia subsessilis Cham. Linnaea 8: 119 (1833); Johnston,
Contr. Gray Herb. 92: 85 (1920).— Type from Brazil.

Known only from eastern Brazil and Trinidad.

Trrnipap: Balandra Bay, June 21, 1921, Freeman, Herb. Trin. 10431 (T):;
Balandra Bay, Dec. 8, 1933, shrub near sea, corolla reddish, Broadway 9331
(G); road near sea between Balandra and Toco, fl. yellow, fruit drupaceous,
lobed, apparently ripening white, Aug. 20, 1940, Cheesman & Baker 370,
Herb. Trin. 13735 (T).

I am unable to distinguish the above cited collections from the plant of
eastern Brazil (Bahia) known as T. subsessilis. The plant, though related
to T. volubilis, has a more compact, more elongate, and less branched in-
florescence that is made up of short, much more densely flowered scorpioid
cymes. Furthermore, its leaves are also very much larger. Indeed, in
general appearance the leaves are most suggestive of those of T. bicolor.
The blade is ovate to elliptic, 4-10 cm. broad, 8-18 cm. long, somewhat
leathery in texture, lustrous and practically glabrous above, and dull and
glabrescent beneath. Balandra Bay is on the Atlantic coast of Trinidad
about 10 miles from the northeast corner of the island. It lies about 2000
miles from the Brazilian localities at which the species has been previously
known. No plant readily confused with it is known north of the Amazon.

4. Tournefortia volubilis L. Sp. Pl. 140 (1753). — Type West Indian.
Tournefortia sericea Vahl, Ecol. 1: 17 (1796).— Type from Montserrat,

yan.
Messerschmidia punctata Spreng. Neue Entdeck. 3: 28 (1822). — Type
from Martinique, Steber.
Tournefortia punctata Spreng. Syst. 1: 643 (1825).
Tournefortia psilostachya, var. caribaea DC. Prodr. 9: 525 (1845).—
on material from Guadeloupe, Martinique, and Trinidad.
Tournefortia caribaea (DC.) Griseb. Fl. Brit. W. I. 484 (1861).

132 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Tournefortia barbadensis N. E. Brown ex Britton, Bull. Torr. Bot, Club
48: 343 (1922).— Type from Barbados, Bovell & Freeman 404.
Tournefortia trinitatis Riley, Kew Bull. 1925: 139 (1925).— Type from
Trinidad, Riley 210.

An extremely variable species ranging from Florida and Mexico south
through the West Indies and Central America into South America.

MonrTserrat: Ryan (G, photo of TYPE). GUADELOUPE: Stehlé 2685, 2686,
2832, and 2841 (G). Marte GaLante: Stehlé 2836 (G). Dominica:
Hodge 2533, 2542, 3083, 3084, and 3152 (G); Eggers 689 (G); Imray (G).
MARTINIQUE: Duss 4504 (G); Hahn 415 (G); Sieber 63 (G). GRENADA:
Broadway (G). St. VINcENT: Eggers 6600 (G); Smith 127, 1277, and
1540 (G). Barpavos: Eggers 7097 (G). Toraco: Broadway 3665 (G);
Eggers 5914 (G). Trtnipap: Chacachacare, Herb. Trin. 1858 (1 )s Fates
Island, Broadway 8908 (G) ; Gasparee Island, Britton 451 and 2781 CG
St. Anne, Broadway, Herb. Trin. 9203 eine

In one form or another this extremely variable species is to be expected
on all the islands in our area. In addition to islands from which speci-
mens are cited above, it has been reported from St. Martin, St. Barthélemy,
Saba, St. Eustatius, Desirade, St. Lucia, Bequia, and Mustique. The
species is the most variable member of its genus. Its leaves vary greatly
in size and shape as well as in quality, quantity, distribution and color of
pubescence. Forms of the plant may differ conspicuously in gross appear-
ance. With so many diverse phases it is not surprising that the species
has accumulated the most extensive synonymy in the genus. Above I
have given only those names that are based on plants originating in our
area. These apply to phases of the plant that are neither unusual in
appearance nor endemic to our region. As with most of the phases named
in other regions, the named ones from the West Indies can be expected to
appear here and there erratically at widely separated places within the total

geographical range of the species.

5. Tournefortia filiflora Griseb. Fl. Brit. W. I. 483 (1861); Urban,
Symb, Ant. 4: 522 (1910). — Type from Dominica, Imray.

Ranging from Porto Rico south to St. Vincent.

GUADELOUPE: Stellé 2697 and 2839 (G). Martz GALAnte: Stehlé 2838

1). Dominica: Hodge 1592 (G); Imray 322 (G). St. Vincent:
Smith 1542 (G).

As originally published 7. filiflora included not only our present plant
but also material of another species from Jamaica. I am following Urban,
I. c., who limited the species to the plant distributed from Porto Rico to
St. Vincent. With the type to be selected from among the non-Jamaican
specimens cited by Grisebach, the best choice is Imray’s collection from
Dominica. The plant grows in Porto Rico and the Virgins and is reported
from St. Martin, Saba, St. Eustatius, St. Kitts, and St. Lucia. It replaces
T. foetidissima L. in the West Indies south of Hispaniola. The various
reports of the latter species in the Lesser Antilles probably all apply to
T. filiflora.

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 133

6. Tournefortia bicolor Sw. Prodr. 40 (1788) and Fl. Ind. Occ. 1: 344
(1797). — Type from Jamaica.
Tournefortia laevigata Lam. Tab. Encyc. 1: 416 (1791); Poir Encyc. 5:
356 (1804).— Type from Guadeloupe, Badier,

Widely distributed in the American tropics.

GUADELOUPE: Stehlé 710, 2831, and 2835 (G), Duss 2579 (G). Marie
GALANTE: Stehlé 2694 and 2834 (G). Dominica: Cooper 195 (G); Eggers
Ooo (Gi, Madge 1522120, 2547, 3100, 31/0, 3002, shot and 3790 ™(G):
MartTINIQUE: Hahn 719 (G); Duss 4503 (G). St. VINCENT: Sauth 434

. Grenapna: Alexander, Herb. Trin. 434 (T). Trinitpap: Broadway
9104 and 9844 (G); Britton 2206 (G); Britton & Hazen 409 (G

The plant occurs on most of the islands in our area. Besides those from
which specimens are cited, it has been reported from Saba, St. Kitts, An-
tigua, Montserrat, Barbados, and Tobago.

7. Tournefortia hirsutissima L. Sp. Pl. 140 (1753).— Type from Haiti.

Widely distributed in tropical America.

GUADELOUPE: Steltlé 2840 (G). Martinique: Duss 1207 (G). Tosaco:
Eggers 5609 (G); Williams 11113 (G). Trinipap: Broadway (G); John-
ston 79 (G).

This species appears to have a limited distribution in our area. I have
seen specimens from only a very few islands and have found reports from
only two additional ones, St. Martin and St. Vincent.

The original description of 7. hirsutissima L. was based upon a drawing
made by Plumier, which Linnaeus studied in Holland previous to the pub-
lication of the Species Plantarum. The drawing was no doubt that sub-
sequently published by Burman, Pl. Am. Plumier 226, ¢. 229 (1760).
Notes associated with Plumier’s original drawings at parc indicate that
the plant concerned was collected near Léogane, Haiti.

8. Tournefortia cuspidata HBK. Nov. Gen. et Sp. 3: 83 (1818); John-
ston, Jour. Arnold Arb. 16: 54 (1936).— Type from Ecuador?
eps obscura DC. Prodr. 9: 517 (1845).— Type from British

Tee setifera Urban & Ekman, Arkiv Bot. 22A: no, 17: 94 (1930).
— Type from Haiti, Ekman H102 04.
Tropical South America, southern Central America, Trinidad, and Haiti.
TRINIDAD: Cedros at St. Anna, low cliff, seashore, Broadway, Herb. Trin.
9407 (T); Irois forest, Williams, Herb. Trin. 12046 (T); Cap de Cille,
shrub in roadside bush, Baker, Herb. Trin. 14205 (T).
Se tee OTRO PLUM
KEY TO THE SPECIES
Flowers scattered along the elongate leafy stems; fruiting pedicels elongate.
-10 mm. long; fruit conic above the middle........ di . lagoense.

Flowers aggregated into crowded unilateral spikes; fruiting pedicels stout,
—1 mm. long; fruit rounded or depressed at summit.

134 JOURNAL OF THE ARNOLD ARBORETUM [von Sex

Plant completely glabrous, very succulent, usually somewhat glaucous. .
i ee Nay he dea Nea eseice Gadling's Gian grade eas Gate Ge a . curassavicum.
Plant hairy; herbage not decidedly succulent, never with a pallid waxy
bloom.

Petiole short, 1-4 mm. long; spike bearing scattered small bracts:
fruit dividing into 4 Pe single-seeded nutlets, strigose.
Leaf-blade 2-5 mm. long; plant a low much-branched perennial,
usually Scie eas and silvery-silky; style very
short and obscure................ ioe ie

Leat-blade 10-30 mm. oo plant laxly er ched, erec
Plant sears shrubby, 3-15 dm. tall; leaf-blade ee
ontrac into a “oaiiole 1-2 mm. long, style elon-
ees ev eee Cin ake es See A. ternatum.
Plant annual, 1-3 dm. tall; leaf-blade very gradually con-
tracted into a petiole 2-4 mm. long fe very short
an SOUT aaah desc cunapees Ai. filiforme.
Petiole well developed, evident, 5 mm. long or longer; spike with-

out bracts among the flow wers; plants usually an
Corolla usually ae or violet, ie tube 24.5 mm. a and
m

angulate, cleft vertically, the two lobes divergent and
each breaking up into a pair of single-seeded nut tle Tie sa
Ura GAR Ase she wha wt ee oils cake gpa wed aa gh taa ain oe k AI. indicum.
Corolla white, tube short, scarcely if at all surpassing e small
calyx; fruit not cleft.
Fruit breaking up into 4 single-seeded nutlets, ag ee
pikes commonly paired and usually 2-5 en 1. long;
eee usually densely strigose and cinereous......
VEN hed erie Soloed aes 7. H. procumbens.
Fruit pacts in half, the halves each 2-seeded, epidermis
roughened by minute bladder-like swellings that col-
lapse oF appear scale-like in age; spikes commonly
single and usually becoming 10 cm. long; plant spar-
ingly appressed hairy, green..8. H. a anglospermumni.
ie Heliotropium lagoense Soe Girke in Engler & Prantl, Nat.
Pflanzenf. iv. Abt. 3a: 97 (1893).
Schleidenia lagoensis Warming, Kjoeb, Vidensk. Medde 1, 1867: 15 (1868).
— Type from Brazil.
Heliotropium trinitense Urban, Symb. Ant. 7: 350 (1912).— Type from
Piarco Savanna, Trinidad, Luni 6030.

Known from Brazil, eastern Bolivia, Surinam, Venezuela, Panama,
Costa Rica, Guatemala, and Trinidad.

Trinipap: Piarco Savanna, Lunt 6030 (Berlin, tyre; fragment, G);
Piarco Savanna, 1895, Lunt, Herb. Trin. 6030 (i Piarco, airfield, plant
decumbent, fl. white with yellow tube and eye, March 1939, Cheesman, Herb.
Trin. 13375 ele

A broadly dictabausl species known only from a relatively few widely
scattered localities. Its wiry, slender, ascendingly branched stems are
Jaxly decumbent or prostrate, usually spring from a clearly annual root,

1949] JOHNSTON, STUDIES IN THE BORAGINACEAE, XVIII 135

and usually become 1—2 dm. long. It is a plant of damp soil and has been
collected most frequently where savanna conditions prevail.

2. Heliotropium curassavicum L. Sp. Pl. 130 (1753).—Type from
Curacao.

Widely distributed in the warmer parts of America, growing along
coasts and in saline soils inland.

St. Martin: Walsh (NY). Sr. BartHéLtemy: Forsstrém (NY).
GUADELOUPE: Duchassaing (G) ; Stehlé 707 and 2833 (G); Duss 2755 (G,
NY). Martrinigue: Duss 1416 (NY); Egler 394 (NY). Sr. VincENT:
Smith 1213 (G). Union: Smith D34 (NY). Grenapba: Broadway (NY).
Bargapos: Dash 353 (NY).

This halophyte is to be expected on all the Lesser Antilles. In addition
to the stations given above, the plant is reported from St. Eustatius, An-
tigua, Marie Galante, Bequia, and Mustique. Spel there are no
reports of its occurrence on either Trinidad or To

3. Heliotropium microphyllum Sw. ex Wikstrém on Guadeloupe. F1. |
K. Vet. Akad. Hand]. 18271: 58 (1828).— Type from Guadeloupe,
Forsstrom.

A variable species ranging is aes the Virgins, Porto Rico, His-
paniola, and Cuba into the Baha

ANGUILLA: Boldingh 3517 (N St. Martin: Boldingh 3034 (NY).

UADELOUPE: coralline rocks along coast, alt. 30 m., Point des Chateau,
Stehlé 2696 (G).

The species is here interpreted in the broadest sense and includes the
habitally similar plants found in the Greater Antilles and the Bahamas.
The plants are moderately variable but hardly: to the degree that would
justify the many segregate species that have been proposed. From Haiti
the plant has been described as H. plumerii Urban, H. elegans Urban, and
H. glomerifolium Urban, and from Porto Rico as H. crispiflorum Urban.
Even more names have been applied to plants of Cuba and the Bahamas.

4. Heliotropium ternatum Vahl, Symb. Bot. 3: 21 (1794).— Type from
the West Indies.
Pioctonon ternatum (Vahl) Raf. Sylva Tellur. 88 (1838).
seis pe Sp. Pl. 141 (1753); Syst. ed. 10, 917 (1759) ; Sp.
I a fake 2 (1762).— Based on a plant of Martinique illustrated by
Plur
nore humile (L.) R. Br. ex R. & S. Syst. 4: 37 (1819), not Lam.
1791

Heliotropium ae Lehm. Neue Schr. Naturf. Ges. Halle 32: 10 (1817),
ad. Caes. Leop. Nat. Cur. 9: 135 (1818), and ae 1:
62 (1818) a & S. Syst. 4: 38 (1819).— Type from Venezuela.
Heliotropium hispidum HBK. Nov. Gen. et Sp. 3: 87 (1817) and 451
(1820).— Type from Venezuela.
Heliotropium fruticosum, var. hispidum (HBK.) DC. Prodr. 9: 543
1845

Scnedenia hispida (HBK.) Fresen, in Mart. Fl. Bras. 8!: 37 (1857).

136 JOURNAL OF THE ARNOLD ARBORETUM [VOL., XXX

Heliotropium ele R. & S, Syst. 4: 37 and 733 (1819), in pt. — Type

les dian; a mixture of H. ternatum and H. fruticosum.

Pioctonon ie Raf. Sylva Tellur. 88 (1838), in pt.— Type West

Indian; a mixture of H. ternatum and H, fruticosum.
Heliotropium fruticosum, var. nee DC, Prodr. 9: 542 (1845).—
Type from Guadeloupe, Bertero.

Heliotropinin fruticosum, var. “ee DC. Prodr. 9: 543 (1845).

ype from Cuba.

Native to the West Indies, Yucatan, and northern Venezuela and
Colombia

ee Box 1291 (G); Rose, Fitch & Russell 3360 (NY). Mont-
SERRAT: Shafer 412 (NY). Destrave: Stehlé 258 (NY), 2829 (G); Duss
2756 (NY). Dominica: Imray (G); Bailey 750 (NY); Lloyd. mead Bry 4
ee 3790 (G). MartTINIQuE: Steber, Fl, Mixta 430 (G, NY); Duss

7 (G, NY), Egler 39-6 and 39-53 (NY); Hahn 416 (G): ey 2219
Zens 3469 (G), 6168 (G). St. Vincent: Smith 977 (G, NY).

A loosely branched shrubby plant frequently forming a bush 3-20 dm.
tall. J have seen one collection from Cuba (Wright 3135) and numerous
collections from the Bahamas, Jamaica, Porto Rico, and the Virgins. The
species has been reported from Haiti, Guadeloupe, St. Lucia, and Bequia.
It is apparently absent from Trinidad and Tobago, but is common along
northern Venezuela and on the adjacent Margaritas and Dutch islands.

Distinctive of the species is its bushy growth and strong root and its
development of opposite or ternate leaves at one or more nodes on vigorous
shoots. It varies considerably in size and shape of leaves, in pubescence,
and in quantity of opposite or ternate leaves developed. In the West
Indies it has been confused with the very different H. fruticosum L., cf.
Johnston, Contr. Gray Herb. 81: 66 (1928). That species occurs in Porto
Rico, Haiti, and Jamaica, and on the continent, and is readily distinguished
from our plant by having a lower habit, annual root, and well-developed
leaf-like bracts in the inflorescence. In an earlier paper I failed to dis-
tinguish H. ternatum from the related South American H. Ottoni Lehm.
That latter species has synonyms in H. strictum HBK., Tournefortia mono-
stachya Willd., H. Ottonianum R. & S., H. Ottonis DC., H. strictissima
Moric., and Schleidenia subracemosa Warm. It differs from our species in
its short-lived chiefly annual root, few usually sparingly branched and
frequently strict stems, uniformly alternate leaves, and more southerly
range. The plant has been found from northern Venezuela to eastern
Brazil. Though formerly confused with H. salicoides Cham., both H.
Ottoni and H. ternatum are readily distinguished by their white rather
than decidedly yellow corollas, cf. Johnston, Jour. Arnold Arb. 16: 62
(1935). In Central America H. ternatum is confined to the Yucatan
peninsula, being replaced in Mexico and other parts of northern Central
America by the closely related but lower-growing and less strongly peren-
nial H. oaxacana DC. and H. mexicanum Greenm. ‘These two latter
species have only alternate leaves. The more northerly ranging H. mexi
cana occurs from Oaxaca to Sonora and Nuevo Leon and has retrorse

1949] JOHNSTON, STUDIES IN THE BOR AGINACEAE, XVIII 137

hairs on its leafy twigs, as well as leaves that are usually broadest at or

above the middle.

For the present species Urban, Symb. Ant. 8: 590 (1921), has used the
name “H. humile (L.) R. Br. Prodr. I (1810) p. 497 (non Lam.).” The
name is based upon Tournefortia humilis L., which has its ultimate basis
in a plant illustrated by Plumier. It is known that Linnaeus, while in
Holland, studied copies of Plumier’s plates during the preparation of the
Species Plantarum. Among these was the representation of the present
species later published by Burnam, Pl. Amer. Plumier. LIAN ey! foie Seed
(1760). Notes accompanying the original drawing at Paris state that the
plant concerned was collected on Martinique (south end) in the area
known as “La pointe des Salines au Cul de Sac Marin.” It is a very
coarse atypical form of H. ternatum. Urban indicates that Robert Brown
made the combination “H. humile.” This is incorrect. The actual trans-
fer was first published by Roemer & Schultes, Syst. 4: 37 (1819), and even
then only as a synonym. This binomial, in any case, is not a valid name
for our plant, since it is a late homonym of H. humile Lam. Tab. Encyc.
1: 393 (1791). The application of Lamarck’s binomial has been uncer-
tain. I have referred it to the synonymy of H. angiospermum Murr.

5. Heliotropium filiforme Lehm. Gotting. Gel. Anzeigen 1817: 1515
(1817), and Asperif. 1: 37 (1818); Johnston, Contr. Gray Herb. 81:
61 (1928) and Jour. Arnold Arb. 16: 62 (1935).— Type from
Venezuela.

Paraguay and eastern Bolivia northward through Brazil to the Guianas
and Venezuela; northern Central America and adjacent Mexico; Trinidad.

Trinipap: Icacos, Broadway 7401 (T),

This is a plant of wet ground. It has been collected most commonly
near streams.

6. Heliotropium indicum L. Sp. Pl. 130 (1753).

A weedy plant widely distributed in the warmer parts of the world.

ANTIGUA: Box 929 (G). Montserrat: Shafer 366 (NY). Dominica:
Lloyd 247 (NY); Hodge 3872 (G). Martinique: Duss 1419 (Gren Y):.
St. Lucia: Walsh (NY). Sr. Vincent: Smith 1213 (G). Beguta:
Joseph 231 (NY). Grenava: Broadway (G, NY). Barsapos: Freeman
12H (NY). Trinipap: Herb. Trin.

Reported from most of the islands in our area and very probably pres-
ent on all of them.

7. Heliotropium procumbens Miller, Dict. ed. 8, no. 10 (1768); John-
ston, Contr. Gray Herb. 81: 52 (1928).— Type from Cartagena,
Colombia.

Heliotropium inundatum Sw. Prodr. 40 (1788) and Fl. Ind. Occ. 1: 343

(1797).— Type from Jamaica.

A species widely distributed in the warmer parts of America but of

restricted occurrence in the West Indies.

138 JOURNAL OF THE ARNOLD ARBORETUM (VOL. Xxx

TIGUA: Box 852 and 1357 (G). Trinipap: Broadway (G, NY);
Britton 2506 (G, NY), Britton & Hazen 720 (NY).

The plant has been found in Jamaica, Cuba, Hispaniola, and Porto Rico,
but is very rare or absent in the Lesser Antilles. It has been reported from
Guadeloupe, but recent collectors have not found it there, and Duss, Fl.
452-3 (1897), does not list it for the French islands. The plant is usually
annual and usually favors places subject to periodic inundations of fresh
water.

8. Heliotropium angiospermum Murray, Prodr. Stirp. Gottingen 217

(1770); Johnston, Contr. Gray Herb. 81: 10 (1928).— Type from
a European botanic garden.

Schobera Oh ee Murray ex Ne Intr. 158 (1777); Britton &
Wils., Bot. Porto Rico 6: 134 (19

Heliotr ore eee L. Mant. 2 ae (1771).

Heliotropium humile Lam, Tab. oe cyc, 1: 393 (1791); Johnston, Jour.
Arnold Arb. 16: 186 (1935).

Widely distributed in the warmer parts of America.

St. Martin: Boldingh 2344 (NY). St. BARTHELEMY: Forsstrom (NY).
St. Eustatius: Walsh (NY). St. Kirts: Kidder (G); Britton & Cowell
279 (NY). Awnticua: Duss 41 (NY); Box 1050 (G). GUADELOUPE:
Stehlé 762 (NY), 2688 (G), 2830 (G); pees (G, NY). Martt-
NIQUE: Stehlé 3641 (G); Duss 1418 (G, NY); Hahn (G); Egler 39-45
(NY). Sr. Vincent: Smith 1582 ce ee Broadway (NY).
Tosaco: Williams (T). Trinipap: Broadway 2717 and 5140 (G

A plant of sunny open places with well-drained soil. Frequently a weed
in waste ground. The species has been reported from Anguilla, Saba,
Dominica, and Barbados.

The name H. Aumile Lam. has long been of uncertain application and a
source of confusion, cf. R. & S. Syst. 4: 38, 733 and 734 (1819) and
Johnston, Contr. Gray Herb. 81: 67 (1928). It has been applied to such
diverse species as H. ternatum Vahl, H. fruticosum L. and H. angio-
spermum. After a study of specimens available to Lamarck at Paris |
am now of the opinion that H. Aumile Lam. belongs in the synonymy of
H. angiospermum.

ARNOLD ARBORETUM,
Harvarp UNIVERSITY.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 139

PLANTAE PAPUANAE ARCHBOLDIANAE, XIX *
Lity M. Prerry

THIS PAPER covers the Cunoniaceae of Papuasia in the Arnold Arbore-
tum herbarium. The following genera are represented: Spiraeanthemum,
Gilbeea, Opocunonia, Spiraeopsis, C eratopetalum, Schizomeria, Aistopeta-
lum, Weinmannia, and Pullea. I have found nothing to match either
Kaernbachia or Stollaea. The latter, from the description, appears ex-
ceedingly close to the genus Opocunonia, which Schlechter accepted as hav-
ing fleshy fruit’ In the abundant material of this genus at hand, it can
now be shown that the fruit is capsular and very similar to that described
for Stollaea. The fruit of Aistopetalum, however, is a drupe; there are,
accordingly, two genera with drupaceous fruits represented in New Guinea.
Some reductions have been made, others may need to be made. Wherever
species have been added to a genus (except in Weinmannia), a brief key is
given to help future workers evaluate the species. In Weinmannia, be-
cause of the interrelationships of the species, it would be necessary to pre-
pare a key including the Malaysian and Polynesian species as well; we
do not have sufficient material for such a key. The genus Gillbeea is rep-
resented by two collections (Clemens 3430, 3769) which practically match
the figure of Schlechter’s G. papuana. The other genera are elaborated
below.

SPIRAEANTHEMUM A. Gray
Key TO THE SPECIES
Reavesndenticulates chartaceous sis ct uct S01 u iene een S. Kajewskit,
Leaves entire or inconspicuously crenulate-serrulate, coriaceous (except in
S. parvifolium).

Pubescence of the branchlets, the under surface of the leaves, and the
inflorescence dense and shaggy; caiyx shortly villous or pilose...
ee Be rT siete nae aaa ahs Sie a bbs Se near eee a Lean

Pubescence of the branchlets and inflorescence shorter, usually = ap-
pressed; calyx glabrous

Leaves four in a verticil, cuspidate-acuminate; scales of the disk

ovat S. integrifolium,

Leaves opposite or ternate, not cuspidate-acuminate; scales of the
isk oblong or quadrate, not ovate.

Leaves small, 2-2.7 x 1-1.7 cm., chartaceous; carpels villous;
SEV es SMDClaVate re Pos ac:9 ace wg ded nee oe . parvifolium.

Leaves larger, 5.5-13 x 2.5-6 cm., coriaceous; carpels puberu-
ous or almost glabrous; styles subulate.

Axis of the inflorescence hirtellous or subtomentose....
Sena an Teena ret 4 Wt: iy Aenea eam sian ee S. reticulatum.

PP OO 8 6 88 O08 Rie eee eee 8 won 8 4 6 ae eee 8 8

*Botanical Results of the Richard Archbold Expeditions. See Jour. Arnold Arb.
30: 39-63, 1949,

140 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Axis of the inflorescence puberulous.
Leaves elliptic or lance-elliptic (broadest at the mid-

e).
sae leaves densely reticulate on both surfaces,
rgins remotely crenulate-serrulate; fol-
ae practically glabrous (here and there
are very scattered minute hairs).........
Boies obese, eae alte) ae Pew) pola: sey 8) ere ce . novogiuimecense.
Mature ae obscurely reticulate on the upper
only, margins entire; follicles
sparsely but evenly puberulous Yee ee
pee pag te gd cetaeead as . idenburgeise.
Leaves lanceolate (broadest below the middle)......
eee Pome Nae eee ee S. lanceolatinn.
Spiraeanthemum Kajewskii Perry, spec. nov.
Arbor usque 20 m. alta; ramulis atro-fuscis, novellis minute pubescenti-
bus, compressis, soto male incrassatis, cito glabratis; stipulis oblongis,
m ongis, . latis, obtusis, extus appresse pubescentibus; foliis
chartaceis, pice - 12 cm. longis, 2.5-5 cm. latis, utrinque angustatis,
apice neice: vel breviter et obtuse acuminatis, basi cuneatis, margine
denticulatis, supra olivaceis subtus pallidioribus, utrinque glabris (subtus
costa venisque hinc inde pilis obsitis), venis primariis utrinsecus +13
oblique patentibus prope marginem furcatis utrinque perspicuis, venulis
+ dense reticulatis utrinque manifestis; petiolo 0.8-2 cm. longo; inflo-
rescentiis +7.5 cm. longis, axi et ramulis puberulis, multifloris; pedicellis
1 mm. longis; calyce 1.5 mm. longo, lobis 4 vel 5, oblongis, 1 mm. longis,
acutiusculis; staminibus longioribus 1 mm. alternis 0.7 mm. longis, an-
theris minutis: disci squamis 4 vel 5, bilobatis apice truncatis vel paulo
retusis, 0.4 mm. longis, 0.5 mm. latis: carpellis 4 vel 5, ovariis 0.7 mm.
longis, pubescentibus, 2-ovulatis, stylis 0.4-0.5 mm. ene vix 0.2 mm.
crassis, stigmate capitata; folliculis 3.5 mm. longis, 0.7 mm. diam., appresse
a tena
SOLOMON ps: Bougainville: Kupei Gold Field, Kajewskr
1700 ony re i 1930, alt. 1000 m., rain-forest (tall tree up to 20 m. high) ;
Koniguru, Buin, Kajewski 2082, Aug. 1930, alt. 950 m., rain-forest (tree up
to 20 m. high; Pawiek white. Wood used by natives for building houses ).
Guadalcanal: Uulolo, Tutuve Mt. Kajewski 2533, April 1931, alt.
200 m., rain eal A 8 or 9 m. 1 gh; flowers minute, green; fruits
small, mostly in fou
Kajewski 2082 is ne from new growth. It is much more pubescent
than the type, the terminal bud is villous as well as two very young leaves
subtending it. Also, the leaves are somewhat larger (10-15 X 5.5-8.5
cm.) and the midrrb and veins of both surfaces are pubescent with fine
hairs, the lower surface more densely so. Between the veins the lamina
is glabrate. Further, the axis and branches of the inflorescence are much
more hairy than those of the type. These collections appear nevertheless
to be conspecific.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 141

Spiraeanthemum Kajewskii is more like some Polynesian species than
those of New Guinea. It should be compared with S. Macgillivrayi Seem.
from Aneityum, which is only very briefly characterized and of which we
have no specimen. It suggests S. samoense A. Gray, but this last species
has leaves rounded at the base, stamens as long as the calyx or longer, and
the parts of the disk as many as the stamens.

Spiraeanthemum Pulleanum Schlechter in Bot. Jahrb. 52: 140. 1914,
in Nova Guin. 12: 491, ¢. 197. 1917; Kaneh. & Hatus. in Bot. Mag.
(Tokyo) 56: 111. 1942.

NETHERLANDS New Guinea: 15 km. southwest of Bernhard Camp, Iden-
burg River, Brass & Versteegh 11923, Brass 12017, 12136, Jan. 1939, alt.
1780 and 1800 m., common in mossy forest (subsidiary tree 15-18 m. high;
flowers light yellow; leaves stiff, convex). |

I am unable to distinguish these collections either from this species (as
defined in the original description) or from Kanehira’s collection taken in
the type locality, except that the leaves are not, or are only very slightly,
bullate. This characteristic appears to be variable within the species.
Spiraeanthemum reticulatum Schlechter in Bot. Jahrb. 52: 140. 1914.

NETHERLANDS NEw GuiINneEA: 18 km. southwest of Bernhard Camp, Iden-
burg River, Brass 12693, Feb. 1939, alt. 2150 m., mossy forest (tree attain-
ing a height of 25 m. and stem diameter of 30 cm.).

This collection agrees fairly well with the original description except
that it is difficult to decide what Schlechter considered a short petiole. In
this specimen the largest leaf is 6.5 cm. long (including a petiole 1 cm.)
and 3.2 cm. wide. The smaller leaves have petioles about 5 mm. long not
including the shortly decurrent (1-3 mm.) base of the lamina. The
lamina tends to be slightly convex between the impressed primary nerves,
and the margins are somewhat revolute. The axis and branchlets of the
inflorescence are more hirtellous or subtomentose than villous, the hairs
are fairly short and crinkly but not matted. The carpels are mostly pu-
berulous in this profusely fruiting specimen. The seeds are about 2.5
mm. long.

Spiraeanthemum idenburgense Perry, spec. nov.

Arbor 16 m. alta, 35 cm. diam.; ramulis atro-fuscis, novellis puberulis,
nodis incrassatis; foliis coriaceis, oppositis et ternatis, ellipticis, (2.3—)
5.5-11 cm. longis, (1.5—) 2.5-6 cm. latis, utrinque angustatis, apice ob-
tusis, basi cuneatis, margine integris, utrinque glabris, supra olivaceis,
subtus pallidioribus, venis primariis utrinsecus 6-8 supra impressis subtus
prominulis, oblique patentibus prope marginem arcuatis, venulis dense
reticulatis supra vix distinctis, subtus perspicuis; petiolo (0.3—) 1-1.5
cm. longo; inflorescentiis 5.5-13 cm. longis, pedunculo 0.5—6 cm. longo,
axi et ramulis puberulis; calyce 2 mm. longo, glabro, lobis 4 vel 5, ovatis
subacutis, 1.3 mm. longis, 1 mm. latis; filamentis 1.2 mm. longis, antheris
ellipsoideis, 0.4 mm. longis, leviter retusis; disci squamis oblongis vel
quadratis 0.6 mm. longis; carpellis saepissime 3, ovariis pubescentibus,

142 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XxX

0.7 mm. longis, 2-ovulatis; stylis 1.1 mm. longis; folliculis 5 mm. longis
++ puberulis; seminibus 2.3 mm. longis.

NETHERLANDS New Guinea: 15 km. southwest of Bernhard Camp, Iden-
burg River, Brass & Versteegh 11938 (type), Jan. 1939, alt. 1680 m., fre-
quent on forested slopes (tree 16 m. high, 35 cm. diameter; flowers light
yellow; fruit yellow-green).

This species shows much variability in the size of the leaves and the
length of the infructescences. The mature leaves are more firmly coria-
ceous than in any other species at hand. When the leaves are ternate,
often one leaflet is much smaller than the others.

Spiraeanthemum novoguineense Perry, spec. nov.

Arbor + 15 m. alta, 20 cm. diam.; ramulis atro-cinereis, novellis puberu-
lis, nodis incrassatis; foliis coriaceis oppositis, interdum ternatis, ellipticis
utrinque angustatis, basi cuneatis, apice anguste obtusis vel subrotundatis,
interdum breviter et obtuse acuminatis, margine remote crenulato-serrulatis
vel integris, utrinque glabris dense reticulatis, venis primariis utrinsecus
5-7, patenti-ascendentibus prope marginem arcuatis, supra inconspicuis,
subtus prominulis, lamina 2.5-8.5 cm. longa, 1.3-4.2 cm. lata, in petiolo
2-5 mm. decurrente; petiolo 3-10 mm. ienea: supra - applanato, inter-
dum puberulo; cae ereseentee usque 9 cm. longis, axi et ramulis + pu-
berulis; pedicellis vix 1.5 mm. longis; calyce vix 2 mm. longo, glabro, lobis
4, 1.3 mm. longis, 0.9-1 mm. latis, acutiusculis vel obtusiusculis; filamentis
1.7—2 mm. longis, glabris; disci squamis 0.5—0.6 mm. longis, 0. 2 mm. latis,
oblongis vel paulo cuneatis vel interdum subquadratis; carpellis 3 vel 4,
immaturis 1.8 mm. longis, sparsim pilosulis 2- or 3-ovulatis, maturis 4 mm.
longis, hinc inde pilis minutis obsitis; stylis subulatis 1.2 mm. longis;
seminibus 2 mm. longis.

NETHERLANDS New GuiIneA: 15 km. southwest of Bernhard Camp, Iden-
burg River, Brass 11860 oe Jan. 1939, alt. 1800 m., mossy forest (abun-
dant subsidiary tree + 15 m. high, 20 cm. diameter ; fruits green).

NortHeast New nee vicinity of Samanzing, Clemens 9476, 9584,
Jan., Feb. 1939, alt. = 2400 m

In these collections the ie are occasionally ternate. Ata glance the
reticulate upper surface of the leaf may be seen. The species differs from
S. parvifolium Schltr. in having somewhat larger and flat coriaceous leaves.
It may be distinguished from S. reticulatum Schltr. by the larger leaves,
the shorter infructescence (not more than half as long again as the leaves),
and the puberulous axis and branchlets. In S. reticulatum Schltr. the in-
florescence is two or three times as long as the leaves, and the peduncle is
villous.

Spiraeanthemum lanceolatum Perry, spec. nov.

Arbor glabra; ramulis fuscis, nodis eucgae alabastris parvis, breviter
tomentosis; foliis coriaceis, opp itis vel interdum ternatis, lanceolatis,
6.5-13 cm. longis, 2.4-4.8 cm. latis, basi ae apice late et obtuse
acuminatis, acumine 6-12 mm. longo, margine integris, na glabris,

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 143

venis primariis utrinsecus + 7, patenti-ascendentibus prope marginem
arcuatis, supra impressis subtus prominulis, venulis dense reticulatis supra
inconspicuis subtus perspicuis; petiolo (incl. lamina 3—7 mm. decurrente)
0.5-1.3 cm. longo, canaliculato; inflorescentiis 3-5 cm. longis, axi et
ramulis puberulis; pedicellis 1-1.8 mm. longis; calyce 1.7 mm. longo, lobis
4-6, 1.2 mm. longis, 0.6-0.7 mm. latis, oblongis vel oblongo-ellipticis,
acutiusculis; filamentis 1.5—2 mm. longis, antheris minutis, late ellipsoideis
minute apiculatis; disci squamis 4—6, bilobis (interdum bipartitis), 0.2—
0.3 mm. longis truncatis vel paulo retusis; carpellis 2-4, ovariis 0.5 mm.
longis, sparsim pubescentibus, 4-ovulatis; stylis 1.5-2 mm. longis.

NETHERLANDS NEw GUIN 2 km. southwest of Bernhard Camp, Iden-
burg River, Brass 13615 Cy Mar. 1939, alt. 700 m., rain-forest (substage
tree; flowers white).

This species is readily distinguished from S. novoguineense by the lanceo-
late entire leaves with a smooth upper surface. The latter character is
also found in S. idenburgense, but the leaves of the latter are elliptic and
somewhat more firmly coriaceous. Spiraeanthemum idenburgense also has
a little larger flowers with shorter filaments and styles but larger disk
scales than S. lanceolatum.

OPOCUNONIA Schlechter

In our material there are at least four species of this genus, and five
different collections in fruit. The fruits are bivalvate capsules instead of
drupes having a fleshy exocarp as indicated in the original description.
The seeds are winged at both ends + 3 mm. long (including the wings) and
about 0.5 mm. broad. With only the original description of the genus
Stollaea Schltr. and the figure of the type-species S. papuana Schltr. for
comparison, I am unable to find any generic characters to separate that
genus from Opocunonia Schltr. However, since we have no material which
matches S. papuana Schltr., I merely mention this in passing, and leave
the actual reduction to the monographer, who should have access to isotype
material, if possible.

Opocunonia kaniensis Schlechter in Bot. Jahrb. 52: 160, fig. 8, A-G.

NETHERLANDS New GuINEA: Bele River, 18 km. northeast of Lake Hab-
bema, Brass & Versteegh 11120, Brass 11478, Nov. 1938, alt. = 2200 and
2300 m., frequent in the forests (large canopy tree; bark reddish brown,
warted; wood soft, flowers white); 6 km. southwest of Bernhard Camp,
Idenburg River, Brass & Versteegh 12563, Feb. 1939, alt. 1230 m., common
in primary forest (tree 32 m. high, 59 cm. gee bark gray; sapwood orange,
heartwood dark red; flowers white; fruits gre

NorTHEAST NEW GUINEA: oe eee Ciene D022;) Jap tose. alt
1770 m.; Samanzing, Clemens 8967, 9297, Oct., Nov. 1938, alt. 1600-1800 m.,
mountain bush (tree 15 in. to 3 ft. dam flower buds pale green; flowers
white); Matap, Clemens 11102, Feb. 1940, alt. 1500-1800 m.

144 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

In these specimens the base of the leaflets is often rounded, then very
shortly cuneate; the ovary is puberulous rather than glabrous; the capsule
is 6-8 mm. long and about 3 mm. diameter. The stipules vary in size
(0.7-2.5 cm. long, 1-3 cm. broad) and are somewhat rectangular in out-
line, with truncate or broadly and shallowly retuse apex. The rounded
sides contract into a broad short base at the attachment to the stem. From
this base the two main veins extend to the outer points of the apex like a
very broad V.

Opocunonia Nymanii (K. Sch.) Schlechter in Bot. Jahrb. 52: 159. 1914.

Ackama Nymanu K. Sch. in K. Schum. & Lauterb. Nachtr. Fl. Deutsch.

Schutzgeb. Stidsee, 272. 1905

Nortrueast New GuiIneA: Morobe District, Sattelberg, Clemens 6654,
June 1937.

This collection, although fragmentary and with only young flower-buds,
matches very closely the original description. At hand is another specimen
collected below Boana in Aug. 1938, at approximately 750 m. alt., Clemens
8704 (tall tree 2 ft. diam.); a few leaflets of this specimen are about the
same size as those of Clemens 6654, but most of the leaflets are larger, the
largest being 12.5 cm. long, 4 cm. broad and chartaceous; those of Clemens
6654 are slightly firmer. The branches of the inflorescence are villous,
even in the infructescence, those of the other specimen (of which we have
only a fragment of the inflorescence) are only minutely puberulous. Clem-
ens 8704 has an inflorescence (young buds only) about 10 cm. long and
an infructescence about the same length. The capsule (without the per-
sistent style) is 5 mm. long and minutely appressed-pubescent.

Opocunonia papuana Kanehira & Hatusima in Bot. Mag. (Tokyo) 56:
108, fig. 4. 1942.

eae New Guinea: Chaban, 30 km. inward from Nabire, Kane-
hira & Hatusima 11756, Feb. 1940, alt. 100 m., in dense rain-forest (tree
30 m. high) ; 15 km. southwest of Bernhard Camp, Idenburg River, Brass &
Versteegh 11950, Jan. 1939, alt. 1900 m., occasional in primary forest (tree

1 m. high, 43 cm. diam.; Date black ; ahon! red-yellow, heartwood dark
red; flowers white); 18 iin, southwest of Bernhard Camp, Idenburg River,
Brass & Versteegh 12501, Feb. 1939, alt. about 2000 m., rare in primary
forest (tree 43 m. high, 67 cm. diam.; bark black, rough; wood red-yellow ;
flowers white; ripe fruit brown) ; 6 km. southwest of Bernhard Camp, Iden-
burg River, Brass & Versteegh 12540, 13105, Feb. 1939, alt. 1200 m., frequent
or occasional in primary forest (tree 20-28 m. high, 43-53 cm. diam. bark
brown, scaly in one, in the other gray, rough; owas white) ; 4 km. Soult ae
west of Bernhard Camp, Idenburg River, Brass & Versteegh 13117, Brass
13309, 13630, March 1939, alt. 850 and 900 m., plentiful in rain forest canopy
and common subsidiary tree in Agathis forest (tree 35 or 25 m. high; flowers
white).

Nortueast NEw GvuINEA: Sambanga, Clemens 6957, Sept. 1937, alt.
1500-1800 m. (tall tree 18 in. len.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 145

There is some variation in the size of the flowers of the collections cited
above. The length of the calyx-lobes varies from 1.4—2 mm. long, and the
pedicels from 1-2 mm. long. The leaves also vary considerably in size,
but all seem to be of similar texture and venation; in some specimens the
lower surface of the leaf is sparsely and minutely pubescent and small
domatia occur in the axils between the primary veins and the midrib. The
capsule of this species is 4-5 mm. long (not including the style), with
scattered minute hairs on the outside. The seeds are + 3 mm. long.

Possibly the following collection belongs here: 2 km. southwest of
Bernhard Camp, Idenburg River, Brass & Versteegh 13505, Apr. 1939,
alt. 750 m. The leaves of this specimen are a little coarser than in most
of the collections cited above, and the inflorescence (flower buds very
young) in one specimen measures 20 & 25 cm
Opocunonia trifoliolata Schlechter in Bot. Jahrb. 52: 161. 1914.

British New Guinea: Central Division, Mafulu, Brass 5247, Sept.-Nov.
1933, alt. 1250 m., old secondary forest (tree 8 m. high, with ee smooth
leaves and white flowers).

The above cited specimen is probably only a pubescent form of this
species. On the lower surface of the leaves there are minute scattered
hairs, the base of the lateral leaflets is rounded, while that of the terminal
ones is usually cuneate. The leaves are both 3- and 5-foliolate. The axis
of the inflorescence is densely pubescent, as are the pedicels, the lower
part of the calyx, and the ovary. The filaments are 6 and 8 mm. long.

SPIRAEOPSIS Miquel

In our herbarium the genus Spiraeopsis Miq. is represented by thirteen
specimens, one from the Celebes, two from Manado, and ten from the
Philippines. The genus was originally described as dioecious by Miquel,
later emended to monoecious by Koorders in his Suppl. Fl. Celebes 1: 22.
1918. Since then B. L. Burtt, in Kew Bull. 1936: 462. 1936, in a discus-
sion of S. celebica Miq., has indicated that all herbarium specimens seen
have been entirely male or entirely female. I have dissected several flowers,
all perfect, i.e., possessing both pollen and ovules, and in appearance they
are like figs. 11, 12, 13, of Koorders’ plates 7a, 76, which in the legend
are designated as ‘“‘masc.’”’ Could it be that the fugacious character of the
petals and stamens gives rise to the impression of a pistillate inflorescence?

The genus Betchea Schlechter, hitherto regarded as endemic in New
Guinea, seems to differ only in having two to five carpels instead of two,
rarely three. In this study it is treated as a synonym of Spiraeopsis.

KEY TO THE SPECIES
Flowers with mostly 2 (occasionally 3)-loculed ova
Lower surface of the leaves glabrous (but eee glandul a except
midpib-and: veins: (assem err eee eee ons ese o ces S. glabrescens.
Lower surface of the leaves not glabrous, or if so, leaves 5—9-foliolate.

146 JOURNAL OF THE ARNOLD ARBORETUM [VoL. xxx

Lower surface of the leaves tomentulose, or if glabrous, leaves
DO AGUGates ie iba ceerariaaeue ees S. celebica sensu lato.
Lower surface of the leaves densely tomentose or pilose-tomentose.
Leaves 3-foliolate; capsule 5-6 mm. long, tomentose; seeds

i S

DeNlOSinyasnnskes hie es ereaaenseeaews ens . Brassu.
Leaves 5-foliolate; capsule 3-4 mm. long, shortly villous-
Beau rucueuaseesees ce cote eneces S. pubescens.

tomen
Flowers with 3-5-loculed ovary.
Flowers with mostly 3 (occasionally 4)-loculed ovary.

Lsep ves a OOle eget Oe eee lata ceastaua wheat id Sa pula,
Leaves 5—11-foliolate.
Leatlets Secsi6. cccdeves akneghe caters awsuees S. canariufolia.
Leaflets petiolulate.
Lower surface of leaflets rufous-tomentose........ S. rufa.

Lower seer of leaflets stellate-puberulous to almost gla-
s and glandular

pane small, 2.5—/ 1Sdoemises 64.4 S. Clemensiae.

Leaflets 7 8-12 x 3.5-5¢
s 7-11, gradually eas. .S. papuana.
ae Dy POI Neetu Se csaag ss - myrtantha.

Flowers with 4- or 5loculed ovary.

Flowers with 4-loculed ovary; under surface of leaves glabrous and

gla Be. cece enience hea veueeea . pometiuformis.
Flowers with 5 (occasionally 4)-loculed ovary; under surface o
leaves densely tomentulose................ . aglaiaeformis.

Spiraeopsis celebica (Bl.) Miq. Fl. Ind. Bot. 1(1): 719. 1856; B. L.
Burtt in Kew Bull. 1936: 462. 1936; sensu lat.

Cunonia celebica Bi, an 868. 1826.

SoLtomMon ISLAN Bougainville: Kupei Gold Field, Kajewski
1/13, Apr. 1930, ne * 1200 m., rain forest (tree up to 30 m. high; fruit brown
when ripe splitting into two parts, length without style 5 mm., diam. 2.5 mm.) ;
Buin, Koniguru, Kajewski 2113, Aug. 1930, alt. 900 m., rain forest (tree up
to 25 m. high; flowers with green sepals, very long siomeas, and a faint
sweet odor). Ysabel: Tiratona, Brass 3214, Nov. 1932, alt. 600 m.,
common in mountain rain forests (tall tree with thin gray bark; fruit brown).

There is considerable variation in the pubescence of these three collec-
tions as is also true of the other material of this species at hand. B
until such time as a critical examination of the species can be made, it
seems preferable to place the material here.

Spiraeopsis aglaiaeformis (Kaneh. & Hatus.) Perry, comb. nov.

Betchea aglaiaeformis Kanehira & Hatusima in Bot. Mag. (Tokyo) 56:

106 2. 1942,

NETHERLANDS New Guinea: Angi, Arfak Mts., Kanehira & Hatusima
13907 (1soryre), Apr. 1940, alt. 1900 m., in forest along Iray River to Lake
Gita (tree 15 m, high, 20 cm. diam.) ; 9 km. northeast of Lake Habbema,
Brass & Versteegh 10476, Oct. 1938, alt. = 2700 m., rare in primary forest
(tree 34 m. high, 30 cm. diam.; wood rose; bark gray, fairly smooth;
flowers and fruits yellow); Bele River, 18 km, northeast of Lake Habbema,

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 147

Brass & Versteegh 11138, Nov. 1938, alt. 2350 m., frequent in Baas forest
(tree 31 m. high, 42 cm. diam.; bark gray, Toney wood rose; flowers
brown) ; 15 km. southwest of Bemiard Camp, Idenburg River, Bae dipafpot ip
Jan. 1939, alt. 1700 m., in rain forest ravine (tree 5 m. high, 10 cm. diam. ;
flowers white).

NortHEAST NEw GuINEA: Sambanga, Clemens 6835, Aug. 1937, alt.
1500-1800 m., margin of mountain forest (frequent big tree).

Among he fruits I have found three, four and five locules but mostly
the last mentioned. The leaves are 3-, 5- or 7-foliolate, with the 3-foliolate
opposite the 5-foliolate ones.

Spiraeopsis rufa (Schlechter) Perry, comb. nov.

Betchea rufa Schlechter in Bot. Jahrb. 52: 148, f. 4, A-G. 1914.

NortHeast New Guinea: Yunzaing, Clemens 3872, Aug. 1936, alt. 1350
m.; Quembung, Clemens 3195, June 1936, alt. 750 m., forest hills (tall tree
10-12 in. diam.; flower purplish, stamens purple-red, anthers pale yellow;
fruits brown, 3-fid, purple inside) ; Sattelberg, Clemens 2216, Apr. 1936, alt.
+ 1000 m., in forest.

In flowers and pubescence these three specimens match Schlechter
17086, one of the numbers cited in the original description. The capsules
are 3 mm. long, valves + 1 mm. broad; seeds including the wings 1 mm.
long.

Spiraeopsis fulva (Schlechter) Perry, comb. nov.

Betchea fulva Schlechter in Bot. Jahrb. 52: 148. 1914.

We have no specimen which matches this species, but it should be com-
pared with S. glabrescens Perry.

Spiraeopsis Brassii Perry, spec. nov.

Arbor; ramulis dense subferrugineo-tomentosis deinde glabratis, cras-
Sargent 5 mm. diam.; stipulis parvis, 5-7 mm. longis, 7-8 mm. latis,
subreniformibus vel subrotundis, utrinque tomentosis; foliis trifoliolatis,
margine plerumque crenato-dentato interdum fere subintegris, supra rugo-
sis, novellis tomentosis cito glabratis, subtus dense tomentosis vel breviter
sublanatis, foliolis ellipticis terminali interdum obovato-elliptico, 7-11
cm. longo, 4-8.5 cm. lato, apice subacuto vel obtuso vel interdum breviter
et obtuse acuminato, basi cuneato vel rotundato, petiolulo 0.5-1.8 cm.
longo, lateralibus ellipticis, saepe minoribus, 3—7 cm. longis, 2.3—-5.4 cm.
latis, apice obtusiusculis, basi plerumque inaequaliter rotundatis interdum
cuneatis, subsessilibus vel sessilibus, venis primariis utrinsecus 8—12
patenti-ascendentibus prope marginem furcatis + anastomosantibus, supra
impressis subtus prominulis, venulis supra perspicuis; petiolo 1.5—-3 cm.
longo, dense tomentoso; inflorescentiis tomentosis, 15 cm. vel. ultra, multi-
floris; pedicello 1-1.5 mm. longo; floribus 4—5-meris; calycis tubo 1 mm.
longo, lobis 1-1.5 mm. longis; petalis circiter 1.5 mm. longis, obovato-
oblongis; filamentis petalis aequalibus, antheris 0.5 mm. longis latisque
apiculatis; disco 10-lobulato; stylis 2, circiter 1 mm. longis; ovario dense

148 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

tomentoso 2-loculari, pluri-ovulatis; capsula 5-6 mm. longa, seminibus
2—2.5 mm. longis (alis inclusis) minute papillosis.

NortHeAst New GuINneEA: Sarawaket, Clemens 6050 (in fruit), 7535a
(leaves and very young flower buds), Apr., Nov. 1937, alt. 2400-2700 and
2100 m.; vicinity of Samanzing, Clemens 9511, Tom. 1939, alt. 2100-2400 m.
(tree with brown fruit).

British New Guinea: Central Division, Mt. Tafa, Brass 4852 (Type),
May-Sept. 1933, alt. 2700 m., on cleared summit (bushy small tree 4 m.;
underside of leaves densely brown pubescent); same locality, Brass 5024,
alt. 2400 m., very common in valley forests (tree up to about 30 m. high;
irregular scant-foliaged crown; branchlets stiff; profusion of small green
flowers; fruit not seen).

In the last specimen cited there is only one old leaf (not as rugose on
the upper surface as those of the other specimens), the rest are immature,
but the lateral leaflets are sessile and the margin is inconspicuously serrate-
dentate; the inflorescence is broken, but the main axis is 26.5 cm. long,
the branches on one side (the others are broken off) are 10-12 cm. long,
the pubescence of the whole plant is definitely tomentose. In the other
specimens the lower surface of the leaves is almost lanate or villous-
tomentose, the upper noticeably rugose. Until more material is at hand
it seems best to accept all as one species. No other species of the genus
which I have seen has papillose seeds or as thick pubescence.
Spiraeopsis pubescens Perry, spec. nov.

Probabiliter arbor; ramulis tomentosis crassiusculis, + 5 mm. diam.;
stipulis non visis; foliis 5-foliolatis, coriaceis, inconspicue ae supra
rugulosis stellate: pilosulis, subtus nee une dense _stellato-pilosis
vel piloso-tomentosis, foliolis 7-17 cm. longis, 4-9 cm. latis, ellipticis vel
ovato-ellipticis, terminali basi cuneato, petiolulo 0.7—2.5 cm. longo, later-
alibus basi rotundatis vel rotundato-cuneatis, apice brevissime et obtuse
acuminatis vel acutiusculis, venis primariis utrinsecus 12-16 oblique
patentibus prope marginem + anastomosantibus, supra impressis subtus
prominulis, aur supra distincte manifestis, metolulis 3-6 mm. longis;
petiolo 3-4 cm. longo et rhachide tomentosis; inflorescentiis multifloris,
juvenilibus = 20 cm. longis, dense tomentosis, in fructu 28 cm. longis,
ramis inferioribus 15 cm. longis; pedicellis 1 mm. longis; floribus 5-meris;
calyce 1.5 mm. longo (incl. lobis vix | mm.); petalis (siccis) 1 mm. longis,
spathulatis; filamentis 2 mm. longis, antheris ovatis apiculatis; disco
5-lobato; ovario dense subvilloso- ee stylis 2 vix 1.5 mm. longis;
capsula 3. 4 mm. longa, + 2 mm. lata, seminibus vix 1.5 mm. longis,
minute rugulosis vel -+ ees

NortHeast New Guinea: Yoangen, Clemens 3429 (type), June 1936,
alt. 1500-1800 m. (specimen with infructescence and one separate leaf) ;
Ogeramnang, Clemens 4802, Dec. 1936, alt. about 1800 m. (specimen with
very young inflorescence, branches not yet elongated).

The pubescence of this species is of longer hairs and more dense than
that of S. rufa (Schltr.) Perry, and practically as long as that of S. Brassii

1949 | PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 149

described above. In fact, the latter differs only in the trifoliolate leaves
(but some species have 3-—7-foliate leaves), the slightly larger capsules
usually with shorter stellate tomentum, and minutely papillose seeds.
Spiraeopsis glabrescens Perry, spec. nov.

Arbor; ramulis minute stellato-puberulis fuscis; stipulis circiter 1 cm.

ngis, 1.5 cm. latis, late rotundatis, basi subtruncatis deinde 3-5 mm.
stipitatis, margine + dentatis utrinque stellato-puberulis; foliis 3—5-folio-
latis, novellis supra consperse stellatis subtus dense glandulosis, maturis
Bubeorincere margine serrulatis, supra fere glabris, costa nervisque —&
puberulis, subtus consperse glandulosis, costa nervisque -+ dense stellato-
puberulis, foliolis ellipticis vel obovato-ellipticis, 8-16 cm. longis, 4.5—-8
cm. latis, terminali basi cuneato, petiolulo 0.5-1.5 cm. longo, lateralibus .
basi inaequaliter rotundato-cuneatis, apice subabrupte et breviter acumi-
natis, venis primariis utrinsecus 11-15 oblique ascendentibus prope mar-
ginem -+ furcatis arcuatim anastomosantibus, supra impressis subtus
prominulis, venulis subtus manifestis, petiolulis + 4 mm. longis; petiolo
2-5 cm. longo et rhachide consperse glandulosis et minute stellato-
puberulis, infructescentiis 17 cm. vel ultra, axi et ramulis dense tomen-
tellis; pedicellis vix 1 mm. longis; floribus 5-meris; calycis lobis vix 1 mm.
longis, acutis; petalis et staminibus non visis; disco 5-lobato, lobis bilo-
bulatis; capsula dense glandulosa et tomentella, 4 mm. longa (incl. stylis
1 mm.), 2(—3)-valvata; seminibus (incl. alis) 1 mm. longis.

NortHeast New GuINEA: Sambanga, Clemens ee (TYPE), Sept.
1937, alt. 1500-1800 m. (an infructescence, a new shoot and a separate 5-
foliolate leaf) ; Sarawaket trail above iaile forest, ane. 4966, Jan. 1937,
alt. 1650-1800 m. (a medium-sized felled tree in a heap of debris).

The nearest approach to this species seems to be S. fulva (Schltr.) Perry
according to the description of the latter. Both have 3-foliolate leaves,
very short pubescence, and very small flowers on very short pedicels. How-
ever, in S. fulva (Schltr.) Perry the lower surface of the leaves is densely
pubescent, the ovary stellate-villous and 3-loculed.

Spiraeopsis Clemensiae Perry, spec. nov.

Arbuscula + 12 cm. diam.; ramulis gracilibus, 3-4 mm. diam., brun-
nescentibus, glandulosis et minute stellato-puberulis; stipulis rotundatis, 5
mm. longis latisque tomentellis; foliis 5—7-foliolatis, 10-14 cm. longis,
novellis dense glandulosis et stellato-tomentosis cito glabratis, maturis
coriaceis serratis, supra pallide olivaceis fere glabris, costa -— puberula,
subtus brunnescentibus, costa et venis minute stellato-puberulis ceterum
glabris et consperse glandulosis, foliolis oblongis, 2.5—7 cm. longis, 1.3-3
cm. latis, apice breviter acuminatis vel acutiusculis vel obtusis, basi ob-
tusis vel cuneatis, venis primariis utrinsecus + 11 patenti-ascendentibus,
arcuatis, supra impressis subtus conspicuis, terminalis petiolulo + 1.5 cm.
longo, lateralibus basi leviter inaequalibus, petiolulis 3-7 mm. longis;
petiolo vix 1.5—2.5 cm. longo et rhachide stellato-puberulis; inflorescentiis
multifloris, 14 (in fructu 24) cm. longis, axi et ramulis dense stellato-
tomentellis; pedicellis 1 mm. longis; calyce minute stellato-puberulo, tubo

150 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

0.5 mm. et lobis 1 mm. longis, 0.8 mm. latis, acutis; petalis et staminibus
non visis; disco 10-lobato; ovario circiter 1 mm. longo tomentello, stylis 3,
vix 1 mm. longis; capsula circiter 4 mm. longa (stylis inclusis), 3-valvata,
seminibus + 1.5 mm. longis (alis inclusis).

Nortueast New Guinea: Morobe District, Clemens 6049, Mar. 1937,
alt. 2400-2700 m. (a fruiting specimen wihout further data); Sattelberg,
Masak R.R., Clemens 7568 (TYPE), Nov. 1937, alt. 1800-2400 m., in mountain
forest (treelet 5 in. diam.).

Spiraeopsis myriantha (Schltr.) Perry, comb. nov.

Betchea myriantha Schlechter in Bot. Jahrb. 52: 150. 1914.

NETHERLANDS NEw GuiINnEA: Bele River, 18 km. northeast of Lake Hab-
bema, Brass & Versteegh 11137, Nov. 1938, alt. = 2350 m., common in old
secondary forest (tree 22 m. high, 33 cm. diam.; bark brown-green, smooth;
wood white; flower-buds red).

ay

This specimen for the most part has very immature flower buds. The
leaves fall within the measure of those given for this species and are 5-
foliolate, but they are almost glabrous on the lower surface. The inflo-
rescence is about 20 cm. long. The three flowers dissected had 4-loculed
ovaries, whereas the original description points to 3-loculed ones. When
there is opportunity to examine the isotype, this collection may be found
to belong elsewhere.

Spiraeopsis papuana (Pulle) Perry, comb. nov.; vel aff.

Ackama papuana Pulle in Nova Guin. 8: 645. 1912.

Betchea mee (Pulle) Schlechter in Bot. Jahrb. 52: 150. 1914, in Nova
Guin. 12: . 1917.

BRITISH | GUINEA: Central Division, Ononge Road, Dieni, Brass
3957, May 1933, alt. 500 m., rain forest (slender sparsely foliaged tree 10 m.
high; indumentum brown; upper surface of leaves shiny; inflorescence
reddish brown).

This collection differs in several respects from the original description,
but the description of the leaflets is within the limits of the specimen at
hand; however, this leaf is 5-(rather than 3-) jugate. The leaflets vary
in size, one measures 16 X 5 cm., and the apex is long-acuminate (2 cm.).
The inflorescence is 50 cm. long (including a peduncle of 14 cm. below
the lowest branches), but in other collections the size of the inflorescence
varies considerably; the number of styles is usually 3, although sometimes
only 2. Another difference which should be pointed out is that Professor
Pulle does not mention the stellate pubescence of the inflorescence, which
is a key character of the genus. In conjunction with this species S.
pometiiformis Ridl. ought to be examined.

CERATOPETALUM Smith
Ceratopetalum tetrapterum Mattfeld in Jour. Arnold Arb. 20: 432.
1939; Kanehira & Hatusima in Bot. Mag. (Tokyo) 56: 108. 1942.

NETHERLANDS NEW GUINEA: 6 km, southwest of Bernhard Camp, Iden-
burg River, Brass & Versteegh 12516, Feb. 1939, alt. 1200 m. (occasional

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX loi

tree 30 m. high, 58 cm. diam.; flowers white; fruits dark red): 4 km. south-
west of Bernhard Camp, Idenburg River, Brass & Versteegh 13162, Mar.
1939, alt. 800 m., occasional, slopes of primary rain forest (tree 32 m. high,
43 cm. diam.; sap red; flowers white) ; 2 km. southwest of Bernhard Camp,
Idenburg River, Brass & Versteegh 13183, alt. 750 m., frequent on slopes in
primary rain forest (tree 20 m.; fruits dark red) ; Boemi, Nabire, Kanehtra
& Hatusima 12801, Mar. 1940, alt. 300 m., in Agathis forest.

NortHeast New Guinea: Wareo, Clemens 1618, Jan. 1936, alt. = 600 m.

British New Guinea: Oroville Camp, Fly River (30 miles above D’Al]-
bertis Junction), Brass 7405 (ISOTYPE).

Although the leaves show considerable variation in size and may be
oblong or obovate-oblong, acute, obtuse, or obtusely acuminate, all appear
to belong to one species. The flowers are rather small and the calyx-tube
densely hirtellous; calyx-lobes four, outside sparsely hirtellous, within
densely pubescent except at the very base, ovate, 2.3 mm. long,
broad, acutish; petals lacking; filaments 1.3 mm. long, the anthers broadly
ovate-elliptic, 0.5 mm. long and broad, the apicula short, not recurved;
disk slightly lobed, glabrous; ovary densely pubescent; styles glabrous.

Ceratopetalum succirubrum C. T. White of North Queensland is prob-
ably the nearest related species, but the leaves are smaller and lanceolate,
and the anthers have a fairly prominent apiculus.

SCHIZOMERIA D. Don

The specific characters of this genus appear to be highly variable and
difficult to define. Some fruits may be ellipsoid or broadly so in the
younger forms and apparently become practically globose at maturity.
The pubescence of the inflorescence is variable. The petals are fugacious
and often irregular. With only the descriptions of S. floribunda Schltr.
and S. Pulleana O. C. Schm., and very poor material of S. serrata Hochr.,
I have been unable to separate some of the named species at hand satis-
factorily from them. For this reason, although these three species have
priority, they are not included in the key.

Key To THE SPECIES
Leaves ovate, subtruncate at base, small (less than 5 cm. long), on very
SITO tae De BIO CS. win te ere ten ee ree nore re, et 0.1.5) 6k, ee S. parvrfolia.
Leaves, if ovate, not subtruncate at base, mostly larger, with longer petioles.
Ovary densely tomentulose or tomentose.
nflorescences not quickly becoming glabrate; outside of flower
tomentulose.

Petioles 2-3.5 cm. long; leaves crenulate-serrulate: flowers
Shorl=pedicellateaiamanyn steer teehee, . Clemensiae.
Petioles 0.5-1.5 cm. long; leaves crenate-serrate; flowers ses-
SUG ic. < Sates tert ia OR ete pels ae . Versteeght.
Inflorescences quickly becoming glabrate; outside of flowers usu-
ally pubescent on the pedicel and the calyx-tube, the calyx-

lobes glabrous or with scattered hairs.

152 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Leaves oblong- eS or ovate-elliptic to lance-elliptic, the
primary veins close (5-9 mm. apart) but widely spread-
ing, the lower ones subhorizontal.

Fruit oe ovoid; leaves not glandular on the lower sur-

Ce eee Tee S. orthophlebia.

Fruit elobose leaves minutely glandular on the lower sur-

le ee ee tera rer ne terete ee NOVOGUINEENSIS.

ra Prone the primary veins rather remote (1-2 em

part), spreading ascending, 4.4 ieraseee i ea as S. Whiter.

Ovary pilose to noe oe: (if somewhat fomentilose, flowers very small,

not more than 4 mm. across).

Leaves glandular on the lower surface, glands visible without a

lens Towers 45er0US.--4ccyensesscdecer cay . adenophlla.

Leaves not glandular on the lower surface, or if glandular, glands

very = and not visible without a lens; flowers mostly
5—6-me

Leaves ene or oblong-lanceolate, gradually acuminate.
Last pedo G eee Pe eee gorumensis.

Leaves elliptic, or if lanceolate only acute or abruptly acuminate.

Fruit ellipsoid; flowers practically glabrous (ovary wit
few scattered a gigkehkaee We katicen . katastega.
Fruit globose or subglobos
eaves coriaceous or enn coriaceous; Ovary
densely pilose; flowers small (calyx lobes 1.5-
2 tit. 1002) 3...n sre eues eee eee S. Brassu,
Leaves firmly coriaceous; ovary entirely glabrous ;
flowers larger (calyx lobes 2.5-3 mm. OnE )
icina,

ae ea Sy a ew a as fee 0 ee Dg me a re er ee ee,

Schizomeria parvifolia Perry, spec. nov.

Arbor 13-14 m. alta; ramulis brunneo-ferrugineis, glabris; stipulis ob-
longis, subtruncatis, 7 mm. longis, 2 mm. latis; foliis valde coriaceis 3—4.5
cm. longis, 1.8-3 cm. latis, ovatis, obtusis, basi subtruncatis, margine
crenulato-serrulatis, sun nitidis, pallide olivaceis, subtus olivaceis, dense
reticulatis, venis primariis utrinsecus 10-12 supra interdum paululo im-
ere subtus non elevatis, oblique patentibus prope marginem arcuatis

t ++ furcatis; petiolo 2-5 mm. longo, 2 mm. lato, brunnescente; infruc-
tescentiae axi 2.5 cm. longo, glabro; fructibus ovoideis, 2.4 cm. longis, 1.8
cm. crassis.

British New Guinea: Central Division, Wharton Range, Murray Pass,
Brass 4576 (vyrr), June-Sept. 1933, alt. 2840 m., common through forests
(straight boled tree up to 13-14 m. with compact, densely foliaged crown ot
stiff erect branches; reddish brown fruit

In foliar characters this species is very distinct from the rest at hand.
The small ovate leaves on very short petioles are pale greenish yellow
with shining upper surface, under a lens very smooth except for a few
slightly impressed lines marking some of the primary veins, yet if examined
without a lens the primary veins may be easily counted. On the lower

1949 |] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 153

surface the venation is the same color, but the small and shallow depres-
sions between the veinlets (fairly coarse in this species) are slightly darker.
Schizomeria adenophylla Perry, spec. nov.

Arbor parva, 2-3 m. alta; ramulis cinereis, glabris; alabastris termi-
nalibus vernicosis; stipulis oblongis, + 4 mm. longis 1.5-2 mm. latis,
obtusis, glabris; foliis tenuiter coriaceis, lanceolate-ellipticis vel ellipticis,
3.5—6 cm. longis, 2—3 cm. latis, apice acutiusculis vel obtusis, basi cuneatis
deinde decurrentibus, margine crenulato-serrulatis, glabris, supra olivaceis,
inter venas dense reticulatis, subtus brunneo- olivaceis, reticulatis, glandu-
losis, glandulis minutis sine lente manifestis, costa utrinque prominula,
venis utrinsécus 6-8, oblique ascendentibus prope marginem furcatis et
arcuatim conjunctis utrinque paululo elevatis; petiolo 2-5 mm. longo;
inflorescentiis 1.5 cm. longis, glabris; floribus 4-meris; pedicello 1 mm.
longo; calycis lobis vix 2 mm. longis latisque, acutiusculis; petalis + 1
mm. longis, trifidis; filamentis 1.5 mm. longis, antheris apiculatis, 0.7
0.5 mm.; disco 4-partito, gibberibus bilobis; stylis 1 mm. longis; ovario
glabro, 2-loculare; loculis 2-ovulatis.

British New Guinea: Central Division, Wharton Range, Murray Pass,
Brass 4661 (Type), June-Sept. 1933, alt. 2840 m.; common in forest borders
(small tree or large bush 2-3 m. tall; leaves pale, stiff; flowers greenish, ovary
brown).

In the size of the leaves this species approaches S. ilicina (Ridl.) Schltr.;
but in the description of the latter there is no indication of glands on the
leaves, whereas in this species they can be seen without a hand lens. Then
too, the nerves are obvious and the stipules oblong.

Schizomeria gorumensis Schlechter in Bot. Jahrb. 52: 157, f. 7, H-O.

NETHERLANDS New GuINEA: 6 km. southwest of Bernhard Camp, Iden-
burg River, Brass & Versteegh 12566, 12599, Feb., Mar. 1939, alt. 1300 and
1250 m., frequent on slopes in primary rain forest (tree + 25 m. high, bark
brown or black, with some colorless resin; sapwood rose, heartwood dark
brown, flowers white; fruits brown); 2 km. southwest of Bernhard Camp,
Idenburg River, Brass & Versteegh 13192, Apr. 1939, alt. 800 m., frequent
in primary rain forest; Bernhard Camp, Idenburg River, Brass & Versteegh
13566, Apr. 1939, alt. 570 m., occasional in primary rain forest (tree 21 m.
high, 53 cm. diam.; bark brown, with a little light red resin; sapwood rose,
heartwood dark brown; flowers light yellow).

The leaves are acuminate but mostly not obtusely so, and there is a
slight variation in both the size (4.5—7.5 cm. long, 1.2—2.5 cm. broad) of
those of the first two collections cited above, and also in the length of the
petiole (5-7 mm. long); nevertheless, the lack of apparent floral differ-
ences or those of foliar texture suggests that these collections belong to
S. gorumensis Schltr. The last two collections cited are almost a perfect
match for the figure of S. Ledermannu Schlechter. However, the ovary is
pilose with fairly short hairs rather than densely villous as given in the

154 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXX

original description, the flowers are hardly more than 1 mm. pedicellate,
although the fruits (ovoid to subglobose + 1 cm. in diameter) are on
pedicels 3-5 mm. long. If S. Ledermannii Schltr. is separable from S.
gorumensis Schltr., these collections are more like S. gorumensis Schltr.
in floral characters. On the under surface of the leaves, with the aid of a
lens, one may find very minute and scattered glands.

Schizomeria pence Perry, spec. nov.

? Arbor -& . diam.; ramulis + atro-cinereis, glabris; stipulis ob-
longis ce ere resinosis, 4 mm. longis, vix 2 mm. latis; foliis coriaceis,
ovato-ellipticis vel ecice aria: ellipticis vel ovatis, 5. 9.5 cm. longis, 2.8— 5
cm. latis, apice anguste obtusis vel acutiusculis, basi subrotundatis vel
cuneatis, margine crenatis, utrinque glabris, novellis puberulis, subtus
consperse et minutissime glandulosis, venis primariis utrinsecus 13-1
supra manifestis, subtus prominulis, late patentibus fere horizontalibus,
bifurcatis prope marginem anastomosantibus, venulis utrinque dense reticu-
latis; petiolo 0.7-1.8 cm. longo; inflorescentiis immaturis 5 cm. longis,
ferrugineo-tomentellis; floribus 5—6-meris fere sessilibus, extus puberulis,
calycis lobis ovatis, + 1.5 mm. longis, acutis; petalis -- 1 mm. longis,
trifidis, staminibus calycem aequantibus, anthers 0.5 mm. longis, ellip-
soideis apiculatis; stylis 0.7 mm. longis, ovario dense tomentello, 2(or 3)-
loculare; loculis 4-ovulatis; infructescentiis 9 cm. longis; fructibus globosis
1.5 mm. diametro.

NortHeast New Guinea: Ogeramnang, Clemens 4763 (Type), Dec.
1936, alt. = 1170 m.; Matap, Clemens 41075, Feb.-Apr. 1940, alt. 1500-1800
m. (2.5 inches diameter; flowers dull cream color)

In the rather close and widely spreading primary veins this species
approaches S. orthophlebia, but the leaves of the latter are oblong-elliptic
and the fruit elongate-ovoid as well as much larger than that of S$
NOVOgUINEENSIS.

Schizomeria orthophlebia Perry, spec. nov.

Arbor usque 32 m. alta, 55 cm. diam.; ramulis glabris -_ compressis,
nodis incrassatis; stipulis triangularibus 5-7 mm. longis, 5 mm. latis,
obtusis, caducis; foliis + valde coriaceis, oblongo-ellipticis, 6-14 cm
longis, 2.5-6.3 cm. latis utrinque angustatis, basi rotundatis, apice obtusis
vel acutiusculis (vel breviter et obtuse acuminatis), margine crenulato-
serratulis, utrinque glabris, novellis crispe pilosulis, cito glabratis, venis
primaris utrinque distinctis utrinsecus costam 17-22 inferioribus sub-
horizontalibus superioribus oblique patentibus prope marginem arcuatim
conjunctis utrinque dense reticulatis vel in vetustioribus supra obscure
reticulatis; petiolo 0.8-1.7 cm. longo; (inflorescentiis immaturis, 3-10 cm.
longis, tomentulosis, bracteis ob!ongis obtusis; alabastris non expansis,
extus tomentulosis, calycis lobis intus versus apscem puberulis; antheris
ellipsoideis apiculatis; ovario tomentello 2-loculari, loculis 4-ovulatis) in-
fructescentiae axi 7-14 cm. longo, tomentello; fructibus 2-4 cm. longis,
1.2—2 cm. latis, oblongo-ovoideis, basi 2 mm. pedicellatis.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 155

NETHERLANDS New GUINEA: 9 km. northeast of Lake Habbema, Brass
& Versteegh 10463 (type), Oct. 1938, alt. = 2750 m., along a stream in
forest, common (tree 32 m. high, 55 cm. diam.; bark brown, fairly smooth;
fruit green); Bele River, 18 km. northeast of Lake Habbema, Brass & Ver-
steegh 11161, Nov. 1938, alt. 2230 m., on slope of a ridge in secondary forest,
common (tree 16 m. high, 34 cm. diam.; bark black, rough, flower-buds soft

In the material at hand this species is easily recognized by the crenulate-
serrulate oblong-elliptic leaves with fairly numerous (straight for 34 of
the distance to the margin) and almost horizontal primary veins. No
other species seen has elongate-ovoid obtusely pointed fruit. Apparently
it develops only one seed. Even though the inflorescences are given as
3—10 cm. long, all are very immature, and it is quite possible that in those
measuring 5-10 cm. long the subtending new leaves have dropped off
in process of drying.

Schizomeria ilicina (Ridl.) Schlechter in Bot. Jahrb. 55: 194. 1918.

Cremnobates ilicina Ridley in Trans. Linn. Soc. II. Bot. 9: 41, pl. 3, figs.

56-63, pl. 4, fig. 55. 1916.

NETHERLANDS New GuINEA: Bele River, 18 km. northeast of Lake Hab-
bema, Brass & Versteegh 11106, Nov. 1938, alt. 2200 m., occasional in
primary forest (tree 18 m. high, 30 cm. diam.; flower buds red; fruit green-
brown); Balim River, Brass & Ver steegh 11188, Dec. 1938, alt. 2160 m.,
forested slopes (tree 15 m. high, 40 cm. diameter; flowers white; aa
green); 15 km. southwest of Bernhz ae Camp, Idenburg River, Brass & Ver-
ae 11932, Brass 12150, Jan. 1939, alt. 1780 and 1800 m., one of the larger
trees in the mossy forest, not common; 6 km. southwest S Bernhard Camp,
Idenburg River, Brass & Versteegh 12571, 12583, 12763; Feb. 1939, alt.
1300 and 1200 m., occasional in primary forest (tree 18-22 m. high, bark
black, rough; fruit brown; flowers white); 4 km. southwest of Bernhard
Camp, Mekatinnes River, Brass & Versteegh 13114A, 13114, Mar. 1939, alt.
900 and 850 m.; frequent in pemrary rain forest (tree 25 m. high, 56 cm.
diam.; bark ray, eee smooth; flowers white; fruits brown

Nort THEAST New GUINEA: @xeranin ang, Clemens 4650, 4794, 4997, 5114a,
6358, Dec. 1936, Jan. 1937, and May 1937, alt. 1500, 1800, 1650, 1740, and
1750 m. respectively; in mossy forest (tree 75-85 ft. high) ; Sambanga,
Clemens 6845, 7744a, Nov. 1937, alt. 1500-1800 m., margin of clearing in
mountain forest (tree 10-12 inches in diameter; fruit pale khaki color) ;
Matap, Clemens 41027, Feb.-Apr. 1940, alt. 1500-1800 m. (tree 2 ft. diam. ;
flower buds immature, pale)

In designating all these collections as Schizomeria ilicina (Rid1.)
Schltr. possibly more than one entity is included; certainly great variabil-
ity is shown. Unfortunately, we have flowers and fruit of only a few
numbers; however, the obvious differences between the specimens and
Ridley’s description and figures will be noted.

In the first place Ridley’s description indicates 5-merous flowers and the
figure a 4-merous one. Of the several flowers examined in this study, all
are either 5- or 6-merous; apart from that, their pistils do not match that
in the figure. Ridley’s description “ovarium ... uniloculare, uniovulatum”

—

rs

156 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

does not fit any Schizomeria which I have examined. Perhaps, Ridley
assumed such a structure from an examination of the fruit; the styles some-
times are very short, but most of them are as long or almost as long as the
calyx-lobes. In Brass & Versteegh 12583 the inflorescence is 1.8 cm.
long and glabrous, with flowers slightly smaller than those in the other
collections. The leaves are 5-7 & 2.5—4 cm., and the reticulations of the
leaf are very dense (the shallow hollows between the veinlets are about
half as large as those in most of the other specimens). This specimen may
represent S. ilicina (Ridl.) Schlechter in the strict sense of the specific
definition. On the other hand, Brass & Versteegh 13114A has inflores-
cences up to 9 cm. long with slightly larger flowers, the axis and branchlets
of the inflorescence ++ densely puberulous, and the leaves with a similarly
dense but less easily distinguished reticulation. Brass & Versteegh 11188,
13114, and Brass 12763 have puberulous or tomentulose inflorescences
+ 3 cm. and 7 cm. long respectively, leaves 4.5-14.6 cm. long, 2-6 cm.
broad, the apex varying from rounded to subacute, the base mostly cuneate
but sometimes rounded-cuneate or rounded, the margin + crenulate-serru-
late, and both surfaces obviously somewhat more loosely reticulate than
in the two specimens discussed above. The calyx-lobes are 2—2.5 mm.
long, the petals about as long, the stamens as long or slightly longer, and
the ovary 2-3-loculed, each locule 4-ovulate. Sometimes when the ovary
is 3-loculed, one locule has only 2 ovules

Brass & Ver steegh 11106, 11188, 11932, and Brass 12150 are specimens
with fruit. The second has very immature fruits, some as small and
narrow as in Ridley’s description while others are almost as broad as long.
In the other collections the fruit is subglobose and up to 2 cm. in diameter.
The material from Northeast New Guinea in some specimens has leaves
that are more coriaceous and larger (7.5-17 > 3.5-10 cm.), but the
flowers and fruit are very like those from Netherlands New Guinea.

Brass & Versteegh 12571 may not belong to this species but is Sa
very closely allied. The fruits are large, one measuring 4 X 3.7 * 3.5
cm. The leaves are firmly coriaceous and rounded at the base. One
very young reddish tomentose inflorescence was collected. The bud-scales
are round rather than oblong or lanceolate as in the other specimens of
the species.

Schizomeria Whitei Mattfeld in Jour. Arnold Arb. 20: 435. 1939.

NortHeast New Guinea: Boana, Clemens 41770, May-Nov. 1940, alt.
750-1350 m. (flowers dull flesh-gray).

After setting up this collection as a new species from New Guinea, I
found I could not separate it in my key from S. Whitei Mattf. collected on
the Atherton Tableland, Queensland. The leaves are slightly less brittle, the
veins stand out a little more on the lower surface, the leaves are cuneate
at base (but this is true in some of those in Kajewski 1135), and the in-
florescence is slightly more tomentulose. These variations do not appear
sufficient to be considered specific.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 157

Schizomeria katastega Mattfeld in Jour. Arnold Arb. 20: 433. 1939.

Schizomeria tegens Mattf. op. cit. 434

British New GuINnEA: Western Division: Oriomo River, Wu-
roi, Brass 5804, Jan.-Mar. 1934, alt. 10-30 m., common in small forest clumps
on savannah (tree 10-12 m. high; flowers white; pale fleshy fruit) ; Oroville
Camp, Fly River, Brass 7429; Lake Daviumbu, Middle Fly River, Brass
7497, 7898; Gaima, Lower Fly River (east bank), Brass 8294.

Perhaps the field notes of Brass indicating two collections as canopy
trees and the other two as substage trees is a little misleading. Without
this I am sure it would be difficult to separate them. There is a slight
variation in the size of the flowers, but such is true even in a single collec-
tion. I am not sure that S. komaliformis Kaneh. & Hatus. is distinct
from this species. On the whole the flowers are smaller and more numer-
ous, but the inflorescences are still young.

Schizomeria Brassii Mattfeld in Jour. Arnold Arb. 20: 435. 1939, vel
aff.

Sotomon IsLanps: Bougainville: Koniguru, Buin, Kajewski 2015,
Aug. 1930, alt. 800 m., rain forest (tree up to 30 m.); Hogoro, Maisua,
Waterhouse 81, Sept. 1932, alt. = 360 m. (tree 40 ft.). Guadalcanal:
Uulolo, Tutuve Mt., Kajewski 2580, Apr. 1931, alt. 1200 m., rain forest (tree
up to 30 m. high; park covered with large corky pustules ; au ee cm. long,
1.5 cm. diam.). Ysabel: Tataba, Brass 3428, Jan. 1933, alt. 50 m., rain
forest ridges, common (very large flanged tree with deeply eee rough
eray bark and hard brown wood; leaves smooth and shining; flowers white).

The type-specimen of S. Brassii Mattf. has only fruits and leaves.
These are a fairly good match for Kajewski 2580 from Guadalcanal.
Until an isotype of S. floribunda Schltr. and good material of S. serrata
Hochr. is available for study with this species, these determinations must
be accepted as tentative.

Schizomeria Clemensiae Perry, spec. nov.

Arbor magna; ramulis atro-cinereis vel brunnescentibus 5-8 mm. dia-
metro, novellis tomentellis; stipulis non visis; foliis coriaceis, ovato-
ellipticis 8.5-19 cm. longis, 4.5—-8.5 cm. latis, apice obtusis vel subrotundatis,
basi rotundatis deinde in petiolo breviter decurrentibus, margine crenu-
latis, in sicco olivaceis, supra inconspicue subtus manifeste reticulatis,
venis primariis utrinsecus 12-17 patenti-ascendentibus, prope marginem
furcatis -- anastomosantibus, supra perspicuis, subtus prominulis, petiolo
1.7-3.5 cm. longo, subrotundato; inflorescentiis + 12 cm. longo, ferrugineo-
tomentellis; bracteis rotundatis; pedicello 1 mm. longo; flore extus
ferrugineo-tomentello vel dense puberulo, expanso 7-8 mm. lato; calycis
lobis 3-3.5 mm. longis, 1.8-2.5 mm. latis, ovatis, acutis; petalis 2.5 mm.
longis longitudine 1% superiore trifidis; filamentis vix calycis lobos aequan-
tibus, antheris 1 mm. longis 0.7 latis, cordatis, apiculatis; disco 10-gibbo,
gibberibus subglobosis 0.8 mm. diametro; ovario 2—3-loculari, dense
tomentoso, loculis 4(—5)-ovulatis; stylis 2 vel 3, circiter 2.5 mm. longis;

158 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

fructibus novellis ellipsoideis (2 x 1.5 cm.), maturis globosis, 3 cm.
diametro.

NortHeast New GurInea: Sattelberg, Clemens 439 (tyPE), 1248, Oct.
1935, alt. = 900 m., margin of forest (tree 50 ft.; flower dull yellow, of the
same odor as the fruit); Quembung, Clemens 2180, March 1936, alt. + 600
m., margin of forest (tree of giant size; fruit russet).

This species is perhaps most closely related to S. Versteeghii, but in the
latter the leaves are crenate-serrate rather than crenulate, the petioles are
not more than half as long, and the flowers are sessile.

Schizomeria Versteeghii Perry, spec. nov.

Arbor 27 m. alta, 46 cm. diam.; ramulis glabris crassis sub inflorescentia
+ 1 cm. diam., cortice cinereo vel brunnescente + rimoso: stipulis non
visis; alabastris axillaribus parvis vernicosis; foliis valde coriaceis ellipticis,
10.5-22 cm. longis, 5.7-14 cm. latis, apice rotundatis, basi rotundatis,
margine crenato-serratis, utrinque glabris, venis primariis utrinsecus 17-19
inferioribus oblique patentibus superioribus patenti-ascendentibus, mar-
ginem versus furcatis vel bi- et tri-furcatis -- anastomosantibus supra
distinctis, subtus prominulis, venis secundariis subtus manifestis, venulis
utrinque dense reticulatis; petiolo 0.5-1.5 cm. longo, crassiusculo; inflo-
rescentiis 10-16.5 cm. longis, 15—18.5 cm. latis, saepe basin prope ramosis,
multifloris, tomentosis; axi et ramulis crassiusculis; bracteis rotundatis,
concavis, 5 mm. longis, 5 mm. latis, caducis; alabastris 7 mm. longis, 5.5
mm. latis, sessilibus; floribus expansis (siccis 9 mm.), 1.2 cm. latis,
5—6-meris; calycis lobis -- 4 mm. longis, 2 mm. latis, ovatis vel lanceolatis,
acutis, intus margine et apice puberulis, ceterum glabris; petalis 3.5—4
mm. longis, trifidis, laciniis exterioribus saepe longioribus; filamentis 3.5—4
mm. longis, antheris ellipsoideis, 1 mm. longis, 0.7 mm. latis, apiculo paulo
recurvato; disco 10-12-gibbo, 1 mm. alto, 0.9 mm. crasso; stylis 2 vel 3,
2—2.5 mm. longis; ovario tomentoso, 2.5 mm. diametro; loculis 2-3,
4—6-ovulatis, ovulis biseriatis; fructibus immaturis, subglobosis vel late
ellipsoideis, + 1.5 cm. diametro.

NETHERLANDS NEW GuINEA: Bele River, 18 km. northeast of Lake Hab-
bema, Brass & Versteegh 11129 (type), Nov. 1938, alt. 2300 m., occasional
in primary forest (tree 27 m. high, 46 cm. diam.; bark 9 mm. thick, gray,
rough, fissured, with a little yellow sap; flowers white; fruits brown).

Schizomeria Versteeghii is most like the collections which have been
placed in S. floribunda Schltr. The former differs in the much shorter
petioles, the crenate-serrate rather than crenulate leaves, and the con-
siderably larger sessile flowers.

AISTOPETALUM Schlechter
At hand are four numbers of the Third Archbold Expedition plants which
seem to belong to Aistopetalum Schltr. Two have fruits, one has both
flowers and fruits, and the other only flower-buds. In addition there is
an isotype of A. tetramerum Kaneh. & Hatus. The flower-buds of the

1949 ] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 159

latter are so young that it is impossible to determine without dissection
whether tetramerous or pentamerous flowers predominate. The specimens
Brass & Versteegh 12513 and Brass 13170 have both 5- and 6-merous
flowers. In one instance a pentamerous flower had six carpels, and in
another a hexamerous flower had five carpels. One of the characters indi-
cated in the original description is two collateral ovules in each carpel, a
character re-affirmed in the description of A. tetramerum Kaneh. & Hatus.
Although several flowers from the above specimens have been dissected,
none of them appear to have more than one pendent ovule in each carpel.
The fruits are ovoid or oblong. The exocarp is fleshy and in some of the
dried fruits irregularly cracked at the apex, the endocarp or putamen is
bony, 5—6-celled, and at the apex 5—6-pointed bony projections cover a
soft area through which probably the germinating seed pushes out. The
seed is tear-shaped, the embryo is very slightly fleshy and embedded in
endosperm with the radicle pointing toward the apex of the fruit.

Aistopetalum viticoides Schltr. in Bot. Jahrb. 52: 144, f. 2, A-F. 1914;

NETHERLANDS NEW GUINEA: 6 km. SW. of Bernhard Camp, Idenburg
River, Brass & Versteegh 12513, 12565, Brass 13286, Feb., Mar. 1939, alt.
1500, 1280, and 1200 m. respectively, primary forest, on slopes of ridge
(occasional tree 15-28 m. high, 15-59 cm. diam.; flowers yellow; fruit orange
or yellow); 2 km. SW. of Bernhard Camp, Idenburg River, Brass & Ver-
steegh 13170, Mar. 1939, alt. 900 m., primary rain forest, on slope of ridge
(frequent tree 20 m. high, 45 cm. diam.; flower-buds yellow-green).

Without more material it is difficult to decide whether the fact that
Brass 13286 and Brass & Versteegh 12565 have ovoid fruits and Brass
12565 oblong fruits constitutes a specific difference or only a variation
within a species. The ovoid fruits are 1.5—-2 cm. long and 1.1-1.9 cm.
in diameter, the oblong are 2.2—2.5 cm. long and 1.5 cm. in diameter.
The foliar differences are intangible; perhaps the venation in Brass 12565
is a little less open than in the other specimens. There are no flowers for
comparison. There are simple, trifoliolate, and 5-foliolate leaves in speci-
mens of the same collection, sometimes on the same specimen. Schlechter
mentions only trifoliolate leaves, but in the figure of the species simple
leaves are also shown.

The seed in the fruit opened was about 1 cm. long, the embryo nearly
as long, with the cotyledons about twice the length of the radicle.

WEINMANNIA L.
Weinmannia purpurea Perry, spec. nov.

Arbor usque 30 m. alta, glabra; ramulis gracilibus 2-3 mm. diametro,
atro-brunnescentibus, levibus, laxe foliatis; stipulis connatis, cito caducis;
foliis simplicibus, tenuiter coriaceis, lanceolatis, 8-17 cm. longis, 2.5—4.5
cm. latis, apice obtuse acuminatis, basi cuneatis, margine serrato-dentatis,
verisimiliter glanduloso-dentatis, utrinque laxe reticulatis, supra + nitidis,

160 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

venis primariis utrinsecus 13-20 utrinque manifestis, non prominulis;
petiolo 0.5-1.5 cm. longo; racemis plerumque ternis in pedunculo com-
muno, usque 16 cm. longis; axi puberulo; pedicellis 2 mm. longis, sparsim
puberulis; floribus 4-meris, pallide purpureis; calyce glabro, lobis 1 mm.
longis, oblongo- sehen oon: petalis ellipticis 1.5 mm. longis, 1 mm.
latis; staminibus 2 mm. longis, anthers apiculatis; glandulis cylindricis,
Vix 1 mm. longis; ovario sparsim puberulo, stylis circiter 1.2 mm. longis;
capsulis oblongis, 6 (stylis inclusis) mm. longis, seminibus (siccis) 1 mm

longis, 0.3 mm. latis, ee barbato-pilosis, pilis circiter 1 mm. longis.

Sotomon IsLtanps: Bougainville: Kupei Gold Field, Kajewski
1738 (TYPE), eel 1930, alt. 100 m., rain forest (tree up to 20 m. high,
found growing at very high a ides) flowers very pale purple on purple
stalks; fruit purple, lengt th including styles 6 mn. 1.5: mim diam, |;

Most of the racemes are in threes at the apex of a common peduncle
3-6 cm. long, but in one instance there are two nodes; the lower one has
two opposite racemes and the upper has three.

From the description of W. marquesana F. Brown, it seems as if these
might be related species. The latter has broader and shorter leaves with
longer petioles; nothing is said of the stipules. In W. purpurea, the short
connate stipules, although loose, still encircle the base of the inflorescence
and the younger branchlets.

Weinmannia Ledermannii Schlechter in Bot. Jahrb. 52: 162. 1914.

NortHeast New Guinea: Wareo, Clemens 1407, Dec. 1935, alt. = 600
m., margin of jungle (tree 100 ft. high; flowers cream-colored; fruits pale
tinged with pink).

The leaves of this collection are 2—5-jugate, the petiole and rhachis are
shortly and somewhat densely pilose, the midrib on the lower surface of
the leaflets is less densely so. The axis of the inflorescence is only pu-
berulous. Apart from these differences in pubescence the specimen agrees
with the description reasonably well.

Weinmannia urdanetensis Elmer, Leafl. Philipp. Bot. 7: 2608. 1915.
NeruHerLANps New Guinea: 15 km. southwest of Bernhard Camp, Iden-
burg River, Brass & Versteegh 11908, Jan. 1939, alt. 1800 m., frequent in
mossy-forest of ridges (tree 26 m. high; flowers white; fruit red); 8 km.
southwest of Bernhard Camp, Idenburg River, Brass 12733, Feb. 1939, alt.
1600 m., one of the principal trees in the mossy forests at this altitude.
For the time being I have placed these collections in W. urdanetensis
Elmer, originally described from the Philippines. There is at hand an
isotype of this species and another collection in young fruit (seeds imma-
ture) from Mt. Palimlim, Luzon. The lower leaves of the latter specimen
are a little larger than in the type and about the same size as the smaller
ones in the New Guinean collections. The leaves in these are 4—9-jugate.
the lateral leaflets are 0.9-1.8 cm. long and the lower surface is more
pilose than in the Philippine material, in which the lower surface is often
glabrous except for a few hairs along the midrib; here also the veins are a
little less easily seen than in the New Guinean collections. In the latter,

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 161

the mature capsules are 44.5 mm. long (including the 1 mm. style), the
seeds 0.8-0.9 mm. long 0.4 mm. wide, not including the tufts of hairs at
the ends of the seeds.

Weinmannia novoguineensis Perry, spec. nov.

Arbor parva, 10-13 cm. diam.; ramulis puberulis vel breviter pilosis;
stipulis subreniformibus intus glabris, venosis, extus puberulis, 6-8 mm.
longis, 1 cm. latis; foliis 4-10 cm. longis, 4—13-jugis, foliolis chartaceis vel
tenuiter coriaceis, serratis, supra costa puberula excepta glabris, subtus
costa praecipue parte inferiore dense et breviter pilosa excepta plerumque
glabris nonnumquam hinc inde pilis conspersis, terminali lanceolato acu-
tiusculo, 2.2-3 cm. longo, 0.7—1 cm. a petiolulo circiter 3 mm. ee
ieeralinuc 1-2.4 cm. longis, 0.5-0.8 . latis, oblongis utrinque angu
tatis, apice acutiusculis vel ee basi inaequaliter obtusis men
bus; petiolo vix 1 cm. longo et rhachide + crispe pilosulis; infructescentiis
+ 8 cm. longis, axi puberulo vel + pilosulo; pedicellis 2-3 mm. longis,
puberulis; Aoribus 4-meris; calycis lobis 1 mm. longis, 0.7 mm. latis,
glabris; petalis et staminibus non visis; glandulis 0.5 mm. longis; capsula
circiter 6 mm. longa (incl. stylis 1.5 mm.), seminibus 1 mm. longis, 0.4
mm. latis, + pilosulis.

NortTHEASt New Guinea: Sarawaket, Clemens 7517 (Type), Nov. 1937,
alt. 1800-2400 m., in mountain forest (small tree 4-5 inches diameter; fruit
pale green).

This species differs from W. Pullei Schltr. in the sessile leaflets with
pubescent midrib on the lower surface, the more heavily pubescent leaf
rhachis, the longer pedicels, and the glabrous calyx.

Weinmannia trichophora Perry, spec. nov.

Arbor 12-15 cm. diam.; ramulis parvis tomentosis dense foliatis; stipulis
subreniformibus, 4 mm. longis, 5 mm. latis, intus basin versus et extus
breviter appresse pilosis; foliis 2-5 cm. longis, 2—7-jugis, foliolis sub-
coriaceis serrulatis, supra consperse puberulis, subtus (costa dense) breviter
appresse pilosis, ‘site crispulis, terminali 1-1.5 cm. longo, 0.5—0.7 cm.
lato, oblongo obtusiusculo vel acutiusculo, petiolulo circiter 2 mm. longo,
lateralibus 0.8-1.6 cm. longis, 0.4-0.6 cm. latis, oblongis acutiusculis vel
obtusis, basi vix inaequaliter obtusis sessilibus; petiolo 3-6 mm. longo et
rhachide tomentosis; inflorescentiis + 5 cm. longis, axi breviter piloso:
pedicellis 1.5 mm. longis puberulis; floribus 4-meris, ante anthesin; calyce
puberulo, lobis ovatis obtusis 1 mm. longis, 0.7 mm. latis; petalis 1.5 mm.
longis, 1 mm. latis, ellipticis; antheris minute apiculatis; glandulis fere
0.4 mm. longis; ovario pubescente, stylis glabris.

NortHeast New GuIneAa: vicinity of Samanzing, Clemens 9498 (TYPE),
Jan. 1939, alt. 2100-2400 m., in mountain bush (tree 5-6 in. diam.; flowers
white-cream).

This species may be related to W. urdanetensis Elmer, but the leaves
are smaller, the leaflets shorter and broader and much more pubescent on
the lower surface.

162 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

Weinmannia ysabelensis Perry, spec. nov.

et patenti-pilosulis, internodiis 1-2 cm. longis; stipulis cito caducis, non
visis; foliis (petiolo incluso) 3.5—7.5 cm. longis, 1—4-jugis, foliolis tenuiter
coriaceis vel chartaceis, crenulato-serratis, supra atro-olivaceis, glabris
subtus pallidioribus, costa + pilosulis, supra laxe subtus dense reticulatis,
terminali majore, 2.7-4 cm. longo, 0.8-1 cm. lato, lanceolato, acuto,
petiolulo 2-3 mm. longo, lateralibus 1-2.3 cm. longis, 0.5-0.9 cm. latis,
oblongo-lanceolatis, apice acutiusculis, basi inaequaliter obtusis, sessilibus
vel subsessilibus; petiolo 0.5-1 cm. longo et rhachide supra subplanis
obscure marginatis margine glabris, ceterum utrinque patenti-pilosulis;
inflorescentiis 5—8 cm. longis, singulis vel binis, axi puberulo; pedicellis +
puberulis, 2 mm. longis; floribus 4-meris; calycis basi hinc inde minute
pilis dispositis, lobis vix 1 mm. longis, ovato-oblongis rotundatis; petalis
1.4 mm. longis, obovato-ellipticis; filamentis 1.4 mm. longis, antheris 0.5
mm. longis minute apiculatis; glandulis cylindricis, 0.4 mm. longis, 0.15-
0.2 mm. latis; stylis fere 1 mm. longis, ovario pubescente, circiter 1 mm.
longo, 0.7 mm. diam.; capsulis (stylis fere 2 mm. inclusis) 5-6 mm. longis,
1.5 mm. diam.; seminibus circiter 1.5 mm. longis, 0.5 mm. latis, +
pilosulis.

SoLoMON Istanps: Ysabel: Tiratona, Brass 3215 (tyre), Nov. 1932,
alt. 600 m., mountain forests, conmon (thick-boled tree 25 m. tall; reddish
brown flaky bark and hard red wood; flower white; fruit pink).

We have no specimens at hand which suggest a close relationship with
this species. Weinmannia virgulata Schltr. has 2—4-jugate leaves, but
3-merous flowers; W. Macgillivrayi Seem. has 5-9-jugate leaves with
larger leaflets. It should be noted that the seeds have scattered hairs
on the surface, not confined to the ends as in W. Versteeghii.

Arbor 25 m. alta; ramulis gracilibus, novellis + 2 mm. diam., breviter

Weinmannia Versteeghii Perry, spec. nov.

Arbor 19 m, alta, 25 cm. diam., fere glabra, ramulis cinereo-brunnes-
centibus, novellis + ferrugineis, minute et sparsim puberulis, ultimis
gracilibus, + 1 mm. diam.; stipulis subrotundatis 4-7 mm. longis latisque,
extus sparsim et minute appresse puberulis; foliis 2.5-5 cm. longis, 1—4-
jugis, foliolis glabris, terminali obovato basi anguste cuneato, subsessili,
lateralibus oblanceolatis, sessilibus, 0.7-1.8 cm. longis 0.4— 7 cm. lati.
apice obtusis basi anguste et inaequaliter obtusis, margine minute crenu-
lato-serratis, supra (in sicco) atro-olivaceis, costa pallidiora, subtus pallidi-
oribus, venis primariis ascendentibus + anastomosantibus utrinque
manifestis; petiolo 6-10 mm. longo et rhachide anguste marginatis; racemis
vulgo singulis 4-8 cm. longis, axi minute puberulo; floribus 4-meris; calycis
lobis 1.3 mm. longis, obtusis, petalis obovato-ellipticis, 1.7 mm. longis,
1.4 mm. latis, rotundatis sparsim ciliolatis; filamentis vix 3 mm. longis;
glandulis Gilindicis vix 1 mm. longis; ovario puberulo, 2 mm. longo,
stylis vix 1 mm. longis; capsulis 6 mm. longis, puberulis, seminibus 1.5
mm. longis utrinque barbato-pilosis.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 163

NETHERLANDS New GUINEA: 9 km. northeast of Lake Habbema, Brass &
Versteegh 10469 (rypr), Oct. 1938, alt. + 2700 m., frequent in forest (sub-
stage tree 19 m. high, 25 cm. diameter; bark 4 mm. cee gray-brown, rough;
outer wood yellow, inner red)

This is the only species reported from New Guinea with the narrowly
margined petiole and rhachis. In habit it suggests W. tannaensis Guil-
laumin, but in the latter the leaflets are narrower at the apex, and the
margin of the petiole and rhachis is broader.

PULLEA Schlechter
In the consideration of the genus Pudlea Schltr. I am indebted to Dr.
F. P. Jonker for very generously providing for comparison two leaves and
a branch of the inflorescence from the original collections of P. glabra
Schltr.
KEY TO THE SPECIES

Flowers tapering into a very short (about 1 mm. long) stipe...... P. glabra.
Flowers sessile
Leaves velvety-pubescent beneath.............e eee eee ee eee P. mollis.

Leaves glabrous or only sparsely puberulous with minute hairs.
Leaves acute or very obtusely short-acuminate.
Flowers 10-12 on a cluster; leaves very obtusely short-acumi-
nate, lower surface fasciculate-barbate in axils between

the primary veins and the midrib........ P. Versteeghit.
Flowers 5-7 in a cluster; leaves acute, not fasciculate-barbate.
sie cit ai Sl aes mene amare abe pete aL dh eRe nas eae ays PB Te eens.

Leaves obtuse, rounded, or emarginate.
Stipules ae leaves small, = 5.4 x 3 cm., reticulations
es aisha sgn. ns anne mee P. papuana.
Stipules eee es obtuse; leaves vee 49 x 3-6
cm., reticulations not prominent......... . Clemensiae.
Pullea decipiens Perry, spec. nov.

Arbor 29 m. alta, 45 cm. diam.; cortice novello nigrescente, vetustiore
brunneo-cinereo, glabro, lenticellato; foliis in sicco brunnescentibus subtus
paullo pallidioribus, coriaceis, lanceolatis vel elliptico-lanceolatis, 3-7.5
cm. longis, 1-3 cm. latis (av. 5-6 X 2.5 cm.), apice acutis, basi cuneatis
in petiolum 3-6 mm. longum angustatis, margine crenulatis, utrinque
glabris, venis primariis utrinsecus 7-10, supra manifestis subtus prominulis,
venulis creberrimis dense reticulatis; inflorescentiis puberulis, paniculatis,
lateralibus vel apicalibus, 2.5-6 cm. longis, saepe 7 ex ramulorum apice
ortis; axi 1-3 cm. longo, ramulis 2—8-natim verticillatis nonnumquam
era verticillatis, 3-15 mm. longis; floribus (post anthesin) parvis,
sessilibus, in capitulis (-+ 5-floris); calycis lobis 1.5-1.9 mm. longis,
0.7-1.0 mm. latis, oblongis, obtusis, membranaceis; staminibus circiter
3 mm. longis; disci squamis 0.2-0.3 mm. longis latisque, apice incrassatis,
emarginatis, bifoveolatis; stylis 2, liberis, circiter 3 mm. longis, basi ex-
ceptis glabris; ovario hemi-infero, extus 1 mm. longo.

104 JOURNAL OF THE ARNOLD ARBORETUM vol. sk

NETHERLANDS New Guinea: 4 km. SW. of Bernhard Camp, Idenburg
River, Brass & Versteegh 13115 (rype), March 1939, alt. 850 m., primary
rain forest on plain; common tree 29 m. high, 45 cm. diam.; bark gray, scaly;
wood red-brown; flowers light yellow).

This species of Pudlea seems to be most like the description of P. papu-
ana Gibbs. It differs in having acute, distinctly crenulate leaves with
lateral veins more raised than the reticulate areas between them, and
flowers with obtuse calyx-lobes. Possibly more material would show the
two are identical. One needs to examine the type to evaluate the differ-
ences. Pullea papuana Gibbs is described as having obtuse minutely
denticulate leaves, and acute calyx-lobes.

Pullea Versteeghii Perry, spec. nov.

Arbor usque 31 m. alta, 48 cm. diam.; cortice novello brunnescente,
vetustiore pallidiore, glabro, lenticellato; foliis ellipticis vel oblongo:
ellipticis, 4.5-10 cm. longis, 2.5-5 cm. latis, apice breviter et obtusissime
acuminatis, acumine 4-6 mm. longo latoque, basi obtusis deinde breviter
cuneatis, margine crenulato-denticulatis, supra olivaceis glabris, subtus
brunnescentibus, venis primariis -- pubescentibus, in axillis inter costam
et venas fasciculato-barbatis, utrinsecus 7-9 patenti-adscendentibus mar-
ginem versus breviter furcatis supra perspicuis, subtus prominulis, venulis
laxe reticulatis utrinque paullo elevatis sine lente manifestis, minoribus
copiose reticulatis supra obscuris subtus sub lente Gianitestis: petiolo
0.5-1 cm. longo; inflorescentiis dense puberulis, paniculatis, acillaribus
et apicalibus, usque 8 cm. longis latisque, circiter 7 ex ramulorum apice
ortis; axi 1.5—4 cm. longo; ramulis 2—8-natim verticillatis, nonnumquam
iterum verticillatis, 6-20 mm. longis, bracteis 1-1.5 mm. ones vix 0.5 mm.
latis, oblongo-ovatis; floribus (post anthesin) parvis eres in capi-
allie (circiter 9-12-floris); calycis lobis 2 mm. longis, 1.5 latis;
staminibus 2.7 mm. longis; disci squamis 0.5 mm. jones: v1x x 04 mm.
latis, apice incrassatis et fere planis; ovario extus -- 1 mm, longo, apice

subconico: stylis 2.5-3 mm. longis inferiore 14 pubescentibus.

NETHERLANDS NEw GuInea: Bele River, 18 km. NE. of Lake oe
Brass & Versteegh 11123, 11146 orale 11147, Nov. 1938, alt. = 2300 n
primary and secondary forest (rare tree 13-31 m. high, 35— 48 cm, a

bark 4-9 mm. thick, rough, white ee in 11147) ; flowers white) ; Balim
River, Brass & Versicegh 11185, Dec. 1938, alt. 2160 m., forest of the slopes.

The immediately distinctive character of this species is found on the
lower surface of the leaves, the tufts of hair in the axils between the midrib
and the primary veins. Also there is a coarse reticulation which can be
seen with the naked eye as well as a finer one visible-on the under surface
only with the aid of a lens. The contrast in the reticulation is much more
marked than in P. decipiens. The flowers are in clusters of 9-12. The
style is more or less pubescent the lower third of its length. Brass &
Versteegh 11147 has lanceolate rather than elliptic leaves, 3-10 cm . long,
1.5—4 cm. wide, and flowers slightly smaller than in the specimen desig-
nated as the type.

1949] PERRY, PLANTAE PAPUANAE ARCHBOLDIANAE, XIX 165

Pullea Clemensiae Perry, spec. nov.

Arbor 45-90 cm. diam.; cortice brunnescente vel cinereo-brunnescente,
lenticellato, glabro, novello minute appresse pubescente; alabastris, stipulis
(obtuse oblongo-lanceolatis cito caducis) et foliis novellis dense cinereo-pu-
bescentibus; foliis coriaceis, ellipticis 4-9 cm. longis, 2.3-6 cm. latis, in sicco
brunnescentibus apice rotundatis vel obtusis, basi cuneatis, margine incon-
spicue serrulatis, supra glabris subtus consperse puberulis vel fere glabris,
venis primariis utrinsecus + 8 oblique ascendentibus prope marginem
furcatis supra distinctis, subtus prominulis, venulis laxe reticulatis utrinque
manifestis non conspicuis; petiolo 1-1.5 cm. longo; inflorescentiis 4—5
cm. longis, fractis dense cinereo-pubescentibus, pilis brevibus + appressis;
capitulis 5—7-floris; floribus basi bracteatis, bracteis 1-1.5 mm. longis, 1
mm. latis, oblongis obtusis; fructibus sessilibus pubescentibus, calycis
lobis 1.5 mm. longis, 1 mm. latis, obtusis vel acutiusculis; filamentis 2
mm. longis; disco squamis + 0.3 mm. longis; ovario maturo 2 mm. diam.,
1.5 mm. longo; stylis 3 mm. longis; semine levi subrotundato 1 mm. longo,
0.9 mm. lato, compresso.

NortHeAst New GuINEA: above ae ee 8989, Oct. 1938,
alt. = 1800 m., in mountain bush (tree 18 in. 2 it. diam.) ; Samanzing,
Clemens 9068 (TYPE), Oct. 1938, alt. = 1800 m., ae 2-3 5 eee Saman-
zing vicinity, Clemens 9283A, es 1938, alt. 1200-1800 m., in mountain
bush (tree 10-12 in. diam.; flower buds yellowish green) ; Wantoa t (Wan-
tot), Clemens 10995, Jan. 1940, Mt "1050. 1800 m. (tree 40-50 ft. high; flowers
pale green-white).

The specimens Clemens 8989 and 9068 are in fruit. On some branchlets
of the inflorescence where the fruits have already fallen the bracts are
still on the axis. In the other two collections the flowers are in bud, but
in Clemens 9283 the bracts subtending the minute flowers have already
fallen. The leaves somewhat resemble those of P. Versteeghi, but are not
fasciculate-barbate in the axils between the primary veins and the midrib,
and the apex is mostly rounded; further, the inflorescences are not nearly
so large or so copiously flowered

ARNOLD ARBORETUM,

Harvarp UNIVERSITY.

166 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. Xxx

STUDIES IN THE THEACEAE XVIII
THE WEST INDIAN SPECIES OF LAPLACEA

CLARENCE E. Kosusk1

THE GENUS Laplacea was named by Humboldt, Bonpland and Kunth
(1822) in honor of Marquis de [Pierre Simon] Laplace (1749-1827), a
French mathematician and astronomer of the first order, and probably
the ranking man of his time in these fields. Born the son of a small
farmer in Beaumont-en-Auge in Normandy on March 28, 1749, Laplace
owed his education to the interest shown in him by wealthy landowners.
In 1767 he left Beaumont, where he had been teaching mathematics, and
accepted the post of Professor of Mathematics at Ecole Militaire in Paris.
His Mécanique céléste (5 vols., 1799-1825) is considered today a monu-
ment of mathematical genius. His complete works were published by the
French government in seven volumes (1843-47), and as late as 1912 a
second edition containing additional matter was published.

A year earlier (1821) two other genera, Wikstroemia and Lindleya,
were described by Schrader and Nees von Esenbeck respectively. These
were named in honor of Dr. J. E. Wikstrom and the English botanist John
Lindley (1799-1865).

A fourth name, Haemocharis, was very casually introduced by Salisbury
in 1806 and taken up by 2} Martius and Zuccarini in 1826.

Of the four names mentioned for the genus, Laplacea and Haemochavris
received the greatest recognition. Considerable controversy, of a mild
nature, resulted among adherents of all four generic names, but Laplacea
and Haemocharis appealed most to the students of botany. Such workers
as Humboldt, Bonpland and Kunth, DeCandolle, Sprengel, Cambessedes,
Spach, Endlicher, G. Don, Walpers, Wawra, Melchior, Schmidt, and Lemée
favored Laplacea, while Martius and Zuccarini, Choisy, Baillon, Kuntze,
Szyszylowicz, and Urban employed the name Haemocharis. Later Rehder
pointed out that Laplacea appeared to be recognized by the majority of
workers and suggested its retention as a nomen conservandum, and as such
it appeared in Kew Bull. 1940: 112. 1940.

Blake (Contrib. Gray Herb. 53: 39. 1918) made a belated plea for
Wikstroemia, the earliest name applied to the genus. He gave an excellent
and very accurate review of the history of the genus. However, the name
Wikstroemia could not be used, since it had already been placed on the
list of nomina conservanda in 1905 as applying to a genus in a different
family, the Thymelaeaceae.

Korthals, in 1842, described the genus Closaschima, which has since
been reduced to synonymy under Laplacea. The study of the Asiatic
material will not be included in this review but will be taken up later under
a series of Asiatic studies.

1949 ] KOBUSKI, STUDIES IN THE THEACEAE, XVIII. 167

The present study of the American species of Laplacea has been divided
into two parts, the first including the West Indian species and the second
those of South and Central America. This paper deals only with the
West Indian species.

Of two species only, namely, Laplacea haematoxylon (Sw.) D. Don and
L. portoricensis (Klug & Urban) Dyer, could one feel that material ample
for complete and satisfactory study had been accumulated in the American
herbaria. Of all the others the material is very sparse, in several instances
the species having been collected only once or twice. Frequently great
difficulty was experienced in preparing accounts of some earlier inade-
quately described species. In the genus as a whole the flowers are fairly
large and conspicuous. Hence, often only a single flower was included
with the collected specimen, and many times this single flower was incom-
plete. In such cases dissections of boiled material could not be made
with a clear conscience, and it was necessary to employ great caution in
working with the material so as not to ruin it for study by future workers.

As far as the West Indian species are concerned, no species have been
reduced to synonymy in this study. In the province of Oriente, Cuba, five
species and one variety are recognized. Oddly enough, the variety (new)
has the most collected specimens. In the case of all the entities included
here, very little material other than the original type was available for
study. It is obvious that all are very closely related, and it is possible
that future collections may prove that only two or perhaps three true
species are represented. The same is true for the three species from Haiti.

An attempt has been made to bring together all the species of the genus,
giving as complete descriptions as possible and supplying the synonymy
and literature in each case. Brief discussions showing relationships be-
tween species and the outstanding characteristics helpful in identification
have also been supplied.

The following abbreviations are used to designate the herbaria cited in
this paper: A=Arnold Arboretum; Ch=Chicago Natural History Mu-
seum; G=Gray Herbarium; Mo= Missouri Botanical Garden; NY=New
York ‘Botanical Garden; and US=United States National Herbarium.

Reece eee Bonpland & Kunth, Nov. Gen. Sp. 5: 161 (207, ed.
folio), ¢. 461. 1822. — DeCandolle, Prodr. Ls /52 7: 1824, — Sprengel,
Syst. Veg. 2: ean 1825. — Cambessedes in St. Hilaire, Fl. Bras.
Merid. 1: 299. 1827. — Spach, Hist. Nat. Veg. 4: 76. 1835. — End-
licher, Gen. Pl. 1020. 1840.— G. Don, Gen. Syst. 1: 569. 1840. —
Walpers, Repert. Bot. Syst. 1: 372. 1842; 2: 801. 1843; 7: 367. 1868.
— Bentham & Hooker, Gen. Pl. 1: 186. 1862. Wawra in Martius,
Fl. Bras. 12(1): 287. 1886.— Melchior in Nat. Pflanzenfam. ed. 2,
21: 135. 1925.—O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg.
Veg. 29: 14. 1931. — Lemée, Dict. Pl. Phan. 3: 947. 1931. — Sprague
et al. in Kew Bull. 1940: 112. 1940. — Marie-Victorin in Contrib.
Inst. Bot. Univ. Montréal 49: 69. 1944. — Kobuski in Jour. Arnold
Arb. 28: 435. 1947.

168 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Wikstroemia Schrader in Gotting. Gel. Anzeig. 1821(71): 710. May 5,
1821.— Pontin [Editor] in Vet. Akad. Handl. Stockholm 1821: 168.
1821, obs. in footnote. — Blake in er Gray Herb. n. s. 53: 36.
1918.— Urban, Fl. Ind. Occ. 8: 759.

Lindieya Nees in Flora 4: 299, May yi eon op. cit. 328 (as syn. of
Wikstroemia). June 7, 1821

Haemocharis R. A. Salisbury, Paradisus Lond. 1: sub ¢, 56. 1806.— Mar-
tius & Zuccarini, Nov. Gen, Sp. 1: 106. 1826.— Choisy in Mém. Soc.
Phys. Hist. Nat. Genéve 1: 142 (Mém. Ternstr. 57). 1855. — Baillon,
Hist. Pl. 4: 253. 1873.—O. Kuntze, Rev. Gen. Pl. 1: 62. 1891.
Szyszylowicz in Nat. Pflanzenfam. III. 6: 185. 1893.— Urban in Bot.
Jahrb. 21: 545. 1896.

Closaschima Korthals, Verh. Nat. Gesch. Bot. ed. Temminck 139, ¢. 28.
1842

Trees or shrubs. Leaves coriaceous, subcoriaceous or membranaceous,
alternate, rarely asymmetrical, usually rounded or obtuse at the apex, often
emarginate, tapering at the base into a short petiole, the margins denticu-
late or crenulate, rarely entire. Flowers solitary in the axils of the upper
branches; bracteoles 2 (or more), sepaloid, arranged along the peduncle,
quickly caducous; sepals 5 (rarely more), thick-coriaceous, unequal, gradu-
ating in size and shape from bracteoles to petals, usually appressed-pubes-
cent on the dorsal surface, deciduous or persistent; petals 5, rarely more,
unequal, usually membranaceous, usually emarginate at the apex and occa-
sionally pubescent on the dorsal surface, the outer petal often resembling
the inner sepal; stamens many, seriate, the filaments adnate to the base of
the petals, the anthers versatile; ovary basically 5-celled, occasionally
4-celled, rarely 6—10-celled, sericeous, the ovules 4 or more in each cell;
the styles usually 5, occasionally 3 or 6, rarely 1, the stigmas usually the
same in number as the styles, 5 when the style is solitary. Fruit an elon-
gate, more or less woody, loculicidal capsule with persistent columella;
seeds flat or compressed, drawn out into an oblong, membranaceous wing.

Type species: Laplacea speciosa HBK.

Urban, in one of his earlier works, in Bot. Jahrb. 21: 545-549. 1896,
designated the flowers of Laplacea (under Haemocharis) as monoecious,
and in all six of the species recorded (L. Wrightii, L. Curtyana, L. haema-
toxylon, L. villosa, L. alpestris and L. portoricensis) he carefully described
both staminate and pistillate flowers. In this paper no reference has been
made to monoecious flowers in either the generic or specific descriptions
because no flowers were examined that were not hermaphroditic, according
to my interpretation. If, in my study of the South and Central American
species, I find material producing monoecious flowers, I will revise my
generic description accordingly. It is interesting to note that in describing
L. cymatoneura at a later date (1925), Urban made no mention of mono-
ecious flowers. O. C. Schmidt reviewed the West Indian species of the
genus Laplacea in 1931 and described L. Samuelssonii without designating
staminate or pistillate flowers. Also in describing L. Urbani in 1925

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 169

Schmidt intimated that the flowers were hermaphroditic. Marie-Victorin
(1944), who described the latest species (L. moaensis) attributed to the
genus, found only hermaphroditic flowers.

As in most other genera of the Theaceae, the terminal leaf-bud in
Laplacea appears to be the true criterion of pubescence. In some species
the mature leaves may often appear. glabrous, yet an examination of the
leaf-bud will show that originally the leaves were appressed-pubescent,
pilose or even villous. The only truly glabrous species in the West Indies
is L. Wrightii.

The texture of the leaf is often very difficult to determine in the dried
state. Seldom does one find a species with thin membranaceous leaves.
The texture is such that when dried the leaf appears subcoriaceous, at
least, even though the collector may have signified the texture to be mem-
branaceous.

The apex of the leaf is usually obtuse or rounded, often retuse or emar-
ginate. The base is usually long tapering, often very finely decurrent into
a rather short petiole, which may appear longer on casual observation
because of the deceptive decurrent base.

The margins are usually finely denticulate, often with a minute point at
the end of the denticulation which is easily and often broken off. More
rare are crenulate or entire margins.

The bracteoles, two or more in number, are as a rule quickly caducous.
These may be distributed along the pedicel or close to the calyx. They are
sepaloid in appearance and grade into the sepals. Like the sepals, they are
usually subligneous and pubescent on the dorsal surface. The sepals them-
selves grade from the bracteoles into the outer petals. The inner sepals
tend to have a more membranaceous margin and occasionally it is difficult
to distinguish between the inner sepal and the outer petal. The inner
petals are considerably larger than the sepals, white in color, thinner in
texture, and usually deeply emarginate at the apex. Occasionally one may
find the outer petal to be entire and narrower at the apex. Pubescence on
the dorsal surface of the petals is common but to a lesser degree than is
found on the sepals.

The stamens of the various species are quite similar, differing mostly in
the length of the filament and the number of series. The ovary, always
sericeous, offers little variation. ‘The accepted number of cells is five, but
occasionally there are four. Laplacea portoricensis is an extreme excep-
tion, having six to ten cells in the ovary.

The style offers considerable variation. In most species the styles
number five with an individual stigma on each style. In L. Samuelssonii
the styles number three, while in L. benitoensis the styles are joined in a
single entire style which is topped by five stigmas. Usually the styles are
erect. Exceptions are found in L. alpestris and L. cymatoneura, where the
styles are horizontal and the stigmas considerably enlarged.

The capsule offers little variation. In most species the number of valves,
corresponding with the number of cells in the ovary, is five. In L. porto-

170 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

ricensis the valves number six to ten. The surface of the capsule is usually
appressed-pubescent or glabrescent, never, as far as I know, sericeous.
The length of the capsule is usually more than two centimeters. I

alpestris and L. cymatoneura the capsule measures approximately one

centimeter.
KEY

TOVTEHE. SPECIES

A. Terminal leaf-bud pubescent.
ary and fruit 4—5-celled.

C. Styles lone the stigmas flattened and considerably dilated.
. Leaves 1- 1.5 cm. wide, long-pilose beneath; petals lightly
ee on the dorsal surface. (Haiti).1. L. alpestris.
DD. Leaves 2-2.5 cm. wide, glabrous beneath; petals glabrous

on the dorsal surface. (Haiti; Dominican Republic

os

Ue mea aula ears ee ees . cymatoneura.

C. Styles erect, the stigmas not flattened and dilated as above
eaves linear-lanceolate or oblong-elliptic, acute at the

apex, about

five times longer than broad. (Cuba)......
Maciea tees hkeeeueeeets 4. L. angustifolia.

DD. Leaves obtuse or rounded at the apex, not linear-lanceo-
late or oblong-elliptic, seldom more than three times longer

than broad.

E. Style entire, topped by several ao (Cub
6.

ee ee ee

hieake maces ote oe ae ee

EE. Styles free to the base.

tyles 3 in number, ca. 0.6 mm. long. (Haiti)
Shi hee eevee ueeeeue eas 3. L. Samuelssonit.

FF. Styles 5 in number, usually ca. 1 cm. long.

. Leaf-margin crenate-undulate; upper pair of

veins arching, first upward and out toward
the margin, then back toward the midrib,
appearing to join near the point of a deep
emargination, forming a heart-shaped fig-
Mes Ube sca sceadet ot einer 7. L. Urbant.

. Leaf-margin entire, denticulate or denticu-
‘late-crenulate; upper pair of veins not as
above.

H. Petals 8-10 mm. long, approximately
equaling the calyx-lobes in length; leaf-
margin denticulate-crenulate along ae
upper half of one side of the margin, the
other side entire; leaves often oblique.
Chee ta4 ay aneueaee . moaensts.

HH. Petals 15-30 mm. iio always greatly

exceeding the calyx-lobes; leaf-margins

entire or denticulate along both sides:

leaves symmetrical.

I. Leaf-margin entire; the under sur-
face villous when young and clearly
showing the original folds of the
leaf. (Cuba)....10. L. Curtyana.

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 171

II. Leaf-margin denticulate; the under
surface pubescent or glabrous but
not showing the original folds of the
leaf.

J. Branchlets and stems villous at
iMatunttves 4(Glainaica) senses cours
Lea eahi et ee ene Lie ie willosa:

. Branchlets and stems pubescent
when very young, glabrescent at
maturity

K. Leaves chartaceous; petals
glabrous or nearly so on
the external surface; sta-
mens 4—5-seriate; pedun-
cles 2-5 mm. long. (Ja-
TULA TG oly ly he eee i niah

L. haematoxylon.

KK. eae coriaceous; petals
pubescent on the external
surface; stamens 3-seriate:
peduncles ca. 1 mm. long.
coe 9. L. Ekman.

BB. Ovary and fruit 6-10-celled. (Porto Rico).. i ie cae oadi ten

AA. Terminal leat-buds glabrous. (Cuba).........5.2.. ay Wrightit.

el
—

FE ee alpestris (Krug & Urban) Dyer in Index Kew. “Suppl. De
1904. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 136. 1925. —
a C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg. 29: 16. 1931.

— Moscoso, Cat. Fl. Doming. 1: 377. 1943.
ae alpestris Krug & Urban in Bot. Jahrb. 21: 547. 1896. — Ur-

Ind. Occ. 8: 436. 1920.
Wikstroema tied (Krug & Urban) Blake 1 a free Gray Herb. n. s.
53: 38. 8. — Urban, Fl. Ind. Occ. 8:

Small tree : —3 m. high. Terminal leaf- ae gray-sericeous. Young
branchlets long-pilose, the older branchlets glabrescent, strigose. Leaves
chartaceous, elliptic to elliptic-obovate, 2.5-4 cm. long, 1-1.5 cm. wide,
obtuse, enh rounded at the apex, fone: -tapering at the base, glabrous,
dark green above, long-pilose and lighter green beneath, the margin quite
flat, occasionally subrevolute, finely denticulate especially toward the apex,
the midrib slightly canaliculate above, the veins 8-10 pairs, conspicuous
above, very outstanding below forming a sharp network over the entire
lower surface, the petiole 2-4 mm. long. Flowers few, solitary, in axils
near the apex of the branchlets; peduncles very brief, ca. 2 mm. long;
bracteoles quickly caducous; sepals ca. 5, broadly ovate to subrotund, the
larger ones 5-6 mm. long, pilose on the external surface; petals 5, obovate,
ca. 15 mm. long, 8-10 mm. wide, unequal, deeply emarginate at the apex,
pubescent on the median portion of the external surface; stamens bi-
seriate, the filaments glabrous, 4-5 mm. long, the anthers subglobose, ca.

172 JOURNAL OF THE ARNOLD ARBORETUM [yous Bax

1 mm. long; ovary ovate, white-pilose, 4—5-celled, with ca. 3 ovules in
each cell, the styles 4-5, horizontal, ca. 1 mm. long, glabrous except at the
extreme base, the stigmas considerably dilated, semiorbicular. Capsule
oblong, ca. 10 mm. long, 4-5 mm. diam., very short-appressed-pubescent,
4—5-celled, 2-3 seeds in each cell; seeds with wing ca. 7-8 mm. long, 2 mm.
wide.

Haitr1: Morne des Commissaires, Gros Cheval, alt. 1500 m., L. R.
Holdridge 1260 (US), June 9, 1942 (tree 10 m. tall; petals white, ca. 2 cm.
long, deeply emarginate; old capsules persistent). — Morne des Commis-
saires, alt. 5800 ft., J. T. Curtis sn. (US), July 24, 1944.— Massif de la
Pelle, Marigot, Jardins Bois-Pin, alt. 1800-1900 m., E. L. Ekman H1622
(US), Aug. 24, 1924.

Laplacea alpestris was the first species of the genus described from the
republics of Haiti and Santo Domingo. Since then two more species, L.
Samuelssonii Schmidt and L. cymatoneura Urban, have been described.
Upon studying the genus as a whole one is struck by the close relationship
of these three species, and the study of further collections may show L.
cymatoneura to be, at the most, only a variety of L. alpestris.

Distinctive characters of this species are: (1) the horizontal styles and
broadened stigmas; (2) the long-pilose hairs on the under surface of the
mature leaves; (3) the petals pubescent on the dorsal surface; and (4) the
fine, extensive and pronounced reticulation on the under surface of the
leaves.

Laplacea Samuelssonii can be separated from this species by the styles
which are three in number and erect and by the stigmas which are some-
what capitate but not broadened. Also the pubescence when present on
the under surface of the leaves is appressed, rather than loose-pilose and
turns dark in color.

Laplacea cymatoneura possesses the five horizontal styles and broadened
stigmas and the pronounced reticulation on the under surface of the leaf.
However, this reticulation is confined more to the outer half of the leaf.
The under surface of the leaf is quite glabrous, as is the dorsal surface of
the petals.

Krug and Urban, in their original description of L. alpestris and Urban
in the description of L. cymatoneura describe in detail staminate and pis-
tillate flowers. In all the flowers I examined or dissected there was no
instance of sexual distinction in the individual flowers.

2. Laplacea cymatoneura Urban in Fedde, Rep. Spec. Nov. Reg. Veg.
20: 34. 1924. —O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg.
29: 16. 1931; 33: 176. 1933. — Moscoso, Cat. Fl. Doming. 1: 377.
1943.

Large forest tree. Terminal leaf-bud appressed-pubescent, the branch-
lets of the current year’s growth appressed-pubescent, glabrous when ma-
ture, erect, brown, terete, lightly striate. Leaves coriaceous, obovate, 4—5
cm. long, 2—2.5 cm. wide, obtuse or rounded at the apex, emarginate, cune-
ate at the base into a petiole 4-5 mm. long, dark green, shining, glabrous

1949 ] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 173

above, lighter green, glabrescent below, pilose especially along the midrib
and margin when young, the margin subrevolute appearing crenulate on
the upper surface “but distinctly denticulate on the under surface, the
midrib deeply canaliculate on the upper surface, broad and saan be-
neath, the veins 8-10 pairs, obvious above, very distinct below, nearly
horizontal with the midrib, anastomosing near the margin forming a sharply
defined network of lesser veins. Flower solitary, axillary near the ends of
the branchlets; peduncles ca. 8 mm. long, appressed-pubescent; bracteoles
quickly ethiaane: the scars prominent on the peduncle; sepals unequal,
subrotund, the outer sepals 5-6 mm. long, 6-7 mm. wide, the inner ones
larger, up to 10 mm. long, 9 mm. wide, appressed-pubescent on the dorsal
surface; petals obovate, glabrous, deeply emarginate, up to 23 mm. long,
13 mm. wide; stamens 2-seriate, the filaments 2.5-3 mm. long, glabrous,
the anthers minute, less than 0.5 mm. long, globular; ovary ovoid, seri-
ceous, 5-celled, the cells 3-ovulate, the styles 5, horizontal, glabrous, ca.
0.5 mm. long, the stigmas expanding considerably beyond the styles.
Capsules narrow-elliptic, ca. 1.5 cm. long, the seeds ca. 3 mm. long with
the es ca. 7 mm. long.

E NICAN Repusiic: Prov. Pacificador, vicinity of San Francisco de
eee is, Quita Espuela, alt. 400-1000 m., W. L. Abbot 2093 (US, IsoTyPE),
April 5-17, 1922 (large forest tree). Hawi Massif du Nord, Port-de-
Paix, Haut-Piton, on laterite, alt. ca. 800 m., FE. L. Ekman H3688 (US),
April 6, 1925.

This species was described by Urban (1924) from very fragmentary ma-
terial and is consequently incomplete. He stated that he had seen only
two leaves. Knowing the variations that are frequently found in this
family, it is surprising that he separated the species from L. alpestris (to
which it is obviously closely related) on the basis of such scanty material.
Abbot’s isotype in the U. S. National Museum is very similar to the ma-
terial seen by Urban. It consists merely of a packet containing two
leaves and a few petals, and on the sheet a piece of wood which evidently
came from a very large tree, as is stated on the label.

Later, in 1933, O. C. Schmidt amplified the description by including the
fruit and by citing two more specimens, Hkman H3688 and H4635.
Ekman H3688 is represented in the U. S. National Herbarium. Before
seeing Schmidt’s work I had difficulty in identifying Ekman H3688,
finally deciding to place it with L. cymatoneura, even though the petals
measured only 11 or 12 mm. in length and the veins rose less horizontally
from the midrib. In the two leaves of the type the veins are strikingly
horizontal. In Ekman H3688, although the majority of the veins appear
to rise subhorizontally to the midrib, many leaves show veins which leave
the midrib at an angle of 60° or even 45°. It is quite probable that Abbot’s
few leaves may have been collected from a tree which exhibited the same
variation in veining as Ekman H3688. 1 was pleased to find that Schmidt
reached the same conclusion as I regarding the Ekman specimen, although
in his added description he did not mention the variations recorded above.

174 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

In his mention of Hkman H4635, a specimen which I have not seen, he
described only the fruit.

This species can be separated from L. Samuelssoni by the horizontal
styles, which number five. In L. Samuelssonii the styles are three in num-
ber and erect, and the stigmas are capitate rather than broadened.

From L. alpestris the species can be separated at present by the larger
(wider) leaves and their glabrous under surface and by the glabrous dorsal
surface of the petals. The leaves of L. alpestris, even in maturity, are
distinctly long-pilose below, and narrower, with an over-all reticulation.
The petals are lightly pubescent on the dorsal surface. In most characters,
however, these two species agree. However, since small petals (11 or 12
mm.) have been found in Hkman H3688, it may eventually be necessary,
on the basis of additional material, to combine the two species, recognizing
this one as a variety of L. alpestris.

3. Laplacea Samuelssonii O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg.
Veg. 29: 16. 1931. — Moscoso, Cat. Fl. Doming. 1: 377. 1943.

Small tree. Terminal leaf-bud and young branchlets of the current
year densely appressed-pilose, the older branchlets terete, glabrous, brown-
gray. Leaves chartaceous, oblanceolate or narrowly obovate, 2-4 cm.
long, 1-1.5 cm. wide, narrowed at the apex, rounded, cceacionally emar-
ginate, cuneate at the base, elongate-tapering at the extreme base into a
very short petiole 1 mm. long, glabrous above, sparsely appressed, very
short pubescent below, darkening with age and appearing black in the
dried state, the margin subrevolute, finely denticulate, the veins 10-12
pairs, rather inconspicuous above, prominent below, rising nearly hori-
zontally from the midrib, anastomosing near the margin, forming a narrow
network of veins. Flowers solitary, axillary near the ends of the branch-
lets; peduncles short, 1-1.5 mm. long; bracteoles quickly caducous; sepals
unequal, suborbicular, 5-7 mm. long, 7-8 mm. wide, appressed-pubescent
on the dorsal surface; petals 5, white (Ekman), obovate, ca. 12 mm. long,
8-10 mm. wide near the apex,. emarginate, lightly appressed-pubescent in
the median portion of the external surface; stamens apparently tri-seriate,
the filaments glabrous, unequal, those of the outer row 3 mm. long and
those of the inner row ca. 4 mm. long, the anthers subglobose, ca. 0.8 mm.
long; ovary subglobose, densely, sericeous, 6-angled a eae the styles
3, thick, ca. 0.6 mm. long, erect, elabious. Fruit not know

Haiti: Massif des Matheux, l’Arcahaie, Morne Delpech, a ca. 4000
m., E. L. Ekman 9320 (US, tsotyPe), Nov. 14, 1927.

ie mounted sheets of Ekman 9320, the type of the species, are de-
posited in the U. S. National Herbarium. One specimen has leaves up
to 4 cm. long and 1.5 cm. wide, while on the second specimen all the leaves
are much smaller, none measuring more than 2 cm. long and 0.8 cm. wide.
In other characters there is ee agreement.

Distinctive characters are: (1) the sharp reticulate veining on the under
surface of the leaf; (2) the short scattered appressed pubescence (eventu-

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 175

ally darkening) on the under surface of the leaf; (3) the three erect, short
styles; and (4) the tri-seriate stamens.

The species is closely allied to L. alpestris (Krug & Urban) Dyer and
to L. cymatoneura Urban. It can be distinguished from L. alpestris by
the three erect, very short (0.6 mm.) styles. In L. alpestris the styles
number ca. 5 and lie horizontal to the ovary, and the stigmas are pro-
nouncedly dilated. The leaves are noticeably much longer pilose on the
under surface of the leaf, especially along the midrib.

4, peo angustifolia (Britton & ye O. C. Schmidt in Fedde,
p. Spec. Nov. Reg. Veg. 22: 94. 1925. — Marie-Victorin in Con-
a Inst. Bot. Univ. Montréal 49: 70, 75. 1944.
Haemocharts angustifolia Britton & Wilson in Bull. Torrey Bot. Club. 50:
43, 1923.

Tree 20-30 m. Terminal leaf-bud slender, appressed-pubescent. Young
branchlets slender, grayish brown, terete, appressed-pubescent, the older
branchlets glabrous. Leaves coriaceous or subcoriaceous, oblong-elliptic
to linear-lanceolate, 3-4 cm. long, 0.6—0.8 cm. wide (also 5.5-6 cm. long
and 1.3-1.4 cm. wide), acute at the apex, not sharply so, finely apiculate,
tapering gradually at the base into a petiole 1 mm. long, glabrous above,
the younger leaves covered underneath with a light spreading pubescence,
at maturity pubescent on the midrib only, the margin entire, subrevolute,
the veins ca. 12 pairs, lightly elevated on both surfaces, anastomosing
near the margin. Flowers solitary, axillary near the apices of the branch-
lets; peduncles very short, terete, 1 mm. long, appressed-pubescent; brac-
teoles 2, sepaloid, broadly ovate, 2-3 mm. long, appressed-pubescent on the
external surface; calyx-lobes 5, suborbicular, 4.5-6 mm. long, 4-6 mm.
wide, appressed-pubescent on the external surface; petals 5, obovate, 1.5—
1.7 cm. long, ca. 1.3 cm. wide, deeply emarginate at the apex, pubescent
on the median portion of the external surface; stamens bi-seriate, ca. 30,
the filaments quite equal, filiform, ca. 5 mm. long, glabrous, the anthers sub-
globular, 0.5 mm. or less long; ovary globose, densely white-villous, 5-
celled, the styles 5, distinct, glabrous, ca. 1 mm. long. Fruit not seen.

Cuspa: Oriente: Pico Turquino, Maestra Ridge, alt. 1300 m., Fr.
Leon 11072 (NY, TYPE), Tae 1922 (tree 20-30 m.).— Pico Turquino, slope
of First Peak, G. C. Bucher 77 (NY), Oct. 12-14, 1924.

The narrow oblong-elliptic leaves of this species are its most distinctive
feature. Closely resembling these leaves in appearance are those of the
narrower-leaved specimens of L. Wrightii Grisebach. The terminal leaf-
buds of the latter species are glabrous, whereas in this species they are
pubescent.

5. Laplacea Wrightii Grisebach in Mem. Amer. Acad. n. s. 8: 166 (PI.
Wright.). 1860. — Walpers, Repert. Bot. Syst. 7: 367. 1868. — Sau-
valle, Fl. Cuba. 10. 1873.— Melchior in Nat. Pflanzenfam. ed. 2,
21: 136. 1925. —O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg.
29: 17. 1931. — Marie-Victorin in Contrib. Inst. Bot. Univ. Montréal
49: 70, 75. 1944.

176 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

Haemocharis Wrighti (Grisebach) Gomez de la Meza, Dicc, Bot. Nom.
Vulg. Cub. Puerto Rig. 15, 1889; Ensayo, Farm, Cuba 26. 1889; in Anal.

Wikstroemia Wrightii (Grisebach) Blake in Contrib, Gray Herb. n. s. 53:
1, 1918.

Slender tree. Terminal buds small, glabrous on the external surface.
Branchlets glabrous (current year’s growth appressed-pubescent), terete.
Leaves coriaceous, obovate to obovate-elliptic, 5-9 cm. long, 2~3 cm. wide,
glabrous on both surfaces, obtuse or rounded at the apex, usually lightly
emarginate, tapering at the base into a short petiole 2-3 mm. long, the
margin entire or rarely minutely serrulate toward the apex, the veins 8-12
pairs, quite inconspicuous on both surfaces. Flowers solitary, axillary;
peduncles glabrous, 4-6 mm. long, terete; bracteoles caducous; sepals 5,
ovate to orbicular, ca. 10 mm. long, sericeous on the external surface,
appressed-pilose on the internal surface, shortly apiculate at the apex;
petals 5, obovate, white, 15-17 mm. long, ca. 10 mm. wide, appressed-
pubescent on the external surface; stamens ca. 50, the filaments bi-seriate,
the anthers ovate; ovary 5-celled, densely sericeous, oval, few-ovulate.
Capsule ovoid, obtuse at the apex, 2-3 cm. long, glabrescent, S-angled,
5-celled; seeds (including wing) 12-15 mm. long.

Cupa: Oriente: near Monte Verde, in forest, C. Wright 48 (1s0-
types, G, NY, Mo), Jan.-July 1859 and 1860-1864 (slender tree with white
flowers). — South of Sierra Moa, Camp La Gloria, J. A. Shafer 8195 (NY,
US), 8203 (NY, US), Dec. 24-30, 1910 (slender shrub 5 ft. high with
white flowers).

The distinguishing characters are: (1) glabrous terminal leaf-buds; (2)
coriaceous leaves; (3) large fruit and seeds; and (4) sepals pubescent on
both the external and internal surfaces. The glabrous terminal buds dis-
tinguish this species from all other Cuban species.

From the Sierra Moa region of Oriente are two specimens, J. A. Shafer
8195:and 8203, which probably belong here but possess neither fruit nor
mature flowers. The leaves are more narrow (sometimes only a single
centimeter wide) and resemble L. angustifolia. However, in the latter
species the terminal buds are distinctly pubescent. I do not believe that
this Shafer material should be given varietal status even though the leaf
difference seems distinctive. Too little material has been examined.

Six sheets of “Wright 48” have been examined in this study. It is very
obvious that this group of specimens does not comprise a single collection.
There are four different labels with this number, all from Oriente:

1. “prope villam Monte Verde dictam, Cuba Orientali, Jan.-Jul. 1859.”
This is printed on a blue label with a written number 48. A blue field
label accompanies the printed label and states that the habit is that of a
slender tree and that the flowers are white. The abbreviation “M. V. Dec.
8” on this field label agrees in locality (M. V. standing for Monte Verde)
but the date “Dec. 8” does not agree with the printed label. A single
specimen is found at the Gray Herbarium.

—

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVII 177

2. “1856-7, in Cuba Orientali.” Another single specimen with a blue
Jabel in the Gray Herbarium. Besides the number “48” is also found
the number “64” both on the label and in a packet at the top of the sheet.

3. “Sept. 1859-Jan. 1860.” This is on a white label with no further in-
formation. There are two sheets of this collection in the herbaria of the
New York Botanical Garden and the Missouri Botanical Garden.

4. “1860-64.” Ona white label, and found at Gray, Missouri, and New
York. The Gray specimen possesses an added field label which states that
the specimen was collected at Cachillas de Baracoa on May 14, and that
it was arborescent with white flowers. However, the Gray and Missouri
specimens possess very mature fruits with no flowers, while: the New York
specimen possesses a few unattached flower-buds.

Fortunately, since this “number 48” is the type of the species, these
specimens all truly belong to the same species. However, until the species
is re-collected we cannot be sure of the true type-locality.

6. Laplacea benitoensis (Britton & Wilson) O. C. Schmidt in Fedde,
Rep. Spec. Nov. Reg. Veg. 24: 78. 1927; 29: 16. 1931.— Marie-
Victorin in Contrib. Inst. Bot. Univ. Montréal 49: 70, 75. 1944.

Haemocharis benitoensis Britton & Wilson in Mem. Torrey Bot. Club 16:
82. 1920

Straggling shrub 1-3 m. high. Terminal leaf-bud small, ca. 5 mm. long,
villous. Branches slender, terete, the very young branchlets appressed-
pubescent. Leaves coriaceous, obovate, 3-5 cm. long, 1-1.6 cm. wide,
rounded at the apex, cuneate at the base, glabrous, light green above,
loosely appressed-pubescent beneath, glabrescent, the margin entire, revo-
lute, the veins few (ca. 8) in number, occasionally evident on both surfaces,
usually not conspicuous, the petiole slender, 5-7 mm. long, pubescent in
the younger leaves. Flowers solitary, axillary near the tips of the’ branch-
lets; bracteoles quickly caducous; sepals 5, small, suborbicular, 3-4 mm.
long and wide, appressed-pubescent on the external surface; petals (fide
Britton & Wilson) 5, white, elliptic-obovate, 10 mm. long, 5-6 mm. wide;
stamens 2- or 3-seriate, glabrous, the filaments unequal, ca. 2 mm. long,
the anthers globular, minute, ca. 0.25 mm. long; ovary subglobose, covered
with a dense white-silver pubescence, glabrous at the apex and tapering
broadly into an entire glabrous style which is topped by several stigmas.
Fruit not seen

CuBa: Oriente: vicinity of Camp San Benito, west of camp, in
thicket, alt. 900 m., J. A. Shafer 4063 (NY, Type), Feb. 24, 1910 (Stragaling
shrub 4 ft. high; flowers white).— South of Sierra Moa, Camp La Gloria,
J. A. Shafer 8208 (Ch, NY, US), 8272 (G, NY), Dec. 24-30, 1910 (shrub
6-10 ft.).

A few of the outstanding characteristics of this species are: (1) the small
calyx-lobes and corolla-lobes; (2) the small pubescent terminal leaf-buds;
(3) the very short peduncles; (4) the entire revolute margins of the small
thick leaves; (5) the very small stamens; and (6) the entire style.

178 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Although six specimens were available for this study, none offered mature
flowers for dissection. Fortunately, the type specimen (J. A. Shafer
4063), deposited in the New York Botanical Garden, possessed remnants
of a partially dissected flower, and it was from this and the diagnosis of the
original authors that the above description was drawn up.

7. Laplacea Urbani O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg.
22: 93. 1925; 29: 17. 1931. — Marie-Victorin in Contrib. Inst. Bot.
Univ. Montréal 49: 70, 75. 1944.

Small tree. Terminal leaf-buds elongate, conspicuously gray-appressed-
pubescent. Branches brown-gray, terete, glabrous, conspicuously marked
by leaf-scars, the very young branchlets short-appressed-pubescent, quickly
glabrous. Leaves disposed at the ends of the branchlets, obovate, 6—10
cm. long, 2.5-4 cm. wide, rounded and deeply emarginate at the apex,
tapering at the base, glabrous above, appressed-pubescent below (on the
midribs only in mature leaves), the margin crenate, nearly undulate, the
veins prominent on both surfaces, ca. 15 pairs, branching and anastomosing
near the margin appearing reticulate, the upper pair arching back and
terminating near the deepest point of emargination, the actual petiole
short, 3-4 mm. long, pubescent below. Flowers solitary, axillary, the
peduncle 10-12 mm. long, terete, pubescent with a number of scars along
the side, the scars probably those of the caducous bracteoles; sepals 5,
obovate to rounded, sericeous-pubescent on the external surface, unequal,
the outer ones 7-8 mm. long and ca. 6 mm. wide, the inner ones 9-12 mm.
long, 6-8 mm. wide; petals 5 (fide Urban), white, obovate, 2.3-2.5 cm.
long, 1.5—1.7 cm. wide, deeply emarginate at the apex, pubescent on the
dorsal surface; stamens many, 3(or more)-seriate, the filaments glabrous,
somewhat unequal, 7-10 mm. long, the anthers oblong-elliptic ca. 1 mm.
long; ovary ovoid, sericeous-pubescent, 2.5-3 mm. long, 5-celled; styles
5, 1-1.5.mm. long, glabrous; stigmas 5. Fruit not seen.

Cupa: Oriente, Sierra Maestra, Pinar de Caridad, southeast of Yara,
pine patch at edge of brook, E. L. Ekman 14687 (NY, isotyre), July 31,
1922.

This species is characterized by large leaves, 6-10 cm. long, 2.5—4 cm.
wide, rounded and deeply emarginate at the apex. The margin is very
outstanding for the genus in that it is softly crenulate, without any evi-
dence of glands. The veining also is distinctive. Conspicuous on both
surfaces, the veins join about half-way to the margin, becoming reticulate.
The upper pair arch upward, then back again toward the midrib, and
appear to join near the point of emargination, forming a nearly heart-
shaped figure. The complete arc is not always obvious or truly formed.

The corolla is deeply emarginate. The stamens are more numerous than
those found in most species and appear to be 3- or 4-seriate. Along the
peduncle were found scars caused by caducous bracteoles, farther spaced,
however, than is usual, scattered over the entire length of the peduncle.

Several characters are mentioned above with an element of uncertainty.
Only a single, partially dissected flower was attached to the specimen. No

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 179

attempt at boiling up the flower or dissections was made. The description
was drawn up from that of the original author and supplemented by ob-
servations on the type whenever possible.

7a. Laplacea Urbani O. C. Schmidt var. subserrulata, var. nov.

A typo recedit foliis margine subserrulatis vel leviter crenulatis et apice
minus profunde emarginatis vel integris.

CusdA: Oriente: Firmesa to Gran Piedra, J. A. Shafer 8974 (NY,
iS) Mar. 4-5, 1911 (shrub 8 ft.).— Range of Sierra Maestra, Loma del
Gatowalt, 900-1000 m., 27, eon, Clement & M. Roca 95598" ( NY, tvPr),
10015 (NY), July 11-Aug. 14, 1921 (small tree 5-6 m.).— Cobre range of
Sierra Maestra, Loma del Gato, Fr. Leon, Clement & M. Roca 10125 (NY),
July 11-Aug. 14, 1921 (shrub 4-5 m.).

The material cited above, like most of the earlier collections, has long
been identified with Laplacea Curtyana, a species of western Cuba. Al-
though it appears to resemble L. Curtyana in many respects, I feel that its
true relationship is with one of the Oriente species, namely, L. Urbani.
Any one of three species, L. Urbani, L. Ekmani, and L. moaensis, might
well be the type from which this variety has derived. However, in L.
moaensis the leaves have a tendency toward obliqueness and in L. Ekmani
the leaves are coriaceous, both characters strong for any species of this
group but lacking in this variety.

From typical L. Urbani the variety differs in the apex, margin, and
venation of the leaf. Otherwise the two entities are very similar. At the
apex of the leaf in the species a deep and very marked emargination is
found, the margin is decidedly crenate, and the top pair of veins turn back
to the midrib to form a heart-shaped figure. In the variety the apex of
the leaf may be slightly emarginate or entire, the margin may be lightly
crenulate, but usually is subserrulate, and the upper pair of veins do not
turn back toward the midrib.

This variety differs nearly as much from the species as do L. Urbani,
L. Ekmani, and L. moaensis from one another. Only four specimens were
available for study of the three species listed above: two for L. Ekmani
and one each for L. Urbani and L. moaensis. As many specimens were
available for this variety alone.

8. Laplacea moaensis Marie-Victorin in Contrib. Inst. Bot. Univ. Mont-
réal 49: 72. 1944.

Tree 3-5 m. high. Terminal leaf-bud with a fulvous or silvery appressed
pubescence. Branchlets terete, reddish brown, pubescent when young,
soon becoming glabrous. Leaves disposed at the ends of the branchlets,
chartaceous, obovate or oblanceolate, 6—12 cm. long, 2—3.5 cm. wide, obtuse
and slightly emarginate at the apex, narrowed at the base, glabrous above,
fulvous-pubescent beneath, especially on the lower portion of the prominent
midrib, the margin subrevolute, lightly denticulate-crenulate along one side
especially toward the apex, entire along the other side, the veins (20-30
pairs) inconspicuous, the petiole very short, ca. 2mm. long. Flowers soli-
tary, axillary; peduncle short, ca. 6 mm. long, with bracteole scars dis-

180 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

posed along the sides, appressed-pubescent; bracteoles quickly caducous
(not seen); sepals subrotund, unequal, 8-10 mm. long, nearly as wide,
appressed-pubescent over the entire external surface; petals (fide Marie-
Victorin) white within, slightly purple without, 8-10 mm. long; stamens
with the filaments (ca. 10 mm. long) gradually attenuated from the base
to the apex; ovary subspherical, 3-4 mm. long, white-sericeous-pubescent,
4—5-celled. Capsule oblong-elliptic, 2.5-3 cm. long, 4—5-celled; seeds
dark-colored, glabrous, ca. 6 mm. long, with wing about 18 mm. long.

Cupa: Oriente: “Region de Moa, chemin des hauteurs a l’ouest du
rio Cayoguan, sur la limonite argileuse recouvrant la serpentine,” Maric-
Victorin, Clément & Alain 21630 (G, 1sotyre), April 16-23, 1943.

Laplacea moaensis, the most recent species of the genus, has been amply
described and illustrated by Marie-Victorin in Contrib. Inst. Bot. Univ.
Montréal 49: 69-75, figs 1 and 2, 1944. Although an isotype of the species
was available for this study, no mature flowers or fruit were to be had for
dissection; hence much of the above description was drawn from that of
Marie-Victorin.

Coming from the serpentine region of Moa in the state of Oriente, like
many other novelties from the same area it appears quite distinct. The
peduncles are short (ca. 6 mm. long) and the bracteole-scars are found
only along the upper half of the peduncle. The veins, although quite in-
conspicuous, are more numerous (20-30 pairs) than those in most other
species of the genus. The margin is subrevolute. Along one side of the
leaf the margin is entire, while along the other half a light crenulation can
be found toward the apex. This margin condition, although not especially
rare in the family, is distinctive for the genus.

A fine reddish dust from the serpentine region overlies most of the speci-
men, giving a brownish red color to the branchlets, midrib of the leaf, etc.,
and may be mistaken for a specific character. This dust can easily be
wiped off with a cloth.

9. Laplacea Ekmani O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg.
22: 94. 1925; 29: 17. 1931. — Marie-Victorin in Contrib. Inst. Bot.
Univ. Montréal 49: 70; 75. 1944.

Small tree. Terminal leaf-bud silver-white sericeous-pubescent, ca. 6
mm. long. Young branchlets appressed-pubescent, glabrous when older.
Leaves coriaceous, obovate, 4-8 cm. long, 2.3-3.5(—4.5) cm. wide, rounded
at the apex, tapering at the base, glabrous. above, appressed-pubescent on
the midrib below, also occasionally under the revolute margin, inconspicu-
ously denticulate with slight evidences of glandular teeth (on the upper
denticulations), the veins inconspicuous on both surfaces, ca. 12 pairs, the
petiole short, 4-5 mm. long. Flowers axillary, solitary; peduncle ca. 1
cm. long, terete, appressed-pubescent; bracteoles several, disposed along
the peduncle, quickly caducous, broadly ovate to rounded, the outer ones
ca. 7 mm. long, sericeous-pubescent on the exterior surface, gradating in
size into the sepals; sepals ovate, ca. 10 mm. long, 6-7 mm. wide, sericeous-
pubescent on the exterior surface; petals (bud) pubescent on the exterior

1949 ] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 181

surface; stamens (bud) 3-seriate, the filaments glabrous, the anthers sub-
globular; ovary (bud) subconical, densely silvery sericeous-pubescent, 5
(or 6)-celled, tapering at the apex into 5 (or 6) styles. Capsule oblong,
2.5—2.8 cm. long, glabrescent, 5 (or 6)-celled, each locule with ca. 5 seeds;
seeds ca. 5 mm. long with the wing 10-12 mm. long.

A: Qriente: Sierra de Nipa, Woodfred, in forest bordering on
pinelands, on left side of Arroya del Medio, E. L. Ekman 15283 (NY, 1so-
TYPE), Sept. 25, 1922 (small tree). — Palmarito de Cauto, A. Cuebelo 6230
(NY), June 2, 1932 (high tree).

This species was described from a fruiting specimen (Ekman 15283),
a duplicate of which was available for this study. A second specimen
(Cuebelo 6230), collected in the same general region as the type, possessed
buds only, and these were very young. Dissections of the buds furnished
only general information, but added somewhat to the original diagnosis of
Schmidt.

The species appears to be most closely allied to L. Urbani Schmidt,
differing mostly in the leaves, which are definitely coriaceous, not charta-
ceous as originally described. They are rounded at the apex and only
rarely even slightly emarginate, not deeply so as is the case in L. Urbant.
Also the leaf-margin is revolute, with hairs protruding from beneath the
revolute portion of the margin. A very slight denticulation occurs along
this margin. In L. Urbani the leaves are more nearly chartaceous, the
margin distinctively crenate, and the pubescence on the mature leaves is
confined to the midrib of the lower surface.

10. Laplacea Curtyana A. Richard, Ess. Fl. Cuba 1: (in Sagra, Hist. Ie
Cuba 2:) 225. 1845.— Walpers, Repert. Bot. Syst. 5: 132. 1846. —
Sauvalle, Fl. Cuba. 10. 1873. — Melchior in Nat. Pflanzenfam. ed. 2,
21: 136. 1925. —O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg.
29: 17. 1931. — Marie-Victorin in Contrib. Inst. Bot. Univ. Mont-
réal 49: 75. 1944.

Haemocharis Curtyana (A. Richard) Maza, Dicc. Bot. Nom. Vulg. Cub.
Puerto-Riq. 15. 1889; Ens. Farm. Cuba 26 1889; in Anal. Hist. Nat.
Madrid 19: 222. 1890; in Estac. Exp. Agron. Cuba Bol Chlba@uba)e22:
74. 1914. — Kuntze Rev. Gen. Pl. 1: 62, 1891, as H. Courtyana. — Urban
in Bot. Jahrb. 21: 546, 1896. — Millspaugh in Field Columb. Mus. ,Bot.
1: 430. 1900.

Laplacea Courtyana Kuntze, Rev. Gen. Pl. 1: 62. 1891.

cela Curtyana (A. Richard) ee in Contrib. Gray Herb. n. s.
53:

Tree up to 15 m. high. Terminal leaf-buds elongate, 1-1.5 cm. long,
densely pubescent. Branches terete, pubescent when very young, later
glabrous. Leaves coriaceous or subcoriaceous, elliptic to obovate, 6-8 cm
long, 2.3-3 cm. wide (occasionally up to 103.8 cm.), obtuse or subob-
tuse at the apex, tapering toward the base, glabrous and shining above
(occasionally pubescent at base of midrib), pubescent below, glabrescent,
the margin entire or subentire, the veins 12-15 pairs, curving upward near

182 JOURNAL OF THE ARNOLD ARBORETUM [VoL. xxx

the margin, the petiole short, 4-7 mm. long, pubescent. Flowers solitary;
peduncle thick, terete, ca. 2 mm. long, pubescent; bracteoles 2, quickly
caducous, when present unequal, the outer one smaller, broadly ovate, ca.
4 mm. long, 3.5 mm. wide, the inner one subrotund, ca. 5 mm. long and
wide, both appressed-pubescent on the external surface; calyx-lobes 5,
imbricate, very unequal in size and shape, outer lobes subrotund 6—8 mm.
long, 6-7 mm. wide, the inner lobes increasing in both length and width,
the innermost lobe ‘broadly obovate to subrotund, ca. 18 mm. long, 10
mm. wide, with a wide membranaceous margin, very slightly emarginate,
all appressed: pubescent on the dorsal surface; petals 5, unequal, the outer
one obovate, the smallest 1.5 cm. long and 1.3 cm. wide, only slightly
emarginate at the apex, the inner four more nearly equal, obovate to ob-
cordate, ca. 2.2 cm. long and 1.5 cm. wide, deeply emarginate at the apex,
all lightly pubescent on the external surface at least at point of emargina-
tion; stamens bi-seriate, ca. 40, the filaments of about equal length, free,
glabrous, 5-6 mm. long, the anthers subglobose, ca. 1.5 mm. long; ovary
semi-globose, densely white pubescent, 5-ribbed, 5-celled, tapering abruptly
at the apex into fine short glabrous styles with 5 stigmas. Fruit woody
ovoid, 5-ribbed, ca. 2 cm. long, 1 cm. wide, glabrous except within the
ridges, the seeds with wings ca. 1.5 cm. long, the wing 9-10 mm. long.

Cupa: Isle of Pines: vicinity of Los Indios, arroya, N. L. Brit-
ton, E. G. Britton & P. Wilson 14247 (Ch, G, Mo, NY, US), Feb. 13, 1916
(tree 15 m., flowers white).—No definite locality, José Blain 22 (Ch).
‘“‘Western Cuba’’: “prope Vuelta de Abajo,’ C. Wright 2109 (G,
Mo, NY, US), April 2, 1865.

Laplacea Curtyana is the first species described from Cuba and as such
has been much confused and recorded over a range far too extensive. Ac-
cording to this study it is confined to the western part of Cuba and the
Isle of Pines.

The species does not possess what might be termed truly outstanding
characteristics. Some of the better characters to help in determination
are: (1) the elongated, narrow, densely pubescent terminal leaf-bud; (2)
the pubescence in waves (parallel to the margin) of density on the lower
surface of the leaf; and (3) the entire or nearly entire margin of the leaf.

The specimens of L. Curtyana offered a good opportunity to observe the
gradual change in form from the outer bracteole to the inner corolla-lobe.
The inner calyx-lobe resembles very much the outer corolla-lobe. All the
corolla-lobes except the outer one are deeply emarginate. The outer lobe
is only slightly so.

Of the specimens cited above no locality was given on the sheet, Wright
2109. The quotation “prope Vuelta de Abajo” given in the citation above
was taken from Urban’s publication.

11. Laplacea villosa (Macfadyen) Grisebach, Fl. Brit. W. Ind. 104.
— Walpers, Repert. Bot. Syst. 7: 367. 1868. — Fawcett, Fl. Pl.
Jamaica 3. 1893.— Melchior in Nat. Pflanzenfam. ed. 2, 21: 136.

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 183

1925. — Fawcett & Rendle, Fl. Jamaica 5(3): 188. 1926. — Schmidt
in Fedde, Rep. Spec. Nov. Reg. Veg. 29: 16. 1931.
Gordonia villosa Macfadyen, Fl]. Jamaica 1: 117. 1837. Walpers, Repert.
Bot. Syst. 1: 375. 1842.
Haemocharis villosa (Macfadyen) eee s Mém. Soc. Phys. Hist. Nat.
Genéve 1: 144 (Mém. Ternstr. 56). —O. Kuntze, Rev. Gen. Pl]
1: 62. 1891. — Urban in Bot. Jahrb. ne oe 1896.
Laplacea viscosa Hooker & Jackson, Index Kew. 2: 30. 1894, lapsu.
Wikstroenia Macfadyenti Blake in Contrib. Gray Herb. n. s. 53: 40. 1918.
Shrub 1.5-3 m. Terminal bud tawny-villous, the young branchlets
terete, villous. Leaves coriaceous, pilose beneath, especially along the
midrib, obovate, 7-10 cm. long, 3.5—5 cm. wide, obtuse or rounded at the
apex, occasionally very short and bluntly acuminate, broadly cuneate at
the base, appearing subauriculate, the margin revolute and crenulate, the
midrib canaliculate above, raised below, widening toward the base, the
veins 8-10 pairs, rather unobtrusive, the petiole thick, 1-3 mm. long,
densely pilose on the under surface. Flowers solitary in the upper axils
of the leaves; peduncles stout, 5-8 mm. long, pilose; bracteoles quickly
caducous; sepals 5, imbricate, deeply concave, rounded, ca. 1.5 cm. long,
1.0-1.7 cm. wide, villous-sericeous, deciduous; petals 5*, obovate, 17—20
mm. long, 10-12 mm. wide, emarginate, pubescent on the external surface;
stamens very numerous, pluri-seriate, the filaments unequal, up. to 9 mm.
long, the anthers oval, 2-2.5 mm. long; ovary villous-sericeous, 5-celled,
the loculi 6—7-ovulate, the styles 5, glabrous, ca. 1 mm. or less long. Cap-
sule subligneous, oblong, 2—2.5 cm. long, 1 cm. diam., obtuse, 5-angled at
the apex, the seeds ‘5 in each cell.”
Jamaica: Newhouse Gap, J. Hart 987 (NY, US), 1886.— Locality in-
definite, J. Hart 351 (US).— Locality indefinite, Macfadyen s. n. (G)
This second species from Jamaica, although quite distinct from L.
haematoxylon (Swartz) G. Don, has been less collected and is not known
as well. A comparison of the two species can be found under the latter.
The size of the petals is probably recorded for the first time above. The
dissection from boiled material shows them to be considerably smaller than
those of L. haematoxylon and much less showy.

12. Laplacea haematoxylon (Swartz) G. Don, Gen. Syst. 1: 569. 1840.
—Grisebach, Fl. Brit. W. Indies, 104. 1859.— Fawcett, Fl. Pl.
Jamaica 3. 1893. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 136.
1925. — Fawcett & Rendle, Fl. Jamaica 5(3): 188, fig. 72, 1926. —
Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg. 29: 17. 1931.

Gordonia haematoxylon Swartz, Fl. Ind. Occ. 2: 1199, 1800. — Lunan,
el: . 1814. — DeCandolle, Prodr. 1: 528. 1824. — Mac-
fadyen, Fl. Jamaica 116. 1837.
decree haematoxylon (Swartz) Choisy in Mém. Soc. Phys. Hist.
Nat. Genéve 1: 144 (Mém. Ternstr. 56). 1855.— O. Kuntze, Rev. Gen.
Pied: 62. 1801. cry ene in Nat. Pflanzenfam. III. 6: 185. 1893.
— Urban in Bot. Jahrb. 21: 546, 1896.

184, JOURNAL OF THE ARNOLD ARBORETUM [VoOL, XXX
Wikstroemia haematoxylon (Swartz) Blake in Contrib. Gray Herb. n, s.
1918.

Tree 10-15 m. high. Terminal buds hirsute. Young branchlets
pubescent. Leaves membranaceous to subcoriaceous, ovate to elliptic,
5-10 cm. long, 2.5—4 cm. wide, obtuse or obtusely acuminate at the apex,
narrowed at the base into a petiole 4-8 mm. long, the margin serrate or
crenate-serrate, especially along the upper half, generally glabrescent with
inconspicuous hairs along the under side of the petiole and the midrib, the
veins rather obscure, 10-12 pairs. Flowers large, solitary in the upper
axils, white, ca. 4-6 cm. in diameter when open, ca. 0.6 cm. in diam. in
bud; peduncles stout, 2-5 mm. long, puberulent; bracteoles apparently 2,
quickly caducous, sepaloid; sepals 5, imbricate, rounded, concave, lightly
puberulous on the outer surface, unequal, varying in size from bracteoles
to petals, ca. 1 cm. long, 1.2 cm. wide, the margin scarious; petals 5-8,
white, obovate, unequal, 2-3 cm. long, varying in width 1-2 cm. on the
same flower, the wider petals deeply emarginate, the narrow petals often
entire at the apex, subglabrous; stamens aur 4- or 5-seriate, the
filaments glabrous, unequal, the outer row ca. 5 mm. long, connate for
various short distances at the base, very re joined the entire length,
the inner filaments ca. 9 mm. long, not joined, the anthers subrotund, ca.
1 mm. across; ovary globose, hirsute, 5-celled, few-ovulate, styles 5, gla-
brous, very short, 1 mm. or less long, channeled, spreading at the apex
into the stigmas. Capsule woody, oblong, subpentagonal, pubescent, ca. 2
cm. long, | cm. or less in diameter, the seeds ca. 3 in each cell, the wing
and seed 8-12 mm. long.

JAMAICA: eastern slope of John Crow Mts., woodlands, alt. 520 m.,
N. L. Britton 4162 (NY), Mar. 9-11, 1909 (tree 12m. high; petals white).

— Vinegar Hill, alt. 1100 m., W. Hows 2493 (Ch, NY, US}, Nov. Z1
1894 (tree 15 ft. high). — Ne ar Woodcutters Gap, alt. 1450 m., W. Harris
6736 (A, Ch, US), Nov. 6, 1896 (tree 5 m. high). — Hardware Gap, alt.
1350 m., W. Harris 10124 (Ch, NY, US), Feb. 19, 1908 (tree 50 ft. high;
flowers white). — Tom’s River Wood, Upper Clarendon, alt. 800 m., WW’.
Harris 10852 (Ch, NY, US), Mar. 1, 1910 ) (tree 40 ft. high; flowers w hite).
— John Crow Mts., alt. 600 m., WV. Fiareis & NL, Petion 10763 (Ch, NY,
US), Mar. 10, 1909 (tree 36 ft. high; corolla pure white).— Blue Mts.,
Marces Gap, alt. 1650 m., J. FR. Perkins 1469 (G), Mar. 31, 1916. — Blue
Mts. near — Gap, alt. 1600 m., A. Rehder s.n. (A), Feb. 10, 1903. —
Marces Gap, sea - n. (NY), Feb. 7, 1906. — Indefinite locality, J.
Hart 609 ion US).

Described under es as in 1800, page is the oldest species in the genus.
It was referred to Laplacea by G. Don in

Some of the salient characters of the aot are: (1) large white flowers,
4—6 cm. across, the petals glabrous or nearly so, the wider petals emar-
ginate at the apex and the narrower petals entire, and (2) the leaves ovate
or elliptic, 5-10 cm. long, quite glabrous beneath with occasional scattered
pubescence along the lower midrib.

1949] KOBUSKI, STUDIES IN THE THEACEAE, XVIII 185

A closely related species, also from Jamaica, is LZ. villosa. The latter
species is characterized by a villous pubescence on most of the parts
(branchlets, lower surface of the leaves, pedicel, calyx-lobes, and external
surface of the corolla). The leaves are obovate, rounded or bluntly acu-
minate at the apex, and broadly cuneate at the base, appearing sub-
auriculate.

13. Laplacea portoricensis (Krug & Urban) Dyer in Index Kew. Suppl.
2: 86. 1904. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 136. 1925. —
O. C. Schmidt in Fedde, Rep. Spec. Nov. Reg. Veg. 29: 17. 1931.
Haemocharis aap a & Urban in Bot. Jahrb. 21: 548, 1896. —
Urban, Fl. Ind. Oce. 4: 411. 1910. — Britton & Wilson, ae as Porto
Rico & Virgin Isl. 5: 582. 1924.— L. R. Holdridge in U. Sobor
Ser, Occulape @lrees bucito. Rico 2) 2: SG0sig. 1942.
Wikstroemia portoricensis (Krug & Urban) Blake in Contrib, Grav Herb.
n. s. 53: 40, 1918.

Tree 5-20 m. high with gray, fissured bark. Terminal leaf-bud ap-
pressed-pubescent, elongate. Younger branchlets finely appressed-pubes-
cent, the older branchlets becoming glabrous. Leaves chartaceous or
subchartaceous, elliptic to obovate, 5-12 cm. long, 2.5—4.5 cm. wide, obtuse
to rounded at the apex, occasionally slightly emarginate, long-tapering at
the base, glabrous above, glabrescent (rarely appressed-pubescent) below,
the margin crenulate along the upper half, entire along the lower portion,
the midrib canaliculate above, the veins 15-20 pairs, rather inconspicuous
above, prominent below, anastomosing near the margin and becoming
reticulate, the petiole 2-5 mm. long. Flowers solitary near the apex, axil-
lary; peduncle short, 1-4 mm. long, appressed-pubescent; bracteoles
quickly caducous; sepals unequal, the larger or inner sepals suborbicular,
10-12 mm. long, sericeous on the external surface; petals 6-9, white,
obovate, unequal, 18-22 mm. long, 15-18 mm. wide, some deeply emar-
ginate at the apex, others rounded, glabrous on the external surface, the
outer petals thickened in the medial area; stamens 3—4-seriate, many
(over 100), the filaments glabrous, unequal, 5-7 mm. long, the anthers
ovate to subglobose, ca. 1 mm. long; ovary globose, white-tomentose,
6-10-celled, tapering at the apex into the 5 or more glabrous styles, the
stigmas recurved, reniform. Capsule woody, narrow-ovate, 15-25 mm.
long, ca. 10 mm. diam., short-appressed-pubescent to glabrescent, 6—10-
celled; seeds 4 or 5 in each cell, 10-13 mm. long with wings.

Porto Rico: El Yunque, F. H. Sargent 541 (US), July 4, 1938 (tree).
— Catalina-Yunque Trail, Luquillo Mts., in forest, alt. 600-850 m., N. L.
Britton & E. M. Bruner 7579 (NY, US), 7604 (NY), Feb. 23-26, 1923
(tree 10 m. ce iad white, fugacious).— Jajoma Alta, WV. E. Hess 5591
CNS): ; — Sierra de Naguabo, Loma Icaco, edge of woods, alt.
210-675 m., of A. ee 3426 (NY, US), July 24, 1914 (tree 30 ft. high:
corolla mite). — Sierra de Naguabo, Rio Icaco and adjacent hills, forest,
alt. 465-720 m., J. A. Shafer 3517 (G, Mo, NY, US), July 30-Aug. 5, 1914
(tree 25 ft. high). — Sierra de Naguabo, Barrio de Maizales, mountain for-
estealt-900 m., Ni... Britton Gi. &. Hess 2277 (Ch, G, NYoUS) March

186 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xxx

9, 1914 (tree 15 m. high; flowers white, 4 cm. wide). — Sierra de Luquillo,
Jimenez, in mountain forest, P. Sintenis 1326 (Ch, G, Mo, NY, US), May
26, 1885. — Sierra de Naguabo, in primary forest, P. Sintenis 5318 (Ch, G
Mo, NY, US), Nov. 5, 1886. — Coco Valley, Maricao Verde, L. E. Gregory
or (NY), Aug, 2, 1940 a 14 ft. high).— Indefinite locality, A. A.
Heller s. n. (Ch, NY),

Of the above cited specimens, Sintenis 1326 and 5318 and Heller s. n.
were cited by Krug & Urban in the original publication of this species.
No specimen was designated as the type. Krug & Urban carefully dis-
tinguished between pistillate and staminate flowers. I sought both types
but was unsuccessful in finding any that could be designated as either pis-
tillate or staminate. The flowers dissected for this study possessed not
only developed ovary, style, and stigma, but also anthers (over 100)
which had produced pollen. Krug & Urban state that in the staminate
flowers the stamens were 3—4-seriate, the filaments 5-7 mm. long, the
styles 1.5 mm. long, and the stigmas scarcely evolute. The pistillate
flowers were described as having filaments 4 mm. long and the anthers
without pollen. The styles were designated as 2.5-3 mm. long, and the
stigmas as reniform, recurved. The individual flowers which I examined
possessed all the fertile characters listed by these authors.

Some of the distinctive characters of this species are: (1) the pubescent
terminal leaf-bud; (2) chartaceous leaves with prominent veins (15-20
pairs) underneath, which anastomose near the margin and form a con-
spicuous network near the margin; (3) the 6-10-celled ovary and fruit
with 5 or 6 styles; (4) glabrous petals, emarginate or obtuse at the apex,
thickened at the median portion; and (5) the very numerous stamens
3—4-seriate. Another leaf-character, usually present, which is distinctive,
is a pair of lines midway between the midrib and the margin and some-
what parallel to the margins of the leaves. This probably is caused by the
folds of the leaf in the bud. In some species of other genera of the
Theaceae, especially in species with a noticeable pubescence on the under
surface, this character is more pronounced because of a heavier growth
of pubescence along this line. Other species exhibit more than a single
pair of lines.

According to L. R. Holdridge (loc. cit.) this species is restricted to the
eastern mountains of Porto Rico above 300 m. in elevation. So far as is
known, L. portoricensis is the sole representative of the genus in Porto
Rico. Vernacular names are Maricao verde and Maricao.

ARNOLD ARBORETUM,

HARVARD UNIVERSITY.

1949] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 187

THE MORPHOLOGY AND RELATIONSHIPS OF
CERCIDIPHYLLUM

B. G. L. Swamy anp I. W. BAILEy
With two plates and eight text-figures
INTRODUCTION

OF THE FIVE FORMERLY ASSOCIATED GENERA, Tvrochodendron, Tetra-
centron, Euptelea, Cercidiphyllum and Eucommia, the first three have
recently been thoroughly re-investigated by A. C. Smith (29, 30), Bailey
and Nast (3) and Nast and Bailey (24, 25). The totality of evidence from
all organs and parts of these plants indicates that, although the vesselless
genera Trochodendron and Tetracentron are related, they exhibit no close
affinities either to Euptelea or to the Magnoliaceae (sensu stricto), Win-
teraceae, Illiciaceae, or Schisandraceae.

There has been much speculation regarding the relationships of Cercidi-
phyllum, and it is desirable that this genus also be re-examined from a
broad morphological point of view.

LEAF: EXTERNAL FORM

One of the most conspicuous characteristics of Cercidiphyllum japonicum
Sieb. et Zucc. is its dimorphic foliage, Fig. 1. The broadly cordate or
reniform, palmately veined leaves with crenate margins, that are referred
to in the generic name as Cercis-like, are borne on short shoots. On the
contrary, the long shoots of the current year’s growth bear leaves which
fluctuate from elliptic to deltoid to broadly ovate and have entire or finely
rounded-serrate margins. In leaves of the short shoots (d-z), the primary
veins diverge from a single locus at the base of the leaf, whereas in leaves
of the long shoots (j-p), they frequently diverge in pairs at successive
levels. Although both kinds of leaves fluctuate considerably in size, the
ranges of variability in form, venation, and character of the margin are
much wider in the case of the leaves of the long shoots than in those of
the short shoots. The leaves of seedlings have fewer marginal glands and
therefore fewer and relatively much coarser appearing crenulations (b-c).
The cotyledons are oblong-obtuse with entire margins (a).

Brown (5) has performed a highly significant paleobotanical task in
re-investigating fossil floras and in demonstrating by the occurrence of
associated fruits, seeds and leaves that Cercidiphyllum was widely dis-
tributed in the Northern Hemisphere during the Upper Cretaceous and
Tertiary. The fossil leaves had previously been assigned to twenty-one
different genera, including Boehmeria, Ceanothus, Cercis, Cissus, Dom-
beyopsis, Ficus, Grewia, Hakea, Hedera, Paliurus, Piper, Populus, Smilax,
Viburnum, and Zizyphus, and the fruits and seeds to suggested relation-

188 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

Fic. 1. a-c. Cotyledon and seedling leaves, a, x 7, 6, X 6, c, X 5.
d-t. Outline tracings of the leaves of the short shoot, X %. j7-p. Same,
of long shoot, x 4%. In each figure, only the prominent veins are shown.

1949] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 189

ships with such diversified plants as the conifers, palms, Leguminosae,
Nyssaceae, Tiliaceae, Proteaceae, etc.

In view of the obvious polymorphism of the leaves of the surviving
species of Cercidiphyllum, and in order to avoid overloading the literature
with a host of fossil species based upon minor variations of leaf mor-
phology, Brown recognized four fossil species based upon norms of foliar
form; (1) elliptic to broad-ovate-elliptic with rounded or cuneate base
in C. ellipticum (Newberry) Brown of the Upper Cretaceous and Paleo-
cene, (2) deltoid with tendency to elongate apex and incipient cordate
base in C. arcticum (Heer) Brown from the Paleocene to the middle Eocene,
(3) elongate broad-lanceolate with cordate base in C. elongatum Brown
from the middle or late Eocene to the upper Oligocene or lower Miocene
and (4) cordate, slightly elongate and asymmetric in C. crenatum (Unger)
Brown from the Oligocene to the late Miocene.

Although all five species of Cercidiphyllum exhibit wide ranges of foliar
variability, it is evident that the frequency of occurrence of specific leaf
forms changes in passing from the Upper Cretaceous C. ellipticum to the
living C. japonicum. The broadly cordate leaves with crenate margins
that Brown adopts as the norm of the latter species occur on the short
shoots, whereas the leaves of the long shoots exhibit a much wider range
of morphological variability which simulates that which occurs in the
older fossil species. It appears likely that one of the extreme variant forms
of the ancestral species may have stabilized on the short shoots of C.
japonicum and possibly also of C. crenatum. This suggests that all of
the leaves of C. ellipticum and C. arcticum may have been borne on long
shoots. In any case, there are two norms of foliar form in C. japonicum,
one characteristic of the short shoots and the other of the long shoots;
the ratios of the two types of leaves fluctuating on a tree during successive
parts of a growing season.

LEAF: VASCULATURE

Each of the cotyledons of C. japonicum is vascularized by two dichoto-
mizing strands, the central branches of the dichotomies fusing to form
the midvein of the cotyledon, Fig. Ja. The two strands are related
to a single gap in the eustele of the cotyledonary node, Fig. 55. On
the contrary, each of the paired primary leaves of the seedling is vas-
cularized by three strands that are related to three separate gaps in the
eustele, Fig. 56. The leaves of subsequently formed long shoots similarly
are vascularized by three strands that are related to trilacunar nodes, Fig.
57. In contrast the leaves of the short shoots differ markedly in having
three strands that are related to unilacunar nodes, Figs. 5, 6. The three
strands behave similarly, however, in the petiole and lamina of leaves of
both long and short shoots. In the basal part of the petiole two branches
of the median strand pass* into an adaxial position with inverted orienta-

*The phraseologies used in this paper are descriptive of changes observed in suc-
cessive serial sections and have no implications regarding ontogenetic or eee
tention of the vascular elements.

190 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

tion of xylem and phloem, Figs. 7 and 58. These small strands with the
three larger ones become aggregated into a vascular cylinder, Figs. 8 and
59, which extends throughout the central part of the petiole. Toward the
base of the lamina, this vascular cylinder becomes invaginated on its
adaxial side, Fig. 60, and gives rise to from five to seven U-shaped or
concentric strands, Figs. 61 and 62, that constitute the primary veins of the
palmate lamina.

Figs. 2-11. Fig. 2. Diagrammatic reconstruction of the pattern of vas-
culature of a short shoot. Figs. 3-8. Transverse sections of the short shoot
at levels as indicated by corresponding numbers in Fig. 2. Figs. 9-11.
Drawings made from cleared preparations of the first, second and third bud
scales to show the method of vascularization. (bs1— outermost bud scale;
bs2 — inner bud scale; bs3 —innermost bud scale; lf — leaf; ax. bd. — axil-
lary bud.)

1949 ] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 191

LEAF: MISCELLANEOUS STRUCTURES

As observed and figured by Harms (13), there is a precocious develop-
ment of concrescent stipules and large marginal glands even in leaf
primordia that are still enclosed within the bud. As the leaf expands the
adaxially oriented, conspicuously forked stipular appendage drops off and
the glistening, hyaline glands abort, leaving more or less conspicuous
depressions in the margin of the matured leaf. As noted by Solereder (31),
the stomata are confined to the lower epidermis and are surrounded by from
four to six ordinary epidermal cells. Rarely, however, one or two of the
latter cells may undergo anticlinal division to form cells of smaller dimen-
sions. The leaves do not contain secretory idioblasts or sclereids, nor do
they form characteristic types of hairs or trichomes.

SPeOWER BUD OF THE SHORT SHOOT

The bud of the sympodially elongating short shoot has three conspicuous
reddish scales. The outermost of these (ds7) is connivent, viz. oriented
with its dorsal side toward the main axis of the preceding year’s growth,
Figs. 2, 6, and 8. The second scale (4s2) is opposite the first and the
third (ds3) in conformity with the first. A leaf (/f) with a bud (ax dd) in
its axil forms opposite the innermost scale. The inflorescence is terminal.
Thus, each year’s elongation of the short shoot is produced by the activity
of an axillary bud.

The outermost bud scale (6s1) contains two main veins that are related
to two widely separated lacunae of the vascular cylinder, Figs. 3-7.
These veins remain unbranched for some distance and then bifurcate,
Fig. 9. Most of the branches arising from the veins tend to spread in a
downward direction. The second bud scale (4s2) has five main veins that
are associated with a corresponding number of lacunae, Figs. 4-8. The
veins branch more or less profusely in the Jamina and the branches anas-
tomose forming a reticulate system, Fig. 10. The innermost bud scale
(6s3) is supplied with from seven to eleven veins, arising from closely
spaced but distinct lacunae, Figs. 4-8. The veins run parallel for a greater
distance in the bud scale and then branch in a predominantly dichotomous
manner, Fig. 11.

The parenchymatous tissue between the stelar bundles that alternate
with those of the third bud scale soon become bridged by the activity of
the cambium, and at a higher level the stele appears as a continuous arc
on the adaxial side, Fig. 5. Conversely, on the side of the second bud scale,
cambium fails to differentiate in the inter-fascicular regions excepting in
the parenchyma flanking the median vein of the second bud scale. Thus,
the vasculature on this side (right hand side in Fig. 5) becomes organized
into three large strands. These enter the petiole, Fig. 6. It is evident,
accordingly, that the three veins of the leaf of the short shoot are related
to a single lacuna of the eustele in contrast to the nodal situation of the
long shoot, where the three veins are related to three separate lacunae,
Figs. 56, 57. The arc-shaped segment of the stele on the side of the third

192 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

bud scale soon becomes cylindrical, Figs. 6, 7, and continues into the
reproductive axis.

THE FEMALE INFLORESCENCE

The reproductive axis is highly condensed and bears either carpels or
stamens in conformity with the unisexuality of the plant. In the female
inflorescence, the axis ae from two to six carpels, whose ventral sutures
are characteristically directed away from the axis, Figs. 12, 16. At first
sight the carpels seem to be aggregated in a cycle, but a careful examina-
tion reveals a more or less decussate arrangement, Fig. 12. Each carpel
is subtended by a membranous bract. The relation of bracts to carpels is
especially significant in those instances where more than four carpels are
involved. As noted by Harms (13) and as is evident in Fig. 20, the
bracts of the inner two “pairs” of carpels are distinct, originate at higher
levels and do not represent parts of a single involucre.

The shape, size and vascular supply of the bracts vary considerably. The
bracts related to the outermost carpels generally are larger and more deeply
lobed than those accompanying the inner carpels. The larger bracts usu-
ally receive three veins from the eustele of the reproductive axis, Figs. 13,
14; however, either the median or a lateral may be absent. The bracts sub-
tending the inner carpels very rarely receive all three veins. Especially
when the bracts are very much reduced, they may be unvascularized.

The carpel of Cercidiphyllum exhibits a particular trend of specialization
of the primitive conduplicate megasporophyll* of the Tasmannia-Degeneria
type. It has acquired a much elongated form with a cylindrical ovule-
bearing part and a long slender style having two extensively “decurrent”
stigmatic ridges, Fig. 19. The external double stigmatic crests of the Tas-
mannia type have been completely eliminated from the fertile part of the
carpel, and the inner ventral surfaces beyond the placentae remain tightly
adpressed, Figs. 12, 16, until dehiscence of the fruit. The stigmatic part of
the carpel exhibits some flaring of the conduplicate surfaces, Fig. 18.

A reproductive axis bearing four carpels is illustrated in Figs. 12-19.
After vascularization of the bracts, the bundles of the eustele resolve them-
selves into four groups, Fig. 14 (i, ti, i, iv). The peripherally situated
bundles of each group (marked v in the figures) form the ventral veins of
the corresponding carpels. The bundles that are situated towards the
interior of the groups (marked m in the figures) give rise to the median
vein. In the majority of cases, a single bundle from the stele departs into
the carpel as its median vein, as in the carpel marked ii in the figures.
Frequently, however, the median vein may be a fused product of two dis-
tinct stelar bundles, as in the carpels marked 7 and iv in the figures. Car-
pels exhibiting one or the other of these types of median veins fluctuate
widely in distribution; some reproductive axes may have one type or both

*It seems wise to retain the terms ee eae and microsporophyll for descrip-
tive purposes in dealing with the broader forms se angiospermic fertile appendages,
without implications of their eke from leav

1949] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 193

types in varying proportions. A few of the remaining stelar bundles of each
group disappear at the bases of the respective carpels, although some of
these veins may occasionally traverse the tissue of the carpel for a short
distance (such veins are indicated by broken lines in Fig. 12)

Figs, 12-20. Fig. 12. Vhree-dimensional drawing of the vascular system
of the female reproductive axis and carpels (vasculature of carpel marked
III is not shown). Figs. 13-18. Transverse sections of the female repro-

ductive axis and carpels as indicated by corresponding numbers in Figs. 12
and 19. Fig. 19. Four carpels from a reproductive axis, showing the gen-
eral form. Vasculature is shown only for the carpel on the right- fend side.
Fig. 20. An optical transverse section of a female reproductive axis bearing
eight car pels , showing the relationship of the carpels and bracts. (m—
median vein of the carpel; ~— ventral vein).

It may be noted that generally a larger number of stelar bundles are
aggregated below the ultimate carpel (number zv in the case illustrated in
Figs. 12, 14). In addition to forming the median and ventral veins of
the carpel, many of these bundles continue into the base of the carpellary
wall (veins indicated by broken lines in carpel iv, Fig. 12). Within the
carpel, the median vein gives off lateral branches that spread in the ovary

194, JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

wall. The ventral veins, after vascularizing the ovules, extend upward in
the style for more than half its length, Fig. 17; (also, carpel on the right
hand side in Fig. 19).

THE MALE INFLORESCENCE

The male reproductive axis terminates in a cluster of stamens, subtended
not infrequently by four membranous bracts corresponding to those associ-
ated with the carpels. It is important to note that the bracts tend to be
arranged in opposite pairs on the reproductive axis, at slightly different
levels and not in a truly cyclic manner, Figs. 21, 22. Furthermore, eac
bract appears to subtend an individual aggregation of stamens, the number
of stamens in each aggregate varying from 8 to 13. The relation of the
bract to the aggregation of stamens is particularly evident in the lower
part of the axis, Fig. 27.

Figs, 21-24. Fig. 21. Male reproductive axis. Note the arrangement
of stamens in bunches, each bunch being related to a bract. Figs. 22-24

Transverse sections of the same from levels as marked by corresponding
numbers in Fig. 27.

Although each bract in general is supplied with three vascular strands
from the stele of the reproductive axis, the range of variation cited in con-
nection with the bracts of the female reproduciive axis is often encountered
here also. After the vascularization of the bracts, the remaining bundles of
the stele soon become centrifugally distributed, and each bundle supplies a
single stamen, Figs. 22-24.

In the young condition, before the opening of the bud scales, the stamens
have relatively long anthers and very short “filaments,” Fig. 21. On the

1949] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 195

contrary, at anthesis the “filaments” become greatly elongated and are
nearly as long or even longer than the anthers. The “connective” is much
reduced and the elongated sporangia are oriented in a latrose position.
However, the microsporophyll broadens beyond the sporangial region form-
ing a short conical apex. The staminal vascular bundle runs unbranched
almost to the distal extremity of the stamen.

NATURE OF THE “FLOWERS”

As emphasized by Solereder (31) and subsequently by Harms (13), the
abaxial orientation of the ventral sutures of the carpels in Cercidiphyllum
japonicum is a serious, if not an insuperable, obstacle to interpreting the
female reproductive structures as a single flower. A detailed study of suc-
cessive stages in the development of carpels and fruits reveals no evidence
of “twisting,” Hutchinson (18), or resupination. Furthermore, the vascu-
larization pattern of the reproductive parts negates any assumption of tor-
sion or rotation. As previously indicated the ventral veins of the carpel are
derived from bundles situated toward the periphery of the axis whereas the
dorsal vein arises from a more centrally situated bundle or bundles, Figs.
14, 15. The decussate arrangement of paired carpels with their subtending
bracts — particularly where six carpels are present — invalidates any in-
terpretation of the bracts as parts of a single involucre, as noted by Harms

15)

That the reproductive structures of Cercidiphyllum japonicum were
actually derived by extreme reduction of an inflorescence is clearly dem-
onstrated by paleobotanical evidence, Brown (5). In Cercidiphyllum
ellipticum, C. arcticum and C. elongatum the fruits were borne on a much
elongated axis and in the case of the first species on a branching inflores-
cense. Indeed, Brown concludes that “The fruits, at least in the early
species, were borne in racemes, the individuals being indiscriminately alter-
nate and opposite in the same raceme.” Thus, each carpel of C. japonicum
represents the vestige of a much reduced female flower and by homology
each of the smaller aggregates of stamens with its subtending bract is the
remains of a single male flower.

Two different trends in the reduction of the female flower to a single
carpel with a subtending bract have been suggested by Solereder (31)
and Harms (13). The two possibilities are graphically illustrated in Fig.
25. According to Solereder, the ancestral female flower of Cercidiphyllum
might have possessed at least two carpels whose ventral sutures faced one
another, Fig. 25,4. In each flower, the particular carpel which has its ven-
tral suture adaxially oriented disappeared during subsequent evolution,
Fig. 25,B, and the carpel which has its ventral suture abaxially oriented
survived, Fig. 25,C.

On the contrary, Harms emphasized the similarity between the orienta-
tion of the carpel on the axis of the reduced inflorescence and that of the
first bud scale on the vegetative axis. He hypothesized that the bud scale
became fertile and transformed into a carpel and that the vegetative leaf

196 JOURNAL OF THE ARNOLD ARBORETUM [Vols SEX

or its stipular appendage became the subtending bract of the carpel, Fig.
Pye Oy oe

Both of these suggestions are obviously highly speculative, but Solereder’s
concept is preferable owing to its directness and simplicity, involving
merely loss of parts and not calling upon complex transformations of bud
scales into carpels and of leaves into bracts. Unfortunately there are no
reliable data available for thoroughly substantiating what actually occurred.
Although the supernumerary bundles at the base of the carpels are indica-

s

~
x
a
-
oe

y
4
se

25
Fic. 25. Schematic representation of the two interpretations of the
“flower” of Cercidiphyllum. A-C conveys Solereder’s view and D-F, that
of Harms.

tive of a loss of parts, reduction has progressed to a stage where it is impos-
sible to determine with certainty the number and character of the append-
ages that may have been eliminated. Through the courtesy of Dr.
Roland W. Brown, we have been able to examine the inflorescences of fossil
species. Some of these bear paired carpels (fruits), but the preservation
of the material is such (compressions and casts) that there is no conclu-
sive evidence to indicate whether the ventral sutures of the paired carpels
face one another or are oriented as in the much compressed inflorescence of
C. japonicum. The discovery of favorably preserved material may clarify
the issue.
EMBRYOLOGY

A single row of ovular primordia originates on each placenta. At first,
the ovules appear as tiny knobs, Fig. 26, and even before the differentia-
tion of the integuments, assume an orientation more or less parallel to the
placenta, Fig. 27. During the development of the integuments, Fig. 28,
the chalazal end of the ovule grows at a very rapid rate in the form of a
flattened tapering projection, Figs. 29-32, which later develops into a wing.
The differentiation of the vascular bundle in the ovule is rather late. The
procambium of the bundle appears first in the funicular region, when the
embryo sac is already in the 2- to 4-nucleate stage, Fig. 29. Even at the

1949 ] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 197

time of fertilization, the procambium does not extend as far as the nucellus,
nor have many of the constituent cells of the bundle become differentiated
as spiral elements, Fig. 30. It is only after fertilization that the bundle
undergoes a rapid and complete development. It first proceeds in the
direction of the chalaza, and in the middle of the wing takes a circuitous
course to reach the base of the nucellus, Figs. 37, 32.

3|

Po ce

LSD O° —

i ee I GE

ONL EE QE
Nea iS

ON ERS

ics. 26-33. Fig. 26. Ovular primordia on a placenta as seen in dis-
sected preparations, x 70. Fig. 27. Longitudinal section of an ovular
primordium at a slightly later stage than in the previous figure, X 35. Fig.
§. Same, at the time of the origin of integuments, X 35. Fig. 29. Longi-
tudinal section of an ovule when the megaspore mother cell is differentiating
in the nucellus, x 35. Fig. 30. Same, at the time of fertilization, x
Fig. 31. Same, at an early stage in the formation of the endosperm, x 35.
Fig, 32. Longitudinal section of a mature seed, x 20. Fig. 33. A section
of the seed passing along the broken line in Fig. 32, x 70.

198 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

The two integuments arise more or less simultaneously, Fig. 28. The
outer integument grows beyond the inner, Fig. 29, and both the integu-
ments take part in the construction of the micropyle, Fig. 30. The inner
integument is from two to three layers of cells in thickness and the outer
from four to five layers of cells at the time of anthesis. During subse-

(>
Se)
or

Ae
KEES
wanraive

AL

a
saoee
an

secatsug
ST EY
eo i ey aes,
er as
eS
a,

ae

nes

A

OY
,

Nt
Wsact

37 38

Fics. 34-39. Fig. 34. Longitudinal section of the nucellus showing the
megaspore mother cell, x 240. Fig. 35. Same, showing linear tetrad of
megaspores, X 240. Fig. 36. Same, showing 4-nucleate embryo sac, x 240.
Fig. 37. Same, showing mature embryo sac, x 240. Fig. 38. Outline
drawing of the polar view of a pollen grain, x 1500. Fig. 39. Portion of
the pollen grain, enclosed within the broken lines in Fig. 38, enlarged to
show the sculpturing, x 3000.

1949 ] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 199

quent development, the inner integument becomes crushed and the outer
one undergoes certain modifications which will be described in connection
with the seed.

The primary archesporial cell differentiates in the hypodermal layer of
the nucellus and divides by a periclinal wall into primary parietal and sporo-
genous cells. The former cell further divides in the same plane to give
rise to from four to five parietal cells, Fig. 34. The epidermal cells cap-
ping the parietal tissue also undergo periclinal divisions to produce two or
three additional layers that merge with the nucellus, Figs. 35-37. The
sporogenous cell functions as the megaspore mother cell, Fig. 34. It di-
vides meiotically to produce a linear tetrad of megaspores, Fig. 35, the
chalazal megaspore further developing into the 8-nucleate embryo sac,
Figs. 36-37. The antipodal nuclei organize into cells and the polar nuclei
fuse before fertilization, Fig. 37.

The structure of the anther and the development of the male gameto-
phyte do not deviate from the usual method known for the majority of
dicotyledons. The anther tapetum is of the binucleate and secretory
type. Microspore tetrads are formed by a simultaneous method of reduc-
tion divisions. After the formation of the vegetative and generative cells,
the microspores are ready to be shed, and the division leading to the forma-
tion of the gametes takes place in the pollen tube.

The spherical and tricolpate pollen grains of Cercidiphvllum are very
distinctive. The furrows are so broad that they occupy more than half
the surface of the grain, Fig. 38. It frequently is difficult to determine
where the furrows begin and where the exine ends. The exine is finely
pitted. The floor of the furrow is flecked with minute granulations, Fig.
39.

After fertilization, the embryo sac expands in a direction parallel to the
long axis of the ovule. The primary endosperm nucleus moves towards
the chalazal end of the sac and divides; the division being followed by the
deposition of a wall. As a result, the embryo sac becomes transversely
partitioned into a larger micropylar chamber and a smaller chalazal cell,

ig. 40. The latter often shows a denser accumulation of cytoplasm with
characteristic vacuolation, Fig. 41. However, by the time the upper cham-
ber divides into four to six cells, Fig. 42, the cytoplasm of the chalazal cell
thins down and the cell begins to divide. The early divisions in both the
cells of the 2-celled endosperm, Figs. 40, 41, take place essentially by trans-
verse walls until from two to twelve endosperm cells are formed, Fig. 43.
Simultaneously with these divisions, the embryo sac undergoes considerable
longitudinal stretching and presents a very much elongated structure as
shown in Figs. 42-44; so that the antipodal end of the embryo sac comes
in contact with the vascular bundle of the ovule, Fig. 37. Subsequent
divisions in the endosperm cells take place in various oblique planes, Fig.
44, and finally produce a compact mass of nutritive tissue, a large propor-
tion of which, however, is used up by the embryo.

200 JOURNAL OF THE ARNOLD ARBORETUM [yobs x

The zygote divides only after the endosperm cells begin dividing in
oblique planes, as represented in Fig. 44. The basal cell of the 2-celled
proembryo, Fig. 45, does not divide further but undergoes conspicuous en-
largement, Figs. 47-50. The terminal cell divides by a wall parallel to
the first, Fig. 46, so that the proembryo now consists of a linear row of
three cells. The next division in the two derivatives of the terminal cell

CVC ES

a

VAs

Wires. 40-50. Figs. 40-44. Stages in the development of the endosperm,
x 140. Figs. 45-50. Stages in the development of the embryo, x 560.

takes place by vertical walls, Figs. 47, 48. Subsequent divisions in these

ment give rise to the mature embryo. At the stage of dispersal of the
seed, the embryo is well differentiated and possesses a long hypocotyl and
two cotyledons that enclose the primary shoot apex. A few layers of endo-
sperm cells may persist as remnants around the embryo, but are insig-
nificant, Figs. 32, 33.

1949 | SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 201

In the mature seed, the outer integument becomes slightly compressed
in the plane of the flattened wing; therefore, the chalazal wing appears
to be extended on either side of the seed, Fig. 33. In the chalazal region
of the wing, the cells are unequally and variously enlarged so as to present
a more or less spongy texture. The vascular bundle retains its downwardly
projecting loop, Fig. 32.

SECONDARY XYLEM AND PHLOEM

The secondary xylem of C. japonicum has been described by a succes-
sion of investigators. Therefore, certain features only need be commented
upon by us. The growth rings as seen in transverse sections of the wood,
Fig. 51, are delimited by three to five layers of radially compressed imper-
forate tracheary elements. The thin-walled, angular vessels are diffusely
scattered, but in some specimens those of the early-wood are conspicuously
larger than those of the late-wood, Fig. 51. The vessel members are very
long with extensively overlapping ends, indicative of their formation by
a relatively primitive form of non-storied cambium. The “perforation
plates” are scalariform with numerous slender bars, Fig. 54. The vessel
members frequently extend beyond the perforated facets in the form of
narrow tapered ends, having “tertiary” helical thickenings, Fig. 54. The
intervascular pitting fluctuates between scalariform, transitional and oppo-
site. The rays of the first-formed secondary xylem are closely spaced,
longitudinally extensive and mostly uniseriate, but in passing outward
radially soon develop biseriate parts, Fig. 52. The rays of the later
formed wood are lower, more widely spaced and typically heterogeneous,
Fig, 53. The ray-vessel pitting varies from scalariform to transitional to
opposite. Relatively large crystals of calcium oxalate occur more or less
sporadically in the erect cells of the heterogeneous rays. Xylem paren-
chyma is apotracheal, scanty, diffuse, and terminal. The imperforate
tracheary cells have conspicuously bordered pits in both their radial and
their tangential walls. Although the cambium, vessels and imperforate
tracheary cells are of relatively primitive dicotyledonous types, the ray
structure is of a form indicative of an advanced stage in the ontogenetic
and phylogenetic reduction in width and height of the multiseriate rays,
Barghoorn (2).

According to Solereder (31), the bark contains isolated strands of
sclerenchyma. In old stems, however, the sclerenchyma of the secondary
phloem occurs in the form of successive concentric bands of varying cir-
cumferential extension.

Among plant fossils from early Tertiary beds of East Greenland, is a
fossil wood from Cape Dalton which was assigned to Corvlopsites groen-
landicus by Mathiesen (21). This wood closely resembles that of Cercidi-
phyllum and of certain species of Corvlopsis. It bears evidences of
crystals in its wood parenchyma and rays, and was referred to Corvlopsis
by Mathiesen owing to the absence of such crystals in wood of Cercidi-
phyllum examined by him. The abundance of leaves of Cercidiphyllum

202 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

in early Tertiary strata of northern lands suggests, as noted by Brown (5),
that Corylopsites groenlandicus is in all probability the wood of Cercidi-
phyllum.

AFFINITIES OF CERCIDIPHYLLUM

Although the genus Cercidiphyllum was established by Siebold and Zuc-
carini in 1846, the plant C. japonicum remained without a specific name
until Hoffman and Schultes (17) formalized its binomial, but no attempt
was made to determine its affinities. In his remarks on the Hama-
melidaceae, Baillon (1) casually raised the question whether Cercidiphyl-
lum should be included under that family, noting foliar similarities with
Disanthus. However, he also called attention to analogies with Spirae-
anthemum of the Cunoniaceae (Saxifragaceae), viz. opposite leaves,
hypogynous and apocarpous ovary and winged seeds. Maximowicz (22)
associated Cercidiphyllum with Trochodendron, and particularly with
Euptelea, in the Magnoliaceae, stressing the occurrence of stipules as evi-
dence of relationship to that family.

Subsequent authors generally follow the lead of either Baillon or
Maximowicz. Thus, Solereder (31), Hallier (8, 9, 10, 11), Lotsy (20),
McLaughlin (23) and Croizat (6) emphasize similarities to the Hama-
melidaceae, whereas Prantl (26), Harms (12, 13, 14), Bessey (4), Hayata
(16), Hutchinson (18), Diels (7), Lemesle (19) and others favor retain-
ing Cercidiphyllum within the Ranales (or Magnoliales) either as a mem-
ber of the Magnoliaceae or the Trochodendraceae or as the representative
of a monotypic family, the Cercidiphyllaceae.

The first comprehensive comparative studies of Trochodendron, Tetra-
centron, Euptelea, Cercidiphyllum and Eucommia (both of their external
morphological characters and their internal structures) are those of Harms
(12) and van Tieghem (33). Although cognizant of outstanding differ-
ences between the five genera, Harms, owing to an evident prejudice
against small families, placed Tetracentron in the Magnoliaceae and the
remaining four genera in the Trochodendraceae. On the contrary, van
Tieghem, utilizing similar comparative evidence, placed each of the five
genera in an independent family, grouping the monotypic Trochoden-
draceae and Tetracentraceae with the vesselless Winteraceae in a new
order, the Homoxylées. It should be noted in this connection that Harms
(13) later recognized the necessity of removing Tetracentron from the
Magnoliaceae, of placing Cercidiphyllum in a family of its own within
the Ranales and (15) of transferring the monotypic Eucommiaceae to the
Urticales where Tippo’s (34) subsequent investigations indicate that it in
all probability belongs. Harms’ (13) comments upon the taxonomist’s
difficulties in dealing with such genera as Cercidiphyllum, Trochodendron
and Tetracentron are particularly significant. He states regarding Cer-
cidiphyllum:

“On the whole the genus gives us many a problem; its morphological struc-
ture shows so many peculiarities that it is hard to find the correct place in the

1949 ] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 203

system for the genus. But so it is with many other genera from eastern
Asia. Often one does not know where to place these odd forms or to what
more widely distributed and more richly developed families one should con-
nect them. We may as well assume that these are old genera whose closer
relations no longer exist. They are remainders of some more richly
developed forms from former geological periods, now unknown to us, pro-
jecting into today’s flora, in which they appear stra ange. Both the Japanese
genus Trochodendron and Tetracentron, a genus growing with COIS NEUE
in the mountains of central oo and placed in the Magnoliaceae, where it
does not belong, are isolated types.’

This tendency among eee and morphologists for gradually rec-
ognizing the necessity of segregating such relic genera in independent
families of their own is in marked contrast to the highly speculative deduc-
tions of Hallier (11), who included Trochodendron, Tetracentron, Eup-
telea, Ceridiphyllum and Eucommia in the Hamamelidaceae along with the
Balanopsidales, Batidales, Buxaceae, Geissolomataceae, Daphniphyllum,
Myrothamnus, Platanus and other plants of doubtful affinities. Such a
procedure stretches the family beyond the elastic limits of even a natural
order.

Much of the confusion during the last fifty years in discussions of the
angiosperms is due to the interjection of phylogenetic concepts into sys-
tems of classification that were not developed originally from an evolu-
tionary point of view. The older systems are logical arrangements de-
signed to facilitate accurate identification and to provide efficiency in the
codification and use of specimens in large herbaria. The groupings of
plants were primarily based upon morphological similarities in one or
another of their parts. Modern attempts to arrange the larger groupings,
viz. families and orders, in phylogenetic series, deriving one grouping
directly from another, lead inevitably to increasing confusion and contro-
versy. This is due to the fact that, in dealing with the survivors of very
ancient angiospermic floras, such phylogenetic arrangements, based upon
the study of specific selected parts of the plants, commonly are invalidated
by evidence from other organs or parts of the same plants. The most that
may be inferred from the totality of evidence from all organs and parts
of the plants is, in a majority of cases, the conclusion that the related groups
of plants were derived from common ancestors, now extinct. Furthermore,
although negations of putative phylogenetic derivations are simple and
conclusive, positive assertions Se evolutionary derivations are diffi-
cult and uncertain owing to the common occurrence of parallel and con-
vergent trends of morphological oe anatomical specializations in all
organs of the vascular plants. That similar end-products, resulting from
such evolutionary trends, are of widespread and frequent occurrence among
animals is becoming increasingly recognized by zoologists in dealing with
both vertebrates and invertebrates.

Before attempting to arrange surviving angiosperms in phylogenetic
series, it is essential to obtain reliable evidence regarding salient trends of
evolutionary specialization in the various organs and internal structures of

204. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

these plants. Such evidence can be acquired only by comprehensive and
time-consuming investigations of the dicotyledons and monocotyledons
as awhole. In the case of the vascular tissues, accumulated data indicate
that an adequate record of the derivation of vessels from modified
tracheids and of the varied trends of specialization of vessel-containing
xylem is preserved among both living dicotyledons and monocotyledons.
However, it is evident from this record that similar or parallel trends of
structural specialization have occurred repeatedly and independently in
remotely related dicotyledonous families.

The voluminous data accumulated by successive generations of taxono-
mists provide abundant material for the study of varied trends of cohesion
and adnation of floral parts and of simplification and reduction of inflores-
cences and flowers. Furthermore, the study of new and neglected plants
from northern Australia, New Guinea, New Caledonia, Fiji and adjacent
regions is yielding significant clues regarding the primitive form of the
angiospermic carpel and stamen. Here again, as also in the case of
angiospermic pollen, it is becoming increasingly evident that similar end
products of specialization frequently develop independently through paral-
lel or convergent evolution. Although much less is known at present
regarding reliable trends of phylogenetic specialization in leaves and seeds
of dicotyledons, sufficient evidence is available to indicate that morpho-
logical and structural similarities, by themselves, are not indicative neces-
sarily of close genetic relationship between plants in which they occur.

Although specific parts of Cercidiphyllum exhibit similarities to homol-
ogous parts of other plants, the totality of evidence from all parts of the
plants is not indicative of close relationship of Cercidiphyllum to any par-
ticular family of the dicotyledons. Many of the similarities stressed by
those desirous of including Cercidiphyllum in the Magnoliaceae, Trocho-
dendraceae or Hamamelidaceae are superficial and lose their significance
when closely analyzed. Stipules, palmate venation, marginal glands and
trilacunar nodal attachments of leaves occur in diverse families of dicoty-
ledons, yet the presence of stipules is utilized by some as evidence of rela-
tionship to the Magnoliaceae and by others to the Hamamelidaceae. Simi-
larly, stipules and palmate venation are stressed by some as indicative of
affinity to Tetracentron and by others to certain selected genera of the
Hamamelidaceae. It should be emphasized in this connection, however,
that more detailed studies of the leaves reveal significant differences in the
vasculature of the petioles and in the stomata of Cercidiphyllum. In addi-
tion, the leaves of this genus do not form types of idioblasts and other
cellular structures that occur characteristically in the Magnoliaceae, Tetra-
centron, Trochodendron and various genera of the Hamamelidaceae.

Excessive emphasis has been placed in the past upon superficial resem-
blances due to analogous stages in the reduction of the perianth and other
floral structures. In Trochodendron and Tetracentron, where the totality
of evidence is clearly indicative of relatively close relationship between the
two genera, the unsealed conduplicate carpels exhibit incipient lateral cohe-

1949] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 205

sion, conspicuous nectariferous dorsal bulges and pronounced abaxial
deformation either preceding or following anthesis. In contrast, the free,
style-less carpels of Euptelea have elongated stipes and the stigmatic mar-
gins of the conduplicate megasporophyll are restricted by concrescence to
a localized, more or less concave part of the ventral side of the carpel. In
the Hamamelidaceae, as in many of the Cunoniaceae and Saxifragaceae,
there is a marked tendency for the carpels to be reduced to pairs which
have more or less extensive cohesion of their ventral surfaces. Thus, there
are distinct and entirely different trends of specialization in the carpels of
Tetracentron and Trochodendron, of Euptelea, and of the Hamamelidaceae,
no evidence of which are detectable in the solitary carpel of Cercidi-
phyllum.

The pollen grains of Tetracentron, Trochodendron and certain of the
Hamamelidaceae, as of many other dicotyledons, are tricolpate. Those of
Cercidiphyllum are tricolpate, but differ markedly in the unusual breadth
of their poorly defined furrows and in the detailed structure of their
exine.

Embryological investigations have revealed no cogent evidence for in-
cluding Cercidiphyllum in the Hamamelidaceae. The epidermal cells of
the nucellus in Cercidiphyllum undergo a few periclinal divisions which
result in an increased number of parietal layers, a phenomenon that has
not been recorded in the Hamamelidaceae (27, 35). Following fertiliza-
tion, the embryo sac of Cercidiphyllum undergoes a conspicuous elongation
lengthwise, a feature which is unknown in the ovules of the Hamamelida-
ceae. The development of endosperm in Hamamelis (27), Fothergilla
(35), and Corylopsis is strictly nuclear, whereas in Cercidiphyllum it is
cellular.

The seeds of the Hamamelidaceae are of two different structural types:
(1) those having a highly opaque, hard testa without wings, and (2) those
that form a less opaque, relatively soft coat, whose margins become flattened
to form a wing-like expansion. Seeds of the latter category are of less
frequent occurrence in the family, being confined to the genera Bucklandia,
Liquidambar and Altingia. The seeds of Cercidiphyllum are winged, but
those who favor hamamelidaceous affinities for Cercidiphyllum have been
too sanguine in citing this character as evidence in support of their con-
tention. Winged seeds are of common occurrence and the presence of a
wing in itself is not indicative necessarily of close relationship.

In the seeds of Liguidambar and Altingia, the outer integument sur-
rounding the micropyle has undergone excessive flattening and considerable
extension longitudinally to form the wing. The wing is radially symmet-
rical and its cells are uniform and more or less homogeneous. In Buck-
landia, the histologically similar wing develops asymmetrically and diag-
onally from the outer integument on the side of the ovule opposite the
raphe. There is no conspicuous extension of the chalazal end of the
ovule or seed in these hamamelidaceous genera and the course of the
raphe bundle is normal. On the contrary, the ovules of Cercidiphyllum

206 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

initiate the development of a subchalazal projection at an early stage of
their ontogeny, viz. megaspore mother cell stage. Subsequent longitudinal
and unilateral expansion of this subchalazal outgrowth forms the con-
spicuous wing of the mature seed. The raphe bundle forms a circuitous
loop in the wing before reaching the chalaza, comparable to the hair-pin
bend of the vascular bundle in the subchalazal prolongations of the ovules
and seeds of Trochodendron and Tetracentron, Nast and Bailey (24).

The seeds of the latter genera, however, differ from those of Cercidi-
phyllum in having slender subchalazal elongations without broad wing-like
modifications, in forming sclerenchymatous layers of characteristically
elongated cells, and in containing a rudimentary embryo embedded in
abundant endosperm. In addition, the epidermis of the outer integument
consists of longitudinally elongated cells which impart a striated appear-
ance to the seeds, whereas in Cercidiphyllum the homologous tissue is
composed of cells which by enlargement produce a spongy texture.

The haploid number of chromosomes is commonly 12 in Hamamelida-
ceae. According to Whitaker (36), the haploid number is 14 in Euptelea,
Illictum and the Schisandraceae, and 19 in Cercidiphyllum, Trochoden-
dron, Tetracentron, the Winteraceae and the Magnoliaceae (sensu stricto).
However, the occurrence of 14 haploid chromosomes in one investigated
species of Michelia (32) suggests that many additional representatives of
the Winteraceae, Magnoliaceae, Schisandraceae, and Illiciaceae should be
studied before relying upon chromosome numbers in discussions of rela-
tionships within the Ranales.

The wood of Cercidiphyllum closely resembles that of certain repre-
sentatives of the Hamamelidaceae, e.g. Corylopsis. However, this simi-
larity, by itself, is not indicative necessarily of close relationship, since
similar combinations of anatomical characters occur in other families which
have retained relatively primitive cambia, vessels and imperforate trache-
ary cells while undergoing analogous reductions in the width and height
of their multiseriate rays. In fact, the similarities in the case of Cory-
lopsis and Cercidiphyllum are no more remarkable than are those between
the woods of such remotely related plants as Maclura pomifera (Raf.)
Schneid. of the Moraceae and Robinia pseudoacacia L. of the Leguminosae.

hus, a summation of evidence from all organs and parts of Cercidi-
phyllum provides no cogent arguments for including this genus in any
particular family of the dicotyledons. Therefore, we agree with van Tieg-
hem (33) and subsequent writers who place the genus in an independent
family of its own, the Cercidiphyllaceae.

Having established an increasing number of new iateneadem families,
taxonomists are faced with the problem of incorporating them in some
logical and useful major system of classification. The role of the mor-
phologist, anatomist, paleobotanist or cytogeneticist should be to provide
significant and helpful data rather than to attempt to dictate solutions of
this difficult problem upon their own ‘terms. As previously stated, prema-
ture efforts to arrange living angiosperms in putative phylogenetic series,

1949] SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 207

deriving one family or order directly from another (Hallier and others)
will lead only to increasing confusion and controversy. The study of
major trends of phylogenetic modification of specific organs or parts of
the dicotyledons and monocotyledons is yielding much significant informa-
tion, but it should be recognized that no single one of these phylogenies
can be utilized by itself in arranging plants in a similar linear series, since
such a procedure is usually negated by evolutionary trends in other parts
of the plants.

The Ranales, as broadly conceived by Engler and Prantl, obviously
associates plants of widely varying degrees of relationship and of morpho-
logical and anatomical specialization. Two different categories of families
may be recognized to advantage within the order: (1) the Winteraceae,
Degeneriaceae, Magnoliaceae (sensu stricto), Himantandraceae, Eupo-
matiaceae, Annonaceae, Myristicaceae, Canellaceae, Calycanthaceae,
Austrobaileyaceae, Monimiaceae, Gomortegaceae, Lauraceae, Hernandia-
ceae and Lactoridaceae, having characteristic secretory cells (‘ethereal oil
cells”) and monocolpate pollen (or phylogenetically modified types of
such pollen), and (2) the Ranunculaceae, Berberidaceae, Lardizabalaceae
and Menispermaceae, having tricolpate pollen (or types derived from such
pollen) and no “ethereal oil cells.”” Monocolpate pollen occurs in many seed
ferns, Bennettitales, Cycadales, Ginkgoales and monocotyledons, whereas
tricolpate pollen and its modifications characterizes most dicotyledons.

The plants of the first ranalian category are predominantly woody (nor-
mal trees and shrubs), having simple pinnately veined leaves (except the
Hernandiaceae), syndetocheilic-appearing stomata, and seeds with copious
endosperm and rudimentary embryos. They retain many early trends in
the differentiation of sepals and petals and in the specialization of primi-
tive conduplicate, unsealed carpels, broad microsporophylls and staminodes,
as well as of vessels and other vascular structures. The second category
of ranalian plants exhibits conspicuous trends toward the acquisition of
herbaceous or scandent habits, have a high ratio of palmate venation, and
form haplocheilic stomata. Although predominantly apocarpous, their
flowers have in general attained advanced stages of modification, and their
vascular tissues are highly specialized. However, they retain seeds with
abundant endosperm and small embryos.

The Illiciaceae and Schisandraceae have abundant ‘ethereal oil cells”
and other characters suggestive of the first category of Ranales, but form
tricolpate or hexacolpate pollen. Furthermore, the Schisandraceae, as
certain of the Monimiaceae and Hernandiaceae, exhibit transitions between
syndetocheilic-appearing stomata and haplocheilic stomata. Conversely,
the Nymphaeaceae are in general more nearly comparable to plants in the
sub-order Ranunculineae, but the Cabomboideae and Nymphaeoideae have
monocolpate types of pollen, whereas the Nelumbonoideae form tricolpate
ones. It should be noted, in this connection, that although the Piperaceae
and Saururaceae have attained a high degree of morphological specializa-
tion in most of their parts, they contain characteristic ‘ethereal oil cells”

208 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

and have monocolpate pollen. They should be included in the Ranales, if
this order is to be retained as broadly conceived of in the Englerian system.

The Tetracentraceae, Trochodendraceae (sensu stricto), Eupteleaceae
and Cercidiphyllaceae have tricolpate pollen, haplocheilic stomata and,
with the exception of Tetracentron, no secretory idioblasts. If these fami-
lies are to be retained in the Ranales, it should not be upon unwarranted
assumptions that they are closely related to, or directly derived from the
Winteraceae, Magnoliaceae, or any other specific family of the order, but
rather upon recognition of the order as a useful repository for relic plants
that retain primitive, ancestral, dicotyledonous characters in one or more
of their organs or parts. Nothing is to be gained by transferring such
genera as Tetracentron, Trochodendron, Euptelea, and Cercidiphyllum
into close relationship with the Hamamelidaceae or Saxifragaceae, since this
would merely serve to expand another order into a less homogeneous
assemblage. The only other alternatives are to include,each relic genus
in a special order of its own or to follow the lead of mycologists and estab-
lish a special assemblage, e.g. “Incognitales,” comparable to the Fungi
Imperfecti.

It should be emphasized in conclusion that much of the present confu-
sion in literature dealing with the relationships and phylogeny of angio-
sperms is due to premature and unduly speculative generalization based
upon excessively restricted and inadequate data. A much broader outlook
is needed, involving the use of evidence from all organs and parts of plants
and an accurate visualization of salient trends of specialization of such
organs and parts in the angiosperms, and in the vascular plants, as a whole.

Lie RA TUR. Clr

Baitton, H. Nouvelles notes sur les Hamamélidacées. Adansonia 10:
120-137. 1871.

Barcuoorn, E. S., Jr. The development and phylogenetic specializa-
tion of rays in the xylem of dicotyledons. II. Modification of the
multiseriate and uniseriate rays. Amer. Jour. Bot. 28: 273-281, 1941.

BaiLey, I. W. & C. G. Nast. Morphology and relationships of Trocho-
dendron and Tetracentron, I. Stem, root, and leaf. Jour. Arnold Arb.
26: 143-154. 1945

Bessey, C. E. The phylogenetic taxonomy of flowering plants. Ann.
Missouri Bot. Gard. 2: 109-164. 19

5. Brown, R. W. Fossil leaves, fruits, and seeds of Cercidiphyllwin.
Jour. Paleontol. 13: 485-499, 1939,

6. Croizat, L. Trochodendron, Tetracentron, and their meaning in phy-
logeny. Bull. Torrey Bot. Club 74: 60- 76. 1947.

7. Dies, L. A. Engler’s Syllabus der Pflanzenfamilien, ed. 11. 1936.

8. Harrier, H. Ueber den Umfang, die Gliederung und die Verwandt-
schaft der Familie der Hamamelidaceen. Beit, Bot. Centralbl. 14:
247-260. 1903

9, ————._ Ueber die Gattung Daphniphyllum ein Ubergangsglied von
den Magnoliaceen und Hamamelidaceen zu den Katzehenbluthlern.
Bot. Mag. (Tokyo) 18: 55-69. 1904.

—

bo

=

=

1949 | SWAMY & BAILEY, MORPHOLOGY OF CERCIDIPHYLLUM 209

10. —————. Provisional scheme of the ie at system of
flowering plants. New Phytol. 4: 151-1 905.
11. ———_. ‘origine et le systéme ce i angiospermes exposés

a Vaide de leur arbre généalogique. Arch. Néerld, Sci. Exact. et
Natur. ser. III. 1: 146-234. 1912

12, Harms, H. Ueber die Stellung der Gattung Tetracentron Oliv. und
die Familie der Trochodendraceen. Ber. Deutsch. Bot, Ges. 15:
350-360. 1897.

13. ————. Ueber die Bliithenverhaltnisse und die Systematische Stellung
der Gattung ee Sieb. & Zucc. Ber. Deutsch. Bot. Ges.
34: 272-283.

14. ————_, Zur eee der ie Cercidiphyllum. Mitteil. Deutsch.
Dendrol.-Ges, 26: 71-87.

15. ————. Zur Kenntnis von oe ulmoides Oliv. Mitteil. Deutsch.
Dendrol. Ges. 45: 1-4. 1933,

16. Hayarta, B. The natural classification of plants according to the dy-
namic system. Ic. Pl. Formos. 10: 97-234. 1921.

17, Horrman, J. & H. Scuuttes. Noms indigénes d’un choix de plantes
du Japon et de la Chine. Paris. Imprimerie Impériale. 1853,
(reprinted from Jour. Asiatique, 1852).

18. Ae a The Families of Flowering Plants. I. Dicotyledons.

London.

19. LEMESLE, a0 s divers types de fibres a ponctuations aréolées chez
les Hee apocarpiques les ee archaiques et leur role dans la
phylogénie. Ann. Sci. Nat. Bot. Sér. XI. 7: 19-40, 1946

20. Lorsy, J. P. Vortrage iiber botanische Stammesgeschichte 3: Cormo-
phyta Siphonogamia. 1

21. MATHIESEN, Fr. J. Notes on some fossil plants from East Greenland.
Meddel. on Grgnland 85(4): 1-62. 1932.

22. Maximowicz, C. J. Diagnoses breves plantarum novarum Japoniae et
Mandshuriae. Bull. Acad. Imp. Sci. St. Petersbourg 17: 142-179.
1872.

23. McLaucuiin, R. P. Systematic anatomy of the woods of the Mag-
noliales. cop: Woods 34: 3-39. 1933.

24. Nast, C. G. & I. W. Batey. Morpholosy and relationships of Trocho-
dendron and Meiracenttort II. Inflorescence, flower, and fruit. Jour.
Arnold Arb. 26: 265-276. 1945.

25. ————— & ————. Morphology of Euptelea and comparison with
Trochodendron. Jour. Arnold Arb. 27: 186-192. 1946.

26. Pranti, K. Trochodendraceae. Engler & Prantl, Nat. Pflanzenfam.
3(2): 21- a 1888.

27. SHOEMAKER, D. N. On the development of Hamamelis virginiana.

Bot. Gaz. 39: 248-266. 1905.

28. Srepotp, P. F. & J. G. Zuccartni. Florae Japonicae, 111-240. (reprint
from Abh. Bayer. Akad. Wiss. Math-Phys. Cl. 4(2): 109-204. 1845;
4(3): 123-240. 1846).

29. Smitru, A. C. A taxonomic review of Trochodendron and Tetracen-
tron. ee Arnold Arb. 26: 123-142. 1945,

——. taxonomic review of Euptelea. Jour. Arnold Arb. 27:

175-185. oe

210 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

31. SoLereperR, H. Zur Morphologie und Systematik der Gattung Cercidt-
phyllum Sieb. & Zucc., mit Beriicksichtigung der Gattung Eucommia
Oliv. Ber. Deutsch. Bot. Ges. 17: 387-405. 1900

32. Suerura, T. A list of chromosome numbers in Angiospermous plants.
Bot. Mag. (Tokyo) 45: 353-355. 1931.

33. TircHem, P. van. Sur les dicotylédones du groupe des Homoxylees.
Jour. de Bot. 14: 259-297, 330-361. 1900.

34. Tippo, O. The comparative anatomy of the secondary xylem and the
phylogeny of the Eucommiaceae. Amer. Jour. Bot. 27: 832-838.
1940

35. Tone, K. Y. Studien tber die Familie der Hamamelidaceae mit
besondenee Berticksichtigung der Systematik und Entwicklungsge-
schichte von Corylopsis. Bull. Dept. Biol. Canton, China, No. 2.
1930.

WaHitaker, T. W. EE number in the Magnoliales. Jour.
Arnold aay, 14: 376-385.

tw
a

EXPLANATION OF PLATES

Plate I. Fig. 51. Transverse section of the mature wood, X or Pig.

52. Tangential section near the early secondary xylem, x 120. Ig. 3.
Same, old secondary xylem, x 120. Fig.54. Part of a vessel na show:
ing the spiral Fiennes | in — tapering end, x 320.

Plate II. Fig. 55. Transverse section of a eu ae at the cotyledonary
node, x 40. Fig. 56. Same, at the next node, x 40. Fig. 57, Transverse
section at the node of a twig, x 10. Figs. 58-60. Transverse sections of
the petiole at successive levels starting from the base, x 18. Fig. 61. Trans-
verse section passing through the base of the lamina, X 18. dig. 62. Same,
at a slightly higher level, x 18.

INSTITUTE FOR RESEARCH IN GENERAL PLANT MorpPuHo.ocy,

HARVARD UNIVERSITY.

PLAtE I

Jour. ARNOLD ARB. VoL. XXX

ee

re
"

a>

<tc aus <

Sy

ci

< ~<@; 7
“age ein
~ cg ie SET ey

< ettacoommne

SE ANE RG ees WS 6 ES te
an ee fre hee a hestnnes.
lel ee ett

ee RS Po? chee Sit OAc D&O A Name
steer yas Per eer es ;

SRE Oe “TE ae UF gE Om on ws, Sess
Tae ee aon OD = Cl tae Von (Fn wel ae

4
m4 4
wSE

vey

~

MorPHOLOGY OF CERCIDIPHYLLUM

> II

PLATE

rk. ARNOLD ArB. Vor. XXX

Jou

ese tetas.

A
3:
¢.
tt
3
;
>
¢
t

ae

ae ne

re
we

Aes

ee

=.
AS)

M

YLLU

F CERCIDIPH

ITORPHOLCGY 6

N

JOURNAL

OF THE

ARNOLD ARBORETUM

VOLS OOO JULY 1949 NUMBER 3

THE MORPHOLOGY AND RELATIONSHIPS OF
AUSTROBAILEYA

I. W. Batrey ann B. G. L. Swamy
With seven plates and three text-figures
INTRODUCTION

THE RELATIONSHIPS of this interesting genus have remained in doubt
since it was first collected in 1929 by Kajewski on the Atherton Table-
land in ee eer In his original description of the genus,
White (11) stat

“As Dr. Diels a as done considerable work on the pes and allied
groups, particularly those of the Papuan region, I sent specimens to him.
He wrote me that he regarded the plant as doe belonging to the
Magnoliaceae and allied to Drimys. He further suggested, though the fruits
were unknown, that I should publish a provisional-description in the account
of Kajewski’s collection. Unfortunately though Kajewski’s specimens con-
sist of several sheets of flowering specimens, the material has become
extremely brittle in the process of drying with the result that the flowers are
very difficult to soften and describe. It seems impossible also to section one
of the carpels to see the number of ovules. Until living or spirit asa
of the flowers and ripe fruits can be obtained it will be impossible
the plant in its correct botanical position. For the time being it is eae
best placed at the end of the Magnoliaceae — even in a restricted sense — with
position uncertain

Subsequently, White modified this opinion in a letter written August
5, 1940 to Croizat: “Your letter of the 14th June to hand, and in reply
I might state that when I was in England I spoke to Mr. Dandy at the
British Museum, the recognized authority on the Magnoliaceae, about my
Austrobaileya. I had, of course, only provisionally placed it in the Mag-
noliaceae. Specimens of the wood parts were examined at the Joddrell
Laboratory by Dr. Metcalfe, and from floral structure and anatomy, we
think there is no doubt the plant is a new family allied to Monimiaceae.

212 JOURNAL OF THE ARNOLD ARBORETUM LvOL. XXX

Dandy and I had intended to prepare a publication on it later on. I
might have done something when I was over the other side, but the war
upset plans a bit. It is most desirable, of course, that fruit should be
found before it is raised to family rank.”

Croizat (5) placed the genus in a new sub-family of the Dilleniaceae,
viz., the Austrobaileyeae, and subsequently (6) in an independent family,
the Austrobaileyaceae, without further discussion of its affinities.

It is evident, accordingly, that three different relationships have been
suggested for Austrobaileya, 1. the Magnoliaceae, 2. the Monimiaceae and
3. the Dilleniaceae. The question whether the genus should be placed
in one of these families or in an independent family of its own can be
satisfactorily answered only by comprehensive morphological investiga-
tions of adequately preserved material. Through the efforts of Mr. L. J.
Brass, and Mr. S. E. Stephens of the Department of Agriculture of North
Queensland, we have succeeded in obtaining such material of the flowers
and vegetative parts, collected at Lamonds Hill, close to the type
locality of Austrobaileya scandens C. T, White.

MATERIAL EXAMINED

Our morphological investigations are based upon examination of the
following material:

1. Isotype specimen of A. scandens C. T. White, Kajewski 1269,
Arnold Arboretum.

2. Two isotype specimens of 4A. scandens C. T. White, Kajewskt 1269,
New York Botanical Garden.

3. Isotype specimen of A. maculata C. T. White (12), White 10734,
Arnold Arboretum.

4. Stems and leaves of A. maculata C. T. White, White 10734, kindly
sent by Mr. White.

5. Stems, terminal shoots and leaves of A. scandens C. T. White, col-
lected and preserved in FAA fixative, Brass 18160.

6. Eleven herbarium specimens of Brass 18160.

7. Flowering shoots of Austrobaileya, preserved in FAA fixative, col-
lected August 26, 1948 by Mr. Stephens in the Lamonds Hill area.

THE STEM

The largest scandent stem of Austrobaileya, Brass 18160, examined by
us is 13 millimeters in diameter. Part of a transverse section of this stem
is illustrated in Plate J, Fig. 1. The pith is composed of parenchymatous
elements having thick, lignified, secondary walls. It is relatively homo-
geneous except for the presence of scattered cells having abundant starch
and dark-colored phenolic material. In longitudinal sections, these cells
are seen to occur in more or less extensive, vertically oriented files. The
parenchymatous elements of the perimedullary region, Fig. 1, are slender,
much elongated vertically, thick-walled and internally septate.

1949 ] BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 213

The cortex is composed of parenchymatous cells having thick, unligni-
fied, primary walls and containing abundant starch and more or less cal-
cium oxalate in the form of “crystal sand.” A considerable number of
the cortical cells also contain dark-colored phenolic material. In the
outer cortex of young stems, before the development of an extensive
periderm, conspicuous secretory cells, i.e., “ethereal oil cells,” are visible.
These idioblasts are more or less spherical and have a tenuous suberized
inner membrane.

The eustele of the stem is composed of numerous discrete strands of
primary xylem and phloem which are faced externally by thick-walled
septate fibres. During enlargement of the stem, as seen in transverse
sections, Fzg. 1, these arcs of phloem fibres form parts of a composite ring
of sclerenchyma; the widening intervening spaces between the arcs of
fibres being bridged by the formation of sclerotic parenchyma. The con-
spicuous multiseriate rays of the secondary body, Fig. 1, extend outward
from the parenchymatous interfascicular lacunae of the eustele and exhibit
considerable flaring in the secondary phloem. As is so frequently the case
in scandent stems, there is a relatively abrupt transition in the secondary
xylem between an inner first-formed zone of denser tissue and a subse-
quently formed, more porous part having conspicuously larger vessels.

The cambium is of a relatively primitive type, containing long (up to
1500), extensively overlapping, fusiform initials. The derived sieve cells,
vessel members and parenchyma strands are of equivalent lengths, but
the imperforate tracheary cells, which elongate during maturation of the
secondary xylem, are longer.

The parts of the secondary phloem between the flaring multiseriate
rays, Fig. 2, are composed largely of alternating transverse bands of sieve
cells and phloem parenchyma strands, the uniseriate rays being incon-
spicuous except where their cells contain dark-colored phenolic material.
The sieve cells have no companion cells. Our observations regarding this
important fact have been verified by Professor Katherine Esau to whom
we sent preserved material for examination. The long, extensively over-
lapping sieve cells have numerous sieve plates in their lateral surfaces,
those at the ends of the cells exhibiting no structural differences that might
be interpreted as evidence of the presence of sieve tubes. There are no
sclerenchymatous elements in the secondary phloem, all of the cells having
primary walls only. Elongated secretory cells, i.e. “ethereal oil cells,”
are of sporadic occurrence.

The imperforate tracheary elements of the secondary xylem, Fig. 3,
are of two types, 1. normal thick-walled tracheids which lose their living
contents at maturity and 2. septate elements which contain several nuclei,
more or less starch and dark-colored phenolic material. The former
tracheary cells have numerous, conspicuously bordered pits (with included
slit-like apertures) in both their radial and their tangential walls. How-
ever, the bordered pits fluctuate markedly in size in different tracheids,
varying in diameter from 3 to as much as 10u. In the septate elements,
the bordering areas of the pits fluctuate from conspicuous to vestigial.

214 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

The vessels of the secondary xylem range in size from diameters approxi-
mating those of the largest tracheids, ie. 30u, to diameters of as much
as 200u. All of the vessels regardless of size have scalariform perforation
plates, but the form of the vessel members, and the number, size and char-
acter of the perforations changes markedly with increasing diameter of
the vessels. In those of minimal cross-sectional area, the constituent
cells are tracheid-like in form, i.e. fusiform with gradually tapered ends
as seen in tangential longitudinal sections of the stem. The aggregations
of scalariform bordered pits in the radial facets of the vessel members,
Fig. 5, differ from those of scalariformly pitted tracheids only in the
dissolution of pit-membranes in a number of the fully bordered pit-pairs.
With increase in the diameter of the vessels, the perforated pit-pairs are
reduced in number, greatly enlarged and lose most of their borders, com-
pare Figs. 5, 6 and 7. In the case of the smaller vessels, the perforated
pits occur in radial walls that are only slightly inclined to the long axiS
of the vessels, and therefore are fully visible in thick, radial longitudinal
sections of the stem. On the contrary, during the ontogenetic expansion
of large vessels, restricted parts of the radial facets, ie. those in which
perforations will ultimately be formed, assume an increasingly diagonal
orientation, Fig. 4, and are fully visible in surface view, Fig. 7, only in
sections cut at an inclination of approximately 45 degrees to the longi-
tudinal axis of the stem.

Most of the vessels are diffusely scattered, but a few of the larger ones
occur in tangentially oriented pairs, Figs. 1 and 3. Where the vessels are
in contact with tracheids, the imperforate bordered pits are relatively
large and circular, where in contact with rays or wood parenchyma strands,
the pits are smaller but clearly bordered. Where vessels are in contact
laterally, the imperforate bordered pits are large, more or less trans-
versely elongated and in opposite or alternating, multiseriate arrangement.

The wood parenchyma is paratracheal in distribution, Fig. 3, and, as
frequently the case in scandent stems, tends to be more abundantly devel-
oped on the abaxial sides of the vessels. The multiseriate rays are extensive
longitudinally, Fig. 4, having slender elongated cells on their flanks and
high-celled uniseriate extensions on their upper and lower margins. The
uniseriate rays exhibit evidences of phylogenetic reduction, particularly in
the first-formed part of the secondary xylem, are low and composed of
vertically much elongated cells.

THE LEAF AND NODAL ANATOMY

During early stages of the ontogeny of the vegetative shoot, the leaves
are arranged in a typical decussate phyllotaxy. However, during subse-
quent growth of a shoot, the leaves of an individual pair may remain truly
opposite or they may become sub-opposite or even widely separated by
irregularities of interstitial elongation. The entire, glabrous, pinnately
veined leaves are coriaceous at maturity, being provided with a thick and
conspicuously striated cuticle in surface view. The relatively large

1949 ] BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 215

stomata of the lower surface of the leaf may have subsidiary cells oriented
parallel to the guard cells (“syndetocheilic” or “rubiaceous” appearing
type) or they may be surrounded by ordinary epidermal cells (‘“haplo-
cheilic” type). In anticlinical sections, cut at right angles to the long

3

TEXT-FIGURES 1-3. Fig. 1. Section of a stoma from the under surface of

a leaf, x 690. Fig. 2. Diagrammatic representation of the transy erse sec-
tion of a node and the vascularization pattern of a leaf. Fig. 3. Mature
pollen grain, x 982.

216 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

axis of the stomata, Text-fig. 1, a highly characteristic pattern of cuticu-
larization is revealed. There is a conspicuous vestibule in the thick
cuticle, and the cuticular substance extends inward along the surfaces of
the guard cells forming two massive plates which subtend them. There
is no clearly defined hypodermis or palisade tissue in the mature leaf.
More or less spherical secretory idioblasts, ie. “ethereal oil cells,” are
scattered through the mesophyll, and are particularly conspicuous in
young developing leaves. Many of the cells of the mesophyll contain
dark-colored, phenolic substances.

Each leaf is vascularized (as are so many cotyledons of angiosperms)
by two discrete strands that are related to a single gap in the eustele,
Text-fig. 2. In other words, the nodal anatomy of ee is of a
unilacunar type. The detailed behavior of the two vascular strands
fluctuates considerably in different mature leaves, even in those of a single
plant. In certain of them, the two strands remain separate throughout
the petiole and the costa of the lamina, each half of the lamina being
vascularized by the ramifications of an independent system, Text-fig. 2.
In other leaves, the two strands may be closely approximated or they
may fuse to form a single arc, particularly in the middle and terminal
parts of the mid-rib. It is significant in this connection, however, that
during early stages of the ontogenetic development of leaves, there are
two separate strands of procambium, subsequent approximations and
localized fusions occurring during the differentiation of metaxylem or of
secondary xylem and phloem. Below the node, the two foliar traces
(when traced downward in serial transverse sections of the stem) remain
independent of each other and become attached to two independent parts
of the eustele. In other words, the two leaf traces do not originate as a
dichotomy of a single vascular bundle.

In mature leaves, the smaller veins are jacketed by thick-walled fibres,
but these fibres decrease in number toward the base of the mid-vein and
commonly are absent in the petiole.

THE: FLOWER

The flower buds of Austrobaileya are borne in the axils of leaves and
are solitary in all specimens examined by us. Those collected by Mr.
Stephens, are globose, Figs. 8-10, the pedicels varying from 0.5 to 1.0
cm. in length, depending upon the age of the bud. The short pedicels of
young flower buds bear 8 to 10 pairs of decussately arranged bracts.
Subsequently during interstitial elongation of the pedicel, the bracts of
one or two of the central pairs tend to be displaced. Thus, in older buds
it is common to find 2 or 3 pairs of compactly and decussately arranged
bracts at the base of the pedicel, 2 or 4 more or less widely spaced bracts
in the middle of the pedicel, and 2 or 3 pairs of compactly and decussately
arranged bracts subtending the perianth, Fig. 9. The penultimate and
ultimate pairs of bracts are larger and more woody than the rest, Figs.
8-11.

The perianth, as also the inner appendages of the flower, are arranged

1949] BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 217

on the floral axis in a much compressed spiral. There is no clearly defined
differentiation of the perianth into calyx and corolla. In both species
of Austrobaileya, the outermost tepal is somewhat larger than the sub-
tending bracts. The central tepals of the series progressively attain larger
dimensions, whereas the inner tepals become reduced in size to dimen-
sions more nearly equivalent to those of the outer members of the series,
Figs. 20-32, The tepals of A. scandens, Kajewski 1269, are much larger
than those of A. maculata, White 10734, and the innermost ones tend
to be of spatulate form.

The androecium of A. maculata, White 10734 as of Stephens’ collec-
tion, consists of 19 to 25 members, the outer 6 to 9 of which differentiate
as stamens and the remainder as staminodia. The two categories merge
into one another, one or two staminodia in the transition region not infre-
quently bearing degenerate sporangia, Fig. 17. The fertile male append-
ages are broad microsporophylls bearing protuberant sporangia on their
conspicuously concave inner surface, Figs. 15, 34, 36.

The fertile microsporophylls and the staminodia of A. maculata, White
10734 as of Stephens’ collection, are characterized by having more or less
numerous and conspicuously embossed purple spots, Figs. 15-19 and 34.
These dark-colored areas tend to be more abundant on the staminodia
than on the stamens, and on the adaxial than on the abaxial surface of
the fertile microsporophylls. In the case of dried herbarium material,
White 10734, the phenolic contents of the papillate epidermal cells, F7g.
42, of the purple areas turn black, and the spots become strikingly em-
bossed, evidently through excessive contraction of surrounding tissue.

The stamens and staminodia of A. scandens, Kajewski 1269, are larger
than those of A. maculata, White 10734, and differ from them in both
form and texture, compare Figs. 33 and 34. They do not exhibit dark-
colored embossed spots. Nor does Kajewski mention the occurrence of
conspicuous purple areas in his field-notes on the color of freshly collected
flowers.

The fact that the flowers collected by Mr. Stephens, close to the type
locality of A. scandens, resemble those of A. maculata from Mt. Spurgeon
raises the question whether there actually are two distinct species of
overlapping ranges. Can the striking differences in size, form, texture
and structure between the flowers of A. scandens, Kajewski 1269, and
those of A. maculata, White 10734 and Mr. Stephens’ collection, be due
to different developmental stages of the flowers at time of collection? In
other words, is the larger size of the flowers of A. scandens and of their
constituent parts due to the fact that the flowers were collected at anthesis,
whereas those of A. maculata and the Stephens’ collection had not attained
their final stages of expansion? In order to answer this question, we
have examined a young flower bud of A. scandens, Kajewski 1269, of
approximately the same dimensions as the largest flower buds of A. macu-
lata, White 10734. The stamens and staminodia of this immature flower
exhibit the same differences in form and texture as the fully developed
flowers, and do not have dark-colored embossed spots.

218 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

It should be noted in this connection, however, that the only conspicuous
morphological difference between the vegetative organs of A. scandens
and A. maculata is in the shape of the leaves. The internal anatomical
characters of the stem and leaves are similar. The leaves of 11 duplicate
sheets of Brass 18160 are of uniform elliptic to lanceolate form regardless
of marked variations in size, and thus resemble those of A. scandens,
Kajewski 1269. The “nitid” character of the latter collection is due
apparently to over-heating during drying. The leaves of A. maculata are
eliiptic-ovate to ovate.

The staminodia of A. scandens and A. maculata are irregularly ridged
or folded longitudinally, Figs. 17-19. As seen in transverse sections, their
basal part is solid and the ridges extend in diverse directions, Fig. 36,
whereas their apical parts are characteristically conduplicate, Fig. 35.

Above the insertion of the staminodia, the floral axis terminates in a
broad mound on the surface of which the free carpels are borne, Figs. 12
and 38. In A. maculata, White 10734 and in Mr. Stephens’ collection,
the number of carpels ranges from 6-8, but as many as 14 may be present
in A. scandens, Kajewski 1269. The dorsal outline of the carpel, Fig. 13,
is more curved than the ventral surface which is nearly straight, the style
being in line with the ventral side. The carpel has a more or less extensive
solid basal part, Figs. 14 and 37, which may possibly represent a much
modified stipe. The locule, situated in the dorsally more bulging part
of the carpel, contains from 8 to 13 anatropous ovules in two series, Figs.
14, 37 and 44.

The carpels of Austrobaileya exhibit extreme phylogenetic modifications
of the primitive conduplicate megasporophyll of the ‘““Tasmannia-Degen-
eria” type, viz. 1. closure of the carpel, 2. differentiation of a hollow
style and 3. elimination of the external, paired, stigmatic crests. Closure
of the ovule-bearing part of the megasporophyll has progressed, as in the
Wintera section of Drimys, Bailey and Nast (2), by concrescence of the
stigmatic ventral surfaces of the conduplicate megasporophyll. Although
closure by concrescence is complete externally, Figs. 36 and 44, evidences
of the suture are preserved internally by a conspicuous cleft in the car-
pellary wall, Fig. 44. This cleft is jacketed by papillate cells that func-
tion in the downward extension of pollen tubes. Similar concrescence
has occurred in the constricted stylar part of the primitively conduplicate
megasporophyll, Figs. 45 and 46, forming a tubular extension whose cavity
communicates with the locule, ig. 37, and is lined by a more or less
papillate and glandular appearing epidermis. The upper part of the style
is bifid, Fig. 13. Just below the level of forking, the style is typically
conduplicate and open on its ventral side. The ovules are anatropous and
have two integuments, the outer being thicker, Fig. 44.

POLLEN

The wall of a young microsporangium of A. maculata in sectional view,
Fig. 39, shows an epidermis subtended by three layers of wall cells which

1949 | BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 219

are subtended in turn by a single, or an irregularly 2-layered, tapetum.
When the microspore mother cells are undergoing reduction divisions, the
tapetal cells become binucleate, Fig. 40. During later stages, the walls
and the cytoplasm of the tapetal cells become homogeneously granulate
and lose cellular organization, Fig. 47. However, neither an amoeboid
shape of the individual cells nor their migration into the cavity of the
sporangium, as often happens with the periplasmodial type of tapetal
organization, was seen. The nutritive layer is soon absorbed after the
stage represented in Fig. 41.

The microspore mother cells undergo reduction divisions in a simul-

This feature facilitates observation of the differentiation of the germinal
furrow on the distal face of the grain. The generative cell is cut off towards
the furrow-end of the grain and the pollen grains are two-celled at the
time of shedding.

The mature pollen grain is spherical with a single germinal furrow
running nearly from pole to pole. The exine is thick and appears to be
finely pitted, occasionally 3-5 pits being arranged in variously curved
lines. The external surface of the furrow is sparsely flecked with minute
protuberances, Text-fig. 3

VASCULARIZATION OF “THE “Be OW

The following observations upon the vasculature of the flowers of
Austrobaileya are based largely upon preserved specimens collected by
Mr. Stephens. They have been supplemented, however, by an examina-
tion of cleared carpels and stamens from herbarium specimens of 4.
scandens, Kajewski 1269, and A. maculata, White 10734.

At the base of the pedicel, there are 10-14 vascular bundles arranged
in four independent aggregations, Text-fig. 4. The parenchymatous
lacunae between these aggregations of bundles correspond with the decus-
sate arrangement of bracts on the pedicel, i.e. four orthostichies. Each
bract receives a pair of strands, the individual members of the pair
arising from independent systems of the eustele, Text-fig. 4, b1-08.

The vasculature of the tepals is similar to that of the bracts. The
distinctness of the individual members of the pair of strands is particularly
clear in the bases of the outer tepals, ¢7-t10,; in the inner tepals, one of
the strands of the pair is either feebly developed, #11, or altogether absent,
t14, 15. Within the tepal, the two strands fuse and then split into a
large number of veins that vascularize the lamina

The stamens and staminodia may show at their point of attachment
either two distinct vascular strands, or a pair of closely approximated
strands, or a single strand. Usually in the outer functional microsporo-
phylls, the strands are distinctly double at least at the base, stu I-stn 6,
while in the inner androecial members, a section at a corresponding level
usually shows a single strand, std 5—-std 11. However, when this strand

220 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

is followed downwards in the axis, it frequently is formed by the fusion
of two bundles from independent systems, std 9, std 10, etc.

When a single vascular strand enters the microsporophyll, it soon tri-
furcates, the lateral strands again splitting at higher levels. On the other
hand, when two strands supply the microsporophyll, they dichotomize at
the base of the structure and the middle two branches fuse; the lateral
branches also bifurcate at a slightly higher level, Text-fig. 5. In both
instances, the strands remain close together as far as the level of the apexes
of the sporangia and then spread in a fan-shaped manner.

In the case of flower buds collected by Mr. Stephens, the exact number
and the behavior of the strands that enter the carpels could not be deter-

cvl om

Text-FiGuRE 4, Pattern of vascularization of a flow The axis is
shown as if longitudinally cut on one side and spread in one > plane, b1-b8,
bracts; t2-t15, tepals; stn I-stn 6, stamens; std 1-std 13, staminodia; C [-C VI,
carpels. The upwe ard terminations of strands correspond sinned to
the level of insertion of the respective appendage.

1949] BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 221

mined with certainty. Even in the oldest bud examined, Text-fig. 4,
some of the vascular strands were still in the process of differentiation
(indicated by stippled lines in the figure) and it is not possible at such a
developmental stage to determine their true relationship either to the floral
axis or to the carpels. However, in the buds examined by us, at least
three vascular strands were. well differentiated in most of the carpels.
These strands occupied positions corresponding to the median and ventral
veins. The median vein in some cases, C J, showed distinct doubleness
at the point of entry into the carpel and when traced downwards, the two
halves fused with two independent systems of bundles in the axis. In
other carpels, the two strands gave indications of ontogenetic fusion,

ie

a
as

eeneee
rail
fies We eararnio® 4

Bee abe

x
9 te

oe
Fo.

ee Fae
£

5

Text-FicurEs 5, 6. Fig. 5. Vascularization pattern of microsporophyll ;
broken lines donee the position of the sporangia. Fig. 6. Vascularization
pattern of carpel; broken line denotes the extent of the locule. Vasculariza-
tion is shown in one longitudinal half of carpel which is cut in the plane of
conduplication.

222 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XxX

C Ill, C IV. In still others, the median strand had lost its paired appear-
ance and its true double nature was demonstrable only far below in the
axis, C H,C VJ.

In the case of carpels from herbarium specimens, Kajewski 1269 and
White 10734, the vasculature appears to be well developed at anthesis.
Three of such carpels showed a distinctly double median strand at the
base, the halves of which fused at a higher level in the “stipe.” In others,
the corresponding strand was single. Along its course through the locular
part of the carpel, the median strand sends out numerous branches laterally
in a pinnate manner and the branches form a diffuse net-work within the
carpellary wall; the ends of the branches usually anastomose with the
ventral strands, Text-fig. 6. The extra strands that often enter the carpel
from the axis also anastomose with the system of the median strand. The
ventral strands supply the ovular traces and extend to varying distances
in the style along with the median strand.

Taking the vascular pattern of the flower as a whole, it appears that
although there may be considerable variation among the different ap-
pendages of the same flower, and even among those of the same whorl,
the paired condition of the median strand is fundamental. In those in-
stances where the strand appears single, it is often possible to demonstrate
its double nature at lower levels of the floral axis. In other words, the
doubleness of the median vascular strands of the floral appendages, as in
the case of the vegetative leaves of Austrobaileya, is not a result of dichot-
omy of a single bundle, but of the approximation of two distinct strands
originating from two separate systems of bundles.

DISCUSSION

The genus Austrobaileya has monocolpate pollen. This structural type
of pollen occurs in many seed ferns, Bennettitales, Cycadales, Ginkgoales,
monocotyledons and in certain families of the Ranales (sensu latu), Bailey
and Nast (1), Swamy and Bailey (13). The rest of the dicotyledons,
including the Dilleniaceae, have tricolpate pollen or phylogenetically modi-
fied types of such pollen. Thus, the occurrence of monocolpate pollen
in Austrobaileya, coupled with the presence of ‘‘ethereal oil cells” and the
absence of raphides, precludes any close relationship of the genus to the
Dilleniaceae.

Among families of general ranalian affinities which have monocolpate
pollen (or phylogenetically modified types of such pollen) all have ‘“‘ethereal
oil cells” of a characteristically similar type with the exception of the
Cabomboideae and the Nymphaeoideae of the Nymphaeaceae. These
secretory cells and monocolpate pollen occur in association in Avustro-
baileya which indicates that the genus belongs among this complex of
families rather than among ranalian families having tricolpate pollen, viz.
Trochodendraceae, Tetracentraceae, Eupteleaceae, Iliciaceae, Schisandra-
ceae, Cercidiphyllaceae, Ranunculaceae, Berberidaceae, Lardizabalaceae or
Menispermaceae.

1949 ] BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 223

The nodal anatomy of Azstrobaileya is of a unilacunar type. Among
ranalian families, having monocolpate pollen (or modified forms of such
pollen) and “ethereal oil cells,” this form of nodal structure occurs in the
Monimiaceae, Gomortegaceae, Lauraceae, Hernandiaceae, Lactoridaceae
and Chloranthaceae (Ascarina and Hedyosmum). On the contrary, the
Winteraceae, Degeneriaceae, Himantandraceae, Magnoliaceae (sensu
stricto), Annonaceae,* Eupomatiaceae, Myristicaceae,* Canellaceae, Piper-
aceae, and Saururaceae have dominantly trilacunar or multilacunar nodes
or anomalous nodal structures, e.g. the Calycanthaceae.

Not only does Austrobaileya differ from the Winteraceae, gains
Annonaceae and other more or less closely related trilacunar and multi-
lacunar ranalian families in its nodal anatomy and in the vascularization
pattern of its leaves and floral appendages, but also in the structure of
its xylem and phloem, its parenchymatous and sclerenchymatous tissues
and in the form and internal structure of its floral appendages. The
totality of general morphological, and of specific anatomical, evidence is
not indicative of relationship to any of these families (i.e. except distantly
as members of a common ranalian grouping), but rather to the assemblage
of unilacunar families enumerated in the preceding paragraph.

mong these unilacunar families, the Monimiaceae are the most highly
diversified, exhibiting numerous trends of morphological and anatomical
specialization, in both their vegetative and reproductive parts, and afford-
ing significant clues regarding phylogenetic changes that occurred in the
evolution of the Gomortegaceae, Lauraceae, and Hernandiaceae.

Austrobaileya is a vine and there is a scandent tendency in a number
of the Monimiaceae. The simple, pinnately veined leaves of Austrobaileya,
as of the Monimiaceae and Chloranthaceae, are opposite or sub-opposite
and exhibit a typical decussate phyllotaxy at least during the earlier stages
of the development of a shoot. The leaves of Austrobaileya, as of Trimenia
and Piptocalyx of the Monimiaceae, Lactoris of the Lactoridaceae and
Ascarina of the Chloranthaceae, are vascularized by two strands that are
independent at the nodal level of the stem and are attached at lower levels
to independent parts of the eustele. The stomata of Austrobaileya, as of

Monimiaceae, are transitional between “syndetocheilic” and
“haplocheilic” appearing, and have a similar pattern of internal cuticulariza-
tion, Money (7). The xylem of Austrobaileya, as we have shown, figs.
5-7, has numerous transitional stages in the development of vessels from
tracheids with unusually large bordered pits. Such imperforate tracheids
occur in the vesselless xylem of Amdborella, Bailey and Swamy (4).
Furthermore, there are numerous transitions between tracheids and septate
fibres, such as are characteristic features of many Monimiaceae. In
Addition: calcium oxalate is deposited in Austrobaileya, as in the Moni-
miaceae and the Lauraceae, in the form of numerous minute crystals, viz.
in the form of “crystal sand” rather than as conspicuous druses or large,
single, rhombic crystals.

* Mistakenly reported as unilacunar by Sinnott (10).

224 JOURNAL OF THE ARNOLD ARBORETUM [voL. xxx

The totality of available evidence from all vegetative parts of Austro-
baileya provides no significant evidence for excluding the genus from the
Monimiaceae as constitued by Perkins (8), particularly if such genera as
Amborella, Trimenia and Piptocalyx are retained within the family. It is
the flowei of Austrobaileya and specifically its multiovulate carpels that is
the chief ubstacle to such a procedure. It should be noted in this connec-
tion, however, that there is such a wide range of morphological variability
in the stamens, staminodes and carpels of the Monimiaceae that excessive
emphasis should not be placed upon differences in the external forms of
these organs in Austrobaileya. Furthermore, the tribes Hortonieae and
Trimenieae were obviously established for florally more primitive repre-
sentatives of the family, specifically for genera having free perianth mem-
bers and a broadly convex or only slightly concave torus. Therefore, it
would not appear to be entirely illogical to include in the family a genus
which has retained broad microsporophylls and carpels in which the
ovules have not been reduced to one, particularly as the vesselless
Amborella is considered to be a member of the family.

As stated by us in a previous paper, Bailey and Swamy (4), if such
genera as Trimenia, Piptocalyx and Amborella are to be retained in the
Monimiaceae then the present concept of the family could be broadened
to include Gomortega and Austrobaileya. Conversely, if a narrow con-
cept of the family is to be maintained, then Trimenia, Piptocalyx and
Amborella should be excluded, as well as such genera as Gomortega and
Austrobaileya. However, before such decisions are attempted, it is essen-
tial to assemble more comprehensive and reliable information regarding
Gomortega, Austrobaileya and the various genera of the Monimiaceae.
As stated by White, in passages quoted in the introduction to this paper,
it is essential to study the fruits and seeds of Austrobaileya before attempt-
ing to segregate the genus in an independent family. Mr. Stephens is
attempting to obtain fruits and seeds for us, and, in addition, material
suitable for critical embryological and cytological investigations. The
separation of Amborella in an independent family, the Amborellaceae,
Pichon (9), upon the basis of our preliminary note (3), without waiting
for the publication of our detailed investigation (4), was premature and
singularly unfortunate since erroneous statements regarding the ray
structure and orthotropous character of the ovule were included in the
brief description of the new family.

ACKNOWLEDGMENTS

We are indebted to the New York Botanical Garden for the loan of
isotypes of Austrobaileva scandens, to Mr. C. T. White for material of
A, maculata, and Mr. L. J. Brass and Mr. S. E. Stephens for their sus-
tained efforts to obtain adequately preserved material of Austrobaileya,
to Mr. Gualterio Loaser and Professor Augusto Pfister for their kindness
in obtaining material of Gomortega and to Professor Katherine Esau for
notes on the occurrence of companion cells in several critical genera.

1949] BAILEY & SWAMY, MORPHOLOGY OF AUSTROBAILEYA 225

LITERATURE CITED

1. Bartey, I. W. & C. G. Nast. The comparative morphology of the
Winteraceae, I. Pollen and Stamens. Jour. Arnold Arb. 24: 340-

346. 1943

2, ———. & ———. The comparative morphology of the Winteraceae,

> II. Carpels, Tous. Arnold Arb. 24: 472-481. 1943,

oy G. L. Swamy. Amborella trichopoda Baill. A new type
of vesselless dicotyledon. Preliminary note. Jour. Arnold Arb. 29:
215, 1948.

A 2 & — — —. “Amborella trichopoda Baill) A. new morphological
type of Weseellcss dicotyledon. Jour. Arnold Arb. 29: 245-253. 1948.

5. Croizat, L. Notes on the Dilleniaceae and their allies: Austrobaileyeae
sub-fam. nov. Jour. Arnold Arb, 21: 397-404. 1940

——. New families. Jour. Cactus and Succulent Soc. Amer. 15:

64. 1943,

7. Money, Littian L, The anatomy and morphology of the Monimiaceae:

leaf, node, stem, pollen. Doctoral Dissertation, Widener Library,
Harvard URiGestan Jan. 24, 1949.
8. Perxins, J. R. Monimiaceae. Pflanzenr. 4 (IV. 101. Nachtr.). 1911.
9. Picuon, P. Les Monimiacées, famille hétérogéne. Bull. Mus. Parts.
20: 383-384. 1948.
10. Sinnott, E. W. Investigations on the phylogeny of the angiosperms.
_ The anatomy of the node as an aid in the classification of angio-
sperms. Amer. Jour. Bot. 1: 303-322. 1914
is W Hope Cer = igneous ae collected for ‘ie Arnold Arboretum in
North Queensl ond by S. F. Kajewski in 1929. Contrib. Arnold Arb.
4: a 1933.
2. ———. A new species of eae Cae pailevaceae) from Aus-
tralia. Jour, Arnold Arb. 29: 255-256, 1948
Se Swan. Go ls & leWe aa EY. The ene and relationships
of Cercidip! hyllum. Jour. Arnold Arb. 30: 187-210. 1949.

226 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

EXPLANATION OF PLATES

PLate I. A. scandens, Brass 18160. Fig. 1. Part of a transverse section
of a large stem, x 20. Fig. 2. Secondary phloem of same more highly mag-
nified, X 130.

Pate Il. A. scandens, Brass 18160. Fig. 3. Secondary xylem of Fig. 1
more highly magnified, x 130. Fig. 4. Tangential longitudinal section of
the secondary xylem, x 50.

LATE TIL. A. scandens, Brass 18160. Figs. 5-7. Scalariform perfora-
tion plates of three vessels of different diameter, x 5

Pirate IV. A. maculata, Stephens’ collection. Figs. 8, 9. Flower buds,
x 3. Fig. 10. Same, as seen from the stalk-end, x 3. Pig. ti, Group of
stamens and nae inner appendages after the removal of the perianth, x 5.
Fig. rroup of carpels after the removal of the other floral appendages,
a. gs. de, oe x 12. Fig. 14. Same, longitudinal section passing
through the plane of conduplication, to show the “stipe,” locule and ovules,
x 22. Figs. 15,16. Adaxial and side views of fertile microsporophyll, x 8.
Figs. 17-19. Sraminodia: in Fig. 17, the structure bears degenerate spo-
rangia, X 8. Figs. 20-32. Per a members of a single dlavier: The order
from the outermost to the innermost runs according to the serial numbers, x 1.

V. fig. 33. A. scandens, Kajewski 1269. Stamen (middle) and
staminodia (right and left), x 5. Fig. 34. A, maculata, White 10734. Sta-
mens (right and left) and staminodia (middle), x 5.5.

PLate VI. A maculata, Stephens’ collection, Fig. 35. Transverse sec-
‘tion of a flower bud at the apical region, x 50. Fig. 36. Same, at a central
level, x 50.

Pirate VII. A. maculata, Stephens’ collection. Fig. 37, Longitudinal sec-
tion of two carpels, the left hand one cut in the plane of conduplication, and
the right hand one, at right angles to the previous plane, x 50. Fig. 38. Lon-
gitudinal section of the floral apex, x 50: Figs. 39-41. Structure of the
microsporangium and tapetal organization, x 130. Fig. 42. A portion of the
epidermis of the microsporophyll, showing on histology of the pigment cells,

Figs. 43-46. Transverse sections of a carpel cut at successi nee
higher levels, x 100

INSTITUTE FOR RESEARCH IN GENERAL PLANT MorPHo_ocy,
Harvarp UNIVERSITY.

*
en a

a
“53
-'

*
wat’
en

4a? .
ow.
hie

BatrEy &

Swamy, MorpHoLocy or AUSTROBAILEYA

oo : eet pe
i &
; a eo
iat
Ae
Se A cS)
at ‘
ut &

#
ha

7

’ 7)
‘ }
A

©.
bn
4
y

at meet see =

XXX IOA ‘Say CIONTY ‘ano

[ FLVIg

. _ if
an,

255

Ce
,
=
tale
eomragea,

Pe Sie)
hy ty 8

Cf

a

>. a
@u®,
A oben

Bey ge

‘4
oe
1)

BaItEy & Swamy, MorpHorocy oF AUSTROBATLEYA

ss ‘ istapaseeen eae
* ny ROR RI ERM, SETS eer
ale sarin - :
. oe
iy
nay
a gl
x

fo en -
cx weecx
aan ®

XXX “I0OA ‘auy aIONuy ‘anof

II M1vIg

Jour. Arnotp Ars. VoL. XXX

PuaTtE ITI

MorRPHOLOGY OF AUSTROBAILEYA

BartEy & Swamy,

Jour. Arnotp Ars. VoL. XXX PLATE IV

24 ek 29) “30

BaILtey & Swamy, MorpHoLoGy or AUSTROBAILEYA

Sl. S2

FULL-TONE—MERIDEN

Jour. ARNOLD ArB. VoL. XXX PLATE V

BatLtEy & Swamy, MorPHOLOGY OF AUSTROBAILEYA

Jour. ARNOLD ArB. VoL. XXX PraTE VI

BaILtEy & Swamy, MorRPHOLOGY OF AUSTROBAILEYA

Pirate VII

Jour. ARNoLD ARB. VoL. XXX

BatLtEy & SwaMy, MorPHOLOGY OF AUSTROBAILEYA

——
:

OO ae
Tse
Pia =

: 7

: : : : _
OS a 7 a a ee -
= 7 7 — —— OO
a 7 -
ee oo Oo oe -

i aes

1949] McKELVEY, A NEW AGAVE FROM ARIZONA 227

A NEW AGAVE FROM ARIZONA
SUSAN DELANO MCKELVEY
With two plates

On May 15, 1934, while visiting the Kaibab Plateau on the northern
side of the Grand Canyon of the Colorado, Coconino County, Arizona,
the writer first saw the Agave here described. Two plants had been col-
lected by the Park authorities at some previous date, but their leaf-rosettes
and long fruiting stalks were still intact. It was not possible to make the
descent into the gorge, but a few scattered plants were discernible some
distance below the rim. Flowers were promised at a later date but were
never forthcoming, unfortunately.

The two plants are shown in the included photograph (McKelvey
152-12); although one was considerably smaller than the other it was
undoubtedly the same species. The dimensions are from the larger plant.
Agave kaibabensis, spec. nov.

A A. utahense differt habitu solitario, non caespitoso; foliis basi rosulatim
confertis, longioribus, latioribus; inflorescentiis elongatis (4.4—7.5
longis), basi ramiferis.

Leaf-rosette 1.4 m. in breadth, 0.7 m. in height, nearly hemispherical,
with ascending and spreading ann leaves which attained 0.4 m. in
length (base included). Leaf-blade concavo-convex for the entire lenge
usually tapered from the point of union with the base to the apex, some-
times constricted above this union, then gradually and slightly broadened
to a point near the middle, thence tapered to an acuminate apex, smooth
on both surfaces; terminal spine 1-1.3 cm. in length, channeled on the
upper surface and decurrent along the margins of the blade for 10-15
cm.; prickles along the leaf-margins irregularly spaced at intervals of
0.7-2 cm., small, 1.6-3.2 mm. in length, curving either backward or
forward following the blade-margins. Leaf-base about 15 cm. in breadth
at the point of union with the stem, gradually narrowed to the point of
union with the blade where about 6.5 cm. in breadth, about 10 cm. in
median length. Inflorescence 4.4 m. in length over all; scape 1.4 m.,
inflorescence proper ca. 3 m.; the last, while spicate in general appearance,
was narrowly ovoid-cylindrical, branched for the entire length, the branch-
lets longer and more complexly branched below than above; branch-
lets near the base 10-12 cm. in length, twice forked; those above gradually
reduced in length and finally at the top of the inflorescence proper simple.
Capsules produced in great numbers, 2.5—3 cm. in length, nearly ovoid
with an acute apex and short-stipitate base, with slightly glossy, red-brown
epidermis; pedicels attaining 0.7 cm. in length, enlarged at the point of
union with the capsule, terete; seeds extremely abundant, small, black;

228 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

bracts near the base of the scape 18-20 cm. in length, only slightly reduced
in length below the inflorescence proper, at insertion 2.5—3.8 cm. in breadth,
tapered from this point to the apex, the upper portion more or less concave,
the lower portion flat, with a short, very sharp, channelled red spine,
tenaciously attached to the scape, eventually reflexed.

ARIZONA: Coconino County, Kaibab Plateau on the north side of the
Grand Canyon of the Colorado River, McKelvey 43817 (Arnold Arboretum,
TYPE), 4381A (Arnold Arboretum), May 15, 1934.

At first glance the plants were thought to be Agave utahensis
Engelm., in colossal form — that species is common in northwestern
Arizona and is said to occur plentifully in the Grand Canyon although
the writer has not happened to see it growing there. But examination
disclosed marked differences in the individual leaf, the leaf-rosette, to some
extent in the inflorescence, and, important, in the habit of the plant.

When examined the leaves of the new species were dry and their margins
were considerably incurled but must once have been much flatter; in form
and size and in the straight, spreading habit they came close to what is
found in the so-called century plant or mescal — with paniculate inflores-
cence —and the rosette in size and near-hemispherical form also. The
leaves of Agave utahensis and of those of its allies with prickles along the
leaf-margins (and producing spicate or subspicate inflorescences) show a
marked tendency to bend inward towards the center of the head, pro-
ducing a spherical or near-spherical rosette. This tendency to spreading
or to curving inward is not altered with the death of the plant; if anything
it becomes more pronounced with drying in each instance. The leaves are
also much smaller, shorter, and narrower.

Agave utahensis shows a tendency towards a panicled inflorescence.
This has been noted by the writer more than once and has been recorded
in the literature. Miss Mulford (Agaves of the United States, 77, 1896)
states that that species “. . . so frequently approaches the paniculate
character, that it might be placed in the Euagave section with almost as
much propriety as in that of the Littaea.”” In the case of the new species
the branching is more complex and not confined to the lower portion of the
inflorescence proper as is usual in the better known plant.

In habit the two would seem to differ radically. The two plants of
Agave kaibabensis which were examined showed no tendency to produce
offsets, while those seen from afar in the gorge were obviously solitary
rather than of clump-habit. Agave utahensis is caespitose, in all instances,
often producing large and involved clumps.

Park officials said that near the base of the gorge flower-stalks attained
7.5 m. in over-all length; while easy to overestimate, yet, from the large
specimen seen, a greater size does not seem improbable. It is not unique
to find a plant growing on the walls of the gorge but not appearing above
the rim — it has been observed in one species of Yucca certainly. It was
said that the Piute Indians called the plant “Oose”’ (this spelling open to
correction); the local name was ‘‘Yant,” the stalk of the inflorescence
called a “Yant-stick” and used to prod cattle.

1949] McKELVEY, A NEW AGAVE FROM ARIZONA 229

Unsuccessful attempts were made to locate the plant in adjacent regions,
and while, from information obtained, it may occur near the lower portion
of Kanab Creek where this empties into the Colorado, it is possible that
the name ‘‘Yant” may be applied as well to Agave utahensis.

It is believed that photographs of Agave kaibabensis have appeared on
three occasions: Once in an unsigned article in Garden and Forest (8: 384,
fig. 53. 1895) as Agave utahensis; but the solitary rosette is far more
suggestive in every respect of the plant under discussion. The short scape
of the photograph would be abnormal for either species, — it may have
been cut for inclusion, but one cannot be sure from the reproduction. The
article notes the leaves as “. . . twelve or fourteen inches long and from
two to three inches broad... .” and “. . . the slender spike . . . as three
feet long, and often raised on a stem seven or eight feet high.” This plant
grew on “. .. the bluffs and cliffs which rise from the slopes of the Grand
Canon...” (while not entirely clear this seems to indicate the Canyon
walls); it is not stated whether on the north or the south side of the river.

Next, two photographs appeared in Mr. Harold E. Anthony’s interest-
ing article published in Natural History (40: 719, 1937), describing the
exploration of Shiva Temple, a near-island in the Canyon connected by a
saddle with the north rim. The first photograph, showing only a rosette
—one which might well belong in the Euagave section —is captioned:
“The Mescal of Shiva is a variety of century plant with stiff, thorn-tipped
foliage. The young plants grow in a compact clump.” The word “clump”
as used here probably refers to the dense rosette — it shows no tendency
to suckering. The second photograph is captioned: ‘“When mature the
mescal, or ‘yant,’ sends up a tall flower stalk twelve feet or more. The
Pueblo Indians used the starchy center of this stalk for food, baking it
in earthen pots.” The rosette is not visible in this picture, but one sees a
tall inflorescence with scape longer than the flowering portion and with
the inflorescence proper considerably broader below than above, although
the general effect is spike-like.

The third picture of this plant has recently (1945) appeared in This
is the Place: Utah, by Maurine Whipple. The caption is The Grand
Canyon of the Colorado, the plate on page 204.

The name Agave kaibabensis indicates the plant’s habitat as now known.
The smaller plant of the picture is McKelvey 4381A. The plant, as noted,
came originally from some distance below the rim on the north wall of
the Canyon. The date of actual collection is not known; it must have
been in the tourist season of 1933; the writer made her specimens on May
15, 1934. Her photographs 152-11, 12, and 153-1, 2 are of these plants;
reproduced here is 153-2 (Pl. I).

Included, to show the different appearance of its rosettes and leaves, is
a photograph (McKelvey 82-1) of Agave utahensis (Pl. I1); the speci-
men (McKelvey 1655) was collected March 31, 1929, some 13 miles east
of Peach Springs, Mohave County, Arizona, in the extreme northwestern
corner of Yavapai County. Over much of this region the plant is plenti-
ful, reaching certainly as far southwest as Valentine, Mohave County.

230 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

The tendency of the leaves to bend inward towards the center of the
head and so form a near-spherical rosette, and the caespitose habit of
the plant are apparent. The six-inch rule indicates the small size of leaf
and leaf-cluster. Miss Mulford (1. c.) described the leaves of Agave
utahensis as “. . .linear-lanceolate ... 12 to 17 cm. long, 2—2.5 cm. wide,
or larger... .” and its inflorescence thus, “... scape 15 to 24 dm. high...
upper 3 to 6 dm. floriferous; panicle narrow . . .”

EXPLANATION OF PLATES
Piate I. Agave kaibabensis McKelvey, showing the near-hemispherical
rosette of leaves and the long-branched inflorescence. Collected on the
Kaibab Plateau, on the north side of the Grand Canyon of the Colorado
River, Coconino County, Arizona.
. Agave utahensis Engelm., showing caespitose habit. Collected
about thirteen miles east of Peach Springs, Mohave County, Arizona.

ARNOLD ARBORETUM,
HarvVARD UNIVERSITY.

PLATE I

Jour. ARNOLD ARB. VoL. XXX

KAIBABENSIS MCKELVEY

E

AGAV

Jour. ARNOLD ARB. VoL. XXX

PLATE II

AGAVE UTAHENSIS ENGELMANN

1949 ] LI, A NEW GENUS OF THE ARALIACEAE 231

A NEW GENUS OF THE ARALIACEAE
Hur-Lin Li

ON THE ISLAND OF TAIWAN (ForMosA), there is an anomalous species
of the family Araliaceae known as Oreopanax formosana Hayata. Oreo-
panax is otherwise an exclusively tropical American genus. Upon careful
study of the ample number of specimens preserved in Formosan herbaria,
I have come to the conclusion that a different genus is represented. Not
being able to refer it to any existing genus, I propose a new genus under
the name Sinopanax, in recognition of its geographical location.

The specimens cited below are those preserved in the Herbarium of
the National Taiwan University except Wilson 9734, which is deposited
in the herbarium of the Arnold Arboretum.

Sinopanax, gen. nov.

Arbuscula vel frutex, ramis cinereo-stellato-tomentosis vel adpresse pilo-
sis inermibus. Folia ampla simpla longe petiolata, petiolis basi dilatatis,
stipulis parum prominulis ad basin petiolorum connatis, laminis late
orbicularibus obsolete 3—5-lobatis vel grosse irregulariter dentatis, subtus
stellato-tomentoso et piloso dense obtectis. Inflorescentia ampla corym-
boso-paniculata terminalia laxa erecta aperta, ramis alternis angulo obtuso
divaricatis. Florum capitula sublobata, capitulis florum pedunculatis
remote instructis, bracteolae florum sub singulo flore 3-nae, 1 subtendente
majore, 2 lateralibus minoribus oppositis. Calyx margine minute dentatus,
basi exarticulatus. Petala 5, aestivatione valvata, caducissima. Stamina
5, filamentis brevissimis; aiscus explanatus. Ovarium 2-loculare, loculis
1-ovulatis, stylis 2 brevibus liberis erectis, stigmatibus terminalibus.
Fructus late globosi, abortu 1-spermi dhupacer stylis persistentibus valde
recurvis. Semina ovoideo subtriquetra, albuminibus ruminatis.

Type SPECIES: Oreopanax formosana Hayata.

One species, Taiwan (Formosa), China.

Sinopanax formosana (Hayata), comb. nov.

Oreopanax formosana Hayata, Fl. Mont. pea: a oe 14, 1908, Icon.
Pl. Form. 2: 61. 1912, Gen. Ind. Fl. Form. 33. 19 Kanehira, Form.
Trees 280. 1917, rev. ed. 524, f. 484. 1936: oe in 1 Jour. Arnold Arb.
5: 19. 1924; Masamune, Short Fl. Form. 155. 1936.

CHINA: Taiwan: Kao-hsiung, Chi-shan, Susuki-Tokio 20930; Ako-
huzi, E. Matuda 1404; Karenko, Nankotaizan, Susuki-Tokio, N. Fukuyama,
& H. Shimida 17547; Hannoki, Heito, S. Suzuki 11127; Arisan, Usyuoko,
S. Suzuki s. n.; Ako, Mt. Buwi, S. Sasaki s. n.; Arisan, E. H. Wilson 9734
(Arnold Arboretum).

A small tree or shrub, in dry sunny places, on ae or in open
forests at high altitudes of over 2000 meters. Endem

For Oreopanax formosana, Hayate originally ane five collections:
N. Konishi 344A; G. Nakahara s. n.; T. Kawakami & U, Mori 1709,

232 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

1871, and 1914. Among these, three are represented by duplicates in
the Herbarium of the Forestry Institute of Taiwan and carry the herbarium
numbers as follows: Kawakami & Mori 18506, Mori 18505, and Konishi
18511. These have also been seen.

Oreopanax is an araliaceous genus of the tropics of the western hemi-
sphere. Two species have been credited to eastern Asia, the only records
of the genus outside America. One of them, Oreopanax chinense Dunn
(Jour. Linn. Soc. Bot. 35: 500. 1906) is a species with digitately compound
leaves, described from southern Yunnan. It belongs to the group of
species of Schefflera with capitate flowers and is now known as Sche fflera
chinensis (Dunn) Li (Sargentia 2: 17, 1942). Another species, Oreopanax
formosana Hayata, a species with simple irregularly and shallowly lobed
leaves and described from Formosa, seems to represent a distinct genus.
The occurrence of O. formosana Hayata in Formosa has often been referred
to as a very anomalous case of distribution in the flora of the island, the
species being the only representative of the otherwise exclusively tropical
American genus. This view thus needs revision.

This Formosan species, here designated as representing the type of a
distinct genus, Sinopanax, resembles Orcopanax in the simple leaves, sessile
and capitate flowers, ruminate endosperm, and the presence of three bracts
under each flower. However, there are some fundamental differences
between the two. The ovary of Oreopanax is generally 5-celled, and the
flowers are polygamo-dioecious or more rarely polygamo-monoecious. In
this Formosan plant the ovary is 2-celled and the flowers are hermaphro-
ditic. There are two styles, and these are so extremely short that the
stigmas can actually be called sessile. In Oreopanax the style is rather
long and bears a flat stigma.

This new genus of Formosa seems to be related to, but evidently not
congeneric with Schefflera. The species of Schefflera have hermaphroditic
flowers. Occasionally simple palmately lobed leaves and capitate flowers
characterize some of its species. Schefflera, however, differs from the
Formosan plant in having a 5—7-celled ovary and uniform or very rarely
slightly ruminated endosperm. The relationship of this Formosan genus is
probably closer to Brassaiopsis than to Schefflera. In Brassaiopsis the
ovary is 2-celled, the leaves are either palmately lobed or digitately com-
pound, and the endosperm is either ruminate or uniform. But in Brassai-
opsis the flowers are polygamous and not capitate, and the two styles, either
long or short, are distinctly united. Sinopanax is also distinct from both
Schefflera and Brassaiopsis in the presence of three small bracts, one larger
than the other two, under each flower.

It thus seems that this Formosan plant is not acceptable to Oreopanax,
Schefflera, or Brassaiopsis. Nor can it be assigned, to the knowledge of
the writer, to any other tropical Asiatic genus of the family. In order to
ascertain its phylogenetic and geographic significance, it is thought best
to designate a separate genus for its accommodation.

NATIONAL TAIWAN UNIVERSITY,

Tarper, TAIwan, CHINA,

1949] HU, THE GENUS ILEX IN CHINA 233

THE GENUS ILEX IN CHINA *
SHIU-YING Hu
With nine text-figures
INTRODUCTION

IN THE SPRING OF 1943, with Prof. W. F. Chu, Head of the Department
of Forestry, College of Agriculture, University of Nanking, and a class
of Agricultural students of Chuan-kang College of Agriculture and Engi-
neering, I made a collecting trip to the charcoal manufacturing area at the
White-water River, Pun-hsien, Szechuan, China. In the dense mixed forest
I found an attractive dioecious evergreen shrub with large dark green
lanceolate leaves, shiny cherry-red fruit, and chocolate-colored small flow-
ers. The beautiful combination of colors, the complete representation of
staminate and pistillate flowers and fruit, the rarity of the plant in the
forests of West China, together created in me a strong interest in the plant.
In my ignorance of the Aquifoliaceae I mistook it for an /lex presumably
a new species. Hence, I collected many sets of the plant, treasuring it
highly and holding it for the time when I could identify it.

In the summer of 1944 I was asked to represent the Chengtu Chapter
of the Science Society of China at its twenty-third annual meeting in
Pai-pei, Chungking, China. I took this opportunity to be my precious
chance to identify the specimens that had been accumulating in the her-
barium of the West China Union University, the institute where I was
working. Those specimens which I considered unusual I carried along,
hoping that I might identify them in the well-equipped herbaria of the
Biological Laboratories of the Science Society of China and that of the
Academia Sinica, two neighboring refugee institutions then at Pai-pei.

After the meeting I stayed on to work at the Biological Laboratories,
and starting with the White-water River shrub I worked through the Aqui-
foliaceae, but was doomed to disappointment. First, there was nothing
that would match my specimen, and secondly, the Aquifoliaceae in that
herbarium were in a discouraging condition. About half of the mounted
sheets were not named. With three months of vacation ahead of me, and
at ihe suggestion and with the encouragement of Prof. 5S. S. Chien, Acting
Director of the Biological Laboratories, and others on the staff, I made a
start in the clarification of the Chinese Aquifoliaceae. At that time I could
not foresee the trouble that lay before me.

Life during those three months will provide me always with precious
memories. I boarded in a mediaeval-type bungalow and worked on the

* A thesis submitted to the Graduate School of Radcliffe College in partial fulfill-
ment of the requirements for the degree of Doctor of Philosophy.

234 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

top floor of a poorly built and unshaded building under the blistering sum-
mer sun in one of the two hottest areas of China. In this environment I
embarked on the study of the Chinese Aquifoliaceae. Though living in
the age when atomic energy had been utilized by man, my equipment was
no better than that used by Linnaeus. All that I had, besides the
specimens and literature, were a hand-made tripod-lens, a ruler, two self-
fashioned dissecting needles, and at night a rapeseed-oil lamp lit by three
pieces of pith from Juncus effusus. My study gradually became so intensi-
fied that I found myself working fourteen or sixteen hours daily. Perspi-
ration and the blood of crushed mosquitoes punctuated my three hundred
pages of manuscript.

On my way back to Chengtu I stopped at the National Central Uni-
versity in Shih-pai, Chungking, then a refugee institution, and asked for
the aquifoliaceous material in the herbaria of the Department of Biology
of the College of Science and the Department of Forestry of the College
of Agriculture. In Chengtu I worked on the material of the Herbarium of
the University of Nanking (which had been transported to this place
during the war), and the Herbarium of the Department of Biology, College
of Science, National Szechuan University.

In the autumn of 1946 I came to the Arnold Arboretum to study tax-
onomy with Professor E. D. Merrill. One day I asked him about a topic
for my thesis. After a good smile he said, “Well, since you have delved
into the problem of //ex, you might as well go more deeply and come out
of it gallantly.” Hearing this, a false joy of holding a secret fell upon me,
for I then thought that I might use my old manuscript, emend it, make
some additions, and use it as a thesis. At that time I still considered
species to be definite entities and so readily separable one from another.

I started with the material in the Arnold Arboretum and the Gray Her-
barium. Dr. A. C. Smith, then Curator of the herbarium of the Arnold
Arboretum, borrowed the Chinese Aquifoliaceae from the U. S. National
Herbarium of the Smithsonian Institution for me. Prof. Merrill, on a
recent trip to Europe, procured either as a loan or a gift, isotypes, photos,
and fragments of types from the herbaria at Kew, the British Museum,
and the Rijksherbarium at Leiden. Later he obtained isotypes and other
authentic specimens from the herbarium of the Muséum National d’Histoire
Naturelle, Paris, France, the National Taiwan University of China, and
the Botanical Institute of the University of Tokyo, Japan. Through Pro-
fessor Johnston I received photographs and the loan of types from Cam-
bridge University and from the British Museum and later the Asiatic types
of /lex from the New York Botanical Garden. In September, 1948, Pro-
fessor Merrill made it possible for me to take a trip to the last mentioned
institution so that I might examine the Chinese Aquifoliaceae in the her-
barium, and as a result of this trip I borrowed about 400 sheets of Jlex for
critical study. As I examined more and more material, my confidence in
the classification of the Chinese Aquifoliaceae waned. ‘First, the study
weakened my concept as to the nature of species. So much variation

1949] HU, THE GENUS ILEX IN CHINA 235

appeared within a species, and such a gradual change from one species
to another was exhibited, that I found it very difficult to determine definite
Specific limitations. I found that my original manuscript prepared in
China was far too premature and that even Loesener’s classical monograph
on the Aquifoliaceae was, in part, based on characters far too superficial.

Besides those whom I have mentioned in the foregoing paragraphs and
all the curators of the herbaria that have facilitated my studies, I would
like to take this opportunity to express my special thanks to Professor
E. D. Merrill who assigned me the problem, who has watched all my
efforts, and has enjoyed with me the completion of the work; to Professor
I. M. Johnston, the present sponsor of my thesis, for helpful criticisms and
suggestions; to Dr. C. E. Kobuski, Curator of the herbarium of the Arnold
Arboretum, for his patience in answering all my questions and, in turn,
questioning me wisely, and for the pains he took in going over my manu-
script; to Professor Karl Sax, Director of the Arnold Arboretum, for all
the facilities he has provided in this study; and to Dr. Lily M. Perry,
Professor A. Rehder, and the whole staff of the Arnold Arboretum for
their willingness and readiness to give any help that a stranger in this New
Land might need. A large part of this work has been done under the
Julia George fellowship granted to me by Radcliffe College. To the admin-
istrators of that college, especially to Dean Bernice Brown Cronkhite, I
wish to extend my deep appreciation.

In the paper I have used the terms East China, Central China, West
China, Southwest China, and South China rather loosely. These names
were taken in the common sense, as they are used in general literature
and the spoken language. There is no definite geographical area assigned
to each. In order to give my reader a picture as to the locality to which
each name approximately refers, I find it necessary to define them. If one
takes Ichang as a center, a circle with a radius of 400 miles will cover
approximately the area between Lat. E. 105—-115° and Long. N. 127-133°.
This I term Central China. East of this area is East China and west of
this area is West China. North of Central and East China is North China.
South of Central and East China is South China. North of West China
is Northwest China and south of West China is Southwest China. North-
east China in general usage refers to Manchuria.

HislToRyY

In Linnaeus’ Species Plantarum (1753), one Asiatic species was de-
scribed, Jlex asiatica from “India.” Because of the very short description,
and because no type was preserved, its status cannot be determined. Lin-
naeus knew no Chinese species of the genus.

The first botanist to publish a species of Chinese Aquifoliaceae was John
Sims. In 1819 he described the staminate plant of /lex chinensis from a
living plant introduced into England. Fourteen years later Hooker and
Arnott (1833) described Ilex pubescens and Prinos asprella (Ilex asprella
[Hook. & Arn.] Champ.) from South China, and in 1850, seventeen years

236 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

later, Lindley and Paxton described two plants introduced by Fortune into
England and named them Jlex cornuta and Ilex microcarpa (= Ilex ro-
tunda Thunb. var. microcarpa |Lindl. & Paxt.| S. Y. Hu). About the same
time (1852), Bentham, working on the flora of Hongkong, described Ilex
cinerea, Ilex graciliflora, Ilex memecylifolia, and Ilex viridis from that
island. These species were treated again (1861) in his Flora Hongkong-
ensis. Maximowicz (1881) in his monographic work on J/lex treated
twenty-two eastern Asiatic species, many of which occur in China. Forbes
& Hemsley (1886), in their Index Florae Sinensis, enumerated nineteen
species, one of which, //ex ficoidea, was new.

Loesener in 1900, in Diels’ Die Flora von Central-China, discussing the
Aquifoliaceae of the area concerned, listed seventeen species and varieties
that occurred there. Three of the specific epithets used, as well as five
names of lower rank, were nomina nuda. In his monumental Monographia
Aquifoliacearum published the following year he treated forty-eight species
and eighteen varieties native to China. Ten of these species and fourteen
of the varieties were described as new. In the supplement to that mono-
graph (1908) he added notes on nine Chinese species and four varieties.
He also described nine other new varieties or forms from China. In 1911,
in Sargent’s Plantae Wilsonianae, he added records for twenty species and
described three from West China as new.

H. Léveillé (1911-1915) created much trouble for later botanists with
an interest in Chinese plants. In his work on //ex he was especially con-
fusing. He described /lex chinensis Sims at various times as a new species
of Celastrus, of Svmplocos, of Embelia, and of Callicarpa. Likewise he
mistook Jlex suaveolens (H. Lévl.) Loes. for a new species of Celastrus,
Ilex metabaptista Loes. for a new species of Maesa and of Embelia and
Ilex macrocarpa Oliv. for a new species of Diospyros and of Celastrus.

Rehder, in his Ligneous Plants of Northern China (1926), recorded /lex
pernyi Franch, from Central and West China and /lex yvunnanensis Franch.
from West China. An important contribution to the study of //ex was his
identification, in 1933, of three species and four varieties improperly classi-
fied by H. Léveiile.

Among the papers I have reviewed in the past six years, there is only
one which is devoted solely to the /lex of China, and this includes only a
portion of the country. This paper was published by Comber in 1933. In
it he published five species and seven varieties or forms from Southwest
China. He also added new records to nineteen species and five varieties.
Handel-Mazzetti in the same year treated nineteen species and eight vari-
eties of /lex in his Symbolae Sinicae. One of these species and four of
the varieties were new.

In 1937, Chen, in his voluminous Illustrated Manual of Chinese Trees
and Shrubs, gave detailed descriptions of eleven species of //ex. Seven of
them were illustrated, with most of the drawings reproduced from sources
not given. But the information concerning each species was quite original
and useful.

1949] HU, THE GENUS ILEX IN CHINA 237

MORPHOLOGICAL CHARACTERS AND THEIR BEARING
N-CEASSIFICATION

The genus //ex is the only representative of the family Aquifoliaceae in
China. The morphological characters discussed below accordingly refer to
the genus /lex only.

I. Habit: The Chinese members of the genus //ex are either evergreen
or deciduous trees or shrubs. The largest of them are trees of the coastal
forests. Plants of Jlex rotunda Thunb. often attain a height of twenty
meters, with a smooth gray trunk of half a meter in diameter. The smallest
species are the shrubs of the Yunnan alpine region. Plants of Jlex intricata
Hook. f. and Jlex perryana S. Y. Hu are creeping, with upright branchlets
only 20-70 cm. high. Unless the contrast is very great and is reinforced
by good supplementary characters, the differences in habit have not been
considered of much importance in this paper. This is because our knowl-
edge of the habit of the plant depends largely on the reports given by the
collectors of specimens in the herbarium. ‘These reports are not always
dependable and are often entirely lacking. One collector may call a plant
a large shrub while another may call it a small tree, depending upon the
collectors’ judgment, experience and knowledge of the plants. Moreover,
in China the vegetation has been so disturbed that a plant with a tree habit
may have been so repeatedly cut by fuel gatherers that it has assumed a
shrubby habit.

Il. Branchlets: Jlex species are generally slow-growing plants. A
small twig usually represents two or more years of growth. In general, the
amount of the annual growth of a twig is 5-15 cm. The current year’s
growth is usually angular. Older growths are terete or subterete.

1. Dimension: Though environmental conditions often produce
variations in the diameter of the branchlets of a species, in general the
dimension attained by the branchlets is correlated with the various series
into which the genus is divided. The Prinifoliae have slender, much ridged
current year’s growth which never exceeds 2 mm. in diameter and with the
terminal buds absent or very poorly developed. On the other hand, the
Lauroilex have stout branchlets 3-4 mm. in diameter with well-developed
terminal buds.

2. ELONGATED AND ABBREVIATED SHooTS: The presence or absence of
the abbreviated shoots is of subgeneric interest in lex. There are no abbre-
viated shoots in the evergreen species. On the other hand, the deciduous
species largely possess this feature. The abbreviated shoots are 0.5—-5 cm.
long, densely covered by persistent bud-scales and scars of former in-
florescences or leaves. It appears that there is a correlation between the
presence of abbreviated shoots and the lenticels on the current year’s
growth of the elongated shoots, for both of them are characteristic of most
members of the subgenus Prinos.

3. Lenticets: The lenticels of the evergreen species of /lex are usu-
ally absent on the first, second, or even the third year’s growth. They are

238 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

small, inconspicuous, circular, or sometimes coalescent. Those of the
deciduous species are usually conspicuous even on the current year’s growth.
They are elliptic, white, and elevated on the second year’s growth.

4. TERMINAL Bups: The members of the genus //ex exhibit a very
interesting variety of terminal buds. Some of them, like /lex atrata W. W.
Sm., have large ovoid buds up to 1 or 2 cm. long with thick coriaceous
densely ciliate scales. Others, like Ilex kobuskiana S. Y. Hu and Ilex
salicina Hand.-Mzt., have narrowly conic acute buds with such loose scales
that the buds look naked. There are still others, like the Prinifoliae and
the Hanceanae, that seldom have any terminal buds. In general, the pres-
ence or absence of the terminal buds, their shape and size, are often of
serial significance. For example, the Prinifoliae possess no terminal buds,
the Hanceanae have poorly developed ones, the Sideroxyloides have thin
buds with loose scales, and the Hookerianae have very well developed ter-
minal buds.

5. INDUMENTUM AND EXCRESCENT GrowTH: Most species of //ex
have smooth, glabrous, or pubescent branchlets; the hairs are always simple
and straight. In general the hairs are very minute and sparse. This I call
puberulous. When the hairs are long enough to be seen by the naked eye,
they are, with the exception of a few species like Jlex pubescens Hook. &
Arn. and /lex aculeolata Nakai, ferruginous in color. Since the variation
in the density of the hairs in a species is often very gradual, unless it is
aided by other characters pubescence has been considered of minor im-
portance in this work.

There are three species of //ex occurring on the high mountains of south-
western China and eastern Himalaya, /lex delavayi Franch., Ilex intricata
Hook. f., and /lex nothofagifolia Ward, that have rugose or warty branch-
lets. In general, these branchlets appear ochraceous and corky. This ex-
crescent growth is denser on the current year’s growth. It varies from
merely pliciform ridges in /lex delavayi to so dense a covering of distinct
warts in /lex nothofagifolia that the current year’s growth appears like a
thin pipe-cleaner. Such excrescent growths form a very convenient means
of distinguishing species.

III. Leaves: The size, shape, margin, and texture of the leaves of Ilex
have always been used as major features in the classification of its species.
Needless to say, the use of these characters has been over-emphasized by
many authors. It is one of the reasons why so much confusion exists in
Loesener’s monograph. For example, because of their agreement in the
number of the pyrenes, the nature of the endocarp, the inflorescences, the
type of the flowers, and the glandularly punctate leaves, /lex crenata
Thunb. and /lex viridis Champ. ex Benth. are very closely related species.
Yet because of differences in the size of the leaves and their relative thick-
ness, Loesener placed the former species in the Paltoria and the latter in the
Aquifolium.

1. Duration: The duration of the leaves, whether evergreen or de-
ciduous, varies in association with other characters and is useful in sub-

1949] HU, THE GENUS ILEX IN CHINA 239

generic, sectional, or serial diagnosis. All the members of the subgenus
Prinos have deciduous leaves. In the warm temperate regions like the
Chengtu Plain, the leaves of those deciduous species appear with the flowers
in mid-April. For six or seven months they persist on the tree and fall in
October or November. In the warmer regions, like Canton, the leaves of
the deciduous species appear a month earlier.

The leaves of Lioprinus rarely remain on the second year’s growth. They
are like many broad-leaved evergreen trees of the Lauraceae in that the
older leaves fall soon after the terminal buds unfold in the spring. Leaves
of the various species of the Dipyrenae are the longest-lived ones in //ex.
Many of them retain leaves on the third or even the fourth year’s growth.

2. TrexTuRE: The Chinese species of J/ex exhibit considerable varia-
tion in the texture of the leaves. In the section Aquifolium are found leaves
of the thick-coriaceous type. Many of them, like the leaves of /lex lati-
folia Thunb., have shiny glabrous surfaces. Some of them, like Jlex
elmeriana S. Y. Hu, have such thick leaves that the lateral nerves are
obscured. A very few of them, like /lex perryana S. Y. Hu and Jlex intri-
cata Hook. f., due to greatly impressed veinlets, have rugose upper surfaces.

The chartaceous leaves of the series Hookerianae present another varia-
tion. Most of these leaves are opaque, with evident lateral nerves. Dis-
tinctive are the membranaceous leaves of the subgenus Prinos. Most of
these leaves are large, with prominent veins. In J/lex tsoi Merr. & Chun
the veinlets are so prominent that the leaf-surfaces are marked with distinct
minute areolae.

Though the texture of leaves is usually greatly influenced by environ-
mental factors, it can be a useful character in defining subgenera and
sections. For example, the subgenus Prinos is characterized by mem-
branaceous or chartaceous leaves, the Denticulatae by coriaceous leaves,
and the Lauroilex by large, entire, thick-coriaceous leaves.

3. SIZE AND SHAPE: ‘The largest leaves of the genus are found in /lex
dolichopoda Merr. & Chun, and the smallest in lex intricata Hook. f. This
variation is from 20 cm. in length in the former to 1 cm. in the latter. Ex-
treme cases, however, are few. In the majority of the species the leaves
vary normally from 2 to 8 cm. in length. In shape they vary from sub-
orbicular, as found in /lex nothofagifolia Ward, to linear-lanceolate, as
found in /lex fargesti Franch., obovate or obcordate as in /lex tutchert
Merr., or ovate as in Zlex macrocarpa Oliv. Ilex cornuta Lindl. & Paxt. is
unique in having subquadrangular leaves. The evergreen species usually
possess leaves elliptic or oblong-elliptic in outline, while in the deciduous
species the leaves are ovate or ovate-elliptic.

Environmental conditions and the age of the plant are factors which
influence very much the variation in size and shape of the leaves. A vig-
orous offshoot of a shrub may have leaves 2—3 times as large as the normal
leaves. Furthermore, the leaf shape is often affected by the position of the
leaves on the branchlets. Often, upwards from the base of the branchlet,
the leaves gradually change in width; thus one finds ovate leaves near the
base and elliptic or even lanceolate leaves toward the apex of a branchlet.

240 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXX

Only when there is a sharp contrast in other characters between two species
may the size and shape of the leaves be used as distinguishing characters.
For example, /lex dolichopoda Merr. & Chun and Ilex kobuskiana S. Y. Hu
both have punctate leaves. The former has leaves 18—25 cm. long and the
latter leaves 4.5—-9 cm. long. This difference, reinforced by differences in
the length of the fruiting pedicels, clearly distinguishes the two species.
Mere size of the leaves alone cannot be used as a specific character.

4. APEX AND Base: In the leaf apices of the Chinese //ex species,
nearly all the common variations are found. The apex may be: (1) long-
caudate and acute; (2) caudate and obtuse; (3) acuminate and serrate;
(4) short-acuminate and retuse; (5) obtuse and retuse; (6) obcordate;
and finally the extreme (7) tricuspidate and spinose. The last-mentioned
apex, although extreme, is probably the best known and the one most
associated with the genus by the general public.

Both the apex and the base of the leaves of a species vary with the
size and shape of the blade and with the position of the leaves on a branch-
let. The leaves on the lower portion of a twig may be ovate with rounded
bases and acute apices, while those of the apical portion of the same twig
may be elliptic with obtuse bases and acuminate apices. This is the case in
Ilex tsou Merr. & Chun and /lex macrocarpa Oliv. For this reason, the
shape of the apex and the base cannot be used alone even for varietal
differentiation. Nevertheless, when associated with other characters, they
are the most obvious characters to employ in the differentiation of species,
or even in characterizing series. For example, /lex venulosa Hook. f. is
differentiated from J/ex omeiensis Hu & Tang by its caudate apex and its
cymose inflorescences. Likewise, the series Longecaudatae is characterized
by the small fruit, entire leaves, and caudate or acuminate leaf apices.

5S. Marcin: Over half of all the species of Chinese //ex possess leaves
with serrate or crenate margins. The teeth may be fine and aristate as in
Ilex pubescens (Hook. & Arn.) Champ. ex Benth. and /lex serrata Thunb.
var. sieboldi (Miq.) Rehd., coarse and remote as in /lex intermedia Loes.
var. fangii (Rehd.) S. Y. Hu, or crenulate and inconspicuous as in Jlex
ficoidea Hemsl. About thirty-seven per cent of the species have entire
leaves. The remaining nine per cent have spinose leaves. Again, the mar-
gin of the leaves in a species may vary with the age of the plant. In species
such as /lex cornuta Lindl. & Paxt., Ilex corallina Franch., and Ilex dipy-
vena Wall., in developing from the juvenile stage to maturity the margin
of the leaf changes from spinose to subentire or entire. For this reason the
margin alone should not be used even for varietal differentiation. How-
ever, when the margin is used as an auxiliary character, it is a very
convenient and obvious means for distinguishing species. It is often used
as a supplementary character to separate sections or series. For example,
over ninety-five per cent of the species in the section Pseudoaquifolium
have entire leaves. The Lauroilex species all have entire leaves and the
Prinoides species all have chartaceous or membranaceous, serrate or crenate.
leaves.

1949 ] HU, THE GENUS ILEX IN CHINA 241

6. VENATION: All the Chinese //ex have pinnate, netted veins. The
midribs are usually impressed above and elevated beneath. Midribs ele-
vated above are often used as supplementary characters for distinguishing
species. The lateral nerves of //ex vary from two, three, or four pairs in
Ilex intricata Hook. f. to ten to twenty pairs in /lex polyneura (Hand.-Mat.)

Y. Hu. An appreciable difference in the number of the lateral nerves is
always a convenient quantitative means for separating closely related
species, for example, //ex micrococca Maxim. is distinguished from Jlex
polyneura by having only six to eight pairs of lateral nerves. About
forty-seven per cent of the Chinese species of //ex have obscure lateral
nerves and veinlets, at least on the upper surfaces. In the case of her-
barium specimens, venation is often affected by the method of preparation
of the plants. In Prinoides all species have distinct veinlets, while most of
the species in Lioprinos have obscure veinlets. In the Hookerianae all the
species have veinlets impressed on the upper surface. Jlex tsoit Merr. &
Chun of the section Prinoides can be readily distinguished even in a sterile
condition because of the very distinct veinlets

7. PrtroLes: Most species of the //ex have canaliculate petioles. Few
of them have plane ones. The presence or absence of grooves on the upper
surface of the petioles is often useful in separating species. For example,
specimens of /lex sterrophylla Merr. & Chun from Hainan have been identi-
fied as representing /lex pedunculosa Mig. But without reference to the
indumentum or pyrene characters, the former can readily be distinguished
from the latter by its plane petioles.

The length of the petiole varies from 2—4 mm. in //ex cinerea Champ. to
15-25 mm. in Ilex sterrophvila Merr. & Chun. The differences in the ratio
between the leaves and the petioles are more striking. In Jlex cinerea
the petioles are one thirty-sixth to one-nineteenth the length of the lamina,
and in Ilex sterrophyila one-fourth to one-half. This proportional differ-
ence is another quantitative measure often used in separating species.

8. SripuLes: Van Steenis in 1948 (in Bull. Bot. Gard. Buitenz. HI.
17: 389) called attention to the fact “that in British botany there is much
controversy as to the occurrence of stipules in Aquifoliaceae.”’ He cited
authors like Hooker, f., Brandis, and Christie. Even as late as 1926,
Hutchinson says, ‘Stipules absent.”

The Chinese representatives of the Aquifoliaceae are assentially stipu-
late, and in ninety per cent of the species, the stipules are present and per-
pint even after the leaves fall. In the other ten per cent the stipules
are obscure or lacking by abortion. In general the stipules are minute,
callose, deltoid, less than one millimeter long. The largest stipules of the
Chinese /lex are found in /lex serrata Thunb. var. sieboldi (Miq.) Rehd.
In this species they are 2 mm. long, pilose, and caducous.

The presence or absence of stipules is of serial significance. It is of great
interest to know that all the species that are devoid of stipules are in the
series Chinenses. Thus the presence or absence of stipules can be used as
a supplementary character for differentiating this series from its closely
related ones.

242 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

The size of the stipules of closely related species is rather definite. Ex-
amples can be found in Jlex polyneura (Hand.-Mat.) S. Y. Hu and Ilex
micrococca Maxim. In the former species the stipules are 1-1.5 mm. long,
and in the latter only 0.2-0.3 mm. In such a case the stipules can be used
as a good supplementary character in distinguishing species.

IV. Inflorescences: The fundamental organization of the inflores-
cences in the genus //ex is a trichotomous axillary cyme as illustrated by
Ilex chinensis Sims and Ilex ferruginea Hand.-Mzt. Through reduction or
multiplication of such cymes, and as a result of the evolution of the shoot
system, directed towards a division of labor, various types (as illustrated
in figures 3 and 4) of inflorescences may be evolved. The types of the
inflorescences and their position on the branchlets are of important con-
sideration in this work.

ZS 5

Fic. 1. Types OF THE SOLITARY INFLORESCENCE IN ILEX: a & b. simple
cymes; c. dichotomous compound cyme; d. trichotomous compound cyme;
e. long-pedicellate solitary flower; f. short-pedicellate solitary flower;
pseudo-umbel; si. head-like inflorescence.

1949] HU, THE GENUS ILEX IN CHINA 243

1. Types OF THE INFLORESCENCE: The Chinese /lex exhibit a very
interesting set of inflorescences. According to their positions on the branch-
lets I discuss them under the following two heads

a. Solitary type: The solitary type of inflorescence occurs only on the
current year’s growth. They are nearly all axillary to a leaf and are often
associated with an adaxial dormant vegetative bud. The simplest kind of
such solitary inflorescences consists of a cyme with three flowers. Its
peduncle is about 5-10 mm. long, and the pedicels of the individual flowers
are usually shorter than the peduncle. Examples are the pistillate in-
florescences of Ilex ferruginea Hand.-Mat. (fig. 1, b) and the staminate
flower of Ilex macrocarpa Oliv. (fig. 1, a).

From the simple cyme evolution can proceed by the multiplication of
branches. This is shown in the dichotomous compound cyme with evident
secondary axes, like those of the pistillate flowers of J/ex macluret Merr.
(fig. 1, c), or a trichotomous compound cyme with rather long secondary
axes and evident tertiary axes, like those found in /lex micrococca Maxim.
(ies ld):

By reduction in the number of flowers in the simple cyme there may be
evolved a solitary flower with a long pedicel and two median or supermedian
prophylla (scales on the pedicel), as, for example, the pistillate flower of
Ilex yunnanensis Franch. (fig. 1, e). Reduction of both the number of
flowers and the length of the pedicel of a simple cyme may occur. Then
there is evolved a subsessile solitary fruit borne on a pedicel 2-3 mm.
long, at the middle of which are the scars of two abortive flowers. The
pistillate flowers of Ilex lancilimba Merr. (fig. 1, f) are an example.

By the reduction of the secondary axes of a compound dichotomous
cyme there can be evolved a pseudo-umbel that has a peduncle 10-20 mm.
long, and many bracteoles at the bases of the pedicels. Examples occur
in the staminate or the pistillate inflorescences of //ex umbellulata (Wall.)
Loes. (fig. 1, g). If such reduction also occurs in the pedicels, then a sub-
capitate inflorescence like that of /lex tugitakayamensis Sasaki (fig. 1, h)
may be evolved.

b. Fasciculate types: The fasciculate types of inflorescence occur
only on the second year’s growth or on older growth. They are axillary
to old leaves and never have a dormant axillary bud at their base. The
commonest kind consists of a sessile fascicle with a basal collar formed of
several persistent cartilaginous or leathery bud-scales, and a short axis
with or without an abortive terminal bud. A membranaceous scale sub-
tends each of the 1-3-flowered individual branches. The staminate in-
florescences of Jlex ficoidea Hemsl. (fig. 2, a) are examples of this type.

Through multiplication by the branching of the individual branches, a
fascicle with trichotomous compound cymes or with pseudo-umbels, as in
the Lauroilex (fig. 2, b & c), may be evolved. By an increase in the num-
ber of flowers on the branches and the elongation of the central axis, a
pseudopaniculate inflorescence such as the staminate inflorescence of Ilex
latifolia Thunb. (fig. 2, d) may result. By a reduction in the number of

244 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

flowers on the branches, a loose fascicle or pseudoraceme of uniflorous
‘cymes may develop. The pedicels each bear one or two submedian pro-
phylla. This is the condition in the flowers of //ex intermedia Loes. (fig.
2, e). By the shortening of both the axis and the pedicels, a compact
fascicle of uniflowered pedicels, each bearing one or two basal prophylla,
may develop, as in the pistillate inflorescences of [lex corallina Franch.
(fig. 2, f). Reduction in the number of individual flowers in the fascicle
produces paired fruits as in /lex kanceana Maxim. and /lex perryana S. Y.
Hu (fig. 2, g). A further reduction of the number of the flowers gives
the solitary-flowered condition, as found in /lex chingiana Hu & Tang

oe

|
SON

Ss

Fic. 2. TYprEs OF THE FASCICULATE INFLORESCENCE IN ILEx: a. fascicle
with 1-3-flowered individual branches; 0. fasciculate compound cymes; c.
fasciculate pseudo-umbels; d. pseudopanicle; e. loose fascicle with uni-
florous individual branches; f. compact fascicle with uniflorous individual
branches; g. paired fruits; /. solitary fruit of a much reduced fascicle.

1949] HU, THE GENUS ILEX IN CHINA 245

(fig. 2, h). In this the scar or scars at the base of the fruiting pedicel
indicate the positions of the aborted flowers.

2. STRUCTURE OF THE INFLORESCENCES:

a. Bud-scales: The fasciculate type of inflorescence is evolved from
an axillary bud. The bud-scales of these inflorescences are often persistent.
They are cartilaginous or coriaceous. Sometimes it is hard to differentiate
bud-scales from the bracts, for the change in texture is gradual.

b. Bracts: The scales that subtend the individual branches of a
fascicle are called bracts in this paper. Those that subtend the secondary
or tertiary branches are called bracteoles. Most bracts have stipule-like
basal appendages. The size and shape of these appendages are sometimes
of serial importance; for example, all the Repandae have bracts with long,
slender, ciliate appendages.

c. Prophylla: The scales on the pedicels, which do not subtend flowers
are called prophylla. Their insertion may be basal, superbasal, median, or
supermedian on the pedicels. The insertion of the prophylla is often used
as an auxiliary character for distinguishing species; for example, besides
the size of the fruit and the texture of the leaves, [lex tephrophylla (Loes.)
S. Y. Hu is differentiated from J/ex corallina Franch. by its minute pro-
phylla, which cover less than half of the fruiting pedicel, while, in contrast,
the prophylla of Zlex corallina reach the calyx.

d. Peduncles: The stalks of all the branched solitary types of in-
florescence and those of the multiflorous individual branches of the fascicu-
late types are called peduncles in this work. The peduncles of the solitary
inflorescences possess no bracts at their bases. Those of the fasciculate
inflorescences all have basal bracts. The length of the peduncle of a species
is a variable character. In a few cases it has been used as an auxiliary
means for distinguishing species. For example, /lex rotunda Thunb. is
differentiated from /lex umbellulata (Wall.) Loes. by its shorter peduncles
(9-13 mm. long), and Ilex sterrophylla Merr. & Chun is differentiated
from its closely related species, Ilex editicostata Hu & Tang, by its
peduncles of a length of 2-3 cm.

e. Pedicels: The stalks of either the solitary individual flowers or of
the ultimate branches of compound inflorescences are called pedicels in this
paper. The longest pedicels, those of Ilex pedunculosa Miq., are 4—4.5 cm.
long. But this length is extremely rare. In general the pedicels of Ilex
are 2-15 mm. long. Most of them possess one or two prophylla at the
base (basal), at the middle (median), or above the middle (supermedian).

The length of the pedicels of //ex is a rather stable character. It is an
obvious quantitative character for distinguishing species. Thus Jlex
asprella (Hook. & Arn.) Champ. is differentiated from all the other Prinos
species by its long slender pedicels. Even between closely related species
like Ilex macropoda Miq. and Ilex tsow Merr. & Chun the length of the
fruiting pedicels is a dependable character for separating them.

f. Central Axis: All the fasciculate types of inflorescence have central

246 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

axes. These may be very short, even almost completely suppressed, as in
Ilex corallina Franch. or Ilex intricata Hook. f. Or they may be quite
long, so long, in fact, that the inflorescences appear pseudoracemose, as
in /lex intermedia Loes., or pseudopaniculate, as in Jlex latifolia Thunb.
The length of the central axes of different inflorescences on a single
branchlet may vary considerably, and thus it is not a good character to
use for distinguishing species. As an auxiliary character it is sometimes
useful in separating a variety from a species; for example, /lex confertiflora
Merr. var. kwangsiensis S. Y. Hu is differentiated from o pee, species
by its large leaves and its central axis of a length of 22 m

g. Secondary and Tertiary Axes: The compound cymes, both solitary
and Me all possess secondary or sometimes tertiary axes. The
presence or absence of the secondary axis and its relative length are con-
venient distinguishing characters. For example, /lex maclurei Merr. is
distinguished from the rest of the Chinese //ex species by having a secondary
axis in the infructescence. Besides having more numerous lateral nerves,
Ilex polyneura (Hand.-Mzt.) S. Y. Hu is also distinguished from its
very close relative, //ex micrococca Maxim., by the fact that its secondary
axis is shorter than the fruiting pedicel.

3. SexuaL DimorpHism: All the Chinese species of //ex are dioecious.
The inflorescences of almost all the species are sexually differentiated.
In over ninety per cent of the species the staminate inflorescences are
more prolific, and the flowers are more showy. This abundant flowering
is attained by the branching of the solitary inflorescence or the fascicle,
or through the specialization of the shoot systems (one species), or
through both.

Nearly all of the more prolific staminate inflorescences result from
increased branching. In //ex serrata Thunb. var. sieboldi (Miq.) Loes. the
inflorescence is solitary. The pistillate flowers are usually solitary or
rarely a much reduced 2- or 3-flowered cyme with a peduncle 1—1.5 mm.
long. The staminate ones are cymose, with nine to twenty-one flowers
and peduncles 3 mm. long. In Jlex chinensis Sims the pistillate cymes
are usually 3- to 7-flowered, with the peduncles usually shorter than the
leaf-petioles, and the staminate cymes usually have twenty-one or more
flowers, with the peduncles always exceeding the leaf-petioles in length.
In all the species of the section Aquifolium the pistillate fascicles have
one-flowered individual branches, while those of the staminate fascicles are
3- to 5-flowered. Many of these staminate fascicles, such as those of lex
latifolia Thunb., are so prolific that they appear pseudopaniculate.

About nine per cent of the more prolific staminate inflorescences are due
to both branching and specialization of the shoot system. In /lex microcarpa
Oliv., for example, the pistillate flowers are solitary axillary, and the
staminate flowers are one- or three-flowered, cymose, and often fasciculate.
Such a condition is also common in /lex yunnanensis Franch., Ilex crenata
Thunb., and their related species.

Ilex asprella (Hook. & Arn.) Champ. is the only species that has solitary

1949] HU, THE GENUS ILEX IN CHINA 247

pistillate flowers and fasciculate staminate flowers with the individual
branches uniflorous.

The inflorescences of the Chinese species of J/ex are as a whole incon-
spicuous. The great reduction of the number of flowers in the pistillate
inflorescences makes them less conspicuous than the staminate ones, and
hence the female plants are less attractive to collectors, especially inexperi-
enced ones. Many of the Chinese //ex are unrepresented in the herbarium
by flowering specimens and specimens with pistillate flowers are especially
uncommon.

Because of this sexual dimorphism in the inflorescences, an unduly large
number of new species have been proposed by workers who either lacked
sufficient material or who did not appreciate the significance of sexual
dimorphism. This has added greatly to the synonymy of certain species.

V. Flowers: The flowers of //ex are all small and inconspicuous.
When fully opened, the majority of them are 5-6 mm. in diameter. The
largest, the staminate flowers of Ilex latifolia Thunb., are only 9 mm. in
diameter. The flowers of Jlex micrococca Maxim. measure only 3-4 mm.
in diameter.

The color of the flowers of //ex is never brilliant. In the total number

of species of which we have floral records, about eighty-four per cent are
greenish yellow or white and about thirteen per cent are pink, lavender,
or red. The flowers of Ilex intricata Hook. f. are chocolate-colored. In
Ilex lancilimba Merr. the pistillate flowers are pink-purple and the stam-
inate flowers greenish white. The color of the flowers of Ilex yunnanensis
Franch. has been reported as greenish yellow, cream-white, or red. It is
most likely that the color of the flowers of this species varies with the age
of the flowers and also with the habitats of the plants. The color of the
flowers of Ilex appears to have some correlation with other characters of
the various groups. All the members of the subgenus Prinos have white
flowers. A majority of the pink, lavender, or red-flowered species are in
the section Lioprinus. All the Aquifolium with the exception of Ilex
intricata Hook. f. have greenish yellow or cream-white flowers. But due
to the small size and the lack of attractive color of the flowers, they have
been overlooked in the field, and thus flowering specimens are often
poorly represented in herbaria. For this reason, in this paper the color
of the flowers has not been taken into consideration in distinguishing
species.
Of the total species of which we have floral records, thirteen per cent
have been reported as fragrant. Personally I have observed bees and
small wasps visiting the white and fragrant flowers of [ex macrocarpa Oliv.
in the woods of West China.

There are no field records concerning the length of the flowering period
of a species. Experience tells us that for a period of two weeks one can
collect good flowering specimens from a tree of /lex macrocarpa Oliv.
growing to a height of ten meters.

The height of the flowering season, as one would expect, is in April,

248 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

with approximately one-third of the species producing flowers at this time.
May and June come next, with March and July following in that order.
A very few species may be found flowering in January and December.

The floral characters which are valuable in the classification of the
species are Summarized here.

1. Catyx: In /lex the calyx is always persistent. At anthesis it is
usually patelliform, but after the fruiting stage it becomes explanate. It
is small, measuring 2—4 mm. in diameter. The dimension of the persistent
calyx is of serial significance. In all members of the series Sideroxyloides
the calyx is large for the genus, its diameter measuring over half that of the
fruits. All representatives of the Repandae have a small quadrangular
calyx 2 mm. in diameter.

The indumentum of the calyx-lobes is a rather variable character. In
the species //ex triflora Blume, for instance, the calyx may be either sparsely
or densely pubescent. Hence this characteristic has been little featured in
the classification of the species. On the other hand, the presence or
absence of cilia on the margin of the calyx-lobes is quite constant. It has
occasionally been employed as a supplementary character in distinguishing
species or varieties.

2, CoroLtta: The shape and size of the corolla of the staminate
flowers of all the species are fairly constant. The corolla is always rotate,
with the petals slightly united at the base. The free parts of the petals
are oblong, 2—3, rarely up to 4 mm. long. The apical half of the petals of
the Repandae is ciliate.

In the pistillate flowers, the corolla is rotate, with the base of the petals
considerably more united in the Paltoria or only slightly so in the Lioprinus
and the Venulosae, or suberect, with the base of the petals either slightly
united or choripetalous in Aquifolium and in Pseudoaquifolium. The
differences in the shape of the corolla and the degree of union of the
petals seem to nave serial or sectional significance. But the inadequate
representation of the pistillate flowers in the herbaria has restricted the
use of these characters in classification.

3. STAMENS AND STAMINODES: At full anthesis the stamens of Jlex
may be longer than the petals, as in the Repandae, equaling the petals,
as in /lex wenchowensis S. ¥Y. Hu, or shorter than the petals, as in Jlex
crenata Thunb. The anthers are usually oblong-ovoid, 0.75-1 mm. long.
As the length of the filament and the shape of the anthers change with
the age of the flower, these characters have not been used in distinguishing
species.

The abortive stamens of the pistillate flowers are called staminodes in
this paper. They are always shorter than the petals. Their anthers are
either sagittate, cordate, or ovate. They are glabrous except in Jlex
memecylifolia Champ. The hairy staminodes of this species have been
used as a supplementary character for distinguishing it from its close
relatives.

4. Ovary AND RUDIMENTARY Ovary: The ovary of the flower of all

1949 ] HU, THE GENUS ILEX IN CHINA 249

the species of //ex is syncarpous, with two up to thirteen carpels. At anthe-
sis it is usually ovoid, 1-2 mm. in diameter, devoid of style, but with a
discoid and lobed stigma, as in /lex serrata Thunb. var. sieboldi (Miq.)
Loes. (tig. 3, a), or with a slightly elevated and pointed mammiform stigma,
as in Ilex kwangtungensis Merr. (fig. 3, b). In /lex tsoi Merr. & Chun the
ovary is subglobose with a capitate stigma (fig. 3, c). In /lex fragilis
Hook, f. the ovary is pulvinate, with an evident style and a capitate or
narrow-cristate stigma (fig. 3, d). In /lex macrocarpa Oliv. the ovary is
ovoid, with an evident style and a columnar stigma (fig. 3, e

The shape of the stigma (especially after the fruit attains maturity) and
the presence of a style are very constant characters in some species of
Ilex. Therefore they are of fundamental importance in the delineation of
species and even of sections. The columnar stigma is characteristic of
Ilex macrocarpa Oliv., Ilex chapaensis Merr., and many members of the
series Sideroxyloides. A majority of the species of the series Denticulatae
have plane or slightly impressed navel-like stigmas.

In general the ovary is glabrous. In /lex pubilimba Merr. and Ilex
wangiana S. Y. Hu the ovary and the fruit are sparsely pubescent. These
characters set the two species off from their close relatives.

The abortive ovary of the staminate flowers is called the rudimentary
ovary in this work. It may be subglobose with obtuse or slightly impressed
center, or Saale with a rostellate and cleft apex as in llex muicrococca
Maxim. (fig. 3, g). The presence or absence of the rostellum on the
rudimentary ovary is of serial or sectional significance. For example,
all the Micrococca and the Umbelliformes species have rostellate rudi-
mentary ovaries. A branchlet of /lex chapaensis Merr. with staminate
flower can be distinguished from one of Jlex macrocarpa Oliv. only by the
rostellum on the ovary. The rudimentary ovary is always glabrous except
in Ilex brachyphylla (Hand.-Mat.) S. Y. Hu (fig. 3, f)

VI. Fruit: Over ninety-five per cent of the Chinese //ex species pos-
sess red fruit. This fruit reaches maturity in autumn and persists on the
tree for a long time, often until the plant flowers again the following
spring. Trees or shrubs with scarlet fruit, whether accompanied by leaves
or not, are striking sights in any landscape. They attract the collectors’
attention. Even the black-fruited //ex is not altogether devoid of means
for attracting attention. Some species, like lex asprella (Hook. & Arn.)
Champ. have their fruit projected from the branchlets on extremely long
pedicels (fig. 3, i), and others, like Z/ex macrocarpa Oliv., have extraor-
dinarily large fruit (fig. 3, h). Since the fruit of over ninety-eight
per cent of our species of J/ex is known, it is of fundamental importance in
the classification of the genus. Thus any character that one can draw
from the fruit, be it the size, the color, the pedicel, the persistent calyx,
the style (when present), or the stigma, is of some aid in distinguishing
the species.

1. Size: Nearly all of the fruit of the Chinese //ex is less than 8 mm.
in diameter (usually 4-7 mm.). The larger fruits are usually limited to

250 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Fic. 3. a-e. STIGMA: discoid stigma of ovary of I. serrata; b. mam-
miform stigma of /. ia eee c. capitate stigma of J. tsott; d.
cristate stigma of /. fragilis; e. columnar stigma of J. macrocarpa. f & g.
RUDIMENTARY OVARY: f. globose and pubescent rudimentary ovary of J.
aging x 12; g. rostellate rudimentary ovary of Ilex micrococca,

Progr? h. a large fruit of /. macrocarpa with nee cee

a long-pedicellate fruit of J. asprella with capitate stigma, x 1;
j. large See calyx of fruit of J. metabaptista, X 2; k. small a of
fruit of J. ficoidea showing the pedicel and prophylla, x 2; J. the same show-
ing the discoid stigma. m & n. SIZE OF PYRENES: m2. a small smooth pyrene
of I. wardii, X 2; . a large rugose pyrene of J. chingiana, X 2. o-u. Our-
LINE OF THE cross- SECTIONS OF PYRENES, all X 3: 0. triangular, smooth with
thin endocarp, /. peduncilosa; p. 3-ridged and eee aaiicuiate pyrene, J,
macrocarpa; & r, U-SHAPED PYRENES WITH SMOOTH THIN ENDOCARP; 5.
triangular thick-walled stony aa of I. subficoidea; t. elliptic thick- walled
woody pyrene of J. dipyrena . suborbicular smooth pyrene of I. perryana.
uv & w. PALMATELY STRIATE YRENES: wv. striate, esulcate pyrene of
perrvana with slightly meee striae; w. striate and sulcate pyrene of
I, ficoidea.

>

1949] HU, THE GENUS ILEX IN CHINA 201

the section Prinoides and to the series Aquifolioides.and Denticulatae of
the section Aquifolium.

2. FRuITING PepicEL: The length of the fruiting pedicels varies from
1-2 mm. long in J/lex ficoidea Hemsl., lex corallina Franch., and many
others in the series Dipyrenae or Repandae to 40-45 mm. long in Jlex
pedunculosa Mig. This difference in the length of the fruiting pedicels
has been used as a supplementary character in distinguishing species and
varieties.

3. Persistent Catyx: The size of the persistent calyx of the fruit
of /lex is constant in the series and sometimes in the sections. The series
Sideroxyloides is characterized by the large persistent calyx, which is over
half the diameter of the fruit (fig. 3, j). On the other hand, the series
Aquifoliotdes and the Repandae are characterized by their quadrangular
persistent calyces (fig. 3, k).

4. Sticma: The stigma of the mature fruit is quite a constant char-
acter, useful as a supplementary character for distinguishing species.
Some of the stigmas, like those of the Denticulatae, are plane or some-
what navel-like, some, as those of //ex ficoidea Hemsl., discoid (fig. 3, 1),
some, like those of Ilex pubescens Hook. & Arn., capitate, and some are
cristate, as in Jlex fragilis Hook. f. (fig. 3, d). |

All members of the genus have fleshy fruit, with chartaceous exocarp,
fleshy and juicy mesocarp, and distinct coriaceous woody or stony endo-
carps, each enclosing a seed. In classification there seems to be no proper
term to apply to this type of fruit. DeCandolle (Prodr. 2: 13. 1825)
called it a berry, ‘‘bacca.” Hooker (Gen. Pl. 1: 357. 1862) referred to it
as a drupe, ‘“‘drupa,” as did Loesener. But the fruit of //ex is neither a
berry nor a drupe, for a berry does not possess a hard endocarp, and a
drupe has only a single stone, since it is derived from a single carpellate
ovary. The hard parts in a berry are the seeds. The hard portion of the
fruit of /lex is the endocarp with the vascular bundles clinging to it or
loosely attached to it. The fruit of the genus might be termed a drupe-
like berry or a berry-like drupe. To make things simple and definite,
however, I would like to propose the term bacco-drupe and define it as a
fruit derived from a syncarpous ovary with chartaceous exocarp, fleshy
mesocarp, and separated coriaceous, woody or stony endocarps (pyrenes),
each enclosing a single seed. It is characteristic of all J/ex species except
Ilex insignis Hook. f. of India.

VII. Pyrenes: The seed of /lex, enclosed by the endocarp, is called
a pyrene in this work. The pyrenes of all the Chinese //ex are in one
whorl embedded in the soft mesocarp. Their abaxial surface is much
broader than the keel-like adaxial surface. The abaxial surface of a
pyrene is called the dorsal surface in this paper.

Like the fruit, the pyrenes are commonly available in the herbaria and
are important in the classification of the genus.

1. Numser: About seventy per cent of the Chinese //ex species have
four pyrenes. The remaining species possess a varying number, from one

252 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

to thirteen. The number in those species producing four pyrenes is con-
stant. The number in those species which have more or less than four
pyrenes is generally variable; for example, a branchlet of Ilex fragilis
ook. f. var. kingit Loes. has many fruits with six to eight pyrenes and
some with nine to thirteen. A single branchlet of Zlex dipyrena Wall.
bears many fruits with two pyrenes and some with one, three, or four. The
large percentage of the four-pyrene species and the variability in the
number of the more or less than four-pyrened species both indicate that
four-merous may have been the fundamental condition of the flower parts,
including the ovary, of //ex, and that the other numbers are derivatives.

The pyrene number has no subgeneric significance. Both the subgenera
Prinos and Euilex have four-pyrened species. Species with five or more
pyrenes are common to both subgenera. Nevertheless, the species that have
less than four pyrenes are all confined to the series Dipyrenes of the Euilex.

The number of the pyrenes has provided a valuable supplementary
character for separating the sections; for example, all of the species of the
Aquifolium have four pyrenes with woody or stony endocarps, and those
of the Pseudoaquifolium, on the other hand, have more than four pyrenes,
with coriaceous or sublignescent endocarps.

2. SIZE AND SHAPE: ‘The smallest pyrenes are found in /lex wardii
Merr. and are ca. 1.5 mm. long (fig. 3, m). The largest are found in
Ilex chingiana Hu & Tang and measure approximately 15 mm. long
(fig. 3, n). The pyrenes in other species of the genus fall between these
extremes.

With a few exceptions, the cross-section of a pyrene of /lex is triangular
in outline. In Jlex pedunculosa Mig. (fig. 3, 0) the dorsal surface is
smooth and convex. In /lex macrocarpa Oliv. (fig. 3, p) the dorsal surface
is 3-ridged and 2-canaliculate. In Jlex lancilimba Merr. (fig. 3, q) the
dorsal surface is widely U-shaped. In /lex kwangtungensis Merr. (fig. 3, r)
the dorsal surface is narrowly U-shaped. In Jlex subficoidea S. Y. Hu
(fig. 3, s) the dorsal surface is rugose and uneven. Other variations in
the outline of the cross-section of the pyrene are found in /lex dipyrena
Wall. (fig. 3, t) where it is oblong-elliptic, and in /lex perrvana S. Y. Hu
(fig. 3, u) where it appears suborbicular.

Viewed from the dorsal surface, the pyrenes show varied shapes, namely
oblong, oblong-elliptic, elliptic, or rarely suborbicular or obovate. The
ends may be pointed or obtuse, rarely retuse.

The shape of the pyrenes is sometimes of significance in the delimiting
of the series. The Umbelliformes have narrow elliptic pyrenes with pointed
ends, while the Repandae have oblong pyrenes, usually with obtuse ends.

3. TEXTURE AND SCULPTURE: The texture of the pyrene of J/lex
varies from coriaceous as found in J/lex crenata Thunb. through sub-
lignescent as found in /lex rotunda Thunb., to woody as found in /lex
pernyi Franch., and finally to stony as found in /lex latifolia Thunb.

The sculpture of the pyrenes of //ex exhibits a very interesting set of
variations. This character is very valuable in distinguishing species. The

1949] HU, THE GENUS ILEX IN CHINA 253

descriptive terms used in this work concerning the surface of the pyrene
need some definition: (1) estriate and esulcate pyrenes are those that are
smooth, coriaceous, free from any vascular bundles of the pericarp, as in
Ilex yunnanensis Franch.; (2) striate and esulcate pyrenes are those that
are smooth, coriaceous, but with vascular bundles loosely attached to the
surface, as in Ilex longecaudata Comber; (3), striate, esulcate pyrenes
with impressed striae are those that are smooth, coriaceous, but with the
vascular bundles slightly sunk into the endocarp, as in //ex crenata Thunb. ;
(4) striate esulcate pyrenes with elevated striae are those that are
coriaceous and with the vascular bundles clinging to the surface of the
pyrenes, as in /lex metabaptista Loes.; (5) striate and sulcate pyrenes
are those that are sublignescent and with three elevated vascular bundles
evenly and longitudinally fused with the dorsal surface of the endocarp,
as in Ilex rotunda Thunb.; (6) palmately striate and sulcate pyrenes are
those that are woody, with the vascular bundles on the surface of the
palmate ridges of the endocarp, as found in //ex ficoidea Hemsl. (fig. 3, w);
(7) rugose, pitted, or wrinkled pyrenes are those that are stony, with irregu-
larly branched and anastomosing vascular bundles which extend over the
protuberances, as in Ilex subficoidea S. Y. Hu and Jlex chingiana Hu &
Tang (fig. 3, n); (8) unicanaliculate pyrenes are those that are sublig-
nescent or coriaceous and have a U-shaped cross-section (fig. 3, q & r),
as in llex lancilimba Merr. or Ilex chinensis Sims; and (9) the three-
ridged and two-canaliculate pyrenes are those that are stony, with the
endocarp so modified that the dorsal surface has two deep canals (fig. 3, p),
as in [lex macrocarpa Oliv.

These various types of pyrenes are of sectional or serial significance. In
the subgenus Prinos, the Pseudoprinos have smooth striate and esulcate
pyrenes with impressed striae. The Micrococca have unicanaliculate
pyrenes, the Prinoides have striate and sulcate or three-ridged and two-
canaliculate pyrenes. In the subgenus Euilex, the Chinenses have uni-
canaliculate pyrenes, the Umbelliformes have three-striate and two-sulcate
pyrenes; the Paltoria have smooth, striate esulcate, or estriate and esulcate
pyrenes, and the Aquifolium have palmately striate and sulcate or rugose
and pitted pyrenes.

GEOGRAPHIC DISTRIBUTION

The geographic distribution of the Chinese Aquifoliaceae reaches its
northern limit in the provinces of the Yangtze River basin; thence it
extends southward into Indo-China and India. Except for a form of Jlex
crenata Thunb., a native of Japan, introduced and cultivated in a park of
Tsing-tao, Shantung, and a few specimens of /lex pernyi Franch., from
Tsing- ifs a mountain range on the border of Szechuan, Kansu, and
Shensi, no Jlex has ever been recorded from either North or Northwest
China. Latitude 33° N. seems to be the northern limit for a natural distri-
bution of the Chinese Aquifoliaceae. From this latitude southward the
number of species occurring in each province is as follows: southern

254 JOURNAL OF THE ARNOLD ARBORETUM [VOL, Xxx

Kiangsu, 3; southern Honan, 1; southern border of Shensi and Kansu, 1
each; Chekiang, 23; Anhwei, 12; Kiangsi, 17; Hupei, 17; Hunan, 12;
Kweichow, 24; Szechuan, 18; Sikang, 9; Yunnan, 38; Kwangsi, 38;
Kwangtung, 26; Fukien, 13; Taiwan, 18; Hainan, 21; and Hongkong, 10.
Though some of the species are widely distributed, covering from six to
twelve provinces, and a few of them extend even to Japan and Korea in
the north, to Himalaya or North India in the west, and one of them to
Java in the south, the majority of them, especially those species in the
sections of Aquifolium, Lauroilex, and Pseudoaquifolium, are localized to
less than five provinces. Some of them, like //ex cinerea Champ. are
endemic even to a small coastal island like Hongkong.

Judging by the large number of species and the many intermediate forms
found there, it seems that South China and Southwest China, or, to be
more specific, the border region of Kwangtung, Kwangsi, Yunnan, and
Indo-China, is the center of the distribution of the Asiatic Jlex. From
there various species disperse radiately to form the Sino-Japanese, the
Sino-Himalayan, and the Sino-Indo-Malaysian elements of the genus.
This assumption is supported by much evidence. In the series Umbelli-
formes, [lex rotunda Thunb. is the most northern of the Sino-Japanese
elements; those occur in East China and Japan. It is an entire-leaved
glabrous species with striate sulcate narrow-elliptic pyrenes. Jlex excelsa
(Wall.) Hook. f. is the most southern of the Sino-Himalayan elements;
those occur in Southwest China and North India. It is also an entire-
leaved species, but with puberulous inflorescences and _ striate-esulcate
smaller but broader pyrenes. /lex rotunda var microcarpa (Lindl. & Paxt.)
Hu is an intermediate form that occurs in South China and northern
Indo-China, and is characterized by its puberulous inflorescences, small
broad-elliptic striate and slightly sulcate pyrenes. Among these three the
changes in the indumentum, the size and shape of the pyrenes, and the
sculpture of the endocarp are so gradual that it is sometimes difficult to
tell to which species a specimen collected from the geographically inter-
mediate provinces like Kwangsi belongs. Similar behavior occurs in the
series Stigmatophorae, involving the Sino-Japanese element Jlex crenata
Thunb., the Sino-Indo-Malaysian element J/ex triflora Blume, and the
intermediate species Jlex viridis Champ. of South China; in the series
Cassinoides, involving the Sino-Himalayan element J/lex yunnanensis
Franch. and the Sino-Japanese element J/ex sugeroki Maxim.; and in the
section Aquifolium, involving the Sino-Japanese element Jlex latifolia
Thunb., the Sino-Himalayan element lex denticulata Wall. ex Wight.,
and the Chinese species [lex intermedia Loes.

The altitudinal range of the Chinese //ex extends from sea-level in the
coastal provinces to a height of 3400 m. in the mountains of Yunnan. In
general, as the habitat changes from low to high altitudes, the habit of
the plants changes from trees to shrubs. In the subalpine Yunnan flora,
three prostrate or creeping species of //ex have been reported. Such creep-
ing forms have otherwise been recorded only from Yezo of northern Japan
and Sakhalin and the Himalaya.

1949] HU, THE GENUS ILEX IN CHINA 255

The Chinese //ex seems to illustrate a very interesting tie between the
flora of Hainan and that of Indo-China, which, from a geographic and
hydrographic standpoint, is to be expected. Examples of this are found in
Ilex pubitlimba Merr., Ilex cochinchinensis (Lour.) Loes., [lex kobuskiana
S. Y. Hu, and /lex hainanensis Merr. & Chun. The flora of Taiwan (For-
mosa) also has close relationship with that on the mainland, especially
that of Chekiang, Fukien, and, for certain high-altitude species, also
Yunnan. Numerous examples in //ex representative of this can be cited,
notably /lex pubescens Hook. & Arn., Ilex formosana Maxim., Ilex yun-
nanensis Franch., and Jlex boiritsensis Hayata.

Another interesting fact is that the Chinese Aquifoliaceae indicate a
certain affinity between the floras of eastern Asia and eastern North Amer-
ica. The fruiting branches of Ilex tsoit Merr. & Chun from East, Central,
and South China, resemble so closely those of lex montana Torr. & Gray
of eastern North America that only a specialist could distinguish them.
The same is true of /lex aculeolata Nakai of China and Ilex decidua Walt.
of eastern North America; /lex asprella (Hook. & Arn.) Champ. of South
China, Taiwan, and Luzon, and /lex longipes Chapm. of eastern North
America; [lex yunnanensis Franch. of China and //ex glabra (Linn.) Gray;
and Ilex serrata Thunb. var. sieboldi (Miq.) Rehd. of China and Japan
and J/lex verticillata (Linn.) Gray of eastern North America. In general,
however, there is a vast difference between the Chinese and the American
Ilex. In China more than ninety per cent of the species are evergreen,
with comparatively few deciduous species, while in eastern North America
most of the species are deciduou

The great concentration of fie in China is shown by statistics giving
the number of species found in Malaysia, as well as in adjoining countries
of Asia. Pitard in 1912 (Lecomte, Fl. Gén. Indo-Chine 1: 850-862)
admitted seventeen species for Indo-China, but by 1948 Tardieu-Blot
(Lecomte, Fl. Gén. Indo-Chine Suppl. 1: 759-781) increased the number
to thirty-seven. In the same general area Craib in 1926 (Fl. Siam. Enum.
1: 277-278) recognized but five species from Siam. The number occur-
ring in Burma is approximately fifteen. Hooker f. in 1875 (FI. Brit. Ind.
1: 598-606) admitted twenty-four species for the area he covered (India
proper, Ceylon, Burma, and the Malay Peninsula), but this number has
now been increased for the area by perhaps fifteen or twenty species.
Ridley in 1922 admitted seventeen species for the Malay Peninsula alone
(Fl. Malay. Penin. 1: 437-442). For the Philippines, Merrill in 1922
(Enum. Philip. Fl. Pl. 1: 476-480) recognized twenty-one species, but
undoubtedly some reductions remain to be made. In the Malay Archi-
pelago, outside of the Philippines, the area extending from Sumatra and
Borneo to New Guinea, about forty-five species have been described. Pos-
sibly a critical consideration of the Malaysian species will somewhat reduce
this number, although from a cursory examination of available material
from the Archipelago it seems apparent that a number of Malaysian species
have as yet not been described; eventually the total may be increased rather
than diminished.

256 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Because of the characteristic local endemism of a high percentage of the
Old World species of /lex and the paucity of species of relatively wide
distribution (and this applies to most of the species in the large genera
of flowering plants for the entire Indo-Malaysian region), no general dis-
cussion of geographic distribution of the Malaysian species seems to be
called for in this study of the Chinese forms. However, for all the species
of British India, as I understand specific limits, only about seven extend
to Southwest China, chiefly Yunnan Province; of the fifteen Burmese
species, ten occur in China; of the thirty-seven Indo-Chinese species about
twenty (many of them are new synonyms) extend to China. This rela-
tively high percentage of species not confined to a single geographic area
is understandable since northern Indo-China borders on southern China,
and - boundary between the two countries is merely one of a political
natur

nee considering those areas more distant from China proper, of the
five Siamese species two occur in Southwest China; only three of the
seventeen Malay Peninsula species extend to China; only two of the
twenty-one Philippine species have this range; and of the larger number
of species (approximately forty-five) characteristic of the Malay Archi-
arc proper and New Guinea, only one is known to extend to China.

As Taiwan (Formosa) is now politically again a part of China, I have
to the best of my ability considered the species of that island. A critical
comparison of our somewhat inadequate Formosan collections with the
vast amount of material available to me from China proper has led me to
reduce a certain number of the proposed Formosan species to synonymy.
As the island lies on the continental shelf, and as it was in geologic times
immediately preceding the present undoubtedly a part of the continent, it
is to be expected that the percentage of endemism for the Formosan flora
would be reduced on the basis of critical comparisons as material became
available for study. I have, in all, reduced eight of the proposed Formosan
species to synonymy. It is expected that the same results will attain in
other genera when it is possible to make critical comparisons of a consider-
able number of the described Formosan species with Chinese material. I
note in passing that both low- and high-altitude species are involved among
those Formosan species which I have reduced to earlier described conti-
nental species and varieties.

In this paper I have not considered the species of the Liukiu Islands and
Japan except in those cases where species originally described from those
areas have later been found on the continent. In the second edition of their
Flora of Japan (1931) Makino and Nemoto treated forty-nine species of
Ilex from Sakhalin, Japan, Liukiu, and Taiwan. In the Supplement Ne-
moto (1936) added nine more species to the flora of Japan. The lack of
comprehensive collections from. the area involved for purposes of com-
parison renders it unsafe for anyone to pass judgment on the validity of
this or that species, except on the basis of a critical study of the types.

In citing the specimens the following abbreviations have been used:
A= Arnold Arboretum, Harvard University; B— British Museum of

1949 ] HU, THE GENUS ILEX IN CHINA 20

Natural History, London, England; CB = Cambridge Botanic Museum,
Cambridge, England; CCC = Canton Christian College, Canton, China,
now Lingnan University; G= Gray Herbarium, Harvard University;
K = Royal Botanic Cardcne Kew, England; LU = Lingnan University,
Canton, China; NY = New York Botanical Garden, New York, N. Y.;
P= WMiniscur National d’Histoire Naturelle, Paris, France; SS = Biolog-
ical Laboratories, Science Society of China, Nanking: China: Sz = College
of Science, National Szechuan DEEVeNsity, Chengtu, China; TU = Na-
tional Taiwan University, Taipei, China; UN = University of Nanking,
Nanking, China; US = United States National Herbarium, Washington,
1D 0 Bee University of Tokyo, Japan.

TAXONOMY

Hex Linn. Sp. Pl. 125. 1753, Gen. Pl. ed. 5, 1754; DC. Prodr. 2: 13. 1825;
Benth. & Hook. f., Gen. Pl. 1: 356. 1862; Maxim. in Mém. Acad.
Sci. St. Pétersb. VII, 29(3): 14-53, pl. 1. 1881; Kronfeld in Engler
& Prantl, Nat. Pflanzenfam. III, 5: 183. 1896; Loes. in Engler &
Prantl, Nat. Pflanzenfam. Nachtr. 217. 1897, in Nov. Act. Acad.
Caes. Leop.-Carol. Nat. Cur. 78: 8-500, pl. 1-15 (Monog. Aquif. 1:)
1901, 89: 20-312 (Monog. Aquif. 2:) 1908, in Sarg. Pl. Wils. 1:
67-82. 1911, in Engler & Prantl, Nat. Pflanzenfam. ed. 2, 20: 36
1942; Comber in Notes Bot. Gard. Edinb. 18: 37-62. 1933.
Flower small, inconspicuous, regular, heteromerous, unisexual by abor-
tion. Staminate flowers: calyx patelliform, 4-6-lobed; corolla rotate, with
petals 4-8, slightly connate at the base; stamens usually as long as the
petals, epipetalous; anthers oblong-ovate with rounded base; rudimentary
ovary subglobose or frequently pulvinate, with a cleft beak, glabrous or
puberulous. Pistillate flowers: 1—18-(usually 4-8)-merous; calyx per-
sistent, 4-8-lobed. Petals connate at the base and spreading or distinct
and suberect; staminodes epipetalous, small, half or two-thirds the length
of the petals, the sterile anthers sagittate or cordate; ovary ovoid, 1—10-
usually 4—-8-loculate, glabrous or rarely pubescent; style rarely developed,
stigma discoid, capitate or columnar. Fruit a bacco-drupe, usually globose,
with a membranous or chartaceous exocarp, a fleshy mesocarp, and distinct
coriaceous, woody or stony endocarps. Pyrenes (endocarps) 1-18, usually
4—6, smooth, striate, striate-sulcate or rugose and pitted, 1-seeded.
Dioecious trees or shrubs. Leaves alternate, rarely opposite, deciduous
or evergreen, the margin entire, serrate or spinose, the stipules minute,
callose, usually persistent. Tnfigrescences cymose, simple or much
branched, solitary and axillary on the current year’s growth or fasciculate
and axillary on the second year’s growth. Flowers white, pink, or red.
Mature fruit red or blac
Over four hundred species distributed in the tropic, subtropic, and warm
temperate zones of both hemispheres. One hundred and twelve species,
belonging in two subgenera, Prinos and Euilex, occur in China. The two
subgenera are well-marked groups and probably represent diverging evolu-

208 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

tionary lines within the genus. Which is the more primitive can be deter-
mined only after the genus has had a thorough anatomical and cytological
study. I have placed Prinos before Euilex and so have departed from
general practice. For publication of my paper this has been most practical.
It does not imply necessarily that I believe Prinos to be the more primitive
subgenus.

KEY TO THE SUBGENERA

A. Deciduous trees or shrubs with membranaceous, chartaceous or rarely
subcoriaceous leaves ; branchlets usually with both elongated and abbre-
viated shoots; lenticels usually conspicuous on the current year’s
Oats ace na eed ae eee arora ee eee ee A. Subgenus Prinos.

AA. Evergreen trees or shrubs with thick-coriaceous, coriaceous, or rarely
chartaceous leaves; branchlets without abbreviated shoots; lenticels usu-
ally absent on the current year's OTOWLN: ca cc edas B. Subgenus Euilex.

A. Suxscenus PRINOS (Linn.) LoESENER

Ilex subgen. Prinos (Linn.) Loes. Vorst. Monog. Aquif.-Diss. 25, 26.
1890 “Prinus,” in Verh. Bot. Ver. Brand. 33: 25, 26. 1891 ‘“Prinus,”
in Engl. & Prantl, Nat. Pflanzenfam. Nachtr. 221. 1897, et in Nov.
Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78: 465 (Monog. Aquif.
1: 465). 1901 ‘‘Prinus’”; Rehd., Man. Cult. Trees & Shrubs, ed.
2, 551. 1940, Bibl. Cult. Trees & Shrubs 402. 1949.

Prinos Linn. Sp. Pl. 330. 1753, Gen. Pl. ed. 5, 153. 1754.

Shrubs or trees usually with both elongated and abbreviated shoots, the
lenticels usually conspicuous on the current year’s growth; leaves deciduous,
membranaceous or chartaceous, rarely subcoriaceous, the margin serrate
or crenulate, rarely subentire, the pistillate inflorescence solitary.

KEY TO THE SECTIONS

A. Mature fruit red; the pyrenes smooth with coriaceous, rarely woody,
endocar

B. Branchlets with definite abbreviated shoots; fruit strongly depressed-

globose with a capitate or cristate stigma, the style evident; pyrenes

ae longitudinally striate, the striae slightly impressed, the endo-

Cat WO0CYscctrebvakd oe ete ete CES Ge wee Sect. I. Pseudoprinos.

BB. Branchlets without abbreviated shoots; fruit globose, with a discoid

stigma, the style ne pyrenes 4-8, smooth, not striate, the
endocarp coriaceou

C. Pistillate Pee in trichotomous cymes or pseudo-umbels;

the inflorescence with 10 or more flowers..................

REEL KELEL EERE R REE RRR EE a Rate aes ect. II. Mucrococca.

CC. Pistillate flowers solitary or in 2- or 3-flowered cymose or

pseudofasciculate inflorescences......... Sect. HI. Euprinos.

AA. Mature oS black; the pyrenes rugose, striate, sulcate or canaliculate ;

the Gncdocal p: StOUy 52. as peace naenwheaie shee Sect. IV. Prinoides.

1949] HU, THE GENUS ILEX IN CHINA 299

Section I. PSEUDOPRINOS, sect. nov.

Frutices vel arbores parvae; ramulis elongatis abbreviatisque, lenticellis
conspicuis; foliis membranaceis vel chartaceis, serratis; inflorescentiis
paucifasciculatis vel solitariis; floribus 6—16-meris; ovario stylifero, stig-
mate capitato vel cristato; Pere ibuc valde depresso- elobosis: pyrenis 6-13,
laevibus, striatis, endocarpio sublignescente.

One species in China.

Aen
—-— PSEUDOPRINOS
— /MICROCACCA
“EU PRINGS
---PRINOIDES

ol. asprella eI. ma crocarpa
o On Chapaensrs al. nis

)
&
StF? Caley AL macropoda +1

I'ic. 4. Geographic distribution of the Eastern Asiatic representatives of
the four sections of the subgenus Prinos with the species of the section
Prinoides shown in detail.

1. Ilex fragilis Hook. f. forma kingii Loes. in Nov. Act. Acad. Caes.
Leop.-Carol. Nat. Cur. 78: 493 (Monog. Aquif. 1: 493). 1901, in
Sarg. Pl. Wils. 1: 82. 1911.

Ilex burmanica Merr. in pee 102. 1941. Syn. nov.

Ilex optenensis S. Y. Hu in Ic. Omei. 2: pl. 173, 1946. Syn. nov.

A deciduous shrub or et tree up to 5 m. high with both elongated
and abbreviated branchlets, membranaceous or chartaceous ovate serrate
leaves with pubescence beneath, Ses short pedicellate fruits with 6-13
small smooth striate woody pyren

Branchlets castaneous or ete elongate shoots 7-15 cm. long or
even longer, angular, smooth, shiny, rarely minutely puberulous in grooves
below the terminal buds, otherwise glabrous, the lenticels conspicuous,
elliptic, white, the terminal buds conic, with glabrous, strongly ciliate scales;
abbreviated shoots 10-15 mm. long, very rugose with persistent bud-scales
and leaf-scars; leaves 15-35 mm. apart on the elongated shoots, 1 or 2,
rarely 3, crowded at the tip of the abbreviated shoots; stipules minute,
broadly deltoid, persistent; petioles 6-12 mm. long, one-ninth to one-eighth

260 JOURNAL OF THE ARNOLD ARBORETUM [yOu Sxe

as long as the lamina, glabrous, narrowly canaliculate above, the distal end
narrowly winged by the decurrent leaf-base; lamina chartaceous or mem-
branous, olivaceous, opaque on both surfaces, hirsute along the veins on
both surfaces, rarely glabrescent, ovate or ovate-elliptic, 5.5-14 cm. long,
2—5.5 cm. wide, rounded or obtuse at the base, acute or acuminate at the
apex, the acumen 3-20 mm. long, serrate, the very tip cuspidate; the
margin serrate, the teeth apiculate and nigrescent; midrib slightly im-
pressed or plane, puberulent or glabrescent above, elevated and hirsute
beneath, rarely glabrescent, the lateral nerves 8 or 9 pairs, slightly
hirsute, rarely glabrescent beneath, the reticulations of the veinlets obscure
above, evident beneath. Staminate inflorescence: fasciculate or solitary
and axillary to scales or basal leaves of the elongate shoots, the individual
branches of the fascicles uniflorous; pedicels 4-6 mm. long, glabrous;
flowers 6—8-merous; the calyx patelliform, glabrous, 3 mm. across, deeply
6—8-lobed, the size and shape of lobes of the same flower varying greatly,
1 mm. long, ciliate, acute, rarely obtuse; corolla rotate, 6 mm. across, the
petals oblong, 2 mm. long, very minutely ciliate, one-sixth connate at the
base; stamens one-half the length of the petals, the anthers ovoid-oblong,
1 mm. long; rudimentary ovary pulvinate, depressed in the center.
Pistillate inflorescence: solitary, axillary to the scales or rarely to the
leaves; the pedicels 2-3 mm. long, up to 5 mm. long after fruiting,
glabrous; flowers 6—16-merous; calyx patelliform, 4 mm. across, deeply
6—8-lobed, the lobes acute, 1 mm. long, ciliate; corolla subrotate, the petals
oblong, 2 mm. long, one-eighth connate at the base; staminodes one-third
as long as the petals, the sterile anthers cordate; ovary pulvinate, the
style evident, up to 1.5 mm. long, glabrous or puberulous, the stigma
capitate or comb-shaped. Fruit depressed-globose or discoid, 4 mm. long,
5-6 mm. in diameter, the persistent calyx explanate, the stigma discoid,
capitate or cristate. Pyrenes eK broadly elliptic or subglobose in
outline, the ends obtuse, 2—2.5 mm. long, 1.5 mm. wide, pen longitudi-
nally striate, the striae slightly eed the endocarp wo

CHINA: Szechuan: Opien-hsien: W. C. Cheng pe hes ee
Law 2233 (A); C. WV ae 2739 (55);.2005 (SS) 4320 (es), 432F he
T. S. Chao 148 (SS), 615 (SS), 734 (SS); Mt. Omei(?), H. C. Chow
12307 (A), 12321 (A); Wa-wu-shan, £2. H. Wtson 692 (A), 334 (A);
C. W,. Yaw. 2300 (35), 3/00 (S38), 3042: (S5)3 danneean, 2. AA iison
4580 (A). Sikang (Southeastern Tibet): Tsarong, G. Forrest 20298
(A), 21808 (A); J. F. Rock 10208 (A, US), 22015 (A, US), 22482 (A);
Ta-chien-lu, &. H. Wilson 892A (A). Yunnan: &. FE. Maire 44 (A);
Wei-hsie, J. F. Rock 17058 (A, NY, US), 17168 (A, NY); Liang-shan,
fee = Tsai 51233 (A); Yi-liang Asien, H. T. Tsai 52128 (A); Ping-pien-
hsien, H. 7. Tsai a (A), 62680 (A); Salwin-Kiukiang Divide, G. For-
te 20298 (US), 21808 (US); T. T. Yu 19237 (A), 19289 (A), 20306 (A);
Tarulaka, T. T. Yu 20920 (A).

UPPER BURMA: Adung Valley, FP. K. Ward 9559 (Type of Ilex
burmanica, A), 9583 (A).

DIA: Sikkim, J. D. Hooker (fragment, A).

1949 ] HU, THE GENUS ILEX IN CHINA 261

Ilex fragilis forma kingi is endemic to Eastern Himalaya. It is common
on the high mountain of Sikkim, Upper Burma, and the Szechuan-Yunnan
border. There it grows as a shrub or small tree in thickets or woods at
altitudes of 1500-3000 m. The flowers appear in June, and the fruit turns
red in October.

Both of Kingdon Ward’s Specimens from Upper Burma are staminate
flowering shoots. The leaves are still young and not fully expanded, hence
they are smaller than those found on the fruiting specimens.

The membranaceous or chartaceous leaves, the variable number of
carpels on one shoot, and the styliferous ovaries of Jlex fragilis indicate a
close relationship with /lex macrocarpa Oliv. The latter, however, has
larger (over 10 mm. in diameter) fruits with ridged and canaliculate
pyrenes. Nevertheless, the staminate branches of the two species are
sometimes hard to distinguish.. The short fruiting pedicels of Ilex fragilis
suggest close relationship with //ex tsaii Merr. & Chun, but the latter has
deeply ridged stony pyrenes. When Loesener published the form kingz
he doubted that this pubescent plant could be a juvenile form of Jlex
fragilis. J have a fragment from Hooker’s cotype (Hooker & Thomson,
Khasia) in fruit. Its leaves are completely glabrous. At the same time
I have Wilson 892A before me. The fruits of this specimen are mature,
also. The lower surfaces of the leaves are pilose. I do not believe that
the presence and absence of hairs in this case is correlated with age.

Section II. MICROCOCCA (LoEs.), STAT. Nov.

Ilex subgen. Byronia (Endl.) Loes. series B. Micrococca Loes. in Nov.
Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78: 91 (Monog. Aquif. 1:
91). 1901.

Large trees with conspicuous lenticels on the current year’s growth,
membranaceous or chartaceous serrate leaves, solitary trichotomous
compound cymes or pseudo-umbels, small maker fruit, and 6-8 smooth
longitudinally canaliculate pyrenes with coriaceous endocarp.

Two species in China.

KEY TO THE SPECIES

A. Inflorescence a compound trichotomous cyme with the secondary axis
longer than the fruiting pedicels; leaves with 6-8 pairs of lateral nerves,
the petioles plane above. (East and Southeast China)..............
Pee ines err RO ee ila st lod gti’ aia 2. I. micrococca.
AA. Inflorescence a pseudo-umbel, the secondary axis usually lacking, when
present shorter than the fruiting pedicels; leaves with 10-20 pairs of
lateral nerves, the petioles canaliculate above. (West and South
(G)ovbn be eer neck i Geek ce eae re ee 3. I. polyneura.

2. Ilex micrococca Maxim. in Mém. Acad. Sci. St. Pétersb. VII, 29: 39,
pl. 1, fig. 6. 1881; Loes. in Nov. Act. Acad. Leop.-Carol. Nat. Cur.

78: 91 (Monog. Aquif. 1: 91). 1901; Chung in Mem. Sci. Soc.
China 1: 141. 1924; Anon. in Notes Bot. Gard. Edinb. 17: 115.

262 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

1926, 156, 268, 303, 365, 403. 1930; Merr. & Chun in Sunyats.
1: 68. 1930; Comber in Notes Bot. Gard. Edinb. 18: 55. 1933;
Masamune Fl. Kainant. [Hainan] 174. 1943; Tardieu-Blot in Fl.
Gén Indo-Chine Suppl. 1: 765. 1948.

Ilex micrococca var. longifolia Hayata Icon, Pl. Form. 3: 55, pl. 9. 1913;

Kanehira, Formosan Trees 377, fig. 334. 1936. Syn. nov.

Tree up to 20 m. high with conspicuous lenticels on the current year's
growth, the inflorescences cymose, trichotomous, solitary, with the second-
ary axis longer than the pedicels, the fruit globose, the pyrenes smooth with
coriaceous endocarp

Branchlets stout, very glabrous, second. year’s growth 1.5—5 mm. thick,
cinereous-brunnescent, longitudinally plicate and rather rugose with con-
spicuous large circular or oblong, often coalescent white lenticels and
slightly elevated obovate leaf-scars often united with the peduncle-scars
immediately above them; current year’s growth 2-4 mm. thick, shiny
castaneous, with conspicuous white lenticels. Leaves occurring only on
the current year’s growth, 1-3 cm. apart; stipules minute, broadly deltoid,
0.2 mm. long, 0.3 mm. wide at the base; petioles slender, 1.5—3.2 cm.
long, one-quarter to one-third the length of the lamina, plane above,
plicate-rugose beneath, glabrous; lamina membranous or chartaceous,
glabrous, olivaceous or brunnescent, dull above, paler beneath, ovate or
ovate- elliptic, 7-13 cm. long, 3-5 cm. wide, rounded or obtuse and often
oblique at the base, acuminate at the apex, the acumen 5-20 mm. long;
margin subentire or sharply aristate-serrate; midrib plane and toward
the base slightly sulcate above, prominent and elevated beneath, the
lateral nerves 5-8 pairs, the tertiary veins prominent, the reticulations
evident on both surfaces. Inflorescence cymose, solitary, or current year’s
growth only, axillary, twice or thrice trichotomous; bracts and bracteoles
on the secondary axis lacking, the bract on the tertiary axis broad deltoid,
minute, 0.4 mm. long, 0.5 mm. wide, acute, the bracts subtending the
individual flowers very minute and warty; peduncle glabrous, longer than

the secondary axis, 9-12 mm. long, plicate and in fruiting branches rugose;
pedicels very short, 2-3 mm. long, shorter than the secondary axis,
glabrous; prophylla 1 or 2, minute, warty, basal. Staminate flowers:
5- or 6-merous; calyx patelliform, 2 mm. Across, shallowy 5- or 6-lobed,
the lobes obtuse, glabrous, eciliate or sparsely ciliates corolla rotate, the
petals oblong, 1.25 mm. long, one-third connate at the base; stamens
equaling the petals in length, the anthers ovoid-oblong, 0.5 mm. long;
rudimentary ovary subglobose, rostellate with a 0.5 mm. beak parted to
the base. Pistillate flowers: 6-S-merous, very small, calyx patelliform,
2 mm. across, deeply 6-lobed, the lobes obtuse, ciliate; corolla rotate, 3 mm.
in diameter, the petals oblong, 1 mm. long, one-fourth connate at the base;
staminodes one-half as long as the petals, the sterile anthers sagittate;
ovary conic-ovoid, 1 mm. in diameter, slightly constricted below the discoid
stigma. Fruits globose, 3 mm. in diameter, the persistent calyx explanate,
orbicalar in outline, ciliate, the stigma thick-discoid, convex, distinctly

1949 ] HU, THE GENUS ILEX IN CHINA 263

6-8-lobed. Pyrenes 6-8, minute, elliptic in outline, 2 mm. lon
wide, the ends obtuse, the dorsal surface slightly roughened, longitudinally
unicanaliculate, the sides smooth, the endocarp coriaceous.

CHINA: Chekiang: Sia-chu, R. C. Ching 1691 (LU, ee se
(US); Tai-sun, R. C. Ching 2181 (A, US). Kiangsi: Lu-sha ,W.,
Chun 4291 (A). Kwangtung: S.S. Sin 9513 (LU, NY); Sin- te
Y. W. Taam 682 (A), 866 (A); Yu-yuen, S. P. 52688 (A); Lo- jane
C. L. Tso 20773 (NY). Hainan: Fan-yah, N. K. Chun & C, L. Tso
44097 (A, NY, US). Taiwan: Lake Jit een Y. Kudo in 1929 (A);
Taihoku, Nakamura 4055 Cae E. H. Wilson 10784 (A); without precise
eee = Seay in 1925 (T

JAPA Pounders Ge 73. (A), SOs Ge

Ilex micrococca was first described from material collected in Japan
and is common in the coastal provinces of east and south China and the
islands of Hainan and Taiwan, occurring also in Indo-China. It is a big
tree up to 20 m. high. The white flowers appear in May, and by October
its red fruit has fully matured.
2a. Ilex micrococca forma pilosa, f. nov.

Ilex ai Ie Franch. in Jour. de Bot. 12: 256. 1898.

Ilex micrococca sensu S. Y. Hu in Ic. Pl. Omei. 2: pl. 155. 1940, in part.

Arbor; pedicellis, cbs et etiam foliis subtus pilosis.

CHINA: Hupei (Hupeh): Enshih, H. C. Chow 1840 (A); Chang-
yang, E. H. Wilson 664 (A, B). Kw peo : Fan-ching-shan, Steward,
Chiao & Cheo 713 (A, NY, US). Szechuan: Tchen-kéou-tin, R. P.
Farges (P); Nan-chuan, W. P. Fang 5656 (A); Mt. Omei, T. S. Liu L259.
CANE TC wiee 40008 Cat pel .. I. ae 15144 (Sz), 15548 (Sz), ae
(Sz) 5076n(S2) TIO. (Sz), 19164 (Sz); So. Asie ore (sz) ek:
Wilson 3317 (K). Yunnan: An-ngy-tsin, Ducloux 2775 (P); Sze-mao,
A. Henry 11974A (NY); Meng-tze, A. Henry 13702 (NY); ¥. pay W,
Wang 85752 (A); Che-li, C. W. Wang 78193 (type, A); Y. Lin & C.W.
Wang 82895 (A); Ping-pien, H. T. Tsai 60843 (A); ee 6827 (A).
Kwangtung: Lok-chong, L. C. Tso 20724 (NY), 20810 ee
Kwangsi: WN. Lu-chen R. C. Ching 5900 (A, US), 6220 (NY);
Hin-yen, R. C. Ching 6915 (LU); Chen-pien, S. P. Ko 55887 (A); On- a
S. P. Ko 55788 (A); Lin-yuin, Steward & Cheo 328 (A, NY), 316 (A,
NY); Shang-sze, W. T. Tsang 22069 (A), 22145 (A), 24149 (A); Ping-

, C. Wang 39316 (A), 39923 (A).
O-CHINA: Tonkin: W. T. Tsang 27284 (A), 29274 (A).

This variety differs from typical //ex micrococca in having pubescent
Pees ae and sometimes the branchlets and the lower surface of
the lea
3. Ilex hore (Hand.-Mzt.), comb. nov.

Ilex micrococca Maxim, var. polyneura Hand.-Mzt. Symb. Sin. 7: 654.

1933

Tree up to 20 m. high (ex Yu), with conspicuous lenticels on current
year’s growth, chartaceous leaves, pseudo-umbelliform inflorescences, small

264 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

globose fruits, each with 6 or 7 pyrenes, the endocarp coriaceous, smooth
and unicanaliculate along the longitudinal median dorsal line.

Branchlets stout, very glabrous; second year’s growth 3.5-4.5 mm.
thick, cinereous-brunnescent, longitudinally plicate, the lenticels conspicu-
ous, white, orbicular or oblong, the leaf-scars semi-orbicular, slightly
elevated, closely associated with the oblong peduncle-scars; current year’s
growth glabrous, shiny, castaneous, 2-4 mm. thick, with conspicuous
elliptic lenticels. Leaves occurring only on current year’s growth, 5-15
mm. apart, the stipules scaly, rather large for the genus, 1.5 mm. long,
1 mm. wide; petioles one-fourth to one-third the length of the lamina,
1.5-2.8 cm. long, cylindric, deeply and narrowly canaliculate above,
puberulent in the groove; lamina chartaceous or thin-coriaceous, glabrous
above, minutely puberulent beneath, especially on the lateral nerves,
olivaceous-brunnescent, oblong-elliptic, rarely ovate-elliptic, 8-15 cm. long,
3.5-6.5 cm. wide; rounded or obtuse, rarely oblique at the base, acuminate
at the apex, the acumen 5—15 mm. long; margin finely and sharply serrate;
midrib slightly impressed above, elevated beneath, the lateral nerves 11—20
pairs, plane above, elevated beneath, the tertiary nerves clear, the reticula-
tion of the veinlets evident on both surfaces. Trnflgneseences pseudo-
SOT solitary, on the current year’s growth, axillary, secondary
axis usually not evident, but if developed, then shorter than the pedicels;
peduncular bracts none, bracteoles subtending the individual flowers
prophylla-like, basal, minute, broadly deltoid, brunnescent; peduncles
slightly compressed, gradually enlarged at the distal end, 6-9 mm. long,
puberulous; flowers 6- or 7-merous. Staminate flowers: pedicels 2-3 mm.
long, puberulous, with 2 minute basal prophylla; calyx patelliform, 2 mm.
across, deeply 6—7-lobed, the lobes deltoid, erose, eciliate; corolla rotate,
4 mm. across, the petals ovate, 2 mm. long, one-fourth connate at the base;
stamens equaling or slightly shorter than the petals, the anthers oblong,

m. long; rudimentary ovary pyramidal, 1 mm. long, the apex rostellate,
the beak cleft. Pistillate flowers: pedicels 3 mm. long, after fruiting
4-5 mm. long, with minute acute basal prophylla; calyx 2 mm. across,
deeply 6- or 7-lobed, the lobes deltoid, acute, erose, eciliate, rarely
minutely and sparsely ciliate; corolla rotate, 4 mm. across, the petals
oblong, 1.5 mm. long, one-fifth connate at the base; staminodes very short,
one-half as long as the petals, the sterile anthers sagittate; ovary ovoid,
2 mm. in diameter, _ stigma discoid. Fruit globose, 4 mm. in diameter,
the persistent calyx 3 mm. across, the stigma discoid, convex. Pyrenes
7, small, elliptic in ae 2—2.5 mm. long, 1 mm. wine on the back and
narrowly unicanaliculate along the longitudinal median line, the endocarp
coriaceous.

CHINA: Sikang: Ya- G. ¥. Chee i254 (A) Yunnan:
Handel-Maszsetti 8407 (A, US): Djiou- ee Handel-Mazzetti 9349 (A);
Mengtze, A. Henry 10329A (B); Sze-mao, A. Henry 11953 (A); Tsing-
pian, H. T. Tsai 52526 (A); without erceice locality, G. Forrest 8003 (A),
8651 (A), 20055 CA): Tsatong, G. Forrest 20623 CA, Us) 2lc0s CA, U5),
22824 (A), 26627 (A, US); Shang-pa, H. T. Tsai 54374 (A), 54810 (A),

1949] HU, THE GENUS ILEX IN CHINA 265

ee (A), 59041 (A); Ping-pien, H. T. Tsai 60434 (A); Kung-shan,

C. W. Wang 66885 (A), 66913 (A); Fo-hai, C. W. Wang 77392 (A);
Shun-ning, T. T. Yu 16010 (A), 16197 (A); Ai-wa, T. T. Yu 22879 (A);
Salwin Valley, T. T. Yu 22918 (Type, A).

The description of the staminate flowers is drawn from Handel-Mazzetti
9349, and that of the pistillate flowers from 7. T. Yu 16010.

This species has a limited range of distribution. It is found only in
the mountainous area of western Yunnan and Sikang at altitudes of from
1500 to 2600 m. It flowers in May, and by December the fruits are
mature. No leaves have been observed on twigs of the second year’s
growth. This may mean that the species is deciduous, or at least that the
old leaves drop before the new ones develop.

In the color of the branchlets, the presence of white lenticels on the
current year’s growth, and the chartaceous leaves, //ex polyneura resembles
Ilex micrococca Maxim. These two species differ chiefly in their venation,
petioles, and inflorescences. In Jlex micrococca the lateral nerves are
6-8 pairs, the petioles plane above, and the inflorescence twice or thrice
trichotomous with prominent secondary axes which are always longer than
the pedicels. In /lex polyneura the lateral nerves vary from 11 to 20 pairs,
the petioles are deeply and narrowly sulcate above, and the inflorescences
are pseudo-umbelliform, but when secondary axes evolve, they are always
shorter than the pedicels. The last character is constant. The differences
impress me as sufficiently strong to warrant giving specific status to
Handel-Mazzetti’s variety.
3a. lex polyneura var. glabra, var. nov.

Arbor; ramulis glabris, castaneis, lenticellis conspicuis; foliis chartaceis
vel pembranacei 7-15 c ongis, 3—7 cm. latis; inflorescentiis pseudo-
umbelliformis, plaice (pee globosis.

CHINA] Youmnan >” Menstze, A. Henry 10329 (AS), 10029C

A); Sze-mao, ie denim ii 953A (A), 119538 (A US), (29556 (ANY ).
11953D (A); without precise locality, G. Forrest 11886 (A), 17981 (A),
18/58 (A); Tali J. F. Rock 6888 (A, NY); Tang-vueh, J. F. Rock 2709
(A, US); Shang-pa-hsien, H. T. Tsai 54500 (A), 54626 (A), 54701 (A),
58988 (A); Tsang-yuan, C. W. Wang 71396 (A); Chen-kang, 7. T. Yu
1/241 (A); Mien: ning, T. T. Yu 17879 (A); Champutong, J. aa Rock
HOTS O MM CAG MIN NG raWiS)) = 11600 (A, US), 22001 (hyPp, A> NY); Tu
19166 (A); Tsarong, G. Forrest 19904 (A), 26652 (A, NY, ay 27417
(A).

This variety differs from typical lex polyneura in having glabrous
leaves and inflorescences. It has been reported only from Yunnan,
especially from the south and southwestern sections of that province.
There it grows as a common tree. The greenish yellow flowers appear in
June, and the fruit becomes scarlet in October.

Section III. EUPRINOS LorEseENER

Ilex subgen. Prinos Loes. sect. Euprinos Loes. in Engler & Prantl, Nat.
Pflanzenfam. Nachtr. 221. 1897, in Nov. Act. Acad. Caes. Leop.-

266 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Carol. Nat. Cur. 78: 465 (Monog. Aquifol. 1: 465). 1901 “Euprinus”’;
Rehd. Man. Cult. Trees Shrubs 551. 1940 “Euprinus,” et Bibl. Cult.
Trees Shrubs 402. 1949.

Ageria Adanson, Fam. Nat. Pl. 2: 166. 1763, pro parte.

Prinos Sect. Ageria (Adans.) DC. Prodr. 2: 17. 1825.

Ilex sect. Prinos (L.) Gray, Man. Bot. N. U.S. ed. 2, 264, 1856; Maxim.
in Mém. Acad. Sci. St. Pétersb. VII, 29(3): 30. 1881.
Shrubs with membranaceous or chartaceous leaves, solitary or sub-

fasciculate red fruits, and 4-6 smooth pyrenes with coriaceous endocarp.
One species in China.

4. Ilex serrata Thunb. var. sieboldi (Migq.) Rehd. in Bailey Cycl. Am.
Hort. [2]: 798. 1900; Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 468 (Monog. Aquif. 1: 468). 1901 “Sieboldii’’;
Makino, Ill. Fl. Nip. 368, fig. 1102. 1940.

Ilex sieboldi Mig. in Versl. Med. Kon. Akad. Wet. IH, 2: 84. 1868 [1866]
( 19. 1866), et in Ann. Mus. Bot. Lugd.-Bat. 3: 104. 1867;
Franch. & Sav. Enum. Pl. Jap. 1: 77. 1873; Maxim. in Mém. Acad.
Sci. St. Pétersb. VII, 29: 30, 48. 1881; Nakagawa in Bot. Mag.
(Tokyo) 13: 108. 1899.

Ilex subtilis Miq. in Versl. Med. Kon. Akad. Wet. II, 2: 84. 1868 [1866]

Repr. 19. 1866), et in Ann. Mus. Bot. Lugd.-Bat. 3: 107. 1867;
Franch. & Sav. Enum. Pl. Jap. 1: 78. 1875.

Ilex serrata var. subtilis (Miq.) Yatabe in Bot. Mag. (Tokyo) 62: 158.
1892; Loes. op. cit. 469 (1901); Koidz. Fl. Symb, Orient.-As. 8. 1930.
Syn. nov.

A very pubescent deciduous shrub 1-2.5 m. high with hirsute
branchlets, conspicuous lenticels, elliptic densely and argutely serrate
leaves, solitary cymose inflorescences borne behind an axillary bud on the
current year’s growth, short-pedicellate fruit, and 4-6 smooth pyrenes.

Branchlets subterete, cinereous, hirsute; third year’s growth 3-4 mm.
in diameter, longitudinally rimulose, the lenticels numerous, circular, leaf-
scars semi-orbicular, not elevated; second year’s growth 3-3.5 mm. in
diameter, puberulous, the lenticels conspicuous; the current year’s growth
1.5—2.5 mm. in diameter, hirsute, longitudinally plicate-canaliculate, the
lenticels conspicuous, the terminal buds ovoid, with puberulous and
crisply ciliate scales. Leaves occurring only on current year’s growth,
expanding a little before or during anthesis, 8-14, rarely 20 mm. apart;
the stipules linear, 2 mm. long, pilose, caducous; petioles 6—8 mm. long, one-
tenth to one-sixth the length of the lamina, plicate-rugose and “irsute,
deeply canaliculate above; lamina membranaceous, griseous-olivavesus,
sometimes brunneous, hirsute on both surfaces, elliptic, rarely ove‘e. or
obovate-elliptic, 2-8.5 cm. long, 1-4 cm. wide (usually 5 cm. long, 2 <m.
wide), acute or rarely obtuse at the base, shortly acuminate or rarely acute
at the apex, the acumen 3—5 mm. long, the very tip mucronate; margin
densely and argutely serrate; midrib elevated on both surfaces, hirsute,
the lateral nerves 6-8 pairs, obscure above, prominent beneath, hirsute
on both surfaces, the reticulation of the veinlets prominent beneath,

1949] HU, THE GENUS ILEX IN CHINA 267

hirsute. Inflorescences essentially cymose, solitary, on current year’s
growth, always behind an axillary bud, hirsute, very rarely subfasciculate
in the pistillate. Staminate inflorescences: cymes twice or thrice dichot-
omous or trichotomous, 9—21-flowered; peduncles 3 mm. long, the secondary
axis 1.5 mm. long, the pedicels 2—2.5 mm. long, with scale-like minute del-
toid basal prophylla; flowers 4- or 5-merous; calyx patelliform, 1.5-2 mm.
across, deeply 4- or 5-lobed, the lobes deltoid, obtuse or rounded, hirsute
and ciliate; corolla rotate, 4.5 mm. across, the petals oblong, erose, eciliate,
one-tenth connate at the base; stamens slightly shorter than the petals, the
anthers oblong, 1.25 mm. long; rudimentary ovary narrowly conic, 0.75
mm. long, acute at the apex, glabrous. Pistillate inflorescences: cymes
with 1-3 flowers, the peduncles 0O-1.5 mm. long, the pedicels 2-3 mm
long; flowers 4—6-merous; calyx as in the staminate flowers; corolla
rotate, 4.5 mm. in diameter, the petals ovate, erose, one-eighth connate
at the base; staminodes one-half as long as the petals, the sterile anthers
sagittate, obtuse at the apex; ovary ovoid, 1.5 mm. in diameter, glabrous,
the stigma discoid. Fruit globose, 5 mm. in diameter, solitary or 2 or 3
in a simple cyme, pedicels less in length than the diameter of the fruit,
2—2.5 mm. long, hirsute, the prophylla sub-basal, persistent, explanate,
suborbicular in outline, 5- or 6-lobed, the lobes round, hirsute and ciliate,
the stigma discoid, 5- or 6-lobed. Pyrenes 4 or 5, rarely 6, broadly elliptic
in outline, 2—2.3 mm. long, 1—-1.25 mm. wide, smooth, estriate-esulcate, the
endocarp coriaceous.

GEENA: Chekiang: Ching-yuen, R. C. Ching 2324 (A). Hunan:
Changsa, Y. Lin in 1947 (A), in 1948 (A). Szechuan: Kia-ting, E. H.
Wilson 3519 (A).

PAIN 2 Po rao Dorsett @ WW’. J. Morse 1490: (A) U0. -haurte-d3 CA),
9 (A); J. G. Jack in 1905 (G); RU ee in 1862 (NY); N. Mochi-
ae in 1904 (G); Stebold (TypE of /lex serrata var. subtilis, fragment, A);
. Shiota 6663 (A); E. H. Wilson 7043 (A), 7555 (G); K. Watanabe in
1892 (G).
CULTIVATED: Europe: Germany, Botanical Garden Darmstadt,
J. A. Purpus (A), America: United States, Arnold Arboretum 892,
5385 (A).

Ilex serrata var. steboldi is common in Japan. It differs from typical
Ilex serrata Thunb. in having hirsute leaf-surfaces. Specimens which
match perfectly the Japanese specimens have been collected at several
places in China.

I am treating Jlex sieboldi Miq. as only a variety of Zlex serrata Thunb.
It appear to differ from the Thunbergian species only in having hirsute
leaf-surfaces. After examining all the available Asiatic material of /lex
serrata, I belive that to be essentially a pubescent species. None of the
specimens has completely glabrous branchlets as stated by Thunberg,
“Rami teretes, laeves, glabri.”” In fact, many of them have pubescent
midribs, and all of them have puberulent pedicels. Nowhere in his
description did Thunberg mention the presence of hairs on the branchlets,
leaves, or inflorescences. Since I have not seen Thunberg’s original

268 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

specimen, I have followed his description and treated the Japanese plant
with sparsely puberulous branchlets, glabrous or glabrescent petioles and
lamina as typical Ilex serrata.

Miquel based his description of lex subtilis on a juvenile branch with
unfolding buds in which the leaves were not fully expanded. A fragment
of the type is in the Arnold Arboretum. It is certainly a juvenile form
of Ilex serrata var. sieboldi.

Ilex serrata Thunb. and J/lex verticillata Gray, the latter a North
American species of deciduous holly, are alike in their crisply hirsute
indumentum, their membranaceous or thinly chartaceous argutely serrate
leaves, their solitary, rarely 2- or 3-flowered infructescences, their ciliate
calyx and their smooth coriaceous pyrenes. ‘The latter species differs in
having 6—9-merous flowers. The texture of the leaves, the shortly pedi-
cellate infructescences, and the indumentum of J/lex serrata var. sieboldi
also simulate [lex pubescens Hook. & Arn. of China, but the latter has
fasciculate inflorescences on the old growth, a styliferous ovary, and striate
and sulcate woody pyrenes.

seoriow TV. PRINOIDES DC.) Gray

Ilex subgen. Prinos Loes. sect. Prinoides (DC.) Gray, Man. Bot.
N.U.S., ed. 2, 264. 1856; Loes. in Engl. & Prantl, Nat. Pflanzenfam.
Nachtr. 221. 1897, et in Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur.
78: 476 (Monog. Aquif. 1: 476). 1901; Rehd. Man. Cult. Trees
Shrubs ed. 2, 552. 1940, et Bibl. Cult. Trees Shrubs 403. 1949.

Prinos sect. Prinoides DC, Prodr. 2: 16. 1825.

Deciduous trees or shrubs with chartaceous, membranaceous or sub-
coriaceous leaves, black mature fruit, 4-9 rugose, striate, sulcate, or
2-canaliculate pyrenes with stony endocarp.

Key TO THE SPECIES

A. Pyrenes 3-ridged and deeply canaliculate on the dorsal surfaces; fruit
with capitate or columnar stigma, the style evident
. Leaves small, (3-)4-5(-7) cm. long; fruit 56 mim, in diameter ;
pedicels slender, 5 or 6 times as long as the diameter of the fruit;
stigma capitate. (East and South China and the near-by y islands)
OLE eee ee ee er ee rer ee eee eee ae aspr se
BB. Leaves large (5—)7-10(- See cm, sess fruit 12-14 mm. in dia

eter; pedicels never as much as 3 times as long as the ee

of ae fruit; stigma columnar
os Rudimentary ovary pulvinate, with slightly depressed center ;
pyrenes 7-9, laterally much compressed, the dorsal sur fate
1-2 mm. wide; style glabrous. (West, southwest, central, and
Oeste adele ous es tae eet aes es . J. macrocarpa.
. Rudimentary ovary conic with a rostellate apex; pyrenes 6 or
7, dorsal surface 4 mm. wide; style puberulous. (South
Chivie: anid iden 5 pe aes wee aes ace . I. chapaensis.

Ss

'?)

1949] HU, THE GENUS ILEX IN CHINA 269

AA. Pyrenes striate and sulcate, the striae of the dorsal surface often retic-
ulate; fruit with discoid (sometimes capitate in J. tsoii) stigma, the
style lacking.

B. Leaves obovate, (2—)3-4(-5) cm. long, cuneate at the base; pyrenes

sea Cane NIE cE Rein tt". 0a sg A Rl ae ae eee . aculeolata.

BB. Leaves ovate or ovate-elliptic, (4-)6-8(-10) cm. long an rounded
base ; pyrenes 5 or 6.

C. Pruiting pedicels 6-7 mm. long; pyrenes 5; calyx 2.5 mm.

; leaves with obscure reticulations on the upper surface.

(Gat and Central China and Japan)....... 9. I, macropoda.

ee eae pedicels 2-3 mm. long; pyrenes 6; calyx 4 mm.

; leaves with sharply defined areoles on both surfaces.

aa nal and. Sout Cilia arte acne koe LOL ts00t,

5. Jlex asprella (Hook. & Arn.) Champ. ex Benth. in Hook. Jour. Bot.
Kew Gard. Misc. 4: 329. 1852; Benth. Fl. Hongkong 65. 1861;
Maxim. in Meém. Acad. Sci. St. Pétersb. VII, 29: 49. 1881; Forbes
& Hemsl. in Jour. Linn. Soc. Bot. 23: 115. 1886; Henry in Transact.
As. Soc. Jap. 24 (Suppl.): 26. 1896; Loes. in Nov. Act. Acad. Caes.
Leop.-Carol. Nat. Cur. 78: 477 (Monog. Aquif. 1: 477). 1901;
Dunn & Tutcher in Kew Bull. Misc. Inf. Add. Ser. 10: 60. 1912;
Merr. in Sunyats. 1: 22. 1930.

Prinos asprellus Hook. & Arn. Bot. Beech. Voy. 176, pl. 36, figs. 1, 2

33; Walp. Rep. 1: 541. 1842.
Ilex ee Mig. in Jour. Bot. aie 1: 124. 1861; Forbes & Hemsl.
im jour: LintwSoc, Bot, 2a: 117. 18s
Ilex Ae Merr. in Philip. i. a Bota 257s lh oUs ron. ooe:
-» Elm: weatl, hilips Bot:'S: 1664. 1913.

Ilex merrillii Brig. in Ann. Conserv. Jard. Bot. Genéve 20: 421. 1919.

A deciduous shrub up to 3 m. high with slender elongate and short
abbreviated branchlets, membranaceous ovate acuminate serrate leaves,
long-pedicellate solitary fruits and 4-6 ridged and deeply sulcate bony
pyrenes.

Branchlets glabrous, castaneous, the lenticels on older portions con-
spicuous and numerous; elongated shoots slender, the third year’s growth
1.5-2 mm. in diameter, the lenticels lacking, the second year’s growth
1 mm. in diameter, the current year’s growth 0.7 mm. in diameter, all
glabrous, the abbreviated shoots 3-15 mm. long, rugose with persistent
bud-scales and scars, the terminal buds very narrowly conic, acute,
glabrous. Leaves 5-10 mm. apart on elongated shoots or 1—4 crowded at
the tips of abbreviated shoots; stipules callose, deltoid, acute, persistent;
petioles 3-8 mm. long, one- ninth to one-sixth the length of the lamina,
narrowly canaliculate and sparsely puberulent above, the rest glabrous;
lamina membranaceous, olivaceous, opaque on both surfaces, minutely
pubescent above (sometimes only on the veins), glabrous beneath, ovate
or ovate-elliptic, 3-7 cm. long, 1.5-3 cm. wide, obtuse, rarely rounded, or
acute at the base, acuminate at the apex, the acumen 3-12 mm. long,
serrulate on the basal portion, the very tips acute or mucronate; margin
serrate; midrib slightly impressed and hirsute above, elevated and glabrous

270 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

beneath, the lateral nerves 6-8 pairs, evident and hirsute above, evident
and glabrous beneath, the reticulation of the veinlets obscure. Staminate
inflorescences: paucifasciculate with 2 or 3 flowers or solitary, axillary to
the leaves or scales, when axillary to the leaves often behind an axillary
bud; pedicels (3—) 5-6 (-—9) mm. long; flowers 4- or 5-merous; calyx
patelliform, 2.5-3 mm. across, glabrous, shallowly 4- or 5-lobed, the
lobes unequal, broadly deltoid or rounded, erose and ciliate; corolla rotate,
6 mm. across, the petals 4 or 5, suborbicular, 2 mm. in diameter, rarely
ciliate, one-fifth connate at the base; stamens 4 or 5, three-fourths as long
as the petals, the anthers oblong, 1 mm. long; rudimentary ovary pulvinate,
the middle shortly rostellate. Pistillate inflorescences: solitary, axillary
to the leaves; pedicels 2—2.5 cm. long, after fruiting up to 3 cm. long,
glabrous, with 2 minute basal prophylla; flowers 4—6-merous; calyx 3 mm.
across, deeply 4—6-lobed, the lobes ciliate; corolla rotate, the petals sub-
orbicular, 2 mm. in diameter, one-fourth connate at the base; staminodes
one-half the length of the petals, the sterile anthers sagittate; ovary
subglobose-ovoid, 1.5 mm. in diameter, the style evident, the stigma thickly
discoid. Fruit globose, 5-6 mm. in diameter, longitudinally striate and
sulcate, the persistent calyx explanate, 3 mm. across, suborbicular in out-
line, ciliate, the stigma capitate, the style evident. Pyrenes 4—6, obovate-
elliptic in outline, the ends obtuse, 5 mm. long, 2 mm. wide, 3-striate,
ridged, canaliculate, the sides almost smooth, striate, the ventral keel
sharp, the endocarp stony.

CHINA: Chekiang: Sia-chu, &. C. Ching 2210 (A, LU, US);
Tai-shun, Y. L. Keng 267 (A). Kiangsi: Woo-kung-shan, H. H. Hu
758 (A); Tsu-chi-hsien, H. H. Hu 1247 (A); Tsoong-jen, Y. Tstang 10218
(NY). Fukien: Central Fukien, Dunn (Herb. Hongk. no. 2463) (A);
Chang-chow, H. H. oe ae (A), 1214 (A, LU); Ing-hok, H. H. Chung
2621 (A), 3227 (A), 7988 (A); Yeng-ping, H. H. Chung 3354 (A);
Ku-tien, H. H. Chung Heh (A); Ku-hang, H. H. Chung 6776 (A oe
Ku-dien, H. H. Chung 7920 (A); Foochow, H. H. Chung 8916 (A
NY); F. P. Metcalf 845 (LU), 3518 (A, LU); Tang Siu-ging i
(LU), 15579 (LU), 75600 (LU). Kwangtung: Canton, C. O. Levine
oe 437 (US), 478 (A, US), 638 (A, US), 2102 (A); Mei-hsien,

_L. Gressitt 1205 (A), 1324 (A); Tung-wan, S. Y. Lau 20040 (LU, NY):
cid shan, Fung-hom 24 (LU 18449) (NY); W. T. Tsang (LU wee)
(NY)? Sung @ McClyre 19 (LU 19325) (NY); ee Rae shan, t. 2,
Merril 10728 (A); Wai-yang, T. ig Tsut 146 (A Pm 2-9 e ee
7. M,. Tsui 195 (US); Ying-tak, M. Tsui 317 (A, — US); Ta-pu,
IW. T. Tsang 21081 (A, NY); ae (2 NY 3 Ho-yuen, W. T. Tsang
28707 (A); Sin-tong, Ko-chow, Y. Tstang 969 (A) ; Lo-fou-shan, Y. Tsiang
1689 (A); Lui-chow, Y. ee 2208 (NY), 2558 (NY); Kwu-dzu, Tse-
kung C. L. Tso 41473 (NY); Sun-wui, Tso & Tsiang 2019 (A). Kwang-
si: Tang-hsien, G. W. Groff (CCC) 4735 (LU). Hongkong :
Hensiers New Territory, W. Y. Chun 6250 (A); without precise locality,
Faber 116 (A); Aberdeen, Y. W. Taam 1985 (A); Bridge Valley, Y.
Tsiang 63 (A); Hwang-tso-kong, Y. Tsiang Ou a NY); without pre-
cise locality, C. Wright (NY). Taiwan Pave 129 (Ast,
Hayashi 21216 (A); A. Henry 221 (A, NY), a (Ady oe (Ae NY:

1949] HU, THE GENUS ILEX IN CHINA 271

1334 (US); T. Hosokawa 9893 (TU); J. L. Gressitt 86 (A, NY); &.
Matuda 44 (A); T. Nonaka & K. Mori in 1932 (TU); S. Sasaki 21569
A); H. Simada 335 (TU); S. Suzuki in 1931 (A); T. Sugzuks in Be

U); T. Tanaka 76 (A, NY); T. Tanaka & Y. Shimada (Herb.
11009) (NY, US), (Herb. no. 11094) (A, NY, US); E. H. Wilson 10063
CAIUS). 10293 CA WS) 2160 CA) -

Ilex asprella is a wide-spread deciduous species in the coastal provinces
of warm temperate and subtropical China including Taiwan (Formosa)
where it occurs at low altitudes. The white flowers appear in March, and
the fruit turns black in October. It is common in Taiwan and extends
to the Philippine Islands where it occurs at altitudes of from 1200 to 1900
meters. It is the only deciduous species of //ex known from the Philippines
and there it flowers also in March. I agree with Merrill in his reduction
of Jlex gracilipes Merr. and Ilex merrillii Brig. as synonyms of the common
Chinese species. Among the many Philippine collections I may cite
Clemens 51888; D. E. Elmer 9515A, 14297; E. C. Leano 22914; A. Loher
13020; Ramos & Edano 38079, 45061, 47271, J. K. Santos 7 and 31735;
all specimens seen in the Arnold Arboretum.

In its membranaceous leaves and its solitary staminate or pistillate
flowers, each behind an axillary bud on elongated branchlets, /lex aspredla
shows close relationship with J/ex serrata Thunb. The latter, however, has
smooth pyrenes and small, shortly pedicellate fruits. The texture of the
leaves and the sculpturing of the pyrenes of Ilex asprella show close
relationship with Z/ex aculeolata Nakai, but the latter has obovate leaves
and very shortly pedicellate fruits.
5a. Ilex asprella var. tapuensis, var. nov.

Frutex; ramulis glabris; foliis membranaceis, ovato-ellipticis, 4-8 cm.
longis, 1.7-3.4 cm. latis, basi obtusis vel rotundatis, apice acuminatis,
acuminibus 3-15 mm. longis, mucronatis, costa supra impressa et puberula;
inflorescentiis fructiferis solitariis, pedicellis 2 cm. longis, glabris; fructibus
ellipsoideis, 9 mm. longis, 8 mm. diametro, calycibus perdistenuibue sub-
explanatis, 4 mm. diametro, ciliatis; stigmate crasse discoideo; pyrenis
6, 8 mm. longis, 3 mm. latis, dorso 3-striatis, 2-canaliculatis, lateralibus
1- vel 2-striatis et sulcatis, endocarnio lapideo.

GHINAY eKavame tune: Ta-pu, 7. 7; Tseng 2izio vee, ak):

The only specimen we have seen was collected at Ta-pu, E. Kwangtung.
There the plant grows as a shrub up to 3 m. high. The fruit is black at
maturity.

This variety could be either a form of lex macrocarpa Oliv. var. longi-
pedunculata S. Y. Hu with small leaves and fruit or a large-fruited variety
of Ilex asprella. I prefer the latter position, since the fruit lacks the
characteristic columnar stigma of J/ex macrocarpa Oliv.

6. lex macrocarpa Oliv. in Hook. Ic. Pl. 8: pl. 1787. 1888; Loes. ex
Diels in Bot. Jahrb. 29: 436. 1900, in Nov. Act. Acad. Caes. Leop.-
Carol. Nat. Cur. 78: 489 (Monog. Aquif. 1: 489). 1901; Pamp.
in Nuov. Giorn. Bot. Ital. n. s. 17: 147. 1900; Dunn & Tutcher in

272 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Kew Bull. Misc. Inf. Add. Ser. 10: 60.1912; Loes. in Sarg. Pl. Wils.
1: 81.1911; Pamp. in Nuov. Giorn. Bot. Ital. n. s. 17: 417. 1910;
Rehd. in Jour. Arnold Arb. 8: 158. 1927, 14: 242. 1933; Chun in
Sunyats. 1: 181. 1933; Hand.-Mzt. Symb. Sin. 7: 659. 1933;
Comber in Notes Bot. Gard. Edinb. 18: 54. 1933; Tardieu-Blot in
Fl. Gén. Indo-Chine Suppl. 1: 780. 1948.
Ilex henryi Loes, in Noy. Act. Acad. Caes. Leop.-Carol. 78: 491 (Monog.
Aquif, 1: 491). 1901. Sam. nov.

lige dasa Oliv. var, trichophylla Loes. ll.ce. 491. 1901, 81. 1913.
Syn. nor

llex macrocarpa Oliv. genuina Loes. ll.cc.; Comber in Notes Bot.
Gard. Edinb. 18: 54. 1933.) Sym. nov.

llex i (G. Don) Trel. var. hupehensis Loes, 1c. 488. 1901. Syn. nov.

flex montana Torr. & Gray var. hupehensis Fern, in Rhodora 41: 428.

: Svit. nov.

Celastris salicifolia H, Lévl. in Fedde, Rep. Spec. Nov. 13: 263. 1914.

Diospyros Bodintert H. Lévl. Fl. Kouy-Tchéou. 144, 1914.

[lex macrocarpa Oliv. var. aa ntpeduncilata S. Y. Hu in Ic. Pl, Omei. 2:

pl. 171, 1946. Syn. nov.

A deciduous tree with both elongated and abbreviated branchlets, ovate,
ovate-elliptic, or rarely oblong-elliptic leaves, solitary or pseudofasciculate
staminate inflorescences, solitary large fruits with 5-9 laterally compressed
striate-sulcate pyrenes.

Branchlets brunneous; third and second years’ growth 3 mm. in diameter,
the lenticels orbicular, conspicuous; current year’s growth glabrous, 2-3
mm. in diameter, the lenticels evident, the terminal buds subglobose-conic,
glabrous; abbreviated shoots weakly developed, 3-10 mm. long. Leaves
10-30 mm. apart on elongated shoots, 1-3 crowded at the tips of the
abbreviated shoots; stipules very minute, callose, often obscure; petioles
9-12 mm. long, one-tenth to one-seventh the length of the lamina, narrowly
canaliculate and minutely puberulous above; lamina chartaceous, oliva-
ceous, opaque on both surfaces, glabrous, glabrescent, or when young
minutely and sparsely pubescent, ovate, ovate-elliptic or rarely oblong-
elliptic, 5-15 cm. long, 3—7 cm. wide, rounded or obtuse at the base, shortly
acuminate at the apex, the acumen 3-10 mm. long; margin serrate, midrib
impressed and usually minutely puberulous above, elevated, glabrous, gla-
brescent or sparsely puberulous beneath, the reticulations of the veinlets
prominent on both surfaces. Staminate inflorescences: pseudofasciculate on
second year’s growth of elongated shoots and on abbreviated shoots, or
solitary, axillary to the basal scales or leaves of elongated shoots, 1- to 5-
flowered, cymose; peduncles 2—4 mm. long, the pedicels 2-7 mm. long, both
glabrous; flowers 5- or 6-merous; calyx patelliform, ca. 3 mm. across, shal-
lowly 5- or 6-lobed, the lobes deltoid-ovate, obtuse or rounded, ciliate;
corolla rotate, 7 mm. across, the petals obovate- oblong, 3 mm. lone, ciliate,
one-sixth connate at the bce stamens equaling the eee in length, the
anthers ovate-oblong, 1 mm. long; rudimentary ovary pulvinate, the center
slightly depressed, inconspicuously lobed. Distillate inflorescences: solitary,

1949 | HU, THE GENUS ILEX IN CHINA 273

axillary; pedicels 6-14 mm. long, glabrous, with 2 sub-basal prophylla;
flowers 7—9-merous; calyx patelliform, 5 mm. across, shallowly 7—9-lobed,
the lobes ovate-deltoid with obtuse or rounded apex, ciliate; corolla rotate,
10-12 mm. across, the petals 4-5 mm. long, one-fifth connate at the base;
staminodes two-thirds as long as the petals, the sterile anthers sagittate,
the apical end obtuse; ovary conical-ovoid, 2 mm. long, 3 mm. in dinmeter
at the base, the style evident, 1-1.5 mm. fone the stigma columnar, 1.5 mm.
long, plabrous: Fruit globose, 12-14 mm. in diameter, the persistent calyx
explanate, 6 mm. in diameter, orbicular in outline, the stigma columnar,
2 mm. long. Pyrenes 7—9, laterally compressed, the ends obtuse, 9 mm.
long, 2 mm. wide, on the dorsal surface longitudinally ridged and deeply
canaliculate, reticulately striate and sulcate on the sides, the endocarp stony.
CHINA: Anhwei: Chu-hwa-shan, R. C. Ching 2722 (A), 2732 (A),
2750 (A, LU). Hupei (Hupeh): A. Henry 2891 (isotype, A), 3445
(G), 3451 (A, US), 4179 (A), 6214 (asoryre of Ilex dubia var. hupehensis,
WS) e4GssuGsonver NY), (362, (A), 7/20 (US): Cees hsien, &. A.
Wilson 151 (A, US); without precise locality, E. H. Wilson 250 (NY),
2695 (NY, US); Pa-tung-hsien, E. H. Wilson 3089 (A, US). Hunan:
Handel-Mazzetti 575 (A). Kweichow: J. Esquirol in 1906 (TYPE of
Celastrus salicifolia, photo & fragment, A), 6094 (A, P); J. Chaffanjon in
1898 (fragment of type of Diospyros Bodinieri, A); Ta-ting, Y. Tsiang
8906 (NY). Cee Chien-yang-hsien, S. S. Chien 5249 (A, SS);
without precise vo E. Faber 27 (NY); Tchen-kéou-tin, R. P. Bares
1397 (CNP) MiOmet, C: VY. Chiao & S.0C. Bon S27C4), 115 (A);
SS Chien 9479) eee H.C. Chow 5842 (Sz), 5882 (Sz), 5883 (Sz), 5897
(Sz), 0902 (Sz), 7020 (AS27611 (A, ee 8247 (NSz)e 9726 (Sz), 11/41
KO) WP ange es4e (CAN 2386 CA, SS) e256 Che Cn. NY.)
P88 2 (US) 1223 0NGSS \.129/70: (SS); nee (Sz), 14484 (Sz), 14825 (Sz),
16421 (Sz), 16473 (Sz), 16663 (Sz), 16976 (Sz), 17232((Sz)7 17261 (Sz),
14/0 (Sz), 1815 (Sz) 18824 (Sz), 18895 (Sz) «PC. bee S157 (Sz) 3368
CS7) 0G alex dere 024 ae C. L. Sun 1898 (Sz) 71989 (Sz), 2007 (32);
Ln Vee Rae GR) . Vad 3320 (SS); $255 (SS es Wao 2723
CSS) ee, Yur 905 ce Te wai: fisien, S..S. Chaen 359% (Sz), 9701 (Sz);
ae TENT HO) TEEN aN) ara pea Ae ue 20797 (A); Chengtu, S. S. Chien 5255
(Sz), W. - Fang 12228 (A, SS, US), 12236 (A, SS), 13385 (SS), 19286
Cz Ga eiane: 7293 a. z H. Wilson 4795 (K); An-hsien, F. T.
WwW ite 22188 (A), 22208 (A); Lo-shan-hsien (Kia-ting), H. C. pee 9649
CA ae enung 216° CA), S. GC. a4 a K. Chang 820 (A); H. H. Tai 5

CA eee (CA ele (A) 3 . Wang 23553 (A); a T. Vu 218
(A); Opien-hsien, S. N. Hsu 30 oe io H. Wilson oe CxS) «S006
(CaF oe) unnan) FF, ee COo7 HX w one precise

locality, E. E. Maire 1487 (K, US); Mi-le, A. Henry Tea (A, US); Yan-
tai-hsien, H. T. Tsat 50810 (A); Hsi- ee ing, Chang-kiang, Y. Tenn &
H. Wang 16243 (A), 16244 (A); Er Ww ae 41495 (A); Kun-ming, C.
Wang 62807 (A), 62843 (A). Kwangtung: Tai-Yong, J. L. Gi a
1736 (A): Yao-shan, N.. River, S. S. Sun 0805 (INYo) a Si yi, Y. Tsiang
2660 (NY). Kwangsi: Luchow, R. C. Ching 5350 (NY).

CULTIVATED: I have seen specimens collected from Ae Arboretum
of ‘“Westonbirt, Glos,” England.

274 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

Ilex macrocarpa was first described from material said to have been
collected in Hupeh and Hongkong. In all the material we have studied
there is no representative from Hongkong. Dunn & Tutcher in Kew Bull.
Misc. Inf. Add. Ser. 10: 60. 1912, and Tutcher in Rep. Bot. For. Dept.
Hongk. Suppl. 20. 1916 may have copied Oliver’s error. Jlex macrocarpa is
essentially a tree of the warm temperate forests or woods and thickets of
the Yangtze Provinces. There its white and fragrant flowers appear in
April. The fruit turns black in November and is eaten by various birds,
especially the common white-head bulbul. Possibly birds have aided in
the dissemination of this widely distributed species.

[lex macrocarpa is closely related to Ilex chapaensis Merr., but the
latter has rostellate rudimentary ovaries, trigonous pyrenes, and_thick-
chartaceous leaves, often with shiny upper surfaces.

After examining many specimens from different parts of China, it is
evident that //ex macrocarpa is a very variable species. The indumentum
of the leaves, for example, varies considerably. Usually the blade of the
leaf is pubescent along the midrib on the upper surface, but in some plants
the blade is entirely glabrous and in others pubescent on both upper and
lower surfaces. The staminate inflorescences vary from solitary cymes
on vigorously growing young shoots to pseudofasciculate on older shoots.
The fruiting pedicels vary in length from those that are slightly shorter
than the petioles to those that are subequal to or longer than the petioles.
Both Oliver’s illustration and his description, “pedunculis axillaribus
petiolo subaequilongis,” indicate that the specimens he examined had
fruits with short pedicels. Henry 2981, an isotype, has leaves on which
the midribs are puberulous on both surfaces and the fruiting pedicels are
slightly longer than the petioles. Based on this evidence, among the host
of variants we can define the typical form of Jlex macrocarpa as the plant
with glabrous branchlets and inflorescences and fruits with pedicels sub-
equaling or slightly exceeding the petioles. In so doing, both /lex henryi
Loes. and flex macrocarpa var. brevipedunculata S. Y. Hu are to be
reduced to synonymy

Loesener ie. Ties dubia (G. Don) Trel. var. hupehensis, based on
A. Henry 6214. This specimen is a fruiting branch of /lex macrocar pa.
Although the fruit is very young, its large size and its columnar stigma are
very significant. I have dissected two fruits. Each has only 5 pyrenes.
This low pyrene number is uncommon in typical /lex macrocarpa. This
particular plant may be a hybrid between /lex macrocarpa and Ilex macro-
poda Miq., as both species are common in Hupei Province. Because of
the large leaves and fruit and the columnar stigma of the plant I think
it is proper to accept it as lex macrocarpa.
6a. Ilex macrocarpa var. reevesae (S. Y. Hu), comb. nov.

llex reevesae S. Y. Hu in Jour. W. China Bord. Res moe, 15(B)s

1945, et in Ic. Pl. Omei. 2: pl. 172. 1946. Syn. no

Branchlets minutely puberulent, leaves ovate, 3-10 cm. long, 2—5.5 cm.
wide, pubescent on both surfaces; inflorescences pubescent; pyrenes
laterally compressed.

1949] HU, THE GENUS ILEX IN CHINA 275

NAY Szechuan. - Chengtu, 5. S. Chren 5925..CA); ae ee Sze
W.P. Fang 12442 (Sz): See th "5354 CWC UW) 2555 B= OW GUE):
The distribution of this variety is very localized. So far, it ae been
recorded only from Chengtu in the province of Szechuan. There its white
and fragrant flowers appear in April. The fruit is black at maturity.
This variety is characterized by the pubescent branchlets, leaves, and
inflorescences.

6b. Liex macrocarpa var. longipedunculata S. Y. Hu in Ic. Pl. Omei.
2: pl. 171. 1946.

Branchlets glabrous; leaves ovate, ovate-elliptic, pubescent at least
along the midrib above; fruiting pedicels 14-33 mm. long, often more than
twice as long as the petioles; pyrenes 7-9, laterally compressed, 3-ridged
and deeply canaliculate, the endocarp bony

CHINA: Kiangsu: I-shing, Ching & Tso 594 (A). eee :
Shiao-fung-hsien, H. H. Hu 559 (A); Chong-hwa-hsien, Y. L. Keng 571
(A); Tien-mu-shan, 7. N. Liou 311 (NY); Yi-tsun, F. N. Meyer 1523 (A,
NY). Anhwei: Chiu-hwa-shan, S. C. Sun 1398 (A). Hupei
(Hupeh): W. Y. Chun 3525 (A), 2526 (A); Pa-tung-hsien, H. C. Chow
512 (A, NY), 665 (A, NY); Chien-shih-hsien, H. C. Chow 1087 (A, NY);
MiechuaneO. 2. wa 07 CA lechane, 2, Ay Wiisonso (Ae Ws) coro
(A). Hunan: Lin-ling, Handel-Massetti 423 (A). Kweichow:
Tsun-yi-hsien, Steward, Chiao & Cheo 108 (A, NY, US); Pi-chien, Tui-po,
Y. ESI GO Sea NY eo eee iim s ait Cha 1918 (SS). Kwang-

Kweilin, W. T. Tsang 28137 (A, US), 28249 (A, US).

This vane is quite wide-spread. It appears that the shortest-pedi-
cellate forms are found in the Upper Yangtze Provinces of Szechuan and
Yunnan. The long-pedicellate forms are found in the Lower Yangtze
Provinces of Kiangsi and Chekiang. A mixture of both the long- and
short-pedicellate forms is found in the Central Provinces of Hupei, Hunan,
and Kweichow.

This variety differs from typical [dex macrocarpa in having long fruiting
pedicels which are usually twice as long as the petioles.

7. Ilex chapaensis Merr. in Jour. Arnold Arb. 21: 373. July 1940;
Tardieu-Blot in Not. Syst. XII, 15: 8. 1945, et in Fl. Gén. Indo-
Chine Suppl. 1: 773. 1948.

Ilex howiit Merr. & Chun ex Tanaka & Odashima in Jour. Agr. Taiwan

: 372. 1938, nomen nudum; Merr. & Chun in Sunyats. 5: 107

August 1940; Masamune, Fl. Kainant. (Hainan) 174. 1943. Syn. now.

Tlex es Merr. in Jour. Arnold Arb. 21: 373. 1940; Tardieu-
n Fl. Gén. Indo-Chine Suppl. 1: 780. 1948. Syn. nov.

A fee tree up to 10 m. high with minutely and sparsely puberulent
or glabrescent branchlets, chartaceous ovate-elliptic or oblong-elliptic
crenate leaves, solitary inflorescences, large fruits with columnar stigmata
and 6 oblong, ridged, and canaliculate pyrenes.

Branchlets castaneous, with numerous conspicuous lenticels; second
year’s growth 4 mm. in diameter, rugose with elevated orbicular lenticels,

276 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

the leaf-scars crescent-shaped, not elevated; current year’s growth 2-3 mm.
in diameter, very sparsely and minutely puberulent or glabrescent, slightly
angular, the lenticels conspicuous; abbreviated shoots few, undeveloped,
2-3 mm. long. Leaves 10-15 mm. (rarely up to 20 mm.) apart on
elongated shoots, 1 or 2 at the very tip of the abbreviated shoots; stipules
deltoid, callose, persistent; petioles slender, 15-30 mm. long, one-fifth to
one-third the length of the lamina, glabrous, narrowly and deeply canalic-
ulate, the distal end slighty winged by the decurrent leaf-base; lamina
chartaceous or thin-coriaceous, brunneous-olivaceous, opaque or very
slightly shiny and puberulous along the midrib and veins above, glabrous
beneath, rarely puberulous on both surfaces in juvenile forms, ovate-
elliptic or oblong-elliptic, 5-11 cm. long, 2.5-5 cm. wide, acute, obtuse,
rarely rounded at the base, acuminate or very rarely (on abbreviated
shoots) obtuse or rounded at the apex; margin finely crenulate-serrulate;
midrib narrowly impressed and puberulous above, glabrous and elevated
beneath, the lateral nerves 8 or 9 pairs, plane or slightly impressed, often
puberulous above, elevated and glabrous beneath, the reticulation of the
veinlets obscure on both surfaces. Staminate inflorescences: pseudo-
fasciculate, occurring at the base of the current year’s growth or along the
sides of the second or third year’s growth, the individual branches 1—5-

oblong, 2 mm. long; rudimentary ovary conic, the apex rostellate and

arted. Pistillate flowers: solitary, axillary to the inner scales of the
terminal buds of the abbreviated shoots, rarely in the leaf-axils; pedicels
6-10 mm. long, minutely puberulous, with 2 sub-basal prophylla; flowers
6- or 7-merous; calyx patelliform, 4 mm. across, the lobes ciliate; corolla
suberect, 8 mm. in diameter, the petals 4 mm. long; staminodes two-thirds
as long as the petals, the sterile anthers sagittate; ovary ovoid, 2 mm. long,
the style evident, puberulous, the stigma columnar, 2 mm. long, con-
spicuously lobed. Fruit subglobose-pomiform, 14 mm. in diameter, the
persistent calyx explanate, 5 mm. in diameter, orbicular in outline, ciliate,
the stigma columnar, 2-3 mm. long. Pyrenes 6 or 7, oblong in outline,
the ends obtuse, 13 mm. long, the dorsal side 4 mm. wide, very deeply
canaliculate and 3-ridged, the sides 1- or 2-striate and sulcate, the endo-
carp bony.

CHINA: Kwangtung: Wung-yuen, S. K. Lau 2795 (A); Fang-
cheng, WW”. T. Tsang 26549 (A), 26715 (A), 26814 (A); Ko-chow, Y. Tsiang
2344 (NY); Sun-yi, Y. Tsiang 2718 (A), 21166 (NY). Kwan S12 Sy
Nan-ning, Seh-feng-dar-shan, R. C. Ching 2766 (NY), 7966 (LU); W.
Po-seh, R. C. Ching 7408 (NY). Hainan: Po-ting, F. C. How 73677
(TYPE of Ilex howii, A); Ting-on, S. K. Lan 28122 (A).

1949] HU, THE GENUS ILEX IN CHINA 277

INDO-CHINA: Tourane, J. & MW. S. Clemens 3389 (A), 4025 (A);
Tonkin, Chapa, .4. Petelot 2303 (tyre of Hex megistocarpa, A), 4596 (NY)
oie (IyPh, A dsObyen, iy )> Hacot,br. F. isang 2/ 280 (A); Tien- ie
W.T. Tsang 30609 (A).

Ilex chapaensis is endemic to the subtropical or tropical forests of Hainan
and the Kwangtung-Kwangsi-Indo-China border and is a shrub or small
tree up to 10 m. high. Its white flowers appear in April, and the fruit
turns black in November.

The large fruit and dorsally 3-ridged and canaliculate pyrenes of /lex
chapaensis indicate very close relationship with //lex macrocarpa Oliv. The
latter, however, has a glabrous style, the pyrenes are laterally compressed
with the dorsal surface 1-2 mm. wide, and the pulvinate rudimentary ovary
is depressed at the middle.

8. Ilex aculeolata Nakai in Bot. Mag. (Tokyo) 44: 12. 1930.
Ilex rhamnifolia Merr. in Sunyats. 1: 201. 1934; H. H. Hu & Tang in
Bull. Fan. Mem. Inst. Biol. Bot. 9: 254. 1940. Syn. nov.
Ilex dubia (G. Don) Britton, Stern & Pogg. var. hiupehensis sensu Hand.-
Mazt. Symb. Sin. 7: 658. 1933, non Loes.

A deciduous shrub up to 2 m. high with both elongated and abbreviated
branchlets, obovate leaves, cuneate and acuminate bases, solitary inflores-
cences, 4- or 5-merous ee pea-sized fruits and 4 deeply wrinkled,
striate and sulcate stony pyre

Branchlets brunneous, os eee on the third and fourth years’
growth numerous and conspicuous; elongated shoots 3—7 cm. long, slender,
pubescent, the hairs short with thick bases, sometimes branched, the
terminal buds conic, sparsely pubescent; abbreviated shoots 3-5 mm.
long, rugose with persistent bud-scales and leaf-scars. Leaves 5-20 mm.
apart on elongate shoots, 1-3 crowded at the tips of abbreviated shoots;
stipules narrowly deltoid, persistent; petioles 10-12 mm. long, one-eighth
to one-third the length of the lamina, narrowly and shallowly impressed
and puberulous above, glabrescent or aculeolate along the side or all over;
Jamina membranaceous or thinly chartaceous, dark olivaceous above, paler
beneath, opaque on both surfaces, pubescent along the midrib and lateral
nerves on both surfaces, glabrescent or sparsely pubescent on both surfaces,
obovate, 2—5 cm. long, 1-3 cm. wide; cuneate and acuminate at the base;
acute or very shortly acuminate or rarely obtuse at the apex, the acumen
2-3 mm. long; margin coarsely and argutely serrate; midrib plane or
slightly impressed above, elevated beneath, pubescent on both surfaces, the
lateral nerves 3 or 4 pairs, evident on both surfaces, the reticulation of
the veinlets obscure. Inflorescences solitary, axillary to scales or leaves
on both elongated and abbreviated shoots; flowers 4- or 5-merous.
Staminate inflorescence: 1—3-flowered, the peduncles 0.5—2 mm. long, the
pedicels 1.5-3 mm. long, glabrous, with 2 basal, deltoid, acute, and ciliate
prophylla; calyx patelliform, 2.5 mm. across, deeply 4-lobed, the lobes
suborbicular-deltoid, ciliate; corolla rotate, 7 mm. across, the petals sub-

y

orbicular-ovate, 3 mm. in diameter, erose, rarely sparsely ciliate, one-third

278 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

connate at the base; stamens 4 or 5, two-thirds as long as the petals, the
anthers oblong, 1.5—-2 mm. long; rudimentary ovary ovoid, shortly rostel-
late and lobed. Pistillate inflorescences: uniflorous, the pedicels 3-4 mm.
long with 2 ciliate basal prophylla; calyx and corolla as in the staminate
flowers; staminodes two-thirds the length of the petals, the sterile anthers
Sagittate; ovary ovoid, 1.5 mm. in diameter, the stigma thickly discoid,
4-lobed

Fruit globose, 7 mm. in diameter, the persistent calyx explanate, ciliate,
subquadrangular in outline, the stigma discoid, 4-lobed. Pyrenes 4, elliptic
in outline, the ends pointed, 6 mm. long, 2.5 mm. wide on the dorsal sur-
face, deeply wrinkled, striate and sulcate, the striae branched and often
reticulate, the endocarp bony.

CHINA: Kiangsi: Nan-chang, H. H. Chung 26 (A), 27 (A);
J. N. Hstung 550 (A); Yung-shing-hsien, H. H. Hu 776 (A); Sung-wu,
Jy Le Gressitt 136 (A); Kien-nan, 5. K. Low 3993 (A, US), 4797 (A);
Ling-chuan, ¥. Tsiang 9852 (NY), Lee (NY); Houng-yang-shan, VY. Tsang
10563 (NY); Ping-hsiang, T. H. Wang (ex Handel-Mazzetti Pl, Sin. 167)

Hunan: Chang-ning a Cyaan OF, ti ioe (A)s- Sins
ning-hsien, C. S. Fan & Y. ¥. Li i 439 (A); Tschangscha, Handel-Maszzetti
2247 = 11344 (A); Da-o-ping, Handel-Maszzetti 2710 = 11725, (type, A);
Yun-schan, 7. H. Wang (ex Handel-Mazzetti Pl. Sin. 70) (A); Yi-chang,
WT. Tsang 23688 (A, ee Fukien: Yen-ping, H. H. Chung 3266
(A). wangtung: Lung-t’au pea lo @ Tsang CU 72797)
(Us )4- (LU 72757 ) ae Loh. -chong, C. L. Tso 21606 (Type of Ilex peas
folia, A); W.T. Tsang 20929 (A, NY); Tee iwa, WV. T. Tsang 26326 (A
Yang-shan, T. M. Tsui 489 (A, NY, US), 649 (A, NY, US). Kw vangsi ;
Lu-chen, Rk. C. Ching 5615 (NY, LU), 6082 (NY); Yung-hsien, Steward &
Cheo 736 (A), 1070 (A, NY).

Ilex aculeolata is native to the Chinese Great Lake Provinces, Hunan
and Kiangsi, and thence south to northern Kwangtung and Kwangsi. It
grows in thickets or woods at an altitude of 250-600 m. The white and
fragrant flowers appear in April. Its fruits become black in late September.

The membranous leaves, solitary fruits, deeply striate and sulcate pyrenes
of Ilex aculeolata indicate close relationship between this species and Jlex
asprella (Hook. & Arn.) Champ. ex Benth., but the latter has ovate leaves
and very long (2.5—3.5 cm.) fruiting pedicels. Some specimens of this
species collected from northern Kwangtung have been mistaken for /lex
pubescens Hook. & Arn. There is much similarity in the indumentum, leaf-
form and margin, and pyrenes of these two species, but //ex pubescens can
easily be distinguished, since it is evergreen and has fasciculate flowers and
fruits in the axil of the leaves, while //ex aculeolata has solitary flowers or
fruits mixed with leaves on abbreviated shoots.
8a. Ilex aculeolata var. kiangsiensis, var. nov.

Arbor parva; ramulis abbreviatis, 3-15 mm. longis, rugosis; foliis
obovatis, 4—9 cm. longis, 1.5-3 cm. latis, apice acutis, breviter acuminatis
vel raro obtusis, basi cuneatis et acuminatis; fructibus solitariis, axillaribus,
ellipsoideis, 15 mm. longis, 7-9 mm. diametro, striato-sulcatis; stigmate

1949] HU, THE GENUS ILEX IN CHINA 279

capitato vel discoideo; pyrenis 5, 9-10 mm. longis, 2.5-3 mm. latis, dorso
canaliculatis, lateralibus striato-sulcatis.

CHINA: Kiangsi: Kien-nan, S. K. Lau 4086 (type, A).

This variety is localized in southern Kiangsi. So far it has been collected
only from near the Kwangsi-Kwangtung border. It grows as a small tree
up to 5 m. in height.

This variety differs from typical J/ex aculeolata in having larger (15 mm.
long) fruit and leaves.

9, Tlex macropoda Miq. in Ann. Mus. Bot. Lugd.-Bat. 3: LO5. 1867;

Franch. & Sav. Enum. Pl. Jap. 1: 77. 1873; Maxim. in Mém. Acad.

Sci. St. Pétersb. VII, 29: 30, 51, pl. 1, fig. 8. 1881; Hara in Bot.
Mag. (Tokyo) 50: 187. 1936.

Ilex costata Bl. ex Mig. Cat. Mus. Bot. Lugd.-Bat. 167. 1870, nomen
nudum; Maxim. l.c. 51. 1881.

Ilex dubia (G. Don) Trel. var. macropoda (Miq.) Loes. in Nov. Act.

Acad, Caes. Leop.-Carol. Nat, Cur. 78: 487 (Monog. Aquif. 1: 487).

901. Syn. nov.

Ilex dubia (G. Don) Trel. var. hupehensis Loes. l.c. 488. 1901, in part.
nm. NOV.

Ilex dubia (G. Don) Britton, Stern & Pogg. var. pseud¢ poda Loes.

in Sarg. Pl. Wils. 1: 82. 1911; Rehd. in Jour. Arnold Arb. 8: 158.
1927: Chien in Contr. Biol. Lab. Sci. Soc. China 3(1): 58. 1927.
Ilex montana Torr. & Gray var. macropoda (Mig.) Fern. in Rhodora 41:

Syn. nov.

A deciduous tree up to 13 m. high with elongated and abbreviated
branchlets, conspicuous lenticels, chartaceous ovate, or broad-elliptic and
serrate leaves, fasciculate staminate and solitary pistillate inflorescences,
globose drupes with discoid stigma, and 5 striate and sulcate bony pyrenes.

Branchlets cinereous or brunneous, glabrous; elongated ones 4-10 cm.
long, 2.5-3 mm. in diameter, brunneous, longitudinally plicate, the lenti-
cels conspicuous, elliptic, the terminal buds ovoid or broadly conic,
glabrous, the inner scales ciliate; abbreviated branchlets 3-50 mm. long
(tenth year’s growth about 3 cm. long), glabrous, rugose with persistent
bud-scales, leaf-scars and scars of the pedicels, the lenticels lacking, the
terminal buds subglobose-ovoid, glabrous, with ciliate inner scales. Leaves
on elongated shoots 10-25 mm. apart, 3-5 crowded at the tips of the
abbreviated shoots; stipules minute, acute, deltoid, often obscure; petioles
11-20 mm. long, one-fifth to one-half the length of the lamina, those on
the elongated shoots shorter, the distal fourth winged by the decurrent
leaf-base, the rest narrowly and deeply canaliculate, hirsute in the groove;
lamina chartaceous or membranaceous, brunneous or olivaceous-brunneous,
opaque on both surfaces, sparsely hirsute especially on the veins above,
glabrous, rarely puberulous beneath, ovate or broad-elliptic, 4-8 cm. long,
2.5-4.7 cm. wide, rounded at the base, but tapering sharply at the petiole,
becoming decurrent, shortly deltoid-acuminate, rarely acute, at the apex,
the acumen 4-7 mm. long; margin argutely serrate; midrib narrowly
impressed and hirsute above, elevated and glabrous beneath, the lateral

280 JOURNAL OF THE ARNOLD ARBORETUM [vou. Xxx

nerves 6-8 pairs, hirsute, often slightly impressed above, prominent
beneath, the reticulation of the veinlets obscure above, evident beneath.
Staminate inflorescence: fasciculate, the fascicles 2—5-flowered, the indi-
vidual branches uniflorous, the pedicels 4-7 mm. long, elabrous: flowers
S-merous; calyx patelliform, 2.5 mm. across, glabrous, deeply 5-lobed, the
lobes ovate-deltoid, 1 mm. long and wide, acute or obtuse, often erose,
ciliate; corolla rotate and reflexed, 5 mm. across, the petals ovate-oblong,
2 mm. long, 1.5 mm. wide, aanusely ciliate, one- fifth connate at the base;
stamens shorter than the petals, the gains oblong, | mm. long; rudi-
mentary ovary compressed, pulvinate, depressed and inconspicuously lobed
at the center. Pistillate inflorescence: solitary, axillary to the scales or
rarely to the leaves of the abbreviated shoots, not uncommonly axillary to
the lower leaves of the elongated shoots; pedicels 6-7 mm. long, glabrous;
flowers 5- or 6-merous; calyx and corolla as in the staminate flowers:
staminodes three-fourths the length of the petals, the sterile anthers
cordate; ovary ovoid, 1.75 mm. long, 1.5 mm. wide, the stigma thick-
discoid. Fruit globose, 5 mm. in diameter, the persistent calyx explanate,
2.5 mm. across, stellate in outline, the lobes 5 or 6, acute, deltoid, ciliate,
the stigma discoid, convex, 5- or 6-lobed. Pyrenes 5, oblong in outline,
the ends obtuse, 4—4.5 mm. long, 2 mm. wide, longitudinally and reticu-
lately striate and sulcate, the endocarp bony.

CHINA: Chekiang: Tien-tai-shan, R. C. Ching 1403 (A); Tien-
mu-shan, T. N. Lint 257 (NY); Chekiang Univ. s.n. (staminate flower, LU;
fruit, LU). Anhwei: Wang-shan, HW’. C, Cheng Gi (Wek. C, Ching
3020 (A), 3048 (A, LU); K. Ling (ex NU Herb. no. 7772) (A): Wu-
Yuan, R. C. Ching 3241 (A); Chiu-hwa-shan, S. C. Sun 1445 (A, NY).
Kiangsi: Lu-shan, H. H. Hu 2354 (LU); VY. Tsiang 10756 (NY).
Hupei (Hupeh): dA. Henry 6107 (isoryre of Mex dubia var. hupehensis,
A, G, NY); western Hupeh, FE. H. Wilson 2695A (K); cee shan-hsien,
i, H. Wilson 3090 (tyre of Ilex dubia var. pseudomacropoda, A

JAPAN: eee (IsoTyYPE, A); J. G. Jack in 1905 (Aug. ea Bo Ges
Sept. 6, A; Oct. 26, A, G); Maximowitcs in 1862 (NY), in 1863 cee NY):
ex Herb. K. Miyabe Cas K he in 1904 (A); K. Shiota 67 (A), 3456
(A), 7207 (A); Siebold ae 2 a: 5: Cade in 1892 (Sept. 2, 24, A;
Oct. 24, G, A); Tschonoski in 1864 i Y); J. H. Veitch (A); K. Uno
21884 (A, NY); K. Watanabe in 1888 (G); E. H. Wilson in 1914 (A).

KOREA: U. Faurie 1634 (A), 1635 (B), 5522 (A); Taquet 629 (A),
1348 (A), 2716 (A), 3183 (A): E. H. Wilson 9503 (A

CULTIVATED: TI have seen two sterile specimens eneuned in the
Arnold Arboretum.

flex macropoda was first described from Japan. The species, however,
is very wide-spread and a fairly common tree in the woods or forests of
the Lower Yangtze Provinces. The flowers appear in May, and the fruit
reddens in September

flex macropoda is closely related to an American deciduous species,
Ilex montana Torr. & Gray, which differs in having very short-pedicellate
fruits (shorter than the diameter of the mature fruit) and a styliferous
ovary.

1949 | HU, THE GENUS ILEX IN CHINA 281

Upon the basis of a specimen collected by Buerger, Miquel in 1867 de-
scribed [lex macropoda as a long-petiolate species with broad-ovate glabrous
leaves. Loesener in 1901 treated it as a variety of the American species,
Ilex dubia (G. Don) B.S.P. But this type of Don’s species has solitary,
subumbelliform inflorescences. It is not the same as /lex macropoda Miq.
and, in fact, is not closely related to it.

Prof. Fernald (Rhodora 41: 428. 1939) treated /lex macropoda Miq.
as a variety of another American species, //ex montana Torr. & Gray. With
this also I cannot agree. The Asiatic species is closely related to Ilex
montana, but geographically and morphologically, since it has distinct
specific characters, it should be retained as a species. The following com-
parison shows these differences.

Ilex montana has (1) an ovary which is styliferous, the styles measuring
1-2 mm. in length; (2) fruiting pedicels 2-4 mm. long and shorter than
the diameter of the mature fruit; (3) rudimentary ovary in the staminate
flowers pulvinate with the center rostellate; and (4) pyrenes 6 mm. long,
deeply striate and sulcate. On the other hand, //ex macropoda has (1) the
stigma sessile and discoid; (2) the fruiting pedicels 6-7 mm. long and
longer than the diameter of the mature fruit; (3) the rudimentary ovary
in the staminate flowers also pulvinate but with a depressed center; and
(4) the pyrenes- smaller (ca. 4.5 mm. long) and less deeply striate and
sulcate.

Loesener in 1913 published /lex dubia var. pseudomacropoda based on
a staminate flower, Wilson 3090.. He distinguished this variety from
typical lex macropoda as follows: “Ramulis abbreviatis crassis usque
4.5 cm. longis, foliis subtus glabris.” These two characters, the length of
the abbreviated shoot and the indumentum of the leaf surface, vary much
in this group of Zlex; indeed, the characters suggested by Loesener for his
variety are found even on Buerger’s specimen, the collection on which
Miquel’s species was based.

10. [lex tsoii Merr. & Chun in Sunyats. 1: 66. 1930; Hu & Chun, Ic.
P). Sin. 4: p7. 78. 1935.

A deciduous shrub or small tree up to 4 m. high with elongated and
abbreviated branchlets, chartaceous, pubescent, ovate, or ovate-elliptic
serrate leaves, very shortly pedicellate fruits, and 5 reticulately striate-
sulcate bony pyrenes.

Branchlets castaneous or atro-griseous, the annual growth of the elon-
gated shoots 2.5-6 cm. long, 2-3 mm. in diameter, castaneous, or ochra-
ceous, the lenticels conspicuous, elliptic, the terminal buds globose ovoid,

labrous, with ovate, acute, or cuspidate, densely long-ciliate scales;
abbreviated shoots up to 3.5 cm. long, rugose with persistent bud-scales
and scars, glabrous. Leaves 15-20 mm. apart on elongate shoots, 1-3
crowded at the ends of the abbreviated shoots; stipules broadly and shortly
deltoid, callose, persistent; petioles 6-10 mm. long, one-sixteenth to one-
ninth the length of the lamina, and narrowly winged by its decurrent base,
glabrous, canaliculate above; lamina chartaceous, olivaceous, opaque,

282 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

prominently marked with small areoles, glabrescent or pubescent, especially
along the veins of both surfaces, ovate or ovate-elliptic, 5-10 cm. long,
3-5 cm. wide, rounded at the base, acute at the apex, the acumen 7-15 mm.
long, serrate and cuspidate; margin finely and argutely serrate; midrib
deeply and narrowly impressed and puberulent above, elevated and hirsute,
glabrescent in age beneath, the lateral nerves 9 or 10 pairs, slightly
elevated, prominent and pubescent on both surfaces, the veinlets densely
reticulate, forming distinct minute areoles on both surfaces. Staminate
inflorescences: fasciculate, the fascicles 1—3-flowered, the individual
branches uniflorous; pedicels 3-4 mm. long, glabrous; flowers 6-merous;
calyx patelliform, 4 mm. across, deeply 6-lobed, the lobes unequal in
size, deltoid or ovate, 1 mm. long, 1 mm. wide, acute, rarely rounded,
ciliate; corolla rotate, 6-7 mm. across, the petals oblong, 2 mm. long,
very minutely ciliate, one-fourth connate at the base; stamens shorter
than the petals, the anthers oblong; rudimentary ovary pulvinate, the
center plane, inconspicuously lobed. Pistillate inflorescences: solitary,
axillary to scales or rarely to leaves on the abbreviated or elongated shoots,
the scales ovate, deltoid, acute or cuspidate, ciliate; pedicels 1-3 mm. long,
glabrous; flowers 5—7-merous; the calyx and corolla as in the staminate
flowers; staminode very small, one-fifth the length of the petals, the
sterile anthers cordate; ovary ovoid, 2 mm. in diameter, the stigma thickly
discoid, convex. Fruit globose, 6-8 mm. in diameter, the persistent calyx
explanate, stellate in outline, the stigma thickly discoid or capitate, promi-
nent. Pyrenes 6, oblong in outline, the ends obtuse, 5 mm. long, 2.5 mm.
wide on the back, deeply ridged and sulcate, the striae reticulate, the
endocarp bony.

CHINA: Chekiang: Tih-tai-shan, R. C. Ching 1522 (A, US):
between Ping-yung and Tai-suan, R. C. Ching 2113 (A, LU, US), Kwei-
chow: Fan-ching-shan, Steward, Chiao & Cheo 405 eae ee OS
Kwangtung: Huang-tung, S. S. Sun 9158 (NY); Jen-hwa, WV. T.
Tsang 26266 (A), 26386 (A), 26391 (material for the description of pistil-
late flower, A), 26396 (A); Ho-yuen, W. T. Tsang 28850 (A); Lok-chong,
C. L. Tso 20778 (type, NY; photo, A). Kwangsi: Shuen-yuen, S. C.
Chung (T. S. Tsoong) 81544 (A); Chuen-yuen S. C. Chung (T. S. Tsoong)
82014 (A); Tsu-yuen, Z. S. Chung (Tsoong) 83507 CS.

Ilex tsoti was first discovered at Lok-chang on the northern Kwangtung
border. Additional material indicates that it is a common species in warm-
temperate South and East China, where it grows as a large shrub in
thickets or as a tree in the woods. The flowers appear in May, and by
July the fruits have become dark purple.

In the size of the leaves, the length of the pedicels, and the sculpture
of the pyrenes //ex tsoii is almost identical with Jlex montana Torr. & Gray,
a native of North America. Besides occurring on different continents, there
are certain technical characters that serve to distinguish these two species.
Ilex montana has broad-elliptic or obovate leaves, acute or acuminate at
he base, 2—5-flowered individual branches of the staminate fascicles,
rostellate rudimentary ovaries, styliferous ovaries and fruits, and larger

1949] HU, THE GENUS ILEX IN CHINA 283

(6 mm. long, 3 mm. wide) pyrenes. Furthermore, the characteristic
minute but distinct areoles on the leaf-surfaces of Jlex tsoii are lacking
in the leaves of //ex montana.

Ilex tsow differs from Ilex macropoda Miq. in having very short fruiting
pedicels (1-2 mm. long). From J/lex fragilis Hook, f. it differs in having
only 5 or rarely 6 pyrenes which are not only larger, but are deeply and
reticulately striate and sulcate.

B. Supscenus EUILEX Logsener

Hex subgenus Euilex Loes. in Verh. Bot. Ver. Brand. 33: 25, 26. 1891, in
Engl. & Prantl, Nat. Pflanzenfam. Nachtr. 218. 1897, et in Nov. Act.
Acad. Caes. Leop.-Carol. Nat. Cur. 78: 94 (Monog. Aquifol. 1:
94). 1901.

Ilex subgenus Aquifolium (Gray) Rehder, Man. Cult. Trees Shrubs 541.
ibli 49.

Ilex §1. Aguifoliam Gray, Man. Bot. N. U. S. 276. 1848.
Ilex sect. Ilex Maxim. in Mém. Acad. Sci. St. Pétersb. VII. 29(3): 26.
1881.

Trees or shrubs with branchlets all elongated, the lenticels usually
lacking on the current, second or even the third year’s growth; leaves
evergreen, thick-coriaceous, coriaceous or rarely chartaceous; inflorescences
solitary or fasciculate.

Key To THE SECTIONS

A. Pistillate and staminate inflorescences both solitary, axillary, on the
urrent year’s growth; pyrenes unicanaliculate or 3-striate and 2-
ee on the dorsal surfaces ; Sere coriaceous or sublignescent.
ee es Oe |e eT Sects ie pens.
. Pistillate flowers solitary, axillary on the current year’s growth (ex-
cept in some /lex triflora Blume), the staminate flowers fasciculate
on the second year’s growth or rarely solitary on the current year’s
growth; pyrenes smooth, or striate and esulcate with insculpt striae, or
slightly roughened; endocarp coriaceous........ Sect. VI. Paltoria.
AAA. Pistillate and staminate inflorescences both fasciculate, axillary, on
the second year’s or even older growth; pyrenes rugose and pitted or
striate with elevated striae; endocarp coriaceous, woody or stony.
B. Pyrenes 4, rarely less; the endocarp woody or stony; individual
branches of the pictitiate ee ALTE ORO US aaupne prassah Maen eer She
ee eRe Cn aT ie area aoe Sectiavatle Aquifolium.
BB. Pyrenes 6 or 7, rarely less or more, the endocarp coriaceous
sublignescent ; ined branches of the pistillate fascicles cymose,
umbelliform, or uniflorous
G. Tneividaal branches ee the staminate and the eae fascicles
trichotomous compound cymes or subumbels, each with more
than 10 flowers; leaves 10-20 cm. long, we cm. wi
re re ae ee ne eee Sect. VIII. Lauroilex.
CC. Individual branches of the fascicles 1-, 3-, or sometimes 5-
flowered; leaves rarely up to 10 cm. long, 3.5 cm. wide....
Se ee ont eed ne a ee Sect. IX. Pseudoaqutfolium.

ax

>

284 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXx

secTION V. LIOPRINUS (Loes.), stat. Nov.

Ilex subgen. Fuilex Loes. ser. Lioprinus Loes. in Verhandl. Bot. Ver.
Prov. Brandenb. 33: 26, 27. 1891, in Engler & Prantl, Nat. Pflanzen-
fam. Nachtr. 218. 1897, et in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 95 (Monog. Aquif. 1: 95). 1901.

Evergreen trees or shrubs with solitary inflorescences in the axils of the
leaves or scales on the current year’s growth.

~---{ |OPRINUS Z
a= RF ALTORIA |
— AQUIFOLIUM ~

-- {AUROILEX Z
o> PSEUDOAQUIFOLIUM

Fig. 5. Geographic distribution of the Eastern Asiatic representatives of
the five sections of the subgenus Euilex.

Key To THE SERIES

A. Pyrenes unicanaliculate on the dorsal surfaces; inflorescences cymose;
leaves usually crenate, serrate, rarely entire..... Series. J; Chinenses.
AA. Pyrenes 3-striate and 2-sulcate on the dorsal surfaces; inflorescences
usually umbelliform, rarely cymose; leaves entire.................
series 2. Umbelliformes.

i

1949 ] HU, THE GENUS ILEX IN CHINA 285

SERIES 1. CHINENSES, ser. nov.

Arbor vel frutex; foliis chartaceis vel coriaceis, brunneis vel atrobrunneis;
stipulis nullis vel minutis; inflorescentiis cymosis vel subumbelliformibus,
solitariis, axillaribus; floribus 4—6-meris; fructibus ellipsoideis raro globo-
sis; pyrenis 4—6, laevibus, dorso longitudinaliter profunde unicanaliculatis,
raro impressis; endocarpio coriaceo vel sublignescente.

KEY TO THE SPECIES

A. Leaves entire, thickly coriaceous (except in I. lonicerifolia of Taiwan).

B. Infructescence a compound cyme, the secondary axis well devel-

oped; leaves large, 13-14 cm. long and 4-8 cm. wide. (Kwang-

BUDE Ce Gi) teenie tela cae 6 g's a 8 Guay. a I, macluret.

BB. Infructescence a simple cyme, the secondary axis mot developed ;
leaves less than 11-12 cm. long.

C. Fruit solitary, 10-12 mm. in diameter; peduncles and pedicels
together less than 4 mm. long, shorter than the diameter of
biel titres CEVA Ia) 5s y's <a k em eee 2. I. lancilimba.

CC. Fruit in simple loose cymes or subcapitate umbels, less than
10 mm. in diameter; peduncles longer than the diameter of the
Pouib

D. Infructescence a subcapitate umbel; fruiting pedicels 0.5—
2 mm. long, much shorter than the peduncles. (Taiwan)..
ee ee oli e! to's dex de . I. tugttakayamensis.
DD. Infructescence a loose umbel or cyme; fruiting pedicels
equaling or longer than the peduncles
E. Plants densely pubescent; petioles 4-10 mm. long, one-
seventeenth to one-eighth the length of the lamina.
. Leaves en less than 6 cm. long, 2.5 cm.
wide. (S.C na) sl decane gt 14. I. dasyphylla.
FF, Leaves eee coriaceous or chartaceous, over 8
Ci Ong oncniy wide: (Taiwan) ois. Se oe ee
SP eae are ae aap caida 15. J. lonicerifolia.
EE. Plants glabrous; petioles 10-25 mm. long, one-fifth to
one-fourth the length of the lamina.
F, Leaves ovate; cymes loose, the peduncles 2- - ane
long, the pedicels 1-1.5 cm. long.
ngtu ne Pe oi a Moyes ce
FF. Leaves clipe cymes compact, the peduncles less
m. long, the pedicels even shorter. (East,
eae and West China) tee l7, fe editicostata:
AA. Leaves crenate or serrate (except in J. manneiensis), subcoriaceous,

s ovate, less than 5 cm. et (Yunnan). .18. ve ferruginea.
BB. Leaves ee elliptic, 8-15 cm. long
s coriaceous ; inflorescences “subumbelliform, the ae
1333 mm. long, ceually longer than the petioles. (Centr
Gna) ice asada amen ene nee ity ioe hE Reade a ee
. Leaves subcoriaceous or chartaceous; inflorescences cymose,
the peduncles less than 15 mm. long, usually shorter than the
petioles.

C

QO?

286 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

D. Leaves entire, the midrib ferruginously and densely pubes-
cent beneath; pedicels very short (1-2 mm. long), shorter
than the peduncles. (Yunnan)....... 20. J. mannetensis.

DD. Leaves crenulate, the midrib glabrous or puberulent; pedi-
cels 4-5 mm. long, usually longer than peduncles
i, Buarichiens glabrous; leaves rarely 10 cm. lone: gla-
brous; fruits ellipsoid; pyrenes narrowly canaliculate.
East, Central, South, and West China, Japan, much
RUlivaled | ccudtieesylcicecaets ease PA ~ chinensis.
: Baciee ashi leaves over 12 cm. long, usu-
ally puberulous; fruits Glices ee pyrenes broadly

U-shaped or triangular in cross-section

F. oe compound, the ee axis well devel-
oped; peduncles 5-12 mm. long, longer than the
Se ia pyrenes broadly U-shaped. (South

WE) ca vio otek dats 22. I. kwangtungensts.
see simple, secondary axis lacking; peduncles
2-3 mm. long, shorter than the pedicels; pyrenes
triangular in cross-section. (Yunnan)........

2

es
es

F

a

aditicostata

axcelsa
ferruginca
odajam

CHOS ARIST AAUAUMMA2DBH>-:

yid ed tied tbe td ag

tugita a ametele
_ umbellulata

Fic. 6. Geographic distribution of the species of the section Lioprinus in
China

11. [lex maclurei Merr. in Lingnan Sci. Jour. 13: 35. 1934.

A small tree or shrub about 3-4 m. high with large coriaceous entire
glossy oblong leaves, several times di- or trichotomous infructescences,
globose red drupes, and 4 pyrenes unicanaliculate on the back

Branchlets stout, glabrous; second year’s growth 5 mm. long, striate
when dry, castaneous, pe lenticels sparse, minute, Bote THEO RS pIe NOUS,
the leaf-scars obovate or deltoid, very slightly elev ated; current year’s

growth angular, 4-5 mm. thick, striate, sulcate, castaneous. Leaves
occurring also on second year’s growth, 4-20 mm. apart; stipules lacking;

1949] HU, THE GENUS ILEX IN CHINA 287

petioles 15-20 mm. long, one-ninth to one-sixth the length of the lamina,
robust, subcylindric, canaliculate above, plicate-rugose beneath, glabrous;
lamina thick-coriaceous, olivaceous-brunneous, shiny above, opaque
beneath, oblong or oblong-elliptic, 9-19 cm. long, 3.5-8.5 cm. wide; base
obtuse or rounded; apex very obtuse, rounded, or rarely shortly and broadly
acuminate; margin entire, slightly recurved, midrib plane above, elevated
beneath, the lateral nerves in 10-12 pairs, obscure above, evident beneath,
the reticulation of the veinlets obscure on both surfaces. Fowers not seen.
Infructescence twice or thrice dichotomous, with 7-9 fruits; peduncles
glabrescent or glabrous, 8-17 mm. long, compressed, enlarged at the distal
end, the secondary axes 2-8 mm. long, the bracts deltoid to lanceolate,
entire, glabrous, the pedicels 3-6 mm. long with 2 basal prophylla; fruiting
calyx 4 mm. across, 4- or 5-lobed, the lobes rounded, very shortly ciliate.
Fruit globose, 6-7 mm. in diameter, when dry black, shiny, plicate, the
stigma mammiform. Pyrenes 4, oblong in outline, U-shape in cross-section,
4.5 mm. long, 3 mm. wide at the back, deeply unicanaliculate on the back,
the endocarp thickly coriaceous.

CHINA: Kwangtung: Lim-kong, F. A. McClure 643 ((=LU
Lobe \eeCDVEE Ne tsOTvre,.U > photo, A).

INDO-CHINA: Tonkin: W. T. Tsang 27164 (A), 30335 (A).

In its oblong leaves and 2 or 3 times di- or trichotomous infructescences
Ilex macluret resembles Jlex thorelit Pierre, but the latter has thinner
leaves and 9-12-celled ovaries. The Tsang Indo-Chinese collections come
from immediately south of the Chinese border.

12. Ilex lancilimba Merr. in Lingnan Sci. Jour. 7: 312. 1929; Tanaka
& Odashima in Jour. Soc. Trop. Agr. 10: 372. 1938; Masamune, FI.
Kainant. [Hainan] 174. 1943

An evergreen tree 10 m. high with yellowish pubescent branchlets and
inflorescences, lanceolate entire leaves, short oo pedicels, large, usually
solitary fruits, and 4 unicanaliculate pyre

unk 20 cm. in diameter; branchlets cer the third year’s growth
ere cath 6 mm. in diameter. plicate and rugose, the large semi-
orbicular leaf-scars united with the scars of the peduncles immediately
above them, the lenticels inconspicuous, cinereous-rufous; current year’s
growth nigrescent, plicate, angular, mm. thick, minutely yellowish

pubescent. Leaves crowded on the current year’s growth, 3-10 mm.

apart; stipules lacking; petioles compressed, 15-25 mm. long, 5 mm. wide,

one-sixth the length of the lamina, sparsely puberulent, the upper half
slightly winged by the decurrent leaf-base; lamina glabrous, thick-
coriaceous, cinnamomeous or fumous, slightly shiny above, opaque beneath,
lanceolate, 8-14 cm. long, 2-4 cm. wide; base cuneate or obtuse; apex
acute or shortly acuminate, the acumen 5 mm. long; margin entire; midrib
plane, slightly elevated and puberulent above, elevated and when young
sparsely hirtulose beneath, the lateral nerves in 10-14 pairs, rather obscure
on both surfaces, the reticulation of the veinlets usually obscure. Inflores-
cences cymose, tomentose, solitary, axillary to the basal scales or to the

288 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

leaves of the current year’s growth; flowers 4-merous. Staminate inflores-
cences: cymes 3 times di- or trichotomous, the peduncles 5-14 mm. long, the
secondary axis often well developed, the pedicels 1.5—2 mm. long; calyx
patelliform, 3 mm. across, 4-lobed, the lobes broadly deltoid, 1 mm. long,
2 mm. wide at the base, the apex rounded or obtuse; corolla pink-purple,
rotate, 7 mm. across, the petals ovate-oblong, 2.5-3 mm. long, one-third
connate at the base; stamens shorter than the petals, the anthers oblong,
1.5 mm. long; rudimentary ovary minute, conic. Pistillate inflorescences:
cymes 3-flowered, the peduncles 2 mm. long, the pedicels 1-2 mm. long;
flowers greenish white, 4- or 5-merous; calyx and corolla as in the staminate
flowers; staminodes one-half the length of the petals, the sterile anthers
cordate; ovary large, ovoid, 2 mm. in diameter, the stigma thickly discoid,
4-lobed. Fruit globose, usually solitary, very large for the genus, 10-12
mm. in diameter; fruiting pedicels 4-6 mm. long, 2—2.5 mm. thick,
yellowish pubescent, the persistent calyx explanate, quadrangular in out-
line; the stigma discoid, 4-lobed. Pyrenes 4, smooth, oblong in outline,
dorsally broadly canaliculate, in cross-section widely U-shaped; the endo-
carp sublignescent.

CHINA: Kwangtung: Ying-tak, McClure 2680 ( = CCC 14541)
(A, 1sotyPE, photo) ; Sin-fung, Y. WV. Taam 541 (A). Hainan: Fan-ya,
N. K. Chun & C. L. Tso 44096 (A, NY, US), 44125 (A, NY); Mo-san-
leng, N. K. Chun 44297 (A, NY), 44301 (A, NY, US); Yai-chow, F. C.
How 70136 (A, NY); Po-ting, F. C. Led 73681 coe Sak-sa, S. K,
25947 (A), 26582 (A), 26588 (A); Lok-tung, S. K. Lau 27303 (A);
Tang, 455 (A); C. Wang 35115 (A, N Y).

Ilex lancilimba is a forest tree in the tropical regions of Kwangtung and
Hainan, where it grows up to an altitude of 1130 m. Its pink-purple
staminate and greenish-white pistillate flowers appear in March, and by
November the red fruits are mature.

In its large coriaceous entire leaves this species resembles the Taiwan
Ilex tugitakayvamensis Sasaki, but the latter has subcapitate-umbelliform
infructescences, smaller fruits, and smooth pyrenes slightly impressed on
the dorsal surfaces.

13. Ilex tugitakayamensis Sasaki in Trans. Nat. Hist. Soc. Form. 21:
153, fig. 3. 1931; Kanehira, Form. Trees 383. 1936.

An evergreen tree with robust branchlets, large coriaceous elliptic entire
leaves, unbelliform infructescences, very shortly pedicellate fruits, and
5 or 6 pyrenes with the dorsal surfaces flat or slightly concave.

Branchlets glabrous, when dry brunneous; rugose with elevated leaf-
scars, annual growth very slight, only 4-5 mm. in the one specimen seen;
third year’s growth 5.5 mm. in diameter, the lenticels lacking, the leaf-
scars deltoid; second and current year’s growth 4—5 mm. in diameter,
angular and plicate-rugose. Leaves rather crowded; stipules obsolete ;

mm

shallowly and broadly canaliculate above, glabrous; lamina thickly coria-
ceous, ochraceous-brunneous, opaque on both surfaces, glabrous, elliptic or

1949 ] HU, THE GENUS ILEX IN CHINA 289

oblong-elliptic, 10-14 cm. long, 3-5 cm. wide; base acute; apex shortly
acuminate, the acumen 5-10 mm. long, the very tip obtuse; midrib
elevated on both surfaces, glabrous, the lateral nerves in 12-14 pairs,
evident on both surfaces, branching and anastomosing near the margin,
the reticulation of the veinlets rather obscure on both surfaces. Infructes-
cences subumbelliform, puberulent, solitary, axillary, on current year’s
growth only; peduncles 8-9 mm. long, flattened, slightly enlarged at the
distal end, umbels 5—7-flowered; bracts linear, 2-3 mm. long, puberulous;
pedicels very short, 0.5-1 mm. long; flowers 5- or 6-merous; persistent
calyx subexplanate, densely pubescent, 4 mm. in diameter, shallowly 5-
lobed, the lobes deltoid, obtuse, pubescent and ciliate. Fruits ellipsoid-
globose 5-7 mm. long, 5 mm. across, the stigma navel-like, inconspicuously
5- or 6-lobed. Pyrenes 4—6, usually elliptic in outline, trigonous in cross-
section, 4 mm. long, 2 mm. wide on the back, the back flattened or slightly
concave, the endorcarp smooth, coriaceous.

CHINA: Taiwan: Mt. Tugitaka, Y. Simada 2538 (TU, Type).

Ilex tugitakayamensis Sasaki is known from a single collection. Through
the courtesy of Dr. H. L. Li, Curator of the Herbarium, National Taiwan
University, I have been able to study the type specimen.

Its large elliptic entire leaves, the flattened peduncles, and the short
pedicels relate it to Ilex lancilimba Merr., but that species differs i nears
(very) short-pedicellate and very ies solitary fruits 10-12 in
diameter and broadly canaliculate pyrenes. Though apparently rare, ie
tugitakayamensis seems to be a valid species.

14. Ilex dasyphylla Merr. in Lingnan Sci. Jour. 7: 311. 1931; Chun in
Sunyats. 2: 68. 1934.

Ilex flaveo-mollissima Metcalf in Lingnan Sci. Jour. 11: 14. 1932.

An evergreen tree up to 9 m. high, with ferrugineous-velutinous branch-
lets, leaves, and inflorescences, entire leaves, simple loose axillary cymes
and deeply unicanaliculate pyrenes.

Branchlets zigzag, subterete; fourth year’s growth 3.5-4 mm. in
diameter, cinereous, pubescent, the lenticels numerous, inconspicuous, the
leaf-scars small, triangular or semi-orbicular, hardly elevated; third and
second years’ growth 2.5-3 mm. in diameter, densely villose, the lenticels
lacking; current year’s growth 1.5-2 mm. thick, thickly ferrugineous-
velutinous. Leaves occurring also on second year’s growth, 2-8 mm. apart;
stipules subulate, covered by the pubescence; petioles short, 4-5 mm.
long, one-seventeenth to one-eighth the length of the lamina, very densely
velutinous; lamina coriaceous, velutinous on both surfaces, brunneous,
ovate, ovate-elliptic or ovate-lancéolate, 3-9 cm. long, 1.3-2.4 cm. wide;
base obtuse, rarely rounded; apex acuminate, rarely acute, the acumen
3-12 mm. long; midrib impressed above, elevated beneath, the lateral
nerves obscure. Inflorescences cymose, solitary, axillary on current year’s
growth, densely ferrugineous-villose; flowers 4- or 5-merous. Staminate
inflorescences: cymes 3—5-flowered pseudumbelliform; peduncles slender,
weak, 4-5 mm. long; bracts to individual flowers deltoid, buried in and

290 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

covered with hairs; pedicels 2 mm. long with minute basal prophylla;
calyx patelliform, densely velutinous and ciliate, 3 mm. across, the lobes
deltoid or rounded; corolla rotate, the petals ovate-oblong, 3 mm. long,
2.5 mm. wide, reflexed at anthesis, one-fifth connate at the base; stamens
equaling the petals in length; the anthers oblong, 1.5 mm. long; rudi-
mentary ovary pyramidal, truncate at the apex. Pistillate inflorescences:
cymes 1—3-flowered; peduncles 5—8 mm. long, the bracts linear-lanceolate,
1.5 mm. long; pedicels 3-8 mm. long, with minute basal prophylla; calyx
and corolla as in the staminate flowers; staminodes one-half the length of
the petals, the sterile anthers sagittate; ovary conic, 2 mm. wide at the
base, the stigma mammiform; style rather prominent, 4-angled. Fruit
globose, 5-6 mm. in diameter, red, when dry brownish red, smooth, shiny;
the persistent calyx explanate, 4-5 mm. in diameter, the stigma thickly
discoid. Pyrenes 4 or 5, short and plump, 4 mm. long, 2.5 mm. wide and
unicanaliculate on the back, the endocarp coriaceous.

Cie : Fukien: Dunn ex Hongkong Herb. no. 2464 (isotype of
lex flaveo-mollissima, A). Kwangtung: Ying-tak, (CCC) 14775 (iso-
TYPE, photo, A); Wung-yuen, S. K. Lau 2796 (A); Sin-fung, W. Y. Taam,
703 (A), 839 (A); Ta-pu, W. T. Tsang 21625 (A, NY). Kwangsi:
Wai-tsap, W. T. Tsang 22725 (A).

This is a shrub or small tree up to 9 m. high. It flowers in May. The
staminate flowers are red and odorless (ex Taam). The fruit turns red
in September.

The species is a very distinct one. There is no other Chinese //ex that
has such a dense ferrugineous indument. Its unicanaliculate pyrenes and
loose infructescences simulate those of /lex chinensis Sims, which differs
in having crenate-serrate glabrous leaves.

15. Ilex lonicerifolia Hayata, Icon. Pl. Form. 3: 54, pl. 8, 1913;
Kanehira, Form. Trees 376, fig. 332. 1936.

A small evergreen tree with ferrugineous pubescence, subcoriaceous
oblong-elliptic or rarely ovate-elliptic entire leaves, solitary subumbelliform
axillary inflorescences, small ovoid-subglobose fruits and 5 or 6 elliptic
unicanaliculate pyrenes, trigonous in cross-section.

Branchlets subterete, second year’s growth 3 mm. in diameter, densely
pubescent, sparsely rimulose, the lenticels very few, the leaf-scars semi-
orbicular-deltoid; current year’s growth 2 mm. in diameter, when dry
castaneous, longitudinally sulcate, densely pubescent, the terminal buds
conic, densely ferrugineously pubescent. Leaves occurring also on the
second year’s growth, 10-15 mm. apart; stipules minute, deltoid, often
obsolete; petioles 4-8 mm. long, one-fifteenth to one-twelfth the length
of the lamina, shallowly canaliculate above, pubescent or sparsely hirsute;
lamina subcoriaceous, castaneous above, brunneous beneath, opaque on
both surfaces, oblong, oblong-elliptic, rarely ovate-elliptic, (4.5—) 8-11
(-12) cm. long, 2—4.5 cm. wide; base obtuse, rarely rounded; apex short-
acuminate, rarely acute, sometimes rounded, the acumen 3—7 mm. long,
the very tip acute or mucronate; margin entire, recurved; midrib plane or

1949] HU, THE GENUS ILEX IN CHINA 291

slightly impressed and hirsute above, elevated beneath, the lateral nerves in
10-12 pairs, prominent on both surfaces, anastomosing near the margin,
the reticulation of the veinlets evident above, obscure beneath. Inflores-
cences subumbelliform, axillary, found only on the current year’s growth,
often behind an axillary bud, hirsute, 3-11-flowered. Pistillate inflores-

ss)
c
(or:
ig]
e2)
Q
ie")
5
ct
a
o)
=
=]
oO
_
wn
B
5B
io”
SS
io)
4
ct
in
oO
<
=
nm
©
—
ct
ion
oD
Ov
od
a
i)
(@)
—
wn
i
ON
3
=
o
=)
oe

with 0-2 deltoid callose basal prophylla; flowers 5- or 6-merous; calyx
patelliform, 3.5 mm. across, pubescent, the lobes deltoid, obtuse, ciliate;
corolla rotate, 7 mm. across, the petals ovate-oblong, 2.5-3 mm. long,
eciliate, one-sixth connate at the base; staminodes three-fourths the length
of the petals, the sterile anthers sagittate; ovary ovoid, 2 mm. long,
1.5 mm. wide at the base, glabrous; the stigma thickly discoid or mammi-
form, convex, 5- or 6-lobed. Staminate flowers unknown. Fruit ovoid-
globose, 5—7 mm. in diameter, the persistent calyx subexplanate-patelliform,
4 mm. across, the stigma mammiform. Pyrenes 5 or 6, elliptic in outline,
trigonous in cross-section, concave on the dorsal surface, 4—4.5 mm. long,
back 2 mm. wide, the endocarp smooth, coriaceous.

FORMOSA: Lake Candidus, J. L. Gressitt 230 (A); E. H. Wilson 9977
(A); Taichu, Lake Jilugetutan, Y. Kudo in 1929 (A); Nokozan, E. Matuda
in 1919 (TU); Sekiin, Kudo & Sasaki 15245 (TU); Taihokusyu, T. Suzuki
17025 (TU); without precise locality, Kudo & Yamamoto in 1930 (AWoy is

T. Suzuki’s specimen has much smaller leaves than the other specimens
examined. Its fruits are solitary. Judging from the length of the annual
growth of the branchlets (which is very short) and the mosses thereon,
the plant may have been a low one from deep forests. E. Matuda’s
specimen is less extreme and transitional to the typical forms of the species
in size of leaves and in inflorescences. It bears both large and small leaves
as well as both solitary and clustered fruits. The species is a variable one
in these characters.

The subcoriaceous leaves, subumbelliform infructescences, and canalic-
ulate pyrenes of //ex lonicerifolia suggest close relationship between this
species and /lex chinensis Sims, which differs in having crenate leaves,
larger fruits, and only 4 pyrenes.
15a. Ilex lonicerifolia var. hakkuensis (Yamamoto), comb. nov.

Ilex hakkuensis Yamamoto Suppl. Ic. Pl. Form. 1: 32, fig. 14. 1925;

ira, Form. Trees 373, fig. 328. 1936.

A completely glabrous tree with conspicuous lenticels on the second
year’s growth; leaves subcoriaceous, castaneous, oblong-elliptic, 5-11 cm.
long, 2-4.5 cm. wide, the base obtuse, the apex shortly acuminate, the mar-
gin entire, the midrib plane or slightly impressed above, elevated beneath;
inflorescences subumbelliform, solitary, axillary, the peduncles 7-15 mm.
long, the pedicels 6-9 mm. long; fruits ovoid-globose, 4—5 mm. in
diameter, the persistent stigma mammiform; pyrenes 5, 4 mm. long,
1.5—2 mm. wide, smooth, deeply unicanaliculate on the back, the endocarp
coriaceous.

292 JOURNAL OF THE ARNOLD ARBORETUM [Vote sax

Cet : Taiwan: Sekiin, Kudo & Sasaki 15244 (TU); Lake Jitu-
gelutan, Y. Kudo in 1929 (A)

The variety hakkuensis differs from typical Ilex lonicerifolia in being
glabrous. There is much variation in the density of the ferrugineous
indumentum of /lex lonicerifolia, but the pubescence is always present at
least in some degree. With species like lex kwangtungensis Merr., Ilex
macrocarpa Oliv., and others, it is evident that the indumentum in the
genus Jlex can be surprisingly variable. Since there are no other morpho-
logical differences nor any marked geographical differences of range
between Jlex lonicerifolia Hayata and /lex hakkuensis Yamamoto, I believe
that this form is not worthy of more than varietal rank.

16. Ilex sterrophylla Merr. & Chun. in Sunyats. 5: 110. 1940.

Ilex pedunculosa sensu Merr. & Chun in Sunyats. 2: 265. 1935, non Miquel.

An entirely glabrous tree up to 15 m. high with a trunk as much as
70 cm. in diameter (ex Chun & Tso), with entire ovate or ovate-elliptic
leaves, simple cymose axillary infructescences, small globose fruit and
shallowly unicanaliculate pyrenes.

Branchlets stout, subterete, plicate-rugose; third year’s growth 5—7 mm
in diameter, cinereous, the lenticels numerous, circular or deltoid; second
year’s growth rufous-castaneous, the lenticels conspicuous; current year’s
growth castaneous-nigrescent, plicate, very glabrous, 3-4 mm. in diameter,
the terminal bud ovoid, the scales densely ciliate. Leaves occurring also
on the second year’s growth, rather crowded, 2-10 mm. apart; stipules
obsolete; petioles 15-25 mm. long, one-fifth to one-fourth the length of
the lamina, entirely glabrous, plane, the distal end evidently winged by
the decurrent leaf-base; lamina coriaceous, completely glabrous, olivaceous
or brunneous, ovate or elliptic, 5-8 cm. long, 2-4 cm. wide; base cuneate
to subrotund, decurrent; apex acuminate, the acumen 5—10 mm. long;
margin entire, very rarely obscurely 1- or 2-toothed; midrib elevated
above, prominent underneath, glabrous, the lateral nerves in 8—10 pairs,
obscure on both surfaces. Inflorescences cymose, glabrous, axillary, on
current year’s growth only; flowers 4- or 5-merous. Staminate inflores-
cences: cymes subumbelliform, 5—13-flowered; peduncles 15—30 mm. long,
twice dichotomous; the secondary axes 1-2 mm. long, the bracts subulate,
1.25 mm. long, the pedicels 3-5 mm. long; calyx patelliform, glabrous,
4- or 5-lobed, the lobes deltoid or rounded, ciliate; corolla rotate, white
(ex Lau), the petals oblong-obovate, one-fifth connate at the base; stamens
4 or 5, shorter than the petals, the anthers oblong, 1 mm. long; rudimentary
ovary minute, ovoid, 1 mm. long, the center rostellate, with 4-parted
apex. Pistillate inflorescences: cymes simple, 3-flowered; peduncles 12-23
mm. long, the pedicels 5-8 mm. long, the calyx ciliate; corolla as in the
staminate flowers; staminodes three-fourths the length ‘of the petals, the
sterile anthers sagittata: ovary ovoid, 2 mm. wide, the stigma thickly
discoid. Fruit ellipsoid, 7-9 mm. long, the persistent calyx explanate,
3 mm. across, rounded, ciliate, the stigma thickly discoid. Pyrenes 4,
oblong, 5—6 mm. long, 3 mm. wide at the back, smooth, estriate, shallowly
concave on the back, the endocarp coriaceous.

1949] HU, THE GENUS ILEX IN CHINA 293

CHINA: Kwangtung: Yao-shan, S. S. Sin 11022 (NY).
Kwangsi: Shang-sze, W. T. Tsang 22376 (A), 24365 (A, NY).
Hainan: Fan-yah, N. K. Chun & C. L. Tso 44093 (A, NY, US); Po-
ting, F. C. How 73683 (A, TYPE); Pak-sa, S. K. Lau 26567 (A); Lok-tung,
S. K. Lau 27324 (A), 27334 (A); Five Finger Mt., F. A. McClure (CCC)
9387 (A

INDO- CHIN: A: Mt. Bana, J. & M. S. Clemens 4380 (A).

Ilex sterrophylla was first described from Hainan Island, where it occurs
as a tree up to 15 m. high with a trunk 70 cm. in diameter (ex Chun &
Tso). Its whitish flowers appear in May (McClure) and its fruits become
dull red in late September

Many of the specimens cited above have been misidentified as Jlex
pedunculosa Miq. or as Ilex purpurea Hassk. (equals I. chinensis Sims).
Because of their dorsally concave pyrenes they are not Jlex pedunculosa,
and ee ee their entire leaves and long petioles they are likewise not
Ilex chine

Ilex cea is closely related to Mex suaveolens (H. Lévl.) Loes.
These two species resemble each other in that they both have simple
cymose infructescences with slender long peduncles and dorsally concave
pyrenes. Jlex suaveolens, however, can easily be distinguished by its
sharp serrate leaves.

17. [lex editicostata H. H. Hu & Tang in Bull. Fan Mem. Inst. Biol.
9: 248

A small tree or shrub up to 6 m. high, with large coriaceous lanceolate
acuminate leaves, simple cymose infructescences, puberulent pedicels, large
fruits, and unicanaliculate pyrenes.

Branchlets stout, nigrescent, plicate; second year’s growth 5—6.5 mm. in
diameter, grayish brown to nigrescent, longitudinally rimulose, the lenticels
Sparse, circular, inconspicuous, the leaf-scars large, semicircular, slightly
elevated; current year’s growth 4-5 mm. in diameter, subterete, plicate

row
ier 12-26 mm. long, one-fifth to one-fourth the length of the lamina,
mm. wide, flattened above, slightly keeled, plicate-rugose beneath;
ae lanceont. 5-12 cm. Tone 2-4 cm. pide base cuneate; apex
acuminate, the acumen 5—15 mm. long, triangular; margins entire, slightly
recurved when dry; midrib elevated on both surfaces, more prominent
above, glabrous, the lateral nerves in 10-12 pairs, usually obscure on both
surfaces, the reticulations sometimes evident. Infructescence a simple
cyme with 1-3 fruits; peduncle 3-6 mm. long, compressed, glabrescent,
the secondary axis completely lacking; bracts linear, obtuse, puberulent,
up to 2.5 mm. long, deciduous; pedicels 4-7 mm. long, plicate-striate,
puberulent or glabrescent with basal deltoid prophylla 0.5 mm. lon
Fruit subglobose, 9-10 mm. in diameter, red, when dry brunneous-
nigrescent, the stigma discoid, convex, its margins lobed; the persistent
calyx explanate, 5-6 mm. in diameter, the lobes 5 or 6, ovate, obtuse, or
rounded, entire, ciliate. Pyrenes 4-6, oblong in outline, 7-8 mm. long,

294. JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

2.5 mm. wide, broadly but shallowly unicanaliculate on the dorsal surface,
the endocarp sublignescent.

CHINA: Chekiang: Tai-suan, R. C. Ching 2219 (A, US). Kw
chow: Fan-ching- chan, mee Chiao & Cheo 691 (ISOTYPE,

US). Kiangsi: Lung- S. K. Lau 4637 (A, US), 4743 (A, ey
Kwangsi: Shang-sze, iV". 7. Tsang 24460 (A); Kwei-ling, IV. T. Tsang
28487 (US).

According to field notes, lex editicostata grows in forests at altitudes
of 600-700 m. It is a small tree or shrub up to 6 m. high and has a
smooth grayish bark. The fruit turns red in late October.

The species agrees with its relative //ex maclurei Merr. in having thick
coriaceous entire leaves and unicanaliculate pyrenes. The latter differs
in that the secondary axes of its infructescences are well developed, its
pedicels are glabrous, and the midrib on the upper surface of its leaf is
not impressed.

17a. Ilex editicostata var. litseaefolia (H. H. Hu & Tang), comb. nov.
Ilex litseaefolia H. H. Hu & Tang in Bull. Fan Mem. Inst. Biol. Bot. 9:
247. 1940.

Leaves smaller, 4-7.5 cm. long, 1.6-3.2 cm. wide, elliptic-lanceolate or
ovate, the apex short-acuminate, the acumen 4-6 mm. long, triangular;
peduncles 4-5 mm. long, glabrescent; pedicels 2-3 mm. long, puberulent
(Keng, Tsang) or glabrescent (Steward); fruit 4-7 mm, in diameter;
pyrenes 5, very shallowly sulcate.

CHINA: Chekiang: Tien-tai-shan, C. Y. Chiao 14522 (A); Y.
Keng 1058 (asotyee of Ilex litseaefolia, A, SS). Kweichow:
chiang-hsien, ne vard & Cheo 962 (A). Kwangtung: Mei-hsien, IV’. 7.
Tsang 21540

This . a a shrub less than 4 m. high which grows at altitudes of
800-2100 m. It differs from the typical form of the species chiefly in its
lower habit and its smaller leaves. When Hu and Tang prcposed Jlex
editicostata they distinguished it from /lex litseaefolia on the basis of its
more robust habit, larger and thicker leaves and larger fruit with 6-8
pyrenes. The number of pyrenes is not a constant character, for in
dissecting fruits of the isotype (Steward, Chiao & Cheo 691) I found many
with only 5 pyrenes. The lanceolate entire leaves with the midrib elevated
on the upper surface, the simple cymose infructescence with puberulent
pedicels, and the shallowly canaliculate pyrenes of /ex litseaefolia, all agree
with the characters of /lex editicostata. Furthermore, in R. C. Ching 2219
the leaf-size is transitional. The lower smaller leaves are like the normal
leaves of Ilex litseaefolia, and the upper larger ones are like those of
Ilex editicostata. 1 therefore interpret Jlex litseaefolia as an ecotype
from a somewhat higher altitude, and place it as a variety of J/lex
editicostata.
17b. Ilex editicostata var. chowii (S. Y. Hu), comb. nov.

Ilex chowii S. Y. Hu in Ic. Pl. Omei. 2: pl. 157. 1946,

Loaves elliptic-lanceolate, 7-15 cm. long, 2.5-5 cm. wide; infructes-

1949 HU, THE GENUS ILEX IN CHINA 295

cence with 3 or 4 fruits; peduncles 6-9 mm. long, compressed; pedicels
3-7 mm. long, entirely SES: calyx 5- or 6-lobed, glabrous, ciliate;
fruits anees 7-9 mm. in diameter, much wrinkled an dry; pyrenes
5 or 6, unicanaliculate on the back.

CHINA: Szechuan: Mt. Omei, H. C. Chow 8138 (tyre of Ilex
chowtt, Sz; isoTyPE, A). Kwangsi: Kwei-lin, HW’. T. Tsang 28487 (A);
Yao-shan, C. Wang 40342 (A).

This variety differs from the species chiefly in being entirely glabrous.
In its coriaceous entire large leaves with acuminate tips, simple cymose
infructescences, and unicanaliculate pyrenes, it conforms to Jlex
editicostata. It is also related to /lex maclurei Merr., which differs in
having well-developed secondary axes in the infructescences.

18. ex ferruginea Hand.-Mat. Symb. Sin. 7: 657, pl. 10, fig. 24. 1933.

A shrub (ex Schneider) or a tree (ex Handel-Mazzetti) with pubescent
branchlets, ovate remotely crenate-serrate leaves, short petioles, simple
cymose axillary infructescences and small unicanaliculate pyrenes.

Branchlets subterete, smooth; fourth year’s growth 3-4 mm. in diameter,
the lenticels minute, circular, sparse, the leaf-scars deltoid, slightly elevated;
third and second years’ growth puberulent, the lenticels lacking; current
year’s growth angular, densely villose, striate-sulcate, 1.5-2.5 mm. in
diameter. Leaves found even on third year’s growth, 8-12 mm. apart;
stipules minute, buried in the indument; petioles short, 4 mm. long, one-
thirteenth to one-tenth the length of the lamina, villose, flat or slightly
grooved above; lamina coriaceous, olivaceous, villose especially along the
midrib, ovate, those near the apex of the shoots ovate-elliptic, 2—5.5 cm.
long, 1.5-3.5 cm. wide; base truncate or rounded, rarely cordate or obtuse;
apex shortly acuminate, the acumen 3-4 mm. long; margins remotely
crenate-serrate, the teeth nigrescent; midrib plane or very slightly impressed
above, elevated beneath, villose on both surfaces, the lateral nerves in
8-10 pairs, obscure above, prominent underneath, the reticulations obscure.
Staminate inflorescences: subumbelliform or cymose, 1—6-flowered, solitary,
axillary to scales or leaves of the current year’s growth; peduncles 5-10
mm. long, the pedicels 1-3 mm. long; secondary axis rarely developed,
1-2 mm. long; bracteoles lanceolate, pubescent; flowers 5—7-merous; calyx
subcampanulate, pubescent, 2—3 mm. across, deeply 5—7-lobed, the lobes
ovate-deltoid, 1 mm. long, obtuse, strongly ciliate; corolla rotate, 6 mm.
across, the petals erose, eciliate, one-fifth connate at the base; stamens 5,
equaling the petals in length, the anthers oblong; rudimentary ovary sub-
globose, the apex shortly rostellate, the beak densely hirsute. Infructes-
cence a simple 3-flowered cyme found only on the current year’s growth,
axillary, villose; peduncles 6-10 mm. long, slightly enlarged at the distal
end; bracts of the individual flowers linear-lanceolate, 5 mm. long, villose;
pedicels 5—9 mm. long; prophylla basal, deltoid, minute. Fruit globose,
5-7 mm. in diameter, when dry shiny, castaneous, wrinkled, the stigma
discoid, the persistent calyx explanate, 3-4 mm. in diameter, villose and
ciliate, the lobes 4 or 5, broadly deltoid, acute. Pyrenes 5 (4—6 ex Handel-

296 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Mazzetti), smooth, 2.5 mm. long, 1 mm. wide, and broadly unicanaliculate
on the back.

CHINA: Kweichow: Ping-chow, Y. Tsiang 7117 (NY). Yun-
nan: FP, Ducloux 148 (NY); Tung- chuan, north of Yunnaniu [Kunming ],
Handel-Massetti 468 (1sotyrr, A); C, Sebnctler 293 (A

The description of the staminate flowers is drawn Ant Busine 148.

Judging by its indumentum, its simple infructescences, and its broadly
sulcate pyrenes, /lex ferruginea is closely related to Ilex kwangtungensis
Merr. which differs in having larger chartaceous elliptic leaves (up to
15 cm. long) and larger pyrenes up to 10 mm. long.

When Handel-Mazzetti first collected in China he traveled with C.
Schneider, starting in March 1914. His number 468, the type of Jlex
ferruginea, was collected on the same day and in the same locality as
C. Schneider 293. Most likely these specimens came from the same plant.
On casual examination one gains the impression that the indumentum is
ferrugineous, particularly along the midrib. On close examination, how-
ever, it is found that the supposedly ferrugineous color of the indumentum
is due to the presence of minute particles of red earth lodged in it. This
extraneous substance can be easily washed off; the pubescence is actual]
gray. Thus the specific name selected for the species is ed ia
In his decupios of the fruits Handei-Mazzetti stated, “pyren a eas
dorso carinatae.” Apparently he confused the dorsal and i cn

e pyrenes are not carinate on the back, but rather unicanaliculate.
Furthermore, he placed /lex ferruginea in the section Microdontae. This
section is characterized by fasciculate inflorescences. He was apparently
misled because the toothing of the leaves resembles that of Jlex corallina
Franch., which does belong in the Microdontae. The presence or absence
of teeth and their shape are very unreliable characters for use in the
classification of lex. Since the inflorescences of this species are cymose,
solitary, and axillary, and the pyrenes are unicanaliculate, its natural
relationship is with /lex kwangtungensis Merr.

19. a suaveolens (H. Lévl.) Loes. in Ber. Deutsch. Bot. Ges. 32:
541. 1914; H. Lévl. Fl. Kouy-tchéou 201. 1914; Rehd. in Jour.
ee Arb. 14: 239, 1933,

Celastrus suaveolens H. Lévl. in Fedde, Rep. Spec. Nov. 13: 263. 1914.

Ilex purpurea sensu Hand.-Mzt. Symb. Sin. 7: 655, 1933, in part; Rehd.

in Jour. Arnold Arb. 14: 239, 1933, in part; non Hassk

A beautiful, entirely glabrous evergreen tree with coriaceous ovate-
lanceolate leaves, subumbelliform, rarely cymose inflorescences, peduncles
mostly exceeding the petioles in length, small globose fruits and smooth
ecanaliculate or slightly dorsally impressed pyrenes

Branchlets straight, the third year’s growth ca. 5 mm. in diameter,
cinereous, the lenticels minute, circular-ellipsoid, very numerous but incon-
spicuous, the leaf-scars obovate-deltoid, plane; current year’s growth

3-4 mm. in diameter, castaneous or brunneous. Leaves occurring also

on second year’s growth, 3-15 mm. apart; stipules lacking or very minute

peal

1949] HU, THE GENUS ILEX IN CHINA 297

and caducous; petioles unusually long, compressed, 15-30 cm. long, 2 mm.
wide, one-fifth to one-fourth the length of the lamina, plane above, nar-
rowly winged by the decurrent leaf-base; lamina coriaceous, olivaceous or
brunneous, ovate, elliptic or lanceolate, 5-10 cm. long, 2.5—4 cm. wide;
base rounded, obtuse or cuneate; apex acuminate, the acumen 5-10 mm.
long; margin crenate-serrate or sometimes subentire, the tips of the teeth
nigrescent; midrib elevated on both surfaces, the lateral nerves in 11-14
pairs, obscure above, evident beneath, the reticulations of the veinlets
inconspicuous. Inflorescence subumbelliform, rarely cymose, solitary,
axillary, on the current year’s growth; cymes 3—7-flowered, twice or more
dichotomous; peduncles glabrous, slender and long for the genus, 15-35
mm. long, usually longer than the petioles, compressed, gradually en-
larged at the distal end; secondary axes usually poorly developed, 0-2 mm.
long; bracts lanceolate, 2.5 mm. long, often ciliate. Staminate inflores-
cences: pedicels slender, 3-8 mm. long, in fruit 1 cm. long; flowers reddish
white (ex Steward and Cheo), 4- or 5-merous; calyx broad subcampanulate,
3 mm. across, the lobes 4 or 5, ovate-deltoidy 1 mm. long, 1.5 mm. wide at
the base, glabrous, ciliate, acute; corolla rotate, 6-7 mm. across, the petals
ovate-oblong, 3 mm, long, 2.5 mm. wide, reflexed, one-eighth connate at
the base; stamens shorter than the petals, the anthers ovoid, oblong,
1,25 mm. long; rudimentary ovary globose, 1.5 mm. in diameter. Pistillate
inflorescences: peduncles 1.5-2.5 cm. long; calyx and corolla as in the
staminate flowers; staminodes one-half the length of the petals, the
sterile anthers cordate; ovary ovoid, globose, 2 mm. in diameter, the
stigma thick-discoid, 4- or 5-lobed. Fruit globose-pomiform, 5-6 mm. in
diameter, red, when dry brownish red, the stigma mammiform or discoid.
Pyrenes 4 or 5, oblong in outline, trigonous in cross-section, 4-5 mm.
long, smooth, estriate, esulcate, sometimes slightly depressed along the
median dorsal line, the endocarp thickly coriaceous or sublignescent.

CHINA 6 hekiane, 2 ‘Sia -chumsiesG Ching. 1/21. CAsIN Yeas),
2611 (A), 2614 (A); Ping-yung, R. C. Ching 2221 (US), 2172 (NY, US);
Yen-tang, H. H. Hu 229 (A); Tai-shun, Y. L. Keng 294 (A). Kiangsi:
Wu-ling, Y, K. Hsiung 5399 (A); Lin-ying-hsien, Steward & Cheo 363
CNM) = oTlupei ( Hupeh): 2a sie iWalson 20st (A, K),3( Veitch Exp: )
2031A (A, K). Kweichow: Kwei-yang (or Kouy-yang), E. Bodinier
2663 (TypPE, fragment and photo, A); Pin-fa, /. Cavalerie 17 bis (fragment
and photo, A); Fan-ching-shan, Steward, Chiao & Cheo 791 (A), 930 (A);
Handel-Mazzetti 10524 (A). Kwangsi: Lin-yuin, Steward & Cheo
363 (A); Kwei-lin, W. T. Tsang 28490 (A). Kwangtung: Sin-fung,
Y. W. Taam 109 (A), 971 (A).

Ilex suaveolens is closely related to and intermediate between Jlex
chinensis Sims and Ilex sterrophylla Merr. In its ovate-elliptic-lanceolate
leaves with crenate-serrate leaf-margins it resembles //ex chinensis, which
differs in having canaliculate elongated pyrenes, 2- or 3-dichotomous cy-
mose inflorescences, and peduncles of the infructescences shorter than the
petioles. In the long-pedunculate loose subumbelliform inflorescences,
depressed-globose fruits, and nearly smooth pyrenes, it resembles //ex
sterrophylia Merr., which differs in having entire leaves.

298 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

20. Ilex manneiensis, sp. nov.

Arbor; ramulis puberulentibus; foliis chartaceis vel tenuiter coriaceis,
ellipticis, 8-16 cm. longis, 2—4.5 cm. latis, basi rotundatis vel cuneatis et an-
guste decurrentibus, apice acuminatis, acuminibus 10 mm. longis deltoideis,
brunneis; costa supra paullo elevata et dense villosa, subtus manifeste
elevata flavo-tomentosa, maturitate glabrescente; nervis lateralibus
utrinqgue 15-17, prominentibus; inflorescentiis solitariis, axillaribus, 3-
floris, pedunculis 5—7 mm. longis, pedicellis 2 mm. longis; floribus 4—6-meris,
calycibus 3 mm. diametro pubescentibus; corolla 6-7 mm. diametro, petalis
sparsim ciliatis, staminodiis quam petalis brevioribus, ovario ovoideo,
stigmate mammiforme; fructibus globosis, 9 mm. diametro, glabris; pyrenis
5 vel 6, laevibus, 8 mm. longis, dorso late sed non profunde 1-canaliculatis.

An evergreen tree up to 9 m. high with puberulent branchlets, charta-
ceous or thin-coriaceous lanceolate entire acuminate leaves, simple
flowered cymose inflorescences, very shortly pedicellate cls; ‘id
unicanaliculate pyrenes.

Branchlets stout, plicate and rugose, brunneous; third year’s growth
5 mm. in diameter, the lenticels numerous, circular, sometimes coalescent,
conspicuous, the leaf-scars elevated, crescent-shaped, deltoid; second year’s
growth plicate, puberulent, the lenticels lacking; current year’s growth
yellowish pubescent with short curly hairs, subterete, plicate, the terminal
buds ovate, 5 mm. long, the outermost scales yellowish pubescent, the inner
ones glabrous, all densely ciliate. Leaves occurring also on second year’s
growth, 2-4 mm. apart; stipules wanting; petioles 15-20 mm. long, one-
eighth to one-sixth the length of the lamina, yellowish pubescent when
young, glabrescent at maturity, plane above, plicate-rugose beneath;
lamina chartaceous or thin-coriaceous, brunneous, glabrous except on the
midrib, elliptic, 8-16 cm. a 2-4.5 cm. wide; base rounded to cuneate,
narrow wly decurrent; a acuminate, the acumen up to 10 mm. one
deltoid; ae slightly ae and densely villose above, the indumen-
tum persistent, elevated and yellowish tomentose beneath, the indumentum
deciduous; the lateral nerves 15-17 pairs, prominent on both surfaces,
parallel, anastomosing near the margin, the reticulations obscure above,
evident beneath. Pistillate inflorescence a simple 3-flowered cyme, soli-
tary, axillary, pubescent, on current year’s growth; peduncles 5-7 mm.
long, greatly compnesced: in fruit 4 mm. wide; bracts minute, deltoid,
ciliate; flowers 4—6-merous, their pedicels very short, 2-3 m ng;
calyx patelliform, 3 mm. across, puberulous, the lobes broadly deltoid or
rounded, ciliate; corolla rotate, 6-7 mm. across, the petals oblong-ovate,
sparsely ciliate, 3 mm. long, 2.5 mm. wide, one-sixth connate at the base:
staminodes three-fourths the length of the petals, the sterile anthers
ovoid; ovary ovoid, 2 mm. long, the stigmas mammiform. Staminate
flowers not seen. Fruit globose, 9 mm. in diameter, castaneous when dry,
the persistent calyx explanate, 5 mm. in diameter, 5- or 6-lobed, the lobes
ciliate, the stigma discoid. Pyrenes 5 or 6, oblong in outline, trigonous
in cross-section, back U-shaped, 8 mm. long, 2.5 mm. wide, smooth,
shallowly but widely unicanaliculate on the back, the endocarp woody.

1949] HU, THE GENUS ILEX IN CHINA 299

CHINA: Yunnan: Mannei, A. Henry 9628 (tvpez, A; 1sotyPeEs, K,
US); Mengtze, A. Henry 11014 (A, K).

The description of the pistillate fewer is drawn from Henry 11014.

In having the midrib yellowish tomentose beneath this species resembles
Ilex atrata W. W. Sm., but the latter differs in having its leaves serrate,
its pedicels longer ar the peduncles, and its branchlets glabrous. In its
entire leaves and indumentum it resembles Jlex lancilimba Merr., which
differs in having very short peduncles and usually large solitary irate:

21. Uex chinensis Sims in Bot. Mag. 46: pl. 2043. 1819; Lindl. in
Donn, Hort. Cantab. ed. 10, 52. 1823; DC. Prodr. 2: 14. 1825;
P. N. Don in Donn, Hort. Cantab. ed. 13. me 1845; Maxim. in
Mém. Acad. Sci. St. Pétersb. VII, 29(3): 40. 1881; Forbes &
Hemsl. in Jour. Linn. Soc. Bot. 23: 115. ae. Loes. in Nov. Act.
Acad. Caes. Leop.-Carol. Nat. Cur. 78: 495 (Monog. Aquif. 1:
495). 1901
Ilex purpurea Hassk. Cat. Hort. Bog. Alt. 230. 1844; Loes. in Nov. Act.
Acad. Caes. Leop.-Carol. Nat. Cur. 78: 111 (Monog. Aquif. 1: 111).
1901; Rehd in Jour. Arnold Arb. 8: 156. 1927, et 14: 239. 1933; Hand.-
Mat. Symb. Sin. 7: 655. 1933; Cheng ex P’ei, in Contr. Biol. Lab.
Sch. Soc ehinany (2)r e711 1934:
llex oldhami Mig. Ann. Mus. Bot. Ludg.-Bat. 3: 105. 1867, et Cat. Mus.
Bot. Ludg.-Bat. 19, 167. 1870; Franch. & Sav. Enum. Pl. Jap. 1: 77.
1873; Maxim, in Mém, Acad. Sci. St. Pétersb. VII, 29(3): 25, 38,
pl. 1, fig. 4. 1881; Forbes & Hemsl. in Jour. Linn. Soc. Bot. 23:
1886; Dunn & atenen in Kew Bull, Add. Ser. 10: 60. 1912; Ghne
in Mem. Sci. Soc. China 1: 141. 1924: Belval in Mus. Heud. Not. Bot.
@hing 22) 11933:
Ilex lucida Blume ex Miq. Cat. Mus. Bot. Lugd.-Bat. 167. 1870, nom, nud.
Tlex Sele var. ae (Miq.) Loes. ex Diels in Bot. Jahrb. 29: 435.
5 nud.; et in Nov. Act. Acad. Caes me -Carol. Nat. Cur. 78:
1B Gee nee 1: 112). 1901, et in Sarg. P . Wils. 1: 76. 1911;
Chung in Mem. Sci. Soc. China 1: 141, 1924; aa Ill. Man. Chinese
Trees 655, fig. 546. 1937. Syn. nov.
Callicarpa cavaleriet H. Lévl. in Fedde, Rep. Sp. Nov. 9: 455. 1911, et
Fl. Kouy-Tchéou 439, 1915.
Celastrus bodinicri H, Lévl. in Fedde, Rep. Sp. Nov. 13: 263. 1914.
Embelia rubro-violacea H. Lévl. in Fedde, Rep. Sp. Nov. 10: 375. 1912,
t Fl. Kouy-Tchéou 285, 1914.
Symplocos courtoisiu H. Lévl. = al Acad, Cienc. Art. Barc. JI; 12:
256 (Cat. Pl. Kiang-sou 22).
Syiroggt vel enone Kaempf. oe Exot. 779. 1717,
Natiome Kaem pic lewSelw eked. Banks. p/, 35, 1/91
A large entirely glabrous evergreen tree up to 134 m. high, with shiny
deep green crenate-serrate leaves, bright red ellipsoid drupes, and dorsally
l-canaliculate pyrenes.
Trunk 30 cm. in diameter; bark gray, smooth (ex Ching & Tso);
branchlets terete, third year’s growth 5 mm. in diameter, cinereous, the
lenticels minute, circular-ellipsoid, numerous and inconspicuous, the leat

300 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

scars crescent-shaped or narrowly deltoid, slightly elevated; current year’s
growth 3 mm. in diameter, angular, plicate, castaneous-nigrescent, gla-
brous. Leaves occurring also on second year’s growth, 3-10 mm. apart;
stipules minute, deciduous; petioles 8-10 mm. long, one-eighth to one-
sixth the length of the coe plane, sometimes slightly sulcate above,
striate beneath, narrowly winged by the decurrent lamina; lamina ia
coriaceous, poeeo elliptic, lanceolate or rarely ovate, 5 11 cm. long,

—4 wide; base obtuse or cuneate; apex acuminate, the acumen
5-10 mm. long: margin crenate or rarely serrate; midrib plane above,
elevated beneath, the lateral nerves 6-9 pairs, rather obscure above,
prominent beneath, glabrous, the reticulations evident beneath. Inflo-
escences cymose, solitary, axillary, glabrous, on current year’s growth.
Staminate inflorescences: cymes 7—15-flowered, 3 or 4 times dichotomous
or sometimes irregularly dichotomous; peduncles 4-14 mm. long, usually
exceeding the petioles; secondary axis well developed, 2-5 mm. long, the
bracts of the ultimate cymules lanceolate, 2 mm. long, often ciliate; flow-
ers lavender or red, 4- or 5-merous; pedicels 2 mm. long, glabrous with
0 or 1 deltoid ciliate prophyllum 0.5 mm. long; calyx subcampanulate,
glabrous, ca. 2.5 mm. across, the lobes broadly deltoid, obtuse or rounded,
ciliate; corolla rotate, the petals ovate, ca. 2.5 mm. long, 1.8 mm. wide,
reflexed at anthesis; stamens shorter than the petals, the anthers ellipsoid,
1.5 mm. long; rudimentary ovary conic, 0.7 mm. long. Pistillate inflo-
rescences: cymes 3—7-flowered, once or twice dichotomous; peduncles
3-10 mm. long, always shorter than the petioles, compressed, usually en-
larged at the distal ends, the secondary axis poorly developed; pedicels
6-10 mm. long; calyx and corolla as in the staminate flowers; staminodes
one-half the length of the petals, the sterile anthers cordate; ovary ovoid,
the stigma thickly discoid, inconspicuously 4- or 5-lobed. Fruit ellipsoid,
10-12 mm. long, 6-8 mm. thick, shiny, red, when dry smooth, shiny,
castaneous-nigrescent, the persistent stigma prominent, discoid, convex.
Pyrenes 4 or 5, narrowly lanceolate in outline, trigonous in cross-section
with a narrow U-shaped back, 9-11 mm. long, 2.5-3 mm. wide, smooth,
deeply 1-canaliculate along the median dorsal line, the endocarp thick-
coriaceous.

CHINA: Kiangsu: without precise locality, Ch. d’Argy in 1844-66
(tyPE of Symplocos courtoistt, photo and fragment, A); Shien-chuan-tze,
Ching & Tso 453 (A); I-shing, Ching & Tso 481 (A); Y. L. Keng 2669
(A); K. King 12293 (NY); Ma-shan of Tai Lake, Ching & Tso 697 (A),
701 (A); Soochow, Ching & Tso 803 (A); Tso 1686 (A); Nanking, Y.
Tsiang 337 (A); Tso 1722 (A). Anhwei: Wang-shan, W. C. eee
4550 (US); N. K. Ip 5116 (A); R. C. Ching 2758 (A); Chemen, ae Oe
Ching 3172 (A); Chiu- -shan, R. C. Ching 7258 (US); Oe ake meg
Y. ¥, Lt 260 (US); Chien- shan, Co. Lae or i‘ dof 2),
Chekiang: MHangchow, C. Y. Chiao 18877 (NY, US), 18820 (US),
18877 (US); Sia-chu, R. C. Ching 1588 (A, NY, US), 2597 (A), 4827
(A); Ping-yung, R. C. Ching 2142; Ning-po, E. Faber (NY); Tsing-tien,
Y. L. Keng 124 (A), 232 (A); Tai-shun, Y. L. Keng 288 (A); I-wu,

1949] HU, THE GENUS ILEX IN CHINA 301

VRC CNG -8 990 (ou ee itaionines 1 lid 049 (CNB Vn ake ied Theses lino
ie LSU OLO een dd Ti iu o/o CA Ne Kelensnatl, Ss a OOee
(A, US), 4281 (A, US), 4397 (A, US); Lung-nan, S. K. Lau 4842 (A
US); Fow-liang, K. Ling, in 1924 (A); Chrtan-Heien, W. T,. Tsang
27585 (A); Ling-chuan, Y. Tsiang 9831 (NY); Tsoong-jen, Y. Tsiang
10204 (NY); Tu-hwa LCI Wasa! (ex Herb. Univ. Nank. 1562) (A).
KMiamesi-hukivem Border: 2.4. Wang 327.(4).) Huper Ceupeh):
Yang-hsien-hsien, H. G. Cheo “18343 (NY); Chien-shih, H. C. Chow 1438
(A, NY), 1554 (A, NY); En-shih, H.C. Chow 1877 (A, NY); western
Hupei, W. Y. Chun 4004 (A); A. Henry 1562 (US), 3911 (B, US), 4440
(GAY) 021 7> GA INNS) ae. a SOW SOG CALUNIS) sel 92 0 GAS NING, oe
LOS SK) eh sae 1K INS) 26994 (K), -2700Ae Oke) adlechens: E.
Wilson 689 (A, US), 3096 ay ce (A, ae Chang- -yang, ieee loa Ve i
LOCK 2097 (Oe, NY 2 US), Hithan) Sinenine so iG. fan ce
Von Vee wl p40 es eye: Handel-Mas gett, 11291 ee. Chang-sha, Handel-Mas-
getitt 11415 (A), 11525 (A, US); Wu-kang, Handel-Mazzeiti 11987 (A).
Kweichow: Kwei-yang (kouy-yang), E. Bodinier 2384 (type of
Celastrus bodiniert, fragment, A); Doo-shan (Tou-shan), J. Cavalerie 2624
(tyPE of Callicarpa cavaleriet, fragment, A); without precise locality, J.
Esquirol 429 (isotyre of Embelia rubro-violacea, A) ; Wong-mo, Chen-ieng,
. W. Teng 90962 (A); Kwei-ting, ¥Y. Tsiang 5558 (NY), 9229 (NY);
Tuh- shan, Y. Tsiang 6603 (NY); Yin-kiang, Y. Tstang 7908 (NY), 7851
(NY); Tu-yun, Handel-Mazzetti 10730. Szech Yuan: O-mei-shan (Mt.
Omen), Ge CiiaovonC SS. fan 271 CA) SC. Chow 0064 (oz) OL01
(S52); 7172 CSZ eer eC. Chow S222 (A), 12173 ay 12 Faber 20344 (NY);
WE PONG IIPUaNest) yl 9927. (Sz), 150L9 (Sz) 5003 (Sz athena. (sz),
19080 (32), FOA0G Comyn lsMte (oz) 29119 (Sz). 19149 (Sz ).e 1G. aces 4or
C32) 549356, (Sz). 461s (Sz 5-462 (Sz) > C, L Sum [2249(S2 2 S.C: ae S
Chang 713 (A), 818 (A), 1548 (A); L, ¥. Tat 473 (A), 1135 (A);
Wang 23005, (CA 25093 (A): C. ¥. Yao 5058. (SS) Chien-wei- ae
re H. Chung 241; without precise locality, Faber 20 (A); Kwan-hsien,
wea ROnGuIAOlORCSz sy. Le Li 1927 (A) Pienshsien a. S. iw 2000
Oe Lo-shan-hsien, F. T. Wang 23546 (A); Kia-ting, H. H. Tai 135
(A), 243 (A); L. Y. Tat 1441 (A); western China (probably Szechuan),
Beare Wal someoconthe) oikange ¢ -C. Y. Chiao 168s, Youina n-
fale pee 344 (NY). Fukien: Yen-ping, H. H. Chung 3532 (A);
Ku-tien, H. H. Chung 7915 (A); Sin-kai-kau, Dunn 2472 (A); Foochow,
Tang Siu eee 7232 (A). Kwangtung: Tai-ping, W. Y. Chun 5669
A); Lin-hsien, C. O. Levine (CCC) 3186 (A); Loh-ch’ang, W. T. Tsang
20771 (A, NY); Ch’uan-hsien, W. T. Chang 27585 (US); Lok-chong,
C. L. Tso 20644 (NY); Yang-shan, 7. M. Tsui 718 (NY), 497 (A, NY,
US). Kwangsi: Kwei-ling, W. T. a 27 60L CAs Us) + Yao-shan,

C. Wang 40621 (A). Hainan: Lok-tung, S. K. Lau 27274 (A).
APAN: Buerger ae R. Oldham 109 (A), 150 CA) 25 (isoryees
of Ilex oldham, A, G, NY); Herb. Lugd.-Bat. (1soryre of Ilex lucida,

NY); G. Masamune in "1915 ONY ),.1m 1923 (NY), in 1940 (NY); Maxi-
mowics (G, NY); Kenzo Shiota 5341 (A); K. Sakurai (A); Jashiro for
Wilson (A).

Ilex chinensis is very well known among the Chinese, as it is extensively
used as an ornamental tree. The branchlets with their deep green shiny

302 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

leaves and bright red fruits are used to decorate temple-courts and large
halls during the period from December to February which includes the
Chinese New Year. It is known as “Tung-ching,”’ meaning “wintergreen,”
or ‘“‘wan-sho-hong,”’ meaning “everlasting red.” It has been collected in
thirteen Chinese provinces and on Hainan Island, at altitudes ranging
from sea level in East China up to 2000 m. in West China. It is often
cultivated, developing into a tree up to 12 m. high and forming a trunk
up to 30 cm. in diameter. Its lavender flowers appear in May, and its
fruits begin to turn red in September. Quantities of fruiting branches are
sold for ornamental purposes at the time of the Chinese New Year
festival in February. The young shoots are sometimes blanched and used
for salad. The crushed pyrenes are reputed to have tonic effects and are
used in medicine.

This beautiful ornamental plant attracted the attention of early Euro-
pean collectors in the coastal cities. It was introduced into England in
1810 (ex Donn). Sims in July, 1814, studied and illustrated a flowering
male plant growing in Messrs. Malcolm & Sweet’s Nursery said to have
been introduced from China. He named it /lex chinensis and published
his description and colored plate in 1819. Apparently only male plants
were then available in England and the species soon disappeared from
cultivation there. DeCandolle in 1825 knew it only from Sims’ illustra-
tion and suggested that it might be a synonym of /lex dahoon Walt. =
Ilex cassine Linn. The species was soon forgotten. Maximowicz in 1881
remarked concerning /lex chinensis, “Planta nunc obsoleta videtur, a
Goeppert in enumeratione /licum hortensium omissa .. . omnino omitti-
tur.” Forbes Hemsley in their Enumeration of Chinese Flowering
Plants regarded it as a doubtful species, stating, “The cymose, . . . inflores-
cence on peduncles exceeding the petioles is unknown to us among Chinese
hollies, and there may be some error as to the origin of this species.” Even
Loesener in his Monographia Aquifoliacearum also listed Jlex chinensis as
a species of doubtful status.

Before the eighteenth century Japan had essentially a Chinese culture.
Many cultigens were introduced into that country from China, and among
them this //ex which I interpret as /lex chinensis. In 1844 Hasskarl
based his /lex purpurea on specimens from Nanawe, Japan. Twenty-
three years later Miquel based lex oldhami on specimens collected by
Oldham at Nagasaki. Fruiting specimens received from various parts
of China, of course, did not match Sims’ illustration of the male plant,
flex chinensis. Some were named Jlex purpurea and others [lex oldhami,
but finally the older of these two names became very generally accepted for
this widely distributed Chinese species. Being now convinced that these
numerous staminate specimens of the present species do represent the
long obscure and misunderstood /lex chinensis, I accept this name as the
oldest valid one for the species and reduce Ilex purpurea Hassk. (1884),
Ilex oldhami Mig. (1867), and lex purpurea Hassk. var. oldhami (Miq.)
Loes. (1901) to synonymy. I agree with Rehder in the reduction of the
several Léveillé binominals.

1949] HU, THE GENUS ILEX IN CHINA 303

22. Ilex kwangtungensis Merr. in Jour. Arnold Arb. 8: 8. 1927;
Groff in Lingnan Sci. Bull. 2: 64. 1930; Chun in Sunyats. 2: 71.
1934; Tanaka & Odashima in Jour. Soc. Trop. Agr. 10: 372. 1938;
Chun in Sunyats. 4: 224. 1940; Merr. & Chun. in Sunyats. 5: 106.
1940; Masamune, Fl. Kainant. [Hainan] 174. 1943.

lex fea var. pilostor Hand.-Mzt. Symb. Sin. 7: 654. 1933.

Syn. 4

lex kz Se Senne var. pilosissima Hand.-Mat. op. cit. 655. Syn. nov.

Ilex shweliensis Comber in Notes Bot. Gard. Edinb. 18: 57. 1933. Syn.

nov.

Ilex latifrons Chun in Sunyats. 2: 69. 1934. Syn. nov.

Ilex phanerophlebia Merr. in Lingnan Sci. Jour. 13: 36. 1934. Syn. nov.

Ilex latifrons var. pilosissima (Hand.-Mzt.) Chun op. cit. 70. Syn. now.

mall evergreen tree up to 9 m. high with puberulent branchlets and
inflorescences, large ovate-elliptic finely serrate or subentire pubescent
leaves, iong lanceolate bracts, large ellipsoid fruits and deeply and broadly
canaliculate pyrenes.

Branchlets straight, terete, the third year’s growth 4—5 mm. in diameter,
cinereous-brunneous, glabrescent, the lenticels minute, dense, circular,
rather evident, the feat -scars semicircular, slightly clevatece second year’s
grow mm. ‘thick, smooth, nigrescent or brunneous, cinereously
puberulent under a lens; current year’s PA angular, 2—3 mm. thick,
ferrugineous or fulvous-pubescent. Leaves found even on third year’s
growth, 5-30 mm. apart; stipules lacking; petioles robust, 10-18 mm. long,
2 mm. thick, one-ninth to one-seventh the length of the lamina, triangular,
narrowly sulcate above, fulvous-pubescent; lamina subcoriaceous, brun-
neous or atro-olivaceous, curly villose on both surfaces, sparsely so above,
densely so beneath and along the midrib, ovate-elliptic, oblong or lanceolate,
7-16 cm. long, 3-6 cm. wide; base rounded or obtuse; apex acuminate,
the acumen 5-10 mm. long; margin minutely serrate or subentire, slightly
recurved; midrib plane or slightly elevated and densely villose AROS very
prominent, elevated and villose beneath, the lateral nerves 9-11 pairs;

s growth only. Staminate inflorescences: cymes 12-—20-
flowered, 3 or 4 times dichotomous, pubescent; peduncles 9-12 mm. long,
secondary axes usually present 3-6 mm. long; bracts filiform, lanceolate,
fulvous-villose, often up to 7 mm. long, basal to the tertiary axis; flower
4- or 5-merous, red or pink, 7-8 mm. across; pedicels 2 mm. long, fulvous-
villose becoming nearly glabrous; calyx patelliform, 2.5-3 mm. across,
the lobes orbicular-ovate, rounded, puberulous, entire, ciliate; corolla
rotate, the petals oblong, 3.5 mm. long, 2 mm. wide, one-sixth connate at
the base; stamens charter than the petals, the anthers oblong, 1.5 mm.
ong; Ir dimen ntary ovary narrowly conical, 1.5 mm. long, the apex
rostellate. Pistillate inflorescences: cymes a2 7-flowered, 1 or 2 times
dichotomous, pubescent, the secondary axes 3-4 mm. lone. the bracts
lanceolate, submedian on the secondary axis; flowers 4-merous; pedicels

304 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

4—7 mm. long, flavous-villose or glabrescent; calyx ciliate; corolla rotate,
the petals ovate, 2.5 mm. long; staminodes shorter than the petals, the
sterile anthers ovoid; ovary ovoid, 2 mm. in diameter, the stigma mammi-
form, 4-lobed. Fruit red, ellipsoid, 12 mm. long, 9 mm. in diameter,
nigrescent when dry, brown, wrinkled, shiny, the persistent calyx explanate,
ciliate, the stigma prominent, 4-lobed and slightly ridged. Pyrenes 4
smooth, 6 mm. long, 3 mm. wide on the back, dorsally unicanaliculate, the
canal 2 mm. deep, 1.7 mm. wide, the endocarp sublignescent.

CHINA: Chekiang: Tai-suan, R. C. Ching 2142 (A, US). Fu-
kien: Dunn, (Hongk. Herb. no. 2476) (A); Tien-hwa-schan (Ting-
chow), Handel-Maszzeits ves of llex kwangtungensis var. pilosior,
A). Kwangtung: Lung-t’au Mt. To, Tsang & Tsang (CCC) 12741
CAS, 12383 (A. LS), Di64 (A, LU, US); Pok-lor, Fung Hom ok
(<= LU 16971, TPE e llex phanerophlebta, NY); Wung- -yuen, S. K,
eo0o Te): Sin- fung, Y. W. Taam 199 (A), 307 (A), 383 (A); Kow ioe
Y. W. Taam 2300 . Ta-pu, W’. T. Tsang 21188 (A); Hwei- — WV.
Tsang 25607 (A), 25695 (A), 25899 (A). Hainan: Po-ting, F. é.
How 72489 (A), 73047 (A), 73349 (A), 73730 (A); C. Wang 33520 (A,
US); Lok-tung, S. K. Lau 26978 (A); Bo-ting, S. K. Lau 28021 (A), 28277
(A). Kwangsi: Kwei-lin, W. T. Tsang 28122 (A, US); Yao-shan,
C. Wang 39321 (A), 39376 (A). Hunan: Chang-ning-hsien, C. S. Fan
a Veda 2oy (a. Poweichows, Tah-shan 7. Chun 6977 (A,
US); Y. Tsiang 6977 (NY). Yunnan: G. Forrest 15947 (1soryrE of
Ilex shweliensts, A).

The characters Chun and Handel-Mazzetti used to set. off /lex latifrons
and its varieties and /lex kwangtungensis var. pilosior fall within the range
of variation exhibited by the original material of /lex kwangtungensis. It
is almost identical with To, Tsang & Tsang (CCC) 12383, a paratype of
that species. Wang 33520 from Hainan fits the description of /lex latifrons
Chun very well, it being but a large-leaved form of /lex kwangtungensis.
With abundant material I can find no tangible characters to differentiate
Chun’s species and Handel-Mazzetti’s varieties from /lex kwangtungensis.
The curly indumentum of the branchlets, inflorescences, and vegetative
characters of Forrest 15947 conform to Ilex kwangtungensis; its pyrenes,
however, are smaller, and the canals are shallower.

Ilex kwangtungensis is related to flex chinensis Sims, but the latter has
very glabrous branchlets and inflorescences, smaller pyrenes with narrow
and shallow dorsal canals.

The pink staminate flowers and the red pistillate flowers of Jlex
kwangtungensis appear in May and June. The fruits turn greenish yellow
in August—October and become red at maturity in November.

23. Ilex atrata W. W. Smith in Notes Bot. Gard. Edinb. 10: 40. 1917;
Anon. in Notes Bot. Gard. Edinb. 17: 167. 1930; Comber in op.
cit. 18: 40. 1933.

An evergreen tree with large terminal buds, ciliate bud-scales, large

chartaceous or thin-coriaceous, finely serrate leaves, flavous-tomentose

fee

1949 | HU, THE GENUS ILEX IN CHINA 305

midribs, simple cymose infructescences, and convex pyrenes, slightly
impressed at one end on the

Branchlets subterete, castaneous-nigrescent; second year’s growth 3—4.5
mm. in diameter, when dry plicate-sulcate, the lenticels rather numerous,
obscure or evident, the leaf-scars slightly elevated, crescent-shaped; cur-
rent year’s growth 3-4 mm. in diameter, angular, glabrous, the terminal
buds large (especially when unfolding in October), the bud-scales ovate,
the outer ones 8 mm. long, 5 mm. wide, slightly keeled, their margins entire
and densely ciliate. Leaves 5-10 mm. apart; stipules lacking; petioles
15-25 mm. long, one-eighth to one-fifth the length of the lamina, 3 mm.
wide, plane or slightly canaliculate above, plicate-rugose beneath, glabrous;
lamina chartaceous or thin-coriaceous, castaneous or brunneous, elliptic,
12-16 cm. long, 3.5—-5 cm. wide; base rounded, obtuse or broadly cuneate;
apex acuminate, the acumen 10-12 mm. long with margins finely serrate
or crenate-serrate; midrib prominent, plane, glabrous above, much elevated
and flavous-tomentose beneath, sometimes glabrescent, the lateral nerves
14-18 pairs, prominent on both surfaces, the reticulation dense, evident on
both surfaces. Infructescence a simple cyme, axillary to the leaves of the
current year’s growth, 1—3-flowered, glabrous; peduncles 2-3 mm. long,
the pedicels 5-8 mm. long; persistent calyx explanate, 4 mm. in diameter,
the lobes ovate, minutely and very shortly ciliate. Fruit globose, 6 mm.
in diameter, al when dry brunneous-nigrescent, shiny, wrinkled, the
stigma esau 4- or 5-lobed. Pyrenes 5, oblong in outline, trigonous in
cross-section, 5.5 mm. long, 2.5 mm. aide at the back, smooth, estriate,
esulcate, the dorsal surface slightly impressed at the basal end, the endo-
carp coriaceous.

CHINA: Yunnan: Teng-yueh, G. Forrest 9419 (1soTyPE, A).

UPPER BURMA: Nam Tamai Valley, Kingdon Ward 13228 (B).

In the leaf-form, size, and texture, and in the simple cymose infructes-
cences, /lex atrata closely resembles Ilex kwangtungensis Merr., which differs
in having pubescent inflorescences, large drupes with deeply iniconaliculate
pyrenes. When Smith described the species he mentioned the terminal
bud as 2 cm. long. The terminal bud of the isotype in the Arnold
Arboretum is in the unfolding stage. It is 2 cm: long, and the bud scales
are loose. I think that the ciliate bud scales are a more significant char-
acter than the size of the bud. Although various collectors have visited
the classical locality since this species was described, they have apparently
not rediscovered the plant
23a. Ilex atrata var. wangul, var. nov.

Arbor; ramulis glabris; foliis chartaceis, ellipticis, 15 cm. longis, 4.5—
5.5 cm. latis; costa Hine elevata, tomentosa; inflorescentiis cymosis,
puberulentibus; pedunculis 10 mm. longis; pedicellis 5 mm. longis;
calycibus puberulentibus, ciliatis; pyrenis 7, dorso planis vel canaliculatis.

CHINA: Yunnan: Chiu-kiang, West of Champutung, C. W. Wang
67359 (TYPE, A).

306 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

This variety differs from the species chiefly in having the peduncles
longer than the pedicels, both being puberulent, but the single collection
seen is not fully mature. The infructescence is of the same type as that
of Ilex manneiensis 5. Y. Hu, but because of the serrate margins and
tomentose midrib I place it as a variety of [lex atrata.

Series 2, UMBELLIFORMES (Logs.), stat. Nov.

Ilex subgen. Euilex Loes. ser. Lioprinus Loes. sect. 1. Excelsae Loes.
subsect. a, Umbelliformes Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 95 (Monog. Aquif. 1: 95). 1901.

Evergreen trees; leaves subcoriaceous or chartaceous, entire; inflores-
cences cymose or subumbelliform, solitary, axillary in current year’s
growth; flowers 4—10-merous; petals distinctly one-fourth to one-sixth
connate; stamens equal to or slightly longer than the petals; rudimentary
ovary rostellate, the beak 3—7-cleft at the apex; fruits medium- to small-
sized; pyrenes 4—10, usually 5 or 6, 3-striate, 2-sulcate or rarely esulcate,
the endocarp coriaceous, sublignified or woody

KEY TO THE SPECIES

A. Pyrenes smooth or 3-striate, esulcate; endocarp coriaceous. (Himalaya,
Waa, IWant) oe a eades cwesaeneteccnwietesaieses 4. I. excelsa.
AA. Pyrenes 3-striate and sulcate : endocarp sublignified or woody.

B. Staminate inflorescence a loose cyme, the peduncle 3-13. mm.
long; fruit ellipsoid or globose; calyx erose eciliate (very rarely cili-
ate in /. rotunda var. nucrocarpa).

C. Leaves (4-)5-8(-9) cm. long, 2-4 cm. wide, the petioles 1-2
cm, long, one-fifth to one-fourth the length of the lamina; rudi-
mentary ovary rostellate. (Japan, East and oe China)...
Sie adr a yo Oe as Gok on ae eee dare eee I. rotunda.

CC. Leaves (2-)3(-4.5) cm. long, 1-2 cm. wide, i petioles 4-6

mm. long, one-tenth to one-eighth the length of the lamina.

NAD) c.ceietes teeter epersiatansaesaco. Te. Cima,

BB. Staminate inflorescence an umbel, the peduncle 14-20 mm. long;

fruit pp adtas globose or globose, when dry ochraceous or brun-
neous; calyx c

Cc Seen pubescent; fruit ca. 3.5 mm. in diameter; pyrenes
5 or 6, ca. 2mm. long. (Himalayan region to Hainan) eee

aT

ee ee ee eee ee eo ee ea
icc, i any glabrous; fruit ca. 6 mm. in diameter; pyrenes 6-10,
2.54 mm. long. Caimalaya) Renesas 28. . Wbeuee.

NS
si

Tlex excelsa te Hook. f. Fl. Brit. Ind. 1: 603. 1875; Maxim.
in Mém. Acad. Sci. St. Pétersb. VII, 29 (3): 23. 1881; Loes. in
Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78: 104 (Monog.
Aquif. 1: 104). 1901; Chung in Mem. Sci. Soc. China 1: 140. 1924;
Comber in Notes Bot. Gard. Edinb. 18: 46. 1933.

Cassine excelsa Wall. in Roxb. Fl. Ind. 2: 376. 1824; Spreng. Syst. 4:
Cur. Post. 124. 1827.

1949] HU, THE GENUS ILEX IN CHINA 307

Ilex elliptica D. Don, Prodr. Fl. Nep. 189. 1825; DC. Prodr. 2: 15. 1825.
Ilex rotunda sensu D. Don, l.c., non Thunb.

Ilex nepalensis Spreng. Syst. 4: Cur. Post. 48. 1827.
Ilex exsulca Wall. List no. 4328. 1830; Brandis, For. Fl. Brit. Ind. 76.
1874.

An evergreen tree up to 10 m. high with chartaceous or subcoriaceous
elliptic entire leaves, solitary cymose inflorescences, small ovoid-ellipsoid
fruits and 5, rarely 4 or 6, smooth or striate, esulcate pyrenes.

Branchlets rugose and angular; third year’s growth 4 mm. in diameter,
cinereous, with few lenticels; second year’s growth 3 mm. in diameter,
plicate-rugose; current year’s growth 2 mm. in diameter, ridged, glabrous,
the terminal buds poorly developed, with loose glabrous scales. Leaves on
current year’s growth only, 3-10 mm. apart; stipules lanceolate, 1-1.5
mm. long, acute; petioles slender, 10-12 mm. long, about one-seventh the
length of the lamina, glabrous, narrowly canaliculate above; lamina char-
taceous or subcoriaceous, brunneous-olivaceous, opaque on both surfaces,
elliptic or oblong-elliptic, 5-10 cm. long, 2-3.5, rarely 5 cm. wide; base
cuneate, obtuse or rarely rounded; apex acuminate, the acumen 5—12 mm.
long, mucronate; margin entire; midrib impressed above, elevated beneath,
glabrous, the lateral nerves 7 or 8 pairs, evident on both surfaces, anas-
tomosing near the margin, the reticulations of the veinlets obscure.
Inflorescences cymose, solitary, cymes 3-7 (-15) -flowered, the bracts
deltoid, acute, the flowers 4-6-merous. Staminate inflorescences: peduncles
4-8 mm. long, hirsute, the pedicels 2-5 mm. long, puberulous with 1 or
2 basal prophylla; calyx patelliform, deeply 4- or 5-lobed, the lobes
rounded, eciliate; corolla rotate, 5 mm. across, the petals oblong, 2 mm.
long, eciliate, one-sixth connate at the base; stamens equal or slightly
longer than the petals, the anthers oblong, 1 mm. long; rudimentary ovary
pulvinate, long-rostellate, the beak evident, 1 mm. long, with distinctly
5-cleft apex. Pistillate inflorescences: peduncles 5—12 mm. long, minutely
puberulous, the pedicels 3-4 mm. long; calyx 2.5 mm. across, 5- or 6-lobed,
the lobes deltoid, eciliate; corolla rotate, 5 mm. across; the petals broad-
ovate, 2 mm. long; staminodes one-half the length of the petals, the
anthers sagittate; ovary ovoid, 1.5 mm. in diameter, the stigma discoid.
Fruit ovoid-ellipsoid, 5 mm. in diameter, the persistent calyx explanate,
2.5 mm. across, the stigma thickly discoid, convex. Pyrenes usually 5,
occasionally 4 or 6, elliptic in outline, trigonous in cross-section, 2.75 mm.
long, 1.25 mm. wide, smooth or 2- or 3-striate, esulcate, the endocarp
coriaceous.

CHINA: Kwangsi: Yung-hsien, Steward & Cheo 757 (A, NY),
838 (A, NY). Yunnan: Shweli-Salween Divide, G. Forrest 15865 (A),
16062 (A); Mengtze, A. Henry 13691 (A); Lung-ling-hsien, H. T. Psat
55633 (A); without precise locality, Wang, Chang & Liu 85048 (A);
Mien-ning, T. T. Yu 17829 (A).

NEPAL: Wallich 4328 (1sotypr, G; fragment, A), 4328A (A).

INDIA: Himalayan region, Griffith 2008 (A, G), Thomson (G);

308 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XXX

Kangra, Parker 3317 (A); Soca Strachey & Winterbottom I (G);
Khasia Mountains, Hooker & Thomson (G).

The chartaceous elliptic acuminate entire leaves, the cymose inflores-
cences, and the rostellate rudimentary ovary of Jlex excelsa indicate a close
relationship between this species and //ex rotunda Thunb. of Japan and
eastern China. The latter usually has larger fruits and striate-sulcate
pyrenes. The variations in the size of the fruits of the Japanese species
and in the striation of the pyrenes of the Chinese material are so great that
it is difficult to decide to which species specimens collected from areas like
Kwangsi, which are geographically intermediate between the two, should
belong.

24a. Ilex excelsa var. hypotricha (Loes.), comb. nov.
flex hypotricha Loes. in Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur.
78: 103 (Monog. Aquif. 1: 103). 1901,

Branchlets puberulous; leaves chartaceous, puberulous on the lower
surfaces or only along the midrib benc.xth, oblong-elliptic, 5-8 cm. long,
2.5-4 cm. wide, shortly acuminate; inoue ui Sa: the
flowers 4- or 5-merous; fruit small, ellipsoid, ca. 5 mm. long, 3 mm. wide;
pyrenes 5, 4 mm. long, 1.25 mm. wide, 3-striate, ren the ane
coriaceous.

CHINA: Yunnan: without precise locality, G. Forrest 17938 (A),
18014 (A), 27200 (tTypE, A); Ping-pien-hsien, H. T. Tsai 60502 (A

NDIA: East Bengal, Griffith 2008 (A, fragment; G, 1sotyrE of Ilex
hypotricha).

This variety differs from typical /lex excelsa in its puberulent branchlets
and lower leaf-surfaces. The Yunnan form is a shrub or small tree

25. ex rotunda Thunb. Fl. Jap. 77. 1784; Willd. Sp. Pl. 1 (2): 711.
1797; Pers. Syst. Veg. 174. 1797, et Syn. Pl. 1: 151. 1805; Poir. in
Lam. Encycl. Suppl. 3: 67, 1813; Roem. & Schult. Syst. 3: 492.
1818; DC. Prodr. 2: 16. 1825; Spreng. Syst. 1: 496. 1825; Dietr.
Syn. Pl. 1: 555. 1839; Sieb. & Zucc. in Abh. Bay. Ak. Wiss. IV,
2: 149. 1845; Mig. Ann. Mus. Bot. Lugd.-Bat. 3: 106 (Prol. FI.
Jap. 106). 1867; Franch. & Sav. Enum. Pl. Jap. 1: 77. 1873;
Maxim. in Mém. Acad. Sci. St. Pétersb. VII, 29(3): 23. 36, pi. 7
fig. 5. 1881; Hance in Jour. Bot. 21: 296. 1883; Forbes & Hemel:
in Jour. Lin. Soc. Bot. 23: 118. 1886; Loes. in Nov. Act. Acad. Caes.
Leop.-Carol. Nat. Cur. 78: 106 (Monog. Aquif. 1: 106). 1901;
Hayata, Ic. Pl. Form. 1: 134. 1911; Lecomte, Fl. Gén. Indo-Chine
IT: 851. 1912; Chung in Mem. Sci. Soc. China 1: 141. 1924;
Kanehira, Form. Trees 386, fig. 337. 1936; Kia, Pl. Sin. Ill. 492,
fig. 833. 1937; Chen, Ill. Man. Chin. Trees 656, fig. 547. 1937;
Masamune Fl. Kainant. [Hainan] 175. 1943; Tardieu-Blot in FI.
Gen. Indo-Chine Suppl. 1: 767. 1948.

Ilex laevigata Blume ex Mig. Cat. Mus. Bot. Lugd.-Bat. 167. 1870.
llex kosunensis Yamamoto, Suppl. Icon. Pl. Formos. 1: 36, fig. 16. 1925.
Syn. nov.

1949 ] HU, THE GENUS ILEX IN CHINA 309

Ilex sasaku Yamamoto, op. cit. 39, fig. 19. 1925. Syn. now.

A large, entirely glabrous evergreen tree reaching 20 m. in height and
developing a trunk as much as 1 m. thick, with entire leaves, loose simple
umbelliform cymes, ellipsoid fruit, 3-striate and 2-sulcate lanceolate
pyrenes.

Branches straight, rough, with slightly elevated obovate or deltoid leaf-
scars; third year’s growth 4-5 mm. in diameter, cinereous, obscurely
rimulose, the lenticels inconspicuous; current year’s growth 2—3 mm. thick,
fuscous, longitudinally ridged, glabrous. Leaves occurring on current
year’s growth only, 1-10 mm. apart; stipules minute, 0.5-1 mm. long,
subulate, deciduous; petioles 10-20 mm. long, one-sixth to one-fourth as
long as the lamina, glabrous, canaliculate above, striate and rugose
beneath; lamina thin-coriaceous or chartaceous, olivaceous or brunnescent,
entirely glabrous, ovate to obovate or allipies4 —9 cm. long, 1.8—4 cm.
wide; base cuneate or obtuse; apex shortly pcminate. the acumen
deltoid, 5-8 mm. long, the tip cuspidate; margin entire, narrowly recurved;
midrib slightly impressed above, elevated beneath, especially toward the
basal half, the lateral nerves 6—9 pairs, obscure above, visible beneath,
reticulate near the margin, the tertiary veins inconspicuous. Inflorescences
cymose or umbelliform, cymes (3—) 4-6 (—13) -flowered, solitary, axillary,
on current year’s growth only. Staminate inflorescences: peduncles 3-10
mm. long, the pedicels 4-5 mm. long with 0-2 brown deltoid basal acute
prophylla; flowers 4-merous; calyx patelliform, 2 mm. across, shallowly
4-lobed, the lobes deltoid, eciliate; corolla rotate and reflexed, 6 mm.
across, the petals oblong, 2.5 mm. long, 1.5 mm. wide, one-sixth connate

base; stamens longer than the petals, the anthers ellipsoid; rudi-
mentary ovary pulvinate, 1 mm. in diameter, the middle rostellate, with
the beak 5- or 6-cleft at the apex. Pistillate inflorescences: peduncles
9-13 mm. long, 3—7-flowered, glabrous, the pedicels 4-8 mm. long, glab-
rous; flowers white, 5-7-merous; calyx subpatelliform, 2 mm. in diameter,
glabrous, the lobes deltoid, erose, 0.5 mm. long, 0.75 mm. wide; corolla
rotate, the petals one-fifth connate, the lobes obovate-oblong, 2 mm. long,

5 wide; staminodes three-fourths the length of the petals, the
filaments much dilated at the base, the sterile anthers cordiform; ovary
ovoid-conic, 2 mm. long, the stigma subglobose. Fruits ellipsoid) 6-8
mm. long, the stigma capitate when young, discoid and 5-—7-lobed at
maturity. Pyrenes 5—7, lanceolate in outline, trigonous in cross-section,
ca. 6 mm. long, 1.25 mm. wide, 3-striate, 2-sulcate on the dorsal enices:
the sides smooth, the endocarp sublignified.

CHINA: Chekiang: Sia-chu, R. C. Ching 1796 (A, US); Chua-
an-hsien, Y. L. Keng 736 (A). Hunan: Yi-chang, W. T. Tsang 23724

US). Fukien: Hing-hwa, H. H. Chung 9735 (A); Ku-liang, 4. H.
Chung 6030 (A), Kwangtume:. Ren-hwa, WwW. ¥. Chun 5616 (A),

A); Lo- chang, Y. Tsiang 1423 (A); Lok-chang, C. L. Tse 20945
(NY). Kwangsi: Seh-feng-dar- ae eG Cig /sls ACN. WS);
Shang-sze, Shap-man-taai-shan, W. T. Tsang 22102 (A). Hainan:
Mune -lae NV. he Chi Oo Coe aso nee CAs ONY ) --Chine-imai,-¢..L. Ler

310 JOURNAL OF THE ARNOLD ARBORETUM [VOL. xxx
488 (A, NY); Taam-chau, Tsang & Fung 440 (NY); W. T. Tsang 497
(A); without precise locality, H. Y. Liang 64514 (NY, US); L. Tang 410

A); C. Wang 34337 (NY). Taiwan: South Cape, A. Henry 929
(NY), 1973 (US); Takao Province, VY. Kudo in 1928 (A); without precise
locality, N. Fukuyama 7317 (TU), E. Matuda 1212 (TU), Nakamura 3952
(TU) ; Tathoku, S. Sasaki in 1929 (A, TU); Koshu Province, EF. H. Wilson
11042 (A).

APAN: K. Beattie & VY. Kurihara 10452 (A); Buerger (Peta:
Faurie 3844 (A); Goto 4803 (A); Maximowicz (@). m 1863" (NY) 3 :
Miyabe (A); R. Oldham 142 oe) 143 oe in 1862 (1soryPeE of Ilex
gata, NY); Siebold (A): Sargent (A): T. Tanaka ee (A);

Wilson 6157 (A), 6209 (A), 7328 (A), re, (A
U. Faurie 3805 (A); Yokohama Nursery Co. (A NY:

KOREA: U. Faurie 497 (A), 1633 (A, B), 1636 (A, B), 1640 (A):
T. Taquet 144 (A), 2720 (A), 3922 (A), 3923 (A), 4462 (A); E. H. Wilson
9556 (A).

INDO-CHINA: Tonkin: A. Pételot 889 (US), 1200 (A): W.T. Tsang
27364 (A), 29989 (A),

Ilex rotunda was first described from Japan and stated to be an entirely
glabrous plant. This typical form has been collected from a few coastal
provinces in China and Indo-China. In China the plant becomes a tree
and develops a trunk up to 30 cm. in diameter. The white fragrant flowers
appear in April, and the red fruit lasts until February or March of the
following year.

u-kiu Is] ee

25a. Ilex rotunda var. microcarpa (Lindl. ex Paxt.), comb. nov.

Ilex microcarpa Lindl. ex Paxt. Fl. Gard. 1: 43. 1850; Loud. Encycl. Pl.
suppl. 2: 1502. 1855,

Ilex rotunda sensu Forbes & Hemsl. Jour. Linn. Soc. Bot. 23: 118. 1886:
sensu Dunn & Tutcher in Kew Bull. Add. Ser. 10: 59. 1912: sensu
Loes. in Sarg. Pl. Wils. 1: 78. 1911; sensu Chien in Contr. Biol. Sci.
Soc. China 3: 58. 1927; sensu Rehd. in Jour. Arnold Arb. 7: 156. 1927:
sensu Groff in Lingnan Sci. Bull. 2: 64. 1930; sensu Bain. al in Mus.
Heud. Not. Bot. Chin. 2: 22, 1933, non Thunb.

An evergreen tree with chartaceous entire leaves; lamina oblong-elliptic,
rarely ovate, 5-9 cm. long, 2.5—4 cm. wide, the base obtuse, rarely rounded
or cuneate, the apex shortly acuminate and cuspidate ; inflorescences sub-
umbelliform 3—13-flowered; the peduncles 5-12 mm. long, the pedicels
2-6 mm. long, both minutely puberulous; flowers 4—7, usually 6-merous,
the calyx erose, rarely sparsely ciliate; fruit globose, rarely ovoid or Aiea:
soid, 5 mm, in Comeas very rarely 7 mm. in diameter; pyrenes 6, rarely
5 or 7, elliptic in outline, in cross-section trigonous, 4 mm., rarely 5-6 mm.
long, the back 1 mm. wide, 3-striate, 2-sulcate, the sides smooth, the endo-
carp sublignified.

CHINA: Kiangsu: I-shing, Ching & Tso 621 (A). Anhwei: Wang-
shan, W. C, Cheng 3885 (US); R. C. Ching 2914 (A, US); Chimen, RC,
Ching 3157 (A), 3329 (A). Chekiang: Yen-tang-shan, C. Y. Chiao
(ex Herb. Univ. ne no, 14741) (A, US); Hangchow, C. Y. Chiao (ex
Herb, Univ, Nanking no. 18814) (NY, US); Sia-chu, R. C. Ching 1688

1949] HU, THE GENUS ILEX IN CHINA 311

(A, US), 2566 (A, US); Tsing-tien, Y. L. Keng 204 (A); Nin-hai-hsien,
Y.L. Keng 1099 (A); Ning-po, D. MacGregor (A). Kiangsi: Ku-ling,
W.Y. Chun 4301 (A); E. H. Wilson 1611 (A); Shang-yu-hsien, H. H. Hu
912 (A); Ta-yu-hsien, H. H. Hu 977 (A); Kien-nan-hsien, S, K. Lau
SOOse An S))s Peart: Chang-ning-hsien, C. S. Fan & Y. Y. Li 421

ae aed : Fang-chin-shan, Gace Chiao & Cheo 841 (A,

Se wikie tt: n-how-hsien, Tang Siu-ging 6786 (A); H. H. Chung
2080 (A), 6617 (A, a 6764 (A), 6868 (A); J.B. S. Norton 1573 (US a
Kwangtung: Shao-chow, W. Y. Chun 5507 (A); Ta-ching, W.
Chun 5510 (A): Ma-hang, W. Y. Chun 5547 (A); Ren-hwa, W. y, ae
oe CAN 20S 6A): Seay kiang, W. VY. Chun 6115 (A); Ting- wu-shan,

WG lineOneU. (CAmblone OC. OC," Leume (CCC) <7 Cheri) 9 5000 2k):
eae Kan Geese ie 50419 (NY). Wung-yuen-hsien, S. K. Lau 656 (A,
NY US). 0924CAe NY). 2220°CA, US): Kao-yvaowS. ye Caucd2s7 (NY):
@anton Gur Os eeoimen( CCC) 7 (A, US), 056 CAs US )2 120546), 1299
CASEU'S))= sL690 4 GU oer éion (Ay. US): 7830: CA WS) a tO CASS. 2E77
Cay) Minshsien, «GO. Lene (CCC) 3360 (A, Us)e EunscmausMt., 2o,
Tsang & Tsang (CCC) 12206 (US); Po-tau-chai, F. A. McClure 232 (A);
Sam-shu, PF. A. McClure 1348 (US); Loh-fau Mt., Fung Hom 556 (NY);
E. D. Merrill 10249 (A), 10891 (A); Sin-fung-hsien, Y. W. Taam 280
(A), 584 (A), 771 (A), 1049 (A); Chung-shaan, W. T. Tsang 2 (NY);
Ta-pu-hsien, W. T. Tsang 21190 (A, NY); Hwei-yang-hsien, WV. T. Tsang
25732 (A); T. M. Tsui 102 (US); Jen-hwa-hsien, W. T. Tsang 26437
(A); Ting-wu-shan, Y. Tsiang 784 (A), 1495 (A); Ko-chow, Y. Tstang
912 (A); Lo-fou-shan; Y. Tsiang 1679 (A); Sun-wui-hsien, Y. Tstang &
Eso 2017 (CA)a Ho-yuens CL. Tso. 21553 CN ¥ 277504 (Na) Dune-ehun;
CEL Eso-2i0os CN) keo-chone. Cx Lo iso-2043) CN Ne 005e. CNY).
20848 (NY); Yang-shan-hsien, 7. M. Tsui 514 (NY, US), 627 (A, NY,
US); Chu-don, C. Wang 472 (A). Kwangsi: Lu-chen: RF. C. Ching
5221 (NY); Shang-sze: W. T. Tsang 24503 (NY). Hongkong: N.C.
Chun me CNY )G2C 2hord (NY): VY. tstang- 2979. (NSO) Elainan

K. Chun & C. L. Tso 43827 (NY); Yai-chow, F. C. How
70446 ae oe Po-ting Cc, How /1509 CA). f1/37 CA) f2000 CA) :
Lok-tung, S. K. Lau 26757 (A); Bak-sa, S. K. Lau 26296 (A); Ching-mai,
C. L. Lei 1 (NY, US); Seven Finger Mts., H. Y. Liang 61775 (A, NY);
No-dao, F. A. McClure ee ne (A); without precise locality, J. L.
Gressitt 838 (A), 1119 (B); Ch’ang-kiang, S. K. Lau 1899 (A, NY);
Taam-chau, W. T. Tsang 70 a US); 367 “CNY US) eae ONY Us);
C. Wang 34280 (A, NY), 35117 (A, NY), 35727 (NY). Taiwan: Tai-
hoku-syu, K. Odashima (ex Herb. Taihoku Imp. Univ. no. 17750) (A, NY,
Wis eeS= Stucke (cA) «lake pees R. Kanehira 21336 (A); E. H. Wil-
son 11121 CRS ee OO. U,

The specimens from Fukien a rather small inflorescences. They
even suggest Jlex stewardii S. Y. Hu, but for the most part they are poorly
selected and are not to be considered as typical. The variety differs from
typical /lex rotunda in having puberulous peduncles and pedicels. In
general its fruits are smaller and its pyrenes are shorter and less sulcate.

The cultivated plants that Lindley saw in Standish and Noble’s Bagshot
Nursery were introduced by R. Fortune during his second trip to China,
1848-50. On this trip he collected in Shanghai and in the vicinity of

312 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Ning-po. He also proceeded from Ning-po across the province of Chekiang
to southern Anhwei. He states that. his objective was to discover the
secret of manufacturing tea. On this trip he was handicapped by fear
and suspicion. for this reason, he probably could not collect many speci-
mens from the interior parts of Anhwei and Chekiang. It is most likely
that his original material of /lex microcarpa came from the vicinity of
Ning-po, where it is common.

In fruit-size, leaf-texture, and the indumentum of the inflorescences,
the variety resembles //ex excelsa (Wall.) Hook. f. more closely than
lex rotunda. Because of the striate-sulcate pyrenes I place it as a variety
of flex rotunda Thunb. As I interpret the group, /lex rotunda var. micro-
carpa is a variant intermediate between the Japanese /lex rotunda and the
Himalayan /lex excelsa.

26. Ilex angulata Merr. & Chun in Sunyats. 2: 266. 1935.

An evergreen shrub or small tree up to 4 m. high with striate puberulent
twigs, entire chartaceous elliptic-lanceolate leaves, solitary or cymose
infructescences, large ellipsoid fruits and_ striate-sulcate ae with
rough excavated backs.

Branchlets slender, zigzag; fourth year’s growth usually 3-4 mm. in
diameter, angularly ridged, rimose, the lenticels lacking, the leaf-scars
semi-orbicular; third and second year’s growth thinner, puberulent; cur-
rent year’s growth very slender, 1-1.5 mm. in diameter, ridged and nar-
rowly canaliculate, pubescent in the grooves, the terminal buds lacking.
Leaves occurring even on the third year’s growth, 2-8 mm. apart; stipules
minute, subulate and falcate, persistent; petioles short, 4-6 mm. long,
one-tenth to one-eighth the length of the lamina, canaliculate above;
Jamina chartaceous or membranaceous, olivaceous, broad-elliptic, elliptic
or lanceolate, 2.5-4.5 cm. long, 1—2.3 cm. wide; cuneate or acute at the
base; acuminate and cuspidate at the apex, the acumen 5-8 mm. long;
margin entire, very rarely near the apex inconspicuously but sharly
serrulate; midrib sulcate and puberulent above, prominent, elevated and
glabrous beneath, the lateral nerves 5—7 pairs, rather obscure on both
surfaces. Inflorescences solitary, axillary, only on current year’s growth,
pubescent; cymes simple, 1—3-flowered, when 1-flowered the pedicels
short, 3-5 mm. long, with 2 sub-basal srophy ila, when 2- or 3-flowered the
peduncles 3-5 mm. long, the pedicel subequal; bracts to individual flowers
deltoid, sparsely pubescent. Staminate inflorescences: cymes usually
3-flowered, calyx patelliform, 3-5 mm. across, deeply 5-lobed, the lobes
membranaceous, suborbicular, 1.5 mm. long and wide, eciliate, eee
or obtuse; corolla rotate, 6— 8 mm. across, the petals ovate, 3 mm. lon
2—2.5 mm. wide, one-sixth connate at the base; stamens ies: fourths o
length of the petals, the anthers oblong; rudimentary ovary globose, 1 mm.
in diameter. Pistillate inflorescences: calyx and corolla as in the staminate
flowers; staminodes one-third the length of the petals, the sterile anthers
sagittate; ovary ovoid, the stigma mammiform, discoid at the apex, 2.5
mm. long, 2 mm. wide at the base; rudimentary ovary globose, 1 mm. in

1949 ] HU, THE GENUS ILEX IN CHINA 313

diameter. Fruit ellipsoid, 6-7 mm. long, 5—6 mm. in diameter, striate and
sulcate, dark brown, the persistent calyx explanate, 5 mm. in diameter, the
style evident, the stigma discoid, irregularly lobed. Pyrenes 5 or 6, 5 mm.
long, 1.5 mm. wide, the dorsal surface 3-striate and sulcate, the middle
ridge usually deeply impressed, the endocarp lignescent.

CHINA> Hiatian:- ‘Po-ting, &. C. How 71993, A) > Dung-ka, W.-K.
Chun & C. L. Tso, 43779 (1sotypr, A; TYPE, NY), 43781 (A, NY); without
precise locality, C. Wang 34675 (NY, US).

This species is endemic to Hainan Island. There it grows at aititudes
of 450-500 m. Its pink staminate flowers appear in April (ex How).
26a. Ilex angulata var. longipedunculata, var. nov.

Frutex pubescens; foliis late ellipticis, 2.5-4.5 cm. longis, 1.5—2.3 cm.
latis, apice acuminatis; inflorescentiis cymosis, pedunculis 8-10 mm.
longis, pedicellis Reenileneis pedicellis fructibus solitariis 15-18 mm.
longis, pubescentibus; fructibus ellipsoideis, 9-10 mm. longis, 6 mm.
diametro; pyrenis 5 mm. longis, striatis, endocarpio sublignescento.

CHINA: Hainan: Potting, F. C. How 72497 (A), 73020 (tTypE, A).

This variety differs Pee the typical /lex angulata in having larger
fruiting pedicels and fruit

27. ex godajam ae Wall. List. no. 4329. 1839; Hook. f. Fl.
Brit. Ind. 1: 604. 1875; Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 101 (Monog. Aquif. 1: 101). 1901; Maxim. in Mém.
Acad. Sci. St. Pétersb. VIT, 29 (3): 23. 1881; Pitard in Lecomte,
Fl. Gén. Indo-Chine 1: 854. 1912; Tardieu-Blot in Fl. Gén. Indo-
Chine Suppl. 1: 767. 1948.

Prinos godajam Colebr. ex Wall. Pl. As. Rar. 3: 38, pl. 261. 1832.
Ilex godajam var. genuina Kurz in Jour. As. Soc. Bengal. 44(2):. 158.
1

“I

Ilex capitellata Pierre, Fl. For. Cochinch. 4: pl. 278b. 1893.

Ilex godajam var. capitellata (Pierre) Loes. op. cit. 102. 1901. Syn. nov.

Ilex rotunda Thunb. var. piligera Loes. op. cit. 108.

Tlex umbellulata sensu Merr. in Lingnan Sci. Jour. 6: 327. 1928, non Loes.

Ilex fabrilis sensu Merr. & Chun in Sunyats. 5: 106. 1940, non Pierre

Ilex rotunda sensu auct. plur., non Thun

An evergreen shrub or tree up to 8 m. high, with pubescent branchlets,
entire thin-coriaceous or chartaceous leaves, umbelliform inflorescences,
small globose fruit, and minute 3-striate and 2-sulcate pyrenes.

Branchlets crooked, the third year’s growth 5 mm. in diameter, subterete,
light cinereous, the lenticels lacking, the leaf-scars semi-orbicular, not
elevated; second year’s growth 4 mm. in diameter, striate; current year’s
growth 2 mm. in diameter, pubescent, brown, the terminal bud lacking or
very weak. Leaves occurring also on second year’s growth, 10-20 mm.
apart; stipules minute, broadly deltoid, acute; petioles very slender, 10-15
mm. long, one-ninth to one-fourth the length of the lamina, pubescent,
narrowly canaliculate above; lamina thin-coriaceous or chartaceous, oliva-
ceous, when young silky, especially on the lower surface, ovate or oblong,

314. JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

5-8 cm. long, 2.5—4.6 mm. wide; base rounded, seldom obtuse; apex
obtuse or very short-acuminate, the acumen 3-5 mm. long, broadly
deltoid; margin entire; midrib impressed, often puberulent above, strongly
elevated and pilose or glabrescent beneath, the lateral nerves 7—11 pairs,
straight and parallel, curved upward near the margin, the reticulation of
the veinlets obscure. Inflorescences umbelliform, axillary, on current
year’s growth, pilose, often when the shoot fails to develop appearing panic-
ulate; flowers 4-6-merous. Staminate inflorescences: umbels 8—23-
flowered; peduncle 14-18 mm. long, the secondary axes often evident,
1—2 mm. long, the pedicel 2-4 mm. long, with 2 subulate basal prophylla;
flowers 4- or 5-merous; calyx patelliform, 2.5 mm. across, pubescent,
deeply 4- or 5-lobed, the lobes ovate, obtuse, erose and ciliate; corolla
rotate, the petals oblong, ca. 2 mm. long, 1 mm. wide; one-eighth connate
at the base; stamens equaling the petals in length, the anthers ovoid, 0.5
mm. long; rudimentary ovary globose, 0.75 mm. in diameter, rostellate,
the beak 4- or 5-cleft. Pistillate inflorescences: umbels 3—13-flowered;
peduncles 10-13 mm. long, the bracts minute, deltoid, crowded and
imbricate; pedicels 2-5 mm. long, with deltoid acute basal prophylla;
calyx as in the staminate flowers; corolla and staminodes not seen; ovary
ovoid, 1.5 mm. in diameter, constricted below the capitate stigma. Fruit
globose, small, 3.5 mm. in diameter, the exocarp thin. Pyrenes 5 or 6,
small, 2 mm. long, 1.5 mm. wide at the back, 3-striate, 2-sulcate, the
endocarp woody.

CHINA: Hainan: Nodoa, W. Y. Chun 4633 (US); Yai-chow,
F.C. How 70672 (A), 70462 (NY, o) 70473 (NY), 71056 (A, NY);
F.C. How & N. K. Chun 70198 (A, NY, US); C. Wang 33205 (A, US);
Man-ning, F. C. How 71430 (A); —. hsien, Ota Lab Oy TAN Y US)
without precise locality, H. Y. Liang 63135 CNY US). 64052 CNY),
65364 (NY); Taam-chau, W. T. Tsang 0 (LU ee) (A, NY, US);
C. Wang 34295 (NY). Kwangsi: Tian-chen, R. C. Pai 7803 (A),
Pode UP: INV WS). 320 - ONY Ve Tae re en P. @ 29203. Yuwie
nan: Che-li-hsien, C. W. Wang 77582 (A).

INDO-CHINA: Tonkin: Harmand 1297 (1soryre of Jlex capitellata,

P); Du Pasquier 3026 (A); A. Pételot 6400

INDIA: £&. O. Shebbean 11776A (A); Assam, Prain’s Collector 765
(A), Reporter on Economie Products 71223 (A); J. O. Voigt 416 (A).

lex godajam was originally based on material from Silhet and Assam
and described as pubescent. Additional collections indicate that it has
a much wider distribution than the closely related glabrous species /lex
umbellulata (Wall.) Loes. It flowers about eae (from late March to
early May) and matures its red fruit in late Augus

Ilex godajam is closely related to [lex pe a (Wall.) Loes., and is
perhaps merely a pubescent variety of it. The two are alike in their
pseudo-umbelliform, long-pedunculate, pubescent inflorescences. However,
whereas /lex godajam is pubescent on branchlets, petioles, midribs, and
often young leaves, /lex umbellulata is glabrous. In general, Ilex godajam
also has smaller leaves and fruits, thinner exocarp, and less sulcate pyrenes.
It has another close relative in //ex rotunda Thunb. var. microcarpa (Lindl.

1949 ] HU, THE GENUS ILEX IN CHINA 315

& Paxt.) S. Y. Hu, which differs in having glabrous branchlets and elliptic
leaves with cuneate bases and long-acuminate apices.

28. Ilex umbellulata (Wall.) Loes. in Nov. Act. Acad. Caes. Leop.-Carol.

Nat. Cur. 78: 99 (Monog. Aquif. 1: 99). 1901; Pitard in Lecomte,
F]. Gén. Indo-Chine 1: 862. 1912; Tardieu-Blot in Fl. Gén. Indo-
Chine Suppl. 1: 767. 1948.

Ehretia umbellulata Wall. in Roxb. Fl. Ind. 2: 344. 1824; Wall. List no.
4329, 1829. 1832; Spreng. Syst. 4: Cur. Post. 66. 1827,

Ilex sulcata Wall. Tigt no. 4330, nom. nud.; Hook. f. Fl. Brit. Ind. 1:
604. 1875, descr.

Pseudehretia umbellulata (Wall.) Turez. in Bull. Soc. Nat. Mosc. 36:

1863.

Ilex godajam (Colebr. 7 Wall. var. sulcata (Wall.) Kurz in Jour. As. Soc.
Bengal 44(2): 158. 1875.
ee es var. megalophylla Loes. os cit. ca 272. 1908; Chung in
. Sei. Soc. China 1: 141. 1924. Syn.

A — evergreen tree up to 18 m. high, = trunk up to 40 cm. in
diameter, glabrous branchlets, large entire oblong leaves, axillary sub-
umbelliform fee globose small fruits with 6-10 3-striate, 1- or
2-sulcate pyren

Branchlets ae crooked, roughened by elevated scars of former flower-
ing shoots; third year’s growth 4 mm. thick, cinereous, longitudinally
plicate, the lenticels usually small, inconspicuous; current year’s growth
2—3 mm. in diameter, brown or castaneous, glabrous. Leaves occurring
even on the third year’s growth, 5-15 mm. apart; stipules subulate, falcate,
acute, 1.25 mm. long; petioles slender, 8-10 mm. long, one-fifteenth to
one-ninth as long as the lamina, glabrous, canaliculate above, rugose
beneath; lamina chartaceous, membranaceous when young, olivaceous,
opaque on both surfaces, oblong, 10-15 mm. long, 5—6.2 mm. wide; base
rounded or obtuse; apex obtuse or very shortly and abruptly acuminate, the
acumen 5 mm. long, broadly deltoid; margin entire, often undulate, when
dry very narrowly recurved; midrib impressed and glabrous above, strongly
elevated beneath, the lateral nerves in 9-12 pairs, evident above, prominent
beneath, curving upward and anastomosing near the margin, the reticula-
tion of the veinlets obscure. Inflorescences subumbelliform, axillary, at
the basal parts of the newly developed branches and often, when the
branches fail to develop, appearing paniculate. Staminate inflorescences:
individual pseudo-umbels 6—16-flowered, the bracts broad-deltoid, scaly,
1 mm. wide; peduncles 18-30 mm. long, puberulent, the pedicels 3-5 mm.
long, puberulent, with deltoid acute basal prophylla; flowers 4- or 5-merous;
calyx patelliform, puberulent, 2-3 mm. across, deeply 4- or 5-lobed, the
lobes semi-orbicular, 1 mm. long, 1 mm. wide at the base, pubescent and
ciliate; corolla rotate, the petals oblong, 2 mm. long, 1.5 mm. wide, one-
eighth connate at the base; stamens slightly longer than the petals, the
anthers ovoid, 0.75 mm. long: rudimentary ovary pulvinate, 1 mm. in
diameter, rostellate, the beak acute, 4- or 5-cleft at the apex. Pistillate
flowers not seen. Fruit globose or depressed-globose, 6 mm. in diameter,

316 JOURNAL OF THE ARNOLD ARBORETUM One Ras

red, very fleshy, rather smooth or slightly striate-sulcate, the exocarp
rather hard, brown, the stigma thickly discoid or cristate. Pyrenes 6-10,
trigonous or laterally much compressed, suborbicular in outline, 2.5—4 mm.
long, 1.5-2.5 mm. wide on the dorsal surface, 3-striate, 1- or 2-sulcate,
the endocarp woody.

CHINA: Yunnan: Szemao, A. Henry 11926 (A, K), 13456 (TypE of
Ilex umbellulata var. megalophylla, A; K); Lan-tsang-hsien, C. W’. Wang
73370 (A); Fo-hai, C. W. Wang 74938 (A); Che-li-hsien, C. W. Wang
75648 (A), 75956 (A), 75881 (A), 77805 (A), 77815 (A), 77854 (A),
78050 (A), 78765 (A), 79105 (A); Jenn-yeh-hsien, C. W. Wang 80554

(As

INDIA: Tenasserim & Andamans: J, W. Helfer ex Herb.
East India Company no. 1998 (G); Bengal: J. W. Helfer in 1836-38
(duplicates . material used by J. D. Hooker for the description of /lex
sulcata Wall., A); Chittagong Hill Tracts, King’s Collector in 1887 (419)
(A)

—

Ilex umbellulata was originally described from Silhet and Assam, India,
as a glabrous plant with pubescent umbelliform inflorescences. In China
it has been collected in mixed forests of southwestern Yunnan at an alti-
tude of 780-1350 m. There its white flowers appear in April. Its red
fruits mature in late August or September (ex Wang

Both of the Helfer specimens are immature, the owes not fully open.
The leaves of these two specimens are memibranacoais: as J. D. Hooker
described them, but they are not fully grown. The texture and size of the
leaves of King’s Collector 4/9 are the same as in Henry 13486, holotype
of Ilex umbellulata var. megalophylla Loes., as is the case with the other
specimens. I see no reason to maintain the Yunnan form as a large-leaved
variety.

Section VI. PALTORIA (Ruiz & Pavon) Maximowicz

Ilex sect. Paltoria (Ruiz & Pavon) Maxim. in Mém. Acad. Sci. St.
Pétersb. VII, 29 (3): 20, 21. 1881. Type species: Jlex Paltoria
Pers. = Paltoria ovalis Ruiz & Pavon = Ilex ovalis (Ruiz &
Pavon) Loes.

Paltoria Ruiz & Pavon, Fl. Peruv. Prodr, 2, pl. 33, fig. 1-9. 1797, et FI.
Peruv. 1: 54, pl. 84, on b. 1798. Type species: Paltoria ovalis Ruiz &
Pavon.

llex subgenus Eutlex Loes. series Paltoria (Ruiz & Pavon) Loes. in
Verhandl. Bot. Ver. Brandenb. 33: 26. 1891, et in Engler & Prantl, Nat.
Pflanzenfam, Nachtr. 218, 1897.

Ilex subgenus Aquifolium (Gray) Rehd. sect. Lioprinus (Loes.) Rehd.
Man. Cult. Trees Shrubs 343. 1927, et Bibliog. Cult. Trees Shrubs 401.
1949, pro parte.

Shrubs; low creeping or up to 5 m. high; leaves coriaceous or subcoria-
ceous, serrate or crenate, rarely subentire or entire, the lower surface
glandular-punctate or epunctate; staminate inflorescences fasciculate on
the second year’s growth or solitary on the current year’s growth, the

1949 ] HU, THE GENUS ILEX IN CHINA 317

pistillate flowers usually solitary, rarely 3-flowered cymose, axillary on
the current year’s growth (occasionally fasciculate in Jlex triflora) ; flowers
4-merous, rarely 5- or 6-merous, the petals of the pistillate flower one-fifth
connate at the base; fruit 6-9 mm. in diameter, with thin exocarp and
mesocarp; pyrenes 4 (rarely 5, or 6 in /lex sugeroki), large, smooth or
rarely slightly rugose on the dorsal surface, striate, esulcate, the endocarp
coriaceous.
KEY TO THE SERIES

Leaves punctate; pyrenes 4, ca. 4 mm. wide, striate with the striae usu-

Ally SUI celatliy aS GUILD beg ten eee nan a vas Sau eries | ees i
AA. Leaves not Ee: pyrenes 4 or sometimes 5 or 6, 2-3 mm. wide,
estriate or (in Ilex pedunculosa Miq.) longitudinal] y tmistriate on the
DAG ear ee et eR ei.) nd acy Series 2. Cassinoides.

Series 1. STIGMATOPHORAE (Logs.), stat. Nov.

Ilex subgen. III ELuilex Loes. ser. C. Aquifolium (Tournef.) Maxim., sect.
3. Microdontae Loes. subsect. c. Stigmatophorae Loes. in Engler &
Prantl, Nat. Pflanzenfam. Nachtr. 220. 1897, et in Nov. Act. Acad.
Caes. Leop.-Carol. Nat. Cur. 78: 343 (Monog. Aquif. 1: 343). 1901.

Puberulent shrubs or small trees with coriaceous or subcoriaceous,
crenate or serrate, punctate leaves, 1—7-flowered fasciculate, or solitary,
cymose, staminate inflorescences, uniflorous fasciculate or solitary pistillate
inflorescences, 4-merous (rarely 5—7-merous) flowers, depressed-globose
or subglobose black fruits, and 4 large smooth striate coriaceous pyrenes,
sometimes slightly ridged on the back.

The characters of various members of this section are rather uniform.
The chief differences seem to be in leaf-form and indumentum, which in
their case prove very unstable characters for taxonomic work. The section
includes two strongly contrasted elements, the northern, /lex crenata
Thunb., a low shrub with small obtuse or acute leaves, and the southern,
Ilex Winona Blume, a shrub or small tree with regret larger, shortly
acuminate leaves. Variants intermediate between the two occur in the

species or as varieties of either element. The distribution of all species of
this series is illustrated on the following map.

KEY TO THE SPECIES

A. Fruits solitary, axillary to leaves, rarely to scales on current year’s
growth; staminate inflorescences | —7-flowered, when cymose the pe-
duncles (up to 5 mm. long) longer than the pedicels; rudimentary
ovaries subglobose.
eaves shiny above, the midribs depressed; flowers 4-merous; calyx

5 mm. across, rounded and entire; stigma of fruit applanate, con-

spicuous, 2.5 mim. in diameter, rounded; pyrenes subrugose, slightly

ridged on the back. (Anhwei to Kwangtung)......2 9. TL. viridis.

Leaves usually opaque above, the midis plane or slightly im-

pressed; flowers 4-7-merous ; calyx 3 mm. across, erose or dentate ;

B

7

318 JOURNAL OF THE ARNOLD ARBORETUM [VOL, Xxx

stigmas of fruit slightly elevated, small, 1 mm. in diameter, dis-
tinctly 4-lobed; pyrenes smooth, striate, the striae slightly impressed.
C. Leaves small, usually 1-3 cm. long, the apex obtuse or acute,
never acuminate, lateral nerves 2 or 3 on each side of the
midrib; pedicel of fruit 4-5 mm. (rarely longer) long; flowers
4nerous. (Japan, cultivated in North China)............
eapedeew eae eue hs ay pak k ce nae ae Oe 30. J. crenata.
CC. Leaves 4-7 cm. long, the lateral nerves 6 or 7 on each side of
the midrib, the apex shortly acuminate; pedicels of peu as

mm. long; flowers 4-9-merous. (Central = West Chi
ee re are here ere Des Wh sen es
AA. Fruits usually fasciculate, axillary to leaves of second year’s growth,
rarely solitary on undeveloped current year’s branchlets; staminate in-
florescences 1-3 (rarely more)-flowered, when cymose the pedicels
equaling or longer than the peduncles, which are 1-2 mm. long; pyrenes
smooth, estriate, esulcate; rudimentary ovary pyramidal. (Java to
POUIGd ee ti. Soa bokeneaneer ees aaece eee o2. J. irtflora;
As far as our specimens indicate, the northern element, Jlex crenata
Thunb., does not reach Taiwan or even the Liu Kiu Islands. Yamamoto
in 1925 reported Jlex crenata Thunb. var. scoriatum W. W. Sm. [scoriarum |
from Taiwan. Koidzumi in 1929 thought that “The formosan plants
differs from SMITH’S species in having glabrous petioles, . . . fasciculate
female inflorescens, leaves not glanduloso-punctate on the under surfaces.”
He named this Liu Kiu species Ilex scoriatulum. I have not seen his
material, but in Gressitt 608, which I assume to represent the same plant,
the branchlets are puberulous, the leaves obovate, shortly acuminate or
acute, 3-4.5 cm. long, 1.5—2.5 cm. wide, the petioles 7-10 mm. long, one-
fourth the length of the lamina. The long petioles distinguish it from the
rest of the Stigmatophorae. Its punctate leaves and large pomiform-
subglobose fruits are like those of I/ex viridis Champ. ex Benth. If this
is what Koidzumi meant by /lex scoriatulum, then his species is a valid one.
Belonging to this section are three other species occurring beyond the
area covered by this work. They are Ilex thomsonii Hook. f. (1875) from
Himalaya, [lex luzonica Rolfe (1886) from the Philippine Islands, and
Ilex radicans Nakai (1930) from Japan. Merely on the basis of leaf size
and form Loesener in 1901 reduced Hooker’s species to a variety of Jlex
crenata Thunb. and Rolfe’s species to a form of the same species. Of
Ilex thomsonit Hook. f. I have seen three specimens cited by Loesener,
Griffith 2003 from East Bengal (G), Griffith 520 (G), and Hooker &
Thomson ex Herb. Ind. Or. no. G), also a small-leaved form from
Sumatra, Hamel 596 (NY, US), and Vaies 2008 (NY). The leaves of
these specimens are 1.5-3 cm. long, and vary from broadly elliptic to
ovate and even obovate, and the apices from acute to obtuse or even
round. The staminate inflorescences are fasciculate, 1-3-flowered. When
cymose, the peduncles are shorter than the pedicels. The fruits are
solitary only on the current year’s growth. Hooker & Thomson 9 is inter-
mediate between the Bengal and the Sumatra specimens. Since it shares
characters with both the northern and the southern elements in the section

1949] HU, THE GENUS ILEX IN CHINA 319

and has a different geographic distribution, I consider it better to retain
Ilex thomsontu Hook. f. as a distinct small: leaved species of the Indo-
Malayan region.

Concerning Jlex luzonica Rolfe, of which Eurya myrtilloides Elm. is a
synonym, I have examined the following specimens: Elmer 7784, 8810,
9105, 11464, 11670, and Merrill 1571. The tuberculate branchlets, the
retuse apex of the leaf, the 4—6-merous flowers, and the smooth estriate
pyrenes are characters strong enough to warrant retaining it as a distinct
species.

Ilex radicans Nakai (1930) is a creeping plant with leaves like those
of Ilex crenata Thunb., but larger. It occurs in Asia at the northernmost
limits not only of the section Stigmatophorae but also for the genus J/ex,
being known only from Yezo and Sachalin.

29. Ilex viridis Champ. ex Benth. in Hook. Jour. Bot. Kew Gard. Miscel.
4: 329. 1852; Walp. Ann. 4: 430. 1857; Benth. Fl. Hongk. 65.
1861; Maxim. in Mém. Acad. Sci. St. Pétersb. VII, 29 (3): 46.
1881; Forbes & Hemsl. in Jour. Linn. Soc. Bot. 23: 118. 1886;
Dunn & Tutcher in Kew. Bull. Add. Ser. 10: 59. 1912; McClure in
Lingnan Sci. Bull. 3: 25. 1931; Belval, Mus. Heud. Not. Bot. Chin.
2 2221939

Ilex triflora var. viridis (Champ.) Loes. in Nov. Act. Acad. Caes. Leop.-

Carol. Nat. Cur. 78: 345 (Monog. Aquif. 1: 345). 1901; Rehd. in Jour.
Se Arb. 8: 157. 1927; Chien in Contr. Biol. Sci. Soc. China 3:
58. 1927.

An evergreen shrub or small tree up to 5 m. high with ovate, obovate, or
broadly elliptic serrate conspicuously punctate leaves, solitary fruits, con-
spicuous stigmas 2.5 mm. in diameter, and sublignified pyrenes slightly
ridged on the back.

Branchlets subquadrangular; third year’s growth 3-5 mm. in diameter,
subterete, greenish, strongly ridged, the lenticels oblong, elliptic, forming
2 longitudinal lines between the ridges; second year’s growth ridged and
angular, the ridges glabrous, the furrows puberulous, the leaf-scars semi-
orbicular; current year’s growth 2 mm. in diameter, ridged and canalicu-
late, puberulous in the grooves; terminal buds narrowly conic, acute,
glabrous. Leaves occurring also on the second year’s growth, 2-8 mm
apart; stipules narrowly deltoid, 1 mm. long; petioles 4-5 mm. long, one-
thirteenth to one-sixth the length of the lamina, puberulous or glabrescent,
deeply canaliculate above, rugose underneath, narrowly winged by the
decurrent leaf-base; lamina thick-coriaceous, green, shiny above, dull and
conspicuously punctate beneath, ovate, obovate or elliptic, 2.5—7 cm. long,
1. cm. wide, the margins slightly recurved, especially toward the base,
crenate-serrate, the teeth 10-18 (usually 12-15) on each side, their tips
brunnescent; midrib deeply canaliculate, shortly and sparsely puberulous
above, elevated and glabrous beneath, the lateral nerves 5—7 pairs, obscure
on both surfaces. Staminate inflorescences: 1—5-flowered cymose, solitary,
axillary to the scales or lower leaves of the current year’s growth or

320 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

fasciculate on the second year’s growth; peduncles 3-5 mm. long, the
pedicels 2 mm. long with 1 or 2 basal or submedian subulate prophylla,
the uniflorous pedicels 3-6 mm. long, with 1 or 2 medial or supermedial
prophylla; flowers 4-merous; calyx patelliform, 2-3 mm. across, glabrous,
deeply 4-lobed, the lobes broadly deltoid, erose, eciliate; corolla rotate,
ca. 7 mm. across, the petals obovate or orbicular, 2.5 mm. in diameter,
one-fourth connate at the base; stamens two- thirds as long as the netals,
the anthers oblong, 1.5 mm. long: rudimentary ovary narrowly conic,
minute, the apex acute or shortly rostellate. Pistillate inflorescence:
solitary, axillary, on current year’s growth only; pedicels 12-15 mm. long,
glabrous, gradually enlarged toward the distal half, with 2 subulate median
prophylla; calyx 4-5 mm. in diameter, glabrous, 4-lobed, the lobes sub-
orbicular; corolla rotate, 7 mm. across, ovate, 2.5 mm. long, one-fourth
connate at the base; staminodes one-third as long as the petals, the sterile
anthers sagittate; ovary ovoid, 2 mm. long, the stigma discoid, convex.
Fruit globose or depressed-globose, black, 8 mm. long, 9 mm. in diameter,
the persistent calyx explanate, 5 mm. in diameter, the stigma discoid-
mammiform, the exocarp tunicate, the mesocarp very thin. Pyrenes 4,
oval in outline,. trigonous in cross section, back convex, rugose, striate, the
striae slightly elevated, the sides smooth, 4-6 mm. long, 3-5 mm. wide,
the endocarp sublignified or thick-coriaceous.

CHINA: Anhwei: Wang-shan, R. C. Ching 3025 (A, LU). Cheki-
ang: Siachu,z. C. Ching 1747 (A, US), 1795 (A). Kiangsi: Lung-
nan, S. K. Lau 4711 (A). Fukien: Pu-cheng- a R. C. Ching 2504 (A,
NY, US); Ku-liang, H. A. oe 6/60 (LU y% pone ging 6763 (A).
Kwanetu ng: Wung-yuen, S. K. Lau 794 (A, NY); Sin-fung, Y. W.
Taam 609 (A), 0/9 TA); 733 (A), 946 (A); Tseng-shing, W’. T. Tsang
20195 (NY), 20411 (A, NY); Tsung-fa, W. T. Tsang 20601 (NY); Ta-pu,
W. T. Tsang 21256 (A, NY), 21649 (A, NY). Hongkong: Ford
(fruit, A, NY); Ford (pistillate flower, NY).

Ilex viridis Champ. ex Benth. was first described from specimens col-
lected at Hongkong. It grows as a shrub up to 5 m. high. Its white
flowers appear in May and its black fruits mature in October.

Loesener_ placed this species as a variety of Jlex triflora Blume. In
superficial appearance the leaves and fruits of these two species are alike,
but on close examination //ex viridis can readily be distinguished from
Blume’s species by the inflorescences, fruit, and pyrenes. In J/lex triflora
the staminate inflorescences are shortly pedunculate, the pistillate flowers
are fasciculate, the fruits have small elevated stigmas, and the pyrenes are
striate but esulcate. In Jlex viridis the staminate inflorescences have
peduncles longer than the pedicels, the pistillate flowers are always solitary,
the drupes have flat large stigmas, and the pyrenes are slightly rugose.

Professor E. 5. Barghoorn of Harvard University has devised a method
for microscopic study of the epidermis of both living and fossil plants as
a means of identification. At his suggestion, I have examined both the
lower and the upper epidermis of the punctate-leaved species of Jlex.

1949] HU, THE GENUS ILEX IN CHINA 321

The lower epidermis of /lex viridis with its numerous stomata arranged
in groups is very different from that of Ilex triflora with a few diffused
stomata.

30. Ilex crenata Thunb. Fl. Jap. 78. 1784; Willd. Sp. Pl. 1 (2): 710.
1797; Pers. Syst. Veg. 174. 1797, et Syn. Pl. 1: 151. 1805; Poir. in
Lam. Encycl. Suppl. 3: 66.1813; Roem. & Schult. Syst. 3: 491. 1818;
DC. Prodr. 2: 16. 1825; Spreng. Syst. 1: 495. 1825; Dietr. Syn.
Pl. 1: 556. 1839; Sieb. & Zucc. in Abh. Bay. Ak. Wiss. IV, 2: 147.
1845; Regel in Gartenflora 13: 39. 1864; Miq. in Ann. Mus. Bot.
Lugd.-Bat. 3: 104 (Prol. Fl. Jap. 268). 1867; Franch. & Sav. Enum.
Pl. Jap. 1: 76. 1873; Maxim. in Mém. Acad. Sci. St. Pétersb. VII,
29 (3): 21, 33. 1881; Tanaka & Ono, Useful Pl. Jap. 3: 2, fig. 668.
1889; Matsumura, Shokubutsu Mei-i 149. 1895; Ito & Matsum. in
Jour. Sci. Col. Univ. Tokyo 12: 367. 1899; Nakagawa in Bot. Mag.
(Tokyo) 13: 108. 1899; Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 199 (Monog. Aquif. 1: 199). 1901; Dallimore, Holly
Yew Box 9, 120. 1908; Rehd. in Mitt. Deutsch. Dendr. Ges. 1908:
160. 1908, et in Bailey Stand. Cycl. Hort. 3: 1640. 1915; Chung in
Mem. Sci. Soc. China 1: 140. 1924; Kia, Pl. Sin. Ill. 490, fig. 828.
1937; Chen, Ill. Man. Chin. Trees 657, fig. 548. 1937; Merr. &
Chun in Sunyats. 5: 106. 1940; Makino, Ill. Fl. Nip. ed. 2, 369,
fig. 1106. 1940; Rehd. Bibliog. Cult. Trees Shrubs 402. 1949.

lex elliptica Sieb. ex Mig. in Ann. Mus. Bot. Lugd.-Bat. 3: 104. 1867.

Celastrus adenophylla Mig. op. cit. 2: 85. 1865, 3: 104. 1867.

flex crenata Thunb. var.. tvpica Loes. forma genuina Loes. op. cit. 201.

. Syn. nov.
Ilex crenata Thunb. var. typica Loes. forma kusnetzoffii Loes. op. cit. 202.
Syn, nov.

A much-branched evergreen shrub up to 5 m. high with puberulous
branchlets, small (usually less than 3 cm. long) coriaceous obtuse, rounded
or acute, crenate-serrate leaves, 1—7-flowered cymose staminate inflores-
cences and solitary, (rarely in 3-flowered cymes) pistillate flowers.

Branchlets straight, cinereous, rarely castaneous or nigrescent, the side-
shoots sometimes becoming thorn-like; third year’s growth 3-4 mm. in
diameter, plicate-rugose, minutely rimose, the lenticels sparse, oval or
orbicular, the leaf-scars minute, crescent-shaped, elevated; second year’s
growth 2—3 mm. in diameter, puberulous; current year’s growth 1—-1.5 mm.
in diameter, angular, cinereous or castaneous, densely pubescent, the
terminal buds usually poorly developed, broadly conic, the scales loose,
puberulous. Leaves occurring even on second year’s growth, 3-10 mm.
apart; stipules small, 1 mm. long, subulate, persistent; petioles short, 2—3
mm. long, one-eighth to one-sixth the length of the lamina, channeled above,
keeled and transversely rugose beneath, pubescent or puberulous; lamina
thickly coriaceous, olivaceous-brunneous, shiny and plicate-rugose above,
opaque and punctate beneath, obovate, ovate or oblong-elliptic, 1-2.5 cm.
(rarely up to 4 cm.) long, 0.5—1.5 (rarely up to 2) cm. wide; base obtuse,

322 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. Xxx

acute or cuneate; apex rounded, obtuse, or subacute, mucronulate; margin
crenate or serrate, the teeth 6-10 on each side; midribs plane or slightly
impressed and puberulous above, prominent, elevated near the base beneath,
the lateral nerves 2 or 3 pairs, obscure. Staminate inflorescences: cymose,
1—7-flowered, solitary and axillary to the scales or the lower leaves of the
current year’s growth or rarely pseudofasciculate on the second year’s
growth; peduncles 4—9 mm. long, the secondary axis when present 1 mm.
long, the pedicels 2-3 mm. long with 1 or 2 sub-basal prophylla, the uni-
florous pedicels 4-8 mm. long with 1 or 2 submedian prophylla; flowers
4-merous; calyx 2 mm. in diameter, glabrous, 4-lobed, the lobes broadly
deltoid, erose; corolla 4 mm. across, the petals broadly elliptic, 2 mm.
long, one-fifth connate at the base; stamens shorter than the petals, the
anthers ellipsoid, 0.7 mm. long; rudimentary ovary conical, the apex
slightly apiculate. Pistillate inflorescences: flowers solitary, or very rarely
2- or 3-flowered cymose, axillary, on the current year’s growth only; pedi-
cels 4-6 mm. long, club-shaped, ridged, with 1 or 2 submedian prophylla;
flower 4-merous; calyx 3 mm. in diameter, the lobes rounded; corolla
rotate, 6 mm. across, the petals ovate, 3 mm. long, one-third connate at the
base; staminodes one-half the length of the petals, the sterile anthers
sagittate; the ovary ovoid-conic, 2 mm. long, style sometimes evident,
the stigma discoid, 4-lobed. Fruit black, globose, 6-8 mm. in diameter,
the persistent calyx 3 mm. in diameter, explanate, the stigma inuie thickly
discoid, 1 mm. in diameter, distinctly lobed. Pyrenes 4, oblong-elliptic in
outline, 5 mm. long, 3—-3.5 mm. wide on the back, smooth, striate, esulcate,
the striae slightly impressed, the endocarp coriaceous.

CHINA: Fukien: Yen-ping, Dunn 2468 ;

bee Yezo: S. Hori in es (A); T. Tanaka in 1929 (A).
Hondo: S. Arimoto in 1903 (A); R. K. Beattie & Y. Kurihara 10438
(a) Lorena in 1885 (A); J. G. ae in 1905 (A, G); Maximowics in
1861 ee in 1862 (G); K. Miyabe in 1891 (A); Sargent, Aug. 1892 (A),
Oct. 1892 (A); Stebold (A); Shiwaya i “ ae Cree . Shiota 68 (A); K.
Uno 19111 (A); Warburg 8067 (A), 8071 (A); E. H. Wilson 7112 (A),
8195 (A); Watanabe in 1889 (G); C. ae in ce (G); Shikoku:
FE. H. Wilson 7790 (A).

KOREA: U. Faurie (A); E. H. Wilson 9452 (A).

Loesener differentiated his two forms, genuina and kusnetzoffii, by the
shape of their leaves, those of the former being obovate or ovate-elliptic,
and of the latter ovate or ovate-elliptic. Among our numerous specimens
I find many with both forms of leaves on the same branchlet. Thus there
is no reason to separate them.

Ilex crenata Thunb. is a native of Japan, where it is widely distributed.
Various cultivated forms occur there, whence they have been introduced
into Europe and North America. The plant is hardy in the United States
as far north as Boston. It is hoped that the following treatment will
clarify the nomenclature of these cultivated forms, which has been badly
confused.

1949 ] HU, THE GENUS ILEX IN CHINA 323

KEY TO THE VARIETIES AND ForMs

A. Length of the leaves less than 3 times the width; the lamina ovate or
obovate
B. Leaves on normal twigs ovate, obovate, oblong, elliptic, or even
lanceolate, but not suborbicular and not crowded at the tips of
the twigs.
C. Leaves plane on both surfaces, the form and apices variable.
D. Staminate cymes 1-5-flowered; the peduncles 2-5 mm
long; pedicels of the fruit 4-6 mm. long.
FE. Leaves over 15 mm. long, oblong-elliptic, the apex
ODEUSENOn SUDACULE +444 sree ee: 30a. f. latifolia.
EE. Leaves less than 15 mm. long, ovate or cee the
apex rounded
IF, Margin of the leaves serrate, the teeth more than
Ggonmeacht Side... ac) wees 30b. f. microphylla.
FF. Margin of the leaves subentire or with 2-6 teeth
on each side; a greatly dwarfed shrub.........
OR a ilice dace Visca y 4 30c. f. helleri.
DD. Staminate cymes 1-/-flowered; peduncles 10 mm. long;
pedicels of the fruit 10 mm. long. .30d. f. longipedunculata.
CC. Leaves convex, oblong, the apex rounded..... 30e. f. convexa.
BB. Leaves crowded at the ends of the twigs, suborbicular.........
Of. ar, martesit.
AA. Length of the leaves 3-5 times the width, the lamina narrow-elliptic or

i

lanceolat
B. Lamina entirely PRSCM 55-5 4 xn a9 G8 wey ee 30g. var. longtfolia.
BB eamitare vale CA te in Gaww eee wad aa be see 30h. f. luteo-variegata.

30a. Ilex crenata Thunb. forma latifolia (Goldr.) Rehd. Bibliog. Cult.
Trees Shrubs 402. 1949.
aa fortune hort. ex Mig. in Ann, Mus. Bot. Lugd.-Bat. 3: 104 (Prol.
. Jap. 1865, pro syn., non Lindl. 1857.
ne elliptica Siepold ex Mig. l.c., pro syn., non H.B.K., 1825.
Ilex crenata latifolia Goldr. in Caden 31: 129. 1887.
Ilex crenata var. major Nicholson in Kew Hand-list Trees Shrubs 1: 61.
1894, nom.; Dallimore, Holly Yew Box 121. 8.
Ilex crenata var. rotundifolia Maxim, ex Matsumura, Shokubutsu Mei-i,
149. 1895, nom.; Murakoshi, Ic. Encycl. Bot. 5: 72, fig. 152. 1935, descr.
A broad-leaved form with oblong or elliptic mature leaves 2-3 cm. long,
1-1.5 cm. wide, the margin serrate, the apex obtuse or subacute, the
staminate cymes 1—3-flowered, the peduncles 3 mm. long, the pedicels 1—2
mm. long, in fruit 6 mm. long.
JAPAN: Yezo: U. Faurie 6894 (A). Hondo: U. Faurie 6120
oe Goes Sargent in. Sept: 1897 (A), in: Oct, 1892 (A) 3 Se Pf ashiro Aor
. H. Wilson in 1917 (A).
This broad-leaved form is much cultivated in parts of Europe, Great
Britain, and the United States.
30b. Ilex crenata Thunb. forma microphylla Rehd. in Mitt. Deutsch.
Dendr. Ges. 1908: 160. 1908, et Bibliog. Cult. Trees Shrubs 402. 1949.

324 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

Ilex crenata var. unr opel Maxim. ex Matsumura, Shokubutsu Mei-i,
149, 1895, nom.; Rehd. in Bailey, Stand. Cycl. Hort. 3: 1640, 1915.

A dwarf shrub, the leaves oblong, ovate or obovate, the margin serrate
with 6 or more teeth on each side, 9-17 mm. long, 5-6 mm. wide, the
staminate cymes 1—5-flowered, subumbelliform, the peduncles 5 mm. long,
the pedicels 1-3 mm. long, in fruit 5-6 mm. long.

JAPAN? J, iS Jack in 1905 (A); Kotobuki Idrikawa 104 (A); R. Old-
ham 140 (G,N

7 . aurtie 713 (A, B); T. Taquet 624 (A).
30c. Ilex crenata Thunb. forma helleri Rehd. in Jour. Arnold Arb. 20:
417. 1939,

Ilex hellert Verkade’s Nurseries Cat. 1936: 13. 1936; Craig, Deser. Price-
list, 493725). 1937.

Ilex crenata helleri Fleming, Flem. Nurser vi 1937 % 171937, diom.*. Ta,
Chadwick in Am. Nurseryman 79(1): 21, fig. (cover). 1944; Rehd.
Bibliog. Cult. Trees Shrubs 402. 1949

A dwarf shrub of compact and stowed habit, the mature leaves elliptic,
8-11 mm. long, the margin crenate-serrulate with 2—4 teeth on each side.

CULTIVATED: United States: E. Morell in 1937 (A).
30d. Ilex crenata Thunb. forma longipedunculata, f. nov.

Frutex; foliis 2-3 cm. longis, 1-2 cm. latis, oblongis, obovato-ellipticis
vel obovatis, apice rotundatis, obtusis vel sqpaqinc: é inflorescentiis
cymosis, 1—7-floris, pedunculis 10 mm. longis, pedicellis 2-3 mm. longis;
9 floribus solitariis, pedicellis 10-12 mm. longis.

A broad-leaved form with mature leaves 2-3 cm. long, 1—-1.5 cm. wide,
oblong, obovate-elliptic or obovate, the apex rounded, obtuse or subacute,
the staminate cymes 1—7-flowered, the peduncles 10 mm. long, the second-
ary axis well developed, 1-2 mm. long, the pedicels 2-3 mm. long, in fruit
10-12 mm. long

CHINA: Shantung: Tsingtao, First Park, C. Y. Chiao 2584 (tyre,
Rey Uy US. SS

°AN: P.H. Dorsett & W. J. Morse in 1929 (A); J. G. Jack in 1905
(A); Masamune (NY); Rikuzen (G); Jugiyama Setichi in 1931 (A)
K. ee in 1911 (A); C. Wright (G).
30e. [lex crenata Thunb. forma convexa (Makino) Rehd. Bibliog. Cult.

Trees Shrubs 402. 1949,

Ilex crenata var. convexa Makino in Jour. Jap. Bot. 5: 27. 1928: Tera-
saki Nipp. Shokubutsu Suppl. 2866. 1938; Makino, Ill. Fl. Nip. ed. 2
369, fig. 1107. 1940.

Ilex crenata {. bullata Rehd. in Jour. Arnold Arb, 12: 73, 1931.

This convex-leaved shrub is cultivated in Japan (fide Makino) and in

the United States. I have seen specimens collected in the Arnold Arbo-
retum by Judd and Rehder.

30f. lex crenata var. mariesii Bean in Kew Hand-list Tree Shrubs ed.
2, 89. 1902, nom.; Dallimore, Holly Yew Box 122. 1908, descr.

1949 ] HU, THE GENUS ILEX IN CHINA 325

Ilex martes Veitch ex Bean, Trees Shrubs Brit. Isl. 1: 646, 1914, pro syn.

flex numnudaria Franch. & Sav. Enum. Pl. Jap. 2: 311. 1878, non Reissek.
1861.

Ilex crenata var. nummularia (Franch. & Sav.) Yatabe in Bot. Mag.
Tokyo) 6: 157. 1892; Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 202 en Aquif. 1: Bee 1901; Murakoshi, Ic.
Enceycl. Botsed=:3,0ce/2. TAC RG: lol NOG ae Makino, Ill. Fl. Nip.
369, fig. 1107. 1940: Rehd. cates Cult. Trees Shrubs 402. 1949, Syn.

nov.

A small shrub 1-2 m. high with leaves crowded at the ends of the twigs,
the lamina ovate or suborbicular, 3-11 mm. long and wide, subentire or
obscurely crenate, the teeth 1—3 on each side. ,

This form is cultivated in Japan and also at Kew (J. W. Bean!) and at
the Arnold Arboretum (Rehder!). I accept the varietal name mariesti
Bean (1902) as validated by Dallimore (1908) because Yatabe’s varietal
name nummularia was based on an invalid specific name.
30g. [lex crenata Thunb. var. longifolia Goldr. in Garden 31: 129. 1887.

Ilex crenata f. longifolia (Goldr.) ate in Mitt. Deutsch. Dendr. Ges.

1908: 161. 1908, et Bibliog. Cult ees Shrubs 402. 1949.

Mature leaves lanceolate or ie 1-3 cm. long, 4-11 mm. wide,
the normal length 4 or more times the width, the apex acute.

JAPAN: &. H. Wilson 6220 (A); Herb. Sci. College, Imp. Univ. Japan,
in 1886 (A); Sargent in 1892 (A).

This variety is cultivated in Germany, Great Britain, and the United
States

30h. Hex crenata Thunb. forma luteo-variegata (Regel) Rehd. in Mitt.
Deutsch Dendr. Ges. 1908: 161. 1909.
Ilex crenata luteo-variegata Regel in Gartenfl. 13: 39. 1864.
Ilex crenata var. aureo-variegata Goldr. in Garden 31: 129. 1887.
as crenata var. puauiegaie Nichols. in Kew Hand-list Trees Shrubs 1:
1894, nom.; Rehd. in Mitt. Deutsch. Dendr. Ges. 1908: 161. 1908;
nen Trees Shrubs Brit. Isles 1: 646. 1914; Dallimore, Holly Yew

Ilex fortunei foam aureo-variegata hort. ex Schelle in Beissner & al.,
Handb, Laubh.-Ben. 291. 1903, nom.
A form distinguished chiefly by its yellow or golden variegated leaves.
Cultivated at Kew. I have examined specimens collected by J. W. Bean
and G. Nicholson.

31. [lex szechwanensis Loes. in Nov. Act. Acad. Caes. Leop.-Carol. Nat.
Cur. 78: 347 (Monog. Aquif. 1: 347). 1901, et in Sarg. Pl. Wils. 1:
80.1911; S. ¥. Hu in Ic. Pl. Omei. 2: pl. 169. 1946.

Llex ssechwanensis tee gede Loes. ex Diels in Bot. Jahrb. 29: 436.
1900, nom. nud., et in Nov. Act. Acad. Caes. Leop.-Carol]l. Nat. Cur.
78: 348 (Monog. ae re 348). 1901, descr. Syn. nov.

Ilex ssechwanensis forma puberula Loes. lc. Syn. noz

Ilex crenata Thunb. var. scoriaruim W. W. Sm. in Notes Bot. Gard.

Edinb. 10: 41. 1917; Anon. in op. cit. 17: 278, 285, 305. 1930

326 JOURNAL OF THE ARNOLD ARBORETUM [VoL. Xxx

Tlex gee Blume var. viridis sensu Comber in Notes Bot. Gard. Edinb.
18: 58. 1933, non Loes.

An evergreen shrub 1-3 m. high with puberulous subquadrangular
branchlets, ovate-elliptic, ovate-oblong, or sublanceolate serrate, rarely
crenate, punctate leaves, 1—7-flowered staminate cymes, peduncles of cymes
longer than the pedicels, solitary pistillate flowers on current year’s growth,
4—7-merous flowers, small elevated stigmas, and smooth striate esulcate
pyrenes.

Branchlets olivaceous-brunneous, subquadrangular, third year’s growth
3-4 mm. in diameter, the lenticels obscure or obsolete, the leaf-scars
crescent-shaped, slightly elevated; second year’s growth 2.5-3 mm. in
diameter, subquadrangular in cross section, under a lens puberulous; cur-
rent year’s growth ridged and canaliculate, 1.5—2 mm. in diameter, minutely
puberulous, especially the grooves; terminal buds usually well developed,
conic, pubescent. Leaves occurring also on the second year’s growth, 3-15
mm. apart; stipules scale-like, ovate-deltoid, acute, 1 mm. long, persistent;
petioles 4-6 mm. long, one-tenth the length of the lamina, puberulous,
shallowly grooved above, much or slightly winged by the decurrent leaf-
base; lamina subcoriaceous or coriaceous, olivaceous, opaque or sometimes
slightly shiny above, definitely opaque and punctate, sometimes puberulous
especially along the midribs beneath, ovate-elliptic, ovate-oblong or rarely
sublanceolate, 3.5—7 cm. long, 2—4 cm. wide, the base obtuse, rarely acute,
the apex acuminate, the acumen deltoid, 3-5 mm. long; midrib plane or
slightly impressed, puberulous above, elevated and glabrous or puberulous
beneath, the lateral nerves 6 or 7 pairs, usually evident or sometimes
obscure on both surfaces, the reticulation of the veinlets obscure. Stamin-
ate inflorescence: 1—7-flowered, cymose, solitary, axillary to the scales or
basal leaves of the current year’s growth, rarely fasciculate; peduncles
4-8 mm. long, the pedicels 2-3 mm. long, with 1 or 2 median or super-
median prophylla, the uniflorous pedicels 3-5 mm. long; flowers white, 4—7-
merous; the calyx patelliform, glabrous or slightly puberulous, 2—2.5 mm.
across, ‘deeply 4—7-lobed, the lobes ovate-deltoid, 1 mm. long, erose or
dentate, rarely ciliate; the corolla rotate, the metals 4 or 5-lobed, the lobes
ovate, eciliate, one-fourth connate at the base; stamens shorter than the
petals, anthers ovate-oblong; staminodes one-half the length of the petals,
the sterile anthers sagittate; rudimentary ovary subglobose, the center
shortly rostellate, inconspicuously lobed. Pistillate inflorescence: flowers
solitary, axillary, on current year’s growth only; pedicels 8-10 mm. long,
with 2 supermedian prophylla; flowers 4-merous; calyx patelliform, 3 mm.
across, shallowly 4-lobed, the lobes rounded, erose; corolla suberect, 4 mm.
across, the petals ovate, 2.5 mm. long, one-sixth connate at the base;
staminodes very minute, one-fifth the length of the petals, the sterile
anthers sagittate; ovary subglobose-ovoid, 1.5 mm. in diameter, the style
sometimes evident, the stigma thick-discoid, convex. Fruits globose or
depressed-globose, 6 mm. long, 7-8 mm. in diameter, the persistent calyx
explenate, 3.5-4 mm. in diameter, the stigma small, 1 mm. in diameter,

1949] HU, THE GENUS ILEX IN CHINA 327

thickly discoid, distinctly 4-lobed, slightly elevated. Pyrenes 4, oval or
suborbicular in outline, trigonous in cross section, 4.5—5 mm. long, 4 mm.
wide, smooth, striate, the striae slightly impressed, the endocarp coriaceous.

Kew anetunme ce. Lohschong, «Cer e.-sl so <2) 197 (NY):
Kwangsi: Tsu-yuen-hsien, 7. S. Tsoong (Z. S. Chung) 83425 (A);
Yao-shan, C. Wang 40222 (A). Hupei (Hupeh): A. Henry 5808 (iso-
TYPES of Ilex ssechwanensis var. puberula, A, US), 6912 (1sotyPE, G) ;
I-chang, FE. H. Wilson 461 (A, US); Chien-shi, E. H. Wilson 1333 (NY);
Chang-yang, FE. H. Wilson 1333A (fragment, A); without precise locality,
E. AH. Wilson 1964 (A, NY, US). Hunan: Mo-fou-shan, Y. K. Hsiung
5900 ee Kweichow: Fan-ching-shan, Steward, Chiao & Cheo 434
(A, NY), 634 (A, NY, US), 756 (A, NY, US); Kiang-kow, Y. Tsiang
7528 (NY), 7550 (A); Yin-kiang, Y. Tsiang 7845 (NY), 7974 (NY).
Szechuan: Wu-shan, A. Een 5716 (1sotyPEs, G, NY; fragment, A) ;
Chung-king, S. Y. Hu 5001 (WU); Opien, T. S. Chao 595 CSS):
Yao 4438 (SS), 4576 (SS); Hung-ya, C. W. Yao 2266 (SS), 3900 (SS):
Mt. Omei, C. Y. Chiao & C. S. Fan 807 (A); E. H. Wilson 4783 (A);
Kwan- een Bote Veang 2ors07( A). Sikene 2 dienchuan; 3. 75..274
1322 (A); Zayul, Kingdon Ward 10978 (B). Yunnan: Meng-tze, A.
Henry 11012 (A, NY), 11303 (A); G. Forrest 7867 (1soTYPE of Ilex crenata
var. scoriarum, staminate flower, A), 8713 (1soTyPE of Jlex crenata var.
scoriarum, pistillate flower, A), 17642 (A )s TAG) (A), 27391 (A), eo
(2 g-yueh, J. F. Rock 7261 (A, US); Ping-pien-hsien, H. T.
60062 CS 60276 (A), 61434 (A), 61660 (A).

Ilex szechwanensis was based on material collected on the Hupei-
Szechuan border not very far from where the living Metasequoia has been
recently discovered. There it grows as a shrub in thickets or on slopes,
often near bamboo groves. Its white flowers appear in April and the fruit
turns black in October. When Loesener described the species he cited
no specimens. However, he enumerated three forms distinguished on the
bases of the indumentum and leaf-form. I have before me isotypes of the
first two forms, calva and puberula. With the numerous specimens from
the same general area for comparison, I note such gradual variations in the
density of the indumentum and in the leaf-form that I see no reason for
maintaining Loesener’s named forms. As to his forma angulata, I have
seen no authentically named specimens, but I judge that it might prove to
be merely a vigorous growth shoot

The Yunnan and Kweichow specimens are much more pubescent than
the other material; even the leaf-surfaces are puberulous. But the stamin-
ate inflorescences are cymose, and the peduncles are decidely longer than
the pedicels. These characters, plus the larger size of the leaves and the
geographic factors involved, persuade me that it is better placed with /lex
szechwanensis than with lex crenata Thunb.

Ilex szechwanensis is intermediate between /lex crenata Thunb. and
Ilex triflora Blume. Like the former, it has staminate cymes with the
peduncles longer than the pedicels and striate pyrenes with the striae
slightly impressed. Like the latter, it has larger and thinner leaves. Its
solitary fruits indicate kinship with Jlex viridis Champ. ex Benth., but the

328 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXX

latter has shiny leaves, impressed midribs, 4-merous flowers, plane and

broad (2.5 mm.) stigmas and rugose pyrenes. Because of the widely

separated geographic areas concerned, hybrid origin seems to be scarcely
possible.
32. Ilex triflora Blume Bijdr. 1150. 1826; Dietr. Syn. Pl. 1: 555. 1839;
Miq. FI. Ind. Bot. 1 (2): 594. 1859; Loes. in Nov. Act. Acad. Caes.
Leop.-Carol. Nat. Cur. 78: 344 (Monog. Aquif. 1: 344). 1901; Val.
in Med. Dep. Landb. 18: 25 (Koord. & Val. Bijdr. Booms. Java 13).
1914; Groff in Lingnan Sci. Bull. 2: 64. 1930; Pitard in Lecomte,
Fl. Gén. Indo-Chine 1: 852. 1912; Tardieu-Blot in Fl. Gén. Indo-
Chine Suppl. 1: 719. 1948.
Ilex horsfieldii Miq. Fl. Ind. Bat. 1(2): 594. 1895; Maxim. in Mém.
Acad. Sci. St. Pétersb. VII, 29(3): 21. 1881,

llex triflora Blume var. hovsheldn (Miq.) Loes. op. cit. 347. 1901; Vale-
ton in Med. Dep. Landb. 18: 27 (Koord. & Val. Bijdr. Booms. Java 13).
1914; Koord. & Val. Atlas Baum. Java fig. 790, F-H. 1918. Syn. nov.

_ niftor Blume var. javensis Loes. le.; Va ee op. cit. 25. 1914;

& Val. op. cit. fig. 790, A-E. 1918. Syn. no

Hex een Rolfe: in Jour. Linn. Soc. Bot. 21: 309, “1884,

flex triflora Blume var. lobbiana (Rolfe) Loes. op. cit. 346. 1901; Hand.-
Mzt. Symb. Sin. 7: 658. 1933. Syn. nov.

flex griffithit Hook. f. Fl. Br. Ind. 1: 601. 1875.

Ilex triflora var. kurziana Loes. op. cit. 346. 1901. Syn. nov.

flex theicarpa Hand.-Maz. in Sinensia 3: 188, 1933. Syn. nov.

llex fleuryana Vardieu in Not. Syst. XII, 15: 119. 1945. Syn. nov.

An evergreen shrub or small tree with pubescent branchlets, oblong-
elliptic or rarely ovate-elliptic or ovate minutely serrate leaves with
punctate lower surfaces, fasciculate inflorescences, pubescent and ciliate
calyx, globose or ellipsoid fruits and 4 smooth estriate, esulcate pyrenes.

Branchlets zigzag, ridged, subquadrangular in cross-section, rarely sub-
terete; third year’s growth 2 mm. in diameter, puberulous, plicate-rugose,
the lenticels lacking, the leaf-scars subsemi-orbicular, slightly elevated;
the second and current years’ growth 1-1.5 mm. in diameter, ridged and
canaliculate, pubescent; terminal buds poorly or not at all developed.
Leaves even found on third year’s growth, 2-12 mm. apart; stipules deltoid,
acute, 1 mm. long, persistent; petioles 4 mm. long, about one-tenth the
length of the lamina; lamina subcoriaceous, brunneous-olivaceous, opaque
on both surfaces, puberulous or glabrescent at maturity, punctate beneath,
elliptic, oblong or ovate-elliptic, rarely ovate or lanceolate, 3-9 cm. long,
1.5—4 (usually 2.5) cm. wide; base rounded or obtuse; apex acute or very
shortly acuminate, the acumen 3—5 mm. long, very rarely obtuse; margin
serrate, subundulate; midrib impressed and pubescent above, elevated and
puberulent underneath, the lateral nerves 7-11 pairs obscure above, the
reticulation of the veinlets obscure. Inflorescences fasciculate with abortive
or rarely active terminal buds, axillary, distinctly pubescent all over;
flowers 4-merous. Staminate inflorescences: individual branches of the
fascicles 1—3-flowered, the peduncles of the 3-flowered cymes 2 mm. long,

1949 ] HU, THE GENUS ILEX IN CHINA 329

the pedicels 2-3 mm. long, the prophylla 1 or 2, basal or submedian; the
calyx patelliform, 3 mm. across, pubescent and sparsely ciliate; corolla
rotate, 5 mm. across, the petals broadly ovate, one-fourth connate at the
base; stamens slightly shorter than the petals, the anthers elliptic; rudi-
mentary ovary pyramidal, the apex shortly rostellate, cleft. Pistillate
inflorescences: individual branches of the fascicles uniflorous, the pedicels
6-14, rarely up to 18 mm. long with 2 median or submedian prophylla;
calyx as in the staminate flowers; corolla suberect, the petals ovate;
staminodes one-third the length of the petals, the sterile anthers cordiform;
ovary ovoid, 1.5 mm. in diameter, the stigma thickly discoid, inconspicu-
ously lobed. Fruit globose or Ailipecid, 7-8 mm. long, 7 mm. in diameter,
the persistent calyx explanate, 4 mm. across, sparsely ciliate, the stigma
discoid, 4-lobed. Pyrenes 4, oval-elliptic in outline, 6 mm. long, 4 mm.
wide, smooth, striate, the striae slightly impressed, the endocarp coriaceous.
CHINA: Kiangsi: Hong-san, J. L. Gressitt 1545 (A); Yu-du-hsien,
ee 117s (CN) eaksien-nan, S. 1K. Law 3947 (CA WS)i44/4 (A. US).
Kweichow: Cheng-feng, Y. Tsiang 4527 (NY); Tung-tze, Y. Tsiang
4997. (NY); Tuh-shan, Y. Tsiang 6645 (NY), 7031 (NY), 7073 (NY).
Fukien: C. Fukien, Dunn (Herb. Hongk. no. 2467) (A); Hing-hwa,
H. H. Chung 925 (A); Min-how-hsien, H. H. Chung 2190 (A); Ku-tien,
H. H. Chung 4054 (A); Ku-liang, H. H. Chung 6444 (A, LU), 6786 (LU),
6878 (A); Ku-dien, H. H. Ging 7944 (A, LU); Hok-chiang, Tang Siu-
ging. 19092 (BU) a iol43"(1U). 15176 (LU), Kwanetume - ‘Canton;
C. O. Levine (CCC) 1477 (A, G); Lung-t’'un Mt., To, Tsang & Tsang
(GEG) 12319 CUS) wWirtung, 7. ¥. Chan 3/72 (Ade a ee tszenuy. V
Chun 5909 (A, US); Nan-hao, W. Y. Chun 7752 (A); ng-yuen, S. K.
Lau 2145 (A), 2791 (A); Loh-fau Mt., E. D. Merrill Ree ae Y. Tsiang
1665 (A); Yao-shan, S. S. Sun 9968 (NY); Sin-fung, Y. W. Taam 213
GAS) S30 CA ens 09, (A), 853 (A); W. T. Tsang 900A (A); Ta-pu, W. T.
Tsang 21022 (A, NY), 21187 (A, NY); Lung-men, W. T. Tsang 25318
(Gaur cae Ae Pok-lo, T. M. Tsui 80 (NY); Ying-tak, T. M. Tsui 407
(A, NY, US); Yang-shan, T. M. Tsui 780 (NY, US); Sun-wui, Tso &
Tsiang 2043 CONN )e Tomekome : C-Pord (US). Galaga mak
G). Lantao Island: W.T. Tsang (Herb. Lingnan no. 16584) (A
ee US). Hainan: Dung-ka, N. K. Chun & C. L. Tso 43846 (A, NY):
ng-shiu, H. Fung 20212 (NY, US); Po-ting, F. C. How 73163 (A); S. K.
ie 28047 (A); Yai-chow, H. Y. Liang 62081 (A), 62526 (A, NY).
Kwangsi: W. Po-seh, &. C. Ching 7520 (1sotyre of Ilex theicarpa,
NY; photo & pisces A) u-chen, ke. -C.. Ching 5793 (IU), 58997 (Ew),
7058 (LU); Seh-feng-dar-shan, R. C. Ching 7805 (NY); S. Nan-ning,
jie Os OF 5 RG) oe (Aerie us) Lin-yuin-hsien, A. N. Steward & H. G.
Cheo 557 (A, NY); Shang-sze, W. T. Tsang 22071 (A), 22131 (A), 22377
(An Waictsap, WP Tsang 22719 (A), 23235 (A), 24548 (NY) Hane-=
on-yuen, T. S. Tsoong (Z. S. Chung) 81732 (A); Yao-shan, C. Wang
39163 (A), 40592 (A). Yunnan: Sze-mao, A. Henry 12018 (A),
IZOLS AD CEU Orel eOTS Ee ICA MOS) 2018C (A, US); 12078 Da CA swath
out precise locality. E. E. Maire 3874 (photo and fragment, A); Shang-pa-
hsien, 2. f. isa1 54396 (A) Ta-li, €. W. Wang 71723 (A) =Fo-haiG.
Wang 73508 (A), 73521 (A), 73715 (A), 73844 (A), 75058 (A), 8O178A
(A).

330 JOURNAL OF THE ARNOLD ARBORETUM [vOL. Xxx

INDO-CHINA: Tonkin: Fleury 32126 (type of Ilex fleuryana,
fragment, A); A. Pételot 5557 (A), 5756 (A), 5868 (A), 6278 (A);
W.T. Tsang 27008 (A), 27069 (A), 27379 (A), 27488 (A), 29800 (A),
29043 (A), 29153 (A), 29329 (A), 30659 (A

INDIA: Assam: Jenkins (1soTyPE of Hex grifithu, G); Prain’s
Collector 876 (A); Reporter Econ. Prod. Government India 17256 (A).
East Bengal: Griffith 2001 (1soTypE of /lex griffithu, A).

MALAY PENINSULA: UM. R. Henderson 7349 (NY); R. E. Holttum
(A); Griffith (iasotyre of Ilex griffthu, G).

SUMATRA: Rahmat Si Bolla 852 (A).

JAVA: Blwne (A), Blume from Mt. Salak (A, probably a duplicate of
the type); H. O. Forbes 1000A (A, G), 2901 (fragment, A, G); Koorders
9917 (fragment, A); Zollinger (A).

BORNEO: #. Hallier 2471 (fragment, A), 2453 (fragment, A); Sara-

wak, Richards 1963 (A).

The Lobb specimen at Kew labeled as from Luzon did not come from
the Philippines. The Lobb specimens were mislabeled, duplicates of the
same collection distributed as from Java, Borneo, the Malay Peninsula and
Luzon. This specimen is the basis of /lex lobbiana Rolfe, which, because
of its supposed geographic occurrence, Loesener placed as a variety of the
Javan /lex triflora Blume. The specimen undoubtedly was from Java,
though there is a bare possibility that it may have come from the Malay
Peninsula or Borneo. See Merrill, E. D. Philipp. Jour. Sci. 10 (C): 190.
1915.

With all the specimens cited above, I cannot find any key character to
separate the material collected from various geographic areas. I conclude,
therefore, that Jlex triflora is a variable species. The species was first
described from material collected in Java. The few Javan specimens
available exhibit as great variation in indumentum, shape and size of the
leaves, and length of the fruit pedicels as do the specimens from the
several other geographic areas involved

Loesener referred Forbes 1000A and 2901 from Java, as well as Hors-
field’s specimen, to J/lex triflora var. horsfieldti. J have compared the
duplicates of these two numbers with authentic specimens of /lex triflora
and found no differences between them except that Forbes’ specimens
have longer pedicels (10-18 mm.) than those of Blume’s which are in a
flowering stage.

Loesener referred to J/lex triflora var. lobbiana not only the original
Lobb material from “Luzon (?)” (which did not come from that island),
but also a Griffith specimen from Malacca, a Teysmann specimen from
Borneo and a Warburg specimen from Fuchow, China. I have seen
duplicates of several of these. They are all similar to Blume’s specimens
in indumentum, color and texture of the leaves and fruits, and I see no
reason for recognizing even a variety here, much less a species.

Ilex griffithii Hook. f. was based on a Jenkins specimen from Assam and
Griffith specimens from Malacca and Assam. Griffith’s Malaccan speci-

n is normal J/lex triflora. Jenkins’ Assam specimen has rather large
glabrescent oblong-elliptic leaves, more numerous lateral nerves, and

1949 ] HU, THE GENUS ILEX IN CHINA 331

fasciculate inflorescences. It agrees with specimens cited from Yunnan
and Indo-China.

When Handel-Mazzetti described Jlex theicarpa he overlooked the
glands on the lower surfaces of the leaves. He says: “Proximae, quamvis

Te ViGus
I crenata
xxxxx T szechwanensis }%
Sie ethig10ra

—-- [ pedunculosa
Pee er O Cte

——" T yunnanensis -
—-- T sugerokt

Fic. 7. Geographic distribution of the Chinese species of the section
Paltoria.

332 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

folia epunctata, videntur J. triflora Bl. et I. szechwanensis Loes.” Ex-
amination of an isotype shows that the leaves are ee glandular,
but that the glands are obscured by the indumentu

Fleury 32126 (wototyPE for /lex fleuryana Tardieu) is a normal
staminate branch of /lex triflora. Its pyramidal rudimentary ovary is very
characteristic of the species.

Ilex triflora is manifestly closely related to //ex crenata Thunb., which
differs in having smaller and fewer-nerved leaves, solitary fruits, and
peduncles of the staminate cymes longer than the pedicels; yet Loesener
placed the former species in [Sect.] Aquifolium and the latter in [Sect. |
Paltoria.

In an attempt to find tangible differences I have studied intensively all
of the specimens cited. I have made use of high magnification, and have
examined the plants not once, but several times. I have tried to distinguish
entities according to the density of the indumentum, the pattern of the
inflorescences, the length of the pedicels, the form and size of the stigmas,
the characters of the calyces, the form of the rudimentary ovaries, and the
size and pattern of the pyrenes. In none of these could I find a constant
character strong enough to warrant their separation. An attempt to dis-
tinguish them on the basis of their geographical location was no more
successful. Specimens from widely separated areas, such as Hainan
(Lau 28047), Kwangsi (Ching 7879 and 8210), Yunnan (Henry 12018C
and 12018D and Wang 80178A), and Indo-China (Tsang 29800) all
impress me as being exactly like the Javan specimens. It is not logical
to take certain plans and name them //ex triflora Blume, and at the same
time to put their next-door neighbors into another species, when morpho-
logically there is no ground for doing so. My study led me to believe
that what I had was a somewhat variable collective species whose center
of origin was the region about the China-Indo-China border, where it
flourishes, and that from there it had spread southward through the Malay
Peninsula to Sumatra, Borneo, and Java, westward to Assam and Bengal,
and northward and eastward into China.

In certain parts of China and in Taiwan forms which appear to be inter-
mediate between /lex triflora Blume and Jlex crenata Thunb. occur. These
might be interpreted as weak species or as varieties of either of the two
species. A good illustration of this is flex crenata Thunb. var. scoriarum
W. W. Sm. Its staminate inflorescences are essentially those of Jlex
crenata, and the leaves are those of /lex triflora in form and size. Smith
placed the variety under the former species, while Comber interpreted it
as the latter. The form discussed immediately below falls in this same
intermediate category.

32a. ex triflora Blume var. kanehirai (Yamamoto), comb. nov.
Ilex crenata Thunb. var. kanehirat Yamamoto, Suppl. Ic. Pl. Formos. 1:
31, fig. 11. 192
llex alg (Yamamoto) Koidz. in Bot. Mag. (Tokyo) 43: 389. 1929;
Kanehira, Fo . Trees 1936.
llex a ee var. guibn Kanehira Le. fig. S20, 1936:

1949] HU, THE GENUS ILEX IN CHINA 333

An evergreen shrub with puberulous branchlets; leaves coriaceous or
subcoriaceous, oblong, obovate or oblong-elliptic, the apex rounded, obtuse
or rarely acute, never acuminate; inflorescences fasciculate, the staminate
fascicles composed of 1-3-flowered cymes, the pedicels 2-4 mm. long,
peduncles of 3-flowered cymes 1-2 mm. long, shorter than the pedicels;
infructescences fasciculate, fruits globose, 5-6 mm. in diameter, the per-
sistent calyx explanate, mm. across, ciliate; stigma small, elevated,
distinctly 4-lobed.

CHINA: Fukien: near the Chekiang border, R. C. Ching 2225 (B,
LU, US). Kwangtung: Loh-fau-shan, C. O. Levine CECE) 31597
(US); F. P. Metcalf 17507 (LU). Taiwan: R. Kanehira (photo of
type of Ilex crenata var. kanehirai, TU): E. H. Wilson 10150 Can):
Hainan: Five Finger Mt., F. A. McClure (CCE 05 94 = (AS

Llex triflora var. kanehirai is intermediate between Jlex triflora and
flex crenata Thunb. It resembles the former in having fasciculate infruc-
tescences on the second year’s growth and peduncles of the staminate
cymes shorter than the pedicels. It resembles the latter only in its oblong-
obovate leaves rounded or obtuse at the apex. Since Jlex crenata is
characterized by solitary fruits found only on the current year’s growth,
it seems preferable to place this as a variety of Ilex triflora.

SERIES 2. CASSINOIDES (Losgs.), coms. nov.

Ilex subgen. Euilex, ser. A. Lioprinus, sect. 2. Cassinoides Loes. in Engler
& Prantl, Nat. Pflanzenfam. Nachtr, 218. 1897, et in Nov. Act. Acad.
Caes. Leop.-Carol. Nat. Cur. 78: 131 (Monog. Aquif. 1: 131). 1901.

Branchlets puberulous; leaves coriaceous or subcoriaceous, serrate or
subentire, rarely entire; inflorescences solitary, axillary, on current year’s
growth only, staminate ones cymose, 3—9-flowered, pistillate ones usually
uniflorous, rarely 3-flowered, the pedicels of the fruit 1—5.5 cm. long, rarely
shorter; flowers 4—6-merous; calyx usually ciliate, the petals well united,
the stamens shorter than the petals, the rudimentary ovary globose; fruit
red, globose; pyrenes 4—6, smooth, neither striate nor sulcate, the endocarp
thickly coriaceous or sublignescent.

KEY TO THE SPECIES

A. Leaves 5-8 cm. long, usually entire, sometimes serrate, the midribs
plane or slightly impressed; leaf-apices acuminate; pedicels of the fruit
25-55 mm. long. (Central China and Japan)....... 33. I. pedunculosa.

AA. Leaves less than 4 cm. long, serrate or rarely subentire, the midrib
elevated and pubescent above; leaf-apices acute, obtuse or rounded;
pedicels of the fruit 6-15 mm. long.

cS

334 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

BB. Leaves ovate, ovate-lanceolate or elliptic; petioles 2-5 mm. long,
one- fourteenth to one-fifth the length of the lamina.

Cs. heaeeir pedicels less than 4 mm. long; leaves obovate or ob-

; the apex rounded. (Alpine Yunnan)..... . J. rocki.

GO, oe pedicels 10-15 mm. long (6 mm. long in a variety of

i, iis leaves ovate or ovate-elliptic, rarely lanceo-

>. Pyeacs 4; branchlets thickly ferruginous- pubescent ; leaves
aristate-serrate nearly to the base (except in one variety).
(Western and Central China and Taiwan)............
io Miia itp id add neta ee ee oo ae 35. J. yunnanensis.
DD. Pyrenes 4-6; branchlets puberulous; leaves serrate or

crenate, the lower half entire. (Japan and Taiwan
See ee ee hee eee en re eee eee 36. TI. er
33. Ilex pedunculosa Miq. in Versl. Med. Kon. Akad. Wet. II, 2: 83.
1868 [1866] (Repr. 19. 1866), et in Ann. Mus. Bot. Lugd.-Bat. 3
106. 1867; Franch. & Sav. Enum. Pl. Jap. 1: 77. 1873; Maxim.
in Mém. Acad. Sci. St. Pétersb. VII, 29 (3): 25, 37. 1881; Loes.
ex Diels in Bot. Jahrb. 29: 435. 1900, et Nov. Act. Acad. Caes. Leop.-
Carol. Nat. Cur. 78: 108 (Monog. Aquif. 1: 108). 1901; Kia, Pl.
Sin. Ill. 492, fig. 832.1937; Makino, Ill. Fl. Nip. 367, fig. 1100. 1940.
lex ee genuina Loes. op. cit. 110; Chien in Contr. Biol. Sci.
a 3: 58. 1927; Hand.-Mzt. Symb. Sin. 7: 655. 1933. Syn.

i eee f. continentalis Loes. ex Diels in Bot. Jahrb. 29: 435.
1900, nom. nud., et in Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78:
110 (Monog. Aquif. 1: 110). 1901, descr., et in Sarg. Pl. Wils. 1: 76.
1913; Chien in Contr. Biol. Sci. Soc. China 3: 58. 1927; Rehd. in
Jour. Arnold Arb. 7: 156. ely 14: 240. 1933; Masamune Fl. Kainant.
[Hainan] 174. 1943. Syn. no
Ilex purpurea Hassk. var. eee Loes. in Léveillé, Fl. Kouy-Tchéou
201. 1914,
Ilex morii Yamamoto, Suppl. Ic. Pl. Form. 1: 38, fig. 18. 1925.
Ilex impressivena Yamamoto, Suppl. Ic. Pl. Form. 1: 34, fig. 15. 1925;
Kanehira, Form. Trees 373, fig. 329. 1936. Syn. nov.
An evergreen shrub up to 5 m. high with ovate subentire acuminate
leaves, long slender fruit-Stalks, and smooth esulcate and estriate pyrenes.
Branchlets straight, subterete; second year’s growth 4-5 mm. thick,
plicate-striate, the bark cinereous, the lenticels minute, circular, incon-
spicuous, the leaf-scars semi-orbicular; current year’s growth 2-3 mm. in
diameter, angular, brunneous-nigrescent, minutely puberulent especially
— the nodes. Leaves occurring even on the third year’s growth, 3-5
m. apart; stipules deltoid, falcate-acute, 1 mm. long, persistent; petioles
ae 12-17 mm. long, one-fifth to one-third the length of the lamina,
ininutely ee narrowly and deeply canaliculate above, plicate-
rugose beneath; lamina thinly coriaceous, castaneous-nigrescent above,
brunneous beneath, ovate to oblong-elliptic, 5-8 cm. long, 2-3 cm. wide;
base rounded, rarely obtuse; apex acuminate, the acumen 5-10 mm. long;

1949 ] HU, THE GENUS ILEX IN CHINA 335

margin entire, plane, slightly revolute near the base, often obscurely serrate
near the apex; midrib slightly impressed above, puberulent or rarely gla-
brous, the lateral nerves very obscure on both surfaces. Inflorescences
cymose, solitary, axillary, only on current year’s growth, minutely puberu-
lent; flowers 4- or 5-merous. Staminate inflorescences: cymes 3-9-
flowered; peduncles 25 mm. long, the secondary axes sometimes 3 mm.
long, the pedicels 2—4 mm. long, the bracteoles 1-1.5 mm. long, lanceolate,
puberulent especially on the adaxial surfaces; calyx 1.5 mm. in diameter,
patelliform, 4- or 5-lobed, the lobes deltoid, very acute, glabrous, some-
times sparsely ciliate; corolla rotate, 3 mm. across, the petals 1.8 mm.
long, ovate, one-eighth connate at the base; stamens shorter than the
petals, the anthers ovoid; rudimentary ovary subglobose, 0.5 mm. in
diameter. Pistillate inflorescences: flowers solitary, very rarely 3-flowered
cymose, the pedicels 4—4.5 cm. long, very slender, with 2 subulate super-
median prophylla 1 mm. long; calyx 3 mm. in diameter, 4- or 5-lobed, the
lobes ciliate; corolla 5 mm. across, the petals 2 mm. long, ovate; staminodes
shorter than the petals, the sterile anthers ovate; ovary broad-conic, 2 mm.
in diameter, the stigma mammiform. Fruit globose, 7-8 mm. in diameter,
red, when dry reddish brown, smooth, the stigma prominent, subpyramidal.
Pyrenes 5 (4 ex Miquel, often 5 or 6 ex Loesener), elliptic in outline,
trigonous in cross-section, 6 mm. long, 2.5 mm. wide, smooth, unistriate
along the dorsal median line; the endocarp coriaceous.

CHINA: Launing: Dairen, Man-shu No-san Sho-kai Inc. (A); C.
ue wford (G). Anhwei: S. Chen 2630 (A); Chien-shen-hsien, C. S. Fan

& Y. Y. Li 187 (A); Wang-shan, W.C. Chong 4131 (US); R. C. Ching
3017 (A, LU), A. N. Steward 7158 (A, US); Chu-hwa-shan, R. C. Ching
COLOx GA Se CO a-Sun. 1407, ang: Tien-mo-shan, Rk. C. Ching
5189 (A). Kiangsi: Lu-shan, A. N. Steward 2643 (A); E. H. Wilson
1609 (A). Hupei (Hupeh): ve Henry 1702 (isotypes of Llex peduncu-
losa var. continentalis, A, NY), 4913 (NY), 5168 (G), 5910 (G), 6614
(1soTYPE of var. continentalis, G, US); E. H. Wilson 477 (A, US), 1305
(A), 1325 (A, NY, US), 1976 (A), 1976A (NY), 2257 (A, K, NY, US),
2698 (A, K, NY, US), 3095 (A, US), 2699 (K, NY), 2700 (K, NY),
2700B (KB). Hunan : Mo-fou-shan, Y. K. Hsiung 5898 (A). Sze-
chuan: Kiang-yu-hsien, F. T. Wang 22271 (A). Kweitchow:. Pin-ia,
J. Cavalerie 1066 (tyre of Ilex purpurea var. leveilleana, fragment and
photo, A); Yin-kiang, Y. Tsiang 7976 (NY). Kwangsi: San-chiang-
hsien, Steward & Cheo 982 (A, NY). Taiwan: R. Kanehira 21177 (A);
TBeSusueielicon (nw ).

JAPAN: K. Beattie & Y. ess 10149 (A); ug * (ISOTYPE, ae

Sci. Col. Imp. Univ. Jap. (A); U. Faurie 6895 (A); J. G. Jack in 1905
CAS Co H. Mayer in ae es ee in 1863 ee NY); Siebold
he eS VOLa Ca) Se sangent (A) s ha Uno Ga) Wilson

6024 (A), 6235 (A), bie (A), 7756 (A).

This species was first described on the basis of Japanese specimens,
collected by Buerger, as a glabrous, entire-leaved, long-pedunculate and
tetramerous species. A duplicate of the type collection is in the Gra
Herbarium. The branchlets, petioles, and peduncles are shortly puberu-

336 JOURNAL OF THE ARNOLD ARBORETUM [VOL. xxx

lent. Additional material from Japan shows that even the calyx-lobes
are sparsely ciliate (aximowicz, in 1863). Similar specimens have been
collected from many provinces in China.

This species occurs in cultivation in England and in the United States.

When Loesener differentiated two forms of Ilex pedunculosa he dis-
tinguished the continental from the Japanese form by the size of the nn
and the ciliate calyx. For f. genuina he says, “foliis tantum usque 9 ¢
longis,” and for f. continentalis, “foliis longioribus usque 12 cm. longs,”
With numerous Chinese specimens before me, including isotypes of f.
continentalts, I find no leaves longer than 9 cm. The presence or absence
of cilia and their density when present vary similarly in both Chinese and
Japanese specimens. I see no reason for distinguishing two forms here.
33a. Ilex pedunculosa var. taiwanensis, var. nov.

Frutex; foliis subcoriaceis, integerrimis, suborbiculo-oblongis, 1.5—3 cm.
longis, 1- a 6 cm. latis, basi rotundatis, apice acutis; costa supra plana et
pubescente, nervis lateralibus utrinque 6 vel 7, subtus evidentibus; petiolis
7-9 mm. longis, quam laminis 4-4 brevioribus: floribus femineis soli-
tariis, axillaribus; pedicellis 11-14 mm. longis, floribus 4-meris vel raro
calycibus 5-meris; corolla rotata, 6 mm. diametro; ovario ovoideo: stigmate
discoideo, stylo evidente.

CHINA: Taiwan: Tai-ho-ku-syi, T. Suzuki 18333 (rypr, TU).

This variety is very closely related to Ilex sugeroki var. brevipedunculata,
but the latter has plane pedicels 3-7 mm. long which are one-tenth to one-
seventh the length of the lamina. Because of the proportionally long and
canaliculate petioles, I think it is preferable to place this plant as a small-
leaved variety of Ilex pedunculosa.

34. Ilex rockii, sp. nov.
lex intricata sensu Hand.-Mzt. Symb. Sin. 7: 658. 1933, pro parte; non
Hook. f.

Frutex; ramulis pubescentibus; foliis obovato-oblongis, 1-2.5 cm. longis,
6-14 mm. latis, basi cuneatis, apice rotundatis obtusis vel raro subacutis,
margine crenulato-serratis; costa supra elevata, pubescente, subtus obscura,
glabra, nervis lateralibus utrinque 2—4, obscuris; inflorescentiis solitariis,
raro subfasciculatis; pedicellis 2-4 mm. longis; floribus rubris, 5-meris, raro
4- vel 6-meris; calycibus patelliformibus, 3— 4mm. diametro, alabiis, ciliatis:
corolla rotata, 7 mm. diametro; petalis basi connatis, lobis 3 mm. longis;
staminibus 5; staminodiis quam petalis brevioribus; ovario conico-ovoideo;
stigmate capitato: fructibus globosis, 5—7 mm. diametro: stylo evidente;
pyrenis 5, 5 mm. longis, 3 mm. latis, laevibus, estriatis et esulcatis; endo-
carpio coriaceo.

A small shrub 1-2 m. high with distinctly pubescent branchlets, obovate-
oblong serrate leaves, cuneate at the base, rounded, obtuse or rarely sub-
acute at the apex, solitary pistillate flowers at the base of the current year’s
growth, and very short pedicels.

Branchlets stout, when dry castaneous or smoky; fourth year’s growth

1949] HU, THE GENUS ILEX IN CHINA 337

4 mm. in diameter, longitudinally rimulose, the lenticels lacking; third
year's growth 2-3 mm. in diameter, smooth, densely puberulous; second
year’s growth tomentose, plicate-rugose, 1.5-2 mm. in diameter; current
year’s growth 1—1.5 mm. in diameter, castaneous, hirsute. Leaves occurring
even on the fourth year’s growth, 2-5 mm. apart; stipules minute, callose,
deltoid, often persistent; petioles flattened, 2 mm. long, one-tenth to one-
fifth the length of the lamina, plane above, pilose; lamina coriaceous,
olivaceous-brunneous, shiny above, paler beneath, epunctate, obovate-
oblong, 10-25 mm. long, 6-14 mm. wide; base cuneate; apex rounded,
obtuse or rarely subacute; margins crenulate-serrate, the teeth prominent,
3-6 on each side, the lower half entire; midrib prominent, elevated, hirsute
above, obscure and glabrous beneath, the lateral nerves 2—4 pairs, obscure
above and beneath. Pistillate inflorescence: flowers solitary, axillary to
scales at the base of the current year’s growth; bracts coriaceous, caducous,
ovate, 1—-2.5 mm. long, ciliate; pedicels short and thick, 2-4 mm. long,
glabrous or sparsely puberulous; prophylla none; flowers red (ex Rock),
rather large for the genus, 7 mm. across, 5-merous or rarely 4- or 6-merous;
calyx patelliform, 3-4 mm. in diameter, glabrous, rugose, the lobes nor-
mally 5, imbricate, deltoid, 1.5 mm. long and wide, rounded, obtuse or
acute, erose and ciliate; corolla rotate, the petals ovate-oblong, 3 mm.
long, 1.5 mm. wide, one-fifth connate at the base; staminodes one-half
the length of the petals, the sterile anthers very minute, cordiform; ovary
large, 2.5 mm. long and wide at the base, conic- Bod stigma capitate,
distinctly 5-lobed. Staminate nfloresconcet fasciculate, the stamens 5;
only buds seen. Fruits globose, 5-7 mm. in diameter, the persistent stigma
elevated, discoid, the style short, pedicels 2-4 mm. long. Pyrenes 5
(sometimes 3 or 4), 5 mm. long, 3 mm. wide, smooth, estriate, esulcate,
the endocarp coriaceous.

CHINA: Yunnan: Mekong-Salwin Divide, Handel- eee 9617
(A); Kun-shan, Solola, J. F. Rock 22299 (type, A; isotypes NY, US), 22651
(A); C. W. Wang 66727 (A); Salwin-Kiukiang Divide, T. T. Yu 20228
CA) 202412 CA) 2237507).

The description of the fruit is drawn from Yu 20241.

Ilex rockii is a low alpine shrub occurring at altitudes of 4100-4330 m.
in the high mountains of northwestern Yunnan. The red pistillate and pink
staminate flowers appear in May and June, and its fruits turn red in
September

The plant is closely related to Zlex crenata Thunb., Jlex luzonica Rolfe,
Ilex yunnanensis Franch., and Jlex intricata Hook. f. Superficially its
leaves suggest those of /lex crenata. That species, however, differs in
having punctate lower leaf-surfaces, impressed midribs, 4-merous flowers,
longer pedicels and striate pyrenes. Jlex rockii resembles /lex luzonica in
its 5-merous flowers, but differs in its lack of greatly elevated leaf-scars and
punctate leaves with retuse tips. In the thick indumentum of the branch-
lets, elevated hairy midribs, flattened petioles, and smooth pyrenes, it
resembles flex yvunnanensis, which differs in having 4-merous flowers, ovate
or ovate-elliptic leaves with slightly produced acute apices, longer pedicels,

338 JOURNAL OF THE ARNOLD ARBORETUM [vVOL. Xxx

and cymose staminate inflorescences. In its short pedicels and 5-merous
flowers, it resembles /lex intricata, which differs in its completely glabrous
and peculiarly warty branchlets, impressed venation, and striate-sulcate
pyrenes.

35. Ilex yunnanensis Franch. Pl. Delav. 2: 128. 1889; lLoes. in Nov.
Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78: 131 (Monog. Aquif. 1:
131). 1901, et in Sarg. Pl. Wils. 1: 76. 1911; Rehd. & Wils. in Sarg.
Pl. Wils. 3: 425. 1917; Rehd. in Jour. Arnold Arb. 7: 199. 1926;
Anon. in Notes Bot. Gard. Edinb. 17: 29, 38, 48. 1929, 262, 381.
1930; Hand.-Mzt. Symb. Sin. 7: 655. 1933; — Ill. Man. Gita,

Trees 656. 1937; S. Y. Hu in Ic. Pl. Omei. 2: pl. 158. 1946.
Ilex sugeroki Maxim. forma brevipedunculata sensu ee Le. 134. 1901;
on (Maxim.) Hu.

A ne or a ‘small tree up to 12 m. high with ferrugineous-villose
branchlets, small coriaceous ovate or ovate-lanceolate leaves, elevated and
pubescent midribs, aristate-crenulate-serrate margins, cymose staminate
inflorescences, solitary pistillate flowers on the current year’s growth,
4-merous flowers, globose fruits, elevated discoid stigmas, and smooth
coriaceous pyre

Branchlets a fourth year’s growth 3-4 mm. in diameter, cinereous,
pubescent, smooth, minutely and longitudinally rimose, the lenticels lack-
ing; second year’s growth 1-2 mm. in diameter, ferrugineous-villose; cur-
rent year’s growth | mm. in diameter, densely villose, hairs golden, the
terminal buds well developed, ovoid, acute, ferrugineous-villose. Leaves
occurring even on the fourth year’s growth, 3-10 mm. apart; stipules
callose, buried in the indumentum, persistent; petioles 2—5 mm. long, one-
fifteenth to one-seventh the length of the lamina, pubescent, slightly

brunneous, usually shiny above, paler beneath, when young villose,
especially near the base and along the margin, becoming glabrescent, ovate
or ovate-lanceolate, rarely elliptic, 2-3.5 cm. long, 1-2 cm. wide; base
rounded, obtuse or rarely narrowly subacute; apex acute, mucronate,
sometimes ciliate; margin recurved, crenulate-serrate, teeth often aristate;
midrib elevated and pubescent above, plane or slightly elevated, usually
villose beneath, the lateral nerves obscure. Inflorescences cymose, solitary,
axillary to leaves or scales of current year’s growth, pubescent. Staminate
inflorescences: cymes 1-3, usually 3-flowered, peduncles 8-14 mm. long,
the pedicels 2-4 mm. long; flowers small, normally white, those from high
altitudes pink or red, dmicrous: calyx subpatelliform, 2 mm. across, deeply
4-lobed, the lobes deltoid, acute or obtuse, ciliate; corolla rotate, the
petals ovate, obtuse, 2 mm. long, 1.5 mm. wide, one-fourth connate at the
base; stamens shorter than the petals, the anthers broadly ovoid; rudi-
mentary ovary conic, the apex obtuse. Pistillate inflorescences: flowers
solitary, very rarely 2- or 3-flowered cymose, the pedicels 8-14 mm. long
with 1 or 2 supermedian prophylla; calyx as in the staminate flowers;
corolla suberect, the petals 1.5 mm. long, one-third connate at the base;

1949] HU, THE GENUS ILEX IN CHINA 339

staminodes one-half the length of the petals, the sterile anthers sagittate;
ovary globose, 1 mm. in diameter, with 4 longitudinal grooves, style
distinct, 0.5 mm. long, the stigma ficeor: 4-lobed. Fruits red, globose,
5—6 mm. in diameter, the persistent calyx explanate, quadrangular in
outline, ciliate, the stigma elevated, discoid. Pyrenes 4, oblong-oval in
outline, eubieemeular in Bee cian: 5 mm. long, 3 mm. wide, estriate,
esulcate, the endocarp smooth, thickly coriaceous.

CLUINAS Szechwan -~ Mt: Onmel, Ho CC how 850 (Ane 12c0r (Kh):
W. P. Fang 11609 (Sz), 16912 (Sz), 23387 (A); A. Henry 7144 (NY);
SCS on Ree Cliong s40.CA )? O-pien, Cy te Sumit LeGsz ). 15s (Sz) 5
without precise locality, C. W. Wao STT#t CSS eels GSS) 4254-088 ¥,
4328 (SS); Wen-chuan, EE. H. Wilson 1024 (A, US), 3092 (US), 3094
(A, US); Kwan-hsien, E. H. Wilson 3093 (A, US). Sikang: Pao-shien
(Muspins jek eh. Chi 3720 (SS), 3097 (SS)=) Lucting(ha-chien-li),
WP. Fang 3755 (A); BE. AH. Wilson 4135 (A); Tien-tsian, Y. S. Lin 1815
(A). Yund-shi Mountains (southez sea ete) saniet eee Rock 23151 (A),
23094 (A, NY) Munian-: Li-kiane, R. C. Ching 200374.CA) < S. Chune-
tien, K. M. Feng 3254 (A): western ee G. Forrest 4193 (K), 10247
(CA) 13932 CUS a CA), 19074 (A), 19459 (A, US): Handel-Maz-
setit 8497 (A, US), 9141 (A, US); J. F. Rock 4196 (A, US), 8945 (A,
NY, US)s92853 “CA, US)5 Weirsi,.A. T. Tsat37995) (A), 59809 (A), S99T!
(A), 63045 (A): C. W. Wang 63544 (A), 64002 (A), 64061 (A), 67868
(A); A-tun-tse, T. T. Yu 7864 (A); Ne-wah-lung, T. T. Yu 19230 (A);
Tseh-chune, 7.7, Vu l90is (A).

UPPER BURMA: Nam Tamai Valley, Kingdon Ward 13296 (B).

Ilex yunnanensis is a large hardy shrub common in the high mountains
of western China. There it grows at altitudes from 1500 to 3100 m.
The fragrant flowers appear in June. They are usually white, but on
the snow range of northwestern Yunnan they are said to be pink or even
red in color. Several varieties of the species have been proposed. These
can be distinguished by the following key.

A. Leaves ovate or oblong, the apex obtuse, the margin crenate: staminate
rowers 4-6-merous. (Shensi, Hupei and N. E. Kweichow)........
Re omen tr iat uth tena Wawa ao welded Niue poeea a ee Sane, MVC WOME OS.

AA. Leaves ovate, oblong or lanceolate, the apex acute, the margin serrate,
the teeth often aristate: flowers always 4-merous.

B. Fruiting pedicels less than 7 mm. long, the fruit nodding. (Borde

of Szechuan, Sikang and Yunnan)... .3 var. eas
BB. Fruiting pedicels 8-15 mm. long, the fruit erect

C. Calyx ciliate.

Leaves entire or 1—3-toothed on each side (teeth spinose ),

thickly coriaceous, 2-3 cm. long, 1-1. - cm. wide. “un-

TUES) ita orcs 2a, Ree gear eae ear a oer, 5c; . paucidentata.

DD. nae serrate, the teeth aristate; coriaceous, 1.2-2 cm.

longs “OS =0)/cunswicles eat, atl) ea ares ae we Sede

ich asd ingen ine we PREM rae cs 35d. var. parvifolia.

CGr Calysceci lates (enemy tei eee eee Oey, VAI CeCIOld,

340 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

35a. Ilex yunnanensis var. gentilis (Franch.) Loes. ex Diels in Bot.
Jahrb. 29: 435. 1900, nom. nud. et in Nov. Act. Acad. Caes. Leop.-
Carol. Nat. Cur. 78: 132 (Monog. Aquif. 1: 132). 1901, descr., et
in Sarg. Pl. Wils. 3: 425. 1917; Chen, Ill. Man. Chin. Trees 656.
1937
flex gentilis Franch. ex Loes. in Nov. Act. Ac Caes. Leop.-Carol. Nat.
Cur. 78: 133 (Monog, Aquif. 1: 133). 1901,

Leaves thin-coriaceous, ovate or oblong, the apex | ine or rarely sub-
acute, the base round, the margin crenate; staminate inflorescence cymose,
the flowers 4-, rarely 5- or 6-merous

CHINA: Shensi: Tai-pei-shan, WV. Purdon 1114 (A). Hupei
(Hupeh): W. Y. Chun 4214 (A); 4. Henry 6901 (A, US); E. H. Wilson
(Veitch Exp.) 2344 (A, NY, US), 4458 (A). So ean : Mt. Omei,
. S. Lin 1195 (A); Wen-chuan, E. H. Wilson 4195 (A, US). Taiwan:
Arisan, E. H. Wilson 9781 (A, US).

Several flowers (staminate) of Wilson 2344 are 5- or 6-merous. In
this respect the plant more or less resembles /lex sugeroki Maxim. var.
brevipedunculata (Maxim.) S. Y. Hu. Because of the leaf-form and size,
the indumentum and the geographic range, I think it is better to place it
as a variety of //ex yunnanensis rather than to transfer it to lex sugeroki
Maxim. Farges 129, from Tschen-kéou-tin, Szechuan, is within the range
of the distribution of this variety, and I suspect that it belongs here.

I have examined several specimens collected by Prof. A. Rehder in the
Arnold Arboretum from plants grown from seeds obtained by Wilson in
Hupei Province. The pubescence on the twigs of these cultivated forms is
shorter and thinner than that found on the specimens of the wild Jlex
yunnanensis. The venation of the leaves is more prominent than on the
normal form. In appearance it looks more like /lex crenata Thunb. than
flex yunnanensis. Perhaps this is a demonstration of how environmental
factors can affect the appearance and characters of the species.

i

35b. Lex yunnanensis var. brevipedunculata S. Y. Hu in Ic. Pl. Omei.
2: pl. 158. 1946

Leaves oval, ovate or elliptic, 2-3.5 cm. long, 1-2 cm. wide; fruiting
pedicels 3—6, rarely up to 7 mm. long, the fruit nodding.

ae Kweichow: Fan-chin-shan, a ot & Cheo 842
ay US). Szechuan: Mt. Omei, T. C. Lee 6449 (Sz); O-pien,

WW, a 2834 (SS). Yunnan: Lap-ping it. T. Tsat 54016 (A);
Shang: pa, H. T. Tsai 54484 (A); Wei-si, H. C Tsai 57918 (A), aly
(A); C. W. Wang 63808 (A), 67842 (A), 68710 (A), 70457 (A): Tal
C. W. Wang 63344 (A), 63345 (A),

This variety occurs only at higher altitudes, 2600-3600 m. Its purplish
red flowers appear in May and June. Its fruits are red by November.

35c. Dex yunnanensis var. paucidentata, var. nov.
Arbor, 5 m. alta; foliis crasse coriaceis, ovato-lanceolatis, 2-3.5 cm.
longis, 1-1.5 cm. latis, supra nitidis, basi rotundis vel obtusis, apice acutis

1949 | HU, THE GENUS ILEX IN CHINA 341

spinosisque, margine integris vel 1—3-denticulato-serrulatis; pedicellis 12
mm. longis; fructibus globosis, 4-5 mm. diametro, calycibus ciliatis, stylis
perspicuis, stigmate discoideo, 4-lobato.
HINA: Yunnan: Wei-si, C. HW’. Vang 67855 (rypz, A).
The leaves of this variety are much thicker and more rigid than the
typical /lex yunnanensis. The leaf-tips terminate in a weak spine. More
material may show it to be worthy of specific rank.

35d. Ilex yunnanensis var. parvifolia (Hayata), comb. nov.
flex parvifolia Hayata in Jour. Coll. Sti. Univ. Tokyo 30: 57. 1911, et Ic.
Pl. Form. 1: 134, fig. 19. 1911; Kanehira, Form. Trees 378, fig. 336.
1936. Syn. nov.

Leaves coriaceous, oblong-lanceolate, small, 11-20 mm. long, 6 mm.
wide, serrate, teeth aristate.

NA: Taiwan: Arisan, Faurie 914 oe 1ST ADE EL bP al con
9470 GAS US); 20°07 CAS) Kamiyama CEU.

In the nature and density of the indumentum, the texture and form of
the leaves, the venation, the inflorescences, the fruits, and the pyrenes,
the Taiwan and Yunnan specimens are alike. The only difference I have
noted between them is that the Taiwan form has somewhat smaller leaves.
I judge it to be worthy of varietal rank only.
35e. [lex yunnanensis var. eciliata, var. nov.

Frutex, 1-3 m. altus; foliis tenuiter coriaceis, ellipticis, raro ovatis,
1—2.8 cm. longis, 4-12 mm. latis, basi acutis, raro obtusis, apice acutis vel
obtusis; floribus 4—6-meris; eh cibu: batelliforminne 2- 3 mm. diametro,
glabris, lobis 4—6, raro usque ad 9, deltoideis, acutis, eciliatis.

CHINA: Szechuan: Wen-chuan, £. H. Wilson 3092 (A); Hong-
ya, Wa-wu-shan, C. WW. Yao 2379 (SS), 2377 (SS), Bo/eetos i Wt Omel,
S.N. Hsu 16 (3S)).#. T. Wang 23390 (A); O-pien, CW Yao 4328 (rven,
SS). Sikang: Pao-shien (Mu-ping), K. L. Chu 2960 (SS), 2957 (SS);
MienechuatCw ee Wee (eco CSs.. UN), 12254 (SSi.U Nees be Gnu. 2628

(SS).

The local name reported is “Shui-cha-tze,” which means “water tea”
or “the tea growing by the water.” The leaves are collected by natives of
the Sino-Tibetan Border and used as a substitute for tea. It is of interest
to note that species of //ex belonging to the same section have been used in
a similar way by Indians of both North and South America.

Ilex yunnanensis is closely related to Jlex sugeroki Maxim. var.
brevipedunculata (Maxim.) S. Y. Hu. The typical forms of the former
have densely ferrugineous-villose branchlets, 4-merous flowers, 4 pyrenes,
aristately crenulate-serrate leaves and acute apices. The variety of /lex
sugeroki has puberulous branchlets, 4-6-merous flowers, 5 or 6 pyrenes,
and leaves coarsely serrate only on the distal half and obtuse at the apex.
The former occurs in western China and the latter is found in Japan. In
Taiwan the two elements meet. Sometimes it is difficult to decide to which
species the Taiwan specimens belong.

342 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

36. Ilex sugeroki Maxim. in Mém. Acad. Sci. St. Pétersb. VII, 29 (3):
35, pl. 1, fig. 7 e. 1881; Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 133 (Monog. Aquif. 1: 133). 1901; Makino, Ill. FI.
Nip. 367, fig. 1101. 1940.

flex sugeroki forma Seeman Maxim. lc. 36. 1881; Loes. lc.
134. 1901 yn. no

Ilex sugerokt subsp. longipedencntt (Maxim.) Makino in Bot. Mag.
(Tokyo) 27: 78. 1913 . nov

An evergreen shrub up to aA m. high with puberulous branchlets, coria-
ceous ovate or ovate-elliptic leaves with the margin serrate on the apical
half, solitary axillary inflorescences, long-pedicellate (2—3.5 cm.) fruits,
elevated discoid stigmas, and 4—6 smooth estriate esulcate pyrenes.

Branchlets puberulous, brunneous, older growth roughened by many
elevated leaf-scars; third and second year’s growth 2 mm. in diameter,
the lenticels lacking; current year’s growth 2 mm. in diameter, angular,
pubescent, the terminal buds well developed, ovoid, 3 mm. long, 2 m
wide, acute, with sparsely puberulous and ciliate scales. Leaves occurring
even on the fourth year’s growth, 5-10 mm. apart; stipules minute, callose,
deltoid, persistent; petioles 4-7 mm. long, one-tenth to one-seventh the

ceous, shiny above, opaque beneath, ovate or ovate-elliptic, 2-4 cm. long,
1-2.8 cm. wide; base rounded or obtuse; apex acute or shortly acuminate,
the acumen 2—5 mm. long; margin serrate only on the distal half, midrib
elevated and puberulous above, plane and glabrous beneath, the lateral
nerves 6 or 7 pairs, obscure on both surfaces. Inflorescences solitary,
axillary, on current year’s growth only, puberulous; flowers 4-, 5-, or 6-
merous. Staminate inflorescences: cymes 2~—7-flowered, the peduncles
12-18 mm. long, the secondary axis 1-2 mm. long, the pedicels 2-5 mm.
long, the bracteoles linear, 2 mm. long, puberulous; calyx patelliform,
.5 mm. across, deeply 4-6-lobed, the lobes deltoid, densely long-ciliate;
corolla white, rotate, 5-6 mm. across, the petals oblong, eciliate, one- -sixth
connate at the base; stamens shorter than the petals, the anthers oblong;
rudimentary ovary ovoid, acute or shortly rostellate. Pistillate inflores-
cences: flowers solitary, the pedicels slender, erect, 3.5 cm. long, with 1
or 2 prophylla above the middle; calyx and corolla as in the staminate
flowers; staminodes three-fourths the length of the petals, the sterile
anthers oblong-sagittate, obtuse at the apex; ovary ovoid, 1.5 mm. long,
glabrous, stigma elevated, discoid, plane and explanate. Fruits globose,
5-6 mm. in diameter, persistent calyx explanate, 3 mm. across, ciliate,
stigma elevated, discoid. Pyrenes 4-6, oblong in outline, 3.5 mm. long,
2 mm. wide, neither striate nor sulcate, the ends obtuse, endoeanp smooth,
coriaceous.
APAN: Fujiyama region, H. Dorsett & W. J. Morse 384 (A).
Hondo: Mino, K. Shiota 77 as 4363 (A), 6785 (A); E. H. Wilson
(A)

Ilex sugeroki is endemic to the Fujiyama region of Japan. There it

1949] HU, THE GENUS ILEX IN CHINA 343

grows as a shrub at altitudes of 1000-1500 m. Its white flowers appear
in June. The fruit turns red in October and remains on the plants until
the flowers appear the following June.

Maximowicz named two forms, distinguishing them by the length of
the peduncles and the position of the fruits. He cited no specimens under
these forms. Since the plant he illustrated as /lex sugeroki is the large-
leaved, long-pedicellate form, I am accepting it as typical of the species in
the strict sense. According to this illustration, the leaves of the species are
essentially ovate with a rounded base. However, most of our specimens
have neither ovate leaves nor rounded leaf-bases. The fruiting pedicels of
these latter are less than 2 cm. long. I interpret them as representing
Ilex sugeroki Maxim. var. brevipedunculata (Maxim.) S. Y. Hu.
36a. Ilex sugeroki var. brevipedunculata (Maxim.), comb. nov.

Ilex sugeroki forma brevipedunculata Maxim. in Mém. Acad. Sci. St.
Pétersb. VII, 29(3): 36, pl. 7, fig. d. 1881; Loes. in Nov. Act. Acad.
Caes. Leop.-Carol. Nat. Cur. 78: 134 (Monog. Aquif. 1: 134). 1901.
Syn. nov.

Ilex sugeroki subsp. brevipedunculata Makino in Bot. Mag. (Tokyo) ZL:
8. 1913. Syn. nov.

Ilex sugeroki sensu Hara in Bot. Mag. (Tokyo) 50: 188. 1936, non Maxim.

Ilex taisanensis Hayata in Jour. Coll. Sci. Imp. Univ. Tokyo 30: 57.
1911, et Ic. Pl. Form. 1: 134, 1911; Kanehira, Form. Trees 382, fig.
339. 1936, Syn. nov.

A much-branched shrub up to 3 m. high with puberulous branchlets;
leaves elliptic or ovate-elliptic, 2-3.5 cm. long, 1-1.5 cm. wide, cuneate or
obtuse at the base, acute or obtuse at the apex, the margin serrate only in
the apical half, the midrib elevated and puberulous above, plane or slightly
elevated underneath, the petiole 3-8 mm. long, one-tenth to one-fourth
the length of the lamina; inflorescences solitary, axillary, staminate 3-
flowered, cymose, the peduncles 10 mm. long, the pedicels 1-5 mm. long,
the prophyllum none or 1 near the calyx, puberulous; pistillate flowers
solitary, pedicels 12-15 mm. long, prophylla 1 or 2, supermedian; calyx
deeply 4-6-lobed, long-ciliate; fruit globose, 8 mm. in diameter, the
stigma elevated, discoid.

CHINA: Taiwan: Bioritsu, 7. Kawakami & U. Mori in 1908 (TYPE
of Ilex taisanensis, photo and fragment, A; fragment, TU); Taihoku, S.
Suzuki in Aug. 1928 (A, TU); Mt. Taiheizan, S. Suzuki in Aug. 1928 (TU).

JAPAN: Mt. Hakusan, (Herb. Sci. Coll. Imp. Univ. Jap.) (A); Shiri-
beshi, T. Ishikawa in 1893 (A); Oshima, K. Miyabe & Y. Fukubucht in
1890 (A); Mt. Hakkoda, C. S Sargent in 1892 (A); Mt. Nasu, K. Sakurat
in 1909 (A); Echizen, K. Shiota 3453 (A); Mino, K. Shiota 5932 (A),
6393 (A); E. H. Wilson 7100 (A), 7183 (A), 7285 (A), 7632 (A), 7723
(A); Kitamifuji, Hokkaido, K. Uno 1581 (A), 1698 (A).

The color of the dry Taiwan specimen cited above is brunneous-
nigrescent. This is probably due to the methods of preparation. At
sight, specimens of this variety might easily be mistaken for [lex yun-

344 JOURNAL OF THE ARNOLD ARBORETUM B20) Wee. 6

nanensis Franch., but the latter has densely ferrugineous-villose branchlets
and its leaves are serrate or crenulate to near the rounded or obtuse base,
while /lex sugeroki var. brevipedunculata has puberulous branchlets and
leaves which are serrate only on the upper half.

To BE CONTINUED

FRONTISPIECE Jour. ARNOLD ARB. VoL. XXX

JOHN GEORGE JACK

JOURNAL

OF THE

ARNOLD ARBORETUM

VoL. XXX OCTOBER 1949 NUMBER 4

JOHN GEORGE JACK. 1861-1949.
With portrait.

On May 20, John George Jack, Assistant Professor of Dendrology,
Emeritus, at the Arnold Arboretum, died at the age of eighty-eight years
at his home, “Folly Farm,” in East Walpole. He had been confined
to his bed since August of the preceding year, having broken his hip while
working in his orchard.

John George Jack was born in Chateauguay, Quebec, Canada, on
April 15, 1861. His mother, Mrs. Annie L. Jack, originally a school-
teacher, was a well-known horticulturist. She produced a series of articles
under the title of “Garden Talks” and wrote a handbook on ‘The Cana-
dian Garden.” Undoubtedly his mother’s influence was responsible for
Professor Jack’s early interest in horticulture. Although Professor Jack’s
formal education did not extend beyond high school he acquired a thorough
knowledge of plants from practical experience and long study. He was
outstanding as a field worker in detecting variations and species-hybrids.

In 1886, Professor Jack first came to the Arnold Arboretum to work
and study under Professor C. S. Sargent and in 1891 he was appointed
Lecturer at the Arnold Arboretum. He served as Instructor in Forestry
at Harvard from 1903 to 1908, and as Lecturer in Forestry at the Massa-
chusetts Institute of Technology from 1899 to 1908. He was made an
Assistant Professor of Dendrology in 1908. Although Professor Jack
spent most of his time at the Arnold Arboretum where he was a staff
member from 1891 until he was retired in 1935, he was also on the staff
both of the Bussey Institution and the Harvard Forest.

Even though most of Professor Jack’s time was spent on the Arnold
Arboretum grounds checking identifications of the trees and shrubs, lec-
turing to field classes, and supervising the plantings, he made many
collecting trips to various parts of the world. The early trips were financed
by funds obtained by Professor Jack himself from outside sources since
at first the Arnold Arboretum had little money for exploration and Pro-
fessor Jack’s salary was only $500.00 per year. In 1891 he visited the
botanic gardens and nurseries in France, Germany, Italy, Denmark and

346 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

England and in 1904 with Professor Alfred Rehder he made extensive
collections in the western United States and Canada. In 1905 he went to
Japan, Korea and China, where he studied both native and cultivated
ornamental plants sending seeds and cuttings back to the Arnold Arbo-
retum. Nearly every year from 1926 to 1936 Professor Jack went to
the Atkins Botanical Garden in Cuba for a few months where he worked
with the collections in the garden and collected native species in the
vicinity of Soledad.

In 1907 Professor Jack married Cerise Emily Agnes Carmen, daughter
of Elbert S. Carmen, who published the “Rural New Yorker.” They
adopted two children. Mrs. Jack died in 1935. Their daughter, Betty
Wirth, and her husband lived with Professor Jack in his later years at
“Folly Farm” in East Walpole.

Professor Jack was especially interested in trees, but like all old-time
naturalists his interests covered the entire range of biology. His knowledge
of plants, his spontaneous enthusiasm and sincerity, made him an out-
standing teacher of field classes. He made many friends for himself and
for the Arnold Arboretum.

Taxonomists are a hardy race, but few could keep up with Professor
Jack. While spending a month with him at the Atkins Garden in 1936,
I discovered that although he was 75 years old Professor Jack was up at
6 A.M., worked all day, often traveling into the surrounding country on
horseback, and continued to work until 11 or 12 o’clock at night. Later
I visited him at his farm on his 85th birthday where I found him spraying
his orchard. He complained that he couldn’t work after supper as formerly
when he was younger.

Professor Jack published several hundred papers beginning in 1888.
Most of these were notes on various plants grown in the Arnold Arboretum
and descriptions of field work. His earlier contributions were published
in “Garden and Forest” and in various horticultural journals. Many of
his later publications appeared in the Bulletin of Popular Information
published by the Arnold Arboretum.

In recognition of his work in taxonomy the following species were named
for Professor Jack: —

Sinojackia Hu, a new genus with two species from China.
Alnus Jackii Hu (= A. trabeculosa Hand.-Mazz.).
Amelasorbus Jackii Rehd.

Betula Jacktu Schneid.

Crataegus Jackii Sarg.

Juniperus communis var. Jackii Rehd.

« Populus Jacki Sarg.

x Quercus Jackiana Schneid.

x

x

x X

Rosa Maximowiczii var. Jackii (Rehd.) Rehd. (R. Jackii Rehd.).
Sorbaronia Jackii Rehd. (Pyrus Jackii [Rehd.| Fernald).
Viburnum Jackii Rehd.

The hybrids (>) were discovered by Professor Jack.

1949] JOHN GEORGE JACK 347

A SELECTED BIBLIOGRAPHY OF THE WRITINGS OF JOHN GEORGE JACK.

1888-1889. Notes from the Arnold Arboretum. (Garden and Forest, vol.
1 eZ

1890-1892. Notes on shrubs. (Garden and Forest, vol. 3 & 5.)
1891-1893. Notes of a summer journey in Europe, I-X XIV. (Garden and
-)

Forest, vol.
1891. Notes on some hardy wild roses, I-V. (Garden and Forest,
vol. 4

1894. Notes on some injurious insects. (Trans. Mass. Hort. Soc. 1:
1-20.)

1894. Notes on trees and shrubs. (Garden and Forest, vol. 7.)

— Native trees and shrubs about Montreal, Canada, I-V. (Garden
and Forest, vol. 7

1897. The fructification of Juniperus. (Bot. Gaz. 18: 369-375.)

Notes on the Eastern American Spruces. (Garden and Forest,

vol.

1900. The Pikes Peak, Plum Creek, and South Platte aie reserves.
(20th Annual Rept. U. S. Geol. Survey 1898-99: pt. 5.)

1904. The Arnold ae tum, a lecture delivered before ae Massa-
chusetts Horticultural Society. (Trans. Mass. Hort. Soc. 1:
59-75.)

1911. Trees and other woody plants found in the Harvard Forest,
Petersham, Massachusetts. (Bull. Harvard Forestry Club

1: 10-26.)
1930. Planting trees in autumn. (Bull. Pop. Inf. ser. 3, 4: 65-68.
illustr.

1931. Forsythias. (Bull. Pop. Inf. ser. 3, 5: 9-12. illustr.)

— Flowering cherries. (Bull. Pop. Inf. ser. 3, 5: 13-16. illustr.)

1935. Winter injuries among trees and shrubs. (Scientific Monthly,
338.)

1936. Arboreta, old and new. (Scientific Monthly, 41: 541-550.)

Professor Jack was a member of the Society of Poresters Botanical
Society of America, Massachusetts Horticultural Society, American Acad-
emy, Boston Society of Natural History, and Deutsche Dendrologische
Gesellschaft.

With the passing of Professor Jack, the era of the old-time naturalist
is drawing to a close. Largely self taught, with enthusiastic interest in all
phases of nature, and with apparently unlimited energy, these early bot-
anists made the contact between botanical science and the horticultural
public which led to the establishment and maintenance of so many of our
botanic gardens and parks.

KARL SAX

348 JOURNAL OF THE ARNOLD ARBORETUM [VoL. xxx

THE GENUS ILEX IN CHINA, II

SHIU-YING Hu

Continued from page 344

SECTION VII. AQUIFOLIUM Gray

Ilex § Aquifolium Gray, Man. Bot. N. U. S. 276. 1848; Maxim. in
Mém. Acad. Sci. St. Pétersb. VII. 29(3): 26. 1881; Rehd. Bibliog.
Cult. Trees Shrubs 399. 1949,

Ilex subgen, Euilex Loes. series C. a (Maxim.) Loes. in Verh.
Bot. Ver. Brand. 33: 26. 1891, in Engler & Prantl, Nat. Pflanzenfam.
Nachtr. 219. 1897, et in Nov. ren Acad, Caes. Teaan -Carol. Nat. Cur.
78: 227 (Monog. Aquif. 1: 227). 1901

Evergreen trees or shrubs with bscieilats inflorescences in the axil of

the leaves of the second year’s or even older growth, the pyrenes 4, rarely

less, with stony or woody striate and sulcate or rugose and pitted endo-
carps.

Forty-seven species in five series occur in China. Most of the species

are locally endemic. The range of the five series is as illustrated in

figure 8.

KEY TO THE SERIES

A. Leaves spinose or entire with the apex ending in a spine (except /lex

hylonomia ).
B. Pyrenes stony, ve aad wrinkled and pitted; fruit always con-
TED ED EC CSAC: am DN'2 tc) |<< ne ae meer a meries 1. st aad
BB. Pyrenes woody, aay striate; fruit containing 2 pyrenes....
Cigtwiutuddeearstietaw er ttawee eames sees wer ee ee penne

AA. Leaves entire, serrate or crenate, those from the mature plant never

B. Pyrenes irregularly wrinkled and pitted, the endocarp stony; fruit

8-12 mm. in diameter; stigma navel-like, rarely discoid.....

oe Lee Te eee eee eee Le ee eee ee ee erics.\ 0: Deaueinloiae,

BB. Pyrenes palmately striate and sulcate; fruit 4-6 rarely up to 8 mm.
in diameter; stigma discoid, capitate, rarely navel-like.

C. Leaves chartaceous or subcoriaceous, when dry brunneous,
with impressed veinlets on the upper a fruiting pedi-
eels 427 mm, lONP cass ds bec eens aes Series 4. Hookerianae.

CC. Leaves thick coriaceous, coriaceous, sore ea the
veinlets obscure or platie above, never impressed; fruiting
pedicels 24 mm, long.................4. Series 5. Repandae.

1949] HU, THE GENUS ILEX IN CHINA, I 349

Series 1. AQUIFOLIOIDES (Loes.), stat. nov.

Ilex subgen. IIT. Euilex Loes. ser. C. Aquifolium (Gray) Loes. sect. 2.
Aqui folioides Loes. in Engler & Prantl, Nat. Pflanzenfam. Nachtr. 219.
1897, et in Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78: 244
(Monog. Aquif. 1: 244). 1901.

Trees or shrubs with spinose or coarsely serrate leaves, fasciculate
inflorescences, the individual branches of the staminate fascicles 1-3-
flowered, those of the pistillate fascicles uniflorous, the flowers 4-merous,
the fruiting pedicels 2-12 mm. long, the fruit large, 6-12 mm. in diameter,
the pyrenes wrinkled and pitted, the dorsal surfaces convex, and the
endocarp stony.

“Aquifolioides
Dipyrenae

—Denticu/atae
Sapre Hookerranae

~~ Re pandae

Fic. 8. Geographic distribution of the Chinese representatives of the five
series of the section Aquifolium.

KEY TO THE SPECIES

A. Fruiting pedicel s 2 mm. long; leaves oblong-elliptic -or elliptic-lanceo-

late, the margin spinosely dentate or coarsely serrate
B. sp ies branches of the staminate fascicles 3-flowered; leaves
6-12.5 cm. long, 2.4-4.5 cm. wide, the margin coarsely serrate,
ere the teeth ending in weak spines; fruit 10-12 mm. in
diameter with a discoid or mammiform stigma. (Szechuan, Kwei-
CHOW ANGPIVAN OG it hat a4 veww y dade a eoete 37. I. hylonoma.
BB. Individual branches of the staminate fascicles uniflorous; leaves
6-7 cm. long, 1.5-2.8 cm. wide, the margin 4-10 spinose-dentate ;
fruit 6-7 mm. in diameter, with a thin-discoid age een
S7echMameDOrden | pays eres sit wo oti dea tl oa eee 38. 1. ochinensts.

350 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

AA. Fruiting pedicels 5-18 mm. long; leaves quadrangular-oblong, rarely
ovate, the margin entire or sinuate with 1-3 rigid spines on each side.
3¢

(Ease and: Central (Chine oceans thenea en ewaedooisees . 1. cornuta.
37. Ulex hylonoma Hu & Tang in Bull. Fan. Mem. Inst. Biol. Bot. 9:
250. 1940,

llex intermedia sensu S. Y. Hu in Ic. Pl. Omet. 2: pl. 164. 1946, non Loes,

An evergreen tree up to 10 m. high with glabrescent or glabrous
branchlets, elliptic or oblong-elliptic, coarsely and sharply serrate, slightly
acuminate leaves, large (10-12 mm. in diameter) ellipsoid fruits, and 4
large (7-9 mm. long) bony wrinkled and pitted pyrenes, each with a
median longitudinal ridge.

Branchlets straight, castaneous; second year’s growth subterete, 3 mm.
in diameter, plicate; current year’s growth 2 mm. in diameter, slightly
angular, striate and sulcate, the terminal buds slender, acute, conic,
glabrous, with ciliate bud-scales. Leaves occurring also on second year’s
growth (3—)25(—37) mm. apart; stipules minute, deltoid, 1 mm. long,
acute; petiole 8-14 mm. long, one-eleventh to one-seventh the length of
the lamina, canaliculate and minutely puberulent above, rugose beneath;
lamina thin-coriaceous, elliptic or oblong-elliptic, 6-12.5 cm. long, 2.4—4.5
cm. wide, acute or obtuse, rarely cuneate at the base, shortly acuminate
at the apex, the very tip mucronate; margin coarsely and sharply serrate,
sometimes the teeth ending in weak spines; the midrib impressed above,
pilose or glabrescent above, elevated and prominent beneath, the lateral
nerves 9 pairs, obscurely impressed above, elevated and prominent be-
neath, branched and anastomosing near the margin, the reticulations of
the veinlets evident above, rather obscure beneath. Inflorescences fascicu-
late, axillary on second year’s growth, the bracts deltoid, acute, glabrous
and ciliate; flowers 4-merous. Staminate inflorescences: individual
branches of the fascicle 3-flowered, sparsely puberulent; peduncles 1 mm.
long, the pedicel 3 mm. long, with 2 ciliate basal prophylla; calyx patelli-
form, 1.8 mm. in diameter, glabrous, the lobes broadly deltoid, obtuse,
ciliate, 0.5 mm. long, 0.5-1 mm. wide; corolla rotate, the petals obovate-
elliptic, 3.3-3.5 mm. long, 1.8 mm. wide, one-eighth connate at the base;
stamens slightly longer than the petals, the anthers ovoid; rudimentary
ovary subglobose. Pistillate flower not seen. Infructescences paucifascicu-
late, the pedicels 2 mm. long, with 2 sub-basal prophylla. Fruit ellipsoid-
subglobose, 10-12 mm. in diameter, quite wrinkled when dry, the persistent
calyx explanate, 3 mm. across, the stigma prominent, thickly discoid or
almost mammiform. Pyrenes 4, obovate in outline, trigonous in cross-
section, the apical end obliquely retuse, 7-9 mm. long, 3-4 mm. wide,
the dorsal surface irregularly wrinkled and pitted, with a median longi-
tudinal ridge, the endocarp bony.

CHINA: Szechuan: Mt. Omei, H.C. Chow 11831 (A); W. P. Fang
12579 (A, Sz), 14971 (Sz), 15296 (Sz), 15517 (Sz), 16496 (Sz), 18229
(Sz), 18540 (Sz); S. N. Hsu 678 (SS): T. C. Lee 3710 (Sz), 4438 (Sz):
W. W. Ma 2691 (Sz), 2339 (Sz), 2762 (Sz), 2771 (Sz); G. L. Sun 135

1949] HU, THE GENUS ILEX IN CHINA, II 351

(Sz), 149 (Sz); T. H. Tu 397 (asoryre, SS) ; without precise locality, C. W.
Yao 3659 (SS). Kweichow: Wong-Mou, Chen-feng, S. W. Tseng
90969 (A), 90969B (A).

Ilex hylonoma was first reported from Mount Omei, the sacred mountain
in West China. There it grows as a small tree 3-4 m. high in thickets at
altitudes of 950-1200 m., or as a big tree 10 m. high in mixed forests at
altitudes of 1300-1700 m. Its yellowish flowers appear in March. The
large fruits become brown in October and red in November (ex Fang).

Ilex hylonoma is closely related to Ilex centrochinensis. In the thin-
coriaceous texture of the leaves, the fasciculate inflorescences, the very
shortly pedicellate fruits, and in the sculpturing of the pyrenes, they are
almost alike. They differ in that [lex centrochinensis has smaller spinose
leaves, uniflorous branches of the staminate fascicles, and smaller pyrenes.
37a. Ilex hylonoma var. glabra, var. nov.

Arbor glabra; foliis coriaceis vel crasse coriaceis, lanceolatis, oblanceo-
latis, ovato-lanceolatis vel ellipticis, 6-10 cm. longis, 1.8—4.2 cm. latis,
margine serratis vel tenuiter spinosis; fructibus ellipsoideis vel subglobosis,
8-10 mm. diametro, pedicellis 2-3 mm. longis; pyrenis 4, oblongis, trigonis,
apice oblique retusis, 6-8 mm. longis, endocarpio lapideo.

Branchlets, terminal buds, pedicels, midrib and petioles all glabrous;
leaves coriaceous or thickly coriaceous, lanceolate, oblanceolate, ovate-
lanceolate or elliptic, 6-10 cm. long, 1.8—4.2 cm. wide, coarsely serrate or
weakly spinose; fruits ellipsoid or subglobose, 8-10 mm. in diameter;
pedicels 2-3 mm. long, prophylla 2, median, ciliate; pyrenes 4, oblong-
trigonous in outline, 6-8 mm. long, with 1 median longitudinal ridge
along the back, the apical end obliquely retuse, the endocarp bony.

CHINA: Chekiang: Chu-an-hsien, Y. L. Keng 726 (A). Hunan:

ang-sa, Handel-Mazzetti 11501 (A, ON Kwangsi: Hu-chen, Rk. C.

Fe 5386 CINDY): oe wun, S. K. Lau 28662 (A); Kwei-lin, W. T. Tsang

6 (typr, A; US), 2799 8 (AUS); 28063 “CA. US): Hine-on, 2.5.
he (Gisey Tsoong) 83691.

Geographically this variety ranges over a wider area than does the
typical West China species. It extends from the coastal provinces,
Chekiang, westward to the Great Lake Province, Hunan, and southward
to Kwangsi. It is a large tree 10 m. high

This variety differs from the typical West China species in having
glabrous branchlets and midrib. The leaves are thicker in texture.

38. Ilex centrochinensis, nom. nov.
Ilex aquifolium L. var. chinensis Loes. ex Diels in Bot. Jahrb. 29: 435.
.nud., in Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78:

236 (Monog. Aquif. 1: 236). 1901, descr., et in Sarg. Pl. Wils. :
1913:.. Hand.-Mzt. Symb, Sin. 7: 655. 1933;. Chen, Ill. Man. Chin.
Trees 658. 1937.

Tlex ean sensu Franch. in Bull. Soc. me fs 33: 453. 1886;
sensu. Hemsl. in Gard. & Forest 2: 124. 1889, n

Ilex dipyrena Wall. var. leptacantha Loes, ll.cc. tage 1900, nom. nud.,
78. 1901, descr. Syn. nov

—_
NI

352 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

An evergreen shrub up to 3 m. high, with slender angular and ridged
sparsely puberulent or glabrescent branchlets, lanceolate spinose leaves,
fasciculate inflorescences, short- a fruits, and 4 wrinkled-pitted
and median longitudinally ridged pyren

Branchlets slender, longitudinally ree and rugose, brown or nigres-
cent; third year’s érowth 3 mm. in diameter, the lenticels lacking, the
leaf-scars deltoid; second year’s growth 2—2.5 mm. in diameter, ridged;
current year’s growth angular, 1.5 mm. in diameter, puberulent or
glabrescent, sharply ridged from the stipules downward, the terminal buds
thin, acute-conic, puberulent or glabrescent with ciliate bud-scales.
Leaves occurring also on the second year’s growth, 5-15 mm. apart;
stipules deltoid or obliquely deltoid; petioles 5-8 mm. long, one-eleventh
to one-ninth the length of the lamina, very shallowly impressed above;
lamina coriaceous, brown, shiny above, less so beneath, elliptic-lanceolate,
rarely ovate-elliptic, 5—9 (usually 6-7) cm. long; 1.5—2.8 cm. wide; obtuse
or rarely rounded at the base; acuminate at the apex, the acumen deltoid,
6-12 mm. long, terminating in a spine; margin spinose-dentate, the spines
4-10 on each side, 2-4 mm. long, the tips nigrescent; midrib shallowly
impressed above, puberulent near the base, elevated beneath, the lateral
nerves 6-8 pairs, obscure, rarely evident above, obscure or rarely promi-
nent beneath. Inflorescences fasciculate, axillary on the second year’s
growth; the bud-scars often persistent, broadly ovate, cartilaginous,
puberulent or glabrescent, the individual branches of the fascicles uni-
flowered, the bracts ovate-lanceolate, pubescent, ciliate; flowers 4-merous.
Staminate flowers: pedicels 1-2 mm. long, puberulent, with 2 median
ciliate prophylla reaching the calyx; calyx patelliform, 2.5 mm. across,
deeply lobed, the lobes ovate, deltoid, ciliate and often puberulent; corolla
rotate, 6 mm. across, the petals oblong, 3 mm. long, the distal half ciliate,
one-eighth connate at the base; stamens longer than the petals, the anthers
oblong-ovate; rudimentary ovary subglobose, the apex rounded. Pistil-
late flower not known. Infructescences paucifasciculate, the fascicles with
1-3 fruits, the pedicels 2 mm. long, puberulent, with 2 sub-basal ciliate
prophylla; persistent calyx explanate, quadrangular in outline, 2 mm.
across, the lobes ciliate. Fruit globose, 6-7 mm. in diameter, the stigma
thin- -discoid, 4-lobed. Pyrenes 4, oblong-trigonous in authue, ca. 6 mm.
long, 3 mm. wide at the back, wrinkled and pitted all over, the dorsal
surfaces with a median longitudinal ridge, the ends obtuse, the endocarp
ston

CHINA: Hupei (Hupeh): Chien-shih-hsien, H. C. Chow 1523 (A,
NY); Nan-yang-ho, western Hupei, W. Y. Chun 3857 (A); Ichang, A.
Henry 1084 (Tyre, _ G), 3299 (A, G, US); 4239 (TYPE of Ilex dipyrena
var. leptacantha, A); E. H. Wilson 108 (A, NY), 295B (A), 3100 (A, US).
Szechuan: £. Paper 87 (NY); Wu-shan-hsien, E. H. Wilson 1028 (A);
W.Y. Chun 4098 (A).

This species is endemic to the Hupei-Szechuan border where the Yangtze
River cuts through the limestone mountains forming the famous huge
deep gorges. It is in this area that Metasequoia has recently been dis-

1949] HU, THE GENUS ILEX IN CHINA, II 353

covered. There [lex centrochinensis grows as a shrub along the roadside,
margin of the woods or by small streams at altitudes of 500-700 m. The
flowers appear in March and April, and the fruit turns purplish red in
middle August (ex Chun) or is still green in September (ex Chow).

In indumentum and leaf characters, /lex centrochinensis is most closely
allied with /lex aquifolium Linn., but the latter has 3-flowered branches
in the staminate inflorescences, fruiting pedicels as long as or longer than
the diameter of the mature fruits, reticulately striate and sulcate pyrenes
which lack a distinct median longitudinal ridge on the back. Moreover,
Ilex aquifolium Linn. does not occur in any part of China unless cultivated.
In the color and texture of the dry leaves, the nature of the inflorescences,
the short fruiting pedicels, and in the character of the pyrenes, //ex centro-
chinensis is very Closely related to Ilex hylonoma Hu and Tang, but the
latter has larger (up to 12 cm. X 4 cm.) serrate leaves, and much larger
fruits (up to 12 mm. in diameter).

Loesener with the specimen A. Henry 4239 created the variety Ilex
dipyrena Wall. var. leptacantha, basing it upon Ilex leptacantha Lindl. &
Paxt. But the specimen he examined does not belong to the species as
he supposed. It is a fruiting specimen. Each fruit has 4 stony, wrinkled-
pitted, longitudinally ridged pyrenes which are typical of Ilex centro-
chinensis. It is apparently a branch from a vigorously growing plant. The
leaves are larger and thinner, and the lateral nerves are more evident
above than in the other Henry specimens. But this variation is very
common in any growing shrub.

39. Ilex cornuta Lindl. & Paxt. Flow. Gard. 1: 43, fig. 27. 1850, in
Gard. Chron. 1850: 311. 1850; Loud. Encyc. Pl. Suppl. 2: 1302.
1855; Hook. in Bot. Mag. 14: pl. 5059. 1858; Moore in Jour.
Bot. 16: 137. 1878; Maxim. in Mém. Acad. oa St. Pétersb. VII,
29(3): 44. 1881; Forbes & Hemsl. in Jour. Linn. Soc. Bot. 23:
115. 1886; Loes. ex Diels in Bot. Jahrb. 29: 436. 1900, in Nov.
Act. Acad. Caes. Leop.-Carol. 78: 280 (Monog. Aquif. 1: 280).
1901, et in Sarg. Pl. Wils. 1: 78. 1911; Schneid. Ill. Hand. Laubh.
2: 164. 1912; H. Lévl. in Mém. Real Acad. Cien. y Art. 12: 13.
1916; Gee in Sci. [China] 6: 212. 1921; Chung in Mem. Sci.
Soc. China 1: 139. 1924; Liu in Bull. Pek. Soc. Nat. Hist. 2: 126.
1928; Belval, Mus. Heude Not. Bot. Chin. 2: 21.1933; Hand.-Mat.
Symb. Sin. 7: 656. 1933; Cheng in Contr. Biol. Lab. Sci. Soc. China
9: 171. 1934; Chen, Ill. Man. Chin. Trees 657. 1937

Ilex cornuta forma typica Loes. op. cit. 281. 1901. Syn. nov.

Ilex cornuta forma gactana Loes. lc. Syn. nov.

Ilex furcata Lindl. in hortis: Gonnent: in Gartenfl. 1853: 322. 1854.

Tlex burfordtt S. R. Howell, Descr. Cat. Howell abe 19, 1935, ve nud.

Ilex cornuta var. burfordti De France in i Hort. Mag. 13: 1934;
Clarke in Jour. Hort. Soc. 6: 284.

Ilex cornuta forma burfordt (De F1 se Rehd. in Bibl. Cult. Trees
Shrubs 400. 1949. Syn. now.

354 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

An evergreen shrub or small tree with firm thick-coriaceous dimorphic
leaves varying from cordate or oblong-entire to quadrangular and sinuate-
spinose, fasciculate inflorescences, globose drupes, and 4 wrinkled-pitted
stony pyrenes, entirely or half-grooved along the dorsal longitudinal line.

Branchlets stout, subterete, cinereous when dry; third year’s growth
5 mm. in diameter, longitudinally minutely rimulose, the nodes prominent
with elevated leaf-scars, the jenticels lacking; second year’s growth 4 mm.
in diameter, pphevecous or brown . longitudinally striate-rugose; current
year’s ae 2-3 mm. long, longitudinally ridged, minutely puberulent
in the grooves or glabrescent, the terminal bud small, thin-conical, puberu-
lent or glabrescent. Leaves occurring even on the third year’s growth,
5-10 mm. apart; stipules broadly deltoid, callose, ca. 0.4 mm. long, 0.5
mm. wide; petioles 4-8 mm. long, one-seventeenth to one-ninth the length
of the lamina, puberulent, narrowly impressed above; lamina thick-
coriaceous, olivaceous, shiny above, opaque beneath, quadrangular-oblong,
rarely ovate, (3—-)5-6(-8) cm. long, (2—)2.5-3(—4) cm. wide; round or
truncate at the base; acute or shortly acuminate at the apex, the very tip
always terminating in a strong spine; margin entire or sinuate with 1-3
spines on each side, rigid, thickened and of lighter color than the lamina;
midrib very slightly impressed (almost plane on the distal half) above,
prominent and elevated beneath, the lateral nerves 5 or 6 pairs, anastomos-
ing near the margin, obscure above, evident beneath, the reticulation of
the veinlets obscure on both surfaces. Inflorescences fasciculate, sessile,
axillary, on second year’s growth, the persistent scales suborbicular, car-
tilaginous, ciliate and sparsely puberulent, 1.5 mm. long, the individual
branches of the fascicles uniflowered, the bracts ovate, pubescent, ciliate,
obtuse or mucronate, with 2 subulate basal appendages; flowers 4-merous.
Staminate inflorescences: pedicels 5-6 mm. long, glabrous, with 1 or 2
sub-basal minute broadly deltoid prophylla; calyx patelliform, 2.5 mm.
across, the lobes membranaceous, broadly deltoid, sparsely aubenulene
ciliate, 0.75 mm. long, 1.5 mm. wide. corolla rotate, 7 mm. across, the
petals oblong-ovate, 3.5 mm. long, with the apical ends very spatsely
ciliate, one-tenth connate at the base; stamens subequaling or slightly
longer than the petals, the anthers ovoid-oblong, 1 mm. long; rudimentary
ovary subglobose, the apex obtuse or rounded, indistinctly 4-lobed. Pistil-
late ido: seca: pedicels 8-9 mm. long, glabrous, after fruiting 13-14

mm. long with minute broad deltoid basal prophyl!a; calyx patelliform,
the lobes obtuse, sparsely puberulent and ciliate; corolla rotate, 7 mm.
across, Shor oeialons. the petals oblong-ovate, a5 mm. long, sparsely
ciliate; staminodes four-fifths the length of the petals, slightly longer than
the ovary, the sterile anthers ovate-sagittate; ovary oblong-ovoid, 3—4 mm.
long, 2 mm. wide, the stigma discoid. Fruit globose (ellipsoid before
reaching maturity), 8-10 mm. in diameter, the persistent calyx quad-
rangular in outline, the stigma discoid, distinctly 4- sg ae Pyrenes 4,
obovate or elliptic in outline, the ends acute, 7-8 mm. long, 5 mm. wide
at the back, wrinkled-pitted, rugose all over, the cee surface entirely

1949] HU, THE GENUS ILEX IN CHINA, II 355

or partly canaliculate along the median longitudinal line, the endocarp
bony.

CHINA: Kiangsu: Nanking, Chen & Teng 3960 (A, US) 3961
CNN) C. Cheng 853 (SS) 17. P. aoe LOFIO “(55 ), 61108 CSS).
Soochow: HH. F. Feng in 1925 (G); V. L. Keng 1692 (CASEIN) 2B
Faber (NY); Chiu-yong, K. King 2175 (G); A. N. Steward (ex Herb.
Univ. Nanking 1959) (A), 1201 (NY (US); J A Sunsl0or CSS\-- Tso
28 and 210 (A). Chekiang: Moh-kan-san, S. P. Barchet (US); Tien-
iMucshany WC. Cheng.3024 (US) ‘ih talon R. C. Ching 1374 (A,

US); Hangchow, C. Y. Chiao (ex Herb. Univ. Nanking 7966) (US);
Nin-po, C. K. Chao. im 1935,-(CUB); S. S. Lad in 1935-(CUB) ising -yun-
hsien, Y. L. Keng 449 (A, SS); Chang-hwa-hsien, Y. L. Keng 574 (A);
Chun-an-hsien, ¥. L. Keng 712 (A); Lin-an- shsien’ eth Elie L902 (CAL)
locality not given, D. Macgregor in 1908 (A); Tang-si, F. H. Meyer 229
(A); Tien-mo-shan, T. Tang & W. Y. Hsia 428 (A). eaten ei: Hwang
shan, S. S. Chien 1336 (SS); Bau-hua-shan, W.C. Cheng 434 (SS, US),
DOU COs ic C. met,.2408" (SS). Kiwne si : Nan-chang, H. H. Chung
658 (LU); J. N. Hsiung 504 (A); Kian, F. A. McClure 3421 CEC):
Kao-ping, Y. Tsiang 9893 (NY); Kiu-kiang, E. H. Wilson 1608 (A, Sous
Hupei (Hupeh) Wu-chang, S.C. Sun 1078 (A, NY); Ichang, A. Henry
3292A (G); FE. H. Wilson 3101A (A, US). Hunan: Tsing-tschou
(Chang-sa), PigndelNigeeens 364, 637, 442 We

KOREA: Senra, Taquet 2721 (A); same province, E. H. Wilson 11242
(A).
CULTIVATED: I have seen specimens from the gardens of Kew, Eng-
land, Hannover and Dahlem, Germany, Locarno, Switzerland, and from
ee and nurseries of vince Tennessee, and Georgia in the United
State

ie cornuta is endemic to the hilly regions of the lower Yangtze prov-
inces from the sea up to western Hupei. There its yellow flowers appear
in April and the fruit turns red in November.

Ilex cornuta was first introduced to England from the vicinity of
Shanghai by R. Fortune in 1846. The first published record of it appeared
in 1850. From the very beginning Lindley and Paxton had clearly stated
the variation in leaf forms, “Foliis . . . in planta vegetiore grossé sinuato-
dentatis spinosis in adulta 3-cornibus integrisque.” In 1908 Dallimore
commented on Fortune’s specimen deposited in Kew, saying, “it appears
that the upper leaves on mature trees vary in shape, sometimes being
spineless and sometimes bearing but one or two spines.”’ This is correct.

Ilex cornuta is like many other Asiatic spinose species of /Jex in that
maturity in the plant and aridity in the environment are associated with
the reduction of the number of spines on the leaves, often even with entire
leaves. Such a dimorphism in leaves is very common with wild J/lex
cornuta. It has been reported from Chekiang by Keng, from Anhwei by
Ching, and in Nanking I have personally seen trees about 5 m. high with
some branches bearing entire leaves and others bearing spinose leaves.
The type or parent plant of /lex cornuta var. burfordii De France in the
West View Cemetery, Atlanta, Georgia, must have been propagated

356 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

from such an entire-leaved shoot. De France admitted, ‘“When seeds of
Ilex cornuta var. Burfordit germinate they produce seedlings similar to
Ilex cornuta ...” Therefore, his variety may be considered as a horti-
cultural clone. but not a taxonomic form or variety.

The branchlets of Ilex cornuta are rather flexible. In China they are
used by farmers to make nose-rings for cattle. The bark and the leafy
shoots are used as medicine. The plant is believed to be an excellent
remedy for diseases of the kidney and is commonly administered as a
tonic. It is also widely cultivated as an ornamental shrub.

39a. [lex cornuta var. fortunei (Lindl.), comb. nov.
mis Bhi Lindl. in Gard. Chron. 1857: 868. 1857; Chung in Mem.
. China 1: 140. 1924.

| sensu Rehd. in: Jour. Arnold Arb. 8: 156. 1927, pro parte.

Branchlets minutely puberulent or glabrescent; leaves thick-coriaceous,
oblong, obovate-oblong, or quadrangular, the margin entire, detasonelly
with 2 or 3 strong spines on each side, rigidly thickened; inflorescences
fasciculate or in the staminate inflorescence pseudopaniculate with a rachis

p to 10 mm. long; individual branches of the staminate fascicles 1—3-

flowered, the pedicels 5-15 mm. long, when 3-flowered the peduncles
1—2 mm. long; individual branches of the pistillate fascicles uniflorous,
the pedicels 12 mm. long, after fruiting up to 18 mm. long; fruit globose,
10-11 mm. in diameter, the stigma plane-discoid or almost capitate.

CHINA: Kiangsu: Nanking, £. D. Merrill 11371 (A). Honan:
Ki-Kiang-shan, A. N. Steward 5239 (US), 9838 (A, US). Anhwei:
Chu-wha-shan, R. C. Ching 2719 (A, LU); os chu-wan, R. C. Ching 2736
(A5- US): Chien- shan-hsien, = - Fan & Y. VY. Li/72 (A). Chekiang
Tung-yang-hsien, Y. L. Keng 912 (A). a ei: Ichang, A. Henry 3292

A, G, US); £. H. Wilson re (A, 55); eee precise locality, F. B.
Forbes 854 (A).

Wilson 3101 is a mixture of staminate flowers and young fruits. The
former have long pedicels, often with 3-flowered branches. The latter
have short pedicels. In the Arnold Arboretum, I have numbered the
latter 310A.

Cultivated: I have seen specimens collected from La Mortola, Italy,
by A. Rehder and also by C. Schneider; from United States National
Botanic Gardens by P. Russell, and from Australia by J. F. Bailey.

Geographically, there is no clear demarcation between this variety and
the typical /lex cornuta. However, so far as is known, the variety has
not been reported in those coastal areas beyond Long. 120°E. It com-
monly occurs in the hilly high lands on both sides of the Yangtze River
from Kiangsu west to Hupei. The flowers appear in April and the fruit
becomes red in November, persisting for a long time on the branches.

This variety differs from typical Jlex cornuta in having 3-flowered
branches of the staminate inflorescences and longer fruiting pedicels
(12-18 mm.). It was first collected by Fortune in Hwuy-chow, S. Anhwei.

1949] HU, THE GENUS ILEX IN CHINA, II 357

His specimen was from an old plant with entire leaves. About it Lindley
wrote, “In its young state it is much like J. cornuta, but in the adult con-
dition it acquires quite another appearance, resembling a very broad-leaved
entire-leaved European Holly ... in the axil of each leaf a sessile umbel of
from 6-10 stalks [of fruits] each about three-quarters of an inch long...
In S. Anhwei it was again collected by R. C. Ching (2719 with entire
leaves only and 2736 with both entire and spinose leaves). Merrill’s
specimen from Purple Mountain in Nanking and Henry’s and Wilson’s
specimens from western Hupei both possess spinose leaves only. It is
obvious that in nature leaf dimorphism is as common in this long-pedicel-
late variety as in the typical /lex cornuta.

SERIES 2. DIPYRENAE, ser. nov.

Arbor vel frutex, ramulis puberulis; foliis coriaceis, rigidis, juvenilibus
pinosis, in maturitate interdum partim integris; inflorescentiis pauci-
Pcccalatig fasciculis 1—-5-floris; fructus pedicellis 1-3 mm. longis; fructi-
us ellipsoi eis vel depresso- plabosis: saepe in paribus; renis 1-4,
plerumque 2, palmatim striatis et sulcatis, endocarpio crasso, lignescente.

KEY TO THE SPECIES
A. Trees; leaves comparatively large, 6-9 cm. long, often entire; petioles
Ajo mine ones o Clndia YUNNAN). sre wad acear sad hers 40, I. dipyrena.
AA. Shrubs; leaves 2-4.5 em. long; petioles less than 3 mm. long.
B. Pyrenes 4; leaves ovate

C. Leaves 3-4.8 cm. long, the margin sinuate, with 3-7 spines

on each side; pyrenes suborbicular in outline, 4.5 mm. wide,

broadly grooved on the dorsal surface. ae Peat ate

winaile soho Mehr ulol sic ouemn oleate etien.cirehc'at ake "5" Tacjert oo 4-4. se: “a 0 See coe peeetee I. WEUNCHOWCNSIS.

CC. dbeaves: 13-257 rarelyup: to .3 em: ae the margin with

1 or 2, rarely 3 spines; pyrenes obovate or oblong, less than

4 mm. wide, impressed only at the broader end. (West and

MOEN est © litle een eee ere ees tens ae Wl OTS

BB. Pyrenes usually 2, rarely 1; leaves lanceolate, elliptic, rhomboid,

quadrangular, rarely ovate, usually over 4 cm. long (except J.
perryana).

C. Leaves lanceolate, es numerous, spines; fruit obovoid- ellip-

4

soid, 5-6 mm. in diat 1eter. ( “unnan) eee: 43. I. george.
CC. Leaves elliptic, rromboid or quadran gular, rarely ovate; fruit
ellipsoid or subglobose, 8-10 mm. in Gemciee

Direct Saute over 1 m. high: leaves elliptic or ovate,
truncate or rounded at the base, acuminate at the apex:
the vere indistinct above; pvrenes 2, palmately striate
and sulca

yee res velliptic or ovate; margin with 4-6 weak spines
on each side; pedicels, of fruits eset (Sze-
CIA Tce acee ee en doa Sma eer? I. ciliospinosa.
EE. Leaves ovate or quadrangular, margin with 2-3
strong spines on each side; pedicels of fruits glab-
rO (Yunnan to Taiwan)..... 45. I. bioritsensis.

358 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

DD. Prostrate shrubs less than 30 cm. high; leaves sub-
rhomboid, rugose, obtuse at the base, broadly deltoid and
acute at the apex, the upper surface with ae se
veinlets: pyrene 1, palmately striate and esulcate. (Yu
nan, high altitude ) eee oe rere eer 46. J, oe

40. lex dipyrena Wall. in Roxb. Fl. Ind. ed. Carey, 1: 473. 1820;
D. Don, Prod. Fl. Nep. 188. 1825; Wall. List, 4327. 1830, et PI.
As. Rar. 3: pl. 292, 1832; Walp. Rep. 1: 540. 1842; Loud. Encyc.
Pl. Suppl. 2: 1302. 1855; Brandis, For. Flor. 76. 1874; Hook. f.
Fl. Br. Ind. 1: 599. 1875; Franch. Pl. Delav. 127. 1889; Loes. in
Nov. Act. Acad. Caes. Leop.-Carol. Nat. Cur. 78: 276 (Monog.
Aquif. 1: 276). 1901; Chung in Mem. Sci. Soc. China 1: 140, 1924;
Hand.-Mzt. Symb. Sin. 7: 656. 1933; Comber in Notes Bot. Gard.
Edinb. 18: 44. 1933; Chen, Ill. Man. Chin. Trees 659. 1937.
lex eae ena var. connexiva W. W. Sm. in Notes Bot. Gard. Edinb. 10:
41. 1917; Comber op. cit. 61. 1933. Syn. nov.
ire - ntoni i ort. ee Letts hae
Ilex dipyrena var. de aia Loes. op. cit. 89: 283. 1908; Comber op.
cit. 18: 45, 1933. Syn. no
Ilex monopyrena Watt. ex ioe op. cit. 275, a Hu and Tang in Bull.
n Mem. Inst. Biol. Bot. 9: 252. 1940.
Ilex bioritsensis Havata var. integra Comber op. in 43. 1933. Syn. nov.
A puberulent evergreen tree up to 14 m. high with coriaceous elliptic-
oblong, rarely ovate or lanceolate, spinose (or entire with an apical spine)
leaves, fasciculate inflorescences, globose or ellipsoid fruits, 1-4 pyrenes
with thick woody palmately striate and sulcate endocarp
Branchlets straight, stout; third year’s growth 4-6 mm. in diameter,
cinereous, minutely longitudinally rimulose, the lenticels lacking or obscure,
the leaf-scars semi-orbicular, not elevated; current year’s growth sub-
angular, longitudinally striate rugose or smooth, minutely puberulent or
glabrescent, cinereous or cinnamon, the terminal buds conic, acute, puberu-
lent. Leaves occurring also on second year’s growth, 5-20 mm. apart;
stipules callose, broadly deltoid, sometimes obscure; petioles 4-6 mm.
long, one-twentieth to one-thirteenth the length of the lamina, shallowly
canaliculate above, puberulent; lamina thick-coriaceous, shiny above or
opaque on both surfaces, olivaceous, brown or even yellow, elliptic-
oblong or rarely ovate, 4-10 cm. long, 2—4 cm. wide; base rounded; apex
shortly acuminate, the acumen 3-10 mm. long, terminated by a sharp
spine 2 mm. long; margin entire or subentire with few spines or spinose
with up to 14 spines on each side; midrib sulcate, puberulent or glabres-
cent above, elevated beneath; lateral nerves 7—9 pairs, impressed above,
elevated beneath, the reticulation of the veinlets obscure. Inflorescences
fasciculate, axillary on second year’s growth, the individual branch of the
fascicles uniflorous; the bracts ovate-lanceolate, long-ciliate, the outer
ones cuspidate, the inner ones obtuse, the basal appendages subulate, flow-
ers 2—4-merous. Staminate flowers: pedicels 2-3 mm. long, sparsely and
minutely puberulent or glabrescent, with 2 sub-basal lanceolate acute

1949] HU, THE GENUS ILEX IN CHINA, II 359

Bae 1.5 mm. long, which reach the calyx; the calyx patelliform, 3

. across, deeply lobed: the lobes membranous, ovate-deltoid, obtuse or
ne Ged pilin: corolla rotate, 7 mm. across, the petals ovate,
3 mm. long, ciliate, one-fifth connate at the base; stamens 4, longer than
the petals, the anthers oblong-ovoid, 0.75 mm. long; the rudimentary
ovary subglobose-ovoid, 1.5 mm. in diameter, the apex truncate or obtuse.
Pistillate flowers: pedicels 1-3 mm. long; calyx and corolla as in the
staminate flowers; staminodes slightly shorter than the petals, the sterile
anthers ovate; ovary ovoid, the stigma discoid, plane, distinctly lobed.
Fruit red, globose, 6-9 mm. in diameter, appearing subsessile, the per-
sistent calyx explanate, quadrangular, the stigma 2—4-lobed, discoid.
Pyrenes 1-4, often 2; when two, oblong-elliptic or subcircular in outline,
dorso-ventrally compressed, the ends obtuse or rounded, the back convex,
longitudinally and subpalmately striate and sulcate, 5-7 mm. long, 5 mm.
wide, the ventral side also striate-sulcate; when four, oblong in outline,
3.5 mm. wide at the back; the endocarp thick and woody.

CHINA: Yunnan: Li-kiang, R. C. Ching 20243 (A), 21576 (A);
ToM sheng 3101 (A). 3179 (Ada. Rock 3427 (AWS) 516. Schneider
2845 (A); T. T. Yu 14921 (A); Hokin near Sung-kwei, K. M. Feng 6921
(A); western Yunnan, G. Forrest 10171 (A), 10224 (A), 10426 (A), 11071
(A), 12528 (TYPE of Ilex dipyrena var. connexiva, fragment in A), 19817
(A, US), 20096 (A, US), 20101 (A, US), 21023 (isotype of I. bioritsensis
var. integra, A), 25362 (A); Cai-pou, E. E. Maire 125 (A); west of Ta-li,
J. F. Rock 6781 (A, US); mountain ee Londjre, J. F. Rock 8906 Ee Soe
Wei-hsi, J. F. Rock 11701 (A, US); locality not given, H. T. Tsai 57186
(A), 57360 (A), 57339 (A); Wei-si, C. W. Wang 67917 (A), 70400 (A);
Atungtze, T. T. Yu 5773 (A), 7826 (A), 7831 (A), 8486 CS: 10610 (A),
11171 (A). Sikang: Zayul, Kingdon Ward 10990

INDIA: East Himalaya, 1Gr ifith 2011 (NY); Simla Gr ifith (NY);
Chakrata, M. A. Hanud (NY); Ramilchet, G. os (NY); Punjab,
W. Koelz 1990 (NY); Chamba, W. Koelz 8857 (NY); Mussoorie, R. R.
Steward 14895 (NY); Kumson, R. Strachey & J. E. Winterbottom 3 (G) ;
Wallich 4327 (typr, fragments in #

UPPER BURMA: F. K. Ward 342 (NY).

CULTIVATED: Many specimens from botanical gardens of America
and Europe have been examined.

Ilex dipyrena was first described from Nepal, where it is called “‘caulah.”
According to Wallich, it “blossoms from April to July.” Now, as our
knowledge of the flora of southwestern China has increased, we know that
it is a common and wide-spreading species in Yunnan Province, where it
grows as a tree (sometimes as a shrub, after cutting by fuel-gatherers) in
mixed forests, in ravines, or on slopes of mountains at an altitude of
2000-3000 m. The fruit is greenish yellow in October and red in
December.

As Wallich remarked, ‘This species varies considerably in the shape
and serratures of the leaves. Generally they are oblong and remotely
serrate; sometimes, however, they are of a more ovate form, rounded at
the base, the margins sinuated and marked with pretty strong unequal

360 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xxx
spinous serratures.” This statement is true; varieties based on differences
in leaf margin are not worthy of taxonomic recognition. Rock 11701
serves as a good example. On that sheet, there are subentire leaves which
resemble Ilex dipyrena var. connexiva, less spinose leaves which resemble
Ilex dipyrena var. paucispinosa, and very spinose leaves. In specimens
collected from Yunnan, the pyrenes vary from one to four. In the
double-pyrened fruits there are often one or two small globose masses of
stone-cells which may indicate aborted pyrenes. The fruits of H. 7. Tsai
57360 contain three or four pyrenes. Yet its vegetative characters show
it to be an example of /lex dipyrena.

41. Ilex wenchowensis, sp. nov.

Frutex parvus; ramulis geara foliis spinosis, coriaceis, ovatis,
3—4.8 cm. longis, 1.2—2.5 cm. latis; petiolo 1-2 mm. longo; costa supra
impressa et puberulente; nervis lateralibus utrinque 4 vel 5, supra evidenti-
bus, subtus obscuris; inflorescentiis fasciculatis, unifloribus; pedicellis 1
mm. longis, prophyllis ciliatis; calycibus 4-lobis, ciliatis; corolla rotata,
6—7 mm. diametro, petalis oblongis, 3 mm. longis, 2.5 mm. latis, ciliatis,
one-fifth connatis; staminibus 4, quam petalos aequilongis; ovario abortu
turbinato, 1 mm. longo, apice obtuso; fructibus depresso-globosis, 5-6 mm.
longis, 8 mm. diametro, calycibus 2.5 mm. diametro donatis; stigmate
umbilicato vel discoideo; pyrenis 4, suborbicularibus, 5 mm. longis, 4.5
mm. latis, dorso palmatim striatis et depressis, lateralibus striatis et
sulcatis.

An evergreen shrub up to 2 mm. high with pubescent branchlets, spinose
ovate leaves, fasciculate inflorescences, depressed globose fruits and 4
palmately striate and dorsally depressed pyrenes.

Branchlets pubescent, castaneous; third year’s growth 3.5 mm. in di-
ameter, the lenticels lacking, the leaf-scars suborbicular, not elevated;
current year’s growth 3 mm. in diameter, angular and pubescent, the ter-
minal buds narrowly conic, pubescent, unfolding after anthesis. Leaves
occurring even on the third year’s growth, 5-10 mm. apart; stipules
minute, obscure; petioles 2 mm. long, one-twenty-fifth to one-eighteenth
the length of the lamina, puberulous; lamina coriaceous, olivaceous, slightly
shiny above, ovate, 3-4.8 cm. long, 1.2—2.5 cm. wide, truncate or rounded
at the base, acuminate and spinose at the apex; margin sinuate, with 3-7
spines on each side, the midrib slightly impressed and pubescent above,
slightly elevated beneath, the lateral nerves 4-5 pairs, evident above,
obscure below. Inflorescence fasciculate, axillary on second year’s growth,
the individual branches of the fascicles all uniflorous; flowers all 4-merous.
Staminate flowers: pedicels 1 mm. long, puberniods: with 2 large ciliate
basal prophylla extending to the calyx; calyx patelliform, 2 mm. in di-
ameter, deeply 4-lobed, the lobes deltoid, obtuse, ciliate; corolla rotate,
6-7 mm. across, the petals oblong, 3 mm. long, 2.5 mm. wide, sparsely
ciliate, one-fifth connate at the base; stamens 4, equaling the petals in
length, the anthers ovoid; rudimentary ovary conic, 1 mm. long, the apex
obtuse. Pistillate flowers not seen. Infructescences fasciculate, the pedi-

1949] HU, THE GENUS ILEX IN CHINA, II 361

cels 4-5 mm. long, puberulous, with 2 sub-basal ovate ciliate prophylla,
the persistent calyx explanate, 2.5 mm. in diameter. Fruit (immature)
depressed-globose, longitudinally grooved, 5-6 mm. long, 8 mm. in di-
ameter, the stigma thin-discoid or navel-like. Pyrenes 4, suborbicular in
outline, 5 mm. long, 4.5 mm. wide, the dorsal surface palmately striate,
esulcate and longitudinally depressed, the sides reticulately striate and
sulcate, the endocarp woody.

CLilAys S@hekiane 7 Wenchow, &. C.-Chingoi slo 1(tver A US),
1520 (fruit, A; staminate flowers, US).

Ilex wenchowensis is localized in southeastern Chekiang, where it occurs
on bushy slopes at an altitude of 600 m. There it flowers in May. It is
closely related to /lex pernyi Franch. which differs in having .smaller
leaves (1.3-2.5, rarely 3 mm. long) with one or two spines on each side
and oblong or obovate pyrenes, impressed only at the broader ends. From
Ilex bioritsensis Hayata it differs in having 4 pyrenes. From Jlex centro-
chinensis S. Y. Hu it differs in its very short petioles and median longi-
tudinally impressed pyrenes.

42. lex pernyi Franch. in Nouv. Arch. Mus. Hist. Nat. I, 5: 221. 1883,
Pl. Dav. 1: 69. 1884; Oliv. in Hook. Ic. Pl. III, 6: pl. 1539. 1886;
Forbes & Hemsl. in Jour. Linn. Soc. Bot. 23: 117. 1886; lLoes. ex
Diels in Bot. Jahrb. 29: 436. 1900, in Nov. Act. Acad. Caes. Leop.-
Carol. Nat. Cur. 78: 278 (Monog. Aquif. 1: 278). 1901, et in Sarg.
Pl. Wils. 1: 78. 1911; Farrer in Jour. Roy. Hort. Soc. 42: 77. 1916;
Chun, Chin. Econ. Trees 223. 1921; Chung, Mem. Sci. Soc. China
1: 141. 1924; Rehd. in Jour. Arnold Arb. 7: 199. 1926, 8: 156.
1927; Comber in Notes Bot. Gard. Edinb. 18: 75. 1933; Chen,
Ill. Man. Chin. Trees 658, fig. 550. 1937; S. Y. Hu in Ic. Pl. Omei.
2: pl. 162. 1946.

An evergreen shrub or small tree up to 8 m. high with pubescent branch-
lets, very short petioles, small coriaceous ovate-lanceolate spinose leaves,
fasciculate inflorescences, rather large globose drupes up to 8 mm. in
diameter, and 4 palmately striate and sulcate pyrenes.

Branchlets straight, terete or subterete; third year’s growth 4 mm. in
diameter, apricot-colored, covered with a dirty gray pubescence, minutely
rimulose, the lenticels lacking; second year’s growth 3 mm. in diameter,
dirty gray, densely pubescent; current year’s growth 2 mm. in diameter,
angular, pubescent, fulvous, the terminal buds ovoid-conic, acute, pubes-
cent. Leaves occurring even on the fifth year’s growth, crowded together,
appearing sessile; stipules callose, deltoid, acute, 0.75 mm. long, 0.5 mm.
wide; petioles very short, 2 mm. long, one-twelfth to one-eighth the length
of the lamina, pubescent; lamina coriaceous, olivaceous, ovate or ovate-
lanceolate, 1.3-2.5 cm., rarely 3 cm. long, 5-14 mm. wide, broadest
between the two anterior spines; base rounded or truncate; apex tri-
angularly acuminate, the acumen 12-14 mm. long, terminated by a strong
spine 3 mm. long; margin sinuate-dentate, 1-3 spines (usually 2) on each
side, the two anterior ones strongest; midrib slightly impressed above,

362 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

puberulent near the base, gradually becoming obscure toward the apex,
elevated beneath, the lateral nerves 1-3 pairs, obscure beneath. In-
florescence fasciculate, axillary, on second year’s growth; individual
branches of the fascicles all uniflorous, the bracts lanceolate, mem-
branous; pedicels sparsely puberulent, ciliate; flowers all 4-merous.
Siaminaté inflorescence: pedicels 1 mm. long, glabrous, with 2 suborbic-
ular ciliate super-median prophylla; the calyx 2 mm. across, deeply
4-lobed, the lobes broad-deltoid or semi-orbicular, ciliate; the corolla
rotate, 7 mm. across, the petals 3 mm. long, broad-elliptic, the apical half
ciliate; stamens slightly longer than the petals, the anthers 1 mm. long:
the rudimentary ovary conic-ovoid, 1.5 mm. long, the apex obtuse. Pistil-
late inflorescence: pedicels 2 mm. long; calyx as in the staminate flowers:
corolla choripetalous, erect, the Beale ovate, 2.5 mm. long; the staminode
two-thirds as long as the petals, the sterile nies ovate; the ovary ovoid,
the stigma discoid. Fruit globose or depressed- globose, 7-8 mm.
diameter, the persistent calyx quadrangular in outline, 2.5 mm. across,
ciliate, the stigma thick-discoid, distinctly 4-lobed. Pye 4, obovate
or ehione | in outline, 4.5-5.5 mm. long, 3-3.5 mm. wide, impressed on the
back at the wider end, palmately striate and sulcate on the back, reticularly
striate and sulcate alone the sides, the endocarp thick, woody.

CHINA: Anhwei: Wu-yuan, R. C. Ching 3245 (A). Klee S32
ee lung- Snen, Y de eu 5627 (A ee Hupei (Hupeh): Pa-tung-
hsien. H. C. Chow 337 (A, NY); W. Hupei, W. VY. Chun 3808 (A), 4172
(A), 4153 (US); A. Henry 5298 (US), 52084 Ce. ee Tchans. 2. 7.
Wilson 119 (staminate, A, US), 119A (fruit, A, US), 271 (hy Na US
Kweicho wi: Perny (ex Franchet). Kansu: Feng-s’an-ling Pome
16, She : Tsing-ling, David. Szechuan: Mt. Uo-mi SG [ Mt.
Omei J, Py. ” Hugh in 1899 (B); Kwang-yun-hsien, F. T. W ang 22595.
Sikang: Pau-hsien-hsien (Mu-pin), K. L. Chu 2456 (SS), 3090 (3513
Tien-chuan, Y. S. Lin 1313 (A),

CULTIVATED: J have seen cultivated specimens from the Royal
Botanic Gardens at Kew, England, from J. A. Havemeyer, and from the
Arnold Arboretum.

/lex pernyi is endemic to the Tsing-ling Range, the climatic barrier in
central China responsible for the very different aspects of the floras of
North and South China. Jlex pernyi is found on both sides of this range,
from Kansu down to the foothills of Hwang-shan in Anhwei. On the
western limit of its area of distribution, the plant, as Farrer reported,
grows as an “ugly and graceless little dumpy shrub . . . woodland of the
drier ranges.” On the eastern and southern sections of its area of dis-
tribution it is a shrub or small tree in forests growing at an altitude of
1800 m. (ex Chun). It flowers in April and May, and the fruit becomes
red in October.

On the field notes attached to the specimen thi W. Y. Chun 4172,
the collector is erroneously recorded as “R. C. Chin

The fruits of Y. S. Liv 1313 contain either two i Sees pyrenes. This
may indicate that the two-pyrened /lex bioritsensis Hayata is nothing

1949 | HU, THE GENUS ILEX IN CHINA, II 363

but a variety of Jlex pernyi. The small ovate-lanceolate shiny leaves and
the palmately striate-sulcate, thickly woody endocarp of /lex pernyi closely
relate it to [lex georget Comber, which differs in having only two pyrenes.

43. Ilex georgei Comber in Notes Bot. Gard. Edinb. 18: 50. 1933;
Hand.-Mzt. Symb. Sin. 7: 656. 1933; Merr. in Brittonia 4: 100.
1941.

Ilex pernyt Franch. var. manipurensis Loes. in Nov. Act. Acad. Caes.

Leop.-Carol. Nat. Cur. 78: 279 (Monog. Aquif. 1: 279). 190

flex pernyt sensu Loes. in Sarg. Pl. Wils. 1: 78. 1911, in part.

A spinose-leaved evergreen shrub up to 6 m. high with pubescent
branchlets, lanceolate (rarely ovate) leaves, fasciculate inflorescences, obo-
void-ellipsoid red fruits each with 1 or 2 pyrenes.

Branchlets rather stout, pubescent, cinereous or light brown; third
year’s growth 4 mm. in diameter, hirsute, the lenticels lacking; second
year’s growth 3.5 mm. in diameter; current year’s growth subterete, 2—2.5
mm. in diameter, longitudinally shallowly striate-sulcate, densely pubescent,
the terminal bud conic, pubescent. Leaves occurring even on the fourth
year’s growth, 2-5 mm. apart; stipules broadly ovate, acute; petioles
1-2 mm. long, one-twentieth to one-fifteenth the ieneth of the lamina,
pubescent, often covered by the decurrent base of the leaf; lamina thickly
coriaceous, olivaceous, shiny above, opaque beneath, lanceolate, ovate-
lanceolate or rarely ovate, 2—4.5 mm. long, 7-15 mm. wide; rounded or
cordate at the base; acuminate at the apex, the acumen 10 mm. long,
terminating in a spine 3 mm. long; margin thickened, recurved, sub-
entire, with 4-7 spines on each side; midrib pubescent, shallowly im-
pressed above, elevated beneath, the lateral nerves 5—7 on each side,
indistinct above, evident beneath, the reticulation of the veinlets obscure
on both surfaces. Inflorescences fasciculate, axillary on second year’s
growth. Staminate inflorescences: individeel branches of the fascicles
1—3-flowered, the uniflorous pedicels 2—4 mm. long, when 3-flowered the
peduncles 1 mm. long, both sparsely and minutely pubescent, the bracts
membranous, ovate, obtuse, ciliate, the prophylla 2, membranous, sub-
median, pubescent; flowers 4-merous, calyx 1.5—-2 mm. across, 4-lobed,
lobes ovate, obtuse or rounded, ciliate; corolla rotate, 4-5 mm. across, the
petals 2 mm. long, minutely and sparsely ciliate, one-tenth connate at
the base; stamens longer than the petals, the anthers oblong; rudimentary
ovary subglobose or ovoid, obtuse at the apex, sometimes obscurely 2-lobed.
Pistillate flowers not seen. Fruit usually paired, obovoid-ellipsoid, 4—6
mm. long, 3-4 mm. thick, the pedicels 2 mm. long, pubescent, the persistent
calyx ciliate, the stigma discoid. Pyrenes 1 or 2, obovoid-oblong, rarely
suborbicular in outline, slightly dorso-ventrally compressed, flattened on
one side, 4.5 mm. long, 2.5-3 mm. wide, longitudinally (almost palmately )
7—-9-striate and shallowly sulcate on the back, the ends obtuse, the endo-
carp thickly woody.

CHINA: Yunnan: Tengyueh, G. Forrest 7577 (A), 9608 (A), 9712
(A), 16061 (fruit, isoryPr, A), 26251 (staminate flowers, IsoTYPE, A), 26254

364 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

(US); J. F. Rock 7677 (A, NY, US); San-ying-pa, C. Schneider 394 (A);
without precise locality, H. T. Tsai 57247 (A); Mien- ra To Ts YW 17907
(A). Sikang: Ta-chien-lu, EF. H. Wilson 119B (A

NORTH BURMA: F. K. Ward 463 (A, NY).

Ilex georgei is confined to western and southwestern Yunnan and adja-
cent Burma, where it grows as a shrub in dense thickets, on dry slopes,
or in side valleys at altitudes of 1800-2700 m. Its greenish staminate
flowers appear in late March and on into May, and the fruit becomes
red in October.

E. H. Wilson 119 is a mixture of material collected in Hupei and
Sikang. The Hupei material has 4 pyrenes, while the Sikang (Ta-chien-
lu) material has a much prolonged leaf-apex, as well as smaller fruits
maturing only 2 pyrenes each. I have distinguished this latter material
in the Arnold Arboretum herbarium as 119B

The indumentum and the lack of uniformity in the number of pyrenes
of Ilex georgei closely relates it to [lex ciliospinosa Loes., but in the latter
the leaves are less coriaceous and ovate or ovate-elliptic in shape, with a
less prominent apex. The fruit is larger (8-10 mm. long), as are the
pyrenes.

44. ex ciliospinosa Loes. in Sarg. Pl. Wils. 1: 78. 1911; S. Y. Hu
in Ic. Pl. Omei. 2: pl. 167. 1946
Ilex bioritsensis oe var, ciliospinosa (Loes.) Comber in Notes Bot.
Gard. Edinb. 18: 43. 1933,

An evergreen pee or small tree up to 7 m. high, with densely pubescent
branchlets, elliptic or ovate-elliptic weakly spinose leaves, fasciculate
Gafloreseences: — drupes 8 mm. long, and 1-3 oleate striate
and sulcate pyren

Branchlets eae dirty-gray, pubescent; fourth year’s growth 4 mm.
in diameter, reticulat tely rimulose, pubescent, the lenticels lacking, the
leaf-scars broadly deltoid; second year’s growth 3 mm. in diameter, densely
pubescent, rugose; current year’s growth subterete, longitudinally striate,
light brown, densely pubescent, the terminal buds ovoid, pubescent, the
apex cuspidate. Leaves occurring even on the fourth year’s growth, 2-8
mm. apart; stipules callose, often tricuspidate; petioles 2-3 mm. long,
one-fifteenth to one-tenth the length of the lamina, pubescent and rugose;
lamina coriaceous, olivaceous, elliptic or ovate-elliptic, 2.5-4.5 cm. long,
1—2.5 cm. wide; round or rarely obtuse at the base; shortly acuminate or
acute at the apex, the acumen 3-5 mm. long, terminating in a weak spine;
margin serrate, with 4-6 spines on each side; midrib impressed and
pubescent above, elevated and glabrous beneath, the lateral nerves 4—6
on each side, indistinct above, evident beneath, arching upward, near the
margin reticulate. Inflorescences paucifasciculate, fascicles 2—5-flowered,
axillary on last year’s growth, the bud-scales persistent, ovate, acute,
sparsely pubescent, ciliate, the bracts lanceolate, ciliate; the pep 2
median or submedian, not reaching the calyx, individual branches of the
fascicles all anidoreus. the pedicels 2—3.5 mm. long, pubescent or glabres-

1949] HU, THE GENUS ILEX IN CHINA, II 365

cent; flowers 4-merous. Staminate flowers: calyx 3 mm. broad, deeply
lobed, the lobes ovate, deltoid, ciliate, obtuse or rounded at the apex;
corolla rotate, 6 mm. broad, the petals ovate, 3 mm. long, 2 mm. wide,
minutely ciliate at the apex, one-eighth connate at the base; stamens
longer than the petals, the anthers oblong; rudimentary ovary ovoid, 1
mm. in diameter, the apical end indistinctly sulcate. Pistillate flowers:
calyx as in the staminate flowers; corolla erect, choripetalous, 2.5 mm.
long; staminodes about equaling the petals in length, the sterile anthers
sagittate; ovary oblong, 2 mm. in diameter, truncate at the apex, the
stigma discoid. Fruit solitary or paired, seldom 3 in a fascicle, ellipsoid,
8 mm. long, the pedicels 2-4 mm. long, pubescent or glabrescent, the per-
sistent calyx 2-3 mm. across, quadrangular in outline, ciliate; the stigma
thinly discoid. Pyrenes 1-3; obovoid in outline, palmately striate and
sulcate, when 1, suggestive of a wheat-grain with a narrow shallow groove
on the ventral side, when 2, (developing) dorso-ventrally compressed,
slightly convex on the back, flat on the ventral side, when 3, obovoid, 6 mm.
long, 4 mm. wide, the endocarp thick-woody.

CHINA: Szechuan: Mt. Omei, W. C. Cheng 6591 (Sz); (Gimes
Chiao & C. S. Fan 871 (A); W. P. Fang 14800 (Sz); ¥.S. Lin 1447 (A);
PoE GW 2Os Sere al. Wy isdn. 3310. CAa)s Opien-hsien, W. C. Cheng
6216 (Sz), 6456 (Sz); S. N. Hsu 69 (Sz); C. Y¥. Yao 2808 (SS), 4279
(SS), Y. S. Liu 2058 (A); T. T. Yu 639 (A), 663 (A), 835 (A); Chien-
shi-hsien, E. H. Wilson 996 (asotype, A), Wa-shan, £. H. Wilson 996A
(A); locality not clear, E. H. Wilson 3319 (A). Yunnan: Mengtze,
A. Henry 11169 (A).

CULTIVATED: I have seen a specimen from Upper Bank Nurseries,
located at Media, Pennsylvania.

Ilex ciliospinosa is endemic to western Szechuan, where it grows as a
shrub or small tree in bamboo forests or in thickets at an altitude of
1700-2300 m. The greenish flowers appear in early May, and the fruit
becomes red in September.

Comber, presuming that Wilson’s specimens were juvenile forms, reduced
Ilex ciliospinosa to a variety of Ilex bioritsensis Hayata. But Wilson’s
specimens were not juvenile forms, since most of the branchlets show
plainly five years of growth. The leaves on the four-years’ growth are
the same in shape, texture, and margin as those on the current year’s twigs.
Furthermore, the specimen T. T. Yu 663 was collected from a tree 20 feet
high. The spines of the leaves of this specimen are as weak as those
on a 5-foot shrub. In this spinose group of J/ex, the “species” show much
intergradation in the size and dentation of the leaves, in the indumentum,
‘and in the number of pyrenes. Hence the reduction of any of them
to varieties would result, consistently, in further reductions. Probably
all would end up as varieties of Ilex dipyrena Wall., the oldest binomial
of the group. Morphologically there is just as much resemblance and just
as much difference between Ilex ciliospinosa Loes. and Ilex bioritsensis
Hayata as between Jlex bioritsensis and Ilex dipyrena. Until cytological
and genetic techniques can be applied to this group, it is better to keep all
the members in specific rank.

366 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

In its weak and more numerous spines and in the inconsistent number
of its pyrenes (1-3), Jlex ciliospinosa is very Closely related to /lex di-
pyrena, which differs in having large (6-10 cm. long) leaves and glabrescent
branchlets.

45. Ilex bioritsensis Hayata in Jour. Coll. Sci. Tokyo 30: 53. 1911;
Comber in Notes Bot. Gard. Edinb. 18: 42. 1933; Kanehira, For-
mosan Trees 371, fig. 325. 1936.
Ilex veitchii Veitch, New Hard. Pl. W. China 4, 1912, nomen subnudum:
Anon. in Gard. Chron. IIT. 52: 289, 1912.

flex pernyi Franch. var. veitchii Bean, Trees Shrubs Hardy Brit. Isles 1:
650. 1914, nomen nudum; Rehd. in Mitteil. Deutsch. Dendr. Ges. 23:
263, 1914, descr.

Hex diplosperma S. Y. Hu in Ic. Pl. Omei. 2: pl. 163. 1946,

An evergreen shrub or small tree up to 10 m. high, with glabrescent or
sparsely puberulent branchlets, ovate or quadrangular strongly spinose
leaves, fasciculate inflorescences, large fruits (8-10 mm. in diameter) and
two dorso-ventrally compressed, palmately striate and sulcate, thick-woody
pyrenes.

Branchlets glabrescent or sparsely puberulent; third year’s growth sub-
terete, cinereous, 3—4 mm. in diameter, smooth, the lenticels lacking; second
year’s growth 2.5 mm. in diameter, longitudinally striate rugose; current
year’s growth 2 mm. in diameter, striate and ridged, the terminal buds
puberulent, conic, acute at the apex, the scales ciliate. Leaves occurring
even on the fourth year’s growth, 5-15 mm. apart: stipules callose, ovate,
0.8 mm. long, acute, constricted near the base; petioles 3 mm. long, one-
fifteenth to one-twelfth the length of the lamina, pubescent; lamina thick-
coriaceous, olivaceous or ochraceous, shiny above, opaque beneath,
ovate-quadrangular, 2.5-6 cm. long (usually 3 cm. long), 1.5-3.5 cm
(usually 2 cm.) wide; base rounded or truncate; apex acuminate, the acu-
men 5—15 mm. long, terminated by a spine 3 mm. long; margin sinuate,
with 3 or 4 strong spines on each side; midrib sulcate above, puberulent,
elevated beneath, the lateral nerves 4~6 pairs, evidently impressed above,
obscure or slightly elevated beneath; reticulation of the veinlets obscure on
both surfaces. Inflorescences fasciculate, axillary on the second year’s
growth, the individual branches of the fascicles all uniflorous, the bracts
ovate, ciliate, 1.75 mm. long; flowers 2—4-merous. Staminate flowers:
pedicels 2 mm. long, glabrous, with 2 large ovate ciliate prophylla extend-
ing to the calyx; the calyx patelliform, 3 mm. across, the lobes broadly
deltoid, obtuse, sparsely ciliate; corolla rotate, the petals broad-elliptic,
3 mm. long, one-eighth connate at the base; stamens longer than the petals,
the anthers oblong; rudimentary ovary ovoid, 1 mm. in diameter. Pistil-
late flowers: pedicels 2 mm. long, glabrous, with 2 basal prophylla; calyx
as in the staminate flowers; corolla choripetalous; staminodes half the
length of the petals, the sterile anthers cordate; ovary oblong-ovoid, 2—3
mm. long, the stigma thin-discoid. Fruit ellipsoid, 8-10 mm. long; the
stigma discoid. Pyrenes 2, dorso-ventrally compressed, ovate or sub-

1949] HU, THE GENUS ILEX IN CHINA, I 367

orbicular in outline, 5-6 mm. long, 4-5 mm. wide, palmately 8—10-striate
and sulcate on the dorsal surface, 5-striate on the ventral surface, the
endocarp woody.

CHINA: Kweichow: Fan-ching-san, Steward, Chiao & Cheo 390
(A, US), 774 (A, NY, US); P. C. Tsoong 1041 (SS) ; Yin-kiang, Y. Tsiang
7736 (NY), 7965 (NY). Szechuan: Nan-chuan-hsien, W. T. Pang
640 (SS), 976 (A, SS); Mt. ’Omei, W. P, Fang 3097 (A, NY), 12879 (A,
Sz, US), 16431 (Sz) ; O-pien-hsien, T. S. Chao 604 (SS); Muli, J. F. Rock
5211 (A, US), 17368 (A, US); Ning-yuan-fu, C. Schneider 942 (A);
Lang-pa (Lolo-lang), C. Schneider 1070 (A), 1523 (A); Wu-shan, E. H.
Wilson 1028 (A, NY, US); Yung-ching-hsien, Wa-wu-shan, C. WW’. Yao
2224 (SS); without ae locality,.2 °C, Chow 10005. sz). unia.n,.
Iniekianoe i. Vinee, 2 (A); Chung-tien, K. M. Feng 966 (A); Yung-
ning, G. Forrest 205K (ase Che-kia, E. E. Maire 525 (A); Ho-kia-keou,
E. Maire 3330 (NY, US); Yi-liang-hsien, H. T. Tsai 52130 (A); without
precise locality, He ip Yu 5446 (A). Taiwan: N. aro (eae
Suguki 15207 (TU); E. Matuda 1208 (TU); R. Kanehira & S. Sasaki
2180/ CUS) n2bee, Watson 10058 CA, US).

CULE TIVAA ED: = aithave-seen- ster ile specimens of juvenile plants from
Kew and the Arnold Arboretum under the name Ilex pernyt var. veitchi
Rehd.

This plant was first recorded from Bioritsu, Taiwan. Additional mate-
rial has been collected from Kweichow, Szechuan, and Yunnan. Both on
the Island and on the mainland it grows in forests as a small tree up to
10 m. high. It flowers in late April.

All the spinose-leaved J/ex characterized by medium-sized leaves with
rather strong spines and two-pyrened fruits have been placed under this
species. There is great variation in its leaf-form and indumentum. The
stems of some specimens are almost glabrous, but others are as pubescent
as those of Ilex pernyi Franch.

46. ex perryana, sp. nov.

Ilex pernyi sensu Marquand in Jour. Linn. Soc. Bot. 48: 168. 1929;
sensu F. K. Ward in Gard. Chron. III, 92: 232. 1932, non Franch.
lex ae nyi Franch. forma, Anon. in Notes Bot. Gard. Edinb, 17: 270,

1930

Tlex pernyi Franch. var., Anon. in Notes Bot. Gard. Edinb. 17: 304. 1930.
Ilex georgei Comber var. rugosa Comber in Notes Bot. Gard. Edinb. 18:
51. 1933: Merrill in Brittonia 4: 100. 1941.

Frutex prostratus; ramulis brevissime pubescentibus; foliis spinosis,
crasse coriaceis, supra nitidis et rugosis, subrhomboideis, raro ellipticis,
1-3 cm. longis, 7-15 mm. latis, basi obtusis, raro rotundatis, costa supra
impressa, subtus obscura; inflorescentiis paucifasciculatis, axillaribus;
4 unifloribus, pedicellis 23.5 mm. longis, calycibus ciliatis, corolla rotata,
petalis 1.5 mm. longis, eciliatis, staminibus quam petalis aequilongis vel
longioribus; @ unifloribus, pedicellis 3-4 mm. longis, ovario ovoideo,
5 mm. longo; fructibus subgloboso-ellipsoideis, 6-7 mm. diametro;
pyrena 1, subglobosa, dorso-ventraliter compressa, 5 mm. diametro, dorso

368 JOURNAL OF THE ARNOLD ARBORETUM [VOL. xxx

palmatim striata et esulcata, ventro canaliculato, endocarpio lignescente.

A prostrate evergreen spinose shrub 20-30 cm. high with stout puberu-
lent branchlets, small hexangularly subrhomboid rugose leaves, pauci-
fasciculate inflorescences, small subglobose fruits, and a single palmately
striate esulcate subglobose pyrene in each fruit.

Branchlets stout, straight, tuberculate with elevated leaf-scars, normal
annual growth 2-3 mm. long, the third year’s growth and the current
year’s growth almost alike in thickness and texture, 3-4 mm. in diameter,
minutely puberulent, the lenticels lacking, the terminal buds ovoid, puberu-
lent, well protected by crowded leaves or petioles. Leaves occurring even
on sixth year’s growth, 2-3 mm. apart; stipules callose, ovate, ca. 0.6 mm.
long, acute, slightly constricted at the base; petioles 1-2 mm. long, one-
fifteenth to one-tenth as long as the lamina, glabrous or minutely puberu-
lent, dorso-ventrally flattened; lamina thick-coriaceous, olivaceous or
brunneous-olivaceous, very shiny and rugose above, opaque and smooth
beneath, hexangular-subrhomboid, rarely elliptic, 1-3 cm. long, 7-15 mm.
wide; obtuse, rarely rounded at the base; broadly deltoid at the apex,
the tip acute, terminated by a spine 2 mm. long; margin subentire-sinuate,
with 2 or 3 spines on each side; midrib deeply impressed above, glabrous,
slightly elevated beneath, the lateral nerves 3 or 4 pairs, deeply impressed
above, obscure beneath, the reticulation of the veinlets impressed above,
inconspicuous beneath. Inflorescence paucifasciculate, axillary on second
year’s growth; fascicles with 2—4 flowers, the bud-scales persistent,
tilaginous, suborbicular, glabrescent; flowers all 4-merous. Staminate
inflorescences: individual branches of the fascicles uniflorous, the bracts
ovate, obtuse, sparsely puberulent, ciliate; pedicels 2-3.5 mm. long,
glabrous, with 2 submedian or basal glabrous ciliate prophylla; calyx
patelliform, 3 mm. across, deeply 4-lobed, lobes ovate, 0.9 mm. long,
0.75 mm. wide at the base, glabrous, minutely ciliate; corolla rotate,
4-5 mm. across, the petals ovate, 1.5 mm. long, eciliate, one-sixth connate
at the base; stamens equaling or slightly longer than the petals, the anthers
ovoid-oblong, 0.75 mm. long; rudimentary ovary pulvinate, obscurely
lobed. Pistillate inflorescences: individual branches of the fascicles uni-
florous, the calyx and corolla as in the staminate; ovary ovoid, 1.5 mm.
long, constricted below the discoid stigma. Fruit subglobose-ellipsoid,
6-7 mm. in diameter, the fruiting pedicels 3-4 mm. long, the stigma
discoid, orbicular. Pyrene 1, subglobose, slightly dorso-ventrally com-
pressed, 5 mm. in diameter, palmately striate and esulcate on the back,
canaliculate on the ventral side, the endocarp thickly woody.

CHINA: Yunnan: G. Forrest 17538 (A), 18006 (A); Newahlung,
ea Yh 19530 (A) s Barcuhwane. fT. Va 20042 (TYPE, A).

NORTH BURMA: F, K. Ward 224 (A), 9873 (A).

flex perryana is a prostrate dwarf shrub confined to the alpine flora of
eastern Himalaya, where it grows on open rocky slopes at altitudes of
3000-3650 m. The flowers appear in June and the fruit turns red in
November.

an

1949] HU, THE GENUS ILEX IN CHINA, II 369

In its small leaves, few spines, and subglobose pyrenes this alpine Jlex
is closely related to flex pernyi Franch., which differs in having leaves
with a smooth surface, prolonged apex, and truncate base, and fruits which
are consistently 4-pyrened. It can easily be distinguished from all the
other spinose species of //ex by its prostrate habit, and its striate esulcate
pyrenes.

The descriptions of the staminate and pistillate flowers were drawn
respectively from Ward 9873 and Yu 19330.

his species is named in honor of Dr. Lily M. Perry of the Arnold
Arboretum, whose assistance and advice has been of great value.

Series 3. DENTICULATAE, ser. Nov.

Arbor vel frutex; foliis crasse coriaceis, raro subcoriaceis, crenulato-
serratis, serratis vel raro integris; inflorescentiis fasciculatis, pseudopanicu-
latis vel pseudoracemosis; floribus 4-meris; corolla basi connata, ?
choripetala; fructibus globosis, stigmate umbilicato; pyrenis rugosis et
caveatibus, raro palmatim striatis; endocarpio lapidoso

KrEY TO THE SPECIES
A. Ovary and fruit pubescent.
Fruiting pedicels ca. 4 mm. long; leaves elliptic or ovate-elliptic,
3-7 cm. long, 1.3-2.5 cm. wide, the lower pees LU aieabe peti-
oles 3-6 mm. long. (Hainan and Indo-China)..47. 7. pubilvinba.
BB. Fruiting pedicels ca. 8 mm. long; leaves pee or eee elliptic,
. long, 4.5-7.5 cm. wide, the lower surface glabrous;
petioles 20-22 mm. ee (@ck-nbotcho ® pareerwrre re 48. I. kaushue.
AA. Ovary fruit glabro
B. Branchlets, leaves a inflorescences densely hirsute. (Kwangsi)..
ee ee meet rte bs cod nck: 5 Sy an ah eS 49, J. nanningensis.
BB. Branchlets and inflorescences glabrous, glabrescent or puberulous,
the leaves always glabrou
Leaves thick-coriaceous, 10-17 cm. long, 4.5-7.5 cm. wide;
individual branchlets of the staminate re ero 3-9-
East China and apa) cose 50. J. latifolia.
Leaves coriaceous or subcoriaceous, rarely up to 10 cm. long;
individual branchlets of the staminate inflorescence 1-3-
flowered.
Leaves entire; fruits 5 mm. in diameter, smooth when
Chigyge a aA ga 24k gist Seach aah eee ee 1. J. susukit.
DD. ee crenulate-serrulate, serrate or rarely subentire;
fruits 6-15 mm. in diameter, when dry args wrinkled.
Ee. “Fr uits 5-7 mm. in diameter; pyrenes 4-5 1. long.
F. Leaves coriaceous, the apex See or are
shrvpty re shortly acuminate; fruiting pedicels
4-8 mn
Cr ee glabrous; margin of leaves dense-
ly and often irregularly serrate; infructes-
cences pseudoracemose or fasciculate; fruit-
ing pedicels 6-8 mm. long. ndia and
Southwest China)...... D2. We Neen

Q
O

370 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

GG. Branchlets puberulous; margin of leaves re-
motely serrate; infructescences fasciculate,
the fruiting pedicels ca. 4mm, long. (Hong-
kong and vicinity)......53. I. graciliflora.

FF. Leaves subcoriaceous, the apex long-acuminate,
the acumen 10-16 mm. long; fruiting pedicels

10-12 mm. long. (Kwangtung)..54. J. tsangii.

EE. Fruit 10-15 mm. in diameter; pyrenes 7-12 mm. long.
I. Mature fruit solitary; fruiting pedicels 2-4 mm.
long; pyrenes obovate in outline, 10-12 mm. long.

TW acne 5.8 Geueex I. chingiana.

FF, Mature fruits eee fruiting pedicels 6-10
mm. long; pyrenes 7-9 1 ong.

x. Leaves elliptic o eeied = reticulation of the
veinlets a ne on both surfaces; pyrenes
oblong in outline, both ends obtuse or one
end emarginate. (Taiwan)..............
hide hl ete an eases I, uraiensis.

GG. Leaves oblong, ovate- elliptic c or oblanceolate ;
reticulation of the veinlets obscure above:
pyrenes broad-elliptic in outline, the ends
pointed.

H. Branchlets castaneous or nigrescent;:
leaves shiny above with caudate apex,
the acumen 7-16 mm. long; the base ob-
tuse or rounded. ooo eee oree
peewee ens eeds je eee diesen

HH. Branchlets cinereous ; re Ss Opaque on
yoth surfaces, wjth Aearainnte apex, the

acumen 5-10 mm. long; base cuneate.
CT ata se taaeax 58. J. nuculicava.

47. Ilex pubilimba Merr. & Chun in Sunyat. 5: 109. 1940.
flex aio Tardieu-Blot in Not. Syst. se a 120. 1945, et in FI.

Gén. Indo-Chine Suppl. 1: 776. 1948. ‘Syn, no
A Saas pubescent evergreen tree up to 15 n m. hieh with coriaceous,
elliptic, shortly acuminate, crenate-serrate leaves, fasciculate inflorescences,
pubescent ovary, short fruiting pedicels (2—4 mm.), and depressed globose
fruits 7 mm. in diameter

Branchlets straight, subterete, fuscous when dry; third year’s growth
4 mm. in diameter, pubescent, Toncitudinally, rimulose, the lenticels ob-
scure, the leaf-scars semi-orbicular; second year’s growth pubescent;
current year’s growth 2—2.5 mm. in diameter, subterete, very densely
pubescent, the indumentum fuscous; terminal buds small, naked, pubes-
cent. Leaves occurring also on the second year’s growth, 2-12 mm. apart;
stipules callose, deltoid, pubescent at the base; petioles 3-6 mm. long,
one-twentieth to one-twelfth the length of the lamina, densely pubescent,
canaliculate above; lamina thick-coriaceous, when dry cinereous-olivaceous
and shiny above, paler with over-all pubescence beneath, elliptic, rarely

1949] HU, THE GENUS ILEX IN CHINA, II 371

ovate-elliptic or lanceolate, 3-7 cm. long, 1.3—2.5 cm. wide; base rounded,
obtuse or cuneate; apex shortly acuminate, the acumen 3-10 mm. long,
the tip obtuse; margin crenate-serrate, slightly recurved when dry; midrib
impressed and pubescent above, elevated beneath, lateral nerves obscure.
Inflorescence fasciculate, axillary, on last year’s growth, short-pedicellate;
bracts ovate to semi-orbicular, densely pubescent; flowers 4-merous.
Staminate flowers: individual branches of the fascicles 1-—3-flowered,
3-flowered branches cymose, the peduncle 0.5-1 mm. long, the pedicels
2-3 mm. long, with 2 basal prophylla; calyx patelliform, 1.5—-2 mm. in
diameter, rugose, 4-lobed, the lobes rounded, ciliate; corolla rotate, a mm.
in diameter, the petais oblong, 2.5 mm. long, ciliate, one-ninth connate at
the base; stamens slightly shorter than the petals, the anthers ovoid, 0.75
mm. long; rudimentary ovary conical, 1 mm. in diameter, the apex obtuse.
Pistillate flowers: individual branches of the fascicles uniflorous, the calyx
and corolla as in the staminate flowers; staminodes one-third the length
of the petals, the sterile anthers ovate; ovary subglobose, 1.75 mm. in
diameter, eeeeely but distinctly pubescent, the stigma discoid! convex.
Fruit depressed globose, 7 mm. in diameter, the fruiting pedicel e 4 mm.
long, pubescent, the persistent calyx explanate, orbicular in outline, 2 mm.
across, shallowly 4-lobed, the lobes ciliate, the stigma thinly discoid.
orbicular. Pyrenes 4, suborbicular or oblong in outline, 4-5 mm. long,
3.5-4 mm. wide, rugose, flattened or slightly impressed on the back, the
endocarp woody.

CHINA: Hainan: Kan-en, S. K. Law 5974 (A); Bak-sa, S. K. Lau
25742 (material for the description of the staminate flower) (A); S. K. Laut
26069 (material upon which-description of the pistillate flower is drawn)
CAN), Vea-chowse ey weiang 020/79 (A, KK INV Us) 02027 oe A; K,
NY, US); without precise locality, H. Y. Fiepe 63624 (A, NY); C. Wang
35188 (A, NY)

DO Tonkin: W.T. Tsang 30339 (A); Annam: Pot-
lane 8235 (A, isotype of I. hirsuticarpa).

Ilex pubilimba Merr. & Chun has been recorded from Hainan Island
and Indo-China, where it grows as a big tree in mixed forests. The whitish
yellow flowers appear in March.

The vegetative characters, especially the indumentum, and the leaf form
of Ilex pubilimba are very similar to those of the very closely allied species
Ilex nanningensis Hand.-Mzt. Professor Merrill has written on the type
specimen and on other Hainan material “== /lex nanningensis Hand.-
Mzt.” On closer examination, however, these two species can easily be
distinguished from each other by the size of their fruits, the length of the
petioles and fruiting pedicels, and the indumentum of the ovary. In /lex
nanningensis the fruits are 10 mm. in diameter, the petioles one-tenth to
one-sixth as long as the lamina, the fruiting pedicels 6-8 mm. long, and the
ovary glabrous. In Jlex pubilimba the fruits are 7 mm. in diameter, the
petioles one-twentieth to one-twelfth as long as the lamina, the fruiting
pedicels 3-4 mm. long, and the ovary pubescent.

372 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

48. Ilex kaushue, sp. nov.

Arbor; ramulis pubescentibus; foliis oblongis vel oblongo-ellipticis,
10-15 cm. longis, 4.5—7.5 cm. latis, basi obtusis vel cuneatis, apice acutis
vel breviter acuminatis, margine serratis vel biserratis; costa supra impressa
pubescente, subtus elevata, nervis lateralibus 14 vel 15 paribus; inflores-
centiis fasciculatis vel pseudoracemosis, pedicellis 8 mm, longis, pubes-
centibus; fructibus subglobosis, 11 mm. longis, 9-10 mm. diametro;
pubescentibus; calycibus persistentibus ciliatis, stigmate umbilicato;
pyrenis 4, ca. 7.5 mm. longis, 4.5—-5 mm. latis, endocarpio lapidoso,

An evergreen tree with densely puberulent branchlets, large oblong thin-
coriaceous serrate or double-serrate leaves, fasciculate infructescence, large
subglobose puberulent fruits, and 4 reticulately striate sulcate wrinkled
and pitted pyrenes.

Branchlets straight, subterete, brown, longitudinally ridged and canalicu--
late; second year’s growth 4 mm. in diameter, the lenticels lacking, the
leaf-scars deltoid-semi-orbicular; current year’s growth angular, longi-
tudinally plicate, 3.5 mm. in diameter; densely puberulent especially in the
grooves; terminal buds acute-conic, pubescent. Leaves occurring also on
second year’s growth, 10-18 mm. apart: stipules callose, often hidden;
petioles pubescent, slender, 20-22 mm. long, one-sixth to one-fifth the
length of the lamina, subterete and angular, very narrowly canaliculate
above; lamina thin-coriaceous, brunneous-olivaceous, shiny above, paler
and opaque beneath: oblong or oblong-elliptic, 10-18 cm. long, 4.5-7.5
cm. wide; base obtuse or cuneate; apex acute or very shortly acuminate,
the acumen 3-6 mm. long; margin serrate or double-serrate; midrib im-
oressed, sparsely and minutely puberulent above, prominent, elevated
and keeled beneath, the lateral nerves in 14 or 15 pairs, prominent on both
surfaces; the reticulation of the veinlets dense and evident on both
surfaces. Infructescences fasciculate or pseudoracemose, the rachis up to
9 mm. long, densely puberulent; bracts caducous, the lower persistent ones
cartilaginous, broad-deltoid, obtuse, coarsely ciliate, the basal appendages
acute, pubescent; pedicels 8 mm. long, pubescent; persistent calyx ex-
planate, ca. 5 mm. across, puberulent, 4-lobed, the lobes broad-deltoid,
very acute or cuspidate, sparsely ciliate. Fruit subglobose or ellipsoid,
ca. 11 mm. long, 9-10 mm. in diameter, the exocarp densely puberulent, the
stigma navel-like. Pyrenes 4, oblong in outline, ca. 7.5 mm. long, 4.5—5
mm. wide, reticulately striate and sulcate, wrinkled and pitted, the endo-
carp stony.

CHINA:
Ay LU, US)

This species is endemic to Hainan Island. The local name is “kau-
shue.” In September the fruits become yellow.

The large leaves, the pseudoracemose infructescences, and the wrinkled
and pitted pyrenes of /lex kaushue suggest a relationship with Jlex lati-
folia Thunb., but the latter has smaller glabrous fruits, glabrous branchlets,
and thick-coriaceous leaves with no evident reticulations of the veinlets.

Hainan: Taam-chau, W.T. Tsang 864 (LU 16363) \(exre,

1949] HU, THE GENUS ILEX IN CHINA, II 373

On the labels distributed by the Herbarium of Lingnan University, the
notation “flowers yellow” is given. Since the specimens at hand show no
evidence of the flowering stage, I assume that there may have been a
mechanical error in the preparation of the labels. The color yellow as
given undoubtedly refers to the fruit, since the specimen was collected
in September.

49. Ilex nanningensis Hand.-Mat. in Sinensia 5: 2. 1934; Chun in
Sunyat. 4: 225. 1940.

A densely pubescent evergreen tree up to 20 m. high with coriaceous
elliptic or lanceolate, shortly acuminate crenate-serrate leaves, fasciculate
inflorescences, large subglobose fruits, 10 mm. in diameter, and long fruiting
pedicels (8 mm. long).

Branchlets straight, fuscous when dry; third year’s growth 4 mm. in
diameter, pubescent, smooth, inconspicuously longitudinally rimulose, the
lenticels lacking, the leaf-scars semi-orbicular; second year’s growth densely
pubescent; current year’s growth subterete or angular, 2.5 mm. in diameter,
fuscously and densely pubescent, the terminal buds naked, small, pubes-
cent. Leaves occurring even on the fourth year’s growth, 3-10, sometimes
up to 15 mm. apart; stipules callose, deltoid, falcate, very acute, pubescent
at the base; petioles 7-10 mm. long, one-tenth to one-sixth the length of
the lamina, pubescent, the distal half narrowly winged by the decurrent
leaf-base, canaliculate above; lamina thick-coriaceous, cinereous-olivaceous
when dry, shiny above, dusty brown beneath, elliptic, very rarely lanceo-
late or ovate-elliptic, 5-8 cm. long, 1.5-3.5 cm. wide; obtuse or cuneate
at the base; very shortly and gradually acuminate at the apex, the acumen
broad-deltoid, 5-12 mm. long, often nigrescent; margin crenate-serrate,
slightly recurvate when dry; midrib impressed and pubescent above,
elevated beneath, the lateral nerves 7 or 8 pairs, obscure above, evident
beneath, the reticulation of the veinlets sometimes evident beneath. In-
florescence fasciculate, axillary, on the second year’s growth, fascicle
2-5-flowered, the bracts ovate-reniform, mucronate, or cuspidate, ciliate;
flowers 4-merous. Pistillate flowers: individual branches of the fascicles
uniflorous, the pedicels 6-8 mm. long, pubescent, with 2 submedian long-
deltoid pubescent prophylla; calyx patelliform, 3 mm. across, the lobes
shallow, rounded, 0.5 mm. long, 1.5 mm. wide, rugose, very sparsely and
minutely puberulent, ciliate; corolla rotate, 8 mm.. in diameter, the petals
3 mm. long, 2 mm. wide, the apical portion ciliate; staminodes slightly
shorter than the petals, the sterile anthers sagittate; ovary large, ovoid,
3 mm. long, 2.5 mm. wide, the apical end truncate, the stigma discoid.
Staminate flowers not seen. Fruit subglobose, 8 mm. long, 10 mm. in
diameter, the stigma flat, orbicular, 2 mm. across, the persistent calyx
explanate, orbicular, 3 mm. across, ciliate; the fruiting pedicels 8 mm.
long, the exocarp thin, the mesocarp fleshy, abundant. Pyrenes 4, oblong
in outline, 6.5 mm. long, 5.5 mm. wide, rugose, reticulately striate, flattened
and broadly depressed on the dorsal surface, the endocarp woody.

374 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

CHINA: Kwangtung: Sin-fung, Y. W. Taam 271 C2): wea aie
si: Seh-feng-dar-shan, RF. C. Ching 8082 (isotypes LU, NY; fragments,
A); Sup-man-ta-shan, H. Y. Liang 69851 (A); Shap-man-taai-shan, W. T.
Tsang 22073 (A). Hainan: Ba-tung, S. K. Lan 28238 a),

flex nanningensis was first described from a specimen collected in the
high mountains south of Nanning, Kwangsi Province. From the material
at hand, it appears to be endemic to the mountain range between Kwangsi
and Indo-China. There it grows as a big tree in woods or mixed forests
at an altitude of 600 m. The fragrant yellow flowers appear in April and
the fruit turns bright red in October.

It should be noticed that the Kwangsi localities all refer to one mountain
range, the name of which is variously transliterated by persons speaking
different dialects. The preferred spelling is “Shih-wan-ta-shan.”

The large fruit, the big rugose pyrenes, and the form and shape of the
leaves of this species ally it closely to /lex subficoidea S. Y, Hu, which
differs in having glabrous branchlets and leaves.

50. Hex latifolia Thunb., FI. Jap. 79. 1784; Willd. Sp. Pl. EA 708.
1797; Persoon Syst. Veg. 174, L727 2et Sync Phe le151. 1805e- Pair,
in Lam. Encycl. Suppl. 3: 66. 1813; Roem. & Schult., Syst. 3: 488.
1818; DC. Prodr. 2: 16. 1825: Spreng. Syst. 1: 494. 1825: Dietr.
Syn. Pl. 1: 556. 1839; Sieb. & Zucc. in Abh. Bay. Ak. Wiss. IV, 2:
143 (Fl. Jap.). 1845; Regel in Gartenfl. 13: 38. 1864; Hook. in
Bot. Mag. 22: pl. 5597. 1866; Miq. in Ann. Mus. Bot. Lugd.-Bat.
3: 105. 1867; Maxim. in Mém. Acad. St. Petersb. VII, 29(3): 29,
43, pl. 1, fig. 2. 1881; Loes. in Nov. Act. Acad. Caes. Leop.-Carol.
Nat. Cur. 78: 289 (Monog. Aquif. 1: 289). 1901; Moore, Holly
Yew Box 127. 1908; Chung in Mem. Sci. Soc. China 1: 140. 1924;
Chien in Contr. Biol. Lab. Sci. Soc. China 3(1): 58.1927; Rehd.
in Jour. Arnold Arb. 8: 157. 1927; Hand.-Mzt., Symb. Sin. 7:
656. 1933; Belval, Mus. Heud. Nat. Bot. Chin. Deri L035. Kaine
PL. Sin. Ill. 491, fig. 830. 1937; Chen, Ill. Man. China Trees 660,
fig. 552. 1937.
flex macrophylla Blume Bijdr, 1150, 1825; Dietr., Syn. Pl. 1: 556. 1839 ;
Walp. Rep. 1: 540, 1842.

/lex tarajo Hort. Gépp. in Gartenfl. 3: 325. 1854.

Ilex terago Anon. in Gard. Chron, 1859: 508, 972. 1859.

A very large, entirely glabrous evergreen tree up to 20 m. high, trunk
60 cm. in diameter (ex Ching) with large thick-coriaceous oblong serrate
leaves, axillary paniculate staminate and racemose pistillate inflorescences,
globose fruit with thick exocarp, and 4 bony wrinkled pyrenes. .

Branchlets straight, stout, ochraceous or brown; second year’s growth
5-7 mm. in diameter, longitudinally plicate, often rimose, the lenticels
lacking or sparse, the leaf-scars elevated, broad-deltoid or semi-orbicular ;
current year’s growth 4—5 mm. in diameter, striate, sulcate and angular,
smooth and glabrous, tuberculate under a lens. Leaves occurring even
on the third year’s growth, 8-17 mm. apart; stipules very minute and

1949] HU, THE GENUS ILEX IN CHINA, II 375

callose, broad-deltoid, acute; petioles thick, short, rugose, 15-20 mm. long,
one-eighth to one-seventh the length of the lamina, 3 mm. in diameter,
subcylindrical, slightly impressed above, glabrous; lamina thick-coriaceous,
olivaceous, shiny above, opaque beneath, oblong or ovate-oblong, 8-17
cm. long, 4.5-7.5 cm. wide; base rounded; apex obtuse or very shortly
acuminate, the acumen 3-6 mm. long; margin serrate, the teeth nigrescent;
midrib impressed above, strongly elevated beneath, the lateral nerves
15-17 pairs, obscure on both surfaces. Inflorescences pseudopaniculate,
sessile, axillary, on the oe year’s growth; rachis 1-2 cm. long, 2-3 mm.
thick, the bud scales persistent, imbricate, the outer ones cartilaginous,
rounded, ca. 3 mm. long, 5 mm. wide, the inner ones membranous, larger;
flowers 4-merous. Staminate inflorescence: individual branches of the
pseudopanicles 3—9-flowered, cymose, the bracts ovate, 5-7 mm. long, 3-5
mm. wide, the peduncles 2 mm. long, the bracteoles hyaline, ovate or
lanceolate, 2-3 mm. long, the pedicels 6-8 mm. long, with 1 or 2 sheath-
like, broad-deltoid basal prophylla; calyx subcupuliform, 3.5 mm. in
diameter, shallowly 4-lobed, the lobes rounded, ca. 0.75 mm. long, 1.75
mm. wide; corolla rotate, reflexed. ca. 9 mm. across, the petals ovate-oblong,

ca. 3.5 mm. long, 2.5 mm. wide, one-eighth connate at the base; Perens
equaling the petals in length, the anthers ovoid-oblong, ca. 1.25 mm. long,
the filament twice as long; rudimentary ovary subglobose, ca. 1 mm. in
diameter, the apex introrse, 4-lobed. Pistillate inflorescence: individual
branches of the pseudopanicles 1—3-flowered, the peduncles of the 3-flowered
cymes 2 mm. long, the pedicels of the uniflorous ones 5-8 mm. long, with
1 or 2 basal prophylla; calyx patelliform, ca. 3 mm. across; corolla erect,
ca. 5 mm. across, the petals ovate, ca. 3 mm. long, 2 mm. wide; staminodes
one-third the length of the petals, the sterile anthers minute, ovate; ovary
ovoid, ca. 2 mm. wide at the base, the stigma 4-lobed discoid. Fruit
globose, ca. 7 mm. in diameter, brown or reddish when dry, the stigma
thinly discoid, explanate, the exocarp thick, smooth, minutely tuberculate
under a lens, the persistent calyx explanate, 3-4 mm. across. Pyrenes 4,
oblong-elliptic in outline, ca. 5 mm. long, 2.5 wide, recall
wrinkled and pitted, the dorsal surface with 3 sane longitudinal ridges,
the endocarps bony.

CHINA: Kiangsu : ee Ching & Tso 630 (A); K. Ling 12377
(NY). Anhwei: Chu-wha-shan, R. C. Ching 2626 (A, LU); Chi-men,
R. C. Ching 3155 (A); Wang-shan, A. N. Steward 7140 (A, US). Cheki-
ang: Tien-tai-shan, R. C. Ching 1436 (A, US); Chen-chion, R. C. Ching
1/80" (A,, LLU, US) | Tien- a shan, R. C. Ching 5113 (A); Ning-po, E.
Faber (NY). Fukien: n 2474 (A

JAPAN: Honshu, R. K. “ aitie & Y, Kurihara 1014 (A); U. Faurie
ges” CADE eee Maximowicz in 1863 (G, NY); Yokohama, Mavt-
mowics in 1862 (NY); K. Sakurai in 1901 (A), in 1913 (A); Hondo, K.
Shiota 7421 (A); Siebold (A); Mt. Kirishima, E. H. Wilson 6253 (A),
10340 (A); Tokyo, E. H. Wilson 6378 (A); C. S. Sargent (A); K.
Watanabe in 1889 (G).

CULTIVATED: I have seen specimens from various botanical gardens
of Europe and America.

376 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Ilex latifolia was first recorded from Japan. Specimens have been col-
lected from a few of the coastal provinces of China. It has been reported
as common in central Chekiang Province. It grows in woods at an alti-
tude of 100-700 m. and the flowers, which appear in April, are yellowish
green. The fruits turn red in October

In Chekiang and Fukien provinces the wood of this species is used for
rolling-pins and other turned work. The bark is used for making bird-
lime and the leaves are used as a substitute for tea.

The large thick glossy leaves of Jlex latifolia are very characteristic.
There is no other species of Chinese //ex that has the like. In the struc-
ture of the inflorescence and flowers it is closely related to Ilex denticulata
(Wall.) Wight, which differs in having smaller leaves and fruits.

51. Ilex suzukii, sp. nov.

Frutex; ramulis glabris; foliis coriaceis, ellipticis, 2.5-4 cm. longis,
1.5—2.2 cm. latis, integris, basi acutis vel cuneatis, apice breviter acumi-
natis vel obtusis, acuminibus 3 mm. longis, costa supra plana, subtus
elevata, nervis Lendl utrinque 5 vel 6, subtus prominentibus; inflo-
rescentiis fructiferis paucifasciculatis, axillaribus, pedicellis 4 mm. longis,
glabris; fructibus globosis, 5 mm. diametro, stigmate umbilicato, pyrenis
4,3 mm. longis, 2 mm. latis, palmatim striatis, esulcatis, lignescentibus.

An evergreen shrub with glabrous branches, small slintie entire leaves,
fasciculate infructescences, globose fruit with a navel-like stigma and
palmately striate, esulcate pyrenes.

Branchlets nigrescent, glabrous; third year’s growth terete, 2.5 mm. in
diameter, the lenticels lacking, the leaf-scars semi- -orbicular, slightly ae
vated; iment year’s growth 1.5 mm. in diameter, longitudinally ridged,
angular, the axillary buds large, globose, the Seminal bud acute-conic,
glabrous, the scales densely ciliate, Leaves occurring even on the fourth
year’s growth, 2-6 mm. apart; stipules very minute, often obscure; peti-
oles 5-7 mm. long, one-eighth to one-fifth as long as the lamina, namOely
canaliculate above, glabrous; lamina coriaceous, brunneous- Glivaceous when
dry, somewhat aay above, opaque beneath, elliptic, 2.5-4 cm. long,

.5-2.2 cm. wide, acute or cuneate at the base, obtuse or very shortly
acuminate at the apex, \the acumen 3 mm. long; margin entire; midrib
plane above, slightly elevated beneath, the lateral nerves 5 or 6 pairs,
obscure above, prominent beneath, the reticulation of the veinlets obscure
above, prominent beneath. F ower: not seen. Infructescences pauci-
fasciculate, axillary, on the second year’s growth, the fascicles with 2 or 3
fruits, the individual branches of the fascicles uniflorous, the bracts sub-
orbicular or ovate, ciliate; fruiting pedicels 4 mm. long, glabrous, with
2 minute deltoid ee prophylla at the base. Fruit globose, 5 mm. in
diameter, castaneous, smooth and shiny when dry, the persistent calyx
explanate, orbicular in outline, 2.5 mm. in diameter, shallowly 4-lobed, the
lobes rounded, ciliate, the stigma navel-like, 4- lobed. Pyrenes 4, oblong-
obovate in euiline, the ends obtuse, 3 mm. long, 2 mm. wide, palmately
striate and esulcate, the striae slightly elevated, the endocarp woody.

1949] HU, THE GENUS ILEX IN CHINA, II 377
CHINA: Taiwan: Taheizan, S. Suzuki in 1928 (TYPE, A; ISOTYPE,
ae)

The entire leaves, the prominent venation, the striate and esulcate
pyrenes of /lex suzukii, indicate close relationship with Ilex mertensii
Maxim. However, the latter species has large thick-coriaceous leaves
5-9 (-16) cm. long and, also, large fruits 8-9 mm. in diameter. This
species has been mistakenly identned as Ilex hanceana Maxim. in the
past. This last species can readily be distinguished by the pubescent
branchlets, the short petioles (2-5 mm. long), the obscure veins, the
fruiting pedicels 1-1.5 mm. long, and the smaller fruits.

52, Ilex denticulata Wall. List 4333. 1830, nom. nud.; Wight, Il.
Ind. Bot. 2: 147, pl. 149. 1850; Beddome, FI. Sylv. 1: 142. 1869:
Hooker f., Fl. Br. Ind. 1: 600. 1875; Maxim. in Mém. Acad. St.
Petersb. VIL, 29(3): 29. 1881; Loes. in Nov. Act. Acad. Caes. Leop.-
Carol. Nat. Cur. 78: 320 (Monog. Aquif. 1: 320). 1901.

lex milagtrica Mig. ex Hook. f., Ll. c.

A glabrous evergreen tree with coriaceous ovate-elliptic, elliptic, or
oblong leaves, obtuse or shortly acuminate apices, serrate margins, pseudo-
paniculate staminate and pseudoracemose pistillate inflorescences, globose
fruits, and stony, irregularly striate, sulcate and wrinkled pyrenes.

Branchlets straight, glabrous, brunneous, second year’s growth subterete,
3—5 mm. in diameter, minutely rimulose, the lenticels lacking, the leaf-scars
deltoid, not elevated; current year’s growth subangular, striate, 3 mm.
in diameter; terminal buds narrowly conic, the scales cartilaginous, ecili-

eaves occurring also on second year’s growth, 1-2 cm. apart;

stipules lacking or very minute, callose and nigrescent; petioles rather
stout, 10-13 mm. long, one-fifth to one-seventh the length of the lamina,
glabrous, narrowly or not at all sulcate above, rugose beneath; lamina
coriaceous, olivaceous or cinereous-olivaceous, ovate-elliptic or elliptic-
oblong, 5-10 cm. long, 2.5-3.7 cm. wide; base obtuse or cuneate; apex
very shortly and abruptly acuminate or obtuse, the acumen 3-8 mm.
long; margin densely and often irregularly denticulate-serrate, the teeth
nigrescently apiculate; midrib glabrous, narrowly impressed above, ele-
vated beneath, the lateral nerves 6-9 pairs, obscure above, prominent
beneath, the reticulations obscure above, evident beneath. Staminate
inflorescences: pseudo-paniculate, axillary, sessile, only on the second
year’s growth, the rachis 4-14 mm. long, the individual branches 1-3-
flowered, the bracts obovate, acute, 4 mm. long, 2-5 mm. wide; 3-flowered
cymes with peduncles 2 mm. long, the pedicels 2-3 mm. long, the uni-
flowered pedicels 5 mm. long, glabrous, with 2 submedian deltoid prophylla;
flowers 4-merous; calyx patelliform, 2.5 mm. across, glabrous, the lobes
deltoid, ca. 0.57 mm. long, 1 mm. wide, acute or acuminate, entire, eciliate;
corolla rotate, 7 mm. across, the petals oblong, reflex, 3 mm. long,

mm. wide, one-tenth connate at the base; stamens equaling the petals in

length, the anthers small, ovoid, ca. 0.75 mm. long; rudimentary ovary

globose, the apex rounded and slightly depressed. Infructescences fascicu-

378 JOURNAL OF THE ARNOLD ARBORETUM LVOL. XXX

late or pseudoracemose, the central axis 5-10 mm. long, the individual
branches uniflorous, the fruiting pedicels 6-10 mm. long, glabrous or
sparsely puberulous. Fruits globose or depressed globose, 6-7 mm. in
diameter, brown and smooth, the persistent calyx explanate, 3 mm. in
diameter, the stigma navel-like, often depressed, 4-lobed. Pyrenes 4
oblong or elliptic in outline, irregularly striate, sulcate, wrinkled and
pitted, ca. 4-5 mm. long, 3 mm. wide, the endocarp stony.

CHINA: Yunnan: Fo-hai, C. W. Wang 77406 (A).

INDIA: Madras, D. Brandis 666; Nilghiri, Ed. Hohenacker 1455 (A),
1456 (fragment, A); Culgherris, Perrottet (A); Nilghiri Hiliss St,
Wilson, Wight 437 (1sorypr, G, NY), 438 (NY), 490 (fragment, A).

Ilex denticulata was first recorded from Nilghiri Hills of India as a
large tree up to 12 m. high. Specimens collected from the type locality
vary in the degree of their pubescence. Some of them (Wight 437, 490;
Hohenacker 1455) are entirely glabrous, and others (Hohenacker 1456)
are with slightly pubescent pedicels. In the form, size, and margin of the
leaves, pubescence of the inflorescences, size of the fruit, and number and
nature of the pyrenes, the Chinese specimens collected from southwestern
Yunnan match the Indian type very well, with the exception of the more
pubescent pedicels. In Yunnan, the plant grows to bea big tree up to 12m
high. It occurs in mixed forests at an altitude of 2000 m.

In its irregularly wrinkled and pitted stony pyrenes and the navel-like
stigma, Ilex denticulata is closely related to /lex latifolia Thunb., but the
latter has thick-coriaceous larger (8-17 cm. long, 4.5-7.5 cm. wide)
leaves and 3-9-flowered individual branches in the staminate pseudo-
panicles. The leaves of /lex denticulata resemble superficially those of
Ilex chieniana S. Y. Hu, but the latter has palmately striate and sulcate
pyrenes, thickly discoid or capitate stigmata, and ciliate calyx-lobes.

53. Hex graciliflora Champ. in Jour. Bot. Kew Gard. Miscel. 4: 328.
1852; Walp. Ann. 4: 429. 1857; Benth. Fl. Hongk. 65. 1861;
Maxim. in Mém. Acad. Sci. St. Pétersb. VII, 29(3): 43. 1881; Forbes
& Hemsl. in Jour. Linn. Soc. Bot. 1: 116. 1886; Loes. in Nov. Act.
Acad. Caes. Leop.-Carol. Nat. Cur. 78: 338 (Monog. Aquif. 1: 338).
1901; Dunn & Tutcher in Kew Bull. Misc. Inf. Add. Ser. 10: 60.
19%2.

An evergreen tree up to 6—9 m. high with rather stout branchlets, obo-
vate or oblong-elliptic leaves, minutely and loosely serrulate or subentire
leaves, fasciculate inflorescences, globose fruit with plane and thin-discoid
stigmata and 4 wrinkled and rugose, irregularly striate-sulcate pyrenes.

Branchlets subterete, castaneous or rarely brunneous when dry, third
year’s growth 3-4 mm. in diameter, longitudinally rimulose, the lenticels
lacking, the leaf-scars deltoid, not elevated; second year’s growth 2.5 mm.
in diameter, longitudinally plicate, rugose, glabrous; current year’s growth
1.8-2 mm. in diameter, angular, longitudinally ridged, minutely and
sparsely puberulous, glabrescent or glabrous, the terminal bud conic,
acute, pubescent. Leaves occurring even on the third year’s growth, 5-20

1949] HU, THE GENUS ILEX IN CHINA, II 379

mm. apart; stipules minute, often hidden; petioles 10-15 mm. long, one-
fifth to one-half the length of the lamina, glabrous or minutely puberulous
above, narrowly and deeply canaliculate, the distal half winged by the
decurrent leaf-base; lamina thickly coriaceous, olivaceous-brunneous when
dry, shiny above, less so beneath, obovate- or oblong-elliptic, 2—7.5 cm.
long, 1.5-3.5 cm. wide; obtuse or acute at the base; obtuse, rarely acute
or retuse at the apex; margin minutely and remotely serrulate or rarely
subentire; midrib impressed and glabrous above, elevated beneath, the
lateral nerves 5—7 pairs, obscure above, prominent beneath, loosely reticu-
late, the veinlets evident on the lower surface. Inflorescences fasciculate,
axillary, on last year’s growth, with abortive terminal buds, the bracts
ovate-deltoid, acute or tricuspidate, puberulous; flowers 4-merous. Stami-
nate flower: individual branches of the fascicles 3-flowered; peduncles
shorter than the pedicels, 1-2 mm. long, the pedicels 3-6 mm. long, both
sparsely puberulent or glabrescent, the prophylla 1 or 2, basal or sub-
basal; calyx patelliform, 2 mm. across, shallowly 4-lobed, the lobes deltoid,
obtuse or rounded, ciliate; corolla rotate, reflexed, 6 mm. across, the petals
oblong, 2.5 mm. long, one-eighth connate at the base, ciliate; stamens
shorter than the petals, the anthers oblong-ovoid; rudimentary ovary sub-
globose-ovoid, obtuse at the apex, inconspicuously 4-lobed. Pistillate
flower: individual branches of the fascicles uniflorous, the bracts small,
ovate, cuspidate or tricuspidate; pedicels 4 mm. long (after fruiting 6
mm. long) with 2 basal prophylla: calyx as in the staminate flowers;
corolla choripetalous, the petals oblong-obovate, 2.5 mm. long; staminodes
two-thirds as long as the petals, the sterile anthers sagittate; ovary globose-
ovoid, 1.5 mm. in diameter, the stigma discoid. Fruit globose, 5-6 mm.
in diameter, the persistent calyx explanate, 2 mm. in diameter, suborbicular
in outline, ciliate, the stigma plane, thin-discoid. Pyrenes 4, suborbicular
in outline, the ends rounded, 4 mm. long, 3 mm. wide, wrinkled and rugose,
irregularly striate-sulcate, the striae reticulate, the back shallowly im-
pressed, the endocarp woody.

CHINA: Hongkong: Champion (fragment of type, A); Wu-kau-
tin, IW. Y. Chun 5039 (A); Ford (staminate flowers, A, Ge scabiinberss
branches, A); Lamoné (B); Lugard Road, ¥. W. Taam 1603 (A); Wong-
neichong, Y. W. Taam 2013 (A): along Bride Valley, Y. Tsiang 49 (A);
Peak, Y. Tsiang 256 (A, NY); Mt. Victoria, W. J. futcher (ex Herb.
Hongkong no. 10213) (A); C. Wright GCN\YRU Seo bantao-lsland.: °C.
TEE TES (2 UTOS. CAS)

flex graciliflora is endemic to Hongkong and the near-by islands, where
it can be found growing in woods or in thickets as a tree 6—9 m. high. The
white flowers appear in April. The red fruits persist long on the tree, even
into February or later of the following year.

The glabrescent branchlets, the coriaceous leaves, the fasciculate inflo-
rescences, and the wrinkled and rugose pyrenes of Ilex graciliflora indicate
close relationship between this species and Jlex ficoidea Hemsl., which
differs in having leaves with caudate apices, 9 or 10 indistinct parallel
lateral nerves, median or submedian prophylla, and very short fruiting
pedicels (2-3 mm. long).

anny

380 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXX

54. [lex tsangil, sp. nov.

Arbor glaberrima; ramulis nigrescentibus; foliis subcoriaceis, ellipticis
vel ovato-ellipticis, 5-8 cm. longis, 2~3 cm. latis, crenulatis vel subintegris,
basi acutis vel cuneatis, apice acuminatis, eeuminibne 10-15 mm. longis,
costa supra impressa, subtus elevata, nervis lateralibus, utrinque 6-8,
supra obscuris, subtus evidentibus; inflorescentiis fructiferis fasciculatis,
axillaribus; pedicellis 10-12 mm. longis; fructibus depresso- -globosis, 5
mm. longis, 6 mm. diametro; calycibus persistentibus 1.5 mm. diametro,
stigmate minute discoideo; pyrenis 4, 4 mm. longis, 3 mm. latis, striatis,
ae et sulcatis, endocarpio coriaceo.

ntirely glabrous evergreen tree up to 8 m. high with nigrescent
pean elliptic leaves with minutely crenulate or subentire margin,
long-acuminate apices, fasciculate infructescences, slender fruiting pedi-
cels 10-12 mm. long, and rather small globose fruits.

Branchlets slender, subterete, glabrous; third year’s growth 2 mm. in
diameter, sparsely and minutely rimulose, the lenticels lacking, the leaf-
scars elevated; current year’s growth 1-1. 5 mm. in diameter, angular, the
terminal bud narrowly conic, acute, glabrous. Leaves occurring even on
the third year’s growth, 5-12 mm. apart; stipules very minute, often ob-
scure; petioles slender, 10-16 mm. long, one-sixth to one-fifth the length
of the Jamina, narrowly canaliculate above; lamina subcoriaceous, brun-
neous-olivaceous when dry, ovate-elliptic or elliptic, 5-8 cm. long, 2-3
cm. wide; acute or cuneate at the base; acuminate at the apex, the acumen
10-15 mm. long; margin remotely and inconspicuously crenulate or sub-
entire; midrib narrowly and shallowly impressed above, elevated beneath,
the lateral nerves 6-8 pairs, obscure above, evident beneath. Flowers not
seen. Infr fasciculate, axillary, on second year’s growth, fasci-
cles with 2—4 fruits, the fruiting pedicels slender, 10-12 mm. long with 2 mi-
nute ciliate basal prophylla. Fruit depressed globose, 5 mm. long, 6 mm. in
diameter, when dry black and rugose, the persistent calyx: minute, 1.5
mm. in diameter, the stigma minute, thin-discoid. Pyrenes 4, obovate in
outline, 4 mm. long, 3 mm. wide, one end slightly pointed, reticulately
striate, irregularly rugose and sulcate, the endocarp thick-coriaceous (de-
scribed from immature fruit).

CHINA: Kwangtung: Ta-pu, W. T. Tsang 21232 (TYPE A; ie Be

This species is endemic to northeastern Kwangtung, where it is a com-
mon tree in woods or thickets. The fruit has been reported as yellow in
July. The species is closely related to /lex graciliflora Champ., which
differs in having thick-coriaceous leaves with obtuse apices.

59: Oe chingiana Hu & Tang in Bull. Fan. Mem. Biol. Bot. 9: 252.

A ae evergreen tree up to 12 m. high with large (11-15 cm. long)
coriaceous, oblong-elliptic, remotely serrate and acuminate leaves, very
large (15 mm. in diameter) globose fruit, and 4 large wrinkled and pitted
pyrenes.

1949] HU, THE GENUS ILEX IN CHINA, II 381

Branchlets robust, straight, glabrous, castaneous when dry; third year’s
growth 3.5-4.5 mm. in diameter, transversely plicate-rugose when dry, the
lenticels.lacking, the leaf-scars semi-orbicular; second year’s growth longi-
tudinally plicate-rugose, subterete, 3-4 mm. in diameter; current year’s
growth angular, longitudinally striate, 2.5-3 mm. in diameter, the terminal
buds small, conical, the scales glabrous, ciliate, the axillary buds globose,
glabrous. Leaves occurring even on the third year’s growth, 5-20 mm.
apart; stipules obscure; petioles 10-15 mm. long, one-tenth to one-seventh
the length of the lamina, glabrous, rugose, narrowly canaliculate above;
lamina thick-coriaceous, brunneous-olivaceous when dry, glabrous, rather
opaque on both surfaces, oblong-elliptic or very rarely oblanceolate, 11-14
cm. long, 4-5 cm. wide; base obtuse, rarely rounded or cuneate; apex
acuminate, the acumen 7-12 mm. long; margin remotely serrate; midrib
impressed above, sparsely and minutely puberulent, very prominent and
elevated beneath, the lateral nerves 8-12 pairs, slightly elevated above,
evident beneath, the reticulation of the veinlets conspicuous above, less so
beneath. Infructescence a small fascicle, usually only one fruit maturing ;
bracts cartilaginous, glabrous, the apex acute, callose, the basal appendage
stipule-like, acute, ciliate; fruiting pedicels 2-4 mm. long, minutely and
sparsely puberulent when young, glabrescent later, with 2 sub-basal pro-
phylla; persistent calyx explanate, orbicular in outline, 3.5 mm. in diameter,
inconspicuously 4-lobed, the lobes rounded, sparsely puberulent and cili-
ate; fruits globose, 15 mm. in diameter, the stigma navel-like, orbicular in
outline when young, later appearing quadrangular, 4-5 mm. in diameter.
Pyrenes 4, obovate in outline, 10-12 mm. long, 6-8 mm. wide at the base,
reticulately striate, sulcate, wrinkled and pitted, the back flattened and
broadly impressed, the endocarp stony.

CHINA: Kweichow: San-hoa, Y. Tsiang 6396 (NY). Kwan g-
si: Lu-chen-hsien, Rk. C. Ching 6011 (isotype, A, US), Chuen-yuen, 7. S.
Lsoong (= Z. S. Chung) 82033 (A); Ping-nan-hsien, C. Wang 40344
(A), 40388 (A).

Ilex chingiana is endemic to Kwangsi Province, where it grows as a tree
in mixed forests. Specimens collected in June show very young fruits.
This may indicate that it flowers in May. Those collected in November
have large red fruits.

Because of the large leaves with remotely serrate margins, the large
fruit with navel-like stigmas, and the large reticulately striate, sulcate,
wrinkled and pitted pyrenes, this Kwangsi species is closely related to
the Hainan species Ilex nuculicava Hu. The latter species has puberulent
branchlets and petioles, pseudoracemose infructescences with smaller fruits,
and longer fruiting pedicels.

The infructescences of the material at hand appear solitary, but on
careful examination the paucifasciculate nature can easily be detected, for
the scars of the falling flowers and bracts are evident.

Only fruiting material has been available for this study. The size of
the fruits on these specimens varies considerably with seasonal variation.

382 JOURNAL OF THE ARNOLD ARBORETUM LVOL. Xxx

As the fruit matures, the mesocarp increases in volume and becomes fleshy.
As the fruits enlarge, the tissue of the exocarp next to the stigmatic lobes is
torn. When dry, this rent and discolored tissue gives the stigmata a star-
like outline. This character, the star-like stigma, is characteristic for this
species and a diagnostic character.

When the ripe fleshy fruit is pressed, the exocarp breaks and the fruit
flattens. On herbarium specimens such pressed fruits appear much larger
than their actual size. This is plainly the case with C. Wang 40344 and
40388.
55a. Ilex chingiana var. puberula, var. nov.

Arbor parva; ramulis glabris, castaneis, gemmis terminalibus pubes-
centibus; foliis oblahceolatis, 7-12 cm. iene. 3 cm. latis, serratis, basi
cuneatis, apice acuminatis, were 8-14 mm. longis, costa et petiole
puberulis; fructibus globosis, 11 mm. diametro, pedicellis pubescentibus,
3 mm. longis.

CHINA: Kwangsi: Ling-wun, S. K. Law 28662 (typr, A).

This variety grows in mixed forest as a tree 10 m. high with a trunk
30 cm. in diameter

The variety differs from the typical //ex chingiana in that it has narrower
predominantly oblanceolate leaves with cuneate bases and puberulent
terminal buds and petioles.

56. ee uraiensis Yamamoto in Jour. Soc. Trop. Agric. Taiwan 4: 486.
932.

on Hay

An evergreen tree up to 25 m. high, with minutely puberulous or
glabrescent branches, elliptic or obovate-elliptic leaves, remotely crenate
or serrate margins, abruptly and shortly acuminate apices, fasciculate inflo-
rescences, fruiting pedicels 6-8 mm. long, large globose fruit, and irregularly
striate and wrinkled stony pyrenes.

Branchlets subterete, cinereous or rarely nigrescent when dry; third
year’s growth 3 mm. in diameter, longitudinally rimulose, the lenticels
lacking, the leaf-scars crescent-shaped, slightly elevated; current year’s
growth 1-2 mm. in diameter, longitudinally ridged, minutely puberulous
or glabrescent, the terminal buds ovoid, pubescent. Leaves occurring also
on second year’s growth, rarely on the third year’s growth, 5-12 mm.
apart; stipules minute, deltoid, persistent; petioles 6-10 mm. long, one-
tenth to one-sixth the length of the lamina, the distal half winged by the
decurrent leaf-base, narrowly canaliculate above; lamina thinly coriaceous,
olivaceous when dry, shiny above, opaque beneath, elliptic or obovate-
elliptic, 3.5-10 cm. long, 1.2-3.5 cm. wide, cuneate at the base, shortly and
abruptly acuminate at the apex, the acumen 3-8 mm. long, the very tip
obtuse, the margin remotely crenate or serrate; midrib very slightly
impressed above, elevated beneath, the lateral nerves 8—9 pairs, prominent
on both surfaces, the reticulation of the veinlets prominent on both sur-

Ilex shades Ley sensu Yamamoto, Suppl. Ic. Pl. Form. 1: 32, fig. 13.
1925, n

1949] HU, THE GENUS ILEX IN CHINA, II 383

faces. Inflorescences fasciculate, axillary on the second or even the third
year’s growth, the individual branches of the fascicles uniflorous, the bracts
suborbicular, ciliate; flowers 4-merous. Staminate flowers: pedicels 3-5
mm. long, minutely puberulous, with 2 minute ciliate prophylla at the
base; calyx patelliform, 2 mm. in diameter, shallowly 4-lobed, the lobes
rounded, ciliate; corolla rotate, 7-8 mm. in diameter, the petals obovate-
oblong, 3-5 mm. long, 2.5 mm. wide, ciliate, slightly connate at the base;
stamens longer than the petals, the anthers oblong-ovoid, 1.25 mm. long;
rudimentary ovary subglobose, the apex rounded or slightly depressed.
Pistillate flowers not seen. Infructescences fasciculate, the bracts oblong-
ovate, puberulous, and ciliate, with 2 acute stipule-like appendages; fruit-
ing pedicels slender, 6-8 mm. long, pubescent, with 2 sub-basal, deltoid,
acute, puberulous, ciliate prophylla. Fruit globose, large, 10-11 mm. in
diameter, the persistent calyx suborbicular in outline, 2 mm. in diameter,
ciliate, the stigma navel-like, 2.5 mm. in diameter. Pyrenes 4, oblong in
outline, the ends obtuse, 7-8 mm. long, 4.5 mm. wide, irregularly striate
and wrinkled, the back flattened and depressed, the endocarp stony.

CHINA: Taiwan: R. Kanehira & S. Sasaki in 1916 (photo of TPE,
TU); Y. Shimada (TU); Sozan, E. H. Wilson 10288 (A, NY, US), 11213
(A, US); Sirin, K. Odashima (Taihoku Imp. Univ. Herb. no. 17751) (A,
NY, US); Kiirun (Taihoku Imp. Univ. Herb. no. 17797) (A, NY, US);
Herb. Univ. Imp. Taihoku no. 2872 (TU).

This species is endemic to Taiwan, where it grows as a common tree up
to 25 m. high with a trunk 0.5-1 m. in diameter. The white flowers appear
in April, and the mature red fruit persists on the tree during the winter
months.

The slender fruiting pedicels, large fruits, navel-like stigma, and the
shiny upper surface of the leaves with prominent lateral nerves of this
species indicate a close relationship with lex mertensii Maxim., which has
thickly coriaceous leaves with obtuse apices, and palmately striate esulcate
pyrenes with sublignified endocarps.
56a. Ilex uraiensis var. formosae, comb. nov.

Ilex mertensii Maxim. var. formosae Loes. in Nov. Act. Acad. Caes. Leop.-

Carol. Nat. Cur. 78: 338 (Monog. Aquifol. 1: 338). 1901.

Branchlets glabrous; leaves coriaceous, obovate, 2.5—4.5 cm. long, 1.5—2.5
cm. wide, cuneate at the base, rounded or rarely obtuse at the apex,
olivaceous when dry, shiny above, the margin remotely crenulate, the mid-
rib plane or slightly impressed above, elevated beneath, the lateral nerves
4—6 pairs, prominent on both surfaces, the reticulation of the veinlets
evident on both surfaces; infructescences fasciculate, uniflorous, the fruit-
ing pedicels 5-10 mm. long, glabrous, with 2 basal prophylla; fruit
globose, 10-12 mm. in diameter, the persistent calyx inconspicuous, 1.5
mm. in diameter, the stigma navel-like, 1.5 mm. in diameter; pyrenes 4,
oblong in outline, obtuse at the base, obtuse and retuse at the apex, 8 mm.
long, 5 mm. wide, irregularly striate and wrinkled, the back flattened and
depressed, the endocarp stony.

384 JOURNAL OF THE ARNOLD ARBORETUM [voL. XXX

CHINA: Taiwan: S. Cape, A. Henry 938 (A, US), 1002 (isosyNn-
TYPE of Ilex mertensit var. formosae) (A), 1251 (asosyNtTYPE of Ilex mer-
tens var. formosae) (A, NY).

The form of the leaf of this variety is like that found in /lex mertensi
Maxim., but the texture resembles that found in //ex uraiensis. Thus super-
ficially it may be considered as a variety of either of these two species.
The large stony pyrenes of this variety, however, are typically those of
Ilex uraiensis; hence it has been transferred to this species.

57. Ilex subficoidea, sp. nov.

Ilex cinerea sensu Groff in Lingnan Sci. Bull. 2: 63. 1930, non Champ.

Arbor; ramulis glabris; foliis coriaceis, supra nitidis, ovato- vel oblongo-
ellipticis, 7-10 cm. longis, 3 cm. latis, crenatis, basi obtusis, raro rotundatis,
apice acuminatis, acuminibus, 7-16 mm. longis, costa supra impressa,
subtus elevata, nervis lateralibus 10-11 paribus, supra obscuris, subtus
evidentibus; inflorescentiis fasciculatis, axillaribus, flores 4-meris; ¢ 3-
floribus, pedunculis 1 mm. longis, pedicellis 2 mm. longis; calycibus
4-lobatis, ciliatis; corolla rotata, 7 mm. diametro, petalis obovato-oblongis,
2.75 mm. longis, sparse ciliatis, staminibus quam petalis longioribus, ovario
aborto conico, apice obtuso; infructescentiis fasciculatis, pedicellis 10 mm.
longis; fructibus globosis, 10-12 mm. diametro, tuberculatis, stigmate
applanato-discoideo; pyrenis 4, 8-9 mm. longis, 5-7 mm. latis, irregulariter
rugosis et caveatibus.

An evergreen tree up to 15 m. high, with glabrous longitudinally striate
branchlets, ovate- or oblong-elliptic coriaceous and glossy leaves with an
abruptly acuminate apex, large (10 mm. in diameter) fasciculate fruits,
and large woody wrinkled pyrenes.

Branchlets straight, cinereous, brunneous, or even olivaceous; third
year’s growth 3—4 mm. in diameter, longitudinally striate and minutely
rimulose, the lenticels lacking, the leaf-scars deltoid, ovate, not elevated;
current year’s growth striate, angular, 1.5-2 mm. in diameter, glabrous,
the terminal bud thinly conical, glabrescent or glabrous, the scales densely
ciliate. Leaves occurring also on second year’s growth, 1-2 cm. apart;
stipules minute, broadly deltoid, callose; petioles 5-12 mm. long, one-
fourteenth to one-seventh as long as the lamina, éanaliculate above, the
distal half narrowly winged by the decurrent base; lamina coriaceous,
olivaceous when dry, ovate- or oblong-elliptic, 7-10 cm. long, 3 cm. wide;
obtuse at the base, seldom rounded, the apex abruptly acuminate, the
acumen 7-16 mm. long, the tip obtuse, shiny above, opaque beneath;
margin undulate, inconspicuously crenate, the teeth marked with black
spots; midrib impressed above, elevated beneath, the lateral nerves 10-11
pairs, obscure above, evident beneath, rather straight, curving upward
and reticulate near the margin. Inflorescence fasciculate, axillary, on sec-
ond year’s growth; flowers 4-merous. Staminate inflorescence: individual
branches of the fascicles 3-flowered, the bracts broadly ovate, mucronate,
ciliate, the basal appendages stipule-like; peduncles 1 mm. long, the
pedicels 2 mm. long, pubescent or glabrescent; calyx patelliform, 5 mm.

1949] HU, THE GENUS ILEX IN CHINA, II 385

across, the lobes glabrous, very sparsely ciliate; corolla rotate, 6 mm.
across, the petals obovate-oblong, 2.75 mm. long, the apical half very
sparsely and weakly ciliate, one-tenth connate at the base; stamens slightly
longer than the petals, the anthers ovoid, 0.75 mm. long; rudimentary
ovary obtusely conical, glabrous. Pistillate flower not seen. Infructes-
cences fasciculate, the pedicels 10 mm. long, with 2 basal or sub-basal
prophylla. Fruit globose, 10-12 mm. in diameter, tuberculate, the per-
sistent calyx 2.5-3 mm. across, 4-lobed, the lobes rounded, ciliate; stigma
thin-discoid, distinctly 4-lobed. Pyrenes 4, ovate-elliptic in outline, the
ends pointed, 8-9 mm. long, 5-7 mm. wide, irregularly wrinkled and pitted,
the endocarp stony.

CHINA: Kiangsi: Kien-nan, S. K. Lau 3979 (rype, fruits, A, US).
Kwangtung: Wung-yuen, S. K. Lau 2304 (A); Lung-t’au Mt., CCC
12736 (US). Kwangsi: S. Nan-ning, R. C. Ching 8220 (NY). Cre
nan: HY. Liang 637/72 (NY), C. Wang 34710 (A, NY), 36535 (A).

INDO-CHINA: Tonkin, W. T. Tsang 26911 (A

Ilex subficoidea is a large subtropical tree found in mixed forests south
of Lat. 25°N. The flowers are white and appear in January

The Hainan plants have long petioles and short leaves. ie these dif-
ferences fall within the range of variations displayed by the plant on the
mainland
58. [lex nuculicava, sp. nov.

Ilex comerea Champ. var., Merr. in Lingnan Sci. Jour. 5: 115. 1928:

Masamune, Fl. Kainant (Hainan) 137. 1940.

Arbor parva; ramulis pubescentibus vel glabrescentibus; foliis coriaceis,
oblongo-ellipticis vel raro oblanceolatis, 8-13.5 cm. longis, 2.2-4.5 cm
latis, subintegris, undulatis, vel crenulatis, basi obtusis, apice breviter
acuminatis, acuminibus 5-10 mm. longis, costa supra impressa, nervis
lateralibus 7-8 paribus, supra obscuris, subtus evidentibus; inflorescentiis
fasciculatis, axillaribus, pubescentibus; floribus 4-meris; ¢ 3-floribus,
pedunculis 1-2 mm. longis, pedicellis 3-4 mm. longis, calycibus ciliatis,
corolla rotata, 7-8 mm. diametro, petalis oblongo-obovatis, 3 mm. longis,
staminibus petalis aequantibus, ovario aborto subgloboso vel ovoideo,
apice obtuso; 9 unifloribus, pedicellis 3 mm. longis, corolla choripetala,
ovario ovoideo, 2.5 mm. longo, 2 mm. diametro, apice truncato, stigmate
discoideo; fructibus globosis, 10 mm. diametro, stigmate umbilicato; py-
renis 4, 6-7 mm. longis, 4—-4.5 mm. latis, dorso maltatinn Striatis et ene
ateralisire rugosis et caveatibus.

An evergreen tree up to 9 m. high with puberulent or glabrescent
branchlets, large oblong-elliptic acuminate leaves, fasciculate flowers, large
globose drupes and large bony wrinkled and pitted pyrenes.

Branchlets straight, cinereous, plicate and rugose when dry; third year’s
growth 5 mm. in diameter, longitudinally plicate, rimulose, the lenticels
lacking, the leaf-scars semi-orbicular, not elevated, closely associated with
the elevated scars of the inflorescences; second year’s growt mm. in
diameter; current year’s growth 1.75—2 mm. in diameter, longitudinally

386 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX
ridged, sparsely pubescent, the terminal buds conic, 3 mm. long, the
scales pubescent. Leaves occurring also on second year’s growth, 2-12
mm. apart; stipules minute, often obscure; petioles slender, 9-15 mm.
long, one-eleventh to one-seventh the length of the lamina, castaneous or
nigrescent when dry, canaliculate, sparsely puberulent above, the distal
half winged by the decurrent base, rugose beneath; lamina coriaceous,
olivaceous when dry, opaque above, paler and opaque beneath, oblong-
elliptic or rarely oblanceolate, 8-13.5 cm. long, 2.2—4.5 cm. wide; base
obtuse; apex short-acuminate, the acumen 5-10 mm. long; margin sub-
entire, undulate, minutely crenate, marked with black spots; midrib im-
pressed above, sparsely and minutely pubescent, prominent and glabrous
beneath, the lateral nerves 7 or 8 pairs, straight, curved upward and
loosely reticulate near the margin, obsolete above, evident or prominent
beneath, the reticulation of the veinlets obscure above, lacking beneath.
Inflorescences fasciculate, axillary, on second year’s growth, the scales of
the flowering buds persistent, reniform, acute, pubescent, 1 mm. long, 1.5
mm. wide; flowers 4-merous. Staminate inflorescence: individual branches
of the fascicles 3-flowered, the bracts broad-ovate, acute, pubescent, with
pubescent stipule-like basal appendages, the peduncles 1-2 mm. long,
densely pubescent, the pedicels 3-4 mm. long, pubescent, with 0-2 basal
or rarely submedian prophylla; calyx patelliform, sparsely pubescent, 2—2.5
mm. in diameter, 4-lobed, the lobes rounded, ciliate, 0.75 mm. long, 1.25
mm. wide; corolla rotate, 7-8 mm. in diameter, the petals oblong-obovate,
3 mm. long, 2 mm. wide, shortly and sparsely ciliate or eciliate, lightly
connate at the base; stamens equaling the petals in length, the anthers
ovoid, 1.25 mm. long; rudimentary ovary subglobose or ovoid, the apex
rounded or obtuse. Pistillate inflorescences: individual branches uniflorous,
the pedicels 3 mm. long with 2 basal prophylla, the calyx and corolla as in
the staminate flowers; staminode one-third the length of the petals, the
sterile anthers sagittate; ovary ovoid, 2.5 mm. long, 2 mm. in diameter,
the apical end truncate, the stigma discoid. Fruit globose, 10 mm. in
diameter, when dry cinereous and wrinkled, the pedicels 5-6 mm. long,
the persistent calyx explanate, 2.5 mm. in diameter, pubescent and ciliate,
the stigma navel-like, 3 mm. across. Pyrenes 4, broad-elliptic in outline,
tapering at the ends, 6-7 mm. long, 4—4.5 mm. wide, palmately striate and
sulcate on the back, wrinkled, pitted and rugose on the sides, the endocarp
woody.

CHINA: Hainan: Fan-yah, N. K. Chun & C. L. Tso, 44244 (TYPE,
A: NY, US, fruit); Po-ting, F. C. How 72490 (A); Man-ning, PF. C. How
73195 (A); Ch’ang-kiang, S. K. Law 1714 (A, NY); Bak-sa, S. K. Lau
26006 (A), 26550 (A), 26697 (A); Five Finger Mt., F. A. McClure (CCC
no. 9424) (A, NY), 9458 (A, US), 9534 (US); without precise locality,
C. Wang 34741 (NY).

This species is endemic to the island of Hainan, where it grows in woods
or forests at altitudes of 500-1800 m. The flowers appear in April or May.

A close relative is Jlex cinerea Champ. Many herbarium sheets have

1949 ] HU, THE GENUS ILEX IN CHINA, II 337

been incorrectly so named. Jlex nuculicava resembles Jlex cinerea in hav-
ing pubescent branchlets and terminal buds, and large oblanceolate or
oblong-elliptic opaque leaves, but the latter can readily be distinguished
by the characteristic short petiole (2-4 mm. long) and the twice tricho-
tomously branched staminate inflorescences.

58a. Ilex nuculicava var. brevipedicellata, var. nov.

Arbor parva; ramulis et gemmulis terminalibus pubescentibus; foliis
oblongo-ellipticis vel oblanceolatis, 7-17 cm. longis, basi cuneatis, apice
acuminatis; fructibus globosis, pedicellis 2 mm. longis; pyrenis 4, 6 mm.
longis, 5 mm. latis, dorso rugosis et caveatibus.

CHINA: Hainan: N.K. Chun & C. L. Tso 43959 (type, A; US).

This variety differs from the species in having shortly pedicellate fruits.
58b. Ilex nuculicava var. glabra, var. nov.

Arbor parva; ramulis et gemmulis terminalibus glabris; foliis lanceo-
latis vel oblanceolatis, 8-11 cm. longis, 2-3 cm. latis, basi cuneatis, apice
acuminatis; fructibus globosis, 9 mm. diametro, calycibus ciliatis; pyrenis
4, 6-8 mm. longis, 5 mm. latis, dorso rugosis et caveatibus.

CHINA Edina ni. Yai-chow, P.-C. How 71094 (Ive EAS NUS):

This variety differs from the species in having glabrous buds, branchlets,
and inflorescences.
58c. Ilex nuculicava var. auctumnalis, var. nov.

Arbor parva; gemmulis terminalibus pubescentibus, foliis coriaceis, supra
nitidis, oblanceolatis, 6-11 cm. longis, 2-3.5 cm. latis, basi cuneatis, apice
acuminatis; inflorescentiis fasciculatis.

CHINA: Hainan: Ling-shui, F. C. How 73800 (A); Po-ting, F.C.
How /3857 (A); Yai-chow, H. Y. Liang 62751 (type, A; NY).

This variety differs from the species in having smaller and narrower
oblanceolate leaves, with cuneate bases, and in its late flowering season.
In Hainan it flowers in the autumn, August-October. No fruiting speci-
mens were available for this study. Additional material may prove it to
be a distinct species.

(To be continued)

388 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

THE GENUS COCCOLOBA IN CUBA
RicuHarp A. Howarp

Tue New Wortp cEenus Coccoloba is found from Bermuda and
Florida on the North American continent, through the West Indies,
Mexico and Central America, into South America with many species in
the Amazon basin and a few species as far south as Argentina. In 1890,
Lindau published a comprehensive monograph of this woody genus (Engl.
Bot. Jahrb. 13: 106-229. 1890) which then included 125 species. Lindau’s
monograph is notoriously difficult to use and Britton, for his work on the
West Indian vegetation, presented a series of treatments of the genus
Coccoloba as it occurred on each of the islands he studied. Britton’s first
paper on the genus, in 1915, revised Coccoloba for Cuba. In following
years Britton described additional new species from Cuba, as did Urban,
and subsequent to their work Schmidt continued studies of the genus in
Cuba and the Caribbean area. Thus, at the present time approximately
50 species have been reported from Cuba alone, either in the literature or
in herbarium records.

The present paper, which considers the genus only in Cuba, is presented
to clarify certain points of the morphology of the genus, to describe several
new species from Cuba and Hispaniola, and to straighten out specific con-
cepts and bring the nomenclature up to date to aid the work of con-
temporary students of the Caribbean vegetation.

The author is grateful to Brother Leon of the Colegio de la Salle in
Havana, Cuba, for the introduction to this problem and also to Mr.
Sandwith of Kew for invaluable help and consultation on certain problems
of limitation of the species. Material for this work was supplied from
the following herbaria: Arnold Arboretum (A), Jardin Botanique de Etat,
Brussels, Belgium (BR), Botanical Museum, Copenhagen, Denmark (C),
Chicago Natural History Museum (formerly known as the Field Museum)
(FM), Gray Herbarium (G), Colegio de la Salle, Havana (HAB),
Sauvalle Herbarium of the Academy of Science in Havana (HABA),
Kew (K), Univ. of Michigan (MICH), Missouri Botanical Garden (MO),
New York Botanical Garden (NY), Herbarium of Juan Roig (R), Natural
History Museum, Stockholm, Sweden (S), Estacion Agronomica de Cuba
(SV), and the United States National Herbarium (US). I am grateful
to the directors and curators of these herbaria for the privilege of study-
ing their material on loan.

In this paper I am following, with qualifications which are mentioned,
the proposal of Little (Madrofa 7: 244-246. 1944) that the name Cocco-
loba should be conserved. Patrick Browne (Civ. Nat. Hist. Jamaica 209,
pl. 14, f. 3. 1756) originated the name Coccolobis (without generic de-
scription) for four species. Linnaeus (Syst. Nat. ed. 10, 1007, ore

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 389

1759) accepted Browne’s four species and applied binomials to them but
changed the spelling of the generic name to Coccoloba in publishing the
first generic description. Recently Sandwith (Jour. Bot. 78: 99. 1940)
has regarded Coccolobis P. Br. and Coccoloba L. as different names and
adopted the original Browne name and spelling. Moreover, Sandwith
states in his argument that Coccolobis is a diminutive name, thus a com-
pletely different word, and therefore transfers and new combinations are
necessary when one name or the other is adopted. Fortunately for the
present monographer Coccolobis is not a diminutive form, as Coccolobion
would be, and the new combinations are not necessary. Actually no
problem exists between these two words for Coccolobis of P. Browne is
invalidly published as a generic name under the present rules of botanical
nomenclature while Coccoloba L. is a valid publication with the first
generic description of the group. Considerable search and bibliographic
work would be necessary to determine who published the first valid
generic description under the name Coccolobis.

However the name for this genus still involves a problem. The generic
name Guaiabara Miller (Gard. Dict. ed. 4, vol. 2. 1754) is validly pub-
lished with a generic description for the pre-Linnaean species later listed
by Patrick Browne. Therefore Coccoloba L., the most widely accepted
name for the group, must be conserved to continue in use. Little’s pro-
posal for conservation has been approved by the committee on nomen-
clature of the American Society of Plant Taxonomists and the case will
be submitted for consideration and action at the International Botanical
Congress in 1950. I am accepting Linneaus’ name Coccoloba as only an
orthographic variant of Coccolobis, thereby eliminating the need for new
combinations of many of Britton’s new species from Cuba, a step which
would be necessary if one concluded that the two spellings of Coccoloba
actually represent two different names.

Lindau’s elaborate monograph of Coccoloba quite unfortunately intro-
duces several very significant errors of interpretation of the morphology of
the genus which has led subsequent workers astray and has built up con-
siderable confusion in this genus.

Lindau considered the flowers of Coccoloba to be perfect, and only rarely
unisexual by abortion. This is not the case, as careful study of herbarium
material will show, and this is easily verified in field studies of the plants.
The flowers are regularly unisexual and only occasionally are flowers of a
male inflorescence perfect. In the course of field studies made in the
Greater Antilles, Florida, and the Bahamas, within the past three years
I have seen many species of Coccoloba. In all specimens examined in the
field the plants were found to be dioecious and the staminate and pistillate
plants readily distinguishable. Furthermore, the natives in the areas
visited recognized the fact that some trees set fruit and other did not.
Recent information from Cuba reports that cultivated specimens of
C’. venosa L. have perfect flowers in some of the staminate inflorescences
so the plants may be described as dioecious, occasionally monoecious.

390 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

This condition has been recognized by a few workers but has not been
incorporated in floristic treatments.

he staminate flowers are usually clustered at each node of the inflores-
cence. Two to five flowers are commonly produced in the axil of a single
bract. In these flowers the stamens elongate and are exserted and the
pistil is rudimentary with short contorted styles. Pistillate plants by
contrast bear only one flower at each node on the inflorescence axis and in
these the stamens are rudimentary, never developing pollen, and are
included while the pistil is larger and exserted with longer divergent styles.
Lindau recognized these differences in form, i.e., stamens exserted or
included, flowers clustered or solitary, but failed to associate a unisexual
condition with this morphological pattern. Thus one finds in Lindau’s
key to the species of Coccoloba the basic separations of flowers clustered
versus flowers solitary, or stamens exserted versus stamens included. In
more than one case the male and female plants of the same species have
been described under different specific names.

The unisexual condition of the flower and the pattern of numerous
flowers per node in the staminate plants may be confused slightly in the
field by the fact that occasionally a male plant will form sterile fruits.
This was observed in several male plants of Coccoloba uvifera growing in
Loomis Park at Miami Beach, Florida. However, in this species the seed
was either imperfectly formed in all the fruits sectioned or the fruit was
hollow. Likewise these “male” fruits dehisced easily from the pedicels and
rarely approached the size of the mature fertile fruits from the pistillate
plants.

The habit of growth of the species of Coccoloba found in the West
Indies has also led to confusion in the interpretation of the species and the
formation of descriptions. Three growth habits can be observed in the field
and can also be recognized on herbarium material. In one growth form,
short shoots are produced. In these the branches are of determinate length
with very short internodes. The short shoots may be terminal or lateral.
The terminal short shoots frequently resume normal growth after a few
seasons as short shoots and so the congested area of short compact inter-
nodes may be in the middle of a shoot of normal appearance.

The second form is the normal one, with vigorous shoots of uniformly
elongated internodes. The third form is found in adventitious shoots
which are produced when the original shoot is cut or damaged. These
three forms are significant because of the different size and shape of the
leaves produced on each. Adventitious shoots with extremely long inter-
nodes will have much larger leaves on larger petioles and with longer ochreae
than those of the normal shoots. The leaves of short shoots are commonly
smaller in these characteristics than those of the normal shoot. Because of
the variation of leaf shape on a single plant, the shape of the leaf, particu-
larly the nature of the leaf apex is not a good taxonomic character and
where used must be considered with caution.

The large number of species described by Britton, Urban and Schmidt
which I have reduced to synonymy in this paper are commonly based on

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 391

the failure of these men to recognize the condition of variant growth. The
sterile Ekman specimens from Cuba which Schmidt referred to C. longi-
folia, a Jamaican species, are adventitious shoots of C. diversifolia. The
species described by Schmidt as C. lineari-lanceolata is material from an
adventitious shoot of C. pallida, and the species C. pilonis, C. acutissima
and C. woodfredensis described by Urban and Britton are adventitious
shoots or short shoots of C. reflexa of Lindau based on a Wright collection.

Once these peculiarities of Coccoloba are recognized the species appear
more clear-cut and it is hoped that the genus may be completely clarified
by the work which is now in progress.

Little described Coccoloba as “a large genus of about 130 species”.
Actually nearly 400 names have been published for this genus. The large
genus has been divided into sections by Wright (Griseb. Cat. Pl. Cub. 62.
1866), Meisner (DC Prod. 14: 150-171. 1856) and Lindau (l.c.) in their
work on the group. The section as defined by those authors in the
original form are not acceptable today in view of our present knowledge
of additional species from intermediate areas. The problem of sub-
dividing this large genus will be considered in a later publication. For
the present paper I'am not dividing the genus although three of the sections
recognized by Lindau are represented in the species which follow.

o attempt will be made in this paper to give other than the most
important bibliographic references. For a complete list of important as
well as casual references to any of the following species, Lindau’s mono-
graph of the genus Coccoloba (Engl. Bot. Jahrb. 13: 106-229. 1890) and
his later paper in Urban, Symbolae Antillanae (1: 215-233. 1899) present
most of the literature prior to 1900. Britton’s earlier treatment of the
genus Coccolobis in Cuba appeared in the Bulletin of the Torrey Botanical
Club, 42: 365-371. 1915
Coccoloba L. Syst. Nat. ed. 10, 1007, 1367. 1759.

Shrubs or trees; branches terete, often geniculate, or arranged in one
plane, short shoots commonly developed laterally or the terminal shoots of
limited growth becoming long shoots; nodes commonly tumid; ochreae
characteristically developed, membranaceous or coriaceous, deciduous or
persistent, glabrous, puberulent or pilose; leaves alternate, minute to
arge, membranaceous, chartaceous or coriaceous, the margin entire to
undulate, flat or revolute, the primary veins straight to the margin, much
branched at the apex becoming reticulate or arcuate and anastomosing or
arcuate and bifurcate-anastomosing, the secondary venation obscure or
coarsely to minutely reticulate, the upper leaf surface commonly pitted,
rarely pubescent, the lower leaf surface glabrous to pilose, short multi-
cellular glands present or the glands depressed in the lamina producing
resinous secretions; petioles terete to stout, broadly and shallowly canalicu-
late above, pilose to glabrous, the base ain tumid, attached at the base
of the ochreae or above the base to two-thirds the length of the ochreae;
plants dioecious or monoecious; inflorescence racemose of sub-spicate,
terminal on the primary or lateral branches, few flowered and short to

392 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

many flowered and several times the length of the leaves, the rhachis
glabrous, puberulent, pilose or with glandular excretion; flowers unisexual,
the staminate flowers in clusters of 2-5 at each node of the inflorescence,
occasionally solitary, the pistillate flowers solitary; bracts subtending each
node, the flowers developing in a membranaceous sheath which ruptures
irregularly becoming an ochreola, the ochreolae membranaceous, | per
flower, rarely stalked, the flowering pedicels shorter than the ochreolae to
many times as long, the flowers articulated at the apex of the pedicel; peri-
anth campanulate at the base forming a hypanthium, the lobes S, imbricate,
the outer three slightly larger than the inner, the stamens 7 or 8 borne
on the hypanthium, the filaments united at the base, the anthers introrse,
the stamens in the pistillate flowers rudimentary; pistil rudimentary in
the staminate flowers, trigonous in the pistillate flower, the styles 3, dilated
at the apices, the ovary 1 celled, the ovule solitary, attached basally:
perianth expanding in fruit, the lobes surrounding the achene or the
hypanthium expanding sinmoundine the achene with the perianth lobes
appressed against the apex of the achene or coronate on the achene, the
achene with a hard shiny outer layer, the inner layer papery; seed with
ruminated endosperm, the major lobes 3, the minor lobes and involutions
numerous, the embryo centrally located, the cotyledons orbicular, flat
rarely folded, the radicle small, terete.
Type species: Polygonum uvifera L.
KeEY TO THE SPECIES

A. Perianth lobes investing the achene; leaves usually membranaceous ;
ochreolar sheath commonly elongating with the flowers; bracts usually
1: 3G

Me ceri ne gk ee EE 4 dd VEDES REECE Cee ; . venosa.
AA. Hypanthium eee the achene, the perianth lobes appressed against
the apex of the achene or coronate on the achene; leaves coriaceous or

chartaceous fee ene eA in C. tenwifolia) ; ochreolae not elongat-
ing with the flowers but broken early; bracts usually straw colored or
brown
B. Inflorescence cai 1-5 flowered, only rarely equalling the leaves ;
leaves 1-1.5 cm
C, Leaves ec at the apex, cordate at the base, turning
black on drying ree eer eee pee OF microphylla.
CC. Leaves spinose-mucronate or a acuminate at the apex,
turning dark but not black on dr
D. Leaves cordate-ovate, the eee cordate, the apex
LOU SpINOSes geet ee new eek ee eeensen.
DD. Leaves ovate to suborbicular, the base rounded to sub-
cordate, the apex rigidly acuminate, not long spinose.
Cede ware Tae ee Os cea eA re OF peniclatD,
BB. Inflorescence several to many flowered, elongate, 2.5-30 cm. long,
shorter than the leaves to several times as long as the leaves.
Leaves spinose tipped or rigidly acuminate; leaf blades ovate
to oblong, glabrous; rhachis of the inflorescence puberulent ;
pedicels slightly surpassing or twice the length of the eciveolac
Mob id Ges 4a cera ahaa eae ee eee eee er C. reflexa

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 393

CC. Leaves acute, Se eee or emarginate, not spinose
or oy acute at the
ves ea Sees at least below
E. Leaves puberulent on bo ee urfaces; veins ; straight to
the margin not anastomosing....... 6. C. Clementis.
EE. Leaves glabrous above, Tees pubescent below.
F. Leaves orbicular, generally umbonate; veins

DD. Leaves

straight to the marcia not anastomosing; pedi-
cels shorter than a ochreolae; rhachis, bracts
and ochreolae DUGSE. A eae pak ee C. Acuna.

. Leaves ovate, opens or elliptic, flat not umbo-

nate; veins arcuate and anastomosing near the
margin
G. @ciieelie pedicellate, borne at the apex of
stout stalks; leaves white pubescent below,
the blade turning black on drying........
Gad acacia Urea eek hoy oe 8. C. caesia.
GG. Ochreolae sessile; leaves not turning black
on ing.
H. Mature flowering pedicels or those of
the fruit shorter than the ochreolae;
leaves pale aa below, the hairs
blunt at the apex.2...473: 9. C. pallida.
HH. Mature coe pedicels or those of
the fruit exceeding the ochreolae; hairs
attenuate at the apex.
I. Flowering and fruiting pedicels 7-15
mm. long; leaves thick coriaceous;
pedicels and flowers bright red..
Laan wey eee 10. . coriacea.
. Flowering and fruiting bes 1- :
mm, long; pedicels brown or tan
flowers oe white.

J. Leaves thin coriaceous, persist-
ently pubescent below on the en-
tire lamina; petioles arising from
the base of the ochreae; inflores-
cence 25-35 CMe lone tema ss
re te ll. C. retirensts.

JJ. Leaves membranaceous, barbate
in the axils of the veins, glabrate
on the lamina; petioles arising
from above the base of the och-
reae; inflorescence 8-18 cm.

eae ee ee . C. tenutfolia.
glabrous below when mature (rarely puberulent

Laem
e

ae young
. Leaves orbiewlar to reniform, rounded or emarginate
at the a

ie

Fruiting pedicels shorter than the ochreolae; peri-
anth lobes coronate in the fruit, equaling the
fruit in length; fruit pilose. .13. C. baracoensis.

394,

JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

FF, Fruiting pedicels exceeding the ochreolae; peri-
anth lobes shorter than the fruit, coronate or ap-
pressed; fruit glabrous.

G. Rhachis of inflorescence keeled below the
nodes; leaves thick coriaceous, the secondary
venation obscure; fruit obione: the perianth

14. C

lobes not coronate......... “, nipensis.
GG. Khachis of inflorescence terete or ang Ted. not
keelec

H. Leaves broader than long, thick coria-
ceous, not turning black on drying ; fruit

HH. Leaves thin coriaceous or chart iceOUus }
fruit ovoid, perianth lobes more or less
coronate.

I. Leaves coarsely reticulate veined on
both surfaces, turning black on dry-

m
ar
an

Rhachis of inflorescence puberu-
lent; primary veins of the leaves
free to the margin.
eee re tee. te MN oviinopine.
. Rhachis of inflorescence gla-
brous; primary veins
leaves arcuate, anastomosing
near the margin. . retusa,
II. Leaves smooth above, finely reticulate

4
Ss

inflorescence puberulent; veins anas-
ene Neer Ne are hc a has
eer Te ee eT rere . C. praecox.
EE. Leaves ovate, oblong, elliptic oh not orbicular.
t os puberulent to pilos
ruiting pedicels eae than the ochreolae
or only slightly exceeding them
H. Inflorescence stout, 10-30 ecm. long;
leaves thick coriaceous, generally more

I. Veins of the leaves straight to the
margins; perianth red, puberulent. .
ey ee ee eee 19. C. Shaferi.

II. Primary veins arcuate, anastomosing
near the margins; perianth white,
PlOPPOUS. ckaricas kale 20. C. costata

HH. Inflorescence slender, 2-6 cm. long;
leaves thin coriaceous, generally less
than 7 cm. long.

I. Fruit obovoid, the lobes appressed
against the apex of the achene; sec-
ondary venation reticulate on both
surfaces, leaves commonly with

LN

1949]

GG. Fruiting pedicels
ochreolae

He.

HOWARD, THE GENUS COCCOLOBA IN CUBA

395

glandular excretions resembling Sy
tate scales on the BES ee

KH
—

pane is.

Fruit ovoid, the oe ae more
or le
reticulate

ss coronate;
mee

leaves smooth,
ee

not
noticeable

glandular excre

JJ. Leaves smooth

Ct
a

aves ae on both surfaces,
the lamina thin coriaceous,
rigid, the apex acute io obtuse
j not Peo ee
pilose; pedicels of the

shorter than a ochreolae
sp Linkage Daun OMe CALPE ESECIG:
above, minutely
reticulate below, the lamina usu-
ally rigidly coriaceous, the apex
acute or acuminate, commonly

short spinose mucronate; pe eti-
oles puberulent; fruiting pedi-
cels equal to or rarely half again
as long as the aes

reflexa.
2-10 times as a as the

Leaves broader than long, thick coria-

eous, without conspicuous reticulate
Werla HlOts yi-adhe eee uvifera
eaves longer than broad : secondary

venation reticulate, at least bel

Mature leaves 3-5 cm. long, eblone:
ovate to orbicular.

J.

JJ.

Primary veins straight and f

the margins, the leaf b oe

usually turning black on drying
shining above, commonly a ee
late on both surfaces..........
5 -, Northropiac.
Primary veins arcuate, dichot-
omous and anastomosing near
the margin, the leaf blades not
turning black. on drying, smooth
above, minutely reticulate below.
K rounded or
perianth age
appressed

=

eaf apex
emarginate ;
invbricate,
against the apex oe the
dehenen.@ sl aarst. proecor
KK. Leaf apex acute to acumi-
nate, often spinose mu-
cronate; perianth lobes

JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

more or less ee
pais bites B.- “G eae
Ii, a leaves 8-25 cm. long.
aves membranaceous, com-
i barbate in the axils when
mature; petioles attached above
the base of the ochreae........
er ee 12. C. tenuifolia.
JJ. Leaves coriaceous, glabrous;
nae a at the base of
the oc
K: Fruiting pedicels 8-15 mm.
{Kove t a raeeeer 10. C. coriacea.
KK. Fruiting pedicels 3-5 mm.
lon

L. Leaves conspicuously
coarsely reticulate on
both surfaces
blade with glandular
excretions below.....
ree 25. C. Wrghtiu.

LL. Leaves smooth or mi-
nutely reticulate on
both surfaces; glands
usually sunken, excre-
tions not evident.

M.Leaf blades
SMIODE: x05 ace He
e0. GeO sore.

MM. bens blades mi-

FF. Rhachis glabrous or with glandular excretions.
G. Rhachis and flowers with glandular excre-
tions appearing as scales or peltate protuber-
ances.
H. Perianth lobes coronate in fruit, the fruit

7-9 mm. long; leaf blades with small

glandular excretions below............

ee ee ee rer 23. C. Mevighia.
HH. Perianth lobes appressed to the apex of
the achene..

I. Leaf blades ovate, cordate or rounded
and unequal at the base, apex rounded
or obtuse; glands sunken in pits or
depressed; truit fusiform, 10 mm.
long or longer....24. C. Leonardit.

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 397

II. Leaf blades oblong, narrowed at the

base, a

glan dulee

e to acuminate at the apex;

excretions appearing as

peltate scales; fruit ovoid to globose,
mm. fone eee ee C. benitensis.

GG. Rhachis and flowers glabro
H. Fruiting pedicels ee than the och-

reolae.

I. Primary veins straight to the margin;
leaf blades thick coriaceous; petioles
19.

glabrous.

II. Primary veins arcuate, anastomosing
near the margin; leaf blades thin
coriaceous.

J. Fruit ovoid, the perianth lobes
coronate; petioles pilose pubes-
c

C. Swartsit {. pubescens.
fusiform, the perianth

lobes appressed and imbricate
against the apex of the achene;
PeHOles 2 1ADLOUS wee eee 2

UA ahaa 24. .C. Leonardii.

HH. Fruiting pedicels longer than the

ochreolae

I. Leaves oe ela auriculate at

the base

flowers red on pedicels

5 mm. loner ates pedicels 10 mm.

eee ee

Re eva ened 26. C. Cowellit.

II. Leaves flat; flowers green or white.

Leaf

hades cordate at the base,

minutely reticulate on both sur-

faces ;

mm.

flowering pedicels 7-10
long; fruiting pedicels to

15 mm. long...10. C. cortacea.

JJ. Leaf blades rounded or nar-
rowed at the base, not cordate.

K. Leaves shining above when

<. Leaves dull, dark a not
lack dr

dry, usually turning black
on drying; fruit ovoid, the
perianth lobes coronate, the
PriittO.O mm 1One 24 aes
tine eas eae ie [Guventsa:

ruit
globose to obpyriform, the
perianth lobes appressed
against the apex of the
achene, the mature fruit
LOR YS minis Non eases het
ene 27. C. diversifolia,

398 JOURNAL OF THE ARNOLD ARBORETUM [VOL Rex

1. Coccoloba venosa L. Syst. ae ed. 10, 1007. 1759; Fawcett and
Rendle, Jour. Bot. 51: 123.
Coccoloba punctata L. Sp. Vl. ed. 7 : 23. 1702.
Uvifera arbor americana, fructu aromatico punctatus, Pluk. Alm. 394. t.
237, fig. 4. 1696, as to leaf only.
Coccoloba nivea Jacq. Hist. Stirp. Amer. 115, pl. 78. 1763; Enum. PI.
9, 1762.
Guaiabara venosa (L.) House, Am. Midl. Nat. 8: 64. 1922, (as Guaibara).
Trees to 10 m. tall; branches terete, glabrous, the nodes not tumid;
ochreae membranaceous, deeply cleft, acuminate on one side, or truncate,
to 2 cm. long, glabrous or with flattened glands; leaf blades oblong-
lanceolate to elliptic, the apex short acuminate, the base narrowed and
slightly cordate or cuneate or obtuse, 8 & 4, 10 & 4.5, 16.5 & 6.5, 21 x
9, 27 X 10.5 cm. long and broad, membranaceous, glabrous except for
clusters of hairs in the axils of the veins, sparsely glandular below; midrib
and primary veins slightly prominent on both surfaces, the primary veins
8-13 pairs, straight or arcuate, bifurcate and anastomosing at the margins,
the petioles 5-10 mm. long, slabrous: leaves of the adventitious shoots
about the same size, the internodes much elongate and the ochreae to
4 cm. long; inflorescence terminal or terminal on short lateral branches,
the rhachis puberulent, angular; staminate flowers in clusters of 2—5, the
pistillate flowers solitary, the bracts lanceolate-ovate, to 1.5 mm. long,
black, puberulent to pilose or commonly with a fringe of hairs at the apex;
ochreolae membranaceous, enlarging with the expanding bud, each flower
with an ochreola, to 2 mm. long, the flowering pedicels 1-2 mm. long,
glabrous; hypanthium less than 0.5 mm. long, the perianth lobes broadly
ovate, 1.5-2 mm. long and broad, slightly unequal, the fertile stamens to
1 mm. long; fruiting pedicels 1.5-2.5 mm. long, the fruit broadly ovoid,
the perianth lobes fleshy, white or pink enclosing the black achene, the
hypanthium scarcely evident in the fruit, the fruit 3-4 mm. long and broad.
DisTrIBUTION: Cuba (introduced), Hispaniola, Puerto Rico, Jamaica,
Virgin Islands, Lesser Antilles, Trinidad.
CUBA: Las Villas: Atkins Garden, Soledad, P. I. 98831 collected
by Walsingham without number (G). In flower and fruit Nov. and March.
The history and synonymy of this species is quite confused. Fawcett
and Rendle did considerable work in publishing their note on this species
for the Flora of Jamaica. While C. nivea Jacq. has a clear-cut description
and is well illustrated and might well be selected as the type, I am follow-
ing Fawcett and Rendle’s interpretation since it is now widely accepted
in the Caribbean area. Lindau was completely confused in his treatments
of C. punctata, C. excoriata and C. venosa and his work is difficult to
untangle.
The specimen from Cuba is an introduction now under cultivation at
the Atkins Garden at Soledad. The records of the garden list this as a
plant introduced by the United States Department of Agriculture with

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 399

the number 98831 from Great Inagua in the Bahamas. There are no
specimens of this species from the Bahamas in the collections I have studied
and either the place of origin of this cultivated material is confused or
the specimen might be growing in the Bahamas, also under cultivation,
without herbarium records available.

According to correspondence with H. G. Walsingham of the Atkins
Garden there is but one specimen in the Garden. That tree flowers and
fruits regularly. From field studies made of other species I have con-
cluded that the genus Coccoloba consists of dioecious plants; however, this
specimen from Cuba must be monoecious. The majority of the ten
herbarium specimens I have for study from the Soledad plant are with
staminate inflorescences and flowers. On four specimens the flowers are
perfect and have developed fruit. No truly pistillate flowers were seen
in this collection.

The fruits of C. venosa are extremely attractive with the white or pink
fleshy perianth lobes surrounding a jet black achene. Jamaica has been
assumed the type locality of this species although, as Fawcett and Rendle
point out, no collections are known from that island.

2. Coccoloba microphylla Griseb. Cat. Pl. Cub. 62. 1866.

Uvifera microphylla O. Ktze. Rev. Gen. 2: 561. 1891.

Virgate shrub 1-2 m. tall to small tree 5 m. tall, branches often
arranged in one plane, commonly with brown shredding bark, the youngest
branches puberulent, nodes not at all tumid; short shoots well developed,
often spine-like at the tips; ochreae subcoriaceous, 1 mm. long, puberulent,
commonly dehiscent from the side opposite the petiole and remaining
attached to the petiole, appearing as auriculate wings; leaf blades orbicular
to reniform, apex emarginate, base slightly cordate, 1 * 1, 5 & 5 mm.
long and broad, coriaceous, turning black on drying, vention reticulate,
conspicuous on both surfaces; inflorescence terminal on short shoots,
rhachis very short, flowers appearing subsessile; bracts ovate, less than
0.5 mm. long, ochreolae membranaceous, less than 0.5 mm. long, flower-
ing pedicels 0.5 mm. long, hypanthium less than 0.5 mm. long, perianth
lobes 1 mm. long and broad, fertile stamens 1 mm. long; fruiting pedicels
less than 1 mm. long, fruit not known.

DISTRIBUTION: Cuba.

CUBA: Oriente: Sierra de Nipe, - Rio Piloto, Ekman 15069 (S);
Gamboa, Ekman 14996 (A, FM, G, NY, US). Matanzas: Pan de
Matanzas, SE of Canasi, Ekman 16513 i E. of Canasi, Cuabal del Espinal
ak 12975 (NY). “bas Villas? Cieneruita, Combs 725 (2M; G, MO;

US); San Marcos, Wright 2249 (type collection, G, MO, NY, S),
ee 9187 (NY); Motembo, Leon, Edmund & Fortin 8554 (NY), Leon
& Laustalot 11368 (NY). Pinar del Rio: Coloma, Britton, ae
Gager 7023 CNB): Roig & Fors 4214 (NY). Isle of Pines: Sigu
Britton, Britton & Wilson 15376 (NY, US), Ekman 12183 (US). ened
in owen: Feb., Aug., Sept. Local name: uwverillo.

400 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXX

3. Coccoloba armata Griseb. Cat. Pl]. Cuba 62, 283. 1866; Sauvalle, FI.
Cub. 139. 1868; Britton, Bull. Torrey Bot. Club 42: 366. 1915.

Uvifera armata O. Ktze, Rev. Gen. 2: 561, 1891,

Much branched shrub 1-3 m. tall becoming a tree 10 m. tall; branches
commonly in one plane, puberulent, nodes slightly tumid; ochreae 1 mm.
long, commonly breaking free on the side away from the petiole; leaf
blades cordate, apex rigid, long spinose mucronate, base cordate, 5 xX
12 * 9 mm. long and broad, coriaceous, glabrous, midrib impressed above,
primary veins inconspicuous, upper surface smooth, lower surface minutely
reticulate, margin slightly erose or entire, flat or recurved; petioles 1 mm.
long, base tumid or corky, leaves of adventitious shoot 18 K 18 mm.
long and broad; flowers in short terminal inflorescence, 2 or 3 flowers
clustered together, rhachis very short or none, bracts ovate, less than 0.5
mm. long, ochreolae membranaceous, 0.5 mm. long, flowering pedicels less
than 0.5 mm. long, perianth yellow-white, hypanthium less than 0.5 mm.
long, perianth lobes 1 mm. long and wide, fertile stamens 1 mm. long;
fruiting pedicels less than 1 mm. long; fruit ovoid, narrowed at both ends,
to 5 mm. long, 4 mm. in diameter, red, mature perianth lobes appressed
against the apex of the achene, achene dark brown or black.

DISTRIBUTION: Cuba.

CUBA: Oriente: Guantanamo Bay, Britton 1974 (NY); La Car-
bonera, Guantanamo, Ekman 2878 (FM, S, US), Ekman 10170 (A, S);
Sierra de Nipe, near Rio Piloto, Ekman 9993 (S); Bay of Sagua, Taname,
Linden 1857 (BR). Las Villas: Motembo, Leon, Edmund & Fortin
8556 (NY), Leon & Roca 8205 (NY); Mordazo, Leon & Cazafias 5975
(NY); Calicita, Loma de Ciego, Combs 531 (FM, G, MO, NY, US); Rio
San Juan, Britton, gies = Wilson 5890 (NY); Manacas, Howard 5503
(G). ava ga, Leon & Cesdtre 8956 (NY); Loma Coca
near Campo Florida, Biman 13217 (S); Mifas, ra 13140 (NY, S).
Pinar del Rio: Sierra de Vinales, Ekman 16587 (S); Cabafias Bay,
Ekman 10930 (NY, S); cn Marcos, Wright 2250 (G, MO, NY, S, US,
type Collection). Collected in flower: May, Aug., Nov. Collected in fruit:
Jan., Aug., Sept., Nov., Dec.

The original description of this species in Cat. Pl. Cuba lists Wright
1250 as the specimen upon which the species is based. This is corrected
in the emended list of Wright collections (l.c. 288) to read Wright 2250.

Field studies on this species are much needed. There are two aspects to
the leaves of herbarium specimens. In some the leaves are smooth, in
others minutely reticulate on both surfaces. In addition the plants of
these two divisions have a slightly different appearance. At present there
is no satisfactory way of handling these two divisions and all are cited as
one species.

The majority of the specimens cited above are sterile. The flowers are
small and the almost complete absence of an inflorescence axis adds to the
difficulty of selecting flowering or fruiting material in the field.

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 401

4. Coccoloba geniculata Lindau, Engl. Bot. Jahrb. 13: 141. 1890.

Uvifera geniculata O. Ktze. Rev. Gen, 2: 561. 1891.

Virgate shrub, 1-2 m. tall; branches commonly short, contorted or
geniculate, the youngest puberulent, nodes slightly tumid, ochreae cylin-
drical, membranaceous, oblique at the apex, 2-3 mm. long, leaf blades
ovate to suborbicular, often umbonate, apex acute, commonly with a spiny
mucro, base rounded, 1.5 & 1, 2.5 & 1.5, 3 X 2.5 cm. long and broad,
coriaceous, glabrous, smooth often shining above, dull beneath, midrib
impressed above, prominent below, primary veins 9-11 pairs, inconspicu-
ous, secondary venation obscure, leaf surface glandular dotted above and
below, margin entire, recurved; petioles 2-4 mm. long, puberulent or
papillose, tumid at the base; leaves of adventitious shoots ovate, 4 < 2
cm. on petioles 5 mm. long; inflorescence terminal, often clustered, 0.5—
1.5 cm. long, rhachis papillose rarely minutely puberulent, staminate
flowers in clusters of 3-8, female flowers solitary, female inflorescence
short, generally consisting of 1-3 flowering nodes; bracts membranaceous,
1.5-2.5 mm. long, 2-3 mm. broad, glabrous; ochreolae membranaceous,
several concentrically arranged, flowering pedicels to 1 mm. long, perianth
pink, hypanthium 0.5 mm. long, perianth lobes 1.5 mm. long and broad,
filaments of fertile stamens less than 1 mm. long; fruit from a staminate
plant ovoid, 5 mm. long, 2.5 mm. diameter, glabrous, perianth lobes
appressed against the apex of the achene.

DISTRIBUTION: Cuba.

CUBA: Oriente: Sabana de Yaba, Yareyal, Holguin, Leon 15783
(NY). Camaguey: Santayana, Ekman 15339 (A, ONG oN Ye oS):
Caobilla, Acufia 8876 (R. #4029, S. V. #8876, Herb. C. E. Baker #13700).
Las Villas: Santa Clara city, Ekman 14042 (S), 16331 (S), 18842
(NY, S), Leon 14950 (NY), Britton & Cowell 10199 (NY), 13325 (NY),
Britton, Britton & Wilson 6189 (NY); Puerto Principe, Wright 2255 (G,
NY, MO, HABA, type Collection). Collected in flower: Feb., Mar., June,
July, Oct. Collected in fruit: July.

Unfortunately the specimen (Wright 2255) on which Lindau based
this species consists of adventitious shoots or fast growing shoots with
larger leaves, and an abnormal inflorescence. Lindau’s use of the specific
epithet geniculata refers to the angular path of this abnormal inflorescence.
It could, however, apply as well to the commonly geniculate nature of
the branch system of the more typical plants now assigned to the species.
The specimens collected by Britton & Cowell 13325 and Britton, Britton
& Wilson 6189 show the relationship between the normal short shoot
geniculate type of growth and the occasional adventitious shoot with the
larger leaves.

In the plants with more normal growth habit than the type collection
the inflorescence is short, rarely exceeding 1 cm. in length. Three to eight
staminate flowers are clustered in concentric ochreolae as in the Wright
specimen but the flowering axis is much shorter. In the female plants
I have examined the flowering axis is likewise very short and may consist

402 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

of only three nodes with a total of three flowers, one per node of the rhachis.

The fruit mentioned in the description above was one of two fruits
attached to the specimen collected by HAman 15339. This specimen has
staminate flowers on the plant but as occasionally happens the ovary
may develop and apparently a mature fruit is produced. In this case,
as in many cases I examined in the field, the fruit was hollow, with no
evidence of an embryo and endosperm being formed. The shape and
size of this fruit may be regarded as indication of that of fertile fruits.
This conclusion is based on evidence obtained from the study of other
fruits of similar abortive nature from other species studied in the field.
5. Coccoloba reflexa Lindau, Engl. Bot. Jahrb. 13: 141. 1890.

Coccoloba pilonts Urb., Fedde Re ee Sp. Nov. 13: 445, 1914; Schmidt,

Fedde Repert. Sp. Now, 32% 79.1933.
Coccolebis woodfredensis Britton, i. Torrey Bot. Club a. ne 1915.
sea nals dna Urb., Fedde Peper. Sp. Novi 18s 1:

Uvifera reflexa O. Ktze. Rey. Gen. 2: 562. 1891

Shrub 1—2 m. tall; branches terete often wihodinis puberulent when
young, nodes slightly tumid; ochreae membranaceous, puberulent 2—5 mm.
long; leaf blades ovate to ovate-elliptic or oblong, rarely lanceolate, apex
abruptly acuminate, often spinose mucronate, base rounded to slightly
cordate, 4 & 2, 5.5 & 2.5,7 * 4,8 & 3 cm. long and wide, coriaceous,
smooth and shining above, dull beneath, glabrous except for the puberulent
midrib, primary veins 5—7 pairs straight to arcuate anastomosing, incon-
spicuous, secondary venation reticulate below, blade punctate glandular
dotted on both surfaces, margin entire, commonly recurved; petiole 3-6
mm. long, puberulent; inflorescence terminal, 2.5-8 cm. long, rhachis
puberulent, staminate flowers in clusters of 2—5, pistillate flowers solitary,
bracts ovate, 0.5 mm. long, ochreolae membranaceous, 0.5 mm. long,
flowering pedicels 0.5-1 mm. long, hypanthium 0.5 mm. long, perianth
lobes 1-1.3 mm. long and broad, fertile stamens 1 mm. long; fruiting
pedicels 0.8-1.2 mm. long, fruit black, ovoid, to 6 mm. long, 3 mm.
diameter, perianth lobes appressed against the apex of the achene.

DISTRIBUTION: Cuba.

CUBA: Oriente: me Pilon near wie Shafer 1232 (type
collection of C. pilonis, NY, US), Ekman 3289 (S); Cerro de Fraile, Hol-
guin, Ekman 7582 (S), fee 15716 (S); Sabana de la Yaba, Yareyal,
a 15702 (NY); Sierra de Nipe, Rio Pilot, Ekman 2718 (S), 19149
(a) oie (ayer (type collection of C. acutissima, S); Loma Mensura,
Sierra de Nipe, Ekman 9733 (S); Cabeza de Nipe, Sierra de Nipe, Ekman
2479 (S); Bio-Trail, co de Nipe, Ekman 4787 (S); ridge between Rio
Bayate and Rio Piedra, Sierra de Nipe, Ekman 15035 (S); Woodfred to
Piedra Gorda, Sierra de Nipe, Ekman 15255 (S), Shafer 3180 (type collec-
tion of C. wocdfredensis, NY); Punta Padre, Curbelo 109 (NY); Playa
Vaca, Moa, Acufia 12408 (SV). Without locality: Wright 2256
(type collection of C. reflexa, G, MO, HABA). Callecicd in flower: July,
Aug., Sept. Collected in fruit: July, Aug., Sept., Oct., Nov., ;

1949 | HOWARD, THE GENUS COCCOLOBA IN CUBA 403

Schmidt pointed out in his paper in Fedde, Rep. 32: 79. 1933, that the
present collection of specimens represents the extremes of a species with
variable leaf shapes. In doing so Schmidt reduced C. woodfredensis and
C. acutissima to the synonymy of C. pilonis. Unfortunately Schmidt did
not look far enough, for C. reflexa of Lindau based on a Wright collection
(2256) represents the same species and being an older name must be
used for this species.

Again the variation in the leaves can be associated with the growth habit
of the plant, the larger and narrowed leaves of the type of C. acutissima
are vigorous shoots while the smaller leaves represent a more normal
growth pattern. All types of leaves can be found on one or a few sheets
of the same collection. The large number of collections from the vicinity
of Holguin and from Sierra de Nipe indicate the range of variation to be
expected in one species in one area.

As Schmidt correctly points out the amount of puberulence on the
rhachis of the inflorescence varies considerably from specimen to specimen.
Also the characters Britton used to separate C. woodfredensis and C.
pilonis in his revision of the genus Coccoloba (Bull. Torrey Bot. Club
42: 365. 1915) are entirely variable and can not be used
6. Coccoloba Clementis, sp. nov.

Frutex ramulis teretibus, nigrescentibus exsiccis, puberulis, ochreis
brunneis, membranaceis, puberulis, petiolos 6-8 mm. excedentibus, uni-
lateraliter perincisis, apice dilatatis; lamina matura obovata vel elliptica,
apice rotundata, basi acuta vel rotundata, chartacea, supra et subtus
persistente puberula et glandulosa, 5 & 3.5 vel 6 & 4 cm. longa et lata,
margine undulato maturitate recurvato, nervo medio supra subconspicuo,
nervis primariis supra inconspicuis, nervo medio et venis subtus promi-
nentibus, nervis primariis 3-5, ad marginem rectis, non anastomosantibus,
venis secondariis obscuris; penn 3-5 mm. longis, supra subcanaliculatis,
puberulis; inflorescentia terminalis 5-12 cm. longa, rhachi dense puberulo,
bracteis late ovatis plus minusve 0.5 mm. longis, 1.5—-2 mm. latis, puberulis,
brunneis, membranaceis, ochreolis circa 0.5 mm. longis, puberulis, pedicello
puberulo ochreolas non excedente; floribus masculinis 2 vel 3 per nodulum
puberulis, hypanthio 1 mm. longo, lobis 5, imbricatis, exterioribus ee aa |
mm. longis et latis, ieenioribus sub- regione staminibus 7 vel 8,
filamentis plerumque 1 mm. non aequantibus nee circa 0.5
longis; pistillo rudimentario ad 1 mm. longo, stylis 3; flos foemina et
fructus ignotus.

CUBA: Oriente: Rio Cromita, Cayoguain, Punta Gorda, Clemente
4097 (vyPE, G; Isotype, HAB)

The specimens cited above were collected in flower on July 25th. The
species is distinct from others from the West Indies in the undulate leaf
margin and the persistent puberulence on both leaf surfaces.

7. Coccoloba Acuna, sp. nov.

Frutex ramulis teretibus, pilosis, nodis tumescentibus; ochreis mem-

branaceis, pilosis, 2-4 mm. longis, ad basin cylindricis, ad apicem sursum

404. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

dilatatem et obliquis; lamina orbiculare, bullata, apice rotundata, basi
rotundata, 1.5 & 1.5, 2 X 2 cm. longa et lata, coriacea, supra glabra,
nitida, subtus pilosa, nervo medio et venis primariis supra impressis vel
obscuris, nervis primariis 4 vel 5, ad marginem rectis, bifurcatis, non
anastomosantibus, venis secondariis subtus dense reticulatis; petiolis
teretibus, 3-4 mm. longis, pilosis; inflorescentia terminalis, 2—4 cm. longa,
rhachi riloséi bracteis ovatis, 1.5 mm. longis, 1.5-2 mm. latis, pilosis;
ochreolis membranaceis, 1 mm. longis, pilosis, pedicellis ochreclas non
excedente; floribus ecu 2 vel 3 per nodulum, hypanthio 0.5 mm.
longo, lobis ovatis, 1 mm. longis, 1 mm. latis, staminibus 8, filamentis
0.5-1 mm. longis, pistillo rudimentario; floribus foemineis 1 per nodulum,
perianthio masculinis similibus, staminibus rudimentariis, pistillo 1 mm.
longo; fructu ovoideo, nigro, glabro, ad 8 mm. longo, 4 mm. diametro,
lobis perianthii coronatis, 1 mm. longis.

CUBA: Oriente, Brefiales de Playa Vaca, Moa, Acuita 13094 (SV),
Clemente 4104 (type, G; 1sotypr, HAB). Collected in flower: Nov.
Collected in fruit: July.

This species resembles C. baracoensis in having orbicular leaves but can
be distinguished by the umbonate character of the leaf blade, the pubes-
cence on the lower surface of the leaf and the glabrous fruits. It also
resembles the small leaf forms of C. geniculata from which it is distinct
on the pubescence of the leaf and the long inflorescence.

8. Coccoloba caesia Ekman ex Schmidt, Fedde Repert Sp. Nov. 24:
74, 1927

Small ieee branches terete, glabrous, nodes not tumid; ochreae mem-
branaceous, 6-8 mm. long, glabrous, turning black on drying; leaf blades
elliptical, rarely ovate-elliptical, apex rounded, base narrowed commonly
oblique, 9 & 3.5, 13 & 4, 17 & 6.5 cm. long and broad, subcoriaceous,
turning black on drying, margin entire, slightly recurved, midrib and
veins slightly prominent above, prominent below, primary veins 8-12
pairs, arcuate anastomosing, secondary venation reticulate, conspicuous
above, obscured below by club shaped papillae arising from the epidermal
cells giving the lamina a gray or ashen appearance; petioles 8-12 mm.
long, glabrous, arising from the middle of the ochreae; inflorescence
terminal, solitary or with a smaller spike at the base, 12-19 cm. long,
rhachis slightly puberulent, staminate flowers in clusters of 2 or 3, pistillate
flowers solitary, bracts broadly ovate, less than 0.5 mm. long, puberulent,
ochreolae borne on stout puberulent stalks 1—-1.5 mm. long in the staminate
flowers, 2~3 mm. long in the pistillate flowers, ochreolae membranaceous,
puberulent, pedicels not evident, hypanthium less than 0.5 mm. long, the
perianth lobes ovate, 1.5 mm. long, filaments of fertile stamens 1.5—1.8 mm.
long; fruit obovoid, 8 mm. long, 4-5 mm. broad, narrowed at the base,
lobes of the perianth appressed against the apex of the achene.

DISTRIBUTION: Cuba.

CUBA: Oriente: Sierra de Nipe, Loma Mensura, Ekman 9894
(TYPE, S; isotypes, G, NY); near Rio Piloto, Ekman 3351 (FM, S), 6420

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 405

(NY, S); Baracoa, Lomas de Cuaba, Ekman 4237 (S); Sierra Maestra,
edge of Arroyo Corojo near Nagua, Ekman 14739 (FM, NY, S); Loma
del Gato, Clemente 3681 (HAB); Florida Blanca, Santiago de Cuba,
Bucher 231 (NY, R). Collected in flower: May, Oct., Nov. Collected in
fruit: Jan

The stalked ochreolae characterize this species which is also distinct on
the gray pubescence of the lower leaf surface. The type specimen is
staminate.

9. Coccoloba pallida Wr. in Griseb. Cat. Pl. Cub. 61. 1866.
a aaa lineart-lanceolata Schmidt, Fedde Repert. Sp. Nov. 24: 76. 1927,
Uvifera pallida O. Ktze. Rev. Gen. 2: 561. 1891.

Shrub or small tree to 7 m. tall; branes terete, papillose or puberulent,
short shoots conspicuously developed, ochreae membranaceous, 1—2 mm.
long, puberulent to pilose; leaf blades ovate to elliptic-ovate, apex rounded
or obtusely acuminate, base rounded to obliquely cordate, 4 * 2,5.5 & 3.5,
6.5 < 2, 9.5 & 6 cm. long and broad, coriaceous, glabrous and pitted
above, minutely and densely white papillose below, the hairs short and
blunt with rounded apices, midrib and primary veins impressed above,
prominent below, primary veins 5-8 pairs, straight or arcuate, bifurcate
and anastomosing near the margin, secondary venation densely reticulate
on both surfaces; petioles 1-4 mm. long, puberulent attached above the
base of the ochreae; leaves of adventitious shoots ovate-lanceolate to
lanceolate-oblong, 11 & 3, to 19 * 4 cm. long and broad, petioles 3—4
mm. long, ochreae 3-8 mm. long; inflorescence terminal, 3.5-6.5 cm.
long, rhachis puberulent, staminate flowers in clusters of 2-4, pistillate
flowers solitary, bracts ovate, less than 0.5 mm. long, puberulent, ochreolae
membranaceous, spreading, to 1 mm. long, puberulent; flowering pedicels
less than 0.5 mm. long; hypanthium less than 0.5 mm. long, perianth lobes
0.5-1 mm. long and broad, fertile stamens less than 1 mm. long; fruiting
pedicels less than 0.5 mm. long; fruit ovoid, to 7 mm. long, 3.5 mm. thick,
perianth lobes appressed against the apex of the achene.

DISTRIBUTION: Cuba.

CUBA: Matanzas: Matanzas City, Ekman 17211 (type Collection
of C. lineari-lanceolata, S). Havana: Jibacoa beach, Leon 13248 (
Pinar del Rio: Cajalbana, Ekman 10488 (S), Leon 4945 (NY);
San Marcos, Wright 2254 (type collection, G, HABA, MO, NY, S$); Loma
Pelada, Cayajabos, Leon 13221 (NY), 13545 (NY), 13812 (G, HAB, Nye )s
Collected in flower: Dec. Collected in fruit: Jan., Mar. Common names:
Uvero blanco, Uverillo.

Schmidt’s statement that Coccoloba lineari-lanceolata is distinct from all
other West Indian species of Coccoloba is evidence he did not recog-
nize the character of adventitious shoots and the accompanying variation
of leaf form that is characteristic of the genus Coccoloba. The Ekman
collection which is the type of C. lineari-lanceolata is of vigorous shoots
probably adventitious in nature. Ekman noted on the collector’s label
that the specimen was rare in Matanzas where he collected it. The char-

406 JOURNAL OF THE ARNOLD ARBORETUM [VOL, Xxx

acteristic pubescence of C. pallida and the striking venation pattern as
well as the high attachment of the petiole on the ochreae allow this sterile
collection to be placed in C. pallida.

10. Coccoloba coriacea Sagra, Fl. Cub. 11: 184. 1853; Lindau, Engl.
Bot. Jahrb. 13: 153. 1890.

Uvifera coriacea O, Ktze. Rev. Gen. 2: 561.

Coccoloba calobotrys Meisner, DC. p rod. 14: gi 1856.

Shrub to slender tree 7 m. tall; branches terete, stout, puberulent at the
nodes, often of determinate growth (short shoots) with several concentric
ochreae at the base, nodes not tumid; ochreae membranaceous, 4—7 mm.
jong, short pilose pubescent; leaf blades elliptic-lanceolate to oblong-
lanceolate, apex acuminate, often with rigid tip, base rounded to cordate,
Ld de Oe Oy Fe 5, 21 & 9 cm. long and broad, coriaceous, ae
brous, adi and veins flat above, conspicuous below, primary veins 5—7
pairs arcuate, bifurcate and anastomosing near the margin, secondary vena-
tion autely reticulate on both surfaces; petioles 5-8 mm. long, puberu-
lent to glabrate; leaves of adventitious shoots 19 « 6, 21 & 9 cm. long
and broad; inflorescence terminal, 10-27 cm. long, rhachis papillose to
puberulent; staminate flowers commonly single, occasionally in clusters
of 2 or 3 flowers, pistillate flowers solitary; bracts ovate, less than 0.5
mm. long, ochreolae membranaceous, 0.5 mm. long; flowering pedicels
7-10 mm. long, glabrous, perianth red, glabrous, hypanthium 1 mm. long,
perianth lobes ovate, 2 mm. long, 1.5 mm. broad, fertile stamens 1 mm.
long; fruiting pedicels to 15 mm. long, fruit ovoid, acuminate at both
ends, to 7 mm. long, 5 mm. thick, perianth lobes sub-coronate

DISTRIBUTION: Cuba.

A: Oriente: Sierra de Nipe, Piedra Gorda to Woodfred, Shafer
3543 (NY, US); Rio Canapé, Ekman 9593 (S); Bio-Trail, Ekman 4778
(NY, S); Pico Turquino, Acufia 10075 (SV); Ensefiada de M fora, Britton,
Cowell & Shafer 13003 (FM, MO, NY, US); Nagua, Mucural Hill, Leon
11031 (NY); Rio Yara, Ekman 14158 (S); 14841 (S); Mayari, Wright
2258 (G, HABA, MO). Pinar del Rio: Loma de Cajalb ana, Ekman
10487 (NY, 8). Cajalbana, Leon & Charles 4932 (NY). Havana:
Vuelta de Abajo, Sagra al (fr.), 290 (f1.) (NY). Collected in flower:
Mar., April, July, Aug., Sept

A specimen collected by Remon de le Sagra in Havana Province and
found in the Meisner herbarium at the New York Botanical Gardens is
the type of Coccoloba calobotrys Meisner. Two packets are in the corners
of this sheet labelled Sagra 544 (fr) and Sagra 290 (fl). The specimen
according to the label was from the DeCandolle Herbarium and may well
be a fragment of the specimen mentioned by Sagra in the original descrip-
tion. Coccoloba calobotrys Meisner therefore may well be based on the
same specimen as C. coricea Sagra. ‘The numbers may have been assigned
to the specimens after the publication of the Flora of Cuba.

This species is difficult to distinguish from C. costata in sterile condition.

1949 ] HOWARD, THE GENUS COCCOLOBA IN CUBA 407

11. Coccoloba retirensis sp. nov.

Arbor parva ramulis teretibus, dense ferrugineo- -pilosis; ochreis cylin-
dricis, coriaceis, dense ferrugineo-pilosis, 3-5 mm. longis; lamina oblonga
vel elliptica. apice acuta, basi rotundata, 5.5 * 2.5, 7 SC See Oe Cin:
longa et lata, coriacea, supra glabra, subtus dense pilosa, nervo medio et
venis primariis supra inconspicuis, subtus prominentibus, venis primariis
5-7, arcuatis, prope marginem bifurcatis, anastomosantibus, venis secon-
dariis supra et subtus dense reticulatis, margine vix recurvato; petiolis
—~8 mm. longis, pilosis; inflorescentia terminalis, 2.5-3.5 cm. ionga, ad
basem plures Soe pilosas gerens, rhachi dense piloso, bracteis ovatis,
0.5 mm. longis, 1 mm. latis, pilosis, ochreolis membranaceis, ad 1 mm
longis, pedicello piloso, 1-1.5 mm. longo; floribus foemineis 1 per nodulum,
glabris, hypanthio 0.5 mm. longo, lobis 5, imbricatis, 1.5 mm. longis, 1-1.5
mm. latis, staminibus rudimentariis, ad 1 mm. longis, pistillo 2 mm. longo,
stylis 3; floribus masculinis et fructu ignotis.

CUBA: Pinar del Rio, Santa Cruz de los Pinos, Retiro, Ekman 18614

Collected in flower March 5, 1924. Ekman’s collectors label bears the
abbreviation “cult.”” The specimen cited is not referable to any species
from the Caribbean, Central or South America now known to me and so I
assume the plant may have been under cultivation as a shade tree and is
native to Cuba.

12. Coccoloba tenuifolia L. Syst. Nat. ed. 10. 1007. 1759, Amoen. 5:
397. 1760; Fawcett and Rendle, Jour. Bot. 51: 124. 1913.

Coccoloba jamaicensis Lindau, Engl. Bot. eat 13: 206. 1890.

Uvifera jamaicensis O. Ktze. Rev. Gen. 2: 561. 1891.

cee leptostachyoides Lindau, Engl. Bot. ae 13: 207. 1890.

Uvifera leptostachyodes O. Ktze. Rev. Gen, 2: 561. 1891.

Coccolobis? frutescens etc. Browne, Hist. Jam. 210, pl. 14, fig. 3. 1750.

Coccolobis bahamensts Bidcn, Bull. N. Y. Bot. Gard. 4: 116. 1905.

Shrub or small tree to 5 m. tall; branches terete, puberulent, becoming
short shoots often well developed, nodes slightly tumid; ochreae mem-
branaceous above, coriaceous and persistent below, to 15 mm. long,
puberulent; leaf blades elliptical, apex short acuminate, base narrowed,
unequally rounded to subcordate, 3.5 X 2,7 4.5, Oe O12 10
cm. long and wide, membranaceous, Subeoriaceus glabrous apore. finely
puberulent below, tomentose in the axils of the primary veins and often
extending onto the secondary venation or the blade, midrib and primary
veins flat or impressed above, prominent below, primary veins 4 or 5
pairs, arcuate, anastomosing | secondary venation finely reticulate; peti-
oles 6-9 mm. long, persistently puberulent, attached above the enlarged
base of the ochreae; inflorescence terminal 8-18 cm. long, weak and
pendent, rhachis puberulent; staminate flowers 1—4 in a cluster, pistillate
flowers solitary, bracts broadly ovate, less than 0.5 mm. long, puberulent,
ochreolae membranaceous about 0.5 mm. long, puberulent; flowering pedi-
cels 1-2 mm. long; hypanthium 0.5-1 mm. long, perianth lobes 1-1.5

408 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

mm. long and broad, filaments of fertile stamens 1 mm. long, rudimentary
stamens of pistillate flowers less than 0.5 mm. long; fruiting pedicels 1-2
mm. long, fruit ovoid to globose, 5-6 mm. long, 4 mm. thick, perianth
lobes appressed to the apex of the achene.

DISTRIBUTION: Jamaica, Cuba, Bahamas. Type from Jamaica.

CUBA: Oriente: S. of Niguero between Rio Nuevo and the coast
in limestone based forests. Ekman 16152 (S).

The specimen collected by Ekman was unidentified in the Stockholm
herbarium. Unfortunately it is a sterile specimen collected Jan. 16, 1923.
It can be referred without question to this species, previously unreported
from Cuba since C. tenuifolia is readily identified by the puberulent, mem-
branaceous leaves and the characteristic attachment of the petiole to the
ochreae sheath.

Lindau listed Coccoloba tenuifolia L. among the uncertain species at
the end of his monograph. The specimen in the Linnaean herbarium is
an excellent flowering specimen, however, and is readily determinable.
Fawcett and Rendle review this situation in Jour. Bot. 51: 124. 1913 and
refer C. jamaicensis and C. leptostachyoides of Lindau to the synonymy of
C. tenuifolia L

Coccolobis bahamensis of Britton must likewise be referred to the syn-
onymy of C. tenuifolia L. The Bahama material is abundant and while
most of the leaves are in the small end of the size range cited in my de-
scription above the species does not merit varietal status.

13. Coccoloba baracoénsis Schmidt, Fedde Repert. Sp. Nov. 24: 73.
1927

Small tree; branches terete, ferruginous pilose-pubescent, nodes not
tumid; ochreae cylindrical, membranaceous, pilose, 1-3 mm. long, deeply
cleft opposite the petiole; leaf blades orbicular, apex rounded, base slightly
cordate, 2.5 & 2.5, 4.5 & 4.5, 8.5 & 6.5 cm. long and broad, coriaceous,
glabrous, smooth and shining above, minutely reticulate and densely
glandular below, margin entire, slightly recurved, midrib and primary
veins inconspicuous above, slightly prominent below, primary veins 3-5

cm. long, rhachis striate, pilose, staminate flowers not known, pistillate
flowers solitary, bracts ovate, 1-1.5 mm. long, pilose, ochreolae membrana-
ceous, flaring, pilose, 1 mm. long, flowering pedicels less than 1 mm. long,
perianth pilose, hypanthium less than 1 mm. long, perianth lobes 1.5 mm.
long, 1 mm. broad, stamens rudimentary, less than 0.5 mm. long, pistil
1-1.5 mm. long; fruiting pedicels shorter than the ochreolae, fruit globose
and crowned by long attenuate perianth lobes, densely pilose, fruit 3 mm.
diameter, the perianth lobes 3 mm. long, achene smooth, shining, golden
in color.

DISTRIBUTION: Cuba.

CUBA: Oriente: Baracoa, Lomas de Cuaba, Ekman 4297 (TYPE, S;

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 409

tsoTYPE, NY); Sierra Azul, Ekman 4427a (S). Collected in flower and
fruit in January
14. Coccoloba nipensis Urban, Fedde Repert. Sp. Nov. 13: 445. 1914.

Coccoloba uviferiella Lundell, Contrib. Univ. Mich. Herb. 6: 12, 1941.

Shrub to tree, 10 m. tail; branches terete, glabrous, nodes tumid, ochreae
membranaceous, tightly cylindrical, 6-8 mm. long; leaf blades orbicular
to elliptical, oval or rarely sub-obovate, apex rounded or slightly emar-
ginate, base cordate, rounded or rarely narrowed, 3 & 2,3 & 3,4 X 4,
5 & 4 cm. long and wide, coriaceous, glabrous, smooth above and below,
densely glandular pitted below; midrib and primary veins flat to obscure
above and below, primary veins 4 or 5 pairs, evident only near the
midrib, margin entire; petioles 3-8 mm. long, glabrous, tumid at the base;
inflorescence terminal 6-9 cm. long, rhachis glabrous, strongly keeled
below the bracts and flowers; staminate flowers solitary or in clusters of
2 or 3, pistillate flowers solitary, bracts ovate, less than 0.5 mm. long,
ochreolae membranaceous, 0.5 mm. long, glabrous, flowering pedicels 1—2
mm. long, hypanthium less than 0.5 mm. long, perianth lobes 0.5—1 mm.
long and broad, fertile stamens less than 1 mm. long; fruiting pedicels
to 4 mm. long, divaricate at a sharp angle from the rhachis, fruit oblong,
rounded at both ends, to 8 mm. long, 5 mm. thick, perianth lobes appressed
against the apex of the achene.

DISTRIBUTION: Cuba.

CUBA: Oriente: Sierra de Nipe, Woodfred, Shafer 3565 (type
Collection, FM, G, MO, NY, US); Rio Piloto, Ekman 15188 (FM, 5S),
LDLOS:. CS) Sierra de Nipe, Ekman 9963 (A, NY); Loma Minqura, Shafer
3815 (FM, MO, NY, US); Rio Joa, Baracoa, Ekman 3661 (S); Moa,
Mrs. Bucher 44 (NY), 45 (NY), 101 (Tyre of C. uviferiella, NY, MICH),
Victorin & Clemente 21712 (G), Howard 5953 (G), Mrs. Bucher oe
number (SV #11106); Playa de Moa, Clemente, Chrysogone & Nestor 4327
(G), Leon & Clemente 23273 (G); Cayo Chico near Moa, een 13093
(SV). Collected in flower: May, June, July. Collected in fruit: July,

Lundell describing C. uviferiella reports “this species suggests the com-
mon sea grape C. uvifera to which it is allied.””. Apparently Lundell over-
looked the description of C. nipensis published 27 years earlier and. the
specimens in the New York Botanical Garden cited above. Coccoloba
uviferiella is a perfect match for C. nipensis and the species must be
placed in synonymy.

A note in Britton’s handwriting in the herbarium in the New York
Botanical Garden reports the type specimen of this species to be in
Berlin. Since that specimen is presumably destroyed I consider the speci-
men at NY to be the new type specimen since it is in the best condition
of the isotypes cited above.

15. Coccoloba uvifera L. Syst. Nat. ed. 10, 1007. 1759.

Polygonum Uvifera L. Sp. Pl. 365. 1753.

Guaiabara Uvifera (L.) House, Am. Midl. Nat. 8: 64. 1922.

410 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Tree of strand areas, 2-15 m. tall, branches terete, stout, papillose to
pilose, nodes not tumid, ochreae rigid coriaceous at the base, membrana-
ceous at the apex, 3-8 mm. long, papillose to pilose; leaf blades orbicular
to reniform, apex rounded, truncate or emarginate, base rounded to
broadly cordate, one lobe often extending around the petiole, 6 x 8,
11 X 13, 13 X& 18 cm. long and broad, thick and fleshy when fresh,
coriaceous when dry, glabrous and minutely punctate on both surfaces,
midrib and primary veins prominent on both surfaces, frequently brightly
colored when fresh, primary veins 3—5 pairs, usually straight, bifurcate
and weakly anastomosing near the margin, commonly barbate in the axils
of the basal veins, secondary venation minutely reticulate or obscure;
petioles stout, 7-10 mm. long, papillose to pilose; leaves of adventitious
or fast growing shoots usually variable in size and shape commonly obo-
vate; inflorescence stout, 15-30 cm. long, rhachis puberulent; staminate
flowers in clusters of 1-7, pistillate flowers solitary, bracts ovate, 1-1.5
mm. long, 2 mm. broad, puberulent, ochreolae membranaceous, 1 mm.
long, puberulent, flowering pedicels 1-2 mm. long, perianth yellow-white
or greenish, hypanthium 2—3 mm. long, perianth lobes 4 mm. long, 3-4
mm. wide, fertile stamens to 4 mm. long; fruiting pedicels 3-4 mm. long,
fruit obpyriform, 1.2-2 cm. long, 8-10 mm. diameter, narrowed at the
base, rounded truncate at the apex, the perianth lobes appressed against
the apex of the achene, perianth rose-purple when mature, the achene
black.

DistripuTion: Along shores, Florida, Bermuda, through the Carib-
bean to S. America.

CUBAS Oriente = Guibara to. Punta Picacos, Shejerd4s> CUS, NY j:
DAltiIago Faure eee NY). dias neo Soledad, Jack 585 (A),
4118 (A), 4533 (A), 5092 (A), 5319 (A); Rancho Luna, E. of Cienfuegos
wee piel 4209 (G); Castillo de Jagua, ; ombs 569 (G, FM, MO, NY).
Nose eget ee ae La number (C, several sheets); Morro Coji-
mar, le gf2 3S va de Marianao, Palmer & Riley 852 (US);
Rancho Bayeros, lV a. pee (i) 9 soaniiaeo las Vesss. aos. or dd, AV
ee 19915 (NY); Playa Mariano, Van Hermann 887 (NY). Cama-
guey: Cayo Sabinal, Shafer 1125 (US, NY). Isle of Pines: Nueva
ae Curtis without NO. CHY }. —— locality: Wright 2259 ae
Collected in flower: Mar., April, Aug., Sept. Collected in fruit: June, July
Aug. Common names: uva, uva de la fe uvero, uva de playa.
15a. Coccoloba uvifera L. C. diversifolia Jacq., hybr. nov.

Shrub or tree; branches terete, minutely pilose; ochreae coriaceous,
persistent, oblique and flaring at the apex, pilose; leaf blades oblong to
obovate-oblong, apex rounded to sub-acute, base obliquely rounded to
obliquely cordate, 11 & 8, 14 & 9.5, 16 X* 10.5 cm. long and broad,
coriaceous, midrib and primary veins prominent on both surfaces, sec-
ondary venation conspicuously reticulate both surfaces, primary veins 5—7

midrib and veins below and in axils of veins and on lamina parallel to
midrib, densely resinous glandular below; petiole stout, 1-1.5 cm. long,

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 411

minutely pilose; inflorescence terminal 7-15 cm. long, rhachis puberulent,
bracts broadly ovate 1 mm. long, puberulent, ochreolae membranaceous,
1 mm. long, puberulent, flowering pedicels 1.5-2 mm. long, puberulent;
staminate flowers in clusters of 2-3, hypanthium 0.5 mm. long, perianth
lobes 0.5-1 mm. long, stamens 1 mm. long; pistillate flowers not seen;
fruiting pedicels 2-4 mm. long, fruits not seen.

CUBA: Pinar del Rio: Sta. Cruz de Los Pinos, Ekman 18615
(S), Range, Roig 11935 (SV).

In the nature of the pubescence, the texture of the leaves and the pattern
of primary venation this material approaches C. uvifera. The secondary
venation and the type of glands found on the leaves are similar to those
of C. diversifolia. The shape of the leaf blade is intermediate between
these two species. Coccoloba uvifera apparently hybridizes freely with
many other species of the same genus. Unfortunately single collections in
the herbarium are usually of little value in the study of possible hybridiza-
tion and field studies are needed with just this idea in mind.

16. Coccoloba Northropiae Britton, Bahama Flora 117. 1920, (as
Coccolobis).

Coccoloba leoganensis var. parvifolia Griseb. Cat. Pl. Cub. 61. 1866.

Coccoloba retusa Lindau, Engl. Bot. Jahrb. 13: 150. 1890, in part.

Shrub or small tree to 4 m. tall; branches terete, pilose becoming
glabrate, nodes slightly tumid; ochreae membranaceous, 2—3 mm. long,
glabrous; leaf blades orbicular to Soe or elliptic, apex rounded, obtuse
or emarginate, base narrowed, 3 * 2.5, 4 * 3 cm. long and oide: thinly
coriaceous, glabrous, turning black on drying, commonly shining above,
dull beneath, midrib and primary veins equally prominent on both sur-
faces, secondary venation reticulate, primary veins 4-6 pairs, straight,
diffuse branching and becoming reticulate near the margin, not arcuate
and anastomosing, margin entire, flat; petioles 3-5 mm. long, short pilose
pubescent; leaves of adventitious shoots variable in shape, 5 & 3,6
cm. long and wide on petioles to 7 mm. long; inflorescence terminal 5—7
cm. long, rhachis puberulent, the female more so than the male, bracts
ovate, less than 0.5 mm. long, ochreolae membranaceous, less than 0.5
mm. long; male flowers in clusters of 2, female flowers solitary, flowering
pedicels 0.5-1 mm. long, hypanthium 1 mm.‘ long, perianth lobes 1 mm.
long and wide, fertile stamens less than 1 mm. long, stamens of pistillate
flowers rudimentary, less than 0.5 mm. long; fruiting pedicels 1.5 mm.
long, fruit ovoid, to 5 mm. long, 3 mm. thick, perianth lobes appressed
against the apex of the achene.

DISTRIBUTION: Cuba, Bahamas. (Type collection, Curtis without
number from New Providence, Bahamas

CWBRA Oven te. s-csantiaco: ide ent ae LIZ0O" (St an
German, Ekman 6346 (S); Loma Estrella, Sierra de Nipe, Ekman 6392

inar del Rio: Morillo, Ekinan 17376 “(S): Viniales, Sierra de
Vifiales, Ekman 18037 (NY, S): Sierra de Pendejeral, groupa del Rosario;
Sierra de Organos, Ekman 17525 (S). Without location: Wright 2251 (G,

412 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXX

HABA, MO, S). Collected in flower: Mar., May, June, July. Collected
in fruit: Sept. Local name: uverillo.

Both Lindau and Grisebach have misinterpreted C. retusa by including
Wright 2251 in that species. The puberulence of the plant and the char-
acter of the venation of the leaves make this Wright specimen, and the
other specimens cited above, clearly distinct from C. retusa. This repre-
sents the first report of C. Northropiae from Cuba.

17. Coccoloba retusa Griseb., Cat. Pl. Cub. 61. 1866; Lindau, Engl.
Bot. Jahrb. 13: 150. 1890, in part.
Uvifera retusa O. Ktze. Rev. Gen. 2: 562. 1891.
Coccoloba retusa forma acuminata Lindau, Engl. Bot. Jahrb. 13: 151.

Coccolobts colomensis Britton, Bull. Torrey Bot. Club 42: 369. 1915.

Shrub 1-2 m. to tree 17 m. tall; branches terete, glabrous, nodes slightly
tumid, ochreae membranaceous, 5-9 mm. long, glabrous; leaf blades
ovate, oblong, elliptic or obovate, apex acute, rounded or emarginate,
rarely acuminate, base narrowed, rounded or subcordate, commonly un-
equal, 3.5 & 2.5, 7 & 5.5, 10 & 4.cm. long and broad, coriaceous, gla-
brous, shining above, turning black on drying, midrib and primary veins
prominent on both surfaces, primary veins 4—7 pairs, straight towards the
margin bifurcating and anastomosing near the margin, secondary vena-
tion reticulate, conspicuous both surfaces, margin entire or undulate;
petioles 6-10 mm. long, glabrous; inflorescence terminal, 4-14 cm. long,
rhachis glabrous, staminate flowers solitary or in clusters of 2—4, pistillate
flowers solitary, bracts ovate, less than 0.5 mm. long, ochreolae mem-
branaceous less than 0.5 mm. long, flowering pedicels developing after the
flower opens, i.e. flowers sessile at the apex of the inflorescence, or on
pedicels 0.5 mm. long near the base, hypanthium 0.5 mm. long, perianth
lobes 0.5-0.7 mm. long and broad; fertile stamens less than 1 mm. long;
fruiting pedicels 0.5-1 mm. long, fruit ovoid, rounded at the base, to 6
mm. long, 3 mm. thick, perianth lobes coronate.

DISTRIBUTION: Cuba.

CUBA: Oriente: Loma del Gato, Leon, Clemente & Roca 10269
(NY), Leon 12363 (NY), Acura 9820 (SV), Crisostome Sore a) erha
de Nipe, Rio Jimbambay, Ekman 15138 (G, S); Rio Piloto, Ekman 4748
(A, NY, S), 5014 (S), 5794 (NY, S), 6415 (FM, MICH, S), 6691 (A,
55 9767 (NY, S), 9790 (MO, S); Rio Piedra, Ekman 9817 CN)
Nagua on Rio Yara, Ekman 14840 (A, FM, NY, S, US); Cayo del Rey,
ee 4793 (S); Papayo near Sevilla, Ekman 9462 (A, BMG, MOe RY,

5, Uo); Siig near Cauto, Ekman 6192 ee a Arroyo del Cristo,
S O gua, Ekman 14754 (FM, NY, S); rco Largo, Punta Padre,
Curbelo 281 (NY); Bayate near ener . 2m Bee 4668 (S); Mana-
cal, Sierra Maestra, Ekman 9358 (MICH, S); El Paraiso, Holguin, Ekman
7607 (S); Santiago, Clemente 4242 (HAB); Mucural Hill, near Nagua,
Leon 11034 (NY); Rio Seboruco to falls of Rio Mayari, Shafer 3688 (G,
NY, US); Jagueyes de Mulas, Roig 14 (NY); Facallores, Wright 1668
(type Collection of C. retusa f. acuminata, G), Wright 2252 (G, HABA,

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 413

MO, NY, S, type Collection C. retusa, see discussion). Las Villas:
Gaeailen Combs 374 (FM, G, MO, NY); Savana de Motembo, Leon,
Fortun & Edmund 8568 (NY); Cue Ramones, Jucaro Quemado, Peninsula
de Zapata, Roig & Cremata 2255 (FM); Sabana de Manacas, Laguna Asiento
Viejo, Leon 9283 (NY); Central Australia, Peninsula Zapata, Ekman
18328 (S); Lomas de Banao, Luna 110 (NY); El Porvenir to Aguacate,
Britton & Wilson 5361 (NY); Mina Carlotta, Sierra de San Juan, Howard
5628 (G), 5661 (G). Havana: Loma de la Pita, San Miquel, Leon,
Ekman, Johnston, & Roig 9101 (NY), Leon 11529 (NY); Loma de Ne-
Baa eee ae Miquel de Casanova, Leon ae ee Campo ees at

Rio Que “Ekman 13237 GAs HL NY Se WS ab ieartrred elk kat
Loma eee Cayajabos, Leon & Roig: 13538 Gaunt 15959 SENAY e Ton
13205 (NY); Coloma, Britton, Britton & Gager 7037 (NY, US, type
Collection of C. colomensts), Ekman 17828 (NY, S), Britton & Cowell
9548 ee NY, US); Sabalo, Rio Sabalo, Ekman 11447 (NY, S); Savannas
SE of Los Palacios, Leon & Roca 7368 (NY); Los Palacios to San Juan
de Zayas, Shafer 1813 (NY); Rio Blanco, Ekman 17294 (S). Isle of
Pines: Rio Las Casas, Nueva Gerona, Ekman 12005 (S), 12010 (S);
Manati, Blain 185 (FM); Los Indios, Britton, Britton & Wilson 14243
(NY); Coe’s Camp, Ensefiada de Siguanea, Britton & Wilson 14846 (NY,
US); Mal Pais to La Ceiba, Ekman 11915 (S); San Juan, Rotg & Cremata
1783 (NY). Locality Unspecified: Herb. Sauvalle 2173 (HABA). Col-
lected in flower: Feb., Mar., June, July, Aug., Sept., Oct., Dec. Collected
in fruit: Jan., Feb., Sept., Oct., Nov., Dec. Local names: uverillo, Cocuyo
de la maestra

The numerous collections of Coccoloba retusa cited above show a tre-
mendous range of variation in the size and shape of the leaves. Lindau
described one form for the species but since this is just a small part of
the normal variation of leaf shape it is not recognized as valid in this
treatment.

Coccolobis colomensis which Britton based on a Britton & Gager speci-
men from Pinar del Rio is also separated by Britton on characters found
in the shape of the leaf. Leaf shape is an unstable base for specific dis-
tinction in this genus and I am reducing Britton’s species to synonymy of
C. retusa.

The type collection of Coccoloba retusa is Wright 2252. I have seen six
examples of this collection as cited above but in the packets attached to
the sheets of the Gray Herbarium collection are a variety of notes regard-
ing the location where the collection was made. These are “Monte Verde
Sept. 7,” “Potosi, Monte Yow, Oriente, Oct. 10,” “Retin, Oct. 19.” Thus
the collection seems to be mixture. All of the material pound under
the number Wright 2252 however is referable to C. r
18. Coccoloba praecox Wright ex Lindau, Engl. Bot. ates 13: 142.

1890
Uvifera praecox O. Ktze. Rev. Gen. 2: 562. 1891.

Coccoloba Ekmani Urban, Fedde Repert. Sp. Nov. 14: 331. 1916.

Shrub to tree 9 m. tall; branches terete, nodes tumid, twigs puberulent
when young becoming glabrate; ochreae membranaceous, puberulous, 4-6

414 JOURNAL OF THE ARNOLD ARBORETUM Dyes! Sekar

mm. long; leaf blades ovate to oblong-ovate, apex acute with a mucro or
rounded, occasionally slightly emarginate, base rounded to sub-cordate,
3 & 2.5, 4.5 & 3.5, 5 & 4.5 cm. long and broad, thin coriaceous to
chartaceous, glabrous, upper surface smooth or minutely reticulate, lower
surface minutely reticulate, midrib impressed above, slightly prominent
below, puberulent, primary veins 5—7 pairs, inconspicuous, arcuate anas-
tomosing near the margin, margin entire with a prominent edge, flat;
petioles 4-8 mm. long, puberulent; leaves of adventitious shoots, ovate-
oblong, rounded or emarginate at the apex, 5 & 5.5 cm. to 6.5 & 4.5 cm.
long and broad; inflorescence terminal 2—9 cm. long, rhachis puberulent,
staminate flowers solitary or in clusters of 2 or 3, pistillate flowers sania
bracts ovate less than 0.5 mm. long, puberulent, ochreolae less than 1 m
long, flowering pedicels 0.5-1 mm. long, hypanthium 0.5 mm. long, es
anth lobes 1-1.5 mm. long, fertile stamens with filaments 1—-1.5 mm. long;
fruiting pedicels 1—-1.5 mm. long, fruit ovoid, 4 mm. long, 2.5 mm. thick,
perianth lobes sub-coronate.

DISTRIBUTION: Cuba.

CUBA: Oriente: Charco, Wright 2253 (type Collection, G, HABA,
MO, NY, S); Punta Padre, Curbelo 74 (HAB, NY), Roig 5537 (NY);
Mir, between Victoria de las Tunas and Pilea Ekman 7541 (MO, S);
Anate, near Caimito del Guayabal, Ekman 188 (S). Camaguey: Cama-
puey “City, Britton, Briton G@ Cowell I3i15 (NY; VS) Sabana de la
Caridad, Cespedes, Roig, Luaces & Aranjo 910 (NY); Santayana, ae
19036 (3S); Savana de Providencia, Roig 6209 (NY). Las Villa
Mordazo, Leon 9232 (NY), Leon & Casatas 5929 (NY); Casilda, ae
18879 (S). Matanzas: Cuabal del Espinal, Canasi region, Ve eon &
Roig 13354 (HAB, NY), 13355 (NY), Acuta 11290 (SV); Pan de Matan-

zas, Ekman 16510 (S). Havana: Cuabal de Salomon, Minas, Leon
13787 (NY), 13887 (NY); Cuabal de Jestis Maria, Minas, Leon 13334
CNY 5 335 (NY) 2 Madrica, Leon Ge Cesare 8940 (NY); Lomas de
Jatas, Guanabacoa, Ekman 580 (type Collection of C. Ekmani, FM, S),
1OP0D= Ce J0I522 (Cos £09320 (INN, 3) Leon 7sqe NY LAB) Lia
Coca near Campo Florida, Ekman 13220 (S), 16437a (S), 16437b (A, S),
19005 CA, EM, MO. NY, S, US): Pinar del “Rao? -Mayar.5.01
San Cristobal, Fors 4790 (NY). Collected in flower: Apr., May. Collected
in fruit: Apr., May.

The type collection of C. Ekmani, (Ekman 580) was considered distinct
on the ovate to suborbicular leaves however these minor variations in
shape can be matched in many leaves of the original Wright collection.
Coccoloba Ekmani is not a valid species.

19. Coeccoloba Shaferi Britton, Bull. Torrey Bot. Club 42: 369. 1915,
(as Coccolobis).

Coccoloba azulensis Schmidt, Fedde Repert. Sp.’ Nov. 24: 73. 1927,

Shrub or tree to 7 m. tall; branches terete, puberulent or papillose,
contorted, short shoots, commonly developed with concentric ochreae,
ochreae 7-9 mm. long, membranaceous, puberulent; leaf blades ovate to
ovate-elliptic, rarely lanceolate-ovate, apex obtusely acuminate to rounded,

1949 ] HOWARD, THE GENUS COCCOLOBA IN CUBA 415

base cordate, 5 & 3,7 & 5.5,9 & 6, 14 & 9 cm. long and wide, coria-
ceous, glabrous, dull both surfaces, midrib and primary veins slightly
impressed above, prominent below, primary veins 5-7 pairs, straight to
the margins, not arcuate nor anastomosing or rarely bifurcating close to
the margin and reticulate, blades densely pitted on both surfaces, sec-
ondary venation reticulate slightly conspicuous both surfaces, margin
entire, slightly revolute, petioles 7-10 mm. long, glabrous; leaves of vig-
orous or adventitious shoots lanceolate-ovate, apex acuminate, 12 « 3 to
14 x 4.5 cm. long and wide; inflorescence terminal 10-25 cm. long,
rhachis, bracts, ochreolae and perianth papillose or puberulent, flowering
pedicels shorter than the ochreolae, staminate flowers in clusters of 2—4
flowers, rarely ‘solitary, pistillate flowers solitary, bracts ovate, 2 mm.
long, ochreolae membranaceous, 2—3 mm. long, perianth red, hypanthium
2-3 mm. long, perianth lobes oblong, 3 mm. long, 2 mm. broad, fertile
stamens 2 mm. long; fertile pistil 4 mm. long; fruit not known.

DISTRIBUTION: Cuba.

CUBA: Oriente: near mouth of Rio Yamaniguey, Shafer 4252
(NY); Camp Toa to Camp La Barga, Northern Oriente, Shafer 4165 (Tyre,
NY); Sierra de Moa, Ekman 4511 (S), Bucher 104 (NY), 104a (NY),
Shafer 8351 (NY), Howard 6050 (G); Monte de la Brena, Sierra de Moa,
Leon, Clemente & Alain 22567 (G), 22582 (G), Clemente & Crisostomo
4966 (G); Sierra Azul, Ekman 4427b (type collection of C. azulensis, S);
Narave at Baracoa, Ekman 4054 (NY, S). Collected in flower: Jan., Feb.,
Mar., Apr., July, Aug., Sept., Dec.

Following the original description of C. azulensis Schmidt reports his
new species to be similar to C. Skaferi but differing in the longer inflo-
rescence and the glabrous inflorescence rhachis. Apparently this com-
parison was made with the original description of C. Shaferi for the
specimens of C. Shaferi cited above have inflorescences 10-20 cm. long
instead of the 12 cm. originally reported by Britton. The pubescence of
the inflorescence rhachis, bracts, ochreolae and perianths varies consid-
erably from small protuberances of epidermal cells called a papillose
puberulence to short pilose hairs. In all characters the Ekman specimen
selected as the type of C. azulensis seems to grade into the larger collection
of C. Skaferi and may be regarded as an extreme of variation. I have
therefore reduced Schmidt’s species to the synonymy of C. Shaferi.

20. Coccoloba costata Wr. ex Sauvalle, Fl. Cub. 139. 1868; Lindau,
Engl. Bot. Jahrb. 13: 155. 1891; Schmidt, Fedde Repert. Sp. Nov.
DO Oo 1O29:

Uvifera costata O. Ktze. Rev. Gen. 2: 561. 1891.

Coccoloba leoganensis var. cordata Griseb. Cat. Pl. Cub. 61. 1866.

Tree; branches stout, pubescent with a ferrugineous to golden pubes-
cence; ochreae membranaceous, 4-6 mm. long, ferrugineous puberulent;
leaf blades ovate to elliptic, apex obtuse, acute or obtusely acuminate,
base obliquely cordate, 7 « 5, 13 & 8,18 & 12, 24 & 15 cm. long and
broad, coriaceous, golden shining above, dull brown below, midrib and

416 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

veins impressed above, prominent below, primary veins 5-7 pairs, arcu-
ate, anastomosing, blade pitted, slandular above and below, glabrous;
nerole stout, 8-10 mm. long, slightly puberulent; leaves of adventitious
shoots to 35 & 22 cm. long and broad, petioles 1.5 cm. long; inflorescence
terminal, rhachis puberulent, 15-20 cm. long, staminate flowers in clusters
of 2-4, pistillate solitary, bracts ovate, 0.5 mm. long, ochreolae membrana-
ceous, 0.5 mm. long, flowering pedicels 0.5 mm. long, hypanthium 0.5 mm.
long, perianth lobes 0.5-1 mm. long and broad, fertile stamens 1 mm.
long; fruiting pedicels to 1.5 mm. long, fruit globose, to 6 mm. long, 3
mm. thick, perianth lobes coronate.

DisTRIBUTION: Cuba.

CUBA: Oriente: Monte Verde, Wright 1393 (G, HABA, MO,
type Collection) ; Sierra de Nipe, Rio Pp iloto, oe oo (FM, NY, a
Piedra Gorda to Woodfred, Shafer 3084 (FM, G, US )s Arroyo del
Medio, Ekman 15243 (A, Se Loma Mensura oe oe (S)¢ Baracoa,
Lomas de Cuaba, Ekman 4229 (S); Moa, Leon, Clemente & Alain 22518
(G), Clemente 4360 (G); Punta Gorda, Sierra de Moa, Clemente 4053 (G) ; :
Alta Loma Naranja, Bucher without ane (SV #10099). Collected in
flower: Jan., June. Collected in fruit: Aug., Oct. Common name: uvilla.
21. Coccoloba benitensis Britton, cn Torrey Bot. Club 42: 370. 1915,

(as Coccolobis).

Coccolobis monticola Britton, Bull. Torrey Bot. Club 50: 37. 1923.

Coccolobis brevipes Britton, Bull. Torrey Bot. Club 42: 371. 1915.

Shrub or small tree; branches terete, densely puberulent, commonly
branching in one plane, nodes slightly tumid; ochreae 3-4 mm. long,
membranaceous, densely ferrugineous pilose; leaf blades ovate to ovate-
elliptic, apex acute to acuminate, base narrowed or rounded, 2 x 1,
A 2 Oe hyd 0-35 CM, long and wide, coriaceous, elabrous aud
pitted anave. coarsely reticulate and glabrous below but densely to spar-
ingly covered with peltate resinous exudate, margin entire, slightly
recurved, midrib impressed above, slightly keeled when dry, primary veins
3 or 4 pairs, prominent both surfaces, secondary venation reticulate, con-
spicuous both surfaces; petioles 2-4 mm. long, flattened above, ferru-
gineous pilose on the upper surface, commonly covered with peltate
resinous exudate below; leaves of adventitious shoots with ochreae 9-12
mm. long, blades to 10.5 5 cm. long and wide, petioles to 8 mm. long;
inflorescence terminal or terminal on short shoots, 2—3 cm. long, several
concentric ochreolae at the base, rhachis puberulent or with glandular
exudate, bracts minute, less than 0.5 mm. long, ochreolae minute, less
than 0.5 mm. long, flowering pedicels less than 1 mm. long; staminate
flowers in clusters of 2 or 3, pistillate flowers solitary, hypanthium 1 mm.
long, lobes 0.5 mm. long, filaments of staminate flowers 1 mm. long, in
pistillate flowers 0.1 mm. long, rudimentary pistil in male flower 0.3 mm.
long, pistil in female flowers to 3 mm. long; fruiting pedicels 1-1.5 mm.
long, fruit ovoid to globose, 6 mm. long, 4 mm. thick, perianth lobes
appressed at the apex of the achene, perianth of fruit commonly with
glandular exudate.

1949 ] HOWARD, THE GENUS COCCOLOBA IN CUBA 417

DISTRIBUTION: Cuba.

CUBA: Oriente: Camp San Benito, Shafer 4049 (type, NY), 4044
(FM, NY, US); Sierra de Moa, Camp la Gloria, Shafer 8230 (NY, US);
Sierra Maestra, Arroyo Jiménez, Ekman 14802 (NY, S); Rio Yara and
Rio Palmamocha, Ekman 14331 (US, S), 14389 (A, FM, S) ; Pico Turquino,
Leon 10713 (type of C. monticola, NY), Leon 10902 (NY); Sierra Maestra,
Estribo Cardero, Roig, Acuna & Bucher 6590 (NY); Arroyo Veinticinco
Wright 2257 (type Collection of C. brevipes, G, HABA, MO, NY).
lected in flower: July, Aug. Collected in fruit: Aug.

Britton failed to recognize the various patterns of growth found in
specimens of Coccoloba. The presence of short shoots, vigorous growth,
and adventitious shoots each with very different aspects and sizes of
leaves, petioles and ochreae has led to numerous duplication of species
descriptions. Coccoloba benitensis was chosen as the type and name of
this group because the shoots are near the normal to be expected for the
species and because the specimens have female flowers and fruit. The
type specimen of C. brevipes Britton is described by Wright in field
notes as a subscandent bush. This I judge to be vigorous shoots and the
larger internodes, leaves, petioles and ochreae support this view. On the
other hand the material of C. monticola is from a specimen which appears
stunted but which possesses short shoots and has small internodes, leaves,
etc. Britton was impressed with the peltate “black dots” on the lower
leaf surface which are actually glandular excretions. These may be present
on the upper or lower leaf surface of all specimens as well as on the
rhachis of the inflorescence and on the perianth and the fruit. The amount
of resinous exudate is frequently impressive.

22. Coccoloba rufescens Wr. ex Lindau, Engl. Bot. Jahrb. 13: 143.
1890

Uvifera rufescens O. Ktze. Rev. Gen. 2: 562. 1891.

Coccoloba rufescens Wr. forma longifolia Lindau, 1.c. 13: 143. 1890.

Coccoloba punctata var. parvifolia Griseb. Pl. Wright. 175. 1860.

Shrub or small tree, 2-4 m. tall; branches terete, nodes slightly tumid,
youngest branches ferrugineous pilose; ochreae cylindrical, 3-4 mm. long,
cleft almost to the base, pilose; leaf blades ovate, elliptic or oblong-ovate,
apex acute to acuminate, base narrowed, rounded, often oblique, 5 X 2.5
7.5 < 3,9 & 4 cm. long and broad, coriaceous, smooth above, meals,
shining, midrib and primary veins sub-prominent below, secondary vena-
tion obscure, lamina densely and minutely glandular punctate below;
petioles 3-5 mm. long, pilose; inflorescence terminal, 2-6 cm. long,
rhachis pubescent; staminate flowers unknown; pistillate flowers solitary,
bracts broadly ovate, less than 1 mm. high, 2 mm. broad, pilose, ochreolae
membranaceous 0.5 mm. long, strongly bilobed, pilose, pedicels shorter
than the ochreolae, perianth white, glabrous, hypanthium 1—1.5 mm. long,
lobes ovate, 1-1.5 mm. long, stamens rudimentary, pistil oblong triangular;
fruit globose, 5 4 mm., slightly narrowed at apex and the base, perianth
lobes coronate, 1.5—2 mm. long.

418 JOURNAL OF THE ARNOLD ARBORETUM [VOL., XXX

DISTRIBUTION: Cuba.

GUBAS “Oriente> Mi. Verde: een 462 (type Collection, FM,
G, MO, NY, 8S), Wright 1394 (FM, G, HABA, MO, NY, S); Baracoa,
Lomas de Cuaba, Ekman 4220 (NY, Sy. Montes de la Brena Moa, Acuna
13090: (SV ), Acuna 13092 (SV): Camp Toro to Camp La Bares, Shafer
4152 (NY); Cayoquan, Victorin & Clemente 21519 (HAB); Sierra de
Nipe, Bucher without no. (SV #11493). Without location: Wright 462a
(G). Collected in flower: Feb., Apr. Common name: uvero de costa.

23. Coccoloba Wrightii Lindau, Engl. Bot. oe | Kea i cys Pa oe) 8
Schmidt, Fedde Repert. Sp. Nov. abe a

Uvifera Wrightu O. Ktze. Rev. Gen. 2: on

Coccolobis saxicola Britton, Bull. Torrey = Club=50 37-1928.

Shrub or small tree, to 3 m. tall; branches terete, glabrous, ochreae
membranaceous, 4-6 mm. long, puberulent to tomentose; leaf blades
ovate, elliptic, obovate or rarely ovate lanceolate, apex acute to abruptly
short-acuminate, base narrowed to obtuse, usually slightly oblique, 5 & 2.5,
8 X 4,10 * 7,11 X 10 cm. long and broad, coriaceous, young leaves
often shining above, mature leaves generally dull both surfaces, midrib
and primary veins impressed above, the lamina commonly umbonate be-
tween the veins, primary veins 4-6 pairs, prominent below, arcuate,
anastomosing near the margin, secondary venation reticulate, conspicuous
on both surfaces, leaves glandular below, margin entire, slightly recurved;
petioles 4-7. mm. long, papillose to puberulent; leaves of adventitious
shoots to 20 & 17 cm. long and broad, petioles 2.5 cm. long, ochreae to
2 cm. long; inflorescence terminal, 3-10 cm. long, rhachis papillose with
glandular excretions, staminate flowers in clusters of 2—5, pistillate flowers
solitary, bracts ovate, 0.5 mm. long, ochreolae membranaceous 1 mm.
long, papillose or glandular, flowering pedicels 1 mm. long, perianth
greenish white, hypanthium 1 mm. long, perianth lobes 1-1.5 mm. long
and broad, fertile stamens with filaments united below into a tube 1 mm.
long, free portion of filaments 0.5 mm. long; fruiting pedicels 1-3 mm.
long, fruit ovoid, slightly contracted at the base, rounded at the apex, 7-9
mm. long, 4-5 mm. diameter, perianth lobes slightly coronate.

DISTRIBUTION: Cuba.

CUBA: Oriente: Sierra Cristal, at Rio Lebisa, Ekman 6864 (NY, S);
Sierra de Imias, crest of Punton del Maté, Leon 12280 (G, NY); Manacal,
Sierra Maestra, Ekman 9378 (A, NY, S); Cueva del Auta, Sierra Maestra
Roig & Bucher 6688 (NY), Bucher 171 (R, HAB); Loma del Gato, Bacher
259 (NY), Leon, Clemente & Roca 10167 (type Collection of C. saxicola,
NY); Pico Turquino, Sierra Maestra, Acufia 7671 (R); Maestra Range,
Leon 10712 (NY, R); Alto Loma del Nat ranjo, Baracoa, Bucher oe ciiwen
SV); Arroyo Jiménez, Sierra Maestra, Ekman 14785 (NY, S); golo-
songo, Loma del Gato, Clemente 1815 (NY ”); Rio Yara and Rio eo.
divide, Sierra Maestra, Ekman 14386 (S); Loma Mensura, Sierra de Nipe,
Ekman 9922 (NY, S): Monte Real, Sierra del Cobre, er 1803 (5)
Valley of Rio Yaminiguey, Shafer 4229 (NY, US); Monte de la Brena,
Sierra de Moa, Leon, Clemente & Nestor 23325 (G); Punta Gorda, Sierra

1949 | HOWARD, THE GENUS COCCOLOBA IN CUBA 419

de Moa, Clemente 4038 (G); Monte Grande de Centeno, Leon, Clemente &
Alain 22664 (G); Charrascal del Coco, Sierra de Moa, Leon, Alain &
ieee 22031 (G); La Guinea, Wright 1395 (FM, e: MO, NY, type
Collection). leas. Vai lle eons de Banao, Luna 468 (NY); El Burial
and oe Guineos, Lomas del Banao, Ekman 16235 (S); Pico Potrerillo,
Trinidad Mts., Ekman 18947 (S). Collected in flower: June, July, West
Collected in ae: Jan., Mar., July, Aug., Oct

This is a species of mountain areas, epilectedl from 700-1100 meters
altitude. Lindau described the type collection of Wright 1395 as flowering
in June. The Wright material cited above which I have seen was col-
lected Dec. 17, 1859 according to the label attached and the material is
in fruit. Apparently this is another mixture of Wright labels.

24. Coccoloba Leonardii, sp. nov.

Arbor ad 10 m., ramulis teretibus, glabris, nodis vix tumescentibus,
ochreis subcoriaceis, glabris, 3 mm. longis; lamina ovata, apice rotundata
vel acuta, basi rotundata vel subcordata, lateribus inaequalibus, in petiolum
subdecurrente, 6.5 & 4, 8 & 5.5, 11 & 7 cm. longa et lata, coriacea,
glabra, supra cineracea, subtus brunnea, nervo medio et nervis primariis
supra et subtus prominentibus, nervis primariis 5—7, arcuatis, anasto-
mosantibus, venis secondariis reticulatis; petiolis 8-11 mm. longis, supra
late canaliculatis, glabris; inflorescentia terminalis, 5.5-14 cm. longa,
rhachi glabra, bracteis ovatis, 0.5-1 mm. longis, ochreolis membranaceis,
1 mm. longis, pedicello ochreolas non excedente; floribus masculinis 2-4
per nodulum, hypanthio 0.5 mm. longo, lobis 1.5 mm. longis et latis,
filamentis ad 1 mm. longis, pistillo rudimentario circa 0.5 mm. longo,
floribus foemineis perianthio masculinis similibus, staminibus rudimen-
tariis, pistillo 1-1.5 mm. longo; fructu late fusiformi, 10 * 5, 11-*& 7
mm. longo et diametro, supra nigro, subtus brunneo, lobis perianthii sub-
coronato, 1 mm. longo.

DISTRIBUTION: Haiti, Cuba.

HAITI: Tortue Island: Tableland N.E. of Basse Terre, 30’
tree in dry thickets in rocky ravines, E. C. & M. Leonard 12466 (TYPE,
NS ISODVEES¢ VEO. INN LS) = ee Val leCia: 6. & G. M. Leonard 11335
(MICH, US), 11381 (A, GURUS) aie GNSS e425. (GUS)
Pte. Petite Bois, Ekman 4150 iS). Navassa Island: West of the
lighthouse, Ekman 10843 (S, US). Dept. du Nord: Morne la Vigie,
Cap Haitien, Ekman 2706 (S, US); Bayeux, Nash 2os«(EM, NY). Dept.
du Nord-QOuest: Saline Michel near Port au Paix, Ekman 3931 (S).

CUBA: Oriente: Bayate, between Rio Bayate and Arroyo Bibano,
Ekman 9003 (S), 9622 (S).

A new species readily distinct by the fusiform bicolorus fruit from others
of the Antilles with sessile flowers and fruits.

This species is named in honor of its collector, E. C. Leonard of the
Smithsonian Institution in Washington, D. C. who has a continuing inter-
est in the flora of Hispaniola.

The two collections made by Ekman in the Oriente province of Cuba
are both sterile adventitious shoots but they can be referred to this species

420 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

without question on the venation and glandular characters of the leaf
blades.
25. Coccoloba Swartzii Meisner, DC. Prod. 14: 159. 1856.

Coccoloba neglecta oe & Rendle, Jour. Bot. 51: 124. 1913.

Uvifera Swartsii O, . Rev. Gen. 2: 562. 1891.

Tree 8-20 m. tall; eee terete, glabrous, nodes slightly tumid;
ochreae 10-12 mm. lone: basal portion 3-5 mm. long, coriaceous, pe cit.
upper portion 5-7 mm. long, membranaceous, deciduous; jeat blades
ovate to elliptic, apex acute, base narrowed, rounded or slightly cordate,
7X 5,11 & 9,15 & 7.5 cm. long and broad, coriaceous, usually turning
black on drying, glabrous, pit-like depressions on the upper surface, small
glands on the lower surface, midrib and veins inconspicuous or flat above,
prominent below, primary veins 6 or 7 pairs, arcuate anastomosing, sec-
ondary venation conspicuous, reticulate; petioles 10-18 mm. long, gla-
brous, attached at the base of the ochreae; inflorescence terminal, 10-15
cm. long, rhachis glabrous or with glandular exudate, rarely papillose,
staminate flowers in clusters of 3-5 with tightly concentric ochreolae
forming a truncate cylinder after the flowers have fallen, pistillate flowers
solitary, ochreolae erect in flower, flattened against the rhachis in fruit,
bracts ovate, 1-1.5 mm. long, ochreolae membranaceous 1—1.5 mm. long,
flowering pedicels shorter than the ochreolae, hypanthium 0.5 mm. long,
perianth lobes 1-1.5 mm. long, fertile stamens with filaments 1 mm. long;
fruit ovoid, 8-10 mm. long, 6 mm. thick, perianth lobes 1-1.5 mm. long,
coronate.

DISTRIBUTION: Jamaica, Hispaniola, Puerto Rico, Virgin Islands, St.
Croix, Antigua, Montserrat, Guadeloupe, Martinique, St. Lucia, Barbados.
25a. Coccoloba Swartzii forma pubescens, forma nova.

A speciei ramis junioribus, petiolis, laminis ad basin, ochreis et rhachi
inflorescentiae saltem at basin puberulis vel pilosis differt.

ANTIGUA:: Sugar Loat Mt Box 1542 CUS); es ee Blubber
Valley, Box 1411 (type, US), Deize Valley, Box 1184

BARBUDA: Martello Tower, John C. Beard os "a. ae Codring-
ton Village, Fatrchild 3830 (A, US), pe ag es)

GUADELOUPE: De Ponthieu 86 (F

CUBA: Oriente: Punta Padre ae 224 (NY). Local name:
uvillon (Cuba).

The significance of pubescence in the genus Coccoloba has not received
any attention by the earlier monographers. I can do little in interpreting
the variations in pubescence with the material on hand but it is hoped
within the course of this study to have large collections from single plants
and uniform populations to determine what variations are to be expected
within one plant; within the species and to determine what factors cause
pubescence. The specimens cited above differ from the species only in
the presence of a pubescence which varies from minute papillae of epi-
dermal cells to hairs many times the diameter of the epidermal cell. The

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 421

pubescence may be on all parts or only on a few. A long inflorescence
may have a pilose pubescence at the base and only papillae near the apex
of the rhachis. For the present it seems well to designate these variations
as forms until the species can receive more adequate treatment in the
field and perhaps cytological study.

26. Coccoloba Cowellii Britton, Bull. Torrey Bot. Club 42: 368. 1915,
(as Coccolobis).

Shrub to 3 m. tall; branches terete, puberulent when young becoming
glabrate, commonly geniculate, nodes commonly tumid, ochreae sub-
coriaceous, 5—7 mm. tall, dark brown, puberulent; leaf blades ovate-elliptic
to ovate-lanceolate, apex acuminate to rounded, base cordate, 9 & 6 cm.
long and wide on normal shoots, coriaceous, strongly bullate, midrib
and primary veins impressed above, prominent below, primary veins 5—7
pairs, arcuate anastomosing, upper leaf surface smooth, shining, lower
leaf surface sparsely puberulent becoming glabrate, secondary venation
obscure; petioles 3-6 mm. long, puberulent; leaves of vigorous shoots or
of adventitious shoots extremely variable in shape usually lanceolate-
ovate, long attenuate at the apex, cordate at the base, to 23 * 2.5 cm.
long and wide; inflorescence terminal 11-23 cm. long, with several con-
centric ochreae at the base, rhachis glabrous, red in color; staminate flow-
ers in clusters of 2—4, pistillate flowers solitary, bracts ovate, less than 0.5
mm. long, ochreolae membranaceous, to 0.5 mm. long, flowering pedicels
5 mm. long, bright red, perianth red, hypanthium 1 mm. long, perianth
lobes 2 mm. long and broad, fertile stamens with filaments to 2 mm. long,
functional pistil 2.5 mm. long; fruiting pedicels 5-10 mm. long, fruit dark
red, ovoid, rounded at the base, attenuate at the apex, to 6 mm. long, 4 mm.
thick, perianth lobes subcoronate, 1-2 mm. long at apex of fruit.

DISTRIBUTION: Cuba.

CUBA: Oriente: Loma del Mucaral, Nagua, S. of Yara, Leon &
Ekman 11031 (G). Camaguey: sav annahs near Camaguey City, Brit-
ton, Britton & Cowell 13151 (tTypE, NY; isotypes, FM, MO, US); Santa-
yana, Leon 15787 (G, NY), 15789 (NY), Eee 19037 (Suk

This is a striking species with its smooth shining bullate leaves and the
handsome red inflorescence with flowers on long pedicels. Specimens col-
lected in flower in April, June and July; collected in fruit in July.

The variation between the normal leaves and the leaves of vigorous and
adventitious shoots is as striking in this species as any in the genus.
More extensive collections of this species are desired.

Die ete [ae Jacq. Enum. Pl. 19. 1760, Hist. Stirp. Amer.
114, pl. 7 63.
Coccoloba ae Meisn.., DC. Prodr. 14: 162. 1857.
ple cubensis O. Ktze. Rev. Gen. 2: 561. fee
Coccoloba floridana Meisn., DC. Prodr. 14: 165. ‘
ae Curtissii Lindau, Engl. Bot. Jahrb. 13: ce 1891,
Uvifera Curtissii O. Ktze. Rev. Gen. 2: 561. 1891

422 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Coccoloba laurtfolia Lindau, Engl. Bot. Jahrb. 13: 158, 1891, and all recent

authors, not Jacquin.

Coccoloba longifolia Schmidt, Fedde Rep. 24: 73. 1927, not Fischer.

Guaiabara laurifolia (Jacq.) House, Am. Midl. Nat. 8: 64. 1922 (as

Guaibara).

Shrub or tree to 7 m. tall; branches terete, often geniculate by limited
growth, glabrous, nodes rarely slightly tumid; ochreae coriaceous in the
lower portion, this persisting, membranaceous and deciduous above, 3-5
mm. long; leaf blades ovate, oval, oblong, elliptic, lanceolate or obovate,
variable on a single shoot, apex rounded, obtuse, acute or acuminate, base
cuneate to rounded or subcordate, 4 & 3,5, 7 * 5.5,8 * 4.5, 12 *& 8
cm. long and wide, coriaceous, often shining above, dull beneath, glabrous,
midrib and primary veins slightly prominent above, secondary venation
reticulate on both surfaces, primary veins 3-7 pairs, arcuate, anastomosing
before reaching the margin, margin entire, commonly slightly recurved;
petioles glabrous, 7-10 mm. long; leaves of adventitious shoots similar in
shape to those of normal growth but larger in size, 17 & 8, 24 < 13,
32 % 12.5 cm. long and wide, on petioles 1-2.5 cm. long; leaves of wind-
swept specimens often much smaller than those of normal shoots, 2 1.3,
3X 2 cm. long and wide; inflorescence terminal, 4.5, 9, 11 to 18 cm. long,
rhachis glabrous; staminate flowers in clusters of 2-5, pistillate flowers
solitary; bracts ovate, less than 0.5 mm. long, 1 mm. broad, glabrous;
ochreolae membranaceous, less than 0.5 mm. long, glabrous, flowering
pedicels 2-4 mm. long, glabrous; hypanthium 1 mm. long, perianth lobes
2 xX 2 to 3 X 1 mm. long and broad, filaments of stamens from male
flowers 1 mm. long; fruiting pedicels 3-4.5 mm. long, fruit globose to
obpyriform, 10 & 7, 12 & 8, 13 & 8 mm. long and thick, perianth lobes
appressed at the apex of the achene.

DIsTRIBUTION: Florida, Bahamas, Greater and Lesser Antilles.

CUBA? Oriente: Santiago and vicinity, ee 1399 (S), Ekman
7/70 (S), Ekman 8938 (FM, S), Ekman 8962 (MO, S), Havard 20 (NY)
Leon & eo 3045 (G, HAB), Clemente 3792 (G, HAB), 3793 (G,
HAB); yate, Monte Oscuro, Ekman 6110 (NY, S), 6247 (MICH, S,
Ss). 4628 S: Bayate at Paso Est ancia, Ekman 6276 (A, S); Monte
de Ocujal, near Central Manate, Leon 15759 (G, NY, HAB), Leon 16796
ea Punta Maisi, Shafer 7929 (FM, NY, US); Preston, = man 3470
CEM. 2S ; Jiguani, fea 15028 (G, MO, S); Gibara to Punta Hicacos,
Shafer ve (NY, US); Punta Piedra, Nipe Bay, Britton, on & Cowell
1Z48/ IN 3 me Leon 12157 (NY), 12486 (NY); Sabanalamar, near
Central Manati, Leon 15737 (NY); Cupey, Ekman 4942 (FM, S), 6300
(S); Palma Sola, Wright 112 (HABA); Sierra de Nipe, near Rio Piloto,
Ekman 5795 (S); Banes, near Puerto Rico, Ekman 6632 (S); Locality un-
specihed, Linden 2047 (BR, K, type of C. cubensis), Wright 3668 (G

e Roy, without number (NY). Camaguey: eo Gloria, Shafer 415
(PM, G, NY, US); Sante Laces, Shafer 952 CEM, aN YS). ¢ Ganado,
Cayo Sabinal, Shafer 898 (NY, US), Pele ot aa Milian, N. of
Caleta Grande, Roig & Cremata 1853 oe 5S. of Santa Fe, leanne. 650
(G, NY, US). Collected in flower: May, July, Nov. Collected in fruit:

——

1949] HOWARD, THE GENUS COCCOLOBA IN CUBA 423

Feb., Mar., Apr., June, July, Aug., Sept., Nov., Dec. Local names in Cuba:
uvilla, uvillo, uverillo, uva de paloma, fruta de paloma.

For many years a species of Coccoloba with sessile flowers and fruit,
that is, the pedicels shorter than the ochreolae, has been passing under the
name of C. diversifolia Jacq. The drawing of C. diversifolia in the original
publications by Jacquin shows a fruiting specimen with pedicels 3—4 times
the length of the ochreolae. It is obvious therefore that material called
by all recent authors C. laurifolia is really C. diversifolia Jacq. Coccoloba
laurifolia Jacq. is based on a flowering specimen collected near Caracas,
Venezuela. Sufficient material from S. America is not available to me at
the present time to determine if this is identical with the W. Indian C.
diversifolia here treated. It may well be, and C. laurifolia Jacq. may
well belong in the synonymy of the present species. In any case C. laurt-
folia Jacq. is a later name than C. diversifolia which must be used for the
specimens under consideration here. All recent authors have overlooked
this misinterpretation and it is unfortunate that such a widespread species
and one commonly used as:a horticultural plant must have a change of
name.

The type drawing of C. diversifolia Jacq. can be matched in any one of
a dozen of the collections cited above. The leaf variation is tremendous
on a single herbarium sheet and on single trees which I have observed in
the field. Young vigorous shoots may have large leaves. Specimens
growing along the sea coast and exposed to strong winds and salt spray
are often much contorted in form and the leaves are much smaller than
protected plants growing a few yards farther inland. In the Bahamas it
was possible to observe the effect of cutting of these trees to clear garden
patches. Recently cut trees developed large, fast-growing adventitious
shoots with tremendous leaves when compared with the undisturbed trees.
Many of these adventitious shoots and leaves match the specimens which
Schmidt referred to C. longifolia Fisch. ex Lindau (Ekman 1399, 4628).

As in many other species of Coccoloba in the Caribbean the fruits of C.
diversifolia will commonly change shape in the process of maturing. For
that reason it is difficult to use the shape of the fruit as diagnostic of a
species. Fully mature fruits must be used in comparison of size and shape
in this species.

Coccoloba cubensis of Meisner was based on a collection made by Lin-
den. JI have seen two specimens of this collection from the Brussels and
Kew Herbaria. Both are similar and the leaves are small in comparison
with the bulk of the material cited above. If this species were considered
from the standpoint of the Cuban material alone, then C. cubensis would
be comparable to the recent material from the Isle of Pines and might be
worthy of varietal status; however, when material from the Bahamas,
Florida, Jamaica, Puerto Rico and Hispaniola is considered, then the
Linden specimens become one end of a long line of variations.

Lindau described the species Coccoloba Curtissii based on two collec-
tions of Curtiss from Florida, one from the Florida Keys and the other

424, JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

from Merritts Island. Lindau distinguished this species on the presence
of small protuberances (teeth) between the stamens in the flowers. Un-
fortunately Lindau did not recognize the unisexual condition of the flowers
in Coccoloba. The specimens of Curtiss cited by Lindau are staminate
and the protuberances which he mentioned can be found in many flowers
of the specimens cited above.

INSTITUTE FOR RESEARCH IN GENERAL PLANT MoRPHOLOGY,
HARVARD UNIVERSITY.

1949] MANNING, THE GENUS CARYA IN MEXICO 425

THE GENUS CARYA IN MEXICO
Wayne E. MANNING

FIVE SPECIES OF Carya have been reported from Mexico, C. idlinoensis
(Wang.) K. Koch, C. myristicaeformis (Michx. f.) Nutt., C. mexicana
Engelm., C. tetraptera Liebmann, and C. Diguetii Dode. The distribu-
tion of the first three is given in Standley, Trees and Shrubs of Mexico
(Contrib. U. S. Nat. Herb. 23: 165-167. 1920), Martinez, Las Plantas
mas utiles que existen en la republica mexicana (pp. 295-297. 1928), and
Martinez, Catalogo de nombres vulgares y cientificos de plantas mexicanas
(pp. 330-333. 1937). In the last article the distribution is given under
the alphabetically arranged common names such as nogal, nogalillo, nuez.

Various collectors during the past twenty years have added information
concerning the distribution of the species. Studies by the writer on the
types of the last two species mentioned above indicate these to be syno-
nyms of the first species. One species has been reduced to varietal rank
and one new species related to C. cordiformis has been proposed.

The writer wishes to thank the curators of the various herbaria for
their generosity in lending specimens; especially Dr. O. Hagerup, Botanisk
Museum, Copenhagen, Dr. H. Humbert, Museum d’Histoire Naturelle,
Botanique, Paris, and Dr. R. Llamas, Instituto de Biologia, Chapultepec,
Mexico.

Following is a key to the Mexican species of Carya:

1. Leaflets 5, each serration with a dense tuft of hairs on one or both sides
of its apex; bud-scales brown, imbricated; fruit not winged at the sutures,
(eavcodnye toy evel dn yehe 1c) gb (6) cone ee eg ee . C. ovata var. mexicana,

1. Leaflets 7-15, the serrations glabrous or ciliate but without special dense

subapical tufts of hairs; bud-scales valvate; fruit winged at the sutures
Ate Leas tcl Oe, CIE eis Maat MIT ea. dite wy oe. dn aba) acd a eee ee ee Pe ere ce 2
Buds bright vellow; leaflets strongly to densely glandular-lepidote
iis the glands vellow; all leaflets sessile; nut slightly flattened,
not mottled, 4-celled in ees half, thin-shelled; seed probably bitter,
the eae PUNE es soars, aa ee eee ere . Palieri.
Buds grayish-hairy ae yellow) or pine leant ence re-
motely lepidote or densely grayish- or brov ee lepidote beneath;
leaflets sessile or stalked; nut terete, eae 2-celled, or if 4-celled
only at extreme base; seed sweet, the endosperm not ruminating... 3
3. Buds and often twigs very densely grayish- or brownish-lepi-
dote, essentially glabrous; leaflets not falcate, usually sessile,
very densely grayish- or brownish-lepidote beneath, usually ap-
pearing lustrous or silvery white; bud-scale scars crowded,
forming a narrow ring; nut very thick-shelled, without lacunae
in the septa; common peduncles of the staminate catkin clusters
well ney eloped, at the base of the terminal leafy growth only....

N

Se ee ee . - myristicaefo) mus.

4.26 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XxX

=

Terminal buds hairy, the lateral brown with scattered ul rlanes
glands, glabrous or hairy; leaflets remotely yellow-lepidote be-
neath, not lustrous, often strongly falcate, the terminal and often
the lateral stalked; bud-scale scars high, the pairs separate, not
forming a ring; nut mostly thin-shelled, with lacunae in the
primary septa and in the secondary internal ribs; peduncles of
the staminate catkin clusters very short, at base of terminal
leafy shoots and also on special leafless shoots from old wood. .
Se tei oes ers ane reps Ae . C, illinoensis.

A list of collections is recorded for each species in order to validate the
distribution. Also recorded are the herbaria wherein the specimen may
be found. When the material is sterile no symbol is given. Whenever
staminate flowers, pistillate flowers, immature fruit, or mature fruit as
well as leaves are present, iadication is made by the abbreviations stam.,
pist., im. fr., or fr. respectively; fr. only indicates that no leaves are
present. The following are the herbaria and their abbreviations as used
in this paper: AA = Arnold Arboretum; CM = Chicago Natural History
Museum; BPI — Bureau of Plant Industry, U. S. D. A.; GH = Gray
Herbarium; MO = Missouri Botanical Garden; Mich — University of
Michigan; NY == New York Botanical Garden; Tenn — University of
Tennessee; Tex == University of Texas; US — U. S. National Herbarium;
USFS = U. S. Forest Service; WEM = the writer’s personal herbarium;
Cop — Botanisk Museum, Copenhagen, Denmark; Mex = Instituto
de Biologia, Chapultepec, Mexico; Paris —= Museum d’Histoire Naturelle,
Botanique, Paris, France
1. Carya illinoensis (Wang.) K. Koch, Dendr. 1: 593. 1869.

Carya pecan (Marsh.) Engl. & Graebn., Notizbl. Bot. Gard. Mus. Berlin,

pp. 9: 19. 1902,
Carya tetraptera Liebmann, Vidensk. Meddel. Naturh. For. Kjgbenh, 1850:
80. 1850

Carya Digna Dode, Bull. Soc. Bot. France 55: 470, 1908.*

Large tree; leaflets 9-17, the lateral ones sessile or short-stalked, acute
at the base, typically narrow, oblong-lanceolate, falcate, but sometimes
broader, ovate, not falcate; terminal leaflet stalked; twigs, rachises, and
lower leaflet surfaces glabrate to densely puberulent; bud-scales valvate;
terminal bud grayish yellow-hairy, the lateral buds brown, glabrate or
occasionally pubescent, with few to many yellow glands, the bud-scales
frequently broken off and exposing the densely silky-pubescent inner
leaves; bud-scale scars broad, the pairs separated, not forming a ring;
clusters of staminate catkins subsessile or short-stalked (stalk usually
1-3 mm. long), located at the base of an elongate leafy new growth and
also in pairs at the base of special short leafless branches from old wood;
fruit brown, elliptical to oval, mostly 4-winged and -angled to the base;
nut elliptical to oval, cylindrical, not angled or very slightly so above,

* See Little, Amer. Mid]. Nat. 29: 501. 1943 for other synonyms and discussion of
name.

1949 MANNING, THE GENUS CARYA IN MEXICO 427

brown, mottled with irregular darker brown markings; husk and nut-shell
thin; true secondary (dorsal) septa essentially absent, the short projec-
tions from the primary wall not usually reaching the outer wall of the nut,
the nut 2-celled in the lower half or somewhat 4-celled at the extreme
base; primary septa and internal secondary ribs (ridges) with lacunae,
the lacunae of the latter not clearly covered with a hard wall.
VERNACULAR NAMES: La nuez chiguita; nogal liso; liso 0 encarcelado;
nogal morado ; nuez encarcelado,; nogal de Cuilpan,; nogalito; pecan.
Nuevo Leon: Monterrey, C. K. Dodge 19 stam. (GH, US); C. H. & M.T.
Muller 65 fr. (AA, CM, Tex, Mex); C. R. Orcutt HOTOm ti ew S) 446215.
Sargent in 1900 (Canby 222) stam. (AA, US).— Monterrey, Sierra de la
Silla, alt. 1700 ft., C. G. Pringle 11177 fr. (CM, GH, MO, NY, US, Mex). —
Monterrey, Remiate, G. Arséne 6164 (Abbon 179) fr. (AA, GH, MO, US).
—Galeana, along bank of stream, alt. 5400 ft, V. H. Chasec( (S37 ams tf.
(BPI, CM, GH, MO, NY). TaMAuLipas: San Jose, Sierra de San Carlos,
H. H. Bartlett 10394 fr. (US).—La Morita, Marmolejo, H. H. Bartlett
10738 fr. (US).—South Victoria, La Jolla Ranch, R. Runyan 1017 stam.,
pist. (AA, Tex, US); R. Runyan & B. C. Tharp 4070 stam., pist. (Tex). —
Jaumave, H. W. Viereck 302 (US); L. H. Dewey 3/31/1903 (MO). — Circa
jes minas de Victoria a Tula, J. L. Berlandier 855-2275 (GH). Sawn Luis
ost: Tamazunchale, M. T. Edwards 937* (CM, MO, Tex).— Alvarez,
C. C. Parry & E. Palmer 835 % pro parte, fr. (GH, MOUS); JALIsco:
Zapotlan, B. P. Reko 4672 stam. (US). Hupatco: Ixmiquilpan, J. N. Rose,
|. H. Painter & J. S. Rose 8945 fr. (CM, NY, US). — Tecozantla, F. Salazar
in 1913, fr. (US, Mex). Guanayuata: Moist soil in stream-bottom
canyon, 5 km, west of Xichu, alt. 1600 m., E. L. Little Jr. 11074 fr. (USFS,
WEM).— Pamillas, Rio Pamillas, 25 km. northeast of San Luis de la Paz,
moist soil in canyon bottom, alt. 1740 m., E. L. Litile Jr. 11090 fr. broadly
winged (USFS, WEM); same general localitvece clined) roar Oo lit
essentially wingless (USFS, WEM), and E. L. Little Jr. 11092 fr. narrowly
winged (USFS, WEM). CoanultLa: Muzquiz, E. Marsh 6 pist. (Tex).
== Moncleva alt. 2000- tte SS. Witte 702 (Mich), OAxAcAy Oaxaca,
at foot of Cerro de San Felipe, open pasture-like area, A. J. Sharp 45915
(Tenn, WEM).— Oaxaca, Monte Alban, J. N. Rose & W. Hough 4657 fr.
(US). — Cuilapa, F. Liebmann 3777 im. fr., old pist. (Cop, CM).— More-
Los?: (marked “Morelia: Parco”): <Arséne 1/1910 (CM). Mexico:
“Mexique, region de Mexico,” M. L. Diguct in 1908, fr. (Cop, Paris).
This species was reported by both Standley (1920) and Martinez
(1928, 1937) as growing in Nuevo Leon, San Luis Potosi, and Hidalgo.
Both authors indicate also possible distribution in Oaxaca, apparently
referring to the uncertainty of the identification of the collections of
Liebmann. In addition to the above states, the species is now definitely
known from Tamaulipas, Jalisco, Guanajuata, Coahuila, Oaxaca, and pos-
sibly from Morelos and Mexico D. F. I. M. Johnston (Jour. Arnold
Arb. 25: 435. 1944) writes as follows about its presence in Coahuila:
“Reported as growing wild in northeastern Coahuila along the bottom-
lands of the Rio San Diego, Rio Rodrigo, and Rio Sabinas, by Pablo

* Specimen incomplete, hence identification uncertain; it might be Carya Palmeri.

428 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Frick, Mexico Forestal 1: 11-14, fig. (1923), and by Angel Roldan,
Mexico Forestal 3: 30-32, fig. (1923). I have been told of pecan-trees
which formerly grew about Muzquiz and Naciemiento. I have seen no
specimens from Coahuila. However, the species is to be expected in
northeastern Coahuila, for pecans have been collected in Val Verde (Devils
River) and Uvalde Counties in adjoining Texas.” The pecan also grows
wild in southeastern United States.

The natural distribution in Mexico is somewhat uncertain, as the trees
have definitely been introduced in some areas (see Martinez 1928), and
few herbarium labels have clear notes on habitats. On Liebmann 3777
it is recorded that the trees grow wild on hillsides, and on Runyan 1011
it is indicated that the trees are abundant on the sides of mountains in
Tamaulipas. On the other hand, the field notes on Dewey’s specimen
indicate that the many large trees at Jaumave, Tamaulipas, have been
introduced from Texas. It is evident that much of the range indicated
above is a natural one, with part of the range represented by cultivated
trees only, and in other areas by a mixture of native and cultivated trees.

The writer has seen the types of both Carya tetraptera Liebm. and
C. Diguetti Dode. The type of the former, Liebmann 3777 (with four
sheets collected at different times) consists of essentially glabrous speci-
mens of typical pecan with 9-11 (7-13?) narrow stalked leaflets. The
immature, strongly 4-winged fruits and the nearly mature fruit indicate
that the fruit and nut are elongate, essentially elliptical or oval. The
sterile isotype at the Chicago Museum of Natural History has hairy,
broader, less falcate leaflets characteristic of certain trees of C. illinoensis.
Dode (1908) described C. Digueti as belonging to the section EUCARYA
and related to C. texana DC., having 6-10 sessile leaflets, and the stami-
nate catkins on new wood. (Little, Amer. Midl. Nat. 29: 502. 1943, has
pointed out that the name C. texana has been used for two or possibly
three different species, in two different sections; the photograph of the
type of C. texana DC. from Geneva seen by the writer merely shows that
the 9-13 leaflets are very hairy and does not indicate the section to which
it belongs, but the number of leaflets suggests Apocarya.) However, the
type specimen, collected by Diguet in 1908 (poorly pressed by the col-
lector), belongs to the section Apocarya. Some of the staminate catkins
are clearly on special short branches from the old wood, the bud-scales
are valvate, and there are lacunae in the primary septa and in the sec-
ondary internal ribs of the nut. The nut, although small (19-21 mm.
long and 13-15 mm. in diameter) and somewhat oval-cylindrical, is a
typical pecan both externally and internally. The nuts from Copenhagen
examined by the writer were so darkened by oil and age that darker
markings could not be observed. Pringle 10167 (cited by Dode as C.
Pringle, Sierra Madre above Monterrey, Mexique, 2500-3000 ft., 1906,
distributed without fruit as C. myristicaeformis Nutt.), stated by Dode as
undoubtedly belonging to this species, and upon which much of his de-
scripuion of the leaves seems to have been based, actually belongs to C.

1640] MANNING, THE GENUS CARYA IN MEXICO 429

Palmeri described below. The disposition of C. Digueti must, however,
be based on the type specimen. (It should be pointed out that the speci-
men of Pringle 1963 referred by Sargent in Sylva 7: 146. 1895, to C.
myristicaeformis, is true C. myristicaeformis. Dode erroneously cited —
by references — Pringle 10167 and Pringle 1963 as the same.)
2. Carya myristicaeformis (Michx. f.) Nutt. ex Ell. Sketch Bot. S. Car.
& Ga. 2: 628. 1824.
Carya myristicaeformis Nutt. Gen. No. Amer. Pl. 2: 222. 1818; nomen
nudum (see discussion of name in Little, 1943).

Small or large tree; leaflets 7-9, rarely 11, all typically sessile or sub-
sessile, usually not falcate; younger parts of twigs, rachises, and lower
leaflet-surfaces densely brown- or gray-glandular-scaly (lepidote), the
scales often touching each other, the lower surface of the leaflets usually
lustrous brown or silvery white, typically glabrate; bud-scales valvate,
brownish; terminal buds about 7 mm. long, broadly ovate, not flattened;
bud-scale scars narrow, crowded, forming a ciliate ring, though this is not
always distinct; staminate catkins with a definite common peduncle
located only at the base of strong new leafy growth; fruit densely brown-
scaly, 4-wing-valved to base, the husk thin; nut oval, cylindrical, medium
brown, mottled, not angled, with very thick shell, the secondary septa so
low that they appear absent, so nut 2-celled except at the extreme base
where it is 4-celled; lacunae in the nut-shell essentially absent.

VERNACULAR NAME: Nutmeg hickory (in the U. S. A.).

Nuevo Leon: Monterrey: C. H. & M. T. Muller 64 tr (GAAS CM,
Mex); C. H. Muller 2655 fr. (BPI, GH); C. G. Pringle 1963 fie CAAS oll,
GH, MO, NY, US).— Santiago, Horsetail Falls, V. H. Chase 7802 fr.
(BPI, CM, GH, MO, Ney. 15 mi. SW. of Galeana)’¢.-2. oe. 7.
Muller 1142 fr. (AA, CM, Mich, NY, US, Mex).— El Cercado, 30 mi. 5.
of Monterrey, C. H. & M. T. Muller 1352 tr. (AA, CM, Mich, Mex).—
Hacienda Vista Hermosa, 35 mi. S. of Monterrey, alt. 2350 it. &. S. White
1629 fr. (GH, Mich).

Martinez (1928) does not seem to report it, but Standley (1920) and
Martinez (1937) report it from Nuevo Leon. It has so far been reported
from no other state. This species also occurs in southeastern United
States.

The field notes for Muller 2655 state: “Large shrub up to 15 ft. or
small tree to 25 ft. Mature trunks with bark scaling like shagbark. A
common constituent of the oak-hickory wood. Canon Diente.”

3. Carya Palmeri™* sp. nov.

Shrub or medium-sized tree up to 1 < 40 feet; larger leaves 23-37 cm.
long, 15-28 cm. wide; leaflets 9-11, rarely 13, lanceolate to oblong- or
ovate-lanceolate to rarely obovate, sessile, the lateral with rather broad

* Foliolis 9-11, rariter 13, sessilibus, subtus dense aureo-glanduloso-lepidotis; gemmis

ri

aureis, perulis valvatis; fructibus 4-alatis, exocarpio tenue; nuce non variegata, sub-
compressa, putamine dissepimentisque tenuissimis, ad medium 4-loculato.

430 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

obtuse or even subcordate bases, occasionally falcate, finely serrate, strongly
yellow-glandular-scaly beneath, the glands thick, leaflets greenish or more
commonly appearing sironely yellowish brown beneath but not truly
lustrous-sericeous as in C. myristicaeformis; uppermost lateral leaflets
9-20 cm. long, 2-6 cm. wide, the terminal leaflet only slightly larger;
twigs, rachises, and lower leaflet-surfaces glabrate or commonly strongly
pubescent; terminal bud 8-11 mm. long, slender, strongly flattened, all
buds bright yellow because of dense yellow slants: glabrate to sliahtly
puberulent at apex, frequently short-stalked, the bud-scales valvate, appar-
ently only one or two pairs; bud-scale. scars rather broad, the pairs
separated, not forming a ring; staminate catkins about 8 cm. long on
new growth at base of elongate leafy shoots or on special short leafless
shoots from old wood, the common peduncle of the cluster short, 4-8 or
even to 12 mm. long; bract of the flower slender, longer than the bracteoles;
fruit frequently with a short stipe, 3-4 cm. long, 2-3 cm. thick, strongly
4-ridged to the very base, usually winged along the sutures to the base,
strongly yellow-scaly, the husk very thin, 0.5—-1 mm. thick, final dehiscence
uncertain; nut gray-brown, not mottled, not angled, only slightly flattened,
about 2.4 cm. thick parallel to the primary partition, the shell very thin
(about 1 mm. thick), cavity for embryo very large, the secondary septa
strongly developed to about the middle of the nut, so nut 4-celled to about
the middle, primary septa and secondary internal ribs with strong brown-
filled lacunae; endosperm ruminating, hence seed probably bitter. This
species belongs to section APOCARYA.

pee Leon: Monterrey, Sierra Madre, alt. 2600 ft., C. G. Pringle 13957

(GH, Mich, US).— Monterrey, Cafion Diente, C. H. Muller 2643 f1
cor (GEL); = Moniertay, Sierra Madre, alt. 2500- 3000 it., C. G. Pringle
10167 stam. (GH, MO, NY, US, Mex, Pa ris). Monterrey. Sierra Madre,
C. G. Pringle 13200 a (TYPE GH; CM, Mex), Sept. 7, 1904, —
aoe de Villa Santiago, Cafhon Guajuco, Rancho Vista Hermosa,
abundant in upper oak-hickory forest, C. H. Muller 2034 tm: tr. (Ak, CM,
Mic h, Mex). — Villa de Santiago, Horsetal Fails, alt. 2500 ft., Mr. & Mrs.
W. C Leavenworth 811 (CM, MO).

The species is named after Mr. Ernest J. Palmer, one of the foremost
students of Carya, a friend of the writer, and for many years collector and
research assistant at the Arnold Arboretum, Harvard University.

Most of the specimens were originally called C. myristicaeformis, because
of the strong yellow color of the lower surface of the leaflets. Carya
Palmeri differs from C. myristicaeformis in its bright yellow buds, the
yellow-brown color of the lower leaflet-surfaces, and especially in the
external and internal structure of the nut. It differs in these same respects
from C. ilinoensis, and in addition in the sessile leaflets. The species is
very closely allied to C. cordiformis in that the buds, sessile leaflets, and
internal structure of the nut are essentially identical. Carya Palmeri
differs in having: (1) the fruit 4-winged-valved to the base instead of
to the middle; (2) 9-13 instead of 7-9 (rarely to 11) leaflets; (3) more
frequently hairy rachises and twigs; (4) strongly yellow fruit and lower

=

1949 ] MANNING, THE GENUS CARYA IN MEXICO 431

leaflet surface; (5) the leaflets more frequently falcate; (6) a stipe to
the fruit; and (7) a longer staminate floral bract.

There are two different groups of specimens representing the species,
one with strikingly yellow or yellowish brown lower surfaces of the leaflets
(Pringle 13957, Pringle 13200, Muller 2034) and one with leaflets some-
what greener beneath (Muller 2643, Leavenworth 811, Pringle 10167).
This second group is closer to C. cordiformis in its leaves, and it suggests
the possibility that C. Palmeri might be a hybrid between C. cordiformis
and either C. illinoensis or C. myristicaeformis. Certain features, how-
ever, are not intermediate. In each group mentioned above the leaflets
and ‘rachises may be glabrate or strongly pubescent, and the leaflets may
be narrow and small, or broader and larger. In Pringle 13200 and Muller
2034 the uppermost lateral leaflets are 9-11 cm. long and 2-3.5 cm. wide,
the leaves up to 23 cm. long and 15-20 cm. wide; in Leavenworth 811
the leaflets are 15-20 cm. long and 5-6 cm. (terminal 6.5 cm.) wide, the
leaves 28-37 cm. long and 22-28 cm. wide.

4. Carya ovata (Mill.) K. Koch var. mexicana (Engelm.) comb. nov.
Carya mexicana Engelm. ex W. B. Hemsley, Biol. Cent. Amer. Bot.3:
162. 1883.

Tree, 15 to 20 meters high; leaflets 5, mostly obovate, sometimes ob-
lanceolate, at least the terminal one eoacle stalked, each serration with
one or two dense subapical tufts of hairs, the serrations slender appressed
to short, divergent and stout; rachis, twigs, and lower surface of leaflets
glabrate to pubescent; buds brown, the aan ones 9-15 mm. long,
the scales overlapping, the outer persistent in winter; bud-scale scars
crowded into a narrow usually hairy ring; staminate catkins at the base of
long leafy shoots, the common peduncle elongate, ae subglobose to oval,
with a comparatively thin husk (3-6.5 mm. or even 1.5—3 mm. anc),
4-valved to the base, brown, not notably ete nor winged; nut
flattened, slightly to strongly 4-angled, whitish to whitish brown; sec-
ondary partition high, sometimes thick, the ends of the primary septa
frequently much thickened, without lacunae.

VERNACULAR NAMES: Nogalillo de San Luis Potosi; nogal motudo, nogal
Ravyado.

Nuevo Leon: Sierra Madre Oriental: Puerto Blanco to Tarey, 15 miles
SW. of Galeana, C. H. & M. T. Muller 1226 fr. (AA, CM, Mich, NY,
Tex, US, Mex). — El rene 30 miles S. of Monterrey, pene 1355
(Az A, CM, Mich, Tex). San Luis Potosi: Alvarez, C. C. Parry & E.
Palmer 535 % oe parte (see nee C. illinoensis for pikes 835 %), type of
C. mexicana at Kew not seen by the writer, but the following three isotypes
seen: If., fr. MO, If. only GH, fr. only AA. — Parry & Palmer 834% (GH).
— E, Palmer 71 fr. (AA, BPI, GH, CM, MO, NY, US, Mex). — Goodman
1910, 1916, fr. only (AA). Tamautipas: San Jose, H. H. Bartlett 10253
oa GH, US).— Cerro Barril, Bartlett 10488 (US): Bartlett 10490 fr.
CUS: 6. Zamora, El Milagro, Bartlett 11080 fr. (CM, GH, US). —
L Sie Ranch, So. Victoria, R. Runyan 1019 stam. (AA, US). coe
Pinal de eles (Apmatles: )s2. W. Nelson & E, A. Goldman 3934 (GH

432 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

US). Hupatco: near Chapatla, below Alumbres, not far from Zacualtipan,
alt. 6500 ft., A. J. Sharp 46207 (Tenn, WEM). Puesto: slopes above Rio
Necaco, toward Huauchinango, alt. 4700 ft., A. J. Sharp 45377 (Tenn,

EM).— “Parco,”. Nicolas in 1909 (CM).

Both Martinez (1937) and Standley (1920) report the Mexican shag-
bark from San Luis Potosi and Queretaro. In addition to these states,
collections have been made from Nuevo Leon, Tamaulipas, Pueblo, and
Hidalgo. Carya ovata, called shag-bark or shell-bark hickory occurs in
the eastern United States, extending from southern Maine to as far south
as Texas.

This species was separated from C. ovata by Engelmann by its smaller
buds, its more pubescent leaves, and its closer, more appressed serratures
on the leaflets. The buds on the type, it is true, are somewhat shorter
than those found in characteristic C. ovata, but they may represent axillary
buds on fruiting twigs or terminal buds at the tips of short twigs. As
pointed out by Rehder (1935), the buds on some specimens are 15 mm.
long. The serrations are appressed on the type, but on many other speci-
mens the serrations are short, divergent and broad as characteristic in
C. ovata; furthermore some specimens of C. ovata from the northwestern
portion of its range in the United States have slender appressed serrations.

Rehder (1935) pointed out that he saw no character distinguishing C.
mexicana from C. ovata. According to the view of the writer the Mexican
trees are difficult to separate even as a variety, but on certain trees the
nut shell or the partitions are quite thick, the nut shell is often only
slightly angled, and the husk may be quite thin (not on the type). Con-
sequently in the aggregate these trees, all certainly very closely related,
may be considered a geographical variety. Many of these features may,
however, be found separately on individual trees of characteristic C. ovata
in the United States, as far north as Massachusetts.

Rehder’s C. mexicana forma polyneura should be transferred to this
new variety.
4a. Carya ovata (Mill.) K. Koch var. mexicana (Engelmann) Manning

forma polyneura (Rehder) comb. nov.

Carya mexicana Engelm. forma polyneura Rehder, Jour, Arnold Arb. 16:

8: 1935:

This form (C. H. & M. T. Muller 1226, type at AA; and Muller 1355)
appears different at first glance, but is probably a mere ecological form,
possibly a variation occurring on certain parts of a tree. In the leaflets
the veins are closer and more conspicuous. More field study is necessary.
It is interesting that in its glabrous twigs and leaflets and dark buds this
form approaches C. carolinae-septentrionalis (Ashe) Engl. & Graebn.,
the weak North Carolina segregate of C. ovata.

BUCKNELL UNIVERSITY,
LEwisBurRG, PENNSYLVANIA.

1949] COPELAND, NOTES ON FERNS OF FIJI 433

FURTHER NOTES ON THE FERNS OF FIJI

EDWIN B. CoPELAND

THE IMPETUS for the preparation of these notes comes from a study of
a collection of 377 numbers of ferns collected in Fiji by Dr. A. C. Smith
in 1947.1. My earlier publication on Fijian Ferns * contains numerous
binomials which, in my present opinion,® are outmoded; the sequence
followed in my Bulletin of 1929 is also outmoded, but for convenience of
reference it is followed in the present report. In this treatment I insert
brief notes on most of the species which have been reported from Fiji since
my earlier report, and I also indicate the correct binomial for each species
there discussed, in cases where such a change is needed. Five new species
are here described, and several combinations are proposed as new. The
first set of Dr. Smith’s ferns is deposited in the Gray Herbarium and an
essentially complete set, including types of the eae here described,
is in the herbarium of the University of Californ
Botrychium daucifolium Wall. is reported as ee in Fiji by
Christensen in Bishop Mus. Bull. 177: 7. 1943. I have seen no
specimens in support of this record, but such an occurrence is to be

Angiopteris opaca Copel., previously thought to be a Fijian endemic, is
now reported from Samoa by Christensen in Bishop Mus. Bull. 177:
8. 1943.

TRICHOMANES OMPHALODES (Vieill.) C. Chr. = Microgonium ompha-
lodes Vieill.

TRICHOMANES VITIENSE Bake
Copel., ranging from Queensland to Samoa.

TRICHOMANES BIMARGINATUM v. d. B. = Microgonium bimarginatum
v. d

TRICHOMANES SAXIFRAGOIDES Pres] —= Gonocormus minutus (Blume)
v. d. B

TRICHOMANES ERECTUM Brack. and T. ALTERNANS Carr. == Crepidophyl-
lum Endlicherianum (Pres!) Reed in Am. Fern Jour. 38: 89. 1948.

TRICHOMANES HUMILE Forst. = Crepidophyllum humile (Forst.) Reed,
loc. cit.

1 Dr. Smith’s collections were made under the oe of the Arnold Arboretum of
Harvard University and the ge hn pe Guggenheim Memorial Foundation, with the
aid of grants ae the Penrose Fund of the aoe Philosophical Society and the
Bache Fund of the Sint ere my - Sci

2 Bishop Mus. Bull. 59: 1-105. 1929.

3 Copeland, E. B. Genera Filicum. 1947.

434 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

TRICHOMANES BIPUNCTATUM Poiret == Crepidomanes bipunctatum
(Poiret) Copel.

TRICHOMANES MEIFOLIUM Bory —= Macroglena meifolium (Bory) Copel.

TRICHOMANES APHLEBIOIDES Christ —= Vandenboschia aphlebioides
(Christ) Copel., ranging westward to Sumatra.

TRrICHOMANES MAXIMUM Blume == VWandenboschia maxima (Blume)
Copel.

TrICHOMANES CAUDATUM Brack. == Macroglena caudata (Brack.) Copel.,
ranging from Queensland to Tahiti.

TRICHOMANES Asag-Gravt v. d. B. == Macroglena Asae-Grayi (v. d. B.)
Copel., a species structurally similar to the preceding; it is Tricho-
manes longisetum Brack., and Carr., nec Bory; known also from
Samoa and Tahiti.

TRICHOMANES CARTILAGINEUM Vieill. & Pancher, at least in part, as well
as T. peNTATUM v. d. B. and T. SEEMANNII Carr. = Selenodesmium
dentatum (v. d. B.) Copel., common in Fiji and neighboring archi-
pelagos.

TRICHOMANES Harvevi Carr. == Nesopteris intermedia (v. d. B.) Copel.,
ranging from New Guinea to Polynesia.

TRICHOMANES APUFOLIUM Presl = Callistopteris apiifolia (Presl) Copel.

“TRICHOMANES AUSTRALICUM v. d. B.” was an error in citation in Bishop
Mus. Bull. 59: 27. 1929. The correct name is Cephalomanes Bory-
anum (Kunze) v. d. B.

HYMENOPHYLLUM FLABELLATUM Labill. = Mecodium flabellatum (La-
bill.) Copel.

HyMENOPHYLLUM DILATATUM (Forst.) Sw., which == Mecodium dilatatum
(Forst.) Copel., is endemic in New Zealand. Fijian plants passing
under this name and as H. rormosum Brack. —= Mecodium imbri-
catum (Blume) Copel., ranging from Tahiti to Java.

Mecodium polyanthos (Sw.) Copel., a pantropic species, is represented,
among other Fijian collections, by Parks 20614 and 20769.
HyMENOPHYLLUM AUSTRALE Willd., which == Mecodium australe (Willd.)
Copel., is known only from New Zealand and Tasmania and perhaps
Victoria. Fijian plants so identified represent Mecodium samoense
(Baker) Copel., known also from Samoa and Queensland.
Hymenophyllum affine Brack. looks like a real Hymenophyllum.
HYMENOPHYLLUM MULTIFIDUM (Forst.) Sw., which == Meringium multi-
fidum (Forst.) Copel., is endemic in New Zealand. Its Fijian vicar,

HYMENOPHYLLUM FEEJEENSE Brack., —= Meringium feejeense
(Brack.) Copel.

HyMENOPHYLLUM DENTICULATUM Sw. == Meringium denticulatum
(Sw.) Copel.

Meringium Macgillivrayi (Baker) Copel. in Philip. Jour. Sci. 67: 44.

1949] COPELAND, NOTES ON FERNS OF FIJI 435

1938 (Trichomanes Macgillivrayi Baker; Hymenophyllum Macgilli-
vrayi Copel.), was overlooked in Bishop Mus. Bull. 59 (1929).

To Schizaea, in my treatment of 1929, should be added:

Schizaea (Actinostachys) melanesica Selling in Svensk Bot. Tidsk. 38:
208. 1944

Moturiki: Seeman 793. Vanua Levu: Thakaundrove: Maravu,

near Salt Lake, alt. 0-450 m., Degener & Ordones 14144 (in dense wet for-

ester) Miva thai a was “Mudd he Mts.” [Mathuata Range, on mainland
opposite Mathuata Island], eh. Baped.* summit ridge of Mt. Num-
builoa, east of Lambasa, Ae ai -590 m., Smith 6518 (also observed in dry
forest on upper northwestern slopes).

In his original discussion Selling cites only two specimens, the type from
New Caledonia and Seeman 793 (without locality) from Fiji. The species
is now known to occur from New Caledonia to Tonga, but it is rare at
least in Fiji; Dr. Smith reports seeing it in only the one locality cited
above, during his two collecting trips.

GLEICHENIA OCEANICA Kuhn = Sticherus oceanicus (Kuhn) Ching.

GLEICHENIA BRACKENRIDGE! Fournier (in Ann. Sci. Nat. Bot. V. 18: 269.
1873) = Sticherus Brackenridgei (Fournier) comb. nov.

GLEICHENIA LINEARIS (Burm.) Clarke = Dicranopteris linearis (Burm.)
Underw

GLEICHENIA CAUDATA Copel. (in Bishop Mus. Bull. 59: 9. pl. 2. 1929) —=

Dicranopteris caudata (Copel.) comb. nov.

GLEICHENIA JAPONICA Spreng. == Hicriopteris glauca (Thunb.) Copel.
CyaTHEA Hornei (Baker) Copel. (Alsophila eee Baker) —= Gymno-

sphaera Hornei (Baker) Copel. Gen. Fil. 947.

ALSOPHILA DISSITIFOLIA Baker, judging by the Hee must also be a
species of Gymnosphaera.
Cyathea microlepidota sp. nov.

. gregis C. affinis, stipite 60 cm. alto, deorsum atropurpureo griseo-
Pare paleis albidis anguste amecolatie valde attenuatis 15 mm. longis
sparso, sursum fusco glabrescente cicatricibus linearibus ornato; lamina
1.5 m. alta, ovata, rhachibus inferne stramineis subglabris; pinnis infimis
25-30 cm., medialibus 50 cm. longis, 20 cm. latis, sessilibus; pinnulis
10 cm. longis, 2 cm. latis, subacuminatis, infimis brevi-pedicellatis, profunde
pinnatifidis, segmentis infimis basicopicis modo reductis et interdum liberis,
segmentis sequentibus 5 mm. latis, rotundatis, fere integris, costis superne
setulis inflexis atrocastaneis, nieene curenattlis concoloribus minutis in
setulas dissolutis obsitis, sursum glabrescentibus, costulis fere glabris, venis
furcatis; soris medialibus, globosis, indusiis in segmenta magna persistentia
ruptis.

ANUA Levu: Mathuata: Seanggangga Plateau, in drainage of
Korovuli River, vicinity of Natua, alt. 100-200 m., in patches of forest in
open rolling country, Nov. 25, 1947, Smith 6654 (Type in herb. Univ. Calif.)
(“Mbalambala.’”’ Caudex 2-4 m. high, about 5 cm. diam.; fronds about 7-9

436 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

per plant, about 1.5 m. long, the stipe about 60 cm. long, the pinnae 7-9 pairs. )

Well characterized by the minute, dark, dissected squamulae on the
nether face of the costa; and notable, in the group of indusiate species with
broad segments, for the long stipe and only moderately reduced lowest
pinnae.

Cyathea subsessilis Copel. in Philip. Jour. Sci. 6C: 359-1911.

Christensen (in Bishop Mus. Bull. 177: 25. 1943) has identified Smith
1902, from Mt. Ndikeva, Thakaundrove, Vanua Levu, as this species,
described from Samoa. The collection is indeed not typical C. propinqua
Mett., under which name it was distributed; but neither is it typical C.
subsessilts.

Cyathea plagiostegia Copel., previously considered endemic to Fiji, is
now reported from Samoa by Christensen, who lists several collections
in Bishop Mus. Bull. 177: 28. 1943.

Cyathea truncata (Brack.) Copel. is common in Samoa (cf. Christensen
in Bishop Mus. Bull. 177: 30. 1943).

Cyathea subbullata Copel. in Bishop Mus. Occ. Papers 15: 79. 1939.

Vitt Levu: Mba: Immediate vicinity of Nandarivatu, alt. 800-900 m.,
in dense forest along stream, Smith 5045; hills east of Nandala Creek, about
3 miles south of Nandarivatu, alt. 850-970 m., in dense forest, Smith 5925;
Naitasiri: Northern portion of Ration Plateau, between Mt.
Tomanivi [Mt. Victoria] and Nasonggo, alt. 870-970 m., in dense forest,
Suuth 6145.

Previously known from the type, St. John 18304, from mountains west
of Matawailevu, now in the Province of Naitasiri [formerly Tholo East].
The species is very near to the Tahitian C. decurrens (Hook.) Copel. and
the New Caledonian C. alata (Fournier) Copel.

Lastrea Archboldiae sp. nov.

L. rhizomate brevi-repente, basibusque stipitum paleis fuscis lanceolatis
integris 5 mm. longis vestitis; stipitibus gracilibus, 30 cm. longis, sursum
glabrescentibus; lamina 25 cm. longa, basi truncata 16 cm. lata, apice
acuminata, profunde bipinnatifida, rhachi sparse setifera; pinnis fere omni-
bus breviter pedicellatis, infimis basi angustatis, acuminatis, costis superne
velutinis, inferne setis albis sparsis vestitis; segmentis erecto-patentibus,
maximis 1 cm. longis 3-4 mm. latis, rotundatis, integris, sparse ciliolatis,
herbaceis, viridibus, superne et ad venas et ad laminam sparse setosis;
venis ca. 7- paribus, fere omnibus simplicibus, infimis supra sinus excur-
rentibus; soris medialibus, indusiis parvis setulosis.

Vitr Levu: Mba: Vicinity of Nandarivatu, alt. 750-900 m., Feb.-March,
1941, Degener 14273 (TYPE in herb. Univ. Calif.) (in open forest).

Apparently near to Dryopteris savaiensis (Baker) C. Chr., as described
by Christensen in Bishop Mus. Bull. 177: 82. 1943, but with distinctly
stalked pinnae, and without evident aerophores.

DrYOPTERIS PRENTICEI (Carr.) Kuntze — Lastrea Prenticei Carr.

1949 J COPELAND, NOTES ON FERNS OF FIJI 437

DRYOPTERIS BRACKENRIDGE! (Mett.) Kuntze — Lastrea Brackenridgei
(Mett.) Carr.

Dryopteris Harvevi (Mett.) Kuntze — Lastrea Harveyi (Mett.) Carr.

DRYOPTERIS VISCOSA (J. Sm.) Kuntze = Lastrea viscosa J. Sm.

Christensen (in Bishop Mus. Bull. 177: 82. 1943) cites Smith 1669

from Fiji as Dryopteris pubirachis (Baker) C. Chr. This is probably

correct, but I am still unable to distinguish it satisfactorily from Lastrea

viscosa. Two supposed species of New Guinea, Dryopteris subnigra

Brause and D. oligolepia v. A. v. R., represent the same species or group.

DryopTeris Gorponr (Baker) C. Chr. = Ctenitis Gordoni (Baker)
opel. Gen. Fil. 123, 1947.

DRYOPTERIS SETIGERA (Blume) Kuntze probably —= Lastrea Torresiana
Gaud.) Moore

DRYOPTERIS LEUCOLEPIS (Presl) Maxon = Lastrea leucolepis Presl.
DRYOPTERIS SQUAMIGERA (Hook. & Arn.) Kuntze = Ctenitis squamigera
(Hook. & Arn.) Copel. Gen. Fil. 125. 1947.

DRYOPTERIS TENUIFRONS C. Chr. = Ctenitis tenuifrons (C. Chr.) Copel.
Gen. Fil. 125. 1947.
DRYOPTERIS FIJIENSIS (Hook.) C. Chr. = Ctenitis fijiensis (Hook.)

Copel. Gen. Fil. 124. 1947.
Ctenitis davallioides (Brack.) comb. nov.

Lastrea davalltoides Brack. U. S. Expl. Exped. 16: 202. 1854.

Dryopteris davallioides Kuntze, Rev. Gen. 812. 1891; Copel. in Bishop Mus.
Bull. 59: 44. 1929, 93: 34, pl. 6. 1932; C. Chr. in Bishop Mus. Bull.
177: 97, 1943.

Dryopteris microtricha Copel. in Bishop Mus. Bull. 59: 10, 44. 1929.

Ctenitis microtricha Copel. Gen. Fil. 124, 1947.

Without yet feeling certain, I agree with Christensen (loc. cit.) that my
Dryopteris microtricha is probably a synonym of Lastrea davallioides. I
am also uncertain as to the proper genus for this plant. By its pubescence
it seems to be a Ctenitis, but the anadromic dissection of the frond indi-
cates Rumohra. The best developed known specimen is probably Smith
5188, from the summit of Mt. Tomanivi [Mt. Victoria], Mba, Viti Levu,
alt. 1290-1323 m., an epiphyte in mossy forest. This specimen has fronds
more than a meter long, including the stipe, and hexapinnatifid in the
lower part.

DRYOPTERIS GONGYLODES (Schkuhr) Kuntze = Cyclosorus gongylodes
(Schkuhr) Link, the type of its genus

DRYOPTERIS HAENKEANA (Presl) Kuntze, at least as to its Guam type, is
not distinguishable from Cyclosorus unitus (L.) ae see Wagner
& Grether in Bishop Mus. Occ. Papers 19: 52. 1948.

DRYOPTERIS INVISA Coe Kuntze — ee invisus (Forst.)
Copel. Gen. Fil. 142. 1947.

DRYOPTERIS ARIDA (Don) Kuntze = Cyclosorus aridus (Don) Ching.

438 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

DRYOPTERIS ARBUSCULA (Willd.) Kuntze == Cyclosorus arbuscula
(Willd.) Ching; but the presence of this species in Fiji is doubtful.

DRYOPTERIS NYMPHALIS (Forst.) Copel. —= Cyclosorus; but its distinct-
ness from C. parasiticus (L.) Farwell and C. dentatus (Forst.) Ching
is uncertain.

Cyclosorus Degeneri sp. nov.

C. stipitibus fasciculatis, ad vestigia infima pinnarum 2-8 cm., ad
pinnas normales 15 cm. longis, basi paleis paucis parvis lanceolatis atro-
fuscis praeditis, ubique pubescentibus; vestigiis pinnarum 5—15-paribus,
approximatis, triangularibus, 2-4 mm. longis et latis; lamina usque ad 45
cm. longa et 10 cm. lata, acuminata, basi gradatim angustata, pinnata,
rhachi ubique minute setulosa; pinnis approximatis, sessilibus, medialibus
5 cm. longis 9 mm. latis, subacutis, basi truncatis, vix medio ad costam
lobatis, costis venisque setulosis, facie laminare minutissime setulosa, sub-
coriaceis, lobis 2-3 mm. latis obtusis; venis 4—5-paribus, 2-paribus anas-
tomosantibus; soris medialibus, indusiis persistentibus, setosis.

Vanua Levu: Thakaundrove: Maravu, near Salt Lake, alt. 0-450
m., Jan. 1941, Degener & Ordonez 14209 (rypE in herb. Univ. Cali).an
coconut grove near ocean).

Not évidently nearly related to any other known species. The numerous
vestigial lower pinnae are found on several species, the most similar of
which may be Dryopteris Christophersenii C. Chr., which is exindusiate.

DRYOPTERIS MAGNIFICA Copel. == Cyclosorus magnificus (Copel.) Copel.
Gen. Fil. 143. 1947.

DRYOPTERIS MICROSORA Copel. == Cyclosorus microsorus (Copel.) Copel.
Gen. Fil. 143. 1947.

DRYOPTERIS PENNIGERA (Forst.) C. Chr. == Cyclosorus pennigerus

(Forst.) Copel., which probably does not occur in Fiji. The local
plant of this alliance is Cyclosorus costatus (Brack.) Copel. Gen. Fil.

142. 1947.

DRVYOPTERIS UROPHYLLUS (Wall.) C. Chr. == Cyclosorus urophyllus
(Wall.) Copel. Gen. Fil. 143. 1947.

DRVOPTERIS RUBRINERVIS (Mett.) C. Chr. —= Cyclosorus rubrinervis

(Mett.) Copel. Gen. Fil. 143. 1947.

DRYOPTERIS SIMPLICIFOLIA (J. Sm.) Christ == Cyclosorus simplicifolius
(Jom) (Cope. Gen. Fil. 145: 1947,

Dryopterts CESATIANA C. Chr. = Cyclosorus Becearianus (Cesati)
Copel. Gen. Fil. 142. 1947.

In Tectaria, since my publication of 1929, two additional Fijian species

have been described:

Tectaria Degeneri Copel. in Sargentia 1: 3. 1942; related to 7. Gode firoyi
(Luerss.) Copel. but with remarkably slender pinnules and segments.

Tectaria elegans Copel. loc. cit.; a relative of 7. leuzeana (Gaud.) Copel.

1949 ] COPELAND, NOTES ON FERNS OF FIJI 439

POLYSTICHUM ARISTATUM (Forst.) Presl == Rumohra aristata (Forst.)
Ching.

Polystichum aculeatum (L.) Schott is the collective designation of sev-
eral Fijian species which I do not yet venture to identify more
accurately.

Polystichum pilosum sp. nov.

P. P. obtuso J. Sm. et P. mutico Copel. simile, stipitibus caespitosis 30-40
cm. altis stramineis gracilibus, deorsum paleis stramineo-ferrugineis 1 cm.
longis 2 mm. latis acuminatis inconspicue sublaceris squamulisque laceris
albidis vestitis, sursum paleis piliformibus paucis adspersis; lamina 15-20
cm. longa 7-8 cm. lata, bipinnata, rhachibus pilis et paleis piliformibus
basibus parvis dilatatis dilaceratis pallidis densius vestitis; pinnis in-
ferioribus oppositis subacutis, infimis deflexis; pinnulis oblique rhom-
boideis, 7 mm. longis, 4 mm. latis, rotundatis sed interdum minute
Sormulicns lateribus integris vel interdum minute spinuliferis, herbaceis,
pallide viridibus, utraque facie pallide pilosis; soris parvis superficialibus,
indusiis peltatis, laceris, margines versus perpallidis.

Vanua Levu: Mathuata: Southern slopes of Mt. Numbuiloa, east
of eel alt. 350-500 m., Nov. 3, 1947, Smith 6484 (Type in herb. Univ.
Calif.) (on cliffs in thin forest on rocky slope).

ATHYRIUM ACCEDENS (Blume) Milde = Callipteris prolifera (Lam.)
Bory.

ATHYRIUM JAVANICUM (Blume) Copel. == Diplaziopsis javanica
(Blume) C. Chr.

ASPLENIUM REMOTUM Moore. As the oldest name of this extremely poly-
morphic species, Christensen (in Bishop Mus. Bull. 177: 66. 1943)
has chosen Asplenium marattioides (Brack.) C. Chr. Bracken-
ridge gave it three specific names.

ASPLENIUM ADIANTOIDES (L.) C. Chr. is now known by its most familiar
name, Asplenium faleatum Lam.

ASPLENIUM GIBBEROSUM (Forst.) Mett. == Loxoscaphe gibberosum
(Forst.) Moore. Loxoscaphe foeniculaceum (Hook.) Moore is more
finely dissected.

STENOCHLAENA OLEANDRIFOLIA Brack. == Lomariopsis oleandrifolia
(Brack.) Mett.
STENOCHLAENA BRACKENRIDGE! (Carr.) Underw. —= Lomariopsis Brack-

enridgei Carr.
Orthiopteris Copel. in Bishop Mus. Bull. 59: 14, 66. 1929.
Orthiopteris was described with one important error: the apex of the
stem bears small, dark paleae. The later described genus /thycaulon Copel.
is therefore not distinct. For discussion see Copeland, Gen. Fil. 49, 50.
1947.
SACCOLOMA MOLUCCANUM (Blume) Mett. == Orthiopteris minor (Hook.)

440 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

BALANTIUM STRAMINEUM (Labill.) Diels == Culcita straminea (Labill.)
Maxon.

I now have on hand 18 Fijian collections representing this species, ae
a rather firm, entire or lobed or shallowly lacerate indusium. A spec
distinguished by Brackenridge and Maxon, C. blepharodes Maxon, pe
characterized most definitely by “indusium ample, delicately mer bene,
long ciliate.”

Dennstaedtia intermedia sp. nov.

Fronde teste lectore 2—3 m. alta, stipite metrale; rhachi straminea haud
rubescente, inferne glabra fere inerme, superne sulcata breviter sordide
furfuracea; pinnis ca. 8-paribus fere oppositis, subsessilibus, infimis maxi-
mis 70 cm. vel. ultra longis; pinnulis etiam oppositis, infimis (pinnarum
majorum) plus minus reductis, sequentibus usque ad 25 cm. longis, acumi-
natis, herbaceis, glabris, vix bipinnatis; pinnulis” usque ad 4 cm. longis,
acuminatis; pinnulis'! ala angusta confluentibus, oblongis, incisis, 2 mm.
latis; soris parvis (ca. 0.6 mm. latis).

Vitt Levu: Mba: Valley of Nggaliwana Creek, north of the sawmill
at Navai, alt. 725-850 m., in dense forest, July 21, 1947, Smith 5373 (TYPE
in herb. Univ. Calif.).

This genus was overlooked in my Ferns of Fiji (Bishop Mus. Bull.
59: 1929), although Dicksonia incurvata (miscited by Christensen, in
Bishop Mus. Bull. 177: 33. 1943, as D. involucrata) had been described
from Fiji. This species has since been collected by Degener, no. 14701,
from the vicinity of Nandarivatu, alt. 790-900 m. Christensen (Ind. Fil.
Suppl. 3: 70. 1934, and Dansk Bot. Ark. 9, No. 3: 42. 1937) has reduced
D. incurvata (Baker) C. Chr. to D. glabrata (Ces.) C. Chr., but it seems
to me to be a related but easily distinguishable species.

Dennstaedtia intermedia is probably the plant referred to by Christen-
sen (in Bishop Mus. Bull. 177: 34. 1943), “which Baker named Dicksonia
moluccana var. inermis, but it is neither D. moluccana nor D. scandens.”
It is intermediate between the group of D. glabrata and the body of the
genus, being glabrous (except in the groove of the rachis), but without
the rough, maroon axes of D. glabrata.

SCHIZOLOMA ENSIFOLIUM (Sw.) J. Sm. is certainly no Schizoloma; it was
better placed when first named, as Lindsaea ensifolia Sw.

PTERIS DECUSSATA J. Sm. = Pteris mertensioides Willd.

NOTHOLAENA HIRSUTA (Poiret) Desv. is a Chetlanthes, apparently best to
be known as Cheilanthes javensis Moore

ARTHROPTERIS OBLITERATA (R. Br.) J. Sm., more strictly construed, does
not occur in Fiji. The common local plant is Arthropteris repens
(Brack.) C. Chr. in Bishop Mus. Bull. 177: 48. 1943

Arthropteris Archboldiae Copel. in Sargentia 1: 3. 1942, from Viti Levu,
is an indusiate relative of A. tenella (Forst.) J. Sm., of New Zealand,
Norfolk Island, and Australia.

1949] COPELAND, NOTES ON FERNS OF FIJI 441

LASTREA ARTICULATA Brack. = Arthropteris articulata (Brack.) C. Chr.

OLEANDRA WHITMEEI Baker == Oleandra Sibbaldii Grev.

HuMata GAIMARDIANA (Gaud.) J. Sm., as to the Fijian plant, — Humata
Banksii Alston (in Philip. Jour. Sci. 50: 176. 1933). Typical H.
Gaimardiana, of the East Indies, is H. pectinata (J. Sm.) Desv.

POLYPODIUM ACCEDENS Blume —= Weatherbya accedens (Blume) Copel.
Gen. Fil. 191, pl. 6. 1947.

PoLtypopIuM Browni! Wikstr. = Dictymia Brownii (Wikstr.) Copel.

POLYPODIUM LINGUAEFORME Mett. == Microsorium linguaeforme
(Mett.) Copel.

PoOLYPODIUM PUNCTATUM (L.) Sw. = Microsorium punctatum (L.)
Copel.

PoLypopiuMm Parxsir Copel. = Microsorium Parksii (Copel.) Copel.

Gen. Fil. 196. 1947,
POLYPODIUM NIGRESCENS Blume — Microsorium alternifolium (Willd.)
opel. Gen. Fil. 197. 1947.

PoLyPopIUM SCOLOPENDRIA Burm. == Microsorium Scolopendria
(Burm.) Copel.
POLYPODIUM VITIENSE Baker — Microsorium sylvaticum (Brack.)

Copel. Gen. Fil. 196. 1947. ‘Still another name for this i . Polypodium
polynesicum C. Chr. (in Bishop Mus. Bull. 177: 116. 1943).

Potypopium WILKeEsirt C. Chr. == Microsorium alatum ao Copel.
Gen. Fil. 196. 1947.

CAMPIUM PALUSTRE (Brack.) Copel. = Bolbitis lonchophora (Kunze)
Cachr:

CAMPIUM QuOYANUM (Gaud.) Copel. = Bolbitis quoyana (Gaud.)
Ching

CAMPIUM RIVULARE (Brack.) Copel. == Bolbitis rivularis (Brack.) Ching.
Lomagramma cordipinna Holttum in Gard. Bull. Straits Settlem. 9: 202.
1937.

This is a second species of Lomagramma occurring in Fiji, with simply
pinnate fronds, the sterile pinnae with obliquely cordate pinnae or the
acroscopic side truncate. The type is from Samoa, but Holttum cited
Meebold 16856 from Fiji, and a later collection is Degener 14565, from
the vicinity of Nandarivatu, Mba, Viti Levu.

From the group of Elaphoglossum conforme, Krajina (in Studia Bot.
Cech..1: 63-70. 1938) has described five species from Fiji, as follows:
Elaphogossum vanuaense Krajina, typified by Smith 675 and also rep-

resented by Smith 454, both from Thakaundrove, Vanua Levu.
Elaphoglossum Dominii Krajina, typified by a Horne specimen from
Ovalau; this seems to be another name for E. obtusifolium Brack.
Elaphoglossum ovalauense Krajina, typified by a Brackenridge specimen
from Ovalau distributed as E. obtusifolium.

442 JOURNAL OF THE ARNOLD ARBORETUM LVOL, XXX

Elaphoglossum Milnei Krajina, typified by Milve 318, from Viti Levu
without precise locality, and also represented by Horne 808, from the
province of Namosi, Viti Levu.

Elaphoglossum Imthurnii Krajina, typified by im Thurn 315, from
Nandrau, Province of Nandronga & Navosa [formerly Tholo North],
Viti Levu, and also represented by in Thurn 368, from southeastern
Viti Levu.

HYMENOLEPIS MUCRONATA Fée —= Belvisia mucronata (Fée) Copel. Gen.
Fil. 192. 1947. Smith 6093, from northern portion of Rairaimatuku
Plateau, Naitasiri, Viti Levu, is aberrant, or it may represent a new
species.

POLYPODIUM CONFORME Brack., P. L1GuULATUM Baker, P. HIRTELLOIDES
Copel., and P. Hooxeri Brack. are all species of Grammitis.

POLYPODIUM PURPURASCENS Nad. (Enum. Pl. Indig. Tahiti 26. 1873) —=
Ctenopteris purpurascens (Nad.) comb. nov.

POLYPODIUM BLECHNOIDES (Grev.) Hook., as to all the Fijian specimens
at hand, == Ctenopteris Seemannii (J. Sm.) comb. nov. (Crypto-
sorus Seemannii J. Sm. in Benplandia 9: 262. 1861).

PROSAPTIA CONTIGUA Presl, as to the Fijian specimens only, == Prosaptia
pubipes Copel., an endemic species.

MOoONOGRAMMA PARADOXA (lee) Bedd., as to the Fijian specimens, ==
Vaginularia angustissima (Brack.) Mett.

Antrophyum plantagineum (Cav.) Kaulf. Smith 6279, from the escarp-
ment north of Nandarivatu, Mba, Viti Levu, is a small plant with
short stipes. It may be a distinct species.

UNIVERSITY OF CALIFORNIA,
BERKELEY, CALIFORNIA.

1049 ] YUNCKER, FIJIAN SPECIES OF PEPEROMIA 443

ADDITIONAL NOTES ON THE FIJTAN SPECIES OF
PEPEROMIA

T. G. YUNCKER

SINCE MY PREVIOUS PUBLICATIONS dealing with the genus Peperomia in
Fiji,! new collections have resulted in the discovery of noteworthy ma-
terial. Most of the specimens mentioned in the present paper were collected
in 1947 by Dr. A. C. Smith,” but one new species is based upon a collection
by Mr. B. E. V. Parham, of the Department of Agriculture, Suva. It seems
advisable at this time to propose a revised key to the species known to
occur in Fiji; following this, the recently available material is detailed
and three species and three varieties are described as new. Types of novel-
ties are deposited in the herbarium of the Arnold Arboretum, and duplicates
of Dr. Smith’s collections of Peperomia are in the herbarium of DePauw
University.

KEY TO THE FIJIAN SPECIFS OF PEPEROMIA
Leaves predominantly opposite or verticillate.
s moderately to densely hirtellous............... P. leptostachya.
oe glabrous or at most with only a few hairs
Stems with nodal thickenings: leaves 1.2 cm. wide ro) i we

_

SM Ee eR URGES Se ig 4's wai Log ee em een rem mean ees todosa.
olan without nodal thickenings: leaves mostly more on 25 KS
pete eo ress Ah a a ote ee ema moore Pr hk . pilostigiia.
Leaves eee alternate.
Spikes sympodial (leaf-opposed).
Nall hat tc easly 0b Serena Mane ae Pe cr ra aye [Ache nak eave eee P. flexuosa.
Plantsedensely Minbelloussausac cet ute neces P. witilevuensts.

Spikes axillary and/or terminal.
lants obviously and mostly more or less completely hairy.
Plants mostly 5 cm. or less tall; petioles up to 2-2.5 cm. long.
EVE Gh emia g Arles IE aia at ee Pea EE a Te Oe danbe:
Plants much larger; petioles shorter.
Spikes mostly solitary et in branching axillarv clusters).
Leaves palmately 3-nerved, oval or obovate.

Leaves up to : xX 3.5 cm.; stem densely hairy, the
hairs up to 1 mm. long; spikes up to 7 cm. long;
peduncle 12 mm. lon LBs ae ass P. nandartvatensis

Leaves 1.1 X 1.7 cm.; stem minutely puberulent :
spikes up to 1.4 cm. long; peduncle 3 mm. long.
sib liga Gino oie Dea nee renee arse 3 Ady P. curtispica.

1 Bishop Mus. Bull. 141: 25-47. fig. 8-20. 1936; Bishop Mus. Occ. Pap. 17: 215-220.
fig. 1-3. 1943
2 Dr. Smith’s collections were made under the auspices of the Arnold Arboretum
of Harvard University and th ae Simon Guggenheim Memorial Foundation, with
the aid of grants from the Penrose Fund of the pee Philosophical Society and
the Bache Fund of the ve ee of Scienc

—

VY

JOURNAL OF THE ARNOLD ARBORETUM [VOL, Xxx

Leaves palmately 5-nerved.
Leaf apex acute to narrowly acuminate.
Sten hairsserects 4c pacriaee ees ee
Stenr hairesappressed i643 ssc aan F;
Leaf apex obtuse to mee plants ee
HitSUtewiwudecddeece ee auereacee ts P. cilitfolia.
Spikes in branching axillary clusters, occasionally solitary.
Stems long-villous; inflorescence usually of 3 um-

bellately clustered spikes.......... . naitasiriensis.
Stems hirtellous: inflorescence not umbellate.......
tice, (ec abe Gad. gue hie aiesdie saavtel ae tate e eee ee nandalana

Plants essentially glabrous or at most with only a few hairs
Leaves cordate-ovate; fruit longitudinally ribbed... P. pellucida.
Leaves and fruit not as above
Leaves palmately 3-5-nerved, essentially glabrous
Spikes mostly solitary; leaves not exceeding 4 cm. in
length.

Plants suberect, glabrous; ae briefly attenuate,
mostly more than 2.5 l
sain eae arnnnoeeeenilegs vee Endlicheri var. figiana.
Plants decumbent, spreading, minutely ee ae
leaves not attenuate, less than 1.5 c
nee ea raetorieaaee wa eee ees P. fica.
Spikes ae in branching axillary eee leaves
mostly larger.
Leaves ay less than 2.5 cm. wide and 3 or
more times longer than wide (less in var. micro-
limba)

ee err ere P. lasiostigma.
Leaves up to 4 cm. wide and mostly less than 3
s longer than wide; soses5 P. kandavuana.

tin
Leaves mostly plinerved, glabrous or ube ad hairy.
eaves plinerved within the lowermost 5 n
Leaves narrowly and attenuately eee the
upper surface glabrous; UBS branched
and disposed along the stem....... attenuata.
Leaves acute to acuminate, es fae the
nerves on the upper surface; inflorescence
mostly in the upper leaf axils....... P. flavida.
Leaves with the main lateral nerves branching off the
midrib 5-10 mm. above the base
Stems with many ee leaves oe about

wice as long as wide........... lbertiana.
Stems mostly Re leaves 3- : Gare longer
CHAT WiGlO ec ara tee eee ieShes oan ee P. laevilamba.

Peperomia leptostachya Hooker & Arnott in Bot. Beechey 96. 1832.

Vitr Levu: Mba: Vicinity of Nalotawa, eastern base of Mt.
Range, epiphyte in forest along creek, alt. 550-600 m., Smith 4433; Een
portion of Mt. Evans Range, between Mt. Vatuyanita and Mt. Natondra, on
boulders in crest thickets, alt. 700-900 m., Smith 4296.

1949] YUNCKER, FIJIAN SPECIES OF PEPEROMIA 445

Vanua Levu: Mathuata: Mt. Uluimbau [“The Three Sisters” ],
south of Lambasa, stems pink to’ red, on cliffs and boulders on exposed
summit, alt. 150-369 m., Smith 6605.

This species is widely distributed on the islands of the Pacific area. It
is found most commonly on rocks and cliffs at low to moderate altitudes.

Peperomia vitilevuensis sp. nov.

Herba sat parva epiphytica; caule ad 15 cm. alto, conferte hirtulo;
foliis alternis, ellipticis, ca. 1.5 cm. latis, 1.5—4 cm. longis, apice truncato-
acutis, basi acutis, utrinque puberulis, ciliolatis, palmatim 3-—5-nerviis;
petiolo 3-5 mm. longo, hirtulo; spicis oppositifoliis: pedunculo gracili, 3-5
mm. longo, hirtulo; drupa globosa, stigmate subapicali.

A rather small, epinintic herb; stem up to 15 cm. tall, branching, sub-
erect, 2 mm. thick at the base when dry, densely hirtellous, the internodes
mostly 10-15 mm. long; leaves alternate, elliptic, mostly 1.2—-1.8 cm. wide
> 1.5-4 cm. long, the apex bluntly acute, the base acute, puberulent on
both sides, ciliolate, palmately 3—5-nerved, the nerves branched upward,
moderately dark-glandular dotted, drying membranous, translucent; peti-
ole about 5 mm. long, hirtellous; spikes leaf-opposed, 1 mm. thick * 15-20
mm. long, moderately to loosely flowered; peduncle slender, 3—5 mm. long,
hirtellous; rachis glabrous; bracts round-peltate; drupe globose, about
0.5 mm. long, the apex oblique, the stigma slightly subapical.

Vitt Levu: Mba [formerly Tholo North]: Summit of Mt. Nanggara-
nambuluta [Lomalangi], east of Nandarivatu, epiphyte in dense forest, alt.
1100-1120 m., June 23, 1947, Smith 4862 (TyPE, in Arnold Arb. herbarium).
Ra [formerly Tholo North]: ee from Mt. Namama (east of Nandari-
vatu) toward Mt. Tomanivi [Mt. oe stem and leaves frequently
reddish-mottled, epiphyte in dense forest, alt. 1050-1120 m., Samuth 5711.

The comparatively small, densely iivtelgue plants, and especially the
sympodial spikes distinguish this species. From P. flexuosa, which also has
leaf-opposed spikes, it differs because of its densely hirtellous stems.
Peperomia orbiculimba Yuncker var. mathuataensis var. nov.

Foliis orbicularibus vel elliptico-obovatis, 1-2.5 cm. latis, 1-3 cm.
longis, apice rotundatis, obtusis, basi obtusis vel acutis; petiolo gracili, ad
6 cm. longo (pro more 1—2.5 cm.); pedunculo ad 2 cm. long.

Leaves from orbicular to elliptic-obovate, 1-2.5 cm. wide * 1-3 cm
long, the apex rounded, the base obtuse to acute; petioles slender, mostly
1—2.5 cm. or rarely up to 6 cm. long; peduncle up to 2 cm. lon

Vanua Levu: Mathuata: Southern base of Mathuata Range, north
of Natua, on rocky banks along stream in dense forest, alt. 100-250 m.,
Dec. 4, 1947, Smith 6859 (Type, in Arnold Arb. herbarium) ; southern slopes
of Mt. Numbuiloa, east of Lambasa, leaf-blades reddish purple beneath, on
cliffs in steep open forest, alt. 350-500 m., Smith 6569

There is great variation in the size and shape of the leaves on the speci-
mens included here. The leaves on the type specimen of the species, col-
lected near Levuka on Ovalau, are only about 1 cm. long and orbicular or

446 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

nearly so. Some leaves on the type specimen of this variety resemble those

of the species but most of them are much larger and somewhat obovate

while those on no. 6569 are mostly elliptic-obovate. All agree, however, in

being small plants with villous stems and leaves, and in having long, slender

petioles and peduncles, and somewhat pointed fruits.

Peperomia nandarivatensis Yuncker in Bishop Mus. Occ. Pap. 17: 216,
fig. 1

Vitr Levu: Mba [formerly Nandi]: Upper slopes of Mt. Koromba
[Pickering Peak], on boulders in dense forest, alt. 800-1075 m., Smith 4670.
Peperomia curtispica C. DC. in Jour. Linn. Soc. 39: 166. 1909.

Viti Levu: Mba [formerly Tholo North]: Summit of Mt. Nanggara-
nambuluta [Lomalangi], east of Nandarivatu, epiphyte in dense forest, alt.
1100-1120 m., Smith 4861, hills east of Nandala Creek, about 3 miles south
of Nandarivatu, epiphyte in dense mossy forest on ridge, alt. 850-970 m.,
Snuth 5948. Nandronga & Navosa [formerly Tholo North]: North-
ern portion of Rairaimatuku Plateau, between Nandrau and Nanga, epiphyte
in dense forest, alt. 725-825 m., Smith 5515. Ra [formerly Tholo North]:
Ridge from Mt. Namama (east of Nandarivatu) toward Mt. Tomanivi [Mt.
Victoria], epiphyte in dense forest, alt. 1050-1120 m., Simtth 5710. Naita-
siri [formerly Tholo North]: Northern portion of Rairaimatuku Plateau,
between Mt. Tomanivi [Mt. Victoria] and Nasonggo, nodes deep red, epi-
phyte in dense forest, alt. 870-970 m., Smith 5772

This is a small, decumbent, epiphytic species rather common in densely
forested areas. Considerable variation is shown in the size and shape of
the leaves on individual plants.

Peperomia Parhamii sp. nov.

Herba epiphytica statura modica erecta vel suberecta; caulis conferte
hirtulis; foliis alternis vel infimis oppositis vel verticillatis, ellipticis, ellip-
tico-lanceolatis vel subobovatis, infimis obovatis, apice angustatis acutis
vel infimis obtusis, palmatim 5-nerviis, utrinque hirtulis (ad venas saltem) ;
petiolo 3-10 mm. longo, conferte hirtulo; spicis terminalibus et axillaribus;
pedunculo dissite hirtulo; ovario ovoideo, apice obliquo, stigmate sub-
apicali.

A moderate-sized erect or suberect epiphytic herb; stem succulent, 5
mm. thick when dry, up to 30 cm. or more tall, strongly branched upward,
densely hirtellous, hairs erect, the internodes 1 cm. long above, up to

cm. downward; leaves alternate or those on the lower nodes opposite or
whorled, the upper leaves elliptic, lance-elliptic or subobovate, the lower
whorled leaves obovate, quite variable in size, the upper leaves about 1.2
cm. wide X 2.5 cm. long, the lower leaves up to 2-2.5 cm. wide x 3-4
cm. long, the apex narrowed, acute, or lower obovate leaves scarcely pointed
and obtuse, palmately 5-nerved, hirtellous on both sides, especially along
the nerves, dark above, paler and strongly dark-glandular-dotted beneath,
ciliolate above the middle, drying rather thin, translucent; petioles mostly
3-5 mm. Jong above, up to 1 cm. on the larger lower leaves, densely
hirtellous; spikes axillary and terminal, moderately flowered, 1 mm. thick

1940 ] YUNCKER, FIJIAN SPECIES OF PEPEROMIA 447

* 3 cm. long; peduncle about 5 mm. long, loosely hirtellous; bracts
round-peltate; ovary ovoid, the apex oblique, the stigma subapical; fruit
not present.

Vitt Levu: Namosi: Korombasambasanga Mt., alt. 1050 m., Mar.
24, 1940, B. E. Parham 2187 (rype, in Arnold Arb. een aeiur.

This species resembles P. nandarivatensis to some extent but differs
because of the larger size of the plants, shorter pubescence, leaf shape
and pubescence, shorter peduncles, etc. From P. vitiana it differs in the
size and shape of the leaves and the character of the pubescence.
Peperomia naitasiriensis sp. nov.

Herba epiphytica; ramis e basi decumbente radicante ad 35 cm. vel ultra
sat conferte albo-villosis, pilis ad 1 mm. vel ultra longis erectis; foliis
alternis distichis ellipticis vel elliptico-lanceolatis, apice acutis acuminatis
basi acutis, utrinque dissite villosis, paimatim 5-nerviis; petiolo conferte
villoso; inflorescentia umbellata; spicis 5-10 mm. longis; pedunculo
gracili glabro vel dissite villoso.

An epiphytic herb, branching from the rooting, decumbent base, the
branches simple, virgate, 3 mm. thick at the base when dry, up to 35 cm.
or more long, rather densely white-villous, the hairs erect, up to 1 mm.
or more long, the internodes about 1 cm. long; leaves alternate, distichous,
elliptic or lance-elliptic, 1-1.4 cm. wide * 3-4.5 cm. long, gradually
reduced upward, the apex acute to acuminate, the base acute, loosely
villous on both sides, ciliate, palmately ‘5S-nerved, the lateral nerves
slender, dark above, pale beneath, drying thin and membranous; petiole
2-3 mm. long, densely villous; spikes as yet young, scarcely 1 mm. thick
x 5-10 mm. long, moderately flowered, mostly in umbellate clusters of
three, subtended by a lanceolate bract with an apical tuft of hairs, on
pends axillary, loosely villous stalks about 3 mm. long, or uppermost
spikes may be solitary; peduncle slender, up to 5 mm. long, glabrous or
sparsely villous; bracts round-peltate; ovary turbinate (?), the stigma
apical (?); fruit not present.

Viti Levu: Naitasiri [formerly Tholo North]: Northern portion
of Rairaimatuku Plateau, between Mt. Tomanivi [Mt. Victoria] and
Nasonggo, apparently rare and seen only once, stems and lower leaf-surfaces
reddish-tinged, epiphyte in dense forest, alt. 870-970 m., Sept. 18, 1947,
Siith 6144 (Type, unicate in Arnold Arb. herbarium).

The long-villous stems and leaves and the umbellate inflorescence dis-
tinguish this species from other known Fijian species.

—

Peperomia nandalana Yuncker var. nudipeduncula var. nov.

Foliis ellipticis, obovato-ellipticis rhombeisve, ca. 1.5 cm. latis, 2.5-4
cm. longis, apice acutis vel subobtusis, basi acutis, utrinque sat crispo-
puberulis; spicis 5-10 mm. longis; pedunculo glabro.

Leaves elliptic or elliptic-obovate or rhomboidal, 1.2-1.7 cm. wide x
2.5—4 cm. long, the apex acute or obtusish, the base acute, moderately to
sparingly crisp-puberulent on both sides; spikes 1 mm. thick « 5-10 mm.
long; peduncle glabrous.

448 JOURNAL OF THE ARNOLD ARBORETUM [VOL. Xxx

Vanua Levu: Mathuata: Southern base of Mathuata Range, north
of Natua, on rocky banks along stream in dense lr alt. 100-250 m.,
Dec. 4, 1947, Smith 6862 (type, in Arnold Arb. herbari

This variety resembles the species in many respects a differs because
of the smaller size of the plants, shape and size of the leaves, and glabrous
peduncles. It differs from P. vitiana because of the erect stem hairs, less
pointed leaves and shorter spikes which are commonly multiple on axillary
stalks.

Peperomia lasiostigma C. DC. in Jour. Linn. Soc. 39: 165. 1909.

Vitr Levu: Mba [formerly Tholo North]: Hills between Nggaliwana
and Nandala Creeks, south of Nauwanga, nodes deep red, internodes green
or sometimes reddish-flecked, on trees and boulders in dense forest, alt.
725-850 m., Smith 5810; hills between Nandala and Nukunuku Creeks, along
trail from ‘Nandarivatu toward Lewa, nodes purplish, internodes reddish-
tinged, spikes dull red, epiphyte in dense forest, alt. 750-850 m., Smith 6191 ;
slopes of the escarpment north of Na ndarivatu, nodes itn spikes red-
tinged, on rocks in woods along stream, alt. 550-800 m., Smith 6266 (some
leaves and parts of the stems are minutely hirtellous) ; valley of Nggaliwana
Creek, north of the sawmill at Navai, nodes and spikes red, internodes and
eioles red-tinged, epiphyte in dense forest, alt. 725-850 m., Smith 5347;
western and southern slopes of Mt. Tomanivi [Mt. Weta. branches
reddish, epiphyte in dense forest, alt. 850-1150 m., Smith 5268; summit of
Mt. Tomanivi [Mt. Victoria], branches pale green with faint red longi-
tudinal stripes, epiphyte in dense mossy forest, alt. 1290-1323 m., Smith 5197 ;
western slopes of Mt. Nanggaranambuluta [Lomalangi], east of Nandarivatu,
nodes dull red, internodes, leaves, and spikes green, sometimes reddish-tinged,
epiphyte in dense forest, alt. 850-1000 m., Smith 6306a; same locality, alt.
1000-1100 m., Smith 4896; eastern slopes of Mt. Koroyanitu, Mt. Evans
Range, epiphyte in dense low forest, alt. 950-1050 m., Smith 4145. Mba
[formerly Nandi]: Upper slopes of Mt. Koromba [Pickering Peak], on
rocks in dense forest at crest, alt. 800-1075 m., Smith 4699. Nandron
& Navosa [formerly Tholo North]: Northern portion of Rairaimatuku
Plateau, between Nandrau and Rewasau, nodes swollen, deep red, epiphyte
in dense forest, alt. 725-825 m., Smith 5596. Naitasiri [formerly Tholo
North]: Northern portion of Ratsalmmatalen Plateau, between Mt. ee
[Mt. Victoria] and Nasonggo, nodes, internodes and sometimes lower leaf-
surfaces reddish-streaked, epiphyte in ree forest, alt. 870-970 m., a
5/135. G107,

Peperomia lasiostigma var. carnosa (C. DC.) Yuncker in Bishop Mus.
Bull. 141: 36. 1936.

EvU: Ra [formerly Tholo North]: Ridge from Mt. Namama
(east of ea ee toward Mt. Tomanivi [Mt. Victoria], stems often
deep red, . diameter near the base, epiphyte in dense forest, a
1050- 1120 m m., co 5720. Mba eae Tholo North]: Ridge between
Mt. Nang vardnambuluta [Lomalangi ] and Mt. Namama, east of Nandari-
vatu, stems reddish, epiphyte in dense forest, alt. 1050-1120 m., Smith 4985;
western slopes of Mt. Nanggaranambuluta thomalaec east of Nandarivatu,
nodes dull red, internodes, leaves, and spikes green, sometimes reddish-tinged,

1949] YUNCKER, FIJIAN SPECIES OF PEPEROMIA 449

epiphyte in dense forest, alt. 850-1000 m., Smith 6306; summit of Mt.
Tomanivi [Mt. Victoria], stems and branches deep red, succulent, epiphyte
in dense mossy forest, alt. 1290-1323 m., Smith 5148.

There is some question whether the specimens listed above as variety
carnosa are identical with the original P. carnosa C. DC. or not. The
appear to agree sufficiently well with the original description and a photo-
graph of the type, however, to warrant placing them here, at least
questionably. :

Peperomia lasiostigma var. tomaniviensis var. nov.

Herba divaricatim ramosa; caule ad 40 cm. adscendente; foliis ellipticis
vel inferis elliptico-subobovatis, 1-1.5 cm. latis, 2.5-3 cm. longis, apice
abrupte brevi-acutis; petiolo gracili, ca. 5 mm. longo.

Stems divaricately branched upward, up to 30 cm. tall; leaves elliptic
or lower leaves elliptic-subobovate, 1-1.5 cm. wide X 2.5-3 cm. long, the
apex abruptly short-acute; petioles slender, about 5 mm. long.

Vitt Levu: Mba [formerly Tholo North]: Western and southern
slopes of Mt. Tomanivi [Mt. Victoria], epiphyte in dense forest, alt. 850-
1150 m., July 7, 1947, Smith 5114 (ryrr, Arnold Arb. herbarium) ; hills east
of Nandala Creek, about 3 miles south of Nandarivatu, epiphyte in dense
forest, alt. 850-970 m., Smith 6242.

These specimens resemble P. lasiostigma in being glabrous and having
axillary, compound inflorescences. They differ because of the shape and
smaller size of the leaves and more slender petioles. The stems are also
more erect and more widely branched upward.

Peperomia kandavuana Yuncker in Bishop Mus. Bull. 141: 39, fig. 15.
1936.

Viti Levu: Mba [formerly Tholo North]: Hills between Nggaliwana
and Tumbeindreketi Creeks, east of the sawmill at Navai, internodes pinkish,
nodes purplish red, young spikes reddish, on humus-covered boulders in dense
forest, alt. 725-800 m., Smith 5885. Naitasiri [formerly Tholo North]:
Northern portion of Rairaimatuku Plateau, between Mt. Tomanivi [Mt. Vic-
toria] and Nasonggo, internodes green, nodes deep red, epiphyte in dense
forest, alt. 870-970 m., Smith 5782.

The fruit on the type specimen of this species is well developed with the
stigma appearing to be nearly apical. The fruit on no. 5885 is immature
but the ovaries show an oblique apex with the stigma somewhat subapical.
No other difference has been noted, however, and it is believed that it is
the same as this species.

Dept. oF Botany,
DEePpauw UNIVERSITY.

450 JOURNAL OF THE ARNOLD ARBORETUM bo) Pha, 3.

THE ARNOLD ARBORETUM DURING THE FISCAL YEAR
ENDED JUNE 30, 1949

Horticulture. — The Arnold Arboretum plantings have been kept in
good condition during the past year. Extensive pruning work has been
done and weed trees and shrubs have been removed on the South Street
bank in the juniper and yew collections. Commercial fertilizer was applied
to the collections of magnolias, Japanese maples, elms, euonymus, sorbus,
rhododendrons, and to the shrub collection. Mulching of trees and shrubs
with hops has been continued, and has been very effective in promoting
growth, suppressing weeds and preventing grass fires from killing shrubs
and small trees. The control of grass growth by constant mowing with
the “blitzer” has also reduced fire hazards and improved the appearance
of the grounds. Various weed killers, Solvasol, Ammate, Esteron 44, and
Dow 2-4-5 T have been used in an effort to eliminate poison ivy and
shrubby weeds, and to control weeds in the cultivated areas.

Two of the major projects supported by funds from the Friends of the
Arnold Arboretum are nearly completed. More than three hundred azaleas
and other ericaceous shrubs have been planted in the new border along
the Meadow Road. Several hundred ground cover plants were planted in
adjacent areas.

The vine collection has been moved to the Arborway wall, and the un-
sightly trellis in the shrub collection has been removed. A new tree peony
collection was planted near the Wisteria arbor, and includes thirty varieties
not previously grown in the Arboretum.

The rehabilitation of the Peters Hill area, made possible by contribu-
tions from John Ames, is progressing satisfactorily. More than two
thousand superfluous or decrepit trees have been removed, and the small
shrubs and seedlings were sprayed with weed killers. A soil improvement
program was started to insure an adequate sod cover which would prevent
erosion and excessive invasion of weed trees. The hiring of heavy equip-
ment for much of this work was necessary and the cost has been excessive.
Other methods will be used on the alternate contour strips. We hope to
establish a good grass cover and begin a planting program in 1950.

A number of surplus plants were given to Harvard University and Rad-
cliffe College, and several truck loads of surplus nursery stock were given
to the Massachusetts State Hospital for Crippled Children. More than
60 plants of new varieties originating in the Arnold Arboretum were
distributed to cooperating nurserymen. Plants, and seeds of woody plants
have been received from nurserymen and from botanic gardens.

The Case estates in Weston have been of great value in our nursery
trials, testing plots, and for permanent nurseries of surplus stock. Various
types of horticultural experiments are conducted here. About seven acres

1949 | THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 451

are devoted to a forest tree test plot under the auspices of the Cabot
Foundation. The Weston Garden Club did the major work in maintain-
ing the perennial garden, and distributed the flowers to the Veterans’ and
the Waltham Hospitals. The Division of Landscape Architecture con-
tinues to use some of the Case land for demonstration work.

The bulletin of popular information, ‘“Arnoldia” was continued as usual.
A new Guide, “Through the Arnold Arboretum” was written, and five
thousand copies were printed. Dr. Wyman’s new book, “Shrubs and Vines
for American Gardens” was published by the Macmillan Company in
May 1949. The field class held in the fall attracted thirty-five members,
while = spring class held on Saturday mornings drew about thirty
student

Se hundred apple and cherry hybrids have been grown and are now
in test plots. Several of the new apple hybrids selected for further trial
have been found to be apomictic and breed true from seed. New triploid
forsythia hybrids appear to be very promising. Few of the rose hybrids
have been of value. Work on dwarfing rootstocks is being continued.

Comparative Morphology. — Professor Bailey and his co-workers have
continued their investigations of the comparative morphology of various
dicotyledonous families. Carefully preserved material of Azwstrobaileyva
collected by Mr. L. J. Brass and Mr. S. E. Stephens in Queensland enabled
Professor Bailey and Dr. Swamy to make a thorough study of this phylo-
genetically significant genus. Its secondary phloem is remarkable in being
composed of sieve cells and phloem parenchyma strands, and in having
no companion cells. Its nodal anatomy is of an unusual type resembling
that which occurs in Trimenia and Piptocalyx of the Monimiaceae. Its
summation of vegetative and floral characters indicates that it is closely
related to, if not actually a member of this family. Dr. Swamy has found
two peeealics: representatives of the Chloranthaceae. These plants, to-
gether with Amdorella described earlier in the year, raise the number of
surviving primitive, vesselless, dicotyledonous genera to ten in five families.
It is evident that considerable floral diversification in dicotyledons occurred
before the evolution of vessels in the xylem.

Herbarium. — During the past year 11,282 specimens were
mounted and added to the herbarium. The organized herbarium now
contains 645,770 mounted specimens.

Accessions during the fiscal year numbered 46,111 specimens, of which
15,545 came in exchange, 25,016 through purchase or subsidy, 3429 by
gifts, and 2121 in return for identification. Most of this material comes
from southeastern Asia and the Pacific area, or more precisely from con-
tinental Asia (10,629 specimens), northern Australia (12,295 specimens),
and Polynesia, Micronesia, and Malaysia (10,708 specimens). It very
greatly enriches the superb facilities for botanical research on the Pacific
and Asiatic floras already available at the Arboretum. Among the more

452 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. Xxx

important individual accessions are the 11,928 specimens from northern
Australia made by L. J. Brass on the 1948 Richard Archbold Expedition
to Cape York; 2233 plants of Hosokawa collections in Micronesia received
through Dr. H. L. Li from the National Taiwan University; 1338 photo-
graphs of types and other critical specimens in Formosan herbaria obtained
through the courtesy of Prof. Ying Tsiang; 1200 of Mrs. Clemens’ collec-
tions from Papua received from the Botanical Museum, Berlin-Dahlem;
1447 chiefly classic Malaysian collections received from the Rijksher-
barium, Leiden, and 2115 photographs of South American types received
from the Chicago Museum of Natural History. Some other important
accessions are 2784 Polynesian plants from the Bernice P. Bishop Museum,
Honolulu; 2985 plants of Yunnan from the Fan Memorial Institute,
Peiping; and 650 plants of Szechuan from the University of Chengtu,
Szechuan.

A total of 30,591 specimens were sent out from the herbarium during
the year, of which nearly half (6684 mounted and 6992 unmounted
specimens) consisted of American tropical material transferred to the
Gray Herbarium. In exchange, 2892 specimens were sent to American
and 6932 to foreign institutions.

The Arboretum received thirty-five requests for the loan of its herbarium
material. Of these twenty-six came from a total of 14 American institu-
tions and nine from six foreign botanical establishments. A total of 7031
specimens was involved. For the use of the Arboretum staff 6776 speci-
mens were borrowed from other institutions, constituting seventeen loans
from seven American and twelve loans from seven foreign institutions.

Professor Rehder completed the proof reading for his monumental
“Bibliography of Cultivated Trees and Shrubs,” which was finally printed
and issued in the closing fortnight of the fiscal year. The work represents
about ten years of bibliographic research and will long remain an indis-
pensable tool in horticultural and herbarium taxonomy.

Dr. Merrill completed the final proof reading of his definitive catalogue
of the botanical work of Rafinesque. With his work on this large project
finished, he turned to herbarium studies, and especially to the organization
and identification of large collections recently received from the Philip-
pines. Dr. Johnston worked on West Indian and Central American
Boraginaceae and prepared an account of the family for the “Flora of
Trinidad and Tobago.” Dr. Kobuski has progressed in his studies of the
Theaceae despite very heavy editorial duties. Dr. Perry has organized
the very large collections made by Brass on Cape York and now has readied
for distribution many of the named duplicates now on hand from New
Guinea.

Overshadowing in importance all other activities of the herbarium staff
has been the formulation of final plans for the new building in Cambridge
which is to house the combined herbaria and libraries of the Arnold
Arboretum and the Gray Herbarium. Dr. Johnston, Dr. Kobuski, and
Mrs. Schwarten have been particularly active in this project. Their

1949] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 453

routine work has tended more and more to be directed towards eee
the herbarium and library for the anticipated move to Cambri

Two members of the herbarium staff resigned during the fiscal year.
Dr. Allen left in August 1948 because of illness in her family. Dr. Smith
terminated his appointment in September to assume a curatorial post at
the U. S. National Museum.

The Library. — During the fiscal year ended June 30, 1949, 267 bound
volumes were added to the library, bringing the total up to 47,033; 207
pamphlets were catalogued and filed, bringing that total to 14,355. Five
hundred and thirty catalogue cards were added to the main catalogue;
4,885 to the Gray Herbarium new species cards. A shelf list was made
for the monograph section of the library.

The monographs have been reclassified and rearranged according to the
Dalla Torre & Harms system, thus enabling taxonomists to find the neces-
sary work filed under the same number as the herbarium specimen is
filed in the herbarium.

A high school student has worked in the library after school hours and
on Saturdays, dusting the books and shelves with a vacuum cleaner. Minor
book repairs were made at this time; major ones going to the bindery
(some seventy volumes were restored or repaired by the binder).

The photograph collection has been enriched by about 275 pictures
taken in the Arnold Arboretum by the Horticulturist and his assistant.

Inter-library loans have been about the same as in previous years;
numerous requests for typed descriptions, microfilms and photostats, as
well as prints from our collection of Microfilms of war-lost German
periodicals.

Financial report.— The Arnold Arboretum received during the year
$3,333.33 from the Georgiana Wells Sargent bequest, $63,198.45 from the
Louisa W. Case estate, and $1,379.78 from the Marian R. Case estate.
These funds were added to endowment. Gifts for current use included
$8,969.00 from the Friends of the Arnold Arboretum, $340.00 for the
Rafinesque Publication Fund, $2,000.00 for the Chinese Exploration Fund,
and $100.00 for special travel funds.

The total income for the year was $179,328.41 and total expenditures
were $193,897.28. The apparent deficit was due to the fact that special
grants for horticulture and publications built up in former years were
drawn on heavily during the past year.

454 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXX

Bibliography of the Published Writings of the Staff and Students
July 1, 1948 — June 30, 1949
Battey, I. W. Origin of the angiosperms ; need for a broadened outlook.
Jour. Arnold Arb, 30: 64-70. 1948
——. & Swamy, B. G. L.) Ambor lies trichopoda Baill., a new morpho-
logical tvpe of rhe vesselless dicotyledon. Jour. Arnold Arb. 29: 245—
254, pl. 1-5. 1948.

Firtmorre, R. H. Upside down grafting . .
Horticulture 27: 15, lus. 1949
Jomnston, I. M. Boraginaceae, In:

. something new in propagation.

Maguire, B. et al. Plant explorations

in Guiana and the Kaieteur Plateau— V. Bull. Torrey Bot. Club 75:

562. 1948.

———— e es e ee Jose Island (Gulf of Panama). Sargentia 8:
1-306, 7 pi., 2 figs

—-— ae in ae Pieces XVI. Species chiefly ae: Mexico

and Western United States. Jour. Arnold Arb, 29: 227-241. 1948.
Studies in the Boraginaceae, XVII. A. Garda oe Varronia

in Mexico and Central America. Jour, Arnold Arb. 29: 85-104.

The identity of species proposed by Sessé and Mocifo. 104-110, 1940

— Studies in the Boraginaceae, XVIII. Boraginaceae of the soutiera

West Indies. Jour. Arnold Arb. 30: 111-138. 1949.

Kobusxt, C. E. Studies in the Theaceae, XVII,

3onnetia. Jour. Arnold Arb. 29: 393-413, 1948.

———— Studies in the Theaceae, XVIII. The West Indian species of
Laplacea, Jour. Arnold . 30: 166-186. 1949,

———— Ternstroemiaceae. In: Maguire, B. ef al. Plant explorations in
Guiana and the Kaieteur "Plateau LV. Bull. Torrey Bot. Club 75:
412-414, 1948,

Merritt, EK. D. Adventures in locating validly published but Vantiag: bi-
nonuals, fe Symposium on botanical nomenclature V. oe Ou

049,

Bot. an 4-19, 194

ee Rehder, one of this country’s foremost horticulturists.
ican JAumb S(7)* 10, 11, 1948.

oreword [to Rehder, A. Bibliography of cultivated Trees and
ane si

4

A review of the genus

vord [to Western Minerva, or American annals of knowledge
and ene 49,

Metasequoia, a living relic of a fossil genus

a ssil g . Jour. Roy. Hort. Soc.
73: 211-216, f. 87, 88, map. 1947,

——— A note of Philip Miller’ s binomials. Jour. Arnold Arb. 29; 242-
244, 1948.

——— Unlisted fern names of Alphonso Wood. Amer. Fern Jour. 38:
142-147, 1948.

—— & Perry, L. M.
Arnold Arb. 30: 39-63.

Perry, L. M. Plantae Papuanae Archboldianae, XIX. Jour. Arnold Arb.
30: 139-165. 1949,

RenpER, A. Bibliogr aphy of Cultivated Trees and Shrubs, hardy in the cooler
temperate regions of the northern hemisphere. i-xl, 1-825, 1949,

Plantae Papuanae Archboldianae, XVIII. Jour.
1949,

1949] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 455

Sax, K. The Arnold Arboretum during the fiscal year ended June 30, 1948.
Jour. Arnold Arb, 29: 422-428. 1948.

— Genetics and Agriculture. Bull. Atomic Scient, 5: 143, 146. 1949,

ScuwarteNn, L. Index to American botanical literature. Bull. Torrey Bot.

Club 75: 460-468, 586-596, 674-690. 1948; 69-78, 149-160, 223-234,
1949,

Swamy, B. L. Agamospermy in Spiranthes cernua. Lloydia,10: 149-
162, ip a 1948,
— A contribution to the embryology of the Marcgraviaceae. Amer.
Jour. Bot. 35: 628-633, f. 1-21. 1948.
— A contribution to the life history of Casuarina. Proc. Amer. Acad.
Arts & Sci. 77: 1-32, f. 1-77, tables 1, 2, 1948.
Embryological adie inthe ‘Orehidacese I. Gametophytes.
Amer. Midl. Nat. 41: 184-201, f. 1-39, tables ee. II. Embryogeny.
202-232, f. 1-47, tables 1-3. 1949.
— Further contributions to the morphology of the Degeneriaceae.
Jour. Arnold Arb, 30: 10-28, pl. 1-4, f. 1-80. 1948.
—— On the post-fertilization development of Trillium undulatum. Cel-
lule 52: 7-14, pl. 1-3. 1948.
——— A reinvestigation of the embryo sac of Maianthemum canadense.
Bull. Torrey Bot. Club 76: 17-23, f. 1-21. 9
———— Vascular anatomy of orchid flowers. Bot. Mus. Leafl. 13: 61-95,
f. 1-11. ee
—— & Bartrey, I. W. The morphology and ae of Cercidi-
phyllum. ae Arnold Arb, 30: 187-210, pl. 1, 2, f. 1-8. 1
Wyman, D. Azaleas for northern gardens. Amer. Uae "89(10) :
7-8, 32-34, 47-49, 2 f. 1949.
Better hardy vines. Amer. Nurseryman 88(11): 7-9, 48-51, 2 f.
1948,
—_— Crab apple color ... observe fruit before planting. Horticulture
26: 347, 357, illus, 1948.
The Forsythias. Nat. Hort. Mag. 28: 51-64, illus. 1949.
Lilacs. Arnoldia 9: 13-16, 1 col. pl. 1949.
Little care required. Horticulture 26: 379, 397, I fig. 1948.
New or rare ornamental plants recently distributed to commercial
nurserymen by the Arnold Arboretum. Arnoldia 8: 45-64, pl. 14-19.
1948.

Not more varieties, but better ones. Horticulture 27: 133. 1949.
—____— The oriental crab apples. Gardeners’ Chronicle America 53: 90-
91; 103; 3 fig. 1949.
—__—— Pruning rhododendrons. Arnoldia 8: 37-44, pl. 11-13. 1948.
___ Shrubs and vines for American Gardens. Macmillan, 1949. i-iv,
1-442, illus.
Sources for a few new plants. Arnoldia 9: 1-4, pl. 1. 1949.
Spring 1949. “Arnoldia 9: 9-12, pl. 3. ren
—_——— Syringa prestoniae. Arnoldia 8: 29-36
fps = — The wisterias.. Arnoldia-9: (728g 4, a oe

KARL SAX,
Director

456 JOURNAL OF THE ARNOLD ARBORETUM [VOL. xxx

Staff of the Arnold Arboretum
1948 — 49

Kart Sax, S.D., Professor of Botany and Director.

IvAN Murray Jounston, Ph.D., Associate Professor of Botany and
Associate Director. Supervisor of the Library and Herbarium.
JosrepH Horace Fautt, Ph.D., Professor of Forest Pathology, Emeritus.

JOHN GEorGE JAckK,* Assistant Professor of Dendrology, Emeritus.

ELMER Drew MERRILL, S.D., LL.D., Arnold Professor of Botany, Emer-
itus.

ALFRED REeHpER, + A.M., Associate Professor of Dendrology, Emeritus.

CAROLINE KATHRYN ALLEN, Ph.D., Assistant in the Herbarium.
Irvinc WipM_ER BalLey, S.D., Professor of Plant Anatomy.
BEATRIX FARRAND, L.H.D., Consulting Landscape Gardener.
RICHARD Harorp FIttmore, M.S., Propagator.

CONSTANCE MANSFIELD GILMAN, Business Secretary.

CLARENCE EMMEREN KosuskI, Ph.D., Assistant Curator of the Herbarium.
PauL CuristopH MANGE Lsporr, S.D., Professor of Botany.
SusAN DeLano McKe vey, A.B., Research Assistant.

Lity May Perry, Ph.D., Assistant Botanist.

LAZELLA SCHWARTEN, Librarian.

ALBERT CHARLES SMITH, Ph.D.

FRANS VERDOORN, Ph.D., Research Fellow.

RoBerT Gerow WI.ttaMs, B.S., Superintendent.

DonaLtp Wyman, Ph.D., Horticulturist.

* Deceased, May 20, 1949.
fT Deceased, July 21, 1949.

1940] INDEX 457
INDEX
Acacia §) Axonopus compressus,

Ackama acer 144

Additional Notes on Degeneria vitiensis, 1

Additional Notes on the Fijian Species of
Peperomia, 443

Agave from evar A New, 227

Agave kaibabensis,

— viticoide

2,
Allamanda cathartica, 78
Alsomitra peepee 56
— Schultz
Alsophila ee 435
— Hornei, 4
Aly cieatpuie vaginalis, 76
Anchusa depressa, 108
— incana, mee
— mexicana, 109
Angelonia aan 80
Angiopteris opaca, 43
Angiosperms: Need for a Broadened Out-
look, Origin of the, 64
pees plantagineum, 442
Araliaceae, A New Genus of the,
Archboldiane, XVIII, Plantae ipapianad,

Fanaa XIX, Plantae Papuanae,

39
Arnold Arboretum during the Fiscal Year
Ended June 30, 1949, The, 45
Arthropteris Archboldiae, 440
— articulata, 441
— obliterata, 440

—- inate 439

439
Pesan pe as, 80
pe Lamers 439
Ceenae, The Morphology and Rela-
tionships of, 2

Bartey, I. W. Origin ' ie Angiosperms:
Need for a Broadened Rae 64

—& B . SwAMy. The Morphology
and Relationships of susteobaleys, 2 21
—&—T Morphology and Relat
ships 7" Case iaee 187

Balantium stramineum, 440

Belvisia mucronata, 442

Betchea aglaiaeformis, 146

ae of the Published bora
e Staff nee ee July 1, 1948-

Bolts en 441

ee is, 441

Bonnaya veronicaefolia, 79

Boraginaceae, XVII, Studies in the, 8

Boraginaceae, XVIII. Boraginaceae of
the Southern West Indies, Studies in
he, 111

Borreria laevis, ce

Botanical Nam ; Reference to pre-
1753 iene Validating, 71

Botrychium daucifolium, 433

Bourreria calophylla, 107

-epaulae 107
— succulenta, 12
Brachiaria erucaeformis, 84

Callipteris prolifera, 439
Callistopteris apiifolia, 434
Calyptracordia alba, 122
Campium palustre, 441

— quoyanum, 441

— rivulare, 441

458 JOURNAL OF THE ARNOLD ARBORETUM

Carya in Mexico, pe Genus, 425

426
Cassia aos 76
Cassine excelsa
Celastrus adeno 321
— bodinieri, 2
— salicifolia,
— suaveolens, 29
Spe ae Boryanum, 434
Ceratopetalum arene um, 150
Cercidiphyllum, The Morphology
Relationships of, 187
Cercidiphyllum japonicum, 187
Cerdana alliodora, 115
Cerinthe lanceolata, 107
CHATT ee . Validating Botanical
Yames by Reference to pre-1753 Lit-
ree ree 7
Cheilanthes javensis, 440
Chenopodium ambrosioides, 81
Chrysobalanus Icaco, 76
Closaschima, 168
Coccinea cordifolia, 77
Coccoloba i _ ee The Genus, 388
Coccoloba

and

— pina 416
— caes
— see 406

a leoganientis cordata, 415
— — parvifolia, 411

[VOL. XXX

Coccoloba Leonardii, 419

= valida, 405

— uvi ifera ve aversion 410
— eel 409

— brevipes, 416

— saxicola, 418

N B. eis Notes on
the Ferns of Fiji, 433

Cordia Section Varronia in Mexico and

entral America. Studies in the Bora-

ginaceae, XVII. A., 85

Cordia adnata, 91

— alba, 122

— alliodora, 105, 115

— ambigua, 9

— appendiculata, 87

malaca, 99

— Callococca, 123

1949 ]

Cordia cana, 95

— — interrupta,

— — micrantha, I
— subeanescens, 115
— glabra, 1

— globosa, tie

—— humilis, 98

— grandiflora, 116
— graveolens, 120

INDEX

Cordia oaxacana, 87

— panamensis, 126

— sea 105

— pas

— auctor, 104, 106
— per a! a, 87

— rugosa, 120

— socorre

— sphaecepal, 116

— spicata

— sinescens 103, 106
9

— eeraee! 119
Broadwayi, 119

—— tremula, 122

— ulmifolia lineata, 91

— ~~ Wikstromii, 118

— limicola, 88 — humi a

— saake ee Cryptosorus Seemannii, 442
— linea Ctenitis eee 437
— rete om 120 — fijiensis,

— Lockartii, 124 — Gordoni, 43

7
— ee aries 123 Ctenopteris iene ay 442
— mollis, 99 — Seemannii, 442

460 JOURNAL OF THE ARNOLD ARBORETUM

Cuba, The Genus Coccoloba in, 388
Culcita straminea, 440

cimnilicif

s, 43
— ena

Cyrtococcum oo ee
Pt aaa as xanthandru
sig ria__-vitiensis,

Rae Notes

eee: Further Contributions to

Dryopteris arbuscula, 438
—arida, 437

— Brackenridgei, 437

— Cesatiana, 438

— Chris Piaphement 438
— davallioides, 437

— fijiensis, 437

[VOL. Xxx

Dryopteris gongylodes, 437
ordoni, 437

— seti era, 43

— viscosa, 437

Echium ai 109
— longifolium, 109
Betinodiina « Ren 84
Ehretia acanthophora, 128
— Anacuna, 1
—calophylla, 107

— cuneifolia, 107
—exsucca, 107

— lancifolia, oe

— umbellulata, 3
Blaphoglossum. Domini 441

=< obtusifeliuias 441

— ovalauense, 441

— vanuaense, 441

Eleocharis geniculata, 83

Elephantopus mollis, 78

Embelia rubro- wiclacen: 299

Eragrostis pilosa, 8

Erigeron Hane 78

— pusillus, 78

Rinse Peplus, 8

Ferns of Fiji, Ree Notes on the, 433

Fiji, Further Notes on the Ferns of, 433
ijian Species is Peperomia, Additional

443

and Introduced Plants,

Fimbristylis milage, 83

Fruta de paloma, 423

Further renee to the Morphology
of the Degeneriaceae,

Further Notes on the Ferns of Fiji, 433

1949]

eee Pa 435

— nica, 43

Gonocormus minutus, 433

Gordonia eae aes 183

— villosa,

anaes WiLt1aAM. Notes on Some
Fijian Weeds and tenon Plants,
75

Guaiabara pee 422

Gymneeehaers Hornei, 435
Gynostemma He aratee 56
Haemocharis, 168

— portoricensis, 185

— villosa, 183

— Wrightii, 176

Heliotropium angiospermum, 107, 138
107

1

auca, 435

Hollrungia aurantioides, 44

Howarp, RicHarp A. The Genus Cocco-
loba in Cuba, 388

INDEX 461

Hu, eS The Genus Ilex in China,
233,
nen Banksi 441

441
Hymenolepis mucronata, 442
Hymenophy pes affine, 434

— Macgillivrayi, 435

— multifidum, 434

Hyptis pectinata, 80

Identity of ae Propose

and Moci roy

aceae, XVIL B, 10

Tlex 7 ae The ae 233, 348

llex,

d by Sessé
in the Boragin-
4

i tee ics 349

—subsect. Stigmatophorae, 317
— subsect. Umbelliformes, 306

2
See 313
— aa atae
—_—— nee He 1
—asprella, 269

462 JOURNAL OF THE ARNOLD ARBORETUM

Tlex asprella tapuensis, 271
— atrata, 304

— bioritsensis, 366

— cinerea, 384, 385
— cornuta, 353, 356

—— watleeatay 325

— Gaels, 377

— — paucispinosa, 358
— doniana, 30
— dubia hupehensis, 272, 277, 279
iii a 2
geudomacropoda, 276
— eaiticostata, 293
— chowii, 294

Tlex editicostata ce 294
ee 307, 321,

295
— flaveo- Stee aa 289
— fleuryana, 32

— fortunei, 323

— — aureo- eaneoia: 325
— fragilis kingii, 259

=
nee
PS
WwW
S
Ow

— henryi, 272

— horsfieldii, 328

—howii, 27
— hylonoma, 350
—-— glabra, 351

0)
— — hakkuensis, 291

, 271
—— preepedunculen, 272

[VOL. XXX

1949 | INDEX

Ilex macrocarpa genuina, 272 Ilex eee Bae 313
— longi ee 275 — sasakii
— — reevesae, 274 — serrarta era 266
——tr aieetaaee 272 — — subtilis, 266
— macrophylla, 374 — shweliensis, 303
— macropoda, — sieboldi, 266
— malaccensis, 40 — sterrophylla, 292
— manneiensis, 298 suaveolens, 296
— mariesii, 32 — subficoidea, 384
— megistocarpa, 275 — subtilis, 266
— merrillii, 269 — arn 342, 3
— mertensii formosae, 383 — — brevi aN 338, 343
— microcarpa, 310 — — longipedunculata, 342
— micrococca, 261, 263 — sulcata, 315
— — longifolia, 262. — suzukii, 37
— — pilosa, 263 — szechwanensis, 325
— — polyneura, 263 —— calva, 325
— monopyrena, 3 — — puberula, 325
— montana hupehensis, 272 — taisanensis, 343
— — macropoda, 279 — tarajo, 374
— morii — terago, 374
— nanningensis, 373 — theicarpa, 328
— nepalensis, 307 — triflora, 32
— nilagirica, 377 — horsfieldii, 328
— nuculicava, 385 — — javensis, 328
——auctu lis, 387 — — kanehirai, 332
— — brevipedicellata, 387 — — kurziana, 328
— — glabra, — — lobbiana, 328
— nummularia, 325 — viridis, 319, 326
— old , — tsangil, 3
— opienensis, 259 — tsoii, 281
— ovailis, 31 — tugitakayamensis, 288
— oxyphylla, 269 — umbellulata, 313, 315
— paltoria, 316 — — megalophylla, 315
— parvifolia, 341 — uraiensis, 382
— pedunculosa, 292, 334 — — formosae, 383
— continentalis, 334 ers te
gag AC, — viridis, 319

—— ae ee 34

— polyneura, 263
— — glabra, 265
— pseudogodajam, 263

— pubilimba, 370 Iresine Herbstii, 81

— purpurea, 296, 299 Isachne dispar, 84

— — leveilleana, 334 Ischaemum aristatum, 83

— — oldhami, 299 — ciliare,

— reevesae, 274 —rugosum distachyum, 83

— rhamnifolia, 277 — timorense, 83

— rockil, 336 Isotoma longiflor:

= ie 307, 308, 310, 313 Jack. 1861-1949, ae George, 345

— microcarpa, 310 John George Jack. 1861-1949, 345

464 JOURNAL OF THE ARNOLD ARBORETUM

Ivan M. Studies in the Bora-
XVII. A. Cordia Section Var-

JOHNSTON,

The Identity of Species Proposed

by Sessé and Mocifio, 85

—Studies in the Boracinnceads XVIII
Sais of the Southern West
Indies

Reem corollata, 44
OBUSKI, CLARENCE E. Studies in the
Theaceae XVIII. The West Indian
Species of Laplacea, 166

a nuez chiquita, 427

Lantana aes 99

— cory mbosa

a, 95°
Laplacea, Studies in the Theaceae XVIIL
The West Indian Species of, 166

— portoricensis, 185

— Samuelssonii, 174

— Seanad na

— Urban

—— eas 179

— villosa, 182

— viscosa, 183

Lastrea Archboldiae, 436

— oe
sees 437

— oe 437

— Harveyi, 437

Leschenaultia filiformis, 60
1, Hur-Lin. A New Genus of the Ara-
liaceae, 231
ee eee hydrocharoides, 45
— parvifolium, 45
Lindernia anarailis 79
— diffusa, 79
— ier ee 79
Lindleya
Lindsaea erate 440
Liso 0 Sncareslado: 427

[VOL. XXX

Lithocardium corymbosum humile,
— cylindrostachyum Slate oulune 26
— Lockartii, 12

Lithospermum angustifolium, 109

— obtusiflorum, 109
2igraneenalin: 108, 109
— strictum 9

Lobelia conten: 59

— iene 8

Lomagramma cordipinna, 441
eae Brackenridge, 439
— oleandrifolia, 439

Lourea vespertilionis, 76
Loxoscaphe (eae 439
— gibberosum, 4

Lycianthes impar, 52

— Rechingeri

Macroglena Ase. Grayi, 434

— viridiflora, 110
Mallotonia ee
Manni ING, ioe ae Genus Carya

Mex He
Maite me simplex, 55
Mazus japonicus, 80
McKELvey,: en Derano, A New Agave
from Arizona, 2
Mecodium australe, 434
— dilatatum, 434
— flabellatum, 434

. M. Perry. aa
Papuanae “Archboldianae, ae

Messerschmidia gnaphalodes, 1

— punctata,

Mexico, The Genus Carya in, 425

1949]

Microgonium bimarginatum, 433
, 433

sylvati
Mice chomanes vitiense, 433
Mocino. Studies in oraginaceae,

the Bor
XVII. B., The More) of Species Pro-
posed by Sees and, 1
Momordica Charantia, i
Monochoria hastaefolia, 82
—hastata, 82
Monogramma paradoxa, 442
Myosotis mexicana, 10
Namome, 299
Neoalsomitra, 56
— integrifoliola, 56
— Schultzei, 56
Nesopteris intermedia, 4
New Agave from Arizona, A, 227
New Genus of the cae A, 231
Nogal de Cuilpan, 4
Nogal liso, 427

on suai Fijian Weeds and Intro-
duced Pla
ARG Eee hirsuta, 440
oe

Nutmeg hickory, 4

Nymphoides ees 45
— parvifolium,

Oleandra ae 441

— Whitmeei,

Onosmodium Thurber 110

Oreopanax,

— LOrEtOSan 23

Origin of the Angiosperms: Need for a
Broadened Outlook,

Orthiopteris, 439

— minor

Paltoria,

6
Paltoria ovalis, 316

INDEX 465

Passiflora moluccana, 44

Pecan, 427

Peperomia, Additional Notes on the Fijian
pecies of,

Peperomia curtispica, 446
— kandavu

— asostigina, ee

— osa, 4

wos ee ey 449

— leptostachya

— nate cea se

— nandalana nudipeduncula, 447
— nandarivatensis, 446

— orbiculimba mathuataensis, 445

— vitilevuensis, 445

Perry, Liry M. Plantae Papuanae Arch-
boldianae, XIX, 139

Perry, L. M. & E. D. Merricr.
Papuanae soni ris XVIII, 3

pletrtecas Nu 1,

-— uri

Pilea ioe

alarms ancilanum, 136

— ternat 135

—— poe Archboldianae, XVIII,

ae

oe Papuanae Archboldianae, XIX,

139
Plantago depauperata, 54
— lanigera, 53
— major, 80
— stenophylla, 54
— trichophora, 53
Polygonum uvifera, 392
— Uvifera, 409
Polypodium accedens, 441
— blechnoides, 44
— Brownii, 441
— confor

— hirtelloides, 442
— Hookeri, 442
— seUbAtUD 442

,4

nurniracscense

,» 442
— Scolopendria, 441
— vitiense, ne
— Wilkes
Polystichum ues 439
— aristatu 39
— Seatee 439

466 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXx

Pratia Archboldiana, 59 Sinopanax, 231

Prinos, 258 — formosana, 231

— sect. Ageria, 266 Sisyrinchium micranthum, 81

—sect. Prinoides, 268 =e eeeh 299

— vapiian 269 SM A. C. Additional Notes on
— godajam, 313 eee vitiensis, 1

pea Repti 442 Solanum acuminatissimum, 49

— pubipes, 442 — belense, 50

Pseudehretia umbellulata, 315 — heteracanthum, 48

Psidium littorale, 77 — impar

Pteris decussata, 440 — faree alanereis 48

— mertensioides, 440 — incanum, 48

Pullea Clemensiae, 165 — leptacanthum, 45

— deci iens, | - — multifolium, 50

— Verste , 16 — oligolobum, 46

Rochefortia he 128 — Rechingeri, 52

— cuneata, 128 — arene §1

Rumohra aristata, 439 — Scheffe

Saccoloma niglinceaiinl, 439 Se 48

Sagina echinosperma, 40 — torvoideum, 47

— eee 39 — trichostylum, 47

Salix sp., 81 Sorghum halepense, 83

Salvia potanease Spilanthes acmella, 78

a = RL. John Cae Jack. 1861-1949, Spiraeanthemum idenburgense, 141
— Kajewskii, 140

ae a, — lanceolatum, 142
Schizaea ae. 435 — novoguineense, 142
Schizoloma ensifolium, 440 — Pulleanum, 141
Schizomeria, — reticulatum, 141
— adenophylla, 153 Spiraeopsis.
— Brassii, 157 — aglaiaeformis, 146
Clemensiae, 157 — Brassii,
gorumensis, 153 — celebica, 1
— ilicina, 155 — Clemensiae, 149
katastega, 157 — fulva,
— novoguineensis, 154 — glabrescens, 149
— orthophlebia, 154 — myriantha, 15
— parvifolia, 152 — papuana, 150
— tegens, 157 — pubescens, 148
— Versteeghii, 158 —rufa, 147
— Whitei, 156 Staff ol the Arnold Arboretum 1948-49,
Schleidenia a 135 456
— lagoensis, 134 Stenochlaena Brackenridgei, 439
Schobera angiospermum, 138 — oleandrifolia, 439
Scirpus debilis, 8 Sterculia ampla, 43
— Purshian nus, 3 — Clemensiae, 40
Scoparia dulcis — coggygria, 43
Selenedesuiun i gieae: 434 — Conwentzii, 42
Seringia corollata, 44 — gigantifolia, 43
Sessé and Mocifio. Studies in the Bor — multinervia, 42
cage XVII. B,, Oe Identity of — oncinocarpa, 41
es Proposed by, 1 — porphyroclada, 42
eats barbata, 84 — quadrifida, 43
— geniculata, 84 — Shillinglawii, 42
— pallidifusca, 84 Sticherus Brackenridgei, 435

Sida acuta, 76 — oceanicus, 435

1949 ]

Struchium sparganophorum

o an
B. The Identity of
aed Proposed by Sessé and Mocifio,

Sie in the Boraginaceae, XVIII.
et of the Southern West

d

in as Theaceae XVII _ The

West Indian Bee of Laplacea, 166

Swamy, B Further Genrer
to the See eae of the Degeneriaceae,
10

—and I. W. Batrey. The Morphology
and eee of Austrobaileya, 211

orphology and Relationships
of Gatos 187

Symphitum fruticosum, 107

Symplocos courtoisii, 299

Synedrella nodiflora, 77

Talinum paniculatum, 75

Taraxacum officinale, 78

Tectaria Degeneri, 438

aie ee

— elegans, 4
— Godeffroyi, 438
— leuzeana, 4

Theaceae XVIII The West Indian Species
of Laplacea, eee in the, 166
Themeda SENG , 83
Thunbergia alata, is
Tithonia peter 77
Torenia polygonoi
Be eee 132
133

— subsessilis, 131

INDEX 467

Tournefortia one 108
— syringaefolia, 108,
— trinitatis, 132
— volubilis, 131
Trichomanes alternans, 433
434

— Macgillivrayi, 435

— maximum,

— meifolium, 434

— ompualodes, 433
S, 433

saxifr

— neo 434

— vitiense, 433
Trichosanthes dieniensis, 59
— mafuluensis, 5
— papuana, 58
— Pulleana, 5
Triplostegia repens, 55
Turnera ulmifolia, 77
Ulmarronia corymbosa, 91
Uva, 410
Uva de la caleta, 410

oe 399, ae 412, 413, 423

Uvero, 410

Uvero blanco, 405

Uvero de costa, 418

Uvifera arbor americana, fructu aromatico
98

2
— SiGe 417

468 JOURNAL OF THE ARNOLD ARBORETUM [VOLs xxx

Uvifera Swartzii, 420 Varronia martinicensis, 120
— Wrightii, — mexicana, 98
Uvilla, 416, 423 — oaxacana,
Uvillo, 423 — — appendiculata, 87
Uvillon, — paniculata, 1
Vaginularia angustissima, 442 — polycephala, 118
Validating Botanical Names by Reference — Pringlei, 89
to pre-1753 Literature, 71 — sphaerocephala, 116
Vandenboschia aphlebioides, 434 — spinescens, 10
maxima, 434 — tuberosa, 105
Varronia, 85 —urtl 87
— asperrima, 96 Villarsia hydrocharoides, 45
— bullata, 96 Weatherbya accedens, 441
— clarendonensis, 96 Weinmannia Ledermannii, 160
— crenulata, 104, 106 — novoguineensis, 161
— curassavica, 99, 120 — purpurea,
— dasycephala, 116 — trichophora, 161
dichotoma, 90 — urdanetensis, 160
— ferruginea, 103 — Versteeghii, 162
— glo 6 — ysabelensis, 162
— grandiflora, 116 Wikstroemia, 16
guianensis, 120 — alpestris, 171
—hu — Curtyana, 181
—_— icana, 98 — haematoxylon, 184
jacmeliana, 98 — Macfadyenii, 183
— limicola, —portoricensis, 185
— lineata, 91 — Wrightii, 176
sinagitolia: 88, 106 Yuncker, T. G. Additional Notes on the
— macrostachya, 99, 120 Fijian Sandie: of Peperomia, 443