COMPOSITION, ORIGIN, AND AFFINITIES OF THE
MADAGASCAN VASCULAR FLORA
JEAN-F. LEROY'
AISTIRACT
The Madagascan vascular flora which is composed of over 160 families is exceptionally inter-
esting not only because it is prodigiously rich and diversified, but because it has preserved a
number of relicts. In this review, presented roughly according to the phylogenetic system of
Cronquist, every family ias been tentatively, although shortly, analyzed wih regard to its
composition, origin, and affinity. The hypothesis that the Madagascan flora was a part of the
large Cretaceous Gondwanian flora at the time when Angiosperms originated and that the flora
became greatly impoverished on the African continent due to far-reaching climatic changes
accounts for many present characteristics of the Madagascan flora. In this context the existence
of a Madagascan region sensu lato (Madagascar, Mascarene Islands, eastern and southern Africa),
where many austral genera or taxa of higher taxonomic rank exist-some of which belong to
Magnoliidae and to the most primitive parts of Dilleniidae and Rosidae-becomes understand-
able. Equally understandable become many discontinuities of range and the existence of many
endemic uni- or paucispecific genera. Briefly, the present Madagascan vascular flora probably
resulted from a progressive differentiation of the autochthonous Gondwanian stock and natural
introduction of taxa in time through long-distance dispersal.
In this paper, I do not intend, of course, to present an exhaustive enquiry
concerning the origins and affinities of the Madagascan flora. I think an enor-
mous amount of basic research remains to be donc in order to get an accurate
general view on the matter. And how could we do so, since we mostly cannot
establish a satisfactory connection between taxonomic classification and dis-
tributional patterns, in order to express the true taxogenesis?
Since Perrier de la B�thie and Ilumbert's geographical works on the Mada-
gascan flora many papers and books have been published, but no new concept
significant enough to have important implications has been proposed. Some re-
cent publications of great interest pertaining to this matter have appeared, but
they are not especially devoted to Madagascar. A very limited number of
valuable papers have to be mentioned, insofar as they contribute to Madagascan
phytogeography, such as those by Stearn (1971), Dejardin et al. (1973), Koech-
lin et al. (1974), Wild (1975).
In fact, any attempt at a comprehensive synthesis of this kind at this time
would be premature because we need to know much more with respect to
palaeobotany and phylogeny.
Although no important general conclusion has been reached concerning the
biogeography of the Madagascan plants for some decades, the same does not
hold true for strictly taxonomic knowledge, which has much progressed, thanks
in particular to outstanding research by the late forester and botanist R. Capuron.
We are indebted to him for an exceptional improvement in the analysis of
the woody plants of Madagascar. But even in taxonomy we are far from having
1Directeur, Laboratoire de Phan�rogamie, 16 Rue Buffon, Mus�um National d'Histoire
Naturelle, 75005 Paris, France.
ANN. MIssoum BOT. GARD. 65: 535-589. 1978.
0026-6493/78/0535-0589/$05.65/0


ANNALS OF THE MISSOURI BOTANICAL GARDEN
sufficient information, and, in spite of the efforts by Capuron and many others,
among them the botanists of the Mus�um National d'Histoire Naturelle and
O.R.S.T.O.M., many families, certain of which are the largest such as the
Poaceae, Orchidaceae, Euphorbiaceae, and Leguminosae, have not yet been re-
vised in their entirety.
In 1959 Humbert posed a problem, which was in his opinion a very per-
plexing challenge to botanists and zoologists, namely, where and what have
Madagascan taxa originated from? Have they as a whole been derived from
a Cretaceous stock in place for a long time or are they for the most part new-
comers that arrived there during the Tertiary and later. If the latter is the case,
which seems supported by some evidence, how did seed plants manage to
reach Madagascar from Australia or America or, in particular, from Africa with
which so many links are evident? As viewed by Humbert, who principally limited
his considerations to relations with Africa, plants might have reached Madagascar
either by direct land migration (land bridges) or long-distance dispersal (by
means of sea and wind currents, birds, bats, drift wood, rafts, etc.). In his
paper, Humbert, far from discarding long-distance dispersal, discussed at length
two aspects of the question; but in his opinion it was a secondary process that
was not very effective. One can explain the immigration by sea water of a few
plants in this manner, but these are well-known species which always remain con-
fined to, or next to, sandy and rocky beaches. In particular, Humbert believed
that long-distance dispersal could not explain the presence in Madagascar of
certain genera such as Cycas, Nepenthes, and Ceriops. In the same way, al-
though attaching more importance to wind currents, he did not trust in their
efficiency. In order to support his thesis, he invoked evidence drawn from the
Compositae. Let us take, he wrote, the three genera Helichrysum (100 species
in Madagascar), Vernonia (100 species) and Senecio (60 species) which have
a percentage of endemism close to 100%. In comparison with what we know
about the evolution of Tertiary floras in the Mediterranean basin and East
Africa, we must admit a similar differentiation in Madagascar. Any addition
to the flora by long-distance dispersal would have diminished the percentage of
endemism, which is in contradiction with known data. It is clear that in Hum-
bert's opinion, we have to view the increasing number of species as the result of
a differentiation from the ancestral flora, even though the relevant plants are
readily dispersed.
I think Humbert's concept does not fully account for many fundamental
problems with which we have to deal. Perrier de la B�thie (1948), in contrast,
suggested that differentiation might, at least for a small part, have proceeded
from chance immigration, and he referred, in particular, to the case of Calophyl-
lum, several species of which have differentiated in the Mascarenes and the
Madagascan interior, from possibly only a single widespread species, namely
Calophyllum inophyllum.
One could allege that many of the widespread species that we know, such
as many littoral ones, are generally isolated and rather stable, although poly-
morphic, in every country where they occur. But this objection is hardly de-
fensible for at least three reasons. First, every pioneer species is not necessarily
536
[VOL. 65


LEROY-MADAGASCAR
the potential source of a new "taxogenesis." Second, the pioneer species we know
have not possibly had time enough to differentiate. Third, opportunities in re-
lation with changes of environment may have not occurred.
However that may be with respect to present conditions, it is now generally
admitted that past events dealing with biogeography have to be explained
by the theory of plate tectonics. According to this, the history of the earth is a
continuous process, and continents may be rafted as much as 2,000-3,000 km
apart after having been joined, or, on the contrary, to become joined after a
long separation. So, there is no longer any essential difference between long-
distance dispersal and direct migration by land bridges. One can consider both
of these mechanisms as the two extreme limits of a single process. The different
biota of the Globe are presently massive wholes which have been built up
slowly during geological times as plates approached to, or moved away from
each other. Let us not forget that the opening of the South Atlantic began only
125-130 m.y. BP in the Lower Cretaceous. By the Paleocene, the gap between
South America and Africa was only 800 km and "populated with volcanic
islands" (Raven & Axelrod, 1974). Madagascar separated from Africa about
the mid-Cretaceous and from India at the saine time or possibly later, but
opportunities for exchanges with Africa across the Mozambique sea barrier lasted
a long time and are still presently possible. By mid-Cretaceous time, direct mi-
grations between India-Madagascar-Africa and Australia became more and more
difficult as these lands moved away.
If no essential distinction is made between the notion of long-distance dis-
persal and that of direct migration, then we have much more flexible bases
at our disposal for attempting to understand biogeographical facts.
In Humbert's paper (1959), "Origines pr�sum�es et affinit�s de la flore de
Madagascar" all the Madagascan flowering plants are briefly reviewed in turn.
My present contribution is only a complement aimed to bring the Humbert
compilation up to date, not only, of course, in taking new taxonomic and floristic
data into account, but also in considering them in the light of the plate tectonic
theory, which is now reliable enough to be taken as a foundation of biogeog-
raphy. Further, certain families are emphasized either for distributional patterns
that are especially significant in relation to Madagascan biogeography or for
having been recently revised. Thus, such families or orders as Winteraceae,
Canellaceae, Saxifragales, Theales, Tiliales, Malvales, Guttiferae, Celastraceae,
Malpighiaceae, Meliaceae, Combretaceae, Araliaceae, Verbenaceae, Asteraceae,
and others are worth detailed analysis.
With regard to relationships between neotropical and African-Madagascan
floras, authors often concur in the belief that the species of the same genus or the
members of the same species occurring on both sides of the Atlantic Ocean have
not crossed the present ocean, but have always inhabited the site where they
grow today. So, for 90 m.y. BP some genera or species would not have evolved
at ail. In my opinion, one should not forget that by the Lower Eocene (about
50-55 m.y. BP) the Atlantic probably did not yet form an absolute barrier
for living beings. Dejardin et al. (1973) admitted judiciously, among other
possibilities, that such a species as Rhipsalis baccifera could have reached Africa
537
1978]


ANNALS OF THE MISSOURI BOTANICAL GARDEN
a long time after the original separation of the continents. I agree with Raven
& Axelrod (1974) that one must suppose that "numerous islands afforded ready
migration across the opening Atlantic into the Early Tertiary." Later, difficulties
for crossing became stronger and stronger but transmigration continued to occur.
Although this process appears unlikely, just as evolution itself, yet so must it
have happened.
In terms of the above explanation, Christiana africana is possibly a newcomer
in either Africa-Madagascar or America, and the distributional patterns of fam-
ilies mentioned by Dejardin et al. (1973) such as Bromeliaceae, Rapateaceae,
Humiriaceae, and Vochysiaceae, insofar as they can be comparable must also
belong to the same explanation-and this even though several species of Rhipsalis,
manifestly directly derived from the same ancestor, are in existence in Mada-
gascar.
I think that the obstacle in the way of understanding biogeographical facts
is presently not so much what has happened with oceans and sea barriers or
geological considerations, but our own ignorance about plant taxonomy and
phylogenetics. Statistical figures about composition and endemism percentages
classically used since Perrier de la B�thie (1936) and Humbert's (1959) works
are interesting data, as also are the data provided recently by Dejardin et al.
(1973), among others, but we now need multidisciplinary and biological mono-
graphs-why not by international teams?-of only a few especially chosen fam-
ilies.
Taking into consideration a number of recent important biogeographical works
on the one hand, and, on the other, certain data from my own research, I am led
to some general hypothetical principles which can be outlined as follows:
1. Primitive angiosperms originated in Gondwanaland (possibly in West
Gondwanaland, according to Raven & Axelrod, 1974, with whom I agree), then
extended their range northward long before Gondwanaland broke up, possibly
during the last part of the Lower Cretaceous.
2. As soon as the Upper Cretaceous, two large floras must have been estab-
lished: a Laurasian one (Magnoliaceae, Ranuculaceae, Amentifereae, . . .) and
a Gondwanian one (Annonaceae, Winteraceae, Myristicaceae). The concept
that the angiosperms empire is divided into two great floras, was recently, and
rightfully I think, emphasized by Aubr�ville (1974a).
3. At the close of the Cretaceous and by the Early Palaeocene almost all
modern families were in existence.
Every modern family has evolved sufficiently for adapting to new geologic
or climatic conditions. Some have diversified in many directions, and they have
succeeded in colonizing the most diverse regions on the Globe (Orchidaceae,
Compositae, Poaceae, . . .). Others, even though narrowly specialized, have
kept a combination of both advanced or very advanced characters and primitive
or very primitive characters: they represent the so-called "primitive angiosperms."
4. The biogeographical definition of Madagascar should rest on two con-
cepts:
a. The "Grande Ile" is fundamentally a fragment of Africa; its biota, taken
538
[VOL. 65


LEROY-MADAGASCAR
as a whole, is obviously an African one. Dejardin et al. (1973) said that this
idea is now a classic one, and I agree. According to the estimates of Perrier de la
B�thie (1936), accepted by Humbert, the African element in the Madagascan
flora would be around 27%, the pantropical and Asian elements being respec-
tively 42% and 7%. The austral element contributes 3% and the recent element
whose transport is due to long-distance dispersal, 15%. The remaining 6% repre-
sents the true Madagascan clement. The studies of Dejardin et al. (1973) have
complemented earlier ones. In particular, they recognize the existence of a neo-
tropical element. Besides, they unite the pantropical and the "recent exotic"
elements which Perrier had separated. Further, they add a seventh element,
namely the palaeotropical one, which they subdivide. Thus, as one sees, their
analysis goes much farther than the previous ones.
However, such a recognition of groups of elements and of percentages does
not account for several essential factors.
What precise significance is to be given to these elements which we think
are arbitrarily delimited? Their interrelationships should be analyzed more
thoroughly. Insofar as Madagascar is a part of Africa, a Madagascan element
is ipso facto an African one. Let us for example take the genera Garcinia and
Rheedia, which Dejardin et al. (1973) have classified, respectively, in the
palaeotropical element and neotropical clement. These genera are in fact,
not at ail clear-cut and could be merged into one genus, which would then be a
pantropical one. This case shows that any judgement concerning floristic
affinities if founded on percentages of composition depends greatly upon
taxonomic concepts. Moreover, the mere report of a common occurrence of a
taxon in several areas may not reveal the true situation. In their paper, Dejardin
et al. place Symphonia with neotropical elements, although it is in fact a genuine
Madagascan one. Further, they reduced the Madagascan element to the en-
demic one. So, for Dejardin et al. terms such as "Madagascan element" and
"endemic element" are synonymous. In fact, many so-called pantropical or
palaeotropical or neotropical genera rcally belong to Madagascan stock. Al-
though they are strictly Madagascan elements, they occur also in other areas.
b. Madagascar is a fragment of Africa, but it represents with its neighboring
islands and eastern and southern Africa, a natural entity to which could be
given the name "Madagascan region sensu lato." A number of reasons can be
given to support this concept:
-A number of genera arc common to eastern and southern Africa and Mada-
gascar.
-Madagascar and the eastern coasts of Africa are the western border of
the Indian Ocean.
-Madagascar, eastern and southern Africa were linked together, even though
possibly separated by the sea, at the time when Gondwanaland was in existence.
-The uplift of the central mountain range in Africa during Tertiary times
and the climatic changes in relation to this event increased the barrier between
western and eastern Africa so as to reinforce the isolated position of ail the
eastern territories of which Madagascar is one part.
539
1978]


ANNALS OF THE MISSOURI BOTANICAL GARIEN
It is well known (Lebrun, 1961; Perrier de la B�thie, 1936; Aubr�ville, 1955;
Dejardin et al., 1973) that the African tropical flora must be divided into two
parts, the Guineo-Congolcse one having almost no affinity with the Madagascan
one, and the Sudano-Zambezian flora whose relationships with the Madagascan
one are very important. The strong disjunction is rather between the two African
parts, than between Africa and Madagascar.
I shall now give a summary of the more significant taxa which support the
concept of a close relationship betwcen the floras of Madagascar and eastern
and southern Africa. After that, I shall briefly survey the whole of the Mada-
gascan flora by considering each family, except some unrevised ones, in its
turn.
CLOSE RELATIONSHII'S BETWEEN THE FLORAS OF MADAGASCAR AND
EASTERN AND SOUTHERN AFRICA
1. Families, subfamilies or tribes confined to the Madagascan region sensu
lato and having at least one representative on either side of the Mozambique chan-
nel:
Myrothamnaceae (a monotypic family extending somewhat up to Za�re)
allied to the Hamamelidaceae and, possibly, to Didymelaceae, Hydrostachy-
aceae (we need detailed research on these two austral families, which were
considered as primitive and linked within Pittosporales (suborder Brunineae)
by Thorne, 1975; but the Hydrostachyaceae are placed in the Scrophulariaceae
by Cronquist, 1968), Ptaeroxylaceae, Androstachyaceae.
2. Families or groups of plants with wide distributions represented in the
Madagascan regions sensu lato but not elsewhere in Africa:
Winteraceae (a subfamily in Madagascar), Canellaceae, Hamamelidaceae,
Moringaceae, Cunoniaceae, Brexiaceae, Proteaceae, Myoporaceae, Velloziaceae.
3. Some familles or tribes of wide distribution, including West Africa, but
characteristically of the Madagascan region sensu lato:
Monimiaceae, Montiniaceae, Meliaceae-Turraeeae, Portulacaceae, Palmae-
Borassoideae, Pandanaceae, Poaccae-Bambuscae.
4. Some genera confined to the Madagascan region sensu lato (having at
least one species on either side of Mozambique channel):
Otoptera (Leguminosae, 1 sp. in S. Africa, 1 sp. in Madagascar); Xantho-
ceris (Leguminosae, 1 sp. in S. Africa, 1 sp. in Madagascar); Thilachium (Cap-
paraceae, 20 spp., 9 of which are in Madagascar and the Mascarene Islands, 1 sp.
nonendemic); Calantica (Flacourtiaceae, 5-7 spp. in Madagascar and possibly in
E. Africa, Airy Shaw, 1966); Apholia (Flacourtiaceae, 4-5 spp. or only 1 polymor-
phic species which reaches E. Africa, Airy Shaw, 1966); Ludia (Flacourtiaceae, 7
spp. in Madagascar and possibly in E. Africa, Airy Shaw, 1966); Cassinopsis (Cel-
lastraceae, 2 spp. in S. Africa, 4 spp. in Madagascar, ail endemic); Hyalocalyx
(Turneraceae, 1 sp. in Madagascar and Mozambique); Sparmannia (Tiliaceae, 1
sp. in S. Africa, 1 sp. common in Madagascar and E. Africa, 1 sp. in Madagascar;
this genus belongs to an entirely African tribe); Goss!lpioides (Malvaceae, 1 sp. in
540
[VOL. 65


LEROY-MADAGASCAR
E. Africa and Madagascar); Dais (Thymelaeaceae, 1 sp. in E. and S. Africa, 1 sp.
in Madagascar); Sphedamnocarpus (Malphighiaceae, 1 sp. endemic to S.W.
Africa, 15 spp. in Madagascar); Lepidotrichilia (Meliaceae, 1 sp. in E. Africa, 3
spp. in Madagascar); Erythrophysa (Sapindaceac-Koelreuterieae, 3 spp. in E.
Africa, 6 spp. in Madagascar); Macphersonia (Sapindaceae-Schleichereae, sev-
eral species in E. Africa-Madagascar); Camptolepis (Sapindaceae-Schleichereae,
1 sp. common to E. Africa and Madagascar, 2 spp. in Madagascar); Stadmannia
(Sapindaceae-Nephelieae, 1 sp. common to the Mascarene Islands, Madagascar
and E. Africa: S. oppositifolia; Capuron has described 5 Madagascan species rep-
resenting a very distinct group, flowers of which are dioecious and with petals);
Protorhus (Anacardiaceae, 1 sp. in S.W. Africa, 20 spp. in Madagascar); Rhoicissus
(Vitaceae, 10 spp. in E. Africa, one of which reaches the Comores Islands);
Rhigozum (Bignoniaceae-Tecomeae, 3 spp. in S. Africa, 1 sp. in Madagascar);
Hydnora (Hydnoraceae, 10 spp. in S. and E. Africa, only 1 sp. in Madagascar,
which represents a section of its own, Neohydnora; the other genus Prosopanche of
this bitypic family is endemic to S. America-Paraguay, Argentina); Kniphofia
(Liliaceae, many spp. in E. and S. Africa, 1 sp. in Madagascar); Geissorrhiza
(Iridaceae, many spp. in S. Africa, 2 spp. in Madagascar); Gonioma (Apocyna-
ceae-Alstonieae, 1 sp. in S. Africa, 1 sp. in S.W. Madagascar, Markgraf, 1976);
Mascarenhasia (Apocynaceae-Mascarenhasiinae, a monotypic tribe; 12 spp. in
Madagascar, 1 of which reaches E. Africa); Pachypodium (Apocynaceae-Pachy-
podiinae, a monotypic tribe; a genus with 17 spp., 12 of which are in Madagascar,
the other species are in S.E. and S.W. Africa); Hazunta (Apocynaceae-Taber-
naemontanoideae, 7 spp. in Madagascar, 1 of which extends to E. Africa, H.
coffeoides); Uncarina (Pedaliaceae; it is the single genus of the Pedaliaceae en-
demic to Madagascar; this genus is closely allied to Harpagophyton of S. and
E. Africa); Medemia (Palmae, 3 spp. in N.E. Africa and Madagascar).
5. Some genera characteristically of the Madagascan region sensu lato but
nonendemic to this region and absent from western Africa:
Ocotea (Lauraceae, 270 spp. in America, several spp. in S.E. Africa, 16
spp. in Madagascar; Ocotea could have some distant affinity with Hypodaphnis
a monotypic genus of W. Africa); Hirtella (Chrysobalanaceae, 100 spp. in trop-
ical America, 1 in E. Africa, 4 spp. in Madagascar-3 endemic-1 sp. common
to E. Africa and Madagascar; this genus is related to the Madagascan endemic
genus Grangeria); Carpodiptera (Tiliaceae, 3 spp. in Cuba and West Indies, 1 sp.
in E. Africa and Madagascar); Piriqueta sect. Erblichia (Turneraceae, 1 sp. in
Central America, 1 sp. in southern Africa, 3 spp. in Madagascar; the genus
Piriqueta is greatly diversified in America; a Madagascan species, Piriqueta
integrifolia, is an isolated taxon whose status needs restudy); Vepris (Rutaceae,
20 spp. in S. and E. Africa, 24 spp. in Madagascar); Tristellateia (Malpighiaceae,
a palaeotropical genus with 20 spp. in Madagascar, 1 sp. in E. Africa); Nesogenes
(Verbenaceae, 6 spp.; this genus, closely allied to the Madagascan endemic genus
Acharitea (1 sp.), has one endemic sp. in Madagascar, the other species being
in E. Africa and Polynesia; these two genera belong to a group, a family of its
1978]


ANNALS OF THE MISSOURI BOTANICAI GARDEN
own according to some authors [Dicrastylidaceae], which is essentially Aus-
tralian; they are closely related to the Verbenaceae-Stilboideae, a small sub-
family of 5 monotypic genera, endemic to S. Africa); Dicoma (Compositae-
Mutisieae, 35 spp. in E. and S. Africa, 1 sp. in tropical Asia, 4 spp. in Madagascar,
all endemic); there are many genera in the Compositae-Inulae with a compara-
ble distribution: Polycline, Athrixia, Iphiona, Ilunmea, Brachylaena, Stenocline,
etc.).
6. Sone genera characteristically of the Madagascan region sensu lato,
nonendemic to this region and weakly represented in tropical West Africa:
Buxus (Buxaceae, 6 spp. in S. and E. Africa and Madagascar, 1 sp. belong-
ing to another group in W. tropical Africa); Delonix (Leguminosae, a genus of
several spp. in Madagascar and E. Africa); Commiphora (Burseraceae, a genus
with many spp. of warm Africa and India; some spp. in W. tropical Africa, more
than 20 spp. in E. Africa, even more in Madagascar); Philippia (Ericaceae,
40-50 spp. in Madagascar, 10 spp. in E. and S. Africa, 1 sp. on Mt. Cameroon);
Vaccinium (Ericaceae, some spp. in Madagascar, many in S. Africa but none on
mountains in tropical Africa); Dombeya (Sterculiaceae, 200 spp., 20 of which
are in E. and S. Africa and a few elsewhere in tropical Africa); Gravesia
(Melastomataceae, 100 spp. in Madagascar, 7 of which are common to E. Africa,
2-3 spp. in W. Africa); Mimusops (Sapotaceae, a genus of 18-19 spp. in Mada-
gascar, the Comores, Mascarene and Seychelles Islands, 12 in E. and S. Africa,
only 2 in W. Africa); Sphaerothylax (Podostemaceae, 1 sp. in Cameroon, 1 in S.
Africa, reaching Angola and Zambia, 1 in E. Africa, Ethiopia, and Madagascar);
Kalanchoe (Crassulaceae, 1 section in S. Africa, 3 sections in Madagascar-
Dejardin et al., 1973); Streptocarpus (Gesneriaceae-Cyrtandroideae, a genus of
many species in S. and E. Africa, some 40 spp. in Madagascar-where a genus
closely allied, Colpogjne, is confined-some spp. in W. tropical Africa); Heli-
chrysum (Compositae-Inulae, a palaeotropical genus of 500 spp.-extending
to Australia-of which over 100 occur in Madagascar and 200 in S. Africa);
Stoebe (Compositae-Inulae, 30 spp. in S. Africa, 1 sp. in Angola, 1 in Rhodesia,
2 spp. endemic to Madagascar, 1 sp. endemic to R�union); Aloe (Liliaceae, a
genus well developed in Madagascar, where it has 30 endemic spp. also in
E. and S. Africa); Coleotrype (Commelinaceae, 5 spp. in Africa and Mada-
gascar).
7. The littoral formations, mangroves included, along east African coasts
also offer characteristic features of the Madagascan region. These formations
are composed of plants unknown on the Atlantic coast such as Barringtonia
racemosa, Sonneratia alba, Lumnitzera racemosa, Bruguiera gymnorhiza, Rhi-
zophora mucronata, Ceriops tagal, Heritiera littoralis, Xylocarpus (the two spe-
cies) and Cycas thouarsii. A few mangrove species of Madagascar do not reach
the African coast, but many other species of the oriental mangroves reach neither
the Madagascan coast, nor the eastern African one. The Madagascan mangrove
flora is a typical oriental flora, but a strongly impoverished one.
542
[VOL. 65


LEROY--MADAGASCAR
BIOGEOGRAPHICAL SURVEY OF TIIE MADAGASCAN FLORA
(Families are arrangcd roughly according to Cronquist's system)
I
Winteraceae (8 gen., 57 spp.), which are clearly among the most primitive
families of the angiosperms, are represented in Madagascar by a unique species,
the so-called "Bubbia" perrieri, which is the single representative of the family
in Africa-Madagascar. The other genera but one are concentrated in some Pa-
cific islands and Australia: Pseudowintera (3 spp. in New Zealand); Exosper-
mum (1 sp. in New Caledonia); Bubbia (30 spp. in New Caledonia, New Guinea,
and Australia); Belliolum (4 spp. in New Caledonia, 4 spp. in the Solomon Is-
lands); Zygogynum (6 spp. in New Caledonia); Tasmannia (5 spp. from Malesia
to Australia, a genus often treated as a subgenus of Drimys); and Drimys (4 spp.
in S. America-Mexico).
The Madagascan species wrongly described as Bubbia is indeed a quite
different genus, and I proposed (Leroy, 1977a) considering it as a subfamily
of its own. Thus, the distributional pattern of the Winteraceae is not only
typically austral, but the Madagascan species bears witness that the Madagascan
region is a part of what was the cradle of the Winteraceae; therefore, a part
of the cradle of primitive angiosperms.
But more is to be said. "Bubbia" perrieri has an advanced gynoecium that
is a pseudo-monomerous ovary composed of 2 carpels united and opened, a type
only known in two families in the order Magnoliales, the Annonaceae-Mono-
doroideae and the Canellaceae. Let us not forget that all of these plants occur in
one and the same area, except for the Canellaceae which occur also in tropical
America. The Winteraceae, which have no representative in Africa, and the
Canellaceae form an integral whole because of their close relationships, and
they should not be too separated.
Annonaceae 125 gen., only 9 of which are represented in Madagascar: 1
endemic, Ambavia (2 spp.); 3 African-Madagascan, Isolona (19 spp., 5 in
Madagascar), Monanthotaxis (some spp. in Africa and Madagascar), Hexalobus
(6 spp., 1 in Madagascar); 4 palaeotropical, Uvaria (120 spp., 15 in Madagascar),
Artabotrys (100 spp., 6 in Madagascar), Polyalthia (100 spp., 15 in Mada-
gascar), Popowia (50 spp., 16 in Madagascar); 1 pantropical, Xylopia (100 spp.,
25 in Madagascar). All but some of the species (the total number of which is
100) are endemic.
Annonaceae are rather poorly represented in Madagascar. As a whole, their
affinities are palaeotropical and especially African. A remarkable fact is that the
subfamily Monodoroideae, principally characterized by a compound ovary, and
composed of 2 genera (Isolona in Africa and Madagascar, and Monodora in
Africa), palynologically akin (Walker, 1971), is endemic to Africa-Madagascar.
Myristicaceae (16 gen., 5 of which in America-Uphof, 1959-300 spp.).
This family is represented in Madagascar by 3 endemic genera (Capuron, 1972a,
1973b): Haematodendron (1 sp.), Brochoneura (3 spp.), and Mauloutchia (6
spp.); in tropical Africa by 4 endemic genera; in tropical Asia by 4 endemic
genera. Taking the family as a whole, almost 50% of the genera are African-
1978]
543


ANNALS OF THE MISSOURI BOTANICAL GARDEN
Madagascan. None of them is pantropical. The Madagascan genera, like the
African ones, have few species, always less than 7.
Furthermore, the genus Haematodendron is, as compared to the others,
strongly differentiated with some very primitive characters (stigma sessile, sub-
apical-lateral, slightly oblique, with two well-marked papillose lips, separated
by a sutural groove which nearly extends down to the bottom of the carpel;
therefore, certainly a monomerous ovary!) and some very advanced characters
(plants dioecious, flowers with united sepals, ovary stipitate, seeds exarillate,
allumen ruminate).
So it is easy to assign a Gondwanian origin to this very archaic family,
whose specific differentiation has been active in America (Smith & Wodhouse,
1938) and, above ail, in Asia. The African-Madagascan genera appear to be
isolated survivors with few species.
The Canellaceae is a small tropical and subtropical family of 5 genera (11
spp.) whose distribution is restricted to America and Africa. Genera are en-
demic to any parts of the region: 3 of them in America; 1 in S. Africa; 1 in
Madagascar: Cinnamosma (3 spp.). The most striking fact is that the family
is a very natural one and closely allied to the Winteraceae: they have partly
the saine distributional pattern and the saime ovarian type (Leroy, 1977a). It is
probable that they have had a common ancestor (with a monosulcate pollen,
which is the one of present Canellaceae, and free carpels), which inhabited
Gondwanaland.
The Monimiaceae (25 gen.) is one more example of a panaustral family
almost absent from the Guineo-Congolese flora, but richly represented in the
Madagascan region sensu lato. There is a strong concentration of endemic
genera in Madagascar and the Mascarene Islands, possibly 6: Ephippiandra (1
sp.); Phanerogonocarpus (2 spp.); Decarydendron (1 sp.); Hedycaryopsis (1
sp.); Monimia (4 spp.); Tambourissa (26 spp.). Two genera are recognized in
Africa: Xymalos (2-3 spp.) almost restricted to E. and S. Africa but reaching
Mt. Cameroon; Glossocalyx (3-4 spp.) confined to the Guineo-Congolese flora.
The latter genus is strongly differentiated and often considered as the single
African member of the separate family Atherospermataceae. According to
Iutchinson (1964), no representative of the Monimiaceae occurs in Africa, since
he places Xymalos in the Trimeniaceae. Two other genera of Monimiaceae have
been seen as types of separate families. It has been proposed that Hortonia of
Ceylon be raised to the family level, a status which is generally granted to
Amborella of New Caledonia. In spite of the unsatisfactory taxonomic treat-
ment applied to the Monimiaceae, it is clear that this group of related plants,
which is widely dispersed in the Southern Hemisphere, is characterized by its
discontinuons area, the morphological isolation of certain genera, and its near
absence from Africa.
Lauraceae (35-40 gen., 2,000 spp. tropical & subtropical). This primitive
family is very poorly represented in Africa-Madagascar, but its distribution offers
a great interest. Aside from the parasitic vine Cassytha filiformis, present in both
Africa and Madagascar, there are 4 genera in continental Africa, of which one is
an endemic monotypic genus (Hypodaphnis) and belongs to a tribe of its own.
544
[VOL. 65


LEROY-MADAGASCAR
Beilschmiedia, a pantropical genus of 200 species, with 80 species in Africa, is
absent from Madagascar. Two genera, Ocotea and Cryptocarya, are common
to Africa and Madagascar; the latter is a large American genus with 8 species
in Madagascar and some species restricted to E. and S. Africa. The American
genus, Ocotea, has 18 species in Madagascar, and only a few in tropical and E.
and S. Africa. There are 3 other genera in Madagascar: Ravensara, an endemic
genus (18 sp.) related to Cryptocarya (Cryptocaryeae); Potameia, a genus with
many Madagascan species (19) and a few (3) in tropical Africa; Apollonias,
a widely spread genus with 6 endemic species in Madagascar, absent from
continental Africa but present in the Canaries (with Persea and Laurus).
It is to be noted that the tropical African center is very poorly provided with
genera of Lauraceae, the chief centers of distribution being S.E. Asia and S.
America. The family is better represented in the adjoining islands (Canaries,
Madagascar) than on the continent.
Hernandiaceae. This small pantropical family of 4-5 genera and 60 species
has a typically austral distribution. It is of the utmost theoretical importance
in that it offers not only an austral distribution, but also a wide range of isolated
genera with many very primitive characters. There is undoubtedly evidence for
a Gondwanian origin of the whole family (Kubitzki, 1969). Hernandiaceae are
represented in Madagascar by 4 genera: Hernandia subgen. Hernandia (2 spp.,
1 of which is endemic; 1 sp. endemic to the Mascarenes); Hazomalania ( Her-
nandia subgen. Hazomalania (Capuron, 19661; Kubitzki, 1969) (1 sp.); Illigera
(1 endemic sp. ); Gyrocarpus (1 sp., G. americanus). According to Kubitzki, most
primitive components occur in Australia (Hernandia subgen. Valvanthera) and
Madagascar (Hazomalania). The fifth genus, Sparattanthelium, is endemic to
tropical America. Kubitzki has divided Gyrocarpus, a pantropical genus dis-
persed by sea currents, into 8 species, 3 of which are endemic to Madagascar; it,
almost certainly, belongs to the mechanism of diffenentiation to which such spe-
cies as Calophyllum inophyllum are also answerable.
Chloranthaceae. This small family is of exceptional interest for both its mor-
phological characters, some of them being very primitive, and its distributional
pattern. If the Madagascan genus Ascarinopsis (Humbert & Capuron, 1955)
(1 sp.) is reduced to Ascarina, of which, at the best, it could be a special section,
the Chloranthaceae is composed of 4 genera, which are, of course, related with
one another, but only distantly, and their distributional area is extremely discon-
tinuous. And what is more, the only genus Hedyosmum, which has 40 species in
tropical America and the West Indies, and 1 species in S.E. Asia, represents
a strongly differentiated, isolated line with at least special subfamily status. In
particular, it offers evolutionary characters some of which are most advanced and
some others most primitive. It seems its distribution might have resulted, more
or less recently, from a post-Gondwanian extension. The other 2 genera are:
Chloranthus (15 spp., in E. Asia and Indomalaysia) and Sarcandra (3 spp. in
the same area). Obviously, I agree with Smith (1972) that the present genera
of Chloranthaceae are only a few remnants of an old, large Angiosperm flora
which has almost entirely disappeared from the earth. With this interpretation
the presence of one Ascarina species in Madagascar has no clear significance,
1978]
545


ANNAIS 0F THE MISSOURI BOTANICAL GARDEN
as the species may have corne by long-distance dispersal at any time during the
Tertiary.
Piperaceae, Aristolochiaceae, Nymphaeaceae, Ceratophyllaceae, Ranunculaceae,
Papaveraceae (cf. Humbert, 1959).
Menispermaceae (65 gen., 350 spp., chiefly in the tropical Old World). They
are particularly well represented in Africa with 21 genera, 12 of which are en-
demic, and 65 species. There are perhaps 12 genera in Madagascar, 5 of which are
endemic (Orthogynium, Burasaia, Rhaptonenui, Spirospermum, Strychnopsis).
Two genera are common to Madagascar and Africa (Anisocycla, Triclisia). To
summarize, 19 genera are endemic to Africa-Madagascar and the remainder,
except for 2, are palaeotropical. There is no direct link with America. This
pattern of distribution, typically Gondwanian, poses a major problem because the
other families in the order Berberidales are without doubt Laurasian.
II
Hamamelidaceae. This family, composed of 22-25 genera and about 100 spe-
cies, ranges from S. and N. America to Australia, but occurs principally in S.E.
Asia; it has 2 elements in Africa and Madagascar: Dicoryphe in Madagascar with
15 species and Trichocladus in tropical E. Africa and S. Africa with 5 species.
According to Hutchinson this pattern of distribution resembles that of Magnoli-
aceae. There are however some interesting differences in that the latter family
is absent from Australia and Africa and present in S. America. In spite of their
very strong diversification in the Northern Hemisphere, principally in S.E.
Asia, IIamamelidaceae are characterized by a widespread and discontinuons
distribution that includes Africa and Australia. This pattern of distribution
provides evidence of primitiveness and possibly that Hamamelidaceae might
have originated in Gondwanaland according to one of two hypotheses envisaged
by Raven & Axelrod (1974).
Myrothamnaceae is a monotypic family with 2 closely related species: one
endemic to Madagascar, the other in E. and S. Africa (but also in Angola,
Zambia and S. Za�re).
Hydrostachyaceae. A monotypic family with the same pattern of distribu-
tion as the previous one, but with many species (18 endemic spp. in Madagascar,
some 30 in S. Africa).
We need detailed researches on these small austral families considered as
primitive and bunched together within Pittosporales (suborder Brunineae) by
Thorne (1968, 1975), lut the Hydrostachyaceae are placed in the Scrophulariales
by Cronquist (1968).
Did ymelaceae. A monotypic family of 2 species endemic to Madagascar
whose taxonomic position is still uncertain. It is a rather primitive taxon (with
solitary carpel and scalariform perforations) considered sometimes as close to
Euphorbiaceae or Hamamelidales, or sometimes as an order of its own (Takhtajan,
1969).
Myricaceae. (4 gen., 40 spp.). This family, which is widely dispersed in
temperate regions and tropical mountains of both hemispheres, is divided into two
546
[VOL. 65


LEROY-MADAGASCAR
subfamilies, one of which, monotypic and very archaic, the Canacomyricoideae,
is endemic to New Caledonia (Leroy, 1949, 1957). The genera Comptonia (en-
demic to N. America, with 1 sp.), Gale (1 sp. in the Northern Hemisphere),
and Myrica (temperate and tropical zones) are closely allied. The family is pres-
ent in S. America, in Africa (many spp. in S. Africa), and in Madagascar with
the genus Myrica (6 spp.).
Myricaceae is generally considered as a Laurasian family, but two facts
argue against this view. First, the tropical distribution of Myrica, and, in par-
ticular, its abundance in S. Africa. Second, the existence of Canacomyrica in
New Caledonia, where it certainly is a completely isolated survivor. Comptonia
and Gale are possibly Laurasian segregants, but Myrica and, above all Canaco-
myrica, might indicate greater antiquity.
Moraceae. This family divided into 3 subfamilies (Moroideae, Artocarpoideae,
Conocephaloideae), 75 genera, and 2,000 species is widespread in warm regions
and the tropics and is represented in Madagascar by 11 genera and 42 species.
If, following Hutchinson's division, we accept 9 tribes, all of them are represented
in Madagascar, each of them by one or two genera (the small subfamily Cono-
cephaloideae being excluded). Fatoueac: Fatoua, a genus with only 2 species,
of which 1 occurs in Madagascar, the other in Asia, New Caledonia. Dorstenieae:
Dorstenia, a pantropical genus with 100 species of which 2 occur in Madagascar.
Strebleae: Streblus, a palaeotropical genus with 4-5 species, of which 1 occurs in
Madagascar; Maillardia, an endemic genus with 4 species. Moreae: Ampalis
and Pachytrophe, 2 endemic and monotypic genera. Broussonetieae: Allaeanthus,
an endemic and monotypic genus. There are 4 genera in the Artocarpoideae:
Euartocarpeae: Treculia, an African-Madagascan genus with 12 species of which
4 occur in Madagascar. Brosimeae: Bosqueia, another African-Madagascan
genus with 12 species, of which 7 occur in Madagascar. Olmedieae: Antiaris,
a palaeotropical genus with 10 species, of which 2 occur in Madagascar. Ficeae
with 16 species in Madagascar.
The Madagascan Moraceae offer a basis for some interesting comments.
First, we have to emphasize that the whole family is represented in Madagascar
by ail the tribes. Even the endemic genera are distributed into 3 tribes. Judging
from these facts, one has good justification to uphold the hypothesis of Madagascar
as a center of survival. Second, since there are 4 endemic genera, 2 African-
Madagascan ones and 3 palaeotropical ones, links with America are weak. Third,
it is noteworthy that the tribe Moreae, represented by 2 endemic genera, has no
representative of Morus, a widespread genus, one subgenus of which occurs in
tropical Africa, another in Asia, and yet another in America.
Urticaceae. A family of 50 genera (800 spp.) mostly in warm regions and
tropics, is particularly well represented in America. There are 14 indigenous
genera in Madagascar, one of whieh, Neopilea (2 spp.) is endemic, and 51
species (41 endemic) occur on the island.
As with the Moraceae, all tribes (5) of Urticaceae are represented in Mada-
gascar, but endemism is weak, and pantropical or widespread genera are as
many (6 in number) as palaeotropical ones. There is also an American-Mada-
gascan genus, Phenax, with 20 species in America and 3 in Madagascar and the
547
1978]


ANNALS OF THE MISSOURI BOTANICAL GARDEN
Mascarenes. Further, one palaeotropical genus, Pipturus, reaches the Comores (1
nonendemic sp.) but neither Madagascar nor Africa.
Ulnmaceae (15 gen., 150 spp.). This family has 4 genera in Madagascar:
Celtis (3 Madagascan spp.; 1 of which is palaeotropical, the other 2 are related
to African species); Trema (2 Madagascan spp., 1 of which is endemic with
eastern affinities); Aphananthe (1 sp., endemic; the genus is closely related to
Celtis); Chaetacme (1 African-Madagascan sp.).
Casuarinaceae (cf. Humbert, 1959).
III
Nyctaginaceae (cf. Humbert, 1959).
Phytolaccaceae (12 gen., 100 spp.). A family particularly well represented
in America and S. Africa. Four genera are present in Madagascar, 1 of which,
Barbeuia, is endemic, monotypic, and morphologically isolate (ovary 2-carpellate,
syncarpous, 2-loculary). There is an endemic species of Phytolacca.
Didiereaceae. This remarkable family, endemic to South and Southwest Mad-
agascar, consists of 4 genera: Decaryia (1 sp.); Alluaudiopsis (2 spp.); Alluaudia
(6 spp.); Didierea (2 spp.).
Didiereaceae, certain of which have a habit resembling American Fouquieri-
accae, are, with arborescent aphyllous euphorbias, the most striking feature of
the southern Madagascan vegetation; it recalls somewhat the Brazilian "Caatinga"
(Koechlin, 1972). According to Rauh's work (Rauh & Reznik, 1971), the family
has definite affinities with the American family Cactaceae and must be placed
in Centrospermae.
Cactaceae. It is a family of many genera, perhaps over 100, and 1,500 species
almost wholly American. Rhipsalis, the single genus with representatives outside
America, has one species, R. baccifera (J. S. Mill.) Stearn in tropical Africa,
Ceylon, and Madagascar. And what is more, the Madagascan plant is so variable
that several species have been described. According to Guillaumet (1972) there
are at least two well characterized forms, each of them inhabiting a special
ecological region: R. horrida (eastern littoral forest) and R. suarezensis (western
Domain). This pattern of distribution has given rise to discussions, but it is
probable that this differentiation has originated secondarily. A study of the
whole genus Rhipsalis is needed before any opinion can be attempted.
Aizoaceae (Molluginaceae included) (130 gen., 1,200 spp., chiefly S. African).
This unrevised family has some nonendemic genera in Madagascar (Mollugo,
Sesuvium, Trianthema, . . .) with about a dozen species (Baron, quoted by
Humbert, 1959).
Caryophyllaceae. This family is represented by 7-8 nonendemic genera
(subfamily Alsinoideae) and 11 species, 8 of which are endemics. There are
2 endemic species of the palaeotropical genus Corrigiola and 5 species of the trop-
ical and subtropical genus Polycarpaea (Perrier de la B�thie, 1936).
Portulacaceae. This is a very interesting family of about 20 genera, most of
them American. But 7 genera occur in Africa sensu lato, 4 of them endemic to
Madagascar: Talinella (2 spp.), or South Africa: (Ceraria, 4 spp.; Portulacaria,
548
[VOL. 65


LEROY MADAGASCAR
2 spp.; Calyptrotheca, 2 spp.) (Brenan, 1949; McNeill, 1974). The genus
Anacampseros (50 spp.) is common to S. Africa and Australia. The genera
Portulaca and Talinum are represented; the former in Madagascar (3 spp.), the
latter in the Mascarenes (1 sp.). Although the family has two main centers of
distribution, S. America and western N. America, it is as a whole a rather austral
one. Moreover, it is closely related with the large family Aizoaceae, absent from
Madagascar, but primarily represented in S. Africa. With respect to Portulacaceae,
Madagascar and Australia look as if they were parts of a borderland.
Basellaceae, Chenopodiaceae, Amaranthaceae, Polygonaceae, Plumbaginaceae
(cf. Humbert, 1959).
IV
Dilleniaceae (10 gen., 400 spp.). This family is represented in Madagascar
by 3 nonendemic genera: Hibbertia, a genus of 155 species, 130 of which are
in Australia, 24 in New Caledonia and Fiji, 1 (H. coriacea) in Madagascar (this
one is closely related to Australian species); Tetracera, a pantropical genus of
50 species with 3 species in Madagascar, all of them endemic, and about 10
species in Africa. Dillenia, an Asian-Australian genus of 60 species, 1 of which,
nonendemic is in Madagascar, and another endemic to the Seychelles. Apart
from Tetracera, none of the genera of Dilleniaceae occur in Africa. According
to Hutchinson there is one more genus, endemic to the Seychelles (Neowormia),
related to Dillenia. Let us also not forget that the genus Schumacheria (3 spp.) is
endemic to Ceylon. One must, it seems, put apart Hibbertia, which may have
arrived by long-distance dispersal. But with Neowormia and Schumacheria we
have possibly an indication of a much more ancient settlement for the family.
Besidcs, the pantropical genus Tetracera belongs to the tribe Delimeae, largely
American. One feels as if the Dilleniaceae, almost absent from Africa today, must
have been much more numerous in the past and would have become extinct
during the Tertiary. We are also badly in need of a detailed monograph of this
especially interesting family.
Ochnaceae. (30 gen., 250 spp.) is an almost entirely tropical family, especially
well represented in S. America. Its Madagascan representatives, which belong
to subfamily Ochnoideae, tribe Ochneae (Kanis, 1968), are: Ochna, a palaeo-
tropical genus of about 100 species in Africa south of the Sahara, Arabia, Mada-
gascar, the Mascarenes, and 4 species in tropical Asia with perhaps 20 endemic
species in Madagascar. Gomphia (= Campylospermum), a palaeotropical genus
of 12 species, of which 5 are endemic to Madagascar, and 3 in Africa.
Brackenridgea (= Pleuroridgea), another palaeotropical genus, containing 10 spe-
cies, of which 2 endemic to Madagascar.
It is striking that there is no endemic genus. With the Madagascan Ochnaceae
we deal with a part of a palaeotropical group, even though the family is chiefly
South American.
Diegodendraceae (1 gen., 1 sp.). An endemic family (Capuron, 1963b).
Sphaerosepalaceae. This is an endemic family of Madagascar which consists
of 2 genera of woody plants: Rhopalocarpus (15 spp.) and Dialyceras (3 spp.).
549
1978]


ANNALS OF ITHE MISSOURI BOTANI(AL GARDEN
The best place for this family seems next to Diegodendraceae within Ochnales,
but there are also links with Malvales.
Sarcolaenaceae (10 gen., 35 spp.). An endemic family.
Elatinaceae (2 gen., 45 spp.) (cf. Humbert, 1959).
Guttifereae and tHypericaceae represent a chiefly tropical group of 1,000
species in 40 genera, of which 8 are endemic to Africa-Madagascar. If we put
aside Hlypericum, an herbaceous genus of wide distribution throughout tem-
perate regions and tropical mountains, 2 genera offer a widespread distribution:
Garcinia and Calophyllum. However, both are far more highly differentiated in
Asia-Australasia than in America or Africa; Calophyllum, which has 16 Mada-
gascan species, is completely absent from Africa. Symphonia is abundantly
diversified in Madagascar with many species or "morphological entities" merg-
ing into one another, but there are also 1-2 species in Africa and America. Its
affinities are with Montrouziera of New Caledonia. Twenty species or so of
Mammea inhabit Madagascar, ail of them endemic, over 20 other ones occur
in Asia-Indonesia, extending to New Guinea, except for Australia, but only 1
is met in Africa and another in S. America.
There are 3 endemic genera: Eliaea, related to the Asian genus Cratoxylum,
Ochrocarpos, and Pararunmmea (Leroy, 1975, 1977b), each of them belonging
to a different tribe. Finally, the other 2 genera, Ilarungana and Psorospermum,
are common to Africa and Madagascar.
From these data, which are difficult to understand, one can however draw
some conclusions:
Guttifereae are comparatively 1)etter represented in Madagascar than in
Africa. The affinities between certain genera such as Mammea-Paramammea
on the one hand, and Garcinia-Ochrocarpos on the other hand, show that these
genera are probally not newcomers and that they prolably differentiated there.
Such genera as Symphonia and Mammea look as if they had originated and
evolved in Madagascar, although their affinities are with Asian-Pacific genera.
There is some evidence that as a whole, the Guttifereae-Hypericaceae of Mad-
agascar and Africa may have differentiated there, when S. America, Africa, and
Australia were no longer in contact.
The fact that 14 endemic species of Calophyllum (Stevens, 1976) occur in
Madagascar, the genus being unknown in Africa, poses a problem. According
to Perrier de la B�thie (1936) all of the Madagascan species are derived from
the beach species Calophyllum inophyllum. There is a great need for thorough
research in evolutionary biology of this very interesting problem. Wlatever the
answer will be, it is quite certain that Calophyllum bas occurred there for a very
long time.
Dipterocarpaceae. This family of 18 genera, which has until now been con-
sidered as a typically palaeotropical one, ias at least one representative (a mono-
typic subfamily) in Guyana (Maguire et al., 1977). It is basically divided
into 2 very distinct subfamilies: Dipterocarpoideae and Monotoideae. The
former is Asian with 13 genera and 247 species in only Borneo, a number that
quickly decreases going toward India where only 6 genera and 14 species oc-
cur. However, it is surprising enough to meet as many as 10 genera and 44
550
[VOL. 65


LEROY-MADAGASCAR
species in Sri Lanka (Ceylon). Moreover, one genus, Vateriopsis, is endemic
to the Seychelles and several Tertiary fossils have been found in E. Africa
(Aubr�ville, 1976a). The Monotoideae comprises 2 genera: Monotes with 1 en-
demic species in Madagascar, 1 species in western tropical Africa, and 40 species
or so in dry tropical S. Africa; Marquesia, with 4 species in tropical Africa. Some
Tertiary fossils of this subfamily have been described from the temperate North-
ern Hemisphere, but their identity is not always convincingly established.
One can perhaps presume that Dipterocarpaceae has been a pantropical
family which has become particularly impoverished in Africa during the Tertiary
era owing to geological and climate alterations. It would have found a center
of survival in Southeast Asia which has also become a new center for develop-
ment (Meijer, 1974). It is however noteworthy that the family is very poorly
represented in Madagascar, and absent from Australia.
Theaceae: cf. Asteropeiaceae.
Asteropeiaceae. This unigeneric family, endemic to Madagascar, contains
5-6 species which occur both in the western and eastern parts of the country.
It is of great interest due to its close affinities with Theaceae (Melchior, 1925),
from which it is distinguished by minor characters; Asteropeia is clearly a relict.
We have to emphasize that Theaceae are represented in Africa-Madagascar by
only 2 species of the large genus Ternstroemia (Hepper, 1968; Letouzey, 1977b)
and 2 genera endemic to Africa: Melchiora (3 spp.) and Ficalhoa (1 sp. ?).
Medusagynaceae. (An unispecific family endemic to the Seychelles.)
Elaeocarpaceae. This family (8 gen., 400 spp.) is closely allied to Tiliaceac,
of which it is considered often as a subfamily or a tribe, and is typically of
southern distribution (S. America, Australia, New Zealand, New Caledonia, New
Guinea), the only genus Elaeocarpus reaching northward India, China, and
Japan. Curiously enough, there are 2 endemic species of Elaeocarpus in Mada-
gascar (though none in Africa) and 1 endemic species of Sloanea, a widespread
genus (likewise absent from Africa). They "probably reached Madagascar by
long-distance dispersal around the Indian Ocean" (Raven & Axelrod, 1974).
One must not exclude, however, that this mechanism may have operated early
in the Tertiary when Australia was closer to Africa.
Tiliaceae is a predominantly woody family (60 gen., 1,200 spp.) mostly of
tropical and subtropical distribution. Following Capuron (1963a, 1964), 7 gen-
era occur in Madagascar: Christiana (1 African-Madagascan sp.); Carpodiptera
(1 sp. in E. Africa and Madagascar, 3 spp. in Cuba and the West Indies);
Sparrnannia (1 Madagascan sp., 1 Madagascan-E. African, 2 S. African); Grewia
(more than 400 spp. of the tropics and subtropics); Pseudocorchorus (a monotypic
tribe Pseudocorchoreae, 6 spp.); Corchorus (6 spp., nonendemic); Triumfetta
(5 spp., nonendemic).
We also have to note the singular bipolar distribution of the genus Carpodip-
tera, the distribution of Sparrnuannia, a genus of an entirely African tribe,
and the fact that Christiana is a morphologically very isolated genus. The family
is curiously anchored on Madagascar with an endemic tribe. Madagascan
Tiliaceae are, however, strongly linked to Africa of which they are only a part.
1978]
551


ANNALS OF THE MISSOURI BOTANICAL GARDEN
Sterculiaceae is a woody family (38 gen., 1,000 spp.; Hutchinson, 1967)
widely distributed in the tropics and subtropics. Out of 11 tribes recognized by
Hutchinson, 6 are present in the Madagascan region, containing 22 genera (250
spp.). The tribe Dombeyeae, which is wholly palaeotropical, is composed of
11 genera, of which 8 occur in the Madagascan region, 6 being endemic; the
nonendemic genera are: Dombeya, common to Madagascar and Africa but far
more diversified in Madagascar, and Melhania, a widespread genus extending
to Africa, Asia, and Australia.
Another fact must be pointed out: it is that 3 genera offer a very restricted
endemism: Astiria (1 sp.) and Trochetia (5 spp.) to Mauritius, and Ruizia to
R�union. A third remark is the very important development of the African-
Madagascan genus Dombeya in Madagascar and the Mascarenes, where about
200 species have been described, that is four-fifths of the whole.
The trilb Helmiopsideae consists of 3 genera, 2 of which are cndemic to Mad-
agascar, the third, Nesogordonia, having 10-12 species, half of which are Mada-
gascan, the other half African.
The tribe Hermannieae has 3 Madagascan representatives but is deprived,
it seems, of any biogeographical significance. The genus Keraudrenia (Lasio-
petaleae, 12 gencra) has one species in Madagascar (a single and ancient her-
barium number is known of this plant).
More interesting is the tribe Byttnericae with 2 genera indigenous in Mada-
gascar: the pantropical Byttneria and the Asian-Australian Rulingia (25 spp.),
one species of which is endemic to Madagascar.
Sterculieae is present with 1 pantropical genus, Sterculia, and 2 palaeotrop-
ical genera: Firmiana and Pterygota.
There are 2 genera of the palaeotropical tribe Tarrietieae: Hildegardia (2
spp. out of 7) and Heritiera (1 sp. out of 7).
In summary, such isolated representatives as Keraudrenia and Rulingia have
possibly no particular meaning from a biogeographical point of view; as Raven
& Axelrod have noted, these species probably reached Madagascar "by long-
distance dispersal around the Indian Ocean."
But I want to emphasizc some important facts: a vcry strong endemism of a part
of the family in Africa-Madagascar (Dombeyeae, Helmiopsideae); a high de-
gree of endemism in the Mascarene; the affinities of the Madagascan Sterculia-
ceae are with the palaeotropical flora. Many genera both monotypic or paucispe-
cific and endemic, particularly in the Mascarenes, appear to be survivors.
Bombacaceae is a family of 31 genera and 225 species (Hutchinson, 1967),
wholly woody and confined to the tropics, mainly the American ones. The
Durioneae, generally considered as one of the primitive tribes, is Asian-Aus-
tralasian, but the other tribes are either strictly American or represented in
America. Bombacaceae are represented in Africa and Madagascar only by the
tribe Adansonieac, with 15 species distributed in 4 genera: the palaeotropical
Bombax (2 spp. in Africa), the endemic genus Rhodognaphalon, closely re-
lated, if not similar to the American Bombacopsis, with 6 species in Africa, and
Adansonia which extends from Africa (1 sp.) and Madagascar (8 spp.) to
Australia (1 sp.). A third genus (Ceiba, tribe Ceibeae, also exists in Africa, but
552
[VOL. 65


LEROY-MADAGASCAR
it is unanimously considered as an American newcomer. Judging from the
present evidence, one may suggest a Gondwanian (South American-African)
origin for the Adansonicae. A noteworthy fact recalled by Aubr�ville (1975a) is
the African disjunction in the pattern of distribution concerning the whole family.
Malvaceae. This family (88 gen., 2,300 spp.) is world wide both in temperate
and tropical regions. It is represented in Madagascar by 3 tribes: Malveae (5
gen., 17 spp.), Ureneae (3 gen., 5 spp.), Hibisceae (13 gen., 7 of which endemic,
and 80 spp., 60 of which endemic). There are only 3 endemic species within
Malveae (Abutilon-Bates, 1968) and 2-3 endemic species within Ureneae
(Pavonia) in Madagascar-Africa. The tribe Hibisceae, which is composed of
29 genera (Hutchinson, 1967), is well represented in Madagascar and Africa,
even better in Madagascar with 13 genera, of which 7 are endemic, than in
Africa with 12 genera, of which only 3 are endemic. The only genus, Hibiscus,
occurs abundantly in Madagascar with 45 species, 30 of which are endemic.
The endemic genera are paucispecific and not always neatly separated: Megi-
stostegium (3 spp.), Perrierophytum (9 spp.), Humbertiella (3 spp.), Macro-
stelia (3 spp.), Humbertianthus (1 sp.), Helicteropsis (2 spp.), Jumelleanthus
(1 sp.). The genus Gossypioides (2 spp.) is endemic to E. Africa-Madagascar.
Such genera as Cienfuegosia (Fryxell, 1974) (America, Africa, Australia),
Kosteletzkya, and Thespesia are represented respectively by 1, 7, and 1 endemic
species. Two genera of Madagascan Hibisceae have no endemic species: Abel-
moschus and Gossypium.
Nepenthaceae (1 gen., 80 spp.). The genus Nepenthes is confined to the
Indian Ocean with 2 endemic species in Madagascar and 1 endemic in the
Seychelles.
Droseraceae (1 gen., 90 spp.). Five species inhabit Madagascar, one of which
endemic.
Lecythidaceae. Hutchinson (1964) has divided this family into two in ac-
cepting the Barringtoniaceae of Lindley, this one forming "a very natural group"
of 4 genera, Barringtonia (Africa and Madagascar), Combretodendron (Africa),
Careya (Indomalaysia), Planchonia (S.E. Asia, Australia); the genus Foetidia,
confined to the Madagascan region, is retained in Lecythidaceae with Napoleona
and Crateranthus of tropical Africa. For his part, Airy Shaw (1966) considers
Foetidia as the type and the single member of the new family Foetidiaceae.
Foetidia has 5 species (and several to be described), one of which is endemic to
the Mascarenes and another reaching the African coast (Zanzibar). About the
latter "there is no reason why it should not be native, but whether a recent
arrival or a relicts of a one wider distribution is unknown," Sangai (1971) noted.
Barringtonia, with 39 species, ranges from Africa and India to Australia; two
species confined to sandy and rocky beaches, reach E. Africa and the Madagascan
region.
Thus, there are two parts in this whole: a palaeotropical one with Barring-
tonia (and Combretodendron in Africa) and a neotropical one with Foetidia
(distantly allied to some African genera), this one having evolved separately for a
long time. Nevertheless, Barringtonia is to be considered in a special category as
a readily dispersed mangrove genus.
1978]
553


ANNALS OF TIE MISSOURI BOTANICAL GARDEN
Flacourtiaceae. This pantropical family of 84 genera (1,300 spp.) has 28
endmiic genera in Africa sensu lato (including Madagascar and the Mascarenes)
and thereby offers a special interest. It is well represented in Madagascar with
110 species and 14 genera, 8 of which are endemic (but 2 of them, Calantica and
Aphloia reach E. Africa); the most evolved tribe (according to Hutchinson, 1967),
Bembicieae, is endemic to Madagascar. There are 2 large pantropical genera:
Casearia (160 sp.; 1 endemic to Madagascar and 1 to the Mascarene Islands; 11
endemic to Africa-Sleumer, 1971); Homalium (180 spp.; 59 spp. in Africa sensu
lato, 38 of which in Madagascar, all of them endemic). Three genera are palaeo-
tropical: Erythrospermum (5 spp., 2 of which endemic to Madagascar sensu
lato, none in Africa); Scolopia (37 spp., 14 of which in Madagascar, ail endemic);
Flacourtia (60 sp.; 1, nonendemic, in Madagascar). The endemic genera are
frequently pausispecific: Prockiopsis (1 sp.-Capuron, 1968d); Bivinia (2 spp.);
Bembicia (1 sp.); Bembiciopsis (1 sp.); Aphloia (4-5 spp. or 1 polymorphic
species); Calantica (7 spp.); except for Tisonia (15 spp.) and Ludia (23 spp.).
The genus Xylotheca of tropical and E. Africa has 1 species in Madagascar. In
the same way, Aphloia reaches E. Africa (Airy Shaw, 1966).
Such great a number of highly differentiated genera, that is one-third of the
total number, are endemic to Africa-Madagascar that we may suppose the fam-
ily to have originated there.
Violaceae (20 gen.) is represented by 3 nonendemic genera: Viola (1 sp.,
nonendemic); Hybanthus (75 spp., most part tropical, 4 spp. in Madagascar, 3
of which endemic); Rinorea (250 spp. tropical and subtropical, many African,
but also well represented in South America; 27 spp. in the Madagascan region,
all but one endemic).
Onagraceae (= Oenotheraceae) is a family of 21 genera and 640 species
widespread through temperate and tropical regions, mostly as perennial herbs.
There are 2 indigenous genera in Madagascar: Epilobium with 6 species, some
of which are endemic, and Ludwigia with 9 species.
Turneraceae is a family of 8 genera, 6 of which arc endemic to Africa-Mada-
gascar. Two other genera also occur in this area, but are principally American:
Piriqueta and Turnera. There are 3 genera in Madagascar: Piriqueta (3-4 en-
demie spp.); tHyalocalyx, a monotypic genus with herbaceous habit, also present
in Mozambique; Turnera (1 nonendemic sp.). Mathurina (1 sp.) is a monotypic
genus endemic to Rodrigues. A species attributed to Piriqueta, P. integrifolia, is
an isolated taxon whose status needs to be restudied.
Passifloraceae is a family of 12 genera (600 spp.) of shrubs, vincs and herbs
of tropical and warm temperate regions. Two indigenous genera occur in Mad-
agascar, Adenia and Deidamia. The former, a palaeotropical genus of 100 spe-
cies, is represented in Madagascar by 13 endemic species. The latter is an African-
Madagascan genus with 6 species endemic to Madagascar and the Comores and
1 species confined to tropical Africa.
Physena. This is an isolated genus of 2 species endemic to Madagascar
classified in turn in Passifloraceae, Capparidaceae, and Flacourtiaceae.
554
[VOL. 65


LEROY-MADAGASCAR
Begoniaceae. This is a small family of 5 genera particularly well represented
in northern South America, but also in Pacific islands (Barkley & Golding, 1974).
The genus Begonia with over 800 species, mostly American (Baranov & Barkley,
1974), is important in Madagascar where at least 50 species occur (Humbert,
1959).
Cucurbitaceae (116 gen., 1,200 spp.) (data according to Keraudren, 1966).
This family of mainly tropical and subtropical distribution is well represented in
Madagascar with 21 indigenous genera, of which 8 are endemic. Two endemic
genera, Xerosicyos (4 spp.) and Zygosicyos (1-2 spp.), related to the African
genus Gerrardanthus, are considered as primitive. The other endemic genera
are: Seyrigia (5 spp.), related to some American genera; Trochomeriopsis allied
to the Seyrigia; Ampelosicyos (3 spp.); Lemurosicyos (1 sp.); Zombitsia (1 sp.);
and Tricyclandra (1 sp.). One genus, Peponium, is restricted chiefly to Mada-
gascar (12 spp.) but reaches E. and S. Africa. There are 4 other African-Mada-
gascan genera: Oreosyce (5 spp. in Africa, 1 sp. common to Africa and Mada-
gascar); Raphidiocystis (5 spp., 1 of which special to Madagascar); Cyclantheropsis
(3 spp., 1 of which special to Madagascar); and Cucumella (1 sp. in E. and
S. Africa, and Madagascar). Five genera are palaeotropical: Zehneria with 9
Madagascan species (5 endemic); Kedrostis with 6 endemic Madagascan spe-
cies; Muellerargia with 1 of its 2 species endemic to Madagascar, the other en-
demic to Malaysia. The American genus Cayaponia has some species in Africa,
one of which is also in Madagascar. The pantropical (chiefly African) genus
Corallocarpus has 3 Madagascan species, 2 of which endemic.
Ail of these facts support the thesis of an ancient distribution of the family
in Madagascar.
Salicaceae are represented in Madagascar by 2 endemic species of Salix.
Capparaceae. This family of 50 genera and 800 species, primarily pantropical,
ias 8 genera in Madagascar, but none are endemic. Two of them, closely akin,
are common to Madagascar and E. and S. Africa: Boscia (30 spp., 3 endemic in
Madagascar) and Thilachium (20 spp., 9 spp. in Madagascar sensu lato, 8 en-
demie). The other genera are pantropical; Crataeva (3-4 spp. endemic to Mad-
agascar); Capparis (4 spp. in Madagascar, 1-2 endemic); Cleome (7 spp. in
Madagascar, 3 endemic); or palaeotropical: Cadaba (40 spp., 1 endemic in
Madagascar); Maerua (100 spp., 5 endemic in Madagascar); Gynandropsis (1
adventitious sp.).
It is very surprising that this family, particularly well represented in Africa
with 15 endemic genera, has no genus endemic to Madagascar. This lack of
generic endemism may mean that Capparaceae reached Madagascar from Africa
at a late date (about the genus Physena, cf. Passifloraceae).
Cruciferae (cf. Humbert, 1959).
Moringaceae (1 gen., 10 spp.). This monotypic family consists of about 10
species, 2 in India, the others in Africa-Madagascar. Aside from Moringa
oleifera, known to be introduced, there are 2 endemic species in Madagascar,
M. hildebrandtii and M. drouardii. They occur in the south and southwest of
Madagascar with plants such as Adansonia or euphorbias and ail of the genera of
555
1978]


ANNALS OF TUE MISSOURI BOTANICAL GARIEN
Didiereaceae. In this region grow many xerophilous plants such as Xerophyta
dasylirioides, Selaginella nivea, Pachypodium sp., Aloe divaricata, Aloe suzannae,
Rhigozum madagascariensis, Uncarina sp., etc. The African species of Moringa
are distributed in a discontinuons area in the Southwest (M. ovalifolia) and in
the Northeast (M. longituba, M. stenopetala, M. peregrina). Two species are
known in N. India (M. concanensis, M. oleifera) (Keraudren, 1965).
Ericaceae (Vacciniaceae included) (50 gen., 1,350 spp.). This is a family
of cosmopolitan plants or one inhabiting mountains in the tropics. It is divided
into 3 subfamilies, 1 of which, Ericoideae, is well represented in Europe and Africa,
whereas the other 2 subfamilies (Vaccinioideae, Rhododendroideae) are absent
or nearly so from Africa. Three genera occur in the Madagascan region: Agauria
(7 spp.) and especially Philippia, with nearly 40 species, that is half of the whole
genus. The widespread and very distinct genus Vaccinium (a special family for
some authors) has only 4 Madagascan species. This genus, which occurs in S.
Africa, appears absent from tropical African mountains.
According to Stevens (1970, 1971) and others, Agauria is closely related to
the American genus Agarista.
Sapotaceae. Following Aubr�ville (1972, 1974b), the Madagascan Sapotaceae
have 11 genera (85 spp., all endemics) divided into 2 subfamilies. Mimusopo-
ideae: Manilkara is a pantropical genus having more than 60 species, with 7
endemic in Madagascar. Three genera are endemic: Labramia (8 spp.); Fau-
chera (11 spp.); and Labourdonnaisia (3 spp. in Madagascar, 2 spp. in the
Mascarenes). Mimusops is a genus common to E. and S. Africa (12 spp.), the
Comores (1 sp.), Madagascar (13-14 spp.), the Mascarenes (3 spp.), and the
Seychelles (1 sp.) but with 1 species is widely spread over palaeotropics to the
Pacific islands.
Sideroxyloideae: 3 African-Madagascan genera: Sideroxylon, a genus with
about 15 species, well known for its disjunct distribution ( Madeira, Cape Verde
Islands, Teneriffe, E. Africa, Madagascar, Mascarenes); 6 species are in Mada-
gascar and 6 species in the Mascarenes; Donella, 4 species in the Guinco-Con-
golese forest, 1 species in E. Africa, 8 species in Madagascar; Austrogambeya, a
genus from S. Africa and Zambia, close to Donella and Gambleya, with 1 species
in Madagascar. Gambeya, a genus principally African, with 1 species in Mada-
gascar, but with also 1 species in S. America. Two endemic genera: Tsebona
(1 sp.) and Capurodendron (22 spp.).
Judging from these data based on Aubr�ville's work, we must conclude that
Sapotaceae appears to represent a very important and strongly differentiated
Madagascan nucleus. A high degree of generic endemism, a strange disjunct
distribution in Madagascar-Africa (Sideroxylon), an intensive differentiation of
Mimusops species in Madagascar, many strong links with Africa where the
family reaches a very high development, such are the most noteworthy aspects
concerning the distributional pattern of the Madagascan Sapotaceae, from which
we could infer with Raven & Axelrod (1974) that an origin in West Gondwana-
land "appears most probable, almost certainly before the close of the Cretaceous."
Madagascar and its surrounding islands must have taken a considerable part
556
[VOL. 65


LEROY-MADAGASCAR
in saving from extinction many Gondwanian genera which are today mor-
phologically and geographically isolated, such as Sideroxylon, Labourdonnaisia,
and other palaeoendemics. One genus, Northea (5 spp.), is endemic to the
Seychelles.
Ebenaceae (2 gen., 450 spp.). The pantropical Diospyros is represented
in Madagascar by ca. 100 species, 6-7 of which occur on the African mainland
or have close relatives there. The remainder are endemic to Madagascar and do
not seem to be closely related to the mainland species. Euclea (12 spp. in
Africa) has 1 species on the Comores. There are the same number of species
of Diospyros in Madagascar as in tropical Africa, and ca. 80 species in tropical
America. About 250 species are currently accepted from tropical Asia, but this
figure must be much too high (F. White, personal communication). Although
he has not yet given any serious thought to the place of origin of the family,
White inclines to the belief that it might have been the African part of Gondwana-
land (White, pers. comm.).
Myrsinaceae (35 gen., 1,000 spp.). This family is a very natural one whose
7 genera occur in Madagascar: 2 pantropical: Rapanea (190 spp., 1 in the
Seychelles, 1 in Madagascar, 2 in the Comores) and Ardisia (2 spp. in Madagascar,
2 spp. in the Macarene); 2 palaeotropical: Embelia (100 spp. in 8 subgen., 2 of
which in Madagascar containing 8 spp.) and Myrsine (7 spp. in China, India,
Africa, Azores; 1 endemic sp. in Madagascar); 3 endemics: Monoporus (8 spp.
from sea coast to mountains), Oncostemon (97 spp. in Madagascar, 3 in the
Comores, 3 in Mauritius), Badula (14 spp. in the Madagascan region: 3 in
Madagascar, 3 in the R�union, 7 in Mauritius, 1 in Rodriguez).
It is interesting to note that the pantropical genus Ardisia is represented in
Madagascar by a subgenus endemic (Madardisia) and in Africa by another
subgenus (Afrardisia). These facts seem to at least prove some antiquity of the
genus in Africa-Madagascar.
Madagascar must have been a part of the ancient continent where Myrsinaceae
originated. Let us note in passing an unusual intense speciation in the genus
Oncostemon.
Primulaceae (cf. Humbert, 1959).
V
Cunoniaceae. This typically austral family (27 gen., 260 spp.) is represented
in Madagascar by only the genus Weinmnunia and over 20 species. This genus
with 150 species ranges from the Philippines and New Caledonia to tropical
America, where it is particularly abundant, but it is absent from Africa, India,
and Australia. There are 2 genera of Cunoniaceae in S. Africa: Cunonia (about
15 spp. in the Malay Archipelago and New Caledonia, 1 endemic species in
S. Africa); and Platylophus, the latter being monotypic and endemic. All of the
Madagascan species, which Bernardi (1965) has divided into 4 sections, are
endemics.
In spite of its scanty representation in the Madagascan region (Madagascar
and S. Africa), the family Cunoniaceae provides some most valuable facts. First,
there is one and only one endemic genus, and this one is monotypic: a fact
1978]
557


ANNALS OF THE MISSOURI BOTANICAL GARDEN
whose significance is obvious: the genus Platylophus testifies in favor of S. Africa
as a center of survival for Gondwanaland inhabitants. Second, the link between
S. Africa and Madagascar is not less obvious. Third, the genus Weinmannia
has been subjected to very active speciation since Madagascar separated from
Africa. The lack of fossils prevents a definite statement about its possible
presence in Africa in the past.
Pittosporaceae. This typically palaeotropical family of 9 genera, 8 of which
are endemic to Australasia, is represented in tropical Africa sensu lato by the
genus Pittosporum which contains over 150 species, spread from Hawaii to the
Canary Islands. There are perhaps 25 species in tropical Africa, 10 of which
are in Madagascar; ail of them, but 1, endemics.
This family offers a remarkable case of biogeographical meaning. The dis-
tribution pattern of the genus Pittosporum (Cufodontis, 1960), if regarded as an
isolated genus, is almost meaningless, but in relation with the whole of its family,
it becomes probable that its distribution center must have been Australasia, per-
haps in the Upper Cretaceous or later when Australia and Africa were still in close
proximity. However, the relationships of the family being somewhat problem-
atic, it is difficult to know for certain where it originated.
Crassulaceae. This is a family of wide, cosmopolitan distribution consisting
of 35 genera and 1,500 species, often growing in dry regions, chiefly in S. Africa
and the Mediterranean region, but also in Asia and Central America. It is repre-
sented in Madagascar by 3 genera: Crassula (5 ssp. ail of them endemic),
Sedum (1 sp., endemic) and Kalancho� (57 spp. of which 55 endemic). There
is no endemic genus. Affinities are strongly African (P. Boiteau, pers. comm.).
Saxifragaceae. For want of being revised this family proved difficult to in-
terpret. Many genera formerly included in it (Brexia, Roussea, Forgesia, Grevea,
Montinia) are now either incorporated with Escalloniaceae (Hutchinson) or
distributed among small new families. Given our present knowledge, the
latter alternative seems the best one.
Brexiaceae is a small family of 2-3 genera with only a few species: Brexia
in Madagascar and the Seychelles (1-2 spp., 1 of which is also in E. Africa);
Roussea (1 sp. in the Mascarene); Ixerba, the affinities of which are uncertain,
in New Zealand.
Montiniaceae is a family of tropical Africa and South Africa consisting of
two very distinct gencra (each of them should be raised to the subfamily
level): Montinia (1 sp. in South Africa) and Grevea (3 spp., one of which
reacles the western parts of Madagascar). The recently discovered Grevea
bosseri Letouzey by P. Sita in Za�re (Letouzey, 1977a) confirms Africa as
probable center of origin for Montiniaceae and extends the area of the remark-
able genus Grevea. It is noteworthy that Grevea madagascariensis is present
on Madagascar, as, conversely, Brexia imadagascariensis is on the African coast:
the Mozambique canal has been a poor gap between Africa and Madagascar.
Escalloniaceae. Following Airy Shaw (1966), I maintain, at least provision-
ally, the genus Forgesia in this ill-defined family. This genus is monotypic and
endemic to the Mascarene Islands.
Rosaceae is a family of wide distribution which contains 124 genera and over
[VOL. 65
558


LEROY-MADAGASCAR
3,300 species. It is represented in Madagascar by only 6 indigenous genera, one
of which, Grangeria, is endemic. Besides the genus Rubus, they belong in
two tribes: first, Pygeum (3 spp., 1 endemic), second, Parinari (a pantropical
genus with 2 spp. endemic to Madagascar), Licania (a large austral genus with
1 Madagascan nonendemic sp.), Hirtella (another large austral genus, with 4
sp. endemic to Madagascar, 2 of which described by Capuron, 1972b), Crangeria
(a monotypic and endemic genus).
Although very poorly represented, the family Rosaceae is interesting in
that its links with Africa, are well marked by the genera Hirtella and Grangeria,
which belong in one and the same tribe.
Leguminosae (Caesalpiniaceae, Mimosaceae, Papilionaceae). The revision
of the Madagascar Leguminosae by M. Peltier is in progress but still far from
complcte. So, it is impossible to form an accurate opinion about Madagascan
elements. In referring to work by Capuron (1967a, 1968b, 1970b), Viguier
(1944), and Peltier (pers. comm.), there would be approximately 98 genera
and 600 species. Since the Humbert's (1959) essay, Capuron (1967a) has de-
scribed 2 new genera of Caesalpinioideae: Elignocarpus and Mendoravia, bring-
ing the number of endemic genera to 11 out of 25. This is a very strong generic
endemism. And this fact concerning the most primitive subfamily seems ail
the more important as the number of endemic genera is proportionally inferior
in Papilionoideae and reduced to one in Mimosoideae. At the specific level, en-
demism is alout 50% in Papilionioideae, 70-75% in both of the other subfamilies.
One genus, Bremontiera (1 sp.), is endemic to the Mascarenes.
Podostemaceae (40 gen., 200 spp.) (according to C. Cusset, 1974, oral comm.).
This family taken sensu stricto is reprcsented in Madagascar by 4 genera, only
1 of which nonendemic: Sphaerothylax (3 spp.: 1 confined to Cameroon; 1 to S.
Africa, reaching Angola and Zambia; 1 in E. Africa, Ethiopia and Madagascar).
The 3 endemic genera are: Thelethylax (2 spp. endemic), closely allied to
Ledermanniella, an African genus of 47 species, particularly abundant in
Cameroon-Gabon; Endocaulos; and Palaeodicraeia. These last 2 genera are
morphologically noteworthy with their deciduous leaves, the blade falling off
and the persistant sheath becoming thick. This character unknown in African
Podostemaceae exists in Asia ( Hydrobrium, Zeylanidium). With respect to
other characters, affinities are likewise with Asia. There is no link with the
American element (C. Cusset, pers. comm.).
Tristichaceae, a family often merged into Podostemaceae, consists of 4-5
genera: 1 in Amazonia, 2 in Asia, 1 pantropical (with several spp., of which 1,
Tristicha trifaria, in Madagascar and S. Africa; another, T. alternifolia in Mada-
gascar and Africa [and America?]; 1, T. hypnoides in South and Central America;
1 in Australia-possibly a new genus) (C. Cusset, pers. comm.).
This distributional pattern, with 3 endemic genera in Madagascar, seems to
indicate that the Podostemaceae could have had a Gondwanian origin by the
close of the Cretaceous times.
Haloragaceae (6 gen., 120 spp.) (cf. Humbert, 1959).
Lythraceae. A family of herbs (often hygrophilous) and shrubs widespread
throughout the World containing 25 genera and 550 species. It has 6 indigenous
559
1978]


ANNALS OF THE MISSOURI BOTANICAL GARIEN
genera in Madagascar, one of which is endemic, and 20 species, half of which
are endemic. One genus, Tetrataxis (1 sp.), is endemic to the Mascarenes.
No clear meaning is to be drawn from the presence in Madagascar of wide-
spread genera such as Nesaea, Rotala, Ammania, whose 10 or so species are
nevertheless endemic to Madagascar. In contrast, we have to focus interest on
Woodfordia and Pemphis which are small palaeotropical genera, each of them
consisting of 2 species: Woodfordia with 1 African species and a widespread
one ranging from Africa and Madagascar to Indonesia; Pemphis with 1 species,
littoral, reaching the Pacific Islands, and the other endemic to Madagascar.
Moreover, a monotypic genus, endemic to Madagascar, Capuronia, was described
by Lourteig (1960); it is the single dioecious Lythraceae.
Sonneratiaceae. This small family of 2 genera and 7 species is of special in-
terest in that it is centered in South Asia, although the mangrove genus Son-
neratia reaches the African coast with one species. But this very distribution
provides evidence in favor of links between the Madagascan region (including
E. Africa) and E. Asia, to the exclusion of the West African coast. It is one of the
characteristic elements of the floristic oriental belt.
Thymelaeaceae (Aquilarioideae included) (50 gen., 600 spp.). Mainly in
temperate regions (Mediterranean region, S. Africa, Australia . . .), this family
is very akin to the families Penaeaceae (5 gen.) and Geissolomataceae both of
of them endemic to S. Africa. There are 7 genera in Madagascar, 2 of which
palaeotropical (Lasiosiphon, Gnidia). Four genera are common to Madagascar
and Africa: Peddiea, 1 endemic species in Madagascar; Silnaptolepis, 1 endemic
species in Madagascar; Dais, 1 endemic species in Madagascar, 1 species in S.
Africa; Octolepis, 3 species, morphologically isolated and primitive (with a
perianth deprived of tube), represented in Madagascar by 1 species (which
will have to be raised at the subgeneric level), which is dioecious and endemic.
One genus, Stephanodaphne, is composed of 9 species, 1 of which is endemic
in the Comores.
The family, therefore, is strongly attached to Africa. According to L�andri
(1947) the genus Synaptolepis is related to the American genus Lophostoma.
The family is an heterogenous assemblage divided into several phyla, in particular
the Synaptolepis one, the Stephanodaphne one, the Peddiea one.
Myrtaceae (100 gen., 3,000 spp.). This is a family mainly developed in
Australia and tropical America, but with a monotypic, endemic subfamily
( leteropixioideae) in S. and S.E. Africa. There are 3 indigenous genera in Mad-
agascar: Myrtus with 2 endemic species, Eugenia (35 spp.), and Syzygium (20
spp.). Many species also occur in the Mascarenes and Seychelles, and one
monotypic family, possibly related to the Myrtaceae, the Psiloxylaceae, is endemic
to the Mascarenes.
Melastomataceae (250 gen., 4,000 spp.). The Madagascan Melastomataceae
being in badly need of revision, it is not possible to have any clear view about
their composition and affinities. According to Perrier de la B�thie (1951), there
are 10 indigenous genera, 5 of which endemie. Since then, the genus Rousseauxia
has been restated (Jacques-F�lix, 1973) and the geographical distribution of
two genera given as Madagascan has been altered.
560
[VOL. 65


LEROY-MADAGASCAR
In any case, there are 4 genera endemic to Madagascar-Mascarenes: Rhodo-
sepala (1 sp.), Dionycha (3 spp.), Amphorocalyx (5 spp.), and Rousseauxia
(13 spp.), and 5 common to Africa: Gravesia (100 spp. in Madagascar, 7 sp.
reaching E. Africa and some spp. in tropical Africa), Dichaetanthera (20 spp.
in Madagascar, 1 endemic to Africa), Tristemma (15 spp.), Anthostemma (1 sp.
in Africa, 1 in Madagascar), Dissotis (2 African spp.). The other two genera are
palaeotropical: Medinilla (70 spp. in Madagascar) and Memecylon (56 spp. in
Madagascar).
Combretaceae. As stated by Exell (1972) and Exell & Stace (1966), this family
is composed of 19 genera (475 spp.): 13 in Africa-Madagascar, with 7 endemic,
3 palaeotropical, 2 widespread, 8 in Asia with 2 endemic, 8 in America with 4
endemic, 3 in Australia with 1 endemic. There are overall 14 genera endemic
to any region, 2 widespread genera, and 3 mangrove genera, and the family is
centered in Africa (Liben, 1971). There are 62 species in 5 genera in Madagascar
(Capuron, 1973a), 2 of which have a pantropical distribution: Terminalia
(50 spp. in Malaysia, 35 in Madagascar, 30 in Africa, 30 in America, only some
in Australia) and Combretum (100 spp. in Africa, 35 in America, 35 in Asia, and
only 4 in Madagascar). Meiostemon is an African-Madagascan genus of 2 spe-
cies: 1 species is endemic to Madagascar, and 1 species is endemic to the
Sudano-Zambezian area. The endemic genus Calopyxis, which is characterized
by lacking petals and by the corollalike structure of its receptacle, has as many
species as 22, which is a very high number proportionally to the area. Madagas-
car, like Amazonia, and the Malaysian Islands, is a center of strong speciation
(Exell & Stace, 1966).
Some facts must be noted. First, the active speciation of 2 genera (Terminalia,
Calopyxis) might have started after Madagascar separated from Africa. One
can add that Lumnitzera, the fifth genus in Madagascar (1 mangrove sp.), is a
strictly palaeotropical genus. Ail of tribes and subtribes are represented in, one
of which (Pteleopsidineae) is endemic to, the African-Madagascan area. A sub-
family, Strephonematoideae (raised by some authors-Venkateswarlu & Rao,
1971-to familial level) containing 7 species, is endemic to Africa.
Proteaceae (75 gen., 1,000 spp. according to Johnson & Briggs, 1975). This
family is mainly of southern tropical distribution with two major centers: Aus-
tralia (45 gen. with 35 endemic) and South Africa (14 gen.). However, some
genera reach Central America, Ethiopia and China. Johnson and Briggs have
divided the family into 5 subfamilies, 2 of which are present in Madagascar:
Proteoideae (1 tribe, Conospermeae, out of 3) and Grevilleoideae (1 tribe,
Macadamieae, out of 7). Ail of the 14 South African genera, 2 of which (Faurea,
Protea) extend far northward, belong to the Proteoideae.
There are 3 genera in Madagascar with 2 endemic. The monotypic genus
Dilobeia stands apart and represents a subtribe of its own (Conospermeae-
Dilobeiinae). Faurea (18 spp.), a tropical African genus classified in the
Conospermeae-Proteinae, has one species in Madagascar. The third genus,
Malagasia, with only one species, is the only representative of the subtribe
Hicksbeachiinae (Macadamieae) in the African region; its closer ally in the
same tribe could be Heliciopsis the sole genus endemic to Asia to the west of
561
19781


ANNALS OF THE MISSOURI BOTANICAL GARDEN
Wallace's Line. It is noteworthy that another monotypic and isolated genus
in the Macadamieae, Brabeium, inhabits S. Africa.
The Madagascan Proteaceae (besides Faurea) are obviously some rare relicts
of a continuons ancient southern area. Johnson and Briggs think that the
family could have originated in Gondwanaland in the Albian at the latest.
Rhizophoraceae. This family (160 spp.), which includes 18 genera, is basically
palaeotropical. In the Madagascan-African area one finds 9 genera, 4 of which
are endemic: Anopyxis (3 spp.), Macarisia, Comiphyton (1 sp.) (Macarisiae-
Floret, 1974), and Poga (1 sp.) (Anisophylleae). Ail of the tribes are repre-
sented in this area, and the Macarisieae, which reaches America (Sterigmapet-
alum, an endemic genus of 3 spp.) and India (Blepharistemnuw, a monotypic and
endemic genus) has its principal center of differentiation here (Floret, 1976).
There are 7 genera in Madagascar (22 spp. with 17 endemic): 3 mangrove
genera (Rhizophoreae): Ceriops, a genus with 2 species, palaeotropical, 1 of
which, C. tagal, reaches E. Africa, Madagascar, and the Seychelles; Bruguiera,
with 6 species, palaeotropical, 1 of which, B. gymnorrhiza also reaches E. Africa,
the Seychelles, and Madagascar; Rhizophora, a genus of 7 tropical species, 1 of
which, palaeotropical, R. mucronata, occurs in E. Africa, Madagascar, the Sey-
chelles, and the Mascarenes. The palacotropical genus Carallia (Gynotrocheae)
has 10 species of which C. brachiata, reaches the Madagascan rain forests. The
other 3 genera are Macarisia and Cassipourea (Macarisieae) and Anisophyllea
(1 endmnic sp.) (Anisophylleae). The tribe Macarisieae is strongly rooted on
Madagascar with the endemic genus Macarisia (2 spp.) and Cassipourea, a large
genus reaching S. India and S. America, with 10 species endemic to Madagascar.
The family has only 2 gencra endemic to America: Polygonanthus (Anisophyl-
leae) and Sterigmapetalum.
From this distribution, one may suppose that the differentiation of the family
might have begun by the time when America and Africa started separating so
that a small part of the first members of it were trapped in America. Africa-
Madagascar-India might have been the principal primary center of taxogenesis
later in the Cretaceous. In any way, the family is obviously centered around
the India Ocean.
Concerning the position of the Rhizophoraceae in the Myrtales, the recent
work on wood anatomy by van Vlict (1976) suggests the family does not belong
to this order, but rather to Cornales (following the Thorne's [1968] classifica-
tion).
Alangiaceae. This monotypic family which has for a long time been included
in Cornaceae covers a typically oriental area: Vietnam (7 spp.), China (6 spp.),
Sumatra (8 spp.), Borneo (8 spp.), Java (7 spp.), Japan (2 spp.), E. Australia
(1 sp.), New Caledonia (1 sp.). However, the genus Alangium is also repre-
sented in the African-Madagascan area by 3 very distinct species, one of which,
A. griselleoides Capuron, is endemic in Madagascar. Judging from the incom-
plete material at his disposal, Capuron (1962) suggested wrongly that it could
belong to section Conostigma. In all certainty, the Madagascan species is to
be placed with A. chinense and A. barbatum in section Marlea Baill. close to
562
[VOL. 65


LEROY-MADAGASCAR
section Angolum Baill. to which belongs A. salviifolium. The present Indo-
Malesiano-Australian area displays one African-Madagascan branch with 3 spe-
cies. Besides, it is notable that the Madagascan species is dioecious and is the
single one within the genus to be so. In ail probability, this advanced character
must have been acquired after Madagascar separated from Africa, possibly re-
cently. Moreover, the Madagascan species, unlike the species of Conostigma, does
not have a primitive secondary xylem with scalariform perforation plates (Eyde,
1972).
The fossil material of Alangiaceac, Alangioxylon, Alangiopollis (Reitsma,
1970), at our disposal shows that the distribution of the family reached Europe
and N. America during Eocene-Miocene times. For Reitsma, Alangiopollis
eocaenicus Krutzch (Germany, London Clay Flora) could be considered to be
one of the ancestors of the family and, though referring only to pollen, he thinks
A. griselleoides might represent the end product of a line coming from this fossil
taxon through A. barbatum. The placement of A. griselleoides in Marlea supports
this hypothesis.
I agree with Aubr�ville (1976b) that this member of the Madagascan flora
belongs to a Laurasian line. It might have reached Madagascar in the late
Eocene when the Thethys became interrupted, permitting connections between
Africa and the northern continents.
Cornaceae. It is not certain that this family occurs in Madagascar, the two
Madagascan genera formerly attributed to it being now considered to con-
stitute a special endemic family Melanophyllaceae (Airy Shaw, 1972). It is
even possible that this one will have once again to be divided as the morpholog-
ical gap between the two genera appears substantial. Melanophylla has 9 species,
Kaliphora only one.
Olacaceae is a tropical family (24 gen., 230 spp.) that has 4 genera in Mada-
gascar, of which 1 is endemic with 3 species: Phanerodiscus. There is 1 pan-
tropical genus, Ximenia, with 3 species out of 12 over the World, with 2 endemie,
and 1 common to Madagascar and Africa. There are 2 palaeotropical genera:
Olax (50 spp.) with 7 species in Madagascar and the Mascarene Islands of which
6 are endemic and 1 is common to Madagascar and Africa; Anacolosa (16 spp.)
with 2 species in Madagascar (Capuron, 1968a).
Opiliaceae is a palaeotropical family of 8 genera of which 1, Rhopalopilia,
is endemic to Madagascar-tropical Africa: 5 species are endemic to Africa, 2
are endemic to Madagascar, and 1 is common to Madagascar and Africa (Capuron,
1968a).
Santalaceae, Loranthaceae (cf. Humbert, 1959).
Balanophoraceae (18 gen., 45 spp.). This family of parasitic plants has
3 genera in Madagascar, 5 in Africa, 1 of which, Balanophora, is a palaeotrop-
ical genus that occurs in Madagascar and W. Africa. One Madagascan genus,
Ditepalanthus, is endemic and monotypic, the third is Langsdorffia, a genus
classically known as confined to America, but present also in New Guinea (Han-
sen, unpublished paper 1976). Langsdorffia malagasica (= Thonningia mala-
gasica) is a dioecious parasite on roots of Elaeocarpus quadrilobus, an austral
genus absent from Africa.
563
1978]


ANNALS OF THE MISSOURI BOTANICAL GARDEN
What is of particular interest is that 5 genera are endemic to Madagascar-
Africa, distributed in 4 subfamilies of which 2 endemic. Mystropetalon (1-2
spp.) is endemic to S. Africa (according to Airy Shaw, 1966, this genus should
be given family rank). The Sarcophytoideae is composed of 2 genera: Sarcophyte
(2 spp. in E. Africa and Transvaal) and Chlamydophyton (equatorial West
Africa). The subfamily Balanophoroideae has 2 genera in Africa: Balanophora
and Thonningia, an endemic genus. The IIelosioideae, in which the Madagascan
Ditepalanthus is placed, is absent from Africa but has 3 endemic genera in
America (Scybalium, Helosis, Cortnaea) and 3 others in Asia. Ditepalanthus
is related to Rhopalocnemis (Himalaya, Vietnam, Indonesia) (Hansen, unpub-
lished paper, 1976).
It is to be noted that Langsdorff ii malagasica and Ditepalanthus are land-
marks between America and Palaeotropics.
The distributional pattern of the Balanophoraceae is typically austral and
supports the idea that the family must have had a Gondwanian origin.
Hydnoraceae (2 gen., 20 spp.). This is a very interesting bitypic family of
parasites of the Southern IHemisphere restricted to the Madagascan region sensu
lato and America. Ilydnora is composed of about 10 species, one of which is
endemic to Madagascar and represents a section of its own, Neohydnora; the
other species inhabit in S. and E. Africa. The other genus Prosopanche is en-
demie to S. America (Paraguay, Argentina).
The placing Hydnoraceae (Rafflesiales) near Santalales is of special interest
because the Balanophoraceae-Hydnoraceae is a biogeographically homogenous
unit.
Rafflesiaceae (8 gen., 50 spp.). There are 3 endemic species of Cytinus in
Madagascar, a genus confined to Africa and W. Europe.
Celastraceae (llippocrateoideae excluded). This is a family of tropical
and temperate plants dispersed all over the world, although more abundantly in S.
America, and contains 60-90 genera (800-1,000 spp.). According to Lobreau-
Callen's (1975) work, it is, with the Icacinaceae and several small families, a
member of an austral complex. It is of special interest with respect to Mada-
gascar where at least 14 genera occur (4 of them being still undescribed), 10 of
which are endemic: Polycardia, 9 species; Ptelidium, 2 species; Iartogiopsis, 1
species; Salvadoropsis, some species; Brexiella, 2 species; Evonymopsis, 4 species.
The other genera are widespread (Celastrus) or pantropical (Elaeodendron,
Maytenus), or African-Madagascan ( Mystroxylon).
So generic endemism is extraordinarily high, a proportion of 2:3. But this
interpretation is only a provisional account as the family is in badly need of
revision. One monotypic genus (Herrya) is endemic to R�union.
Ilippocrateoideae (N. Hall�, unpublished paper) (24 gen., 320 spp.). This sub-
family is mainly developed in Africa-Madagascar with 17 genera, 9 of them en-
demic, and 150 species. There are only 6 genera (11 sp.) with none endemic in Mad-
agascar: Apodostigma (1-2 spp. common to Madagascar and Africa); Reissantia
(1 sp. in Madagascar, 4 spp. in Africa, 3 spp. in Asia-Melanesia); Loeseneriella
(2 spp. in Madagascar, 12 ? spp. in Africa; 8 ? spp. in Asia-Melanesia); Elachyp-
tera (1 sp. in Madagascar, 3 spp. in Africa, 3 spp. in America); Pristimera (4 spp.
564
[VOL. 65


LEROY-MADAGASCAR
in Madagascar, 5 spp. in Africa, 9 spp. in America); Salacia (2 spp. in Mada-
gascar, 90 spp. in Africa, 30 spp. in Asia-Melanesia, 29 spp. in America).
Salvadoraceae, Aquifoliaceae (cf. Humbert, 1959).
Icacinaceae is a family of pantropical distribution with 58 genera and 400
species. It is represented in Madagascar by 8 genera, of which 4 Madagascan-
African (Cassinopsis, 4 endemic spp., 2 spp. endemic to S. Africa; Leptaulus, 1
endemic Madagascan species, 2 species endemic to tropical Africa; Desmostachys,
1 species in Madagascar, Comores, S. Africa, 2 species endemic to Africa; Raphio-
styles, 6-7 species in Africa, 1 species endemic to Madagascar; 3 palaeotropical
(Apodites, Iodes, Pyrenacantha); 1 endemic, Grisollea, which contains 2 species,
of which 1 is restricted to the Seychelles.
Dichapetalaceae is a small family (3 gen., near 100 spp.) nearly confined
to the tropics. It is represented in Madagascar by 14 species, all of them en-
demic, belonging to genus Dichapetalum, a pantropical one that is by far the largest
in the family, and which is abundantly present in Africa (% of the whole-
Breteler, 1969).
Buxaceae (5 gen., 50 spp.). The plants of this family occur mainly in tem-
perate regions of the northern hemisphere, but also in tropical mountains. The
family reaches the southern hemisphere with 2 genera: Styloceras (tropical S.
America, perhaps a special family) and Buxus sensu lato. The latter is present
in Madagascar with an isolated group of 2 species, B. madagascarica and B.
macrocarpa, which is strongly related to species of E. and S. Africa.
Euphorbiaceae (300 gen., 7,500 spp.). This is a family of almost cosmopolitan
distribution, mainly of the tropics, with two large centers of distribution: tropical
America and Africa. The new taxonomic results that have been obtained for
Madagascan Euphorbiaceae during the last twenty years, part of which arc
not yet published, provide some new insights (L�andri, oral comm.). In par-
ticular, we must cite the Webster's (1975) work, which has modified the former
systematics and suggests new ways of approaching origins and affinities. Fol-
lowing this author, there would be 5 subfamilies in Madagascar: Phyllanthoideae
(7 tribes), Oldfieldioideae (1 tribe), Acalyphoideac (16 tribes), Crotonoideae
(7 tribes), Euphorbioideae (3 tribes). Thus, 34 (out of a whole of 52) tribes
are present in Madagascar. Thanks in particular to Capuron, the number of
Madagascan genera amounts now to 62, 14 of which are endemic. There are
about 600 species, of which nearly 540 are endemic (J. L�andri, oral comm.).
Special mention must be made of the genera Voatarnalo (2 spp., endemic
to Madagascar) and Androstachys which are related to one another and repre-
sent a very singular group (Androstachyaceae Airy Shaw). Androstachys has
several species in Madagascar, one of which (a kind slightly different) is also
present in E. Africa. One monotypic genus, Cordemoya, is endemic to the
Mascarcnes.
This family is well represented in Madagascar, but its revision is not advanced
enough to permit taxogenetic or biogeographical conclusions.
Rhamnaceae. This is a cosmopolitan family of 50 genera and 600 species
which is represented in Madagascar by 10 genera and 22 species (Capuron,
1965a, 1966a). Four genera are pantropical: Colubrina (12 spp., 5 in Madagascar
1978]
565


ANNALS OF THE MIISSOURI BOTANICAL GARDEN
with 1 endemic); Guania (60 spp., 5 in Madagascar with 3 endemic); Scutia
(7-8 spp., 1 common to Madagascar, the Mascarenes, India, S. and E. Africa);
Ziziphus (40 spp., 3 in Madagascar with 1 endemic). Two genera are palaeo-
tropical: Helinum (3 spp., 1 in both Madagascar and S. and E. Africa); Ventilago
(25 spp., 1 common to Madagascar and India). Three genera are common to
Africa and Madagascar: Lasiodiscus (9 spp., 2 of which are endemic to Mada-
gascar-Comores); Phylica (60 spp. in S. and tropical Africa, 1 in Mascarenes, 1
in Madagascar, both endemics); Berchemia (22 spp., 1 of which in Madagascar,
endemic). The 1 endemic genus is Bathiorhamnus (3 spp.).
It is remarkable that the genera are relatively numerous but the species few,
one genus, and 11 species (50%) only being endemics. Only one genus (Bathio-
rhamnus) is really Madagascan, the others (pantropical, Asian or African), it
seems, have come there accidentally.
Leeaceae is a palaeotropical monogeneric family of 80 species with 2 species
in Madagascar: one of which endemic, the other being African-Madagascan
(Descoings, 1967).
Vitaceae. This family of about 12 genera and over 800 species, which ranges
throughout tropical and warm temperate regions, is represented in Madagascar
by 51 species belonging to 5 genera: 3 large pantropical ones (Ampelocissus,
Cyphostemma, Cissus) with 47 species in all of them and but 1 of these species
endemic; 1 palaeotropical (Cayratia) with 3 endemic species; 1 (Rhoicissus)
with 10 species, primarily of E. Africa, but 1 of them reaching the Comores.
There are no endemic genera (Dcscoings, 1967).
Connaraceae. This family of 25 genera and 200 species is of a particular
interest because it principally covers a palaeotropical area and its main center
of distribution is the African-Madagascan area where 16 genera occur, 11 of
which are endemic. It is, however, poorly represented in Madagascar where
only 5 genera occur, none being endemic. Some aspects, however, call for
special mention: 1 African genus, Byrsocarpus, occurs there with 1 endemic spe-
cies and 1 species common to Africa. There are 3 palaeotropical genera: El-
lipanthus (1 sp.), Agelaea (3 spp.), Cnestis (4 spp.), and a pantropical one,
Rourea = Santaloides (2 spp.). Ail of the Madagascan species are endemic.
Sapindaceae. This tropical and subtropical family is composed of woody
plants which number 2,000 species distributed in 150 genera, 28 of which, with
89 species, are present in Madagascar, the Comores, and the Mascarenes
(Capuron, 1969). Nine genera are endemic (Conchopetalum, 2 spp.; Plagioscy-
phus, 10 spp.; Tsingya, 1 sp.; Beguea, 1 sp.; Chouxia, 1 sp.; Tinopsis, 10 sp.;
Pseudopteris, 2 spp.; Tina, 6 spp., Neotina, 2 spp.); Doratoxylon with 5 species
occurs in the Mascarenes; Hornea with 1 species is endemic to the Mascarenes;
Molinea with 6 species has 3 of them in the Mascarenes. One genus, Stadmannia,
reaches the Mascarenes and E. Africa. Three genera are common to Madagascar
and E. Africa: Erytrophysa, Macphersonia, and Camptolepis. Five genera
are widespread in Africa: Zanha, Majidea, Deinbollia, Crossonephelis, Haplo-
coelum. Two genera arc palaeotropical: Filicium, Aphania. Three genera are
pantropical: Allophylus, Dodonea, Cardiospermum. One genus which is mostly
566
[VOL. 65


LEROY-MADAGASCAR
American, but also occurs in Africa, has 1 nonendemic species in Madagascar:
Paullinia pinnata (probably recently introduced). Endemic to the Madagascan
region are 79 species, and 4 species are endemic to Madagascar-E. Africa. It must
be noted that the pantropical genus Dodonaea, highly developed in Australia, has
a species endemic to Madagascar, D. rnadagascariensis. Cossinia (4 spp.) is only
found in New Caledonia and Madagascar (Good, 1950).
The Madagascan Sapindaceae of which a detailed analysis on a biogeograph-
ical basis remains to be done seem to have no direct link with the American
flora.
Burseraceae. This is a tropical, woody family of 20 genera and 550 species
which is represented in Madagascar by 4 genera: Canarium (a palaeotropical
genus of 100 spp.) with 2-3 endemic (?) species. Protium (a pantropical genus
of 85 spp., mostly in tropical America) with 1 species in Mauritius and 1 species
in Madagascar, these 2 species constituting a special section (Capuron, 1968c).
Commiphora and Boswellia, with more than 200 species, are endemic to Africa
and Madagascar occurring in arid or rather dry regions, particularly in north-
eastern Africa. There are more than 40 species of Commiphora in Madagascar,
and only 1 of Boswellia: B. madagascariensis Capuron. This species must be
considered as the type of a remarkable subgenus derived from African stock: it
has unisexual flowers and valvate petals (besides, its stigma is sessile). The
Madagascan Burseraceae are strongly linked to the African flora.
Anacardiaceae is a chicfly tropical family of 600 species distributed in 60
genera. It is represented in Madagascar by 9 indigenous genera, of which 3 are
endemic: Operculicarya (3 spp.), Faguetia (1 sp.), Micronychia (5 spp.). There
are 2 African-Madagascan genera: Protorhus, 16 species with only 1 outside
Madagascar and endeiic to South West Africa; Sorindeia, 1 species endemic
to Madagascar, 1 Madagascan-African species. The other genera are: Gluta, 1
endemic species, 12 Indomalesian species; Campnosperma, 15 tropical species;
Poupartia, 5 Madagascan species, 7 in Africa, 1 in India. Poupartia caffra is with
Ma ytenus sp. a characteristic element of African-Madagascan savannas.
Simaroubaceae (30 gen., 200 spp.). This is a family of a special interest in
the Madagascan region, with 5 genera, 2 of which are endemic. A closely related
(if distinct) family, Ptaeroxylaceae, also occurs in Madagascar with 1 genus, while
the other genus occurs in South Africa. The main center of distribution of
Simaroubaceae is in tropical America, but there is a secondary center in West
Africa (Nooteboom, 1962). The Madagascan genera are: Perriera, 2 endemic
species; Pleiokirkia, 1 endemic species. The other 3 genera are represented
each by one species: Quassia indica, Suriana maritima and, in the Seychelles,
Soulamea terminalioides, an endemic species. This is a surprising distribution as
Soulamea has 6 species endemic to New Caledonia, 1 in Fiji, whereas another
ranges from Polynesia westwards to Borneo. As noted by Nooteboom "Suriana
and Soulamea, both belong to the Barringtonia formation. The fruits are cer-
tainly dispersed by seawater and this seems to have been very effective." Quassia
indica is a palaeotropical species dispersed throughout Asia to Australia.
The Madagascan Simaroubaceae have no link with America, a meaningless link
with Asia, but a strong link to the African center.
567
1978]


ANNALS OF THE MISSOURI BOTANICAL GARDEN
Ptaeroxylaceae (2 gen., ca 4-5 spp.). This family, possibly a subfamily of
Simaroubaceae, is special to the Madagascan region sensu lato: Ptaeroxylon is
a monotypic genus endemic to E. and S. Africa. Cedrelopsi (3-4 spp.) is en-
demic to Madagascar (Leroy, 1959).
Rutaceae. This large family (150 gen., 1,000 spp.) is rather poorly represented
in Madagascar and the Mascarenes with 9 indigenous genera, of which only 1 is
endemic, and 65 species, but it offers some patterns of distribution of great in-
terest. Since 1950, owing to Capuron's (19671)) work, our knowledge about the
Madagascan Rutaceae has much advanced. In particular, he found 3 genera new
for the Madagascan flora, 1 of which, Evodia, is endemic and contains 10 species.
The remarkable genus Chloroxylon, considered until now as Asian, and monotypic,
is present in Madagascar with 2 endemic species (Capuron, 1961). Thus, the
subfamily Flindersioideae, composed of both Chloroxylon and the New Caledo-
nian genus Flindersia, extends to Madagascar. This is a distribution that seems
strongly relict and suggests the possibility of the existence of direct migration be-
tween Madagascar and New Caledonia via India, possibly through rafting.
Toddalicae is represented not only by a meaningless species of Toddalia,
but also by the genus Fagaropsis, which contains 5 species: 3 in tropical Africa
and 2 in Madagascar. It is closely related to the Asian Phellodendron and this
pattern of distribution is to be classified with the precedent one (Capuron, 1961).
Regarding the Xanthoxyleae, we have to put aside the pantropical genus
Xanthoxylum, although it is present in Madagascar with 8 endemic species. The
remaining genera, Evodia and Ivodea have an interesting distribution. The large
palaeotropical Evodia (130 spp.), which is absent from Africa, ias 19 endemic
species in Madagascar, and Ivodea is a Madagascan endemic (Capuron, 1961).
This pattern of distribution is comparable to other two reported above.
The remainder, Vepris (40 spp.), Diphasia (5 spp.), and Teclea 25 spp.)
represent a very natural group nearly confined to Madagascar-Africa, except for
1 Vepris species that extends to India. This genus is particularly well represented
in Madagascar and the Mascarenes with some 30 species, while Teclea has only
1 species in Madagascar. Diphasia consists of 2 species, 1 African, 1 Madagascan.
Meliaceae (51 gen., 550 spp.-Leroy, 1977c). This is a pantropical family
of woody plants divided into 5 subfamilies, all of them represented in Madagascar,
and 3 of them monotypic and restricted to the "Grande Ile": Quivisianthoideae,
Capuronianthoideae, Neomangenotioideae. The Melioideae is richly repre-
sented by the Trichilieae and Turraeeae, but 5 tribes out of 7 are lacking. The
Trichilieae is composed of 10 genera, 9 of which are palaeotropical (3 re-
stricted to Asia, 4 to Africa sensu stricto); there are 4 genera in Madagascar:
Malleatrrum (12 spp.) endemic to Madagascar-Comores-Aldabra; Astrotrichilia
(14 spp.) endemic to Madagascar; Lepidotrichilia (4 spp.) restricted to Mada-
gascar and E. Africa; Trichilia (50 spp. in America, 14 spp. in Africa, some spp.
in Madagascar and the Indo-Malayan region). The palaeotropical tribe Tur-
raeeae is best represented in this part of the world: only 1 genus (Munronia)
is endemic to Asia; Naregamia has 1 species in India, 1 in Angola; Turraea has
24 species in tropical Africa, 30-40 species in Madagascar, Mauritius and the
568
[VOL. 65


LEROY-MADAGASCAR
Comores, and about 6 species in tropical Asia and Australia. There are 2 gen-
era endemic to Madagascar: Humbertioturraea (6-7 spp.) and Calodecaryia
(1 sp.) and 1 genus is endemnic to S. Africa: Nymania (1 sp.). The Swietenio-
idcae-Swietenieae (9 gen.) is represented by genera in Africa and Madagas-
car, all of the endemic: Khaya, Entandophragma, Pseudocedrela and Lovoa in
Africa sensu stricto, Khaya and Neobeguea in Madagascar; there are 2 genera
in Asia-Malesia: Soymida, Chukrasia, and 2 genera in America: Swietenia,
Sclhmardea. The Xylocarpeae is composed of 2 genera: Xylocarpus (2 spp.),
a genus of mangroves and swamps confined to Indian Ocean coasts, included
in the Madagascan region, and one African-American genus, Carapa (2-3 spp.)
(Pennington & Styles, 1975; Leroy, 19771)).
To conclude about Meliaceae, strong endemism exists, at the subfaminial and
generic levels in Madagascar and at the generic level in Africa. The Meliaceae
is an ancient family represented in Africa sensu lato by 24 genera-that is
nearly 50% of the whole family-18 of which are endemic. One genus (Lepido-
trichilia) is common to Madagascar and E. Africa. The tribe Turraeeae seems
centered on the Madagascan region; here Turraea has its highest number of spe-
cies, and 3 genera are endemic to this region sensu lato: 2 to Madagascar, one
to S. Africa. Furthermore, there is a link between Capuronianthus and Xylo-
carpeae.
Zygophyllaceae is a family of 25 genera and 200 species (Hutchinson, 1967),
is mostly woody, and is distributed in dry regions of the two hemispheres, often
in semideserts. It is represented by 2 genera in Madagascar: Tribulus, a widely
spread genus, which there has 2 nonendemic species; Zygophyllum, a large genus
confined in the Old World, especially in southern Africa, with 2 species endemic
to Madagascar, one of them Zygophyllum madagascariense (Baill.) Stauffer,
with a peculiar drupaceous fruit, can be considered as the type of a special
subgenus. The two species occur sympatrically in the South Domain, generally
on sands next to the sea.
Erythroxylaceae is a family of 3(-4) genera, the largest of which is Erythro-
xylum with about 200 species. This pantropical genus has about 30 species in
the Madagascan region, ail endemics. The other genera are: Aneulophus (1 sp.)
and Pinacopodium (2 spp.) in tropical Africa and Nectaropetalum (perhaps a
special family) with 9 species, 5 of which are endemic to Madagascar and still
undescribed, the other 4 being endemic to Africa. Nectaropetalum, whose
Madagascan presence is known thanks to Capuron's work (1963c), is a re-
markable plant with imbricate and persistent sepals and dry and dehiscent
fruits. It occurs in the dry regions in the north and northwest parts of the western
Domain in Madagascar. Aneulophus is sometimes put in the Linaceae.
Lepidobotryaceae. Following Hutchinson (1967), this palaeotropical fam-
ily is composed of 3 genera: Sarcotheca (10 spp.) in tropical Asia, Dapania (3
spp.) in Malaya and Madagascar (Veldkamp, 1967); Lepidobotrys, in Africa
(L�onard, 1950). Thus, it presents an interesting distribution, but its natural-
ness is questioned (Airy Shaw, 1966; Veldkamp, 1967). Whatever it may be,
the Madagascan plant, Dapania pentandra, is a very interesting species sharply
569


ANNALS 0F THE MISSOURI BOTANICAL GARDEN
distinct, with flowers perfect, 5-staminate, and deprived of interstaminate scales
(Huynh, 1971). According to Raven & Axelrod (1974), Dapania might have
reached Madagascar by long-distance dispersal around the Indian Ocean. Let
us note that this hypothesis, rather convincing as regards a genus whose fruits
are readily dispersed, is consistent with a great antiquity for the Madagascan
species.
Geraniaceae. Following Humbert (1959), this family is represented by
3 nonendemic genera: Geranium (2 spp. of the African section Simensia, 1 of
which endemic); Pelargonium (1 endemic sp. of the S. African section Ciconium;
1 endemic species of the section Peristera, which extends from S. Africa to Aus-
tralia, New Caledonia); Monsonia (an African genus with 1 species in secondary
vegetation).
Oxalidaceae. This unrevised family comprises 2 nonendemic genera and 21
species (according to Baron, quoted by Humbert): Biophytum and Oxalis.
Linaceae is represented by 2 nonendemic genera: Linum (2 endemic spp.)
and the palaeotropical genus Hugonia (5 endemic spp.).
Balsaminaceae is a family of 2 genera, 1 of which, Impatiens, occurs abun-
dantly in Madagascar where 100 species or so have been recognized.
Ixonanthaceae, considered by many authors as a subfamily of Linaceae, is
present in Madagascar with an endemic species of the genus Allantospermum
(= Cleistanthopsis), which is bitypic, the other species occurring in Borneo.
These species which were describcd independently in the saine year 1965 by
Forman for Borneo and Capuron for Madagascar, are both very similar to one
another and well distinct: each of them must be considered as a subgenus of
its own.
It is surprising enough that the palaeotropical family Irvingiaceae consisting
of 3 genera, 1 of which, Irvingia, occurs in Africa and Southeast Asia, the other
two being endemic to Africa, is absent from Madagascar. But the related
genus Allantospermum has a distribution parallel to that of Irvingia. Plants of
the group Irvingiaceae-Ixonanthaceae are a fundamental link between Africa-
Madagascar on one hand an Southeast Asia on the other. Hence, I fully agree
with Raven & Axelrod (1974) that it must proceed from "an west Gondwanaland
plexus."
Malpighiaceae. This family of woody plants consists of 5 tribes, 65 genera,
and 1,200 species (Hutchinson, 1967), mostly of South America. Some outliers,
perhaps 15 genera or so, occur far from America. There are 10 genera in Mada-
gascar: Acridocarpus (30-50 spp.) extends from tropical and S. Africa, Arabia,
Madagascar to New Caledonia. Caucanthus is an African-Madagascan genus
of 25 species. The endemic genera are: Trichomariopsis (1 sp.), Philgamia (3
spp.), Calyptostylis (1 sp.), Microsteira (28 spp.), Rhynchophora (1 sp.), Digoni-
opterys (1 sp.). Sphedamnocarpus is a Madagascan-African genus (8 sp. in S.
and E. Africa, 12 sp. in the Madagascan region). Tristellateia is a palaeotropical
genus (40 spp., 20 spp. in Madagascar, 1 in E. Africa, none in Australia). Ail
these genera belong to two tribes: Hiraeeae, which contains 20 genera, 10 of
which are endemie to S. America, and Banistereae (15 gen.). If we consider the
570
[VOL. 65


LEROY-MADAGASCAR
wholly family, 48 genera are endemic to America, 3 to Africa, 3 to Asia, and 6
to Madagascar. Ail Madagascan genera are either endemic or palaeotropical.
On the basis of the present evidence one can retain the idea that a post-
Oligocene improverishment intervened in Africa and perhaps in Asia to eliminate
part of Malphigiaceae, Madagascar playing its well-known role as a center of
survival.
Vochysiaceae and Trigoniaceae. Plants of this group are chiefly represented
in South America with 8 endemic genera and many species. But 1 genus,
Erismadelphus, is present in tropical Africa (with 2 spp.), 1 in Madagascar
Humbertiodendron (1 sp.), and 1 in S.E. Asia Trigoniastrum (1 sp.). Judging
from this pattern of distributions with wide disjunctions and unbalanced pro-
portions concerning the number of species, we may suggest that Vochysiaceae
originated in Gondwana, the extant, extra-American genera clearly being relicts.
Polygalaceae. This family with a wide and mainly pantropical distribution
is an ancient one consisting of 17 genera, 50% of which are endemic to America;
it is poorly represented in Madagascar with only 2 genera: 20 endemic species
of Polygala and 1 endemic species of Carpolobia, a Madagascan-African genus
of 7 species (Breteler, 1969).
Araliaceae is a family (still unrevised for Madagascar) of 600 species dis-
tributed in 50 genera, chiefly tropical, with two principal centers in Southeast
Asia and America. Its representation in Madagascar and the Mascarenes and
the Seychelles is very important and of special interest, with 9 genera, including
4 endemic, and over 50 species. Hutchinson considers the family to be composed
of 7 tribes, 4 of which are present in Madagascar: Cussonieae: Cussonia (25
spp. in tropical and S. Africa, Madagascar, Comores, Mascarenes), Botryopanax
(endemic to the Mascarenes with 6 spp.), Cuphocarpus (a monotypic genus
endemic to Madagascar). Panaceae: Polyscias (palaeotropical genus of 53
spp.; 8 spp. in Africa, 35 spp. in Madagascar-Comores, 8 spp. in the Mascarenes,
1 sp. in the Seychelles, 1 sp. in Ceylon); Gastonia (10 spp. in tropical and E.
Africa, Madagascar, Mascarenes, Borneo, New Guinea . . .); Schefflera (a pan-
tropical genus with 15 spp. in Madagascar-Comores, 13 spp. in Africa, 1 in the
Seychelles, some spp. in Ceylon); Sciadopanax (a monotypic genus endemic to
Madagascar); Geopanax (1 sp.) endemic to the Seychelles. Hedereae: Neocus-
sonia (5 spp. in tropical and S. Africa, Madagascar). Plerandreae: Indokingia
(a monotypic genus endemic to the Seychelles).
Several facts have to be emphasized: (1) the distribution of the genera
into 4 tribes, of which 2 have a single representative. (2) The pattern of en-
demism with one monotypic genus special to the Seychelles, and one polytypic
to the Mascarenes. (3) The very large number of Polyscias species in the Mada-
gascan region. (4) The absence of Schefflera species in the Mascarenes, while
numerous in Madagascar. (5) Two-thirds of the genera are African-Madagascan.
These facts, some of which remain rather enigmatic, support the hypothesis
of the Madagascan region not only as a center of refuge but at the same time
as an active center of speciation. Moreover, the family as a whole is well
rooted in the African region.
Umbelliferae (275 gen., 2,800 spp.). This family is represented in Madagascar
571
1978]


ANNAIS 0F THE MISSOURI BOTANI(AL GARDIEN
by only 9 genera, 2 of which are endemic: Anisopoda (1 sp.) and Phellolophium
(1 sp.); 1 is African-Madagascan, Ileteromorpha (8 spp. endemic to Madagascar
and a few in Africa); 5 are widespread: Carum, Pimpinella, Caucalis, Sanicula,
and Peucedanum. There are 22 species of which 19 are endemic (Humbert, 1959).
The interesting recent discovery of a Lilaeopsis species in Madagascar, Raynal,
(1977) is surprising enough. This genus, which until now had 26 species with
16 American and the others from Australia and New Zealand, is a typically
Conidwanian genus, which is absent from Africa. So a new problem is raised
which is worth thinking about and analyzing.
VI
Loganiaceae. This family (as circumscribed by Leenhouts, [1962] of 26
[28] gen., 600 spp.) is represented in Africa and Madagascan by 11 genera, 5 of
which are endemic. The Madagascan representation is important with 6 tribes
(the family consisting of 7), 8 genera, and 70 species.
The genus Anthocleista (Potalieae) is restricted to tropical Africa and Mada-
gascar with 14 species (4 in Madagascar), but its nearest ally is Fagraea from
Asia and the Pacific. Nuxia (Buddlejeae) consists of 15 species in Africa (Leeu-
wenberg, 1975a), 11 species in the Madagascan region, of which 10 are endemic
(including 1 in the Comores and 1 in the Mascarenes); its closest affinities are
with the monotypic genus Peltanthera from Central and South America, and also
with the monotypic genus Andro ya, endemic to Madagascar (Leenhouts). Bud-
dleja, a large genus of more than 100 species (belonging to the saine tribe),
widespread in the tropics, except for Australia and the Pacific, is highly differen-
tiated in Madagascar (8 spp., 1 of which is endemic to the Reunion) (Leeuwen-
berg, 19751b). Thc trile Loganieae from Malaysia and the Pacifie, reaches the
Mascarenes with 2 endemic species of Ceniostoma: G. borbonicum and G. pedun-
culatum. The small genus (6 spp.) Mitreola (= Cynoctonum) ( American-Asian
tribe, Spigelieae), which bas a very extended and discontinuous area, is present in
Madagascar with 1 endemic species, M. turgida Jovet. Three other species are con-
fined to Asia, 2 of which arc local endemics (Vietnam, Sabah); 1 species is en-
demic to the southern United States, and the sixth, M. petiolata, occurs in N.
America, Brazil, Asia, Australia, and West Africa (Leeuwenberg, 1974). One
African species of Mostuea (;elsemieae), a genus of 8 species, 7 of which are in
tropical Africa, occurs in Madagascar; the eighth species is known from Brazil. The
eighth genus, Strychnos, is widely distributed in the tropics and subtropics with
about 200 species, 15 of which are in Madagascar.
Ail these facts form a complicated whole from which several ideas can be
drawn. First, Loganiaceae is an old family with a worldwide distribution in the
tropies and subtropics of both hemispheres. Second, we have to take care not to
explain every case in terms of the old and discontinuons distribution of the
family. There is one endemic species of Mostuea in Brazil, two endemic species
of Geniostoma in the Mascarenes, and some vast gaps in the area of Mitreola
species, but we think no general and single explanation of this is to be sought.
With respect to Geniostoma species from the Mascarenes, one must have in
572
[VOL. 65


LEROY-MADAGASCAR
mind that an extremely close ally, the genus Labordia, with as many as 20 spe-
cies, is endemic to the relatively young island Hawaii.
Gentianaceae (80 gen., 900 spp.). This family is represented in Madagascar
by 9 genera (50 spp., : of which endemic), of which 2 are endemics: Tachiadenus
(10 spp.) and Gentianothamnus (1 sp.). The genus Chironia is endemic to
Africa-Madagascar. The other genera are palaeotropical (Exacum, 15 spp.,
Sebaea, Canscora) or widely spread (Swertia, Eniscostemma, Neurotheca). Let
us note that the genus Neurotheca is present in the Madagascan swamps, with
the widespread species, N. loeselioides (Guyana, Brazil, tropical and E. Africa)
(Raynal, 1968).
Menyanthaceae (5 gen., 33 spp.). The only genus present in Madagascar
is Nymphoides with 3 endemic species.
Lentibulariaceae (4 gen., 170 spp.) is a family of insectivorous water herbs
widely distributed over the Globe. It has 2 genera in Madagascar: Utricularia
(13 spp., 10 ? endemic) and Genlisea (an African-American genus with 4 spp.
in Madagascar, 3 endemic). Because these plants are readily dispersed, they do
not provide good material for biogeographical analysis.
Apocynaceae (250 gen., 2,000 spp., according to Markgraf, 1976). This is a
family of wide distribution but chiefly pantropical, which is composed of 4
subfamilies, all of them represented in Madagascar. There are 23 genera in
Madagascar (10 endemic) and 146 species (140 endemic). One genus, Oistan-
thera is endemic to the Mascarenes. The palaeotropical genus Carissa (Plume-
rioideae-Carissinae) has 9 species in Madagascar (7 endemic). In the same
subfamily, the Carisseae-Landolphiinae is mainly represented by the pantrop-
ical genus Landolphia, which has a total of 60 species (13 endemic) in Mada-
gascar and over 40 in Africa. There are 2 genera in the Rauwolfieae (a tribe
of 12 genera, mostly in Asia)-Rauwolfiinae (a subtribe of 3 genera with a dis-
junct distribution): the pantropical genus Rauwolfia (5 endemie species) and
the endemic genus Cabucala (18 spp.). The Plumerioideae-Alstonieae is in-
terestingly represented by 4 subtribes: the endemic monotypic genus Craspido-
spermum is one of 2 genera constituting the subtribe Craspidosperminae, the
other inhabiting Malesia. Stenophagia (Aspidospermatinae, a subtribe of 3
genera: 1 in America, 1 in Africa), with 5 species, is endemic. Catharanthus
(Catharanthinae) is an isolated genus with 7 Madagascan species and 1 species
in India. Two genera out of 4 in the Plectaneineae inhabit Madagascar: Plec-
taneia (12 spp.) endemic; Gonioma with 1 species in S.W. of Madagascar and
1 species in S. Africa.
The Apocynoideae has 4 genera in Madagascar, 1 of which, Roupellina (1 sp.),
is endemic; Mascarenhasia is an isolated genus with 12 species, 1 of which reaches
E. Africa; Pachypodium is an isolated, xerophilous genus with 17 species in
Madagascar (12 endemic) and a few in E. and S. Africa.
Stronger endemism is offered by the Tabernaemontanoideae, which contains
4 endemic genera and another one almost endemie with 1 species also in E.
Africa (Hazunta coffeoides).
A fact of special interest is that two monotypic subtribes of Apocynoideae
1978]
573


ANNALS OF THE MISSOURI BOTANICAL GARDEN
( Mascarenhasiinae, Pachypodiinae) are restricted to the Madagascan-E. African
area.
Asclepiadaceae (one monotypic genus, Trichosandra, is endemic to Mauritius),
Solanaceae, Hlydrophyllaceae (cf. Humbcrt, 1959).
Convolvulaceae (Itumbertiaceae included) (55 gen., 1,400 spp.). This is a
family of wide distribution, but mainly of the tropics and subtropics, that is par-
ticularly alundant in America. There are 11 genera in Madagascar, of which
3 are endemie (cf. Humbert, 1959).
Boraginaceae (100 gen., 2,000 spp.) (cf. Humbert, 1959).
Callitrichaceae (1 gen., 30 spp.). The genus Callitriche is represented in
Madagascar by at least 3 endemic species. None of the African or European
species exists in Madagascar. Two species seem to have some morphological
characters of S. America taxa (H. Schotsman, pers. comm.).
Verbenaceae (91 gen., 3,000 spp.). This family is well represented in Mada-
gascar with 12 indigenous genera and 130 species (123 endemies). There are 2
endemie genera: Adelosa (1 sp.) and Acharitea (1 sp.), the latter being very
close to Nesogenes, which also has 1 endemic species in the Madagascan region
while the others occur in E. Africa and Polynesia. Another genus, Coelocarpum,
has 5 Madagascan and 1 Socrotan species. If we add another genus common to
Madagascar and Africa, Chascanum (30 spp. with 2 endemic in Madagascar and
28 in Africa) and 2 palaeotropical genera, Hlolmskioldia (10 spp. with 5 en-
demic in Madagascar) and Premna (200 spp., of which 13 occur in Madagascar,
ail but 2 endemic), we see that 7 out of 12 genera can be considered to be palaeo-
tropical. The other 5 genera, 2 of which, Vitex and Clerodendrum, have respec-
tively 39 species (38 endemic) and 62 species (endemic) in Madagascar, are
pantropical.
In contrast with these large genera, the other 3 are also large genera but
small in their Madagascan representation: Phyla (1 cosmopolitan sp.), Priva
(1 endemic sp., 1 sp. common to Africa, 1 sp. common to Asia), and Callicarpa
(1 endemic sp.).
Some very interesting facts have to be emphasized. First of ail is the presence
of 2 monotypic endemic genera, 1 of which, Acharitea, is akin to Nesogenes (6
spp.), a genus with an extraordinary distribution in E. Africa-Comores-Seychelles-
Mascarenes-Madagascar and Polynesia. The two genera belong to a group, a
family of its own according to some authors (Dicrastylidaceae), which is essen-
tially Australian (Plhysopsis, Lachnostachys, Dicrastylis, Mallophora, . . . ). Ail
of these genera are closely related to Verbenaceae-Stilboideae, a small sub-
family of 5 monotypic genera, endemic to S. Africa. Therefore, there is a link,
which seems ancient, between Madagascan and Australian Verbenaceae.
Another fact is that there is a littoral species of Vitex, V. trifolia, which covers
a large area around the Indian Ocean from E. Africa to Japan and Australia. It
represents, of course, a category different from that of Calophyllum inophyllum.
The former has surely nothing to do with other species such as this one, which
should be considered as a possible pioneer species. Another species, Phyla nodi-
flora, belonging to an American genus, occurs here and there, on sandy beaches
and also at higher elevations, preferentially on calcareous soils.
574
[VOL. 65


LEROY-MADAGASCAR
Labiatae (in particular, one monotypic genus, Mahya, is endemic to R�union),
Avicenniaceae, Plantaginaceae (cf. Humbert, 1959).
Oleaceae. This family of 29 genera and 600 species is widespread over the
World, especially in temperate and tropical Asia. There are 62 Madagascan
species (60 endemic) and 5 genera, of which 2 are endemic to the Madagascan
region: Noronhia (40 spp.) and Comoranthus (2 spp. in Madagascar and one
in the Comores). The other genera are: Olea, an Old World genus with 20
species, of which 4 (3 endemic) occur in the Madagascan region. Two large
tropical and subtropical genera: Linociera (1 sp. endemic to Madagascar and 1
endemic to the Comores), and Jasminum (9 spp., 8 endemic).
Scrophulariaceae (unrevised family; cf. Humbert, 1959; 2 genera, Allocalyx,
1 sp., Bryodes, 3 spp., are endemic to the Mascarenes).
Myoporaceae. This small family of Bignoniales deserves special attention.
It is composed of 4-5 genera, 2-3 of which occur in S. Africa-Madagascar. Only
one genus, Myoporum (30 spp.) has a wide distribution from Australia and the
Pacific Islands to Japan and E. China and even the Mascarenes (1 sp., endemic).
The other genera are: Pholidia (40 spp.), endemic to Australia; Bontia (1 sp.),
endemic to the West Indies; Ranopisoa (1 sp.), endemic to Madagascar (Leroy,
1977d) and, when admitted in Myoporaceae, Oftia, 3 species endemic to S.
Africa. Ranopisoa was described as Oftia by Capuron (1972a), but I have shown
that it is distinguished from the latter by some characters. Whether or not
Oftia belongs in this family or in Scrophulariaceae is not yet established, but in
any case, it is at least an ally.
Thus, there is evidence from the pattern of distribution of the Myoporaceae,
along with those from Cunoniaceae and other austral families, that their geo-
graphical origin might have been in any part of Gondwanaland. Moreover, the
Myoporaceae, although having reached a high taxonomic level, might have
originated before the breakup of Gondwanaland. However, the wide distribu-
tion of the genus Myoporum, obviously post-Gondwanian, might result from
long-distance dispersai during Tertiary times.
Gesneriaceae (120 gen., 2,000 spp.). This is a family divided into 2 sub-
families, Gesneroideae and Cyrtandroideae, the former principally American
(with some Australasian genera), and the latter principally tropical Asian and
African and also European and even Australasian (Cyrtandra, . . .); it is repre-
sented in Madagascar by 3 genera: Streptocarpus, a genus endemic to Africa-
Madagascar, with 41 species in Madagascar, ail of them endemic and 90 species
in Africa, almost all in S. and E. Africa; Didymocarpus, a genus widely distributed
in warm regions (150 spp.) having 1-3 species in Africa and 3 endemic species
in Madagascar; Colpogyne, an endemic monotypic genus (Burtt, 1971).
Bignoniaceae (120 gen., 650 spp.). In Madagascar this family, is represented
by 2 tribes, out of 4 constituting the family (Schuman, 1895), and 10 genera.
It displays an old distributional pattern of great interest with respect to the
tribe Crescentieae, which is composed of 11-12 genera: 5-6 in Madagascar, ail
of them endemic, 5 in America, and 1 in Africa. This distribution is especially
noteworthy as the 5-6 Madagascan genera form a homogenous group, with, it
575
1978]


ANNALS OF THE MISSOURI BOTANICAL GARDIEN
seems, some distinctive evolutionary trends. If we put the genera with unilocular
stamens aside (Ophiocolea, 5 spp. and Colea, 20 spp.), the other 3-4 genera,
although well defined, are very close to one another. In my opinion, Rhodocolea
might be the most primitive, particularly R. perrieri Capuron with its perfect
actinomorph flowers and pentastaminate androecium (it even might be raised
to the generic level). Next to Rhodocolea is Phylloctenium, then Phyllarthron,
a plant whose leaves are reduced to an articulated rachis. Two species of the
latter (discovered and described by Capuron, 1960, 1970a) offer such an extraor-
dinary floral zygomorphism that I (Leroy, 1972) proposed to base a new
genus Paraphyllarthron, upon them. Whether accepted or not as taxon by sub-
sequent authors does not matter, what is important is that it deseended from
Phtllarthron, a taxon still in existence today in the saine country. The genus
Phyllarthron itself was possibly derived from a Rhodocolea-like taxon, now ex-
tinct. The Madagascan Crescentieae are possibly a group of closely related
genera that have originated in Madagascar during post-Gondwanian time.
The tribe Tecomeae has 4 genera in Madagascar: Perichlaena, a monotypic
and endemic genus very close to Fernandoa. Fernandoa, a genus of 6 species
of whieh 3 are endemic to Madagascar and 3 endemic to tropical Africa (Gentry,
1975). Rhigozumn (1 sp. in Madagascar, 3 spp. in South Africa). Stereospermum,
a palaeotropical genus with 9 species in Madagascar (constituting a distinct
section, confined to the western Domain).
The family Bignoniaceae in Madagascar seems to offer a typical case of
post-Gondwanian evolution or taxogenesis from some ancestors, now extinct,
common to Africa and South America.
Acanthaceae (250 gen., 2,500 spp.). In Madagascar there are 57 genera, of
which 22 are endemic, and more than 400 species, 375 of whieh are endemic.
Eleven genera are African-Madagascan with 47 species in Madagascar, 42 of
which are endemics. Five genera are related to African genera. The other
genera are palaeotropical or pantropical: Justicia, 95 species in Madagascar,
90 of which are endemic with African affinities; Ilypoestes 80 species, 78 of
which are endemics. On the whole, affinities with the African flora are obvious;
those with the Asian and Malaysian floras are rather modest (Benoist, quoted by
Humbert, 1959).
Several remarks deserve attention: The very high degree of generic en-
demism is noteworthy in Madagascar. Moreover, the distributional pattern of
several genera is of a particular interest. The genus Stenandriopsis (5 spp. in
tropical west Africa, 1 in E. Africa, 8 in Madagascar) is closely allied to Aphe-
landra (200 spp. in tropical America; Heine, 1963; Stearn, 1971). The genus
Mendoncia, which is mostly S. American (90 spp.), has 4 species in Africa
(Gabon & Ivory Coast) and 3 in Madagascar. It is closely related to the mono-
typic genus Gilletiella, endemic to Congo-Angola, so much so that Bremekamp
proposed to link the two genera in the new family Mendonciaceae. The genus
Justicia occurs likewise in S. America and Africa-Madagascar. Oplonia is Mad-
gascan-American (9 spp. in the West Indies, 5 spp. in Madagascar), but absent
from Africa (Stearn, 1971). All these factors indicate that the family, evolu-
576
[VOL. 65


LEROY-MADAGASCAR
tionarily advanced as it is, might have had a Gondwanian origin (perhaps by
the close of the Cretaceous).
Pedaliaceae. This small family (12 gen.) is almost completely confined to
warm regions of the Old World, especially Africa where 9 genera are endemic.
Uncarina, endemic to Madagascar, is a genus of shrubs with thick branches and
contains 9 species, all of them more or less xerophilous and occuring in the Oc-
cidental Domain (Humbert, 1962). Several species belonging to other genera
have been introduced: Dicerocaryum zanguebarium, a monotypic genus of E.
and S. Africa and Rhodesia, Angola; Sesamum sp.; Martynia (= Martiniaceae
Stapf., an American family of 3 genera) annua, introduced from Mexico into
Africa and Madagascar. An annual herb of African sandy beaches, Pedalium
murex, is also in Madagascar and reaches Ceylon and India.
Campanulaceae (Lobeliaceae included) is a family (60 gen., 1,500 spp.) with
worldwide distribution, but particularly in the Mediterranean region and S.
Africa, and also in the Tropics. There are 30 species (26 endemic) and 7 genera
(3 endemic) in the Madagascan region. Four genera belong to Lobelioideae:
Heterochaenia (a genus consisting of 3 spp. endemie to R�union), Dialypetalum
(endemic to Madagascar, 5 spp.), Monopsis (a genus of tropical and S. Africa
with 1 sp. on the Comores), Lobelia (a widespread genus [300 spp.], mostly
tropical and subtropical, richly represented in Africa and America with 12
species in Madagascar [9 endemie] and 7 species in the Mascarenes [2 endemic] ).
There are 3 genera of Campanulaceae sensu stricto: Berenice (a monotypic
genus endemic to the Reunion-Badr� et al., 1975), Gunillaea (an African-
Madagascar genus with 1 species endemic to Africa, 1 common to E. Africa
and Madagascar), Wahlenbergia (a widespread genus but mainly of the S.
hemisphere with 200 species 11 of which in Madagascar (4 endemic) (M. Thulin,
unpublished paper, 1977).
It is of special interest that 2 genera are endemic to the Mascarenes, and
monotypic: they seem, therefore, to be vestigial.
Goodeniaceae. This family of 14 genera and 300 species is almost exclusively
austral but its Madagascan representation, 1 genus, Scaevola, with 2 species
widespread along sea coasts all over the Globe, has no special significance.
Rubiaceae (500 gen., 5,000 spp.). This is a pantropical and subtropical family
represented in Madagascar by its 3 subfamilies and about 21 tribes (Psycho-
trieae, Morindeae, Triainolepideae, Pauridiantheae, Craterispermeae, Knoxieae,
Paederieae, Hedyotideae, Cruckshanksieae, Lathraeocarpeae, Anthospermeae,
Spermacoceae, Rubieae, Naucleeae, Cinchoneae, Mussaendeae, Gardenieae,
Ixoreae, Alberteae, Vanguerieae, Guettardeae). There are about 70-80 genera,
25-30 of which are endemic (2 new genera recognized by Capuron await de-
scription), and more than 600 species, nearly all of them endemic. As was
emphasized by Bremekamp (Humbert, 1959), the Madagascan Rubiaceae have
much stronger affinities with the African flora than with the Asian one. Some
genera (Triainolepis, Alberta, . . . ) are restricted to Madagascar and E. and S.
Africa. Some tribes are particularly well represented in Madagascar: Psychotrieae
(ca. 11 gen., 5 endemic), Hedyotideae (5 gen., 2-3 endemic), Spermacoceae,
577
1978]


ANNALS OF THE MISSOURI BOTANICAL CARDEN
Naucleeae (a tribe of 8 genera, 5 endemic in Madagascar), Ixoreae (8-12
genera, several of which endemic), Gardenieae (4 gen., 2 endemic-Leroy, 1974).
There is one endemic genus in the Madagascan Cinchoneae, and another one in
the Mussaendeae. Three genera are endemic to the Mascarenes: Scyphochlamys
(1 sp.), Neoschimpera (1 sp., Seychelles), Fernelia (2 spp.).
Many data concerning the Madagascan Rubiaceae are contained in an im-
partant unpublished paper by Capuron, but it is, however, impossible to draw
any serious conclusion before the revision of the family is completed.
Compositae (900 gen., 15,000 spp.). This is a very large family of cosmopoli-
tan distribution. It ranks first in the S. African flora with 174 genera and 2,072
species (Goldblatt, this symposium). It is also first in Madagascar with 87 genera
(13 endemic) and 550 species (500 endemic) distributed in 9 tribes (none of
which endemic). Two monotypic genera are endemic to the Mascarenes
(Eriothrix, Cylindrocline). The Madagascan Compositae are strongly linked
with the E. and S. African ones (cf. Humbert, 1959).
VII
Concerning the Madagascan Monocots I shall be content with recalling some
of the outstanding classical facts, supplemented with some recent results. It is
well known, for example, that the Bambuseae, Palmae, Pandanaceae, and Orchi-
daceae have a particularly important Madagascan representation.
A. Alismatidae.-Among the represented families: Alismataceae, Apono-
getonaceae, Scheuchzeriaceae, Najadaceae, Potamogetonaceae, Hydrocharita-
ceae, Triuridaceae, only the latter has an endemic Madagascan genus, Seychel-
laria (1 sp. in Madagascar, 1 sp. in the Seychelles). The Triuridaceae (7 gen.,
80 spp.) are herbaceous terrestrial mycotrophic plants with apocarpous flowers
and albuminous seeds.
B. Commelinidae.-The Bamboos are impressively represented in Madagascar
with 10 genera, 6 of which endemic. One of the latter, Nastus, has 12 species in
Madagascar, 1 in the R�union. Three are monotypic: Hitchcockella (which
is related to both Nastus and the American genus Chusquea), Decaryochloa and
Pseudocoix. The other 2 genera are Hickelia (2 spp.) and Perrierbambus (2
spp.) (which would be related to an American genus [Humbert, 1959]).
The nonbambusoid tribes of grasses taken as a whole are represented in
Madagascar by 132 genera, 15 of which are endemic, and 500 species. Two
genera constitute by themselves two tribes apart: Cyphochlaena (Boivinelleae)
and Lecomtella (Lecomtelleae) (J. Bosser, pers. comm.). Thus there is some
indication of an active and ancient differentiation of grasses, perlaps since the
Eocene. A study restricted to the Malagasy fodder grasses was published by
Bosser (1969).
Cyperaceae (90 gen., 400 sp.). There are 29 genera and 350 species
(60 endemic) in Madagascar (cf. Humbert, 1959). In particular, the genus
Costularia is to be cited: it is composed of 14 species, 11 of which occur in
Australia, and 3 in S. Africa and Madagascar.
Strelitziaceae (3 gen., 7 spp.). This is a family, formerly merged into
[VOL. 65


LEROY-MADAGASCAR
Musaceae, whose distribution is strictly austral and strongly discontinuous, with
Ravenala, a monotypic genus, in Madagascar, Strelitzia (5 spp.) in S. Africa,
and Phenakospermum (2 spp.) in S. America. One genuine species of the palaeo-
tropical family Musaceae (Ensete sp.) inhabits Madagascar.
Commelinaceae (38 gen., 500 spp.). Two genera are of special interest:
Coleotrype with 5 endemic species in Madagascar, 1 species in S. Africa, and 1
in W. tropical Africa; Pseudoparis (2 spp.), a very isolated endemic genus.
Eriocaulaceae (13 gen., 1,200 spp.) is a family of extreme biogeographical
interest whose area is almost entirely confined to tropical and subtropical
America. There are only 2 genera endemic to Africa-Madagascar: Mesanthe-
mum with 2 species in Madagascar and 6 in Africa; Moldenkeanthus, a mono-
typic genus endemic to Madagascar. Aside from Eriocaulon, a large genus
which has 19 species in Madagascar (10 endemic), the other 2 genera are
American with only 1 or 2 species in Africa and Madagascar (Paepalanthus,
Syngonanthus). The endemic genus Moldenkeanthus is very close to Paepalan-
thus and the strictly American genus Leiothrix (Morat, 1976).
Zingiberaceae (45 gen., 700 sp.). This family, especially well developed in
Indo-Malesia, is represented by 6 nonendemic genera in Madagascar, 1 of
which, Aulotandra, has 5 species (4 Madagascan and 1 in W. tropical Africa).
Typhaceae, Juncaceae (cf. Humbert, 1959).
C. Arecidae.-Palmae (220 gen., 2,500 spp.). The Palmae (Arecales), whose
fossil material is known from the Upper Cretaceous, are one of most noteworthy
features of the Madagascan flora. There are 18 indigenous genera, 15 of which
are endemic, and 120 species, all but 2 endemic (Humbert, 1959).
The palms occur also in the Seychelles (6 endemic gen.: the famous Boras-
soid Lodoicea possibly related to the Malayan Borassodendron, and 5 spiny
Arecoid gen.) and Mascarenes (6 endemic monotypic gen.) (Corner, 1966).
According to Corner, Madagascar "acting as a small continent of long isolation
has evolved new species of undergrowth palms from its Arecoid heritage, but
it has not produced a subfamily or a new manner of palm life." However, such
Borassoid genera as Medemia, Latania, and Lodoicea are Gondwanian relicts
(Corner).
Pandanaceae (3 gen., 700 spp.). The abundance of Pandanus species is
also a striking feature of the Madagascan flora which counts 30 species, nearly
all of them endemic (Mascarene included).
Araceae (100 gen., 2,000 spp.) are of a particular interest with respect to
taxogenesis for, aside from the pantropical genus Pistia and some introduced
species (Alocasia sp., Caladium sp., Zantedeschia sp.), they are represented in
Madagascar by two elements:
1. An autochthonous highly differentiated element composed of 2 endemic
tribes and 4 genera: (a) Typhonodoreae (Philodendroideae): 1 genus, Ty-
phonodorum, which reaches Pemba and Zanzibar. (b) Arophyteae (Aroideae),
including Carlephyton (3 spp.), Colletogyne (1 sp.), and Arophyton (7 spp.).
2. An oriental element marked by the presence of an endemic species of
1978]
579


ANNALS OF THE MISSOURI BOTANICAL GARDEN
Amorphophallus (A. hildebrandtii), and secondarily by some natural, naturalized,
and/or cultivated species (Remusatia vivipara-which reaches Africa, Typhonium
divaricatumn, Pothos scanidens, Amorphophallus campanulatus, and Colocasia
esculentta ).
Ail in all, there are 13 endemic species.
Lemnaceae (cf. Humbert, 1959).
D. Liliidae.-Pontederiaceae (6 gen., 30 spp.) is represented in Madagascar
by one monotypic genus, Scholleropsis. This genus had been considered to be
a Madagascan endemic until 1967, when Letouzey and Bosser (Letouzey, 1967)
reported its presence in Cameroon. The Pontederiaceae are easily dispersed
water plants, however, Scholleropsis might be an ancient genus in its African-
Madagascan area.
Liliales. Any serious biogeographical judgement on the Liliaceae, together
with many other closely allied families, would need a careful review of
a great many publications, among them that by Iluber (1989). I only intend
to give indications on some selected families.
The Herreriaceae (which is a subfamily of Liliaceae for many authors), close
to Asparagaceae, is composed of 3 genera: 1 in Madagascar (Herreriopsis, 1
sp.), 1-2 in S. America (Herreria).
Astcliaceae is a family of some genera (Yucca, Dracaena. . .), 1 of which,
Cohnia, with fleshy fruits, has 3 species ranging from the Mascarene Islands
to New Caledonia.
Dianellaceae is an Australasian family with 2 genera: Dianella in Asia,
Hawaii, and Madagascar (Raven & Axelrod, 1974) and Excremis in S. America.
Asphodelaceae. The genus Arthropodium has 1 species in Madagascar, the
others in Australia and Pacific Islands. Lomatophyllum, an endemic genus with
fleshy fruits, has 10 species in Madagascar, 3 species in the Mascarenes. Aloe, an
African-Madagascan genus, chiefly developed in Madagascar, where it has 30
endemic species, and oriental and austral Africa.
. Liliaceae-Scilloideae has 1 endemic genus: Rhodocodon (8 spp.).
Iridaceae. The endemic monotypic genus, Geosiris, is obviously a relict.
Velloziaceae (5-6 gen., 200 spp.) is a small, typically occidental-austral fam-
ily composed of 2 genera: Xerophyta in Madagascar and Africa and Talbotia S.
Africa and 4 in South and Central America (Smith & Ayensu, 1974). Talbotia
is a monotypic genus. The genus Xerophyta is divided into 3 sections. One of
these, section Xerophyta, is endemic to Madagascar with 2 species. The second
section is common to Africa (1 sp. endemic to Madagascar, 10 endemic spp. in
Africa, some of them reaching Za�re and Ethiopia). The third section is endemic
to Africa (14 spp.).
Dioscoreaceae (4 gen., 650 spp.) is represented in Madagascar by 2 genera:
Dioscorea (over 30 spp., 27 endemic) and 1 monotypic endemic genus, Avetra,
certainly primitive with a vestigial rhizome, which has completely disappeared in
Dioscorea. Another genus, likewise monotypic but Indonesian, Trichopus (Tri-
chopodaceae) is generally considered as related.
Taccaceae (cf. Humbert, 1959).
Orchidaceae (740 gen., 20,000 spp.). This is one of most important families
[Vol. 65
580


LEROY-MADAGASCAR
in the Madagascan flora with about 1,000 spccies and 56 genera, about 10 of
which endemic. Some genera are widely distributed: Bulbophyllum, Eulophia,
Oecoeclades, . . . . Certain others have a much higher number of species in
Madagascar than in Africa: Angraecum, Jumellea, Aeranthes, Cynorkis, Ben-
thamia. The Asian affinities are weaker than the African ones (J. Bosser, pers.
comm.).
In conclusion, many data can be drawn from the analysis of the Madagascan
Monocots to form the basis for certain useful conclusions. The occurrence of
isolated and highly differentiated genera such as Medemia, Latania, Lodoicea,
Seychellaria, Geosiris (a family of its own for some authors), Avetra, as well
as that of endemic tribes, namely Boivinelleae and Lecomtelleae in nonbambusoid
grasses, and Typhonodoreae and Arophyteae in Arales, appears to support the
hypothesis of a Madagascan-African origin for many Monocots by, most likely,
the close of the Godwanian epoch or early in the Tertiary. Further, the fact that
such high level taxa as Bambuseae, Palmae, Pandanaceae, Orchidaceae, are par-
ticularly well represented in Madagascar might, in part, be explained by referring
to a serious impoverishment of the African flora from Miocene times onwards.
In many cases, of course, e.g., those of Orchidaceae (Angraecum, Jumellea,
Aeranthes, Cynorkis, Benthamia), Palmae (the group including Dypsis with 21
spp., Neophloga with 30 spp., Chrysalidocarpus with 20 spp. .. .), and Poaceae,
we do not deal with an ancient differentiation but with a secondary one: it is
the young part of the Madagascan element. The genera Dypsis, Neophloga, and
Chrysalidocarpus are closely allied and have certainly had a common ancestor in
Madagascar. Others, such as those of the fan-leaved subfamily Borassoideae
seem being genuinely Late Gondwanian elements: Medemia (1 sp. in Madagas-
car, 1 sp. in E. Africa); Latania (3 spp.); Lodoicea (1 sp.) related to Borasso-
dendron of S.E. Asia (Corner, 1966); Borassus (5 spp. ? of which 2 in Madagascar,
1 in Africa, 1 in Asia, 1 in New Guinea); Hyphaene (more than 10 spp. in Africa
and Madagascar). The ancestors of these Borassoid genera must have been
dwellers of the Gondwanian Africa.
VIII
Pteridophytes (F. Badr�, pers. comm.; Christenscn, 1932).-Most of the
genera are represented in the Mascarene region, except for the primitive ones
such as Matonia (Malaysia, Borneo, Sumatra), Dipteris (Asia), and Loxsoma
(New Zealand). The Mascarene region is known for its high number of endemic
species: 200 out of a total of 550. The greater part of the Pteridophytes from the
Mascarene region is related to representatives from Asia (and particularly from
Indomalaysia). According to their affinities, Christensen (1932) recognized 7
groups of species:
(1) The Mascarene element: about 70 species, mostly belonging to groups
of species within large genera such as Cyathea, which contains 37 species, almost
all of them endemic. Several large genera have only 1 or a few endemic species.
There are 2 endemic and monotypic genera: Psammiosorus (confined to Mada-
gascar) and Ochropteris.
1978]
581


[VOL. 65
ANNALS OF THE MISSOURI BOTANICAL GARI)EN
(2) The African element: about 92 species (Christensen). Here belongs
also the genus Mohria (2 spp. in Africa, Madagascar, and R�union). Let us
cite also the genera Blotiella with 1 species in America and 16 in Africa, Madagas-
car and the Mascarenes, and Actiniopteris, which is however also present in
India (2 spp. out of 5).
A markedly disjunct area is shown by Elaphoglossum hirtum (tropical
America, R�union, Madeira, Azores) and Pellea calomelanos (Spain, E. and S.
Africa, N. India, Madagascar, R�union).
(3) The Eastern (African-Malesian) element: represented by more than
150 species, that is, nearly M1/ of the total number. Examples include Angiopteris
(from Madagascar and R�union, with 1 species in Japan and Polynesia),
Drynaria (about 20 spp. in Asia and Australia, 1 sp. in Madagascar and Mauri-
tius), Loxogramme (35 spp. in Asia with 2 spp. in the Mascarene region, 1 of
which also in Africa), Microsorium, particularly well represented in tropical
Asia (4 spp. in the Mascarene region, 2 of which are endemic), Davallia (40 spp.
mainly Asian and Polynesian, 1 in Africa, Seychelles, Rodrigues, Mauritius, and
Madagascar). Several species are widely distributed, occuring also in Central
and Eastern Asia. Examples include Dryoathyriuim boranum (Africa, Madagas-
car, Mascarenes, Asia, Polynesia), Diplazium sylvaticum (Mauritius, Asia, Phil-
ippines).
(4) The Western (American-African) element: about 85 species (18% of the
total number of Pteridophytes). Here belong such genera as Elaphoglossum
(400 spp., most of which occur in America, 40 spp. in the Mascarene region),
Doryopteris (35 spp., mainly in S. America, 2 of which are endemic), Pellea (75
spp., most of which occur in S. America, 30 spp. in Africa and the Mascarene
region), Trachypteris (2 spp., 1 of which occurs in S. America and the Galapagos
Islands, the other in Madagascar).
(5) The Southern element: represented by 15 species belonging to such genera
as Blechnum, Schizaea, Gleichenia, ....
(6) The pantropical element with about 20 species.
(7) The cosmopolitan element: about 15 species, including Osmunda regalis,
Lycopodinu clavatum, Pteridium aquilinum, Pteris cretica, Pteris vittata, Equi-
setuim ramosissimu m, Cystopteris fragilis, Adianttum capillu-veneris, Iy menophyl-
lum tunbrigense ....
From that rather artificial classification it becomes clear that the palaeotrop-
ical element is best developed in the Mascarene flora. Also that the number of
endemic species in the Mascarene region is proportionally high, whereas only
a few of the genera are endemic. This is a general feature of the Mascarene
flora which is characterized by a very strong specific endemism.
Gymnoosperms.-Podocarpaceae (8 gen., 150 spp.). This family of Coniferales,
almost confined to the tropical southern hemisphere, is represented in Madagascar
by the only genus Podocarpus (100 spp.) with 6 endemic species. This genus
was divided into 8 sections (Buchholtz & Gray, 1948), 1 of which is endemic
to S. Africa. The Madagascan species belong to section Eupodocarpus (with
transfusion tissue) which is present also in Africa, but also in S. America, Aus-
tralia, South Pacifie Islands and Asia.
582


LEROY-MADAGASCAR
Cycadaceae (1 gen., 10 spp.). The genus Cycas, confined to the Indian Ocean
and western Pacific, is represented in Madagascar, the Comores, and E. Africa
(Tanzania, Mozambic) by 1 special species (C. thouarsii) which grows along
sea coasts, often in the Barringtonia formations.
GENERAL CONCLUSIONS
In every stage of its evolution, whether it be primitive, advanced, or inter-
mediate, the Madagascan vascular flora is of outstanding originality. No doubt,
this is due in part to extremely diversified ecological conditions, but also, and
above all, to the fact that the island which was part of the Gondwanaland con-
tinent, was hardly modified in the course of time, at least since the early Creta-
ceous. In contrast, most of the samie flora that occupied continental Africa became
greatly impoverished, particularly from the Miocene on, due to climatic changes.
Therefore, Madagascar now appears to be an African fragment that has
not only been an extraordinary area for the differentiation of taxa but above ail
both a refuge for every kind of immigrant and a center of survival for archaic
autochthonous plants. Its relationship with East and South Africa is obvious
and represents an especially interesting kind of affinity because it testifies to
the peripheral character of this region compared to the African landmass; it
also testifies to the antiquity of the Madagascan flora. That this relationship
really exists is supported by many facts, by far the most convincing of which
is the occurrence in the Madagascan region sensu lato of a close-knit group of
two austral families with monosulcate or monosulcate-derived pollen, belonging
to Magnoliales: the Winteraceae and Canellaceae. The former is divided into
two subfamilies, one with 7 genera in the Pacifie area and 1 in tropical America;
the other subfamily is unispecific and endemic to Madagascar. The Canellaceae
comprises 5 genera, 1 endemic to Madagascar, 1 to South Africa, and 3 to trop-
ical America. Linked to these two families of Magnoliales, absent from the in-
terior of Africa is the Monimiaceae which is likewise austral, and especially
richly represented in Madagascar. Equally significant in the austral Magnoliales
is the IIernandiaceae (4 gen.); HIernandia consists of several subgenera, 2 of
which are most primitive, 1 of them is endemic to Australia, the other to Mada-
gascar.
It is also noteworthy that the primitive Hamamelididae is represented in the
sanme area by 2 endemic genera of Hlamamelidaceae (none in the interior of
Africa) and, if admitted in this subclass, by several special families such as the
Myrothamnaceae and Didymelaceae.
Another striking fact is the very strong endemism of the Dilleniidae, not only
at the generic level, but also at the familial one: 5 endemic families (Diego-
dendraceae, Sphaerosepalaceae, Sarcolaenaceae, Asteropeiaceae, Medusagyna-
ceae); several families with an exceptional development (Malvales, Flacourtia-
ceae, Sapotaceae, Ebenaceae, ...). The taxonomic level is roughly rather higher
than that of Megnoliidae and in many cases speeciation is active.
The Rosidae are also to be noted. Such families as Brexiaceae, Montiniaceae,
583
1978]


ANNALS OF THE MISSOURI BOTANICAL GARDEN
and Escalloniaceae are austral families and not or hardly represented in con-
tinental Africa sensu stricto. The Proteaceae, an austral family of 75 genera, is
concentrated in Australia (45 gen.) and South Africa (14 gen.), but it has 3
genera in Madagascar, 2 of which are endemic. The Balanophoraceae, again
an austral family, has been divided into 6 subfamilies, 2 of which are endemic
to Africa, principally South Africa; the subfamily Helosioideae to which belongs
the Madagascan genus Ditepalanthus, is absent from Africa and present in Asia
(3 gen.) and America (3 gen.).
The Asteridae which is evolutionarily more advanced is also richly represented
in Madagascar, but contrary to the facts mentioned above, its endemism is in
general rather at a lower level. There is almost a rule: as one goes down tax-
onomic levels, speciation becomes more active. Perhaps not a single endemic
family of Asteridae is present in Madagascar whereas such genera as Clero-
dendron, Justicia, Coffea, Helichrysum, Vernonia, Senecio, etc. have many en-
demic species.
With respect to Monocotyledons, it is difficult to draw any homogenous
conclusion because the taxogenetic knowledge of this group is rather poor.
Roughly, however, the facts are comparable. There is on the one hand, a very
important archaic clement including relictual endemic subfamilies, tribes, and
genera and, on the other hand, a younger element with neo-endemic genera
and a huge number of endemic species that result from an expectionally active
speciation (Orchidaccae, Gramineae, . . .).
As a whole, the Madagascan vascular flora is considered here as an autoch-
thonous flora that has differentiated principally from the original Gondwanian
stock and has, in course of time, grown rich through evolution of its members and
immigration of newcomers through long-distance dispersal.
LITERATURE CITEI
(Many data have been taken from classical Floras, which are not cited here, e.g., Flore
de Madagascar et des Comores, Flore du Congo Belge et du Ruanda-Urundi, Flora Zambesiaca,
Flora of Tropical East Africa, Flora Malesiana, etc.)
AmY SHAW, I. K. 1966. J. C. Willis. A Dictionary of Flowering Plants and Ferns. Ed. 7.
Cambridge Univ. Press, Cambridge.
. 1972. Melanophyllaceae. Kew Bull. 26: 491-493.
AUvu�VI.LE, A. 1955. La disjonction africaine dans la flore foresti�re tropicale. Compt.
Rend. Sommaire S�ances Soc. Biog�ogr. 278: 42-49.
S1972. Etude phytog�ographique de la famille des Sapotac�es malgaches dans le
cadre g�ographique africain. Adansonia, s�r. 2, 12: 55-59.
1974a. Les origines des Angiospermes (1�re partie). Adansonia, s�r. 2, 14: 5-27.
S19741). Origines polytopiques des Angiospermes tropicales (2e partie). Essais cho-
rologiques. Adansonia, s�r. 2, 14: 145-198.
1975a. Essais de g�ophyl�tique des Bombacac�es. Adansonia, s�r. 2, 15: 57-64.
1975b. Madagascar au sein de la Pang�e. Adansonia, s�r. 2, 15: 295-305.
S1976a. Essai d'interpr�tation nouvelle de la distribution des Dipt�rocarpac�es.
Adansonia, s�r. 2, 16: 205-210.
. 19761). Centres tertiaires d'origine, radiations et migrations des flores angiosperm-
iques tropicales. Adansonia, s�r. 2, 16: 297-354.
BADR�, F., T. CADET, G. CUSSET & M. HIDEUX. 1975. Position syst�matique, �tude morpho-
logique et palynologique du genre Berenice. Adansonia, s�r. 2, 15: 139-146.
BARANOV, A. & F. A. BAHKLEY. 1974. The sections of the genus Begonia. Northeastern
Univ., Boston. 28 pp.
584
[VOL. 65


1978]
LEROY-MADAGASCAR
585
BARKLEY, F. A. & J. COLDING. 1974. The species of the Begoniaceae. Northeastern Univ.,
Boston. 144 pp.
BATES, D. 1968. Generic relationships in the Malvaceae, Tribe Malveae. Gentes Herb. 10:
117-135.
BERNARDI, L. 1975. Grandes lignes de diff�renciation et d'affinit�s des Araliac�es et des
Palmiers d'Afrique et des �les de l'oc�an Indien. Boissiera 24: 355-362.
BOSSER, J. 1969. Gramin�es des pasturages et des cultures � Madagascar. M�m. ORSTROM
35: 1-440.
BIENAN, J. P. M. 1949. A contribution to the flora of East Tropical Africa. Kew Bull. 4:
71-95.
BRETELEH, F. J. 1969. The African Dichapetalaceae I. Acta Bot. Neerl. 18: 375-376.
BUCHHOLZ, J. T. & N. E. GRAY. 1948. A taxonomic revision of Podocarpus. J. Arnold
Arbor. 29: 49-63.
Buvrr, B. L. 1971. G�sneriac�es. 180e Famille. In J.-F. Leroy (editor), Flore de Mada-
gascar et des Comores. Mus�um National d'Histoire Naturelle, Paris.
CAPURON, R. 1960. Contribution � l'�tude de la flore foresti�re de Madagascar. 5. Trois
Bignoniac�es nouvelles. Notul. Syst. (Paris) 16: 71-80.
1961. Contribution � l'�tude de la flore foresti�re de Madagascar. Adansonia,
s�r. 2, 1: 65-92.
.1962. Contribution � l'�tude de la flore foresti�re de Madagascar. Adansonia, s�r.
2, 2: 267-284.
.1963a. R�vision des Tiliac�es de Madagascar et des Comores. Adansonia, s�r. 2, 3:
91-129.
.1963b. Contributions � l'�tude de la flore de Madagascar. XI. Pr�sence � Mada-
gascar d'un repr�sentant du genre Macadamia F. v. M. (Prot�ac�es). XII. Pr�sence �
Madagascar d'un nouveau repr�sentant (Bubbhia perrieri R. Capuron) de la famille des
Wint�rac�es.... XIV. Le genre Ardisia Swartz ( Myrsinac�es) � Madagascar. XV. Di-
egodendron R. Capuron gen. nov., type de la nouvelle famille des Diegodendraceae (Och-
nales sensu IHutchinson) .... Adansonia, s�r. 2, 3: 370-400.
. 1963c. Contribution a l'�tude de la flore de Madagascar. X. Pr�sence du genre
Nectaropetalumn Engler � Madagascar. Adansonia, s�r. 2, 3: 141.
S1964. R�vision des Tiliac�es de Madagascar et des Comores (suite). Adansonia,
s�r. 2, 4: 269-300.
S1965a. Les Rhamnac�es arbustives ou arborescentes de Madagascar. Exemplaire
ron�ot�, Centre Technique Forestier Tropical, Tananarive. 52 pp., 6 cartes, 9 pl.
1965b. Une Irvingiac�e malgache. Adansonia, s�r. 2, 5: 213-216.
1966a. Notes sur quelques Rhamnac�es arbustives ou arborescentes de Madagascar.
Adansonia, s�r. 2, 6: 117-141.
S1966b. Hazonalania R. Capuron, nouveau genre malgache de la famille des Her-
nandiac�es. Andansonia, s�r. 2, 6: 375-384.
S1967a. Deux Caesalpinia nouveaux pour Madagascar. Adansonia, s�r. 2, 7: 199-
205.
.19671). Nouvelles observations sur les Rutac�es de Madagascar. Adansonia, s�r. 2,
7: 479-500.
S1968a. Olacac�es, Opiliac�es et Santalac�es arbustives ou arborescentes de Mada-
gascar. Mat�riaux pour l'�tude de la flore foresti�re de Madagascar. Exemplaire ron�ot�,
Centre Technique Forestier Tropical, Tananarive. 69 pp., 2 cartes, 8 pl.
S1968b. Contribution � l'�tude de la flore foresti�re de Madagascar. A. Notes sur
quelques Cassi�es malgaches. Adansonia, s�r. 2, 8: 17-37, 199-27. Les Swartzi�es de
Madagascar, 8: 217-222.
1968c. Sur les Protium (Burs�rac�es) de Madagascar. Adansonia, s�r. 2, 8: 360-
363.
.1968d. Sur le Prockiopsis hildebrandtii Baillon (Flacourtiac�es). Adansonia, s�r. 2,
8: 365-366.
S1969. R�vision des Sapindac�es de Madagascar et des Comores. M�m. Mus. Natl.
IHist. Nat., S�r. B, Bot. 19: 1-189.
S1970a. Deux nouvelles Bignoniac�es. Une Crescenti�e � fleurs r�guli�res: Rhodo-
colea perrieri Capuron. Adansonia, s�r. 2, 10: 501-506.
.1970b. Le genre Alhizzia Durazz (L�gumineuses-Mimosoid�es). Exemplaire
ron�ot�, Centre Technique Forestier Tropical, Tananarive. 145 pp., 14 pl., 4 cartes.


ANNALS OF THE MISSOURI BOTANICAI CARDEN
. 1972a. Contribution � l'�tude de la flore foresti�re de Madagascar. Adansonia, s�r.
2, 12: 375-388.
. 1972b. Contribution � l'�tude de la flore foresti�re de Madagascar. B. Sur deux
nouvelles esp�ces du genre Hirtella L. (Rosac�es). Adansonia, s�r. 2, 12: 379-383.
-. 1973a. Contribution � l'�tude de la flore foresti�re de Madagascar. Notes sur le
genre Terminalia. Bull. Mus. Hist. Nat. (Paris), Bot., s�r. 3, 11(191): 89-179.
. 1973b. Observations sur les Myristicac�es de Madagascar. Les genres Brochoneura
et Mauloutchia. Adansonia, s�r. 2, 13: 203-221.
CHBISTNENS, C. 1932. The composition of the fern-flora of Madagascar. Dansk Bot. Ark.
7: 221-242.
CORNEn, E. J. H. 1966. The Natural Ilistory of Palns. Univ. of California Press, London.
CnoNQulST, A. 1968. Thi Evolution and Classification of Flowering Plants. Thomas Nelson
Ltd., London.
CuFonoNTls, G. 1960. �ber den Umfang und die natirliche Verbreitung der Gattung Pit-
tosporum. Bol. Soc. Brot., ser. 2, 34: 159-176.
CUSSET, C. 1974. Contribution � l'�tude des Podostemaceae IV. Les genres Ledermanniella,
Monandriella et Inversodicrawsia. Adansonia, s�r. 2, 14: 271-275.
DEJAiDIN, J., J.-L. GUILLAUMET & G. MAN(ENOT. 1973. Contribution � la connaissance de
l'�l�ment non end�mique de la flore malgache (v�g�taux vasculaires). Candollea 28:
325-391.
I)ESCOINCS, B. 1967. Vitac�es. 124e Famille. In H. Iumbert (editor), Flore de Madagas-
car et des Comores. Mus�um National d'llistoire Naturelle, Paris.
EXEI.L, A. W. 1972. Patterns of distribution in Combretaceae. Pp. 307-323, in I). H. Valen-
tine (editor), Taxonomy, Phytogeography and Evolution. Academic Press, London.
- & C. A. STACE. 1966. Revision of the Combretaceae. Bol. Soc. Brot., ser. 2, 40:
5-25.
EYni, R. 1972. Pollen of Alangiium: toward a more satisfactory synthesis. Taxon 21:
471-477.
FLORET, J. J. 1974. Comiphytoni genre nouveau gabonais, Rhizophoraceae-Macarisiae.
Adansonia, s�r. 2, 14: 499-506.
. 1976. A propos du Comiplhyton gabonense ( Rhizophoraceae-Macarisieae ). Adan-
sonia, s�r. 2, 16: 39-49.
FORMAN, L. L. 1965. A new genus of Ixonanthaceae with notes on the family. Kew Bull.
19: 517-526.
FRYXELL, P. 1974. Cienfuegosia extended to Madagascar. Ann. Missouri Bot. Gard. 61:
491-493.
(ENTRY, A. II. 1975. Studies in Bignoniaceae 17: Kigelianthe: a synonym of Fernandoa.
Ann. Missouri Bot. Card. 62: 480-483.
Goon, R. 1950. Madagascar and New Caledonia. Blumea 6: 470-479.
G(ILLAUnMET, J.-L. 1972. La variation du genre Rhipsalis (Cactac�es) � Madagascar. Adan-
sonia, s�r. 2, 12: 433-445.
IIEINE, I. 1963. Acanthaceae. In J. HIutehinson & J. M. )alziel (editors). Flora of West
Tropical Africa. Vol. 2: 391-432.
HlEP'PE, F. N. 1965. Preliminary account of the phytogeographical affinities of the flora of
west tropical Africa. Webbia 19: 593-617.
S1968. The occurence of Teinstroemia (Theaceae) in West Africa. Kew Bull. 21:
429-431.
IHUBEn, IH. 1969. Die Samenmerkmale und Verwandtschafsverhiiltnisse der Liliifloren. Mitt.
Bot. Staatssamml. Miinchen 8: 219-538.
HUMBEHT, II. 1959. Origines pr�sum�es et affinit�s de la flore de Madagascar. M�m. Inst.
Sei. Madagascar, S�r. B, Biol. Veg. 9: 149-187.
1962. Les P�daliac�es de Madagascar. Adansonia, s�r. 2, 2: 200-215.
& R. CAPURON. 1955. D�couverte d'une Chloranthac�e � Madagascar: Ascarinopsis
ounrsii gen. nov., sp. nov. Compt. Rend. Hebd. S�ances Acad. Sci. 240: 28-30.
IIUTHIIINSON, J. 1964. The Genera of Flowering Plants. Vol. 1. Oxford Univ. Press, Oxford.
. 1967. The Genera of Flowering Plants. Vol. 2. Oxford Univ. Press, Oxford.
HUYNir , K. L. 1971. Etude du pollen des Oxalidac�es. III. Le pollen du Dapania pentan-
dra Capuron et sa position taxonomique. Bot. Jahrb. Syst. 90: 524-526.
JAcQUES-F�LIX, II. 1973. Contribution � l'�tude du genre Rousseauxia (M�last.) Adansonia,
s�r. 2, 13: 177-193.
586
[VOL. 65


LEROY-MADAGASCAR
587
JOHNSON, L. & B. BRIGGS. 1975. On the Proteaceae-the evolution and classification of a
southern family. Bot. J. Linn. Soc. 70: 83-182.
KANIS, A. 1968. A revision of the Ochnaceae of the Indo-Pacific area. Blumea 16: 1-82.
KERAUDREN, M. 1965. Le genre Moringa en Afrique et � Madagascar (Affinit�s syst�ma-
tiques, int�r�t biog�ographique). Webbia 19: 815-824.
. 1966. Recherches sur les Cucurbitac�es (le Madagascar. M�m. Mus. Natl. Hist.
Nat., S�r. B, Bot. 16: 123-330.
KOECHLIN, J. 1972. Flora and vegetation of Madagascar. Pp. 145-190, in R. Battistini & G.
Richard-Vindard (editors), Biogeography and Ecology of Madagascar. Junk, The Hague.
, J.-L. GUILLAUMET & P. MORAT. 1974. Flore et V�g�tation de Madagascar. J.
Cramer, Berlin.
KUBITSKI, K. 1969. Monographie der Hernandiaceae. Bot. Jahrb. Syst. 89: 78-209.
L�ANDRI, J. 1947. Nouvelles observations sur les Thym�l�ac�es de Madagascar. Notul. Syst.
(Paris). 13: 38-55.
. 1949. Sur la pr�sence d'une Trigoniac�e dans la flore malgache. Compt. Rend.
Hebd. S�ances Acad. Sci. 229: 846-848.
LEBRUN, J. 1961. Les deux flores d'Afrique tropicale. M�m. Acad. Roy. Sci. Belgique, s�r.
2, 32: 1-81.
LEENHOUTS, P. W. 1962. Loganiaceae. In Flora Malesiana, Ser. 1, 6: 293-387.
LEEUWENIIERG, A. J. M. 1974. The Loganiaceae of Africa. XII. A revision of Mitreola L.
Meded. Landbouwhogeschool 74(23): 1-28.
. 1975a. The Loganiaceae of Africa. XIV. A revision of Nuxia Lan. Meded. Land-
bouwhogeschool 75(8): 1-80.
. 1975b. The Loganiaceae of Africa. XII. Buddl�eya L. Acta Bot. Neerl. 24: 83-86.
L�ONARD, J. 1950. Lepidobotrys Engl., type d'une famille, nouvelle deu Spermatophytes:
Les Lepidobotryaceae. Bull. Jard. Bot. �tat 20: 31-40.
LEROY, J.-F. 1949. De la morphologie florale et de la classification des Myricaceae. Compt.
Rend. Ilebd. S�ances Acad. Sci. 229: 1162-1163.
. 1957. Sur deux Amentif�res remarquables de la flore Asiatico-Pacifique et Pacifique.
Proc. Eighth Pacifie Science Congress (Univ. Philippines) 4: 459-464.
. 1959. Sur une petite famille de Sapindales propre � l'Afrique australe et � Mada-
gascar: les Ptaeroxylaceae. Compt. Rend. Hebd. S�ances Acad. Sci. 248: 1001-1003.
. 1972. La gen�se d'un genre chez les Bignoniac�es Cresenti�es de Madagascar.
Compt. Rend. Ilebd. S�ances Acad. Sci. 275: 2675-2677.
. 1974. Recherches sur les Rubiac�es de Madagascar: les genres Mantalania et Pseu-
domantalania (Gard�ni�es). Adansonia, s�r. 2, 14: 29-52.
. 1975. Transpecific evolution; a hypothesis on the phylogenesis of Malagasy Cutti-
fereae. Twelfth Int. Bot. Congr., Abstr. 1: 97.
. 1977a. A compound ovary with open carpels in Winteraceae (Magnoliales): evo-
lutionary implications. Science 196: 977.
S1977b. Taxogen�se de la flore malgache: les genres Mammea et Paramammea
Leroy, gen. nov. (Guttiferae). Compt. Rend. llebd. S�ances Acad. Sci., D, 284: 1521.
. 1977c. Essais de Taxonomie syncr�tique 1) Etude sur les Meliaceae de Madagascar.
Adansonia, s�r. 2, 16: 167-203.
.1977d. Les Myoporaceae � Madagascar: niveau d'end�misme. Adansonia, s�r. 2,
17: 113-118.
LETOUZEY, R. 1967. Pr�sence au Cameroun d'une Pont�deriac�e: Scholleropsis lutea H.
Perr. Adansonia, s�r. 2, 7: 33-37.
.1977a. Pr�sence du genre Grevea Baill. en Afrique centrale occidentale. Adansonia,
s�r. 2, 17: 119-123.
.1977b. Pr�sence du Ternstroenia polypetala Melchior (Th�ac�es) dans les mon-
tagnes camerounaises. Adansonia, s�r. 2, 17: 5-10.
LIBEN, L. 1971. R�partition g�ographique des Combretaceae africaines. Bull. Jard. Bot.
�tat 41: 445-447.
LOBREAU-CALLEN, D. 1975. Deux genres de Celastraceae: Cassine L. et Maytenus Mol.,
revus � la lumi�re de la palynologie. Adansonia, s�r. 2, 15: 215-224.
LOURTEIG, A. 1960. Une Lythrac�e dio�que: Capuronia madagascariensis gen. nov., sp. nov.,
de Madagascar. Compt. Rend. Hebd. S�ances Acad. Sci. 251: 1033-1034.
MCNEILL, J. 1974. Synopsis of a revised classification of the Portulacaceae. Taxon 23: 725-
728.
1978]


588 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VOL. 65
MAGUIHE, B., P. S. ASHTON, C. DE ZEEUW, D. E. CIANNASI & K. J. NIKLAS. 1977. Pakarai-
moideae, Dipterocarpaceae of the western hemisphere. Taxon 26: 341-385.
MARKGnAF, F. 1976. Apocynac�es. 169e Famille. In J.-F. Leroy (editor), Flore de Mada-
gascar et des Comores. Mus�um National d'Histoire Naturelle, Paris.
MEIJEu, W. 1974. Plant geographic studies on Dipterocarpaceae in Malesia. Ann. Missouri
Bot. Gard. 61: 806-818.
MELCIIOR, H. 1925. Theaceae. In A. Engler & K. Prantl, Die natiirlichen Pflanzenfamilien.
Ed. 2. Vol. 21: 109-154.
MORAT, P. 1976. Sur la pr�sence � Madagascar d'un genre end�mique d'Eriocaulac�es:
Moldenkeanthus. Adansonia, s�r. 2, 15: 463-469.
NOOTEnOOM, H. P. 1962. Simaroubaceae. In Flora Malesiana, Ser. 1, 6: 193-226.
PENNINGTrON, T. D. & B. T. STYLES. 1975. A generic monograph of the Meliaceae. Blumea
22: 419-540.
PERInUEI DE LA B�THIE, Il. 1936. Biog�ographie des plantes de Madagascar. Soc. Edit.
G�ogr. Marit. et Col., Paris.
S1948. Notes biog�ographiques sur quelques plantes ayant contribu� au peuplement
de Madagascar. M�m. Inst. Sci. Madagascar, S�r. B, Biol. V�g. 1: 113-132.
. 1951. M�lastomac�es. 153e Famille. In H. Humbert (editor), Flore de Madagascar
et des Comores. Mus�um National d'lHistoire Naturelle, Paris.
RAUII, W. & HI. REZNIK. 1961. Zur Frage der Systematische stellung der Didiereaceae. Bot.
Jahrb. Syst. 81: 94-105.
RAVEN, P. HI. & D. I. AXELROD. 1974. Angiosperm biogeography and past continental move-
ments. Ann. Missouri Bot. Gard. 61: 539-673.
RAYNAL, A. 1968. Les genres Neurotheca Benth. et lHook. et Congolanthus A. Raynal gen.
nov. (Gentianaceae). Adansonia, s�r. 2, 8: 45-68.
. 1977. Le genre Lilaeopsis (Ombellif�res) � Madagascar. Adansonia, s�r. 2, 17:
151-154.
REITSMA, T. 1970. Pollen morphology of the Alangiaceae. Rev. Palaebot. Palynol. 10: 249-
332.
SANGAI, G. R. W. 1971. Lecythidaceae. In Flora of East Tropical Africa. Crown Agents
for Oversea Governments and Administrations, London.
SCHUMAN, K. 1895. Bignoniaceae. In A. Engler & K. Prantl, Die natiirlichen Pflanzenfami-
lien. Vol. IV (3b): 189-252.
SLEUMEH, HI. 1971. Le genre Casearia en Afrique, � Madagascar et aux Mascareignes. Bull.
J. Bot. �tat 41: 397-426.
SMITr, A. C. 1972. An appraisal of the orders and families of primitive extant angiosperms.
J. Indian Bot. Soc. 50 A: 215-226.
-- & R. P. WODHOUSE. 1938. The American species of Myrtaceae. Brittonia 2: 393-510.
SMrrH, L. B. & E. S. AYENSU. 1975 . Velloziaceae. In Flora of Tropical East Africa. Crown
Agents for Oversea Governments and Administrations, London.
STEAHN, W. T. 1971. A survey of the tropical genera Oplonia and Psilanthele (Acanthaceae).
Bull. Brit. Mus. (Nat. Hist.), Bot. 4: 259-323.
STEVENS, P. F. 1970. Agauria and Agarista: an example of tropical trans-Atlantic affinity.
Notes Roy. Bot. Card. Edinburgh 30: 341-359.
. 1971. A classification of the Ericaceae siibfamilies and tribes. Bot. J. Linn. Soc. 64:
1-53.
S1976. The Old World species of Calophylhim I, The Mascarene species. J. Arnold
Arbor. 57: 167-184.
TAKHTAJAN, A. 1969. Flowering Plants. Origin and Dispersal. Transi. by C. Jeffrey. Oliver
& Boyd, Edinburgh.
THORNE, R. F. 1968. Synopsis of a putatively phylogenetic classification of the flowering
plants. Aliso 6: 57-66.
S1975. Angiosperm phylogeny and geography. Ann. Missouri Bot. Card. 62: 362-
367.
UPHOF, J. C. T. 1959. Myristicaceae. In A. Engler & K. Prantl, Die natiirlichen Pflanzenfa-
milien. Vol. 17 a (2): 177-220.
VELDKAMP, J. F. 1967. A revision of Sarcotheca B1. and Dapania Korth. (Oxalidaceae).
Blumea 15: 519-543.
VENKATESWARLU, J. & P. RAO. 1971. Wood anatomy and systematic position of Strephonemna.
New Phytol. 7: 767-771.
ViGuIER, R. 1944. Les L�gumineuses de Madagascar. Arch. Bot. 6: 1-820. [Only one copy.]


1978] LEROY-MADAGASCAR 589
VLIET, G. J. C. M. VAN. 1976. Wood anatomy of the Rhizophoraceae. Leiden Bot. Ser. 3:
20-75.
WALKER, J. W. 1971. Pollen morphology, phytogeography and phylogeny of the Annona-
ceae. Contr. Gray Herb. 202: 3-130.
WEBSTER, G. L. 1975. Conspectus of a new classification of the Euphorbiaceae. Taxon 24:
593-601.
WILI, H. 1975. Phytogeography and the Condwanaland position of Madagascar. Boissiera
24: 107-117.