·:~

·:·<::<!~";;t

Acta Botanica Hungarica 31 (1-4), pp. 3- 34 (1985}

·~· ~
\·,;""' ,.,··

PHYTOGEOGRAPHIC SURVEY OF CUBA

I. THE PHYTOGEOGRAPHIC CHARACTERISTICS
AND EVOLUTION OF THE FLORA OF CUBA

A. BoRHIDI

INSTITUTE OF ECOLOGY AND BOTANY, HUNGARIAN ACADEMY OF SCIENCES,

2163 VACRATOT, HUNGARY

(Received: 10 October, 1984)

Some of the main characteristics of the flora of Cuba are exposed and discussed,
as the dominance of endemics, disjunction, vicariancy, inversion, microphyllia, mic-
ranthia, relict character and vulnerability. The origin, > evolution and presumable
migrations of the flora are also discussed, based on the results of the geological up-to-
date investigations on the plate tectonics. In the process of the flora evolution three
main phases are distinguished, the early plate phase, the middle-tertiary l and-bridge
phase and the late archipelago phase. As the last important period of the flora im-
migrations the middle tertiary land-bridge phase is discussed with its different floras,
as the broad-leave d Honduras-flora, and the sclerophyllous Madro-Tethyan flora . In the
l at e archipelago phase the b road-leaved Guyana-flora, the semideciduous Yucatan-
flora and the extratropical North-American flora influenced the evolution of the flora
and vegetation of Cuba. At last the evolution centres of the ecologically and/or phyto-
sociologically adapted species groups and their migratory routes are designed and
explained.

The phytogeographic characteristics of Cuba

Some outstanding phytogeographic f eatures of Cuba

The floristic analysis and the study of chorological types suggest that
the most important and most typical characteristics of the flora of Cuba are
as follows: 1. Dominance of endemics, 2. Disjunction, 3. Vicariancy, 4. Inver-
sion, 5. Microphyllia, 6. Micranthia, 7. Reli~, and 8. Vulnembility.

The dominance of endemics

As shown earlier (BORHIDI 1982), the endemics co mprise a total of 51.4% of the native
flora. This is outstandingly the largest percentage in the Antilles. MA.RIE-VICTORIN (1942
1944, 1956) attirbuted this to the high Mg and Fe concentrations of "limonite areas". Accord-
ing to ALAIN's (1958 p. 16) hypothesis, the reasons are the early isolation of the flora during
the upper Miocene, the diverse edaphic soil conditions of the country and the arid climate
of certain areas.

The ecology of endemic speciation

In examining the ecological b ack ground of the speciation of endemics MuNIZ (1970)
found that eight habitat types have substantial influence on flora development. These habitats
occur in 25 areas in Cuba. The habitat types with the number of areas in brackets are:

I* Acta Botanica Hungarica 31, 1985

Aka<lemiai Kiad6 , Bu.clapest
4 A. BORHIDI

A. Ferritic soils and tropical brown soils derived from serpentine (6)
B. Ferralitic soils derived from serpentine (5)
C. Oligotrophic quartz-allitic yellow soils (2)
D. White sandy habitats (2)
E. Young, arid, coastal limestone areas ( 4)
F. Young, montane, karstic limestone areas (2)
G. Old, montane limestone karst, "haystack mountains" (1)
H. High montane areas
The geographic distribution of these habitats is shown in the Atlas Nacional of Cuba
p. 60. No doubt that the number of factors facilitating the speciation of endemics is large.
The influence and importance of these factors differed with time and space. Factors potentially
influencing speciation are listed below:

A. Insularity and isolation

a. Geographic factors
- external (insularity)
- internal (isolation)
b. Ecological factors
ba. Orographic factors
- altern'ltion of lowlands and mountains
- great relief-energy
bb. Geological and edaphic factor;
- serpentine and other ultrabasic roch
- frequent occurrence of limestone karsts
- frequent occurrence of acid, slatey bedrocks
- frequent occurrence of acid, white sands
be. Sociological factors
- comn1unity mosaics
- communities as barriers of migration
- interactions between plants and animals
B. Climatic changes
c. Alternation of wet and dry periods
d. Alternation of cool wet and warm dry periods
C. Genetic factors
e. Mutagenic speciation
f. Hybridogenic speciation, introgression
g. Genetic drift

Horizontal distribution of endemics

The study of the horizontal and vertical distribution of endemics yields valuable
information on the ecological effects influencing speciation and geographic range. Figure I
shows the number of endemic species per area. The intertidal mangroves and mangrove
swamps are the poorest in endemics. The lowlands of central Cuba and the southern part of
Isla de Pinos are moderately poor, having 25- 50 endemic species. Moderately rich areas
(50-75) are the younger coastal limestone habitats (Guanahacabibes peninsula, the southern
coast of Las Villas and the northwestern coast of Oriente), the slatey outcrops and white
sands in Isla de Pinos, the hilly regions of Central Cuba and the medium altitude zone of
Sierra Maestra. The rich areas, containing 76-100 regional endemics, are the slatey outcrops
of Pinar del Rio, the ancient serpentines of Cajalbana, the younger serpentine zones of Habana,

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 5

Matanzas and Motembo, the rainforest region of Sierra E scambray, the southern coast at
Sierra Maestra and the northern karstic zone (Guisa-B aire) in Oriente, and the limestones of
the Sagua-Baracoa massif. Most areas v ery rich in endemics (100-150) are on serpentine,
three are younger (Santa Clara, Cam agiiey, Holguin) and three other s are an cient latosol
area s (Nipe, Crista}, Moa-Baracoa). In resp ect of endemic richness only the ancient karsts
with conical formations in the Sierra de los Organos, the montane zone of Sierra Maestra
and the xerotherm semi-deserts of southern Baracoa are commensurable to the above localities.
Outstandigly rich a r eas are found only in the montane zone of ancient serpentine mountains,
such as the Nipe Mts., the surroundings of Pi eo del Crista} and the highlands of Moa (El Toldo,
Iberia, etc.) , where more speciation-inducing factors (age, serpentine, isolation, montane
character) are combined.

Vertical distribution of endemics

After reading the foregoing one would expect that the number of endemics increases
,· as altitude increases, due to the montane effects. However, as our studies in Pico Turquino
showed, the vertical distribution of endemics is a more complex matter. Figure 2 demonstrates
that the number of endemics increase up to the lower limit of the cloud zone and then r apidly
decreases. Its reason is that the vertical distribution of r egional endemics of lowlands and
mountains of m edium height differ s from that of the local montane endemics. The number
of species from the first group increases up to 1000 m , hardly changes up to 1500 m and then
falls on a sudden. The number of montane endemics, however, a bruptly increases between
1000 and 1500 m and does not change any more. The disappearance of lowland regional ende-
mics from the high altitude zone causes the decrease of endemic species (Fig. 2/ A). Contrary
to the absolute figures, the relative proportion of endemics rapidly increases with altitude.
That is, the few endemics occurring b ecome predominant at higher elevations (Fig. 2/ B).

Relationship between aridity and the number of endemics

STEBBINS (1953) pointed out that ecological drought stimulates speciation. The eco-
logical and chorological study of many Cuban endemics provides evidence for this phenomenon.
The largest number and highest density of endemics may be observed in the arid zones (semi-
desert, coast) and in the physiologically dry habitats (serpentines, limestone karsts, white
sand s). Also, the majority of endemics have xeromorphic leaves. Table 1, and Fig. 3 show the

Table I
The leaf-size distribution of 1115 tree and shrub species occurring
in 40 native forest communities of Cuba

L eaf.size N on e ndemics E n d emics Total

category
No. % No. % No. %
- ·- -- ·-· __ _,_ -· -- - -- - - ---- · - -- -- - - -- - · - - · - - - - .
Macrophyll 6 100 0 0 6 100
Mesophyll 58 65 30 35 88 100
Notophyll 96 43 125 57 221 100
Microphyll 163 33 337 67 500 100
N anophyll 45 22 155 80 200 100
Leptophyll 14 20 56 80 70 100
Aphyll 5 18 25 82 30 100

A cta B otanica Hungarica 31, 1985

lllllll~§D
~
..., U1 N
v m (l'1
II
I
Ui ...,' N
0 U1 0
0 U1
U1
0
U1
0

I I I I I

IGIHHOH 'V 9
 PHYTOGEOGRAPHIC SURVEY OF CUBA, L 7

leaf- size di stribution of 1115 tree and shrub species occurring in forty n a tive forest communities
of Cuba. A mong macrophylls and mesophylls the widely distributed species dominate, but in
the microphyll, n anophyll, leptop hyll and aphyll categories the endemics outnumber t h e
others and their proportion increases as leaf- size decreases.

A c
500 50

400 40
a
I b
300
/--- .......,
""0
c
~....::
30 i
li ~LL i
~ / \ ·-·b LL~ i
200
, /
/
i
/ 0
20 ,.~----r ..., c
,..... ....
\
/.
/ 52
\
i \ c
100
"' I
10
I
/ /
/ _,.. /

0 ·-·...,.,·" 0 .-----·""
0 500 1000 1500 2000 0 500 1000 1500 2000
m. s.m. m. s.m.

Fig. 2. Vertical distribution pattern of endemic species in Cub a (A), and their r elative fr e-
quency in the zonal veget ation types (B) along a vertical gradient. - a: total of endemic;,
b: lo cal montane endemics, c: r egional lowland endemics

%
80

70 /
/
60 /
/
Tropical spec ies
50 · / Caribbean sp ecies
I
- i,O · I
Cuban endemi cs

7'-,
30 ''
' '....... _...
!C 1 '~-
I I

~ -I
'j
i
I

i) - I i

Fig. 3. The leaf-siz e class pattern of Cub an trees and shrubs - mg: megaphylls, rna: macro-
phylls, ms: m esophylls, nt: n otoph ylls, mi: microphylls, n a: n anophylls, lp: leptophylls and
· aphylls

Acta Bot.anica Hun gar ica 31, 1985

8 A. BORHIDI

Disjunction
Another chorological characteristic of the flora is disjunction, the separa-
tion of the geographic range of a given taxon into several isolated areas.
Many different types of disjunction, such as bipolarity, hi-, tri-, and multi-
sectorial distribution may he observed in Cuba.

Bipolarity
The most striking type of disjunction is the bipolarity of the geographical range of
certain genera (SAMEK 1973) that occur only in western Cuba and Isla de Pinos, and also in
the mountainous regions of Oriente. This is the case with Purdiaea (Fig. 4), Heptanthus (Fig. 5),
Spathelia (Fig. 6), Pinus, Podocarpus and many other genera. The reason is that the eastern
and westem parts of Cuba has been isolated for a long period of time from the beginning of
the Tertiary. In most cases, the number of species in the disjunct genera is higher in Oriente
than in western Cuba. The larger area and more diversified landscape of Oriente cannot
explain this observation, since the Sierra Maestra emerged only at the end of the Tertiary,
and western Cuba has more div erse geological conditions than Oriente. A more acceptable
interpretation is that Oriente and the continent had been connected through Hispaniola for
a relatively longer period of time , facilitatin g a steadier "gene supply" for this province.

ESSl Purdiaea cubensis

rn Purdioea nipensis, parvifolio, moestrensis, moaensis, velutino eH;;:
Fig. 4. The geographical distribution of the genus Purdiaea in Cuba (after THOMAS 1960,
SAMEK 1973 and BoRHIDI 1973)

He pt onthus

[[]]] H. · cochl eor ifolius

• H. ra nunculoides o o H. cordifo liu s
"" H. brevipes + + H. Shofe~i
D H. lo botus x '-1 yumurien sis

Fig. 5. The geographical distribution of the genus Heptanthus Griseb. (after BoRHIDI 1972)

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I . 9

Geologists (ARNDT 1917- 1922; ScHUCHERT 1935) pointed out that the separation of western
and eastern, Cuba preceded that of Oriente and Hispaniola, although botanists, following
URBAN (1923 p. 52), disagreed. The present analysis of the flora , however, provides a botanical
evidence for the geological results. It is noted that G. SILVA (1979) arrived at the same con-
clusion in his studies on the Chiroptera fauna of Cuba.
At species level only a few examples of bipolar distribution exist. These species are
found mainly on serpentines, e.g., Amyris lineata, Vernonia angustissima, Croton bispinosus,
and H elicteres trapezifolia, but also in limestone karsts, such as Byttneria microphylla and
Neoregnellia cubensis.

Bisectorial geographic distributions

Another type of disjunction is represented by some rainforest species (Magnolia cuben-
sis, Hedyosmum grisebachii, Ocotea ekmanii, see Fig. 7) occurring both in the mountains of
Oriente and the Guamuhaya massif. Up to the early Tertiary, these regions were connected
as still indicated by the presence of the relict block-mountains of Sierra de Najasa. This con-

Spa thelia cubensis 0 S. vernicosa

pinetorum () S. subintegra
ISS3 S. Brittonii Wrightii o S. Leonis
S. stipitata + S. yumuriensis
t::. S. lobulate x S. splendens

Fig. 6. The geographical distribution of the genus Spathelia in Cuba (after MARIE-VICTORIN
1944, SAMEK 1973, complemented by BORHIDI)

Hedyosmum

[[[!] H. Grisebochii
§
•
H. crassifo lium
H . Leon is
.....•
H. subint egrum
H. cuben se
+ H. domingense

Fig. 7. Geographical distribution of the Hedyosmum species in Cuba (after MARIE-VICTORIN

1942, modified)

Acta Botanica Hungarica 31, 1985

10 A. BORHIDI

nection, however , must have come to an end in short time , since most common elements belong
to a n cient dicotyledonous familie s (Magnoliaceae, L aurace ae, Chloranthaceae, etc.). Other
two type s of the bisectorial distribution have developed between the mountainous r egion of
we stern Cuba and the Guamuhay a massif:
l. the plants of k a rstic forests and other limestone areas (Sapiwn leucogynum and
Miconia cubensis), a nd
2. species on serpentines in Pinar del Rio and the Escambray mountains (Coccothrina x
crinita and Linodendron venosum).

Trisectorial distributions
The occurrence of several taxa in three isolated spots (west ern Cuba, Guamuhaya
massif, Oriente) indicates that thes e areas had been separa ted for a long period of time in the
Tertiary as a re sult of shallow-sea transgression of the island. The genera of Lyonia and Vacci-
nium; some montane for est species, such as Tetrazygiopsis laxiflora an d Sapium pallens; and
some acidophilous sp ecies, e.g., Hypericum stypheloides exhibit such a trisectorial di stribution.
As a consequence of long er geographic and ecological isolation of the H ypericum stypheloides
populations, thi s species fall s into three subspecies (LIPPOLD 1971). Some taxa considered
earlier to hav e bi· or tri-sectorial distribution, proved to b e taxonomically heterogeneous, su ch
as Amyris lineata, Croton brittonianus, Ditta myricoides, Lyonia elliptica, and Vaccinium
cubense.

M ultisectorial areas
Several serp entine species (Figs 8- 9) have relatively l arge geographic range falling
into many isolated spots. In these cases the reason of disj unction is the scattered occurren ce
of a cceptable habitats , rather than the prehistoric geo graphic effects. Examples are Phyllan-
thus orbicularis, Neobracea valenzuelana, Rondeletia camarioca, and Jacaranda cowellii. This
type of geographical range is chara cteristic of the common species of the numerous isolated
serpentine areas, a s well a s the plants of the isolated coastal limestones (Castela calcicola,
.M achaonia havanensis, Dendrocereus nudiflorus, and N eobracea angustifolia.

Vicariancy
The third characteristic feature of the flora of Cuba is the abundance of
vicarious taxa. All types of vicariancy are widespread. This fact indicates
prehistoric isolations and allows the surveyor to make inference concerning
the influence of ecological factors on evolution and the development of certain
related taxa.

Geographic vicariancy
Many examp les exist in Cuba. Figure 10 shows the range of two endemic vi carious
gen era of lianes, L escaillea and Harnackia. Both occur in pinewoods and evergreen shrublands
on latosols d erived from serpentine . This serves as an evidence su ggesting a relationship not
only between the r ecent vegetation of the Cajalbana and Nipe Mountains, but also in the
past flora development of thes e regions. Further examples are A nemia coriacea (Fig. 11) and
Moacroton (Fig. 12), among many others. The ranges of Thrinax and Hemithrinax also ex-
emplify geographic vicariancy (Fig. 13). Thrinax radiata is a sp ecies of ro cky and sa ndy b eaches,

A ct a Bota nic<J Hungarica 31 , 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 11

whereas Th. morrisii and Th. drudei occur in the karsts of western Cuba replacing each other.
In central and eastern Cuba this genus is substituted by H emithrinax (H. ek maniana and
H. compacta). The ecological vicariads of them , namely H . riv ularis and H . savannarum,
occur on the serpentin es of Moa. The latter two are cenological vicariads of each other, being
forest and ~crub species, respectively.

• Ph ylla nth us orbi cular is Serpen t:ne a:eas

Oth er spec ies of th e The main migratory route
0 sectio Orbicularia of serpentine pl ants

Fig. 8. Geographical distribution of the serpentine areas and the Orbicularia section of the
genus Phyllanthus in Cuba (after WEBSTER 1958)

D V N. Vaten z uetano 6 ~J. Susann ina

A N. a ngu stlfOli'J • N. Mart ian a
• N. H owo ~di : T N. Ek monii
Fig. 9. The geographical distribution of the genus Neobracea Britt. in Cuba (after BoRHID!
19 73 and LIPPOLD 1979)

•
6 :... esc oillea equt settfo rmi s - Harnac kia bisec to
Fig. 10. The geographical distribution of the genera Lescaillea Wr. in Sauv. and Harnackia
Urb. (after BORHIDI 1973)

A cta Botanica Hun ga rica 31, 1985

12 A. BORHIDI

Anemia coriacea s.l .

& A. caj al ban ica

• A. cor iacea s. str.
o A. coriact=a ssp. moaensis
Fig. 11 . Geographical distribution of Anemia coriacea Griseb. s.l. (after BORHIDI 1973 and
DUEK 1975)

Moacroton

sect. glaucifoliae sect. Moacroton

•
M. trigonocarpus o M. cristalensis
& M. revolutus ·£:. M. Leonis
• M. lanceolatus T M. tetramerus
o M. Ekmanii
Fig. 12. Geographical distribution of the genus Moacroton Croiz. (after BORHIDI 1973)

Thrinax Hemithrinax

IZ2l Th. rarliata o H. Ekmaniana

• Th. morrisii E3 H. compacta
& Th. Drudei (§;! H. savannarum
Ill]] H. rivularis
Fig. 13. Geographical distribution of the genera Thrinax and Hemithrinax m Cuba (after
BORHIDI 1973 and READ 1976)

Acta Botanica Hunga rica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 13

Combinations of geographic and ecological vicariancy

The genus Platygyne (Fig. 14) exhibits both types of vicariancy discussed so far.
P. hexandra, an ubiquitous species found all over the country, is replaced by P. parviflora on
serpentines in Camagiiey and Holguin, by P. dentata in the montane zone of Sierra Maestra,
and by other four species in the serpentine mountains of northeastern Oriente. Furthermore,
the latter four species are also vicarious in their geographic distribution within the province.

Platygyne

~ P. hexondra .A. P. triandra

§ P. porvifolio • P. obovota
!IIIl P. dentate + P. volubilis
x P. Leonis
Fig. 14. Geographical distribution of the genus Platygyne Muell. Arg. (after BORHIDI 1972)

This observation supports the view that speciation is more intensive on serpentines than on
other bedrocks (BORHIDI 1972a). An interesting exception is the genus Neobracea (Fig. 9)
with two species on serpentine, N. ekmanii endemic to Moa and the Pan-Cuban N. valen-
zuelana. Although the latter occurs on all serpentine outcrops in Cuba, the morphological
variability of populations living under different climatic conditions remains within specific
limits. Conversely, the ancestral form of the limestone species has been broken into four
vicarious species along the coasts consisting of old and recent limestones. These species
constitute morphologically well-separated small populations with considerably isolated geo-
graphic distributions.

Ecological vicariancy at the infraspecific level

The ecological factors affecting speciation may be best investigated and demonstrated
through examples of infraspecific vicariancy. This term means that a young species of great
migratory and penetrative power colonizes new habitats which in turn modify the popula-
tions. Mayt enus buxifolia (Fig. 15), an Antille an species, is a good example. It falls into five
vicarious subspecies in Cuba: the widely spread ancestor, two subspecies on serpentines (ssp.
cajalbanica in western Cuba and ssp. serpentini in the east), another adapted to montane
belt (ss,p. monticola), and an extremely drought-tolerant subspecies in the semi-desert belt
(ssp. cochlearifolia). An endemic microphyllous drought-tolerant shrub, Reynosia mucronata
(Fig. 16) exhibits similar vicariancy. The ancestor occurs in the shrublands of central Cuba
with preference in favour of neither limestone nor serpentine. However, the populations on
the latosols of Nipe Mts. (ssp. nipensis) and the stands occurring in the semi-desert zone of
southern Baracoa (ssp. azulensis) have been geographically isolated subspecies (cf. BoRHIDI
and Mu:Nrz 1971).

Acta Botanica Hungarica 31, 1985

14 A. BORHIDI

Maytenus buxif ol ia

• ssp. bux ifolia

.6 ssp. cajalbanica Ji.. ssp. serpentini
o ssp. monticola \1 ssp. cochlearifol ius
F ig. 15. The geographical distribution of the subspecies of Maytenus buxifolia (A. Rich.)
Griseb. in Cuba (after BoRHIDI 1973)

Reynosio mucronat e

• ssp. mucronate Ji.. ssp nipens1s '9' ssp. azulensis

Fig. 16. The geographical distribution of the subspecies of Reynosia mucronata Griseb. in
Cuba (after BoRHIDI 1973)

Geographic vicariancy at the infraspecific level

The infraspecifi c geographic vicariancy indicates climatic changes in th e recent geo-
logi cal past and casts light upon the role of potential cenological barriers. Examples are found
in the serpentine mountains of northeastern Oriente. The climate of this r egion might have
been that of tropi cal rainforests in the Holocene. Since then , a gradually drying climatic
gradient has developed. As a result of this change some species, e.g., Cas asia nigrescens (Fig. 17),
Amyris stromatophylla (BoRHIDI and MUNIZ 1973), Manilkasa mayasensis and Anemia coriacea
(Fig. 11) fall into vicarious subspecies with large leaves and tall habit in Mo a and with small
leaves and body ~ize in the Cristal and Nipe Mountains.

Phytosociological isolation and vicariancy

Development of geo graphically vicarious subspecies may often be facilitated by the
existence of species-saturated communities or vegetation types in different habitats separating
the drifted populations, e.g., those of mountains alternating with more or less wide valleys.
Here, only the distribution of the taxa of Calycogonium rosmarinifolium s. l. (Fig. 18) is shown,
although the vicarious subspecies of Phyllanthus erythrinus (WEBSTER 1959) could have also
heen selected as a good example.

A~hl Botanica Hun_garie« 31 , 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 15

• •
••
• •

Cosasia nigrescens • ss p. nig res cen s .& ssp. moaensis

F ig. 17. The geographical distribution of the subspecies of Casasia n igrescens (Gris eb.) Wr. ex
Urb. in Cuba (after BoRHIDI 1973)

Cal ycogonium rosmar inifoliu m s.l .

• ssp brochyphyllum .& Cal ycogonium moa n urn

Fig. 18. The geographical distribution of the taxa of Calycogonium rosmarinifolium Griseb . s.l.
(after BORHIDI 1973)

Inversion of floristic elements

Another typical feature of the flora , observed mainly in western Cuba,

is inversion (SAMEK 1973). Certain genera of mountains or t emperate regions,
which otherwise occur positively in high mountains at similar latitude, are
found in Cuba at low elevation. Examples are Pinus , Podocarpus, Kalmiella ,
Vaccinium, Lyonia, Befaria cubensis, and Quercus sagraeana, etc. Inv er sion is
probably caused by two factors:
l. One of the cool periods during the Quaternary when these species or
their ancestors reached Cuba and became established in the lowlands and
hills, and

Act« B otanica Hungarica 31, 1985

16 A. BORHIDI

2. The abundance of serpentines and other rocks and soils (e.g., slate
and quartz sand) of poor nutrient supply, which might have facilitated the
adaptation of cool-resistant species to the increasing mean temperature (such
inversions may develop under the same conditions even today).
The vegetation inversion occurring on soils derived from serpentine is
discussed in BoRHIDI (1973, 1976) and BoRHIDI and Mu:Nxz (1980). Of course,
this is accompanied with the inversion of floristic elements constituting the
corresponding vegetation type. In addition to the above mentioned taxa, this
phenomenon is exemplified by the species of Myrica, Ilex, Lobelia, Laplacea,
and Baccharis occurring on serpentines or white sands.

Microphyllia

This is an important feature characteristic not only of the flora of Cuba

but also of the flora of the West Indies. As pointed out in earlier, aridity
must have had an important role in influencing speciation. Table 1 shows that
of the IllS species examined, 800 (72%) proved to be microphyll, nanophyll,
leptophyll and aphyll, most of those (563 species) being endemics. There are
numerous genera represented in Cuba and the Antilles by highly specialized
microphyll, sclerophyll or spiny sections or species, but the corresponding
taxa in the continent are broad leaved. Examples are Plinia, Myrcia, Calyptran-
thes, Eugenia, Rondeletia, Machaonia, Psychotria, Phyllanthus, Croton, Acaly-
pha, Jacaranda, Tabebuia, Byrsonima, Malpighia, etc. Typical West Indian
genera are Catesbaea, Scolosanthus, Ottoschmidtia, Acidocroton, Picrodendron,
Krugiodendron, and Sarcomphalus, Their abundance and substantial cover in
the vegetation allow for the conclusion that during the main period of the
flora development in the West Indies there was a direct relationship to the
continental flora composed of drought-tolerant elements. Then, the flora must
have further evolved in a warm, arid subtropical climate, probably in the
second half of the Miocene.

Micranthia

A striking feature of the flora of the Antilles, and particularly of Cuba,

Is its richness in small-flowered plants, whereas the floras of Palaeotropical
and Neotropical regions in the continents abound in showy, big-flowered
species.

Among many thousands of plant species illustrated in Botanical Magazine and Garden·
flora , Antillean species only occasionally occur. (None of the 3000 endemics in the exceptionally
rich flora of Cuba has big enough flower s to deserve the honour of being the 'national flower' ,
so H edychium coronarium, an Asiatic species, was given this title.)

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 17

The small flowers are pollinated by endemic, highly specialized micro-

scopic sized insects that are usually not capable of long distance flying. To avoid
strong winds blowing from the sea, these insects stay in the shelter of plants.
Thus, gene flow between remote plant populations is occasional, if not im-
possible. This barrier must have contributed significantly to the isolation of
populations subjected to genetic drift, and resulted in the formation of numer-
ous vicarious endemics. The fact that most vicariads differ very little from one
another indicates the influence of genetic drift. Therefore, some taxonomists
consider these vicariads as simple varieties. However, this view is unacceptable
since the morphology of these microspecies is stable, the segregation is complete
and no intermediate forms and genetic relationships exist. Micranthia may be
a partial explanation of the discrepancy that genera represented by only one
or two widespread species in the continent may fall into a dozen of endemic
species with geographic range restricted to a single mountainous region in
the Antilles.

Relict character
The flora of tropical moist vegetation types, such as lowland and sub-
montane rainforests and seasonal evergreen forests in the Antilles is relatively
young and contains few endemics. It is likely that these types had become
widespread only at the end of the Pliocene and during the pluvial periods
synchronous with the Pleistocene glacials. During the Pleistocene the flora
rich in Tertiary xerophilous and sclerophyll elements retreated. This process
still goes on. In my opinion the recent climate is far too moist for the sclero-
phyll vegetation types and the constituting species. The increase in the number
of permanently established deciduous and semi-deciduous species coming from
other community types supports this view. This degradation process is sig-
nificantly accelerated by human impact. The formerly widespread dry ever-
green formations (thickets, woodlands, forests and pinewoods) composed of
Tertiary sclerophylls are now restricted to relict habitats, serpentines and the
slopes and cliffs of conical karsts. The geographic range of endemic sclero-
phylls, which account for 40% of the flora, covers less than 20% of the area
of the island. The relict character of the flora is clearly demonstrated by the
presence of primitive, taxonomically isolated groups (Microcycas, Dracaena,
Cneorum, Spathelia sect. Brittonii and sect. Splendentes, and Harpalyce sect.
Cubenses, etc.), the abundance of disjunct geographical distribution types,
and the large number of local endemics represented by small populations.

Vulnerability
In general, island floras are vulnerable for several reasons:
1. The populations colonizing new biotopes are selected from a reduced
gene pool,

2 Acta Botanica Hung.arica 31 ~ 1985

18 A. BORHIDI

2. There has been no possibility of changing this gene pool for millions
of years,
3. As a consequence of isolation, the ecological tolerance and genetic
flexibility of populations decreased, so that,
4. The competitiveness of species generally diminished or even dis-
appeared in all but one respects,
5. Therefore, they cannot react satisfactorily upon new ecological im-
pacts, cannot take the advantage of succession, and cannot resist to or force
back new competitors.
Thus, island floras are usually composed of ecologically rigid popula-
tions not qualified to adapt to major environmental changes. The social
xecuperability is reduced, the populations and communities are vulnerable to
external effects. The vulnerability of the flora of Cuba is more pronounced
because
a) The island flora itself is constituted by set of ancient, isolated fl01·as,
b) Most endemics have been adapted to the extreme ecological condi-
tions of oligotrophic or bare areas. Thus, the l evel of metabolism became low
in these organisms, and their competitiveness diminished.
c) The sociability of many endemics is low. There are some rare taxa
represented by few populations, and few individuals therein. Consequently,
the flora of Cuba is one of the most endangered island floras of the world.
The list of endangered species of Cuba contains about a thousand items which
amount to 16.5% of the entire set of flowering species in the island. Approx-
imately 30% of the endemics are endangered.

The origin and migration of the flom of Cuba

Palaeobotanical and geological evidences
Our knowledge on the origin and migration of the flora of Cuba is
insufficient. As few as three noteworthy findings have been reported, one
probably from the Eocene and another from the Pleistocene of Sierra de
.. ..
Chorrillo, and the third from the Miocene of Yumuri Valley in Matanzas (see
GRAHAM 1973). The total number of taxa coming from these three periods is
still below fifty. Moreover, no palynological data are available from the
Quaternary of Cuba. Then, the only possibility to give a rough picture on past
changes in the flora is offered by the simultaneous analysis of the recent
flora, geographical distribution types and of the tectonic events r evealed by
structural geology.

Acla Bolanica Hu.ngarica 3 1, 198;)

 PHYTOGEOGRAPHIC SURVEY OF CUBA, L 19

Fig. 19. The geographical position of the Greater Antilles in the late Miocene according to
WEGENER's continental drift theory in the interpretation of CoRRAL (1939)

CORRAL's version of WEGENER's theory

The geological past of the Antilles has b een d es cribed by CoRRAL (1939, 1940) in
accordance with WEGENER's theory (Fig. 19). It is conceivable that portions of the Greater
Antilles us ed to be connected to one another and to South America, as CORRAL claims. How-
ever, these land s must have started to become separated from the continent not la t er than
the Upper Triassic. Due to chronological unreliabilities and consequential inconsistencies in
CoRRAL's theory, it was soon rejected (ALAIN 1958) and ScHUCHERT's (1935) concept s were
accepted.

ScHUCHERT's concepts on the T ertiary in the Antilles

According to ScHUCHERT, the Greater Antilles had been connected to Central America
via a land strip of Honduras from the Triassic to the Late Pliocene, apart from short intervals
in the Upp er Cretaceous and the Upper Oligocen e. Eastern Cuba was connect ed to the con-
tinent via the Cayman Ridge and indirectly via the Nicaragua Ridge, Jamaica and Hispaniola.
In addition, there was also a connection to the Bahamas and Florida during the Upper Creta-
ceous and the Eocene. In the Upper Oligocene Cuba became separated from Hispaniola and
Yucatan, and its connection to the continent ceased to exist any longer. Although in the !\Iiddle
:Vliocene Cuba and Hispaniola had been united again for a short time, no link was formed
between the Antilles and Honduras. In the Upper Miocene Cuba finally became co mpletely
separated from all the other islands of the Antilles and the continent. According to ScHUCHERT,
the Hondura s- .T amaica-Hispaniola land strip was finally broken only at the beginning of the
Pleistocene. In my views, however, this break up must have been completed in the Miocene,
as suggested by the very primitive mammal fauna of the Greater Antilles.

The influence of ScHUCHERT's concepts on phytogeographical inter-

pretations
The r ecent position of the West Indies, which are wedged in between two continents
and are surrounded by narrows and closed bays, makes the impression a s if the West Indies
were a sort of appendage of the continent that b ecame separated not long b efor e. Some
botanist s consider the flora of the West Indies as a descendant of the continental flora in which
the widespread and highly variable continental species are represented by end emic subspecies

2* Acta Botanica Hungarica 31, 1985

 !;-
;;
b;
[:J 1
I ~
0
~
E. Q 2
2
:.:: D 3
,;;• ~ 4
~·
" D 5
~ ~ .6
.....
"'"'"' ~ ~7
~8
~9

\>
tl:1
0
~
8...

Fig. 20. Origins of the flora of Cuba (after ALAIN LJO GIER 1958, stron gly modified). l. Continents and islands, 2. Shallow seas (less than
1000 m deep), 3. Deep seas (depth more than 1000 m), 4. West-Cuban phytogeographical subpr ovince, 5. Middle-Cuban phytogeographical
subprovince, 6. East-Cuban phytogeographical subprovince, 7. Main migratory r outes of the flora, strong floristic affinities, 8. Important
migratory routes, notable floristic relations, 9. Less important m igratory r ou tes of the flora, recognizal>le floristic relations
,. '• c'
 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 21

or varieties. This is, however, an illusory view formed under the influence of SCHUCHERT's
theories discussed above. Given this knowledge, the first fundamental statements regarding
the origin of the flora of Cuba were made by ALAIN (1958 pp. 33-34). He suggests that the
flora is mainly of Neotropical origin and the elements immigrated from five different direc-
tions (Fig. 20). This problem will be considered later in more detail.

The influence of plate tectonics

The results of contemporary plate tectonics gave an impetus to the research of flora
evolution. Several new hypotheses were created concerning the evolution and migration of
primitive floras and the radiation of Gondwana elements, etc. At the same time, controversies
among experts also revived. Several influential biologists, e.g., THORNE (1973) and VAN STEENIS
(1962) in botany, maintain their conservative view. They claim that the division and drift
of continents need not be assumed to understand the recent state of flora s. As VAN STEENIS
asserts, for example, the Pacific and the Indonesian flora may have been directly originated
from Asia via land bridges that have now been covered by sea. Ev en if the land bridge theory
is applicable to the Indonesian flora, there are still many paleogeographic and biogeographic
problems that cannot be solved in this way. It is therefore obvious that most botanists call
in the theoretical possibilities offered by plate tectonics, even though this field of geology has
its own unanswered questions.

Plate tectonics of the West Indies

Despite many controversial details, most geologists agree that the Caribbean r egion is
a separate tectonic unit. After ISACHS et al. (1968) and MoRGAN (1958) the existence of a
separate Caribbean Lithosphere Plate is generally accepted. As far as the origin and the
characteristics of this plate are concerned, the views are varying. ScHUCHERT (1935), WooDRING
(1954), EARDLEY (1954), S UTTERLIN (1956), TIKHOli:IIROV (1967), and KHUDOLEY and FURRAZOLA
(1971) think that the Caribbean Sea took the place of a land mass. H. A. MEYERHOFF (1954),
HEss and MAXWELL (1953), WEYL (1966), A. A. MEYERHOFF (1967), DENGO (1968), MoLNAR
and SYKES (1969), and MACGILLAVRY (1970) presume the existence of a permanent basin in
the ocean surrounded by island bends and geosynchlinals.
Most experts (BALL and HARRISON 1969, DENGO 1969, MoLNAR and SYKES 1969,
MACGILLAVRY 1970, MALFAIT and D~NKELMAN 1972, MATTSON 1973, and ITURRALDE 1975,
1978) explain the development of the Caribbean Plate on the basis of continental drift and,
in general, of the dilatation of ocean floor.
The unanimous view of the above authors is that the Caribbean Plate was primarily
oceanic. They exclude the possibility that the Antilles were connected to either part of the
American continent up to the Middle Miocene. In the Upper Jurassic the Caribbean Plate
may have been located at the northwestern edge of the Afro-American land mass in the Pacific
Ocean at the western mouth of the Tethys Sea (where Nicaragua and Honduras are now
found). In the Cretaceous, simultaneously with the process separating South America and
Africa, the South American continent was significantly rotated causing the Caribbean Plate,
which was connected to the Guyana Plate, to move eastward. Having moved about 800-
1200 kilometres, the Caribbean Pla te drifted away into the Tethys Sea. As from the Upper
Miocene the Antilles exhibited a platform-like development. According to lTURRALDE, Cuba
has been formed of six archaic isolated blocks, Hispaniola probably of three blocks, and
Jamaica and Puerto Rico of one block each. In the Upper Miocene started the general vertical

Acw Botanica Hungarica 31, 1985

22 A. BORHIDI

movement leading to the emergence of the Cayman and Nicaragua Ridge. Thus, the archi-
pelago and Central Ameri ca became connected. This process was completed in the Pliocene
and then the islands started to get separated. Finally, the recent shape of the islands was
formed in the Pleistocene.

The theory of a three-phase evolution of the W est Indian flora

Although Caribbean Plate tectonics poses many unresolved problems

(HowARD 1980 p. 244) it is now attempted to outline the general history of
the Antillean flora. Three major periods are r ecognized, each corresponding to
a given geological stage:
I. Plate phase,
2. Land bridge phase, and
3. Archipelago phase.

Plate phase
The name of the fir st sta ge indicates that at thi s time the Caribbean Pla te was an
indep endent land mass in the sea. It is assumed tha t this p eriod started in t he Middle Jurassic
and ended in the Upper Oligocene. R elatively few elements of the flo ra may be traced b a ck
with complete certainty to this phase. These important taxa are either phytogenetically old
relicts, or endemics with distant relations, or genera with disjunct geographic distribution.

Relict cryptogams

As exp ected, the flora of this early p eriod is m ainly rep res ented by cryptogam s in the
r ecent flor a. Such plants are, for inst ance, the tree ferns . As many as seventeen endemic tree
ferns occur in the Antilles. It is well-known that the fa mily Cyatheaceae includes wid ely
distributed species that are efficiently dispersed through spores. T he fact that one of the evo-
lut ionary centres of the Schizaeaceae family, and its genus Anemia, is in Cuba (12 species of
which sev en are end emics) needs similar interpretation. In this r espect, new bryogeographic
r esults (REYES 1982, BoRHIDI et P6cs 1985) deserve particular attention as they indicate a
d efinite r elati onship to Gondwan aland. For instance, t here are m any more liverwort species
th an mosses in Cuba. Such a partition of bryophytes is known to be t ypical of the islands in
·-
th e southern hemosph ere. The new studies b y P6cs an d R EYES r evealed obvious r elationships
b etween the liverworth flor as of the Guyana Highla nd and of the serpentine plateaux in Cuba,
e.g. b y the occurrences of Plagiochila steyermarkii Robinson and Calypogeia venez uelana . . .:
Fulford. It is well-known that the Guy ana Highland is one of the oldest block mounta ins in
South America d erived undoubtedly from Gondwanaland.

Relict phancrogams

The monotypical gymnosperm genus Mi crocy cas is a " living fossil" representing the
phanerogams of the plate phase. The assumption that at that time the Caribbean Plate was
located at the border of the P a cific Ocean and the T ethys Sea has not only b een jus ti fied by
fo ssil Cretaceous faunas (western Cuba, Viiiales) but h as also been demonstrated b y the
close r elationship between Colpothrinax, a Cuban-Central American genus, a nd the Pacific
P ritchardia. These genera probably ha ve a common ancestor.

A cta B otanica Hu ngarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 23

Peri- Afro- American elements

The evolution of the specially distributed p eri-Afro- American elements may a lso be
dated back to the plate phase. Table 2 lists typical peri-Afro-American phanerogam genera
that occur in the tropical America, East Africa and/or Madagascar but are absent from West
and Central Africa.
Of the fifteen genera listed, eleven occur in Cu ba, four genera (Carpodiptera , S avia,
Oplonia, and Stenandrium) have their evolutionary centres in the A ntilles (BORHIDI 1982).
According to STEARN (1971) this distribution type may h ave ·been develop ed as a
result of dra stic climatic changes. Many taxa became extinct in the interior of the "Afro-
Am erican" supercontinent due to the extreme continental climate. The flora of Central Africa
became relatively poor in species. On the contrary, along the coasts of Madagascar, East
Africa , North Africa and the northern part of South America an extensive coastal zone w as
formed which had a temperate climate. The abundance of diverse habitats in thi s r egion
was advantageous for survival so that a very rich vegetation developed. B esides the Mada-
gascan Plate, the Macaronesian and Caribbean Plates were probably also parts of this coastal zone.
In addition to the genera mentioned so far, a similarly close relationship is indicated
for Dracaena cubensis from Cuba, D. draco from Macaronesia and D. reflexa from Mad agas car.
The primitive Cneoreaceae fa mily may al so be mentioned as a good example (BoRHIDI 1982).
It has only three living species, one in the Mediterranean of Europe, another in the Canar y
Islands and the third in eastern Cuba on the southern slopes of Sierra Maestra. Further
evidence confirming the existence of these relationship s is that amongst the cryptogams there
are not only genera but also some species exhibiting peri-Afro-Am erican distribution (e.g.,
liverworts).

Table 2
Examples of peri-Afro-American genera

Total S. America, W. and E. Africa,

Genus, Family species W. Indies Central Madagascar
Africa

D esmanthus (Fabaceae) 22 15 1 6
Echinochlaena (Poaceae) 7 6 1
Ocotea (Lauraceae} 400 380 3 18
Oliganthes (Asteraceae) 21 12 9
Carpodiptera (Tiliaceae) 6 5 1
Oplonia (Acanthaceae) 18 13 5
Phenax (Urticaceae) 28 25 3
Piriqueta (Turnera ceae) 28 24 1 3
Rheedia (Clusiaceae) 50 37 13
Ravenala (Musaceae) 2 1 1
Savia (Euphorbiaceae) 25 15 1 9
Stenandrium (Acanthaceae) 25 22 3

It is assumed t hat some primitive angiosperms have also b een present in Cuba sin ce
t he plate phase. These are genera with h i- or trisectorial distribution, e.g., Podocarpus, M agno-
lia, Talauma , Ocotea, Persea, Guatteria, Oxandra, Jiedyosmum, Dorstenia, Bonnetia, Laplacea,
Freziera, L agetta, Linodendron, Myrcia, Spathelia, Purdiaea, and Calophyllum and those list ed
in Table 2.

Acta Botanica Hunga rica 31, 1985

24 A. BORHIDI

Land bridge phase

The second period which lasted from the end of the Oligocene up to the end of the
Pliocene may be termed the land bridge phase. At the beginning of the Oligocene started the
Caribbean Plate to emerge. As a result of this process the Greater Antilles became connected
to the continent first via Honduras and the Yucatan peninsula and later via the newly emerged
Lesser Antilles. This is the period of the large scale immigration of species into Cuba because
apart from facilitating migration, the emerged lands offered extensive bare areas for the
pioneer vegetation.

The broad-leaved tropical Honduras flora

During the Oligocene and the Lower Miocene a tropical flora consisting primarily of
broad-leaved evergreen trees and shrubs may have immigrated into the Greater Antilles.
In this period Cuba had been divided up into six isolated blocks: 1. Guane, 2. I sla de Pinos,
3. Villa Clara, 4. Camagiiey, 5. Sierra Maestra, and 6. Moa-Baracoa. In the Middle and Upper
Oligocene Guane was connected to the Yucatan peninsula. The Sierra Maestra block and
Honduras were connected via the Cayman Ridge. The Moa-Baracoa block joined probably
with Hispaniola and formed a part of the Honduras- Jamaica-Hispaniola-Puerto Ri co range
which was the main route of immigration. Obviously, the relatively small Guane block, that
became separated earlier from the continent, accumulated a more limited gene pool than the
larger Oriente blocks (Sierra Maestra and Moa-B aracoa) which had been connected to the
other parts of the Greater Antilles for a longer period of time. The plants of the latter blocks
comprise the so-called Honduras flora which might also be termed the Swietenia-Trichilia-
Zanthoxylum flora after its typical genera. Most Neotropical elements of the Greater Antilles
have been originated from the Honduras flora. According to my hypotheses , it is the phase
when three main groups of tropical genera immigrated. These are:
1. Genera represented by few but widely spread endemic species in the recent flora
( Swietenia, Trichilia, Cedrela, I nga, Albizzia, Bur sera, Dipholis, Pseudo/media, Chlorophora,
and Simaruba).
2. Genera with secondary evolutionary centres evolved due to climatic and geological
changes in the Antilles. These genera have many endemic representatives in the recent flora
(Pithec ellobium, Cassia, Caesalpinia, Copernicia, Eugenia, Calyptranthes, Zanthoxylum, Pilea,
Rondeletia, Psychotria, Guettarda, Miconia, Ossaea, Phyllanthus, etc.), and
3. Presumably extinct genera that are ance stors of typical endemic genera or genus
groups characteristic of the Antilles, e.g., the group Portlandia-Cubanola-Thogsennia; the
group Schmidtottia-Isidorea; the group Phialanthus-Ceratopyxis-Shaferocharis ; the group Antil-
lia-Urbananthus-Spaniopappus; the group Scolosanthus-Eosanthe; the group Pinillosia-Tetra-
perone-Koehneola; and the group of Espadaea-Goetzea-Henoonia, etc.

The sclerophyllous broad-leaved Madro-Tethyan vegetation

AxELROD (197 5) recognizes that the high similarity between the mediterranean vegeta-
tion of North America and Eurasia-N. Africa is not merely a physiognomic resemblance
because most genera occurring are common to both regions. He shows that a sclerophyllous
broad-leaved vegetation developed in the Eocene and Oligocene of Sierra Madre (Mexico)
and the southern United States, and also in the Oligocene and Miocene of southern Eurasia,
with many species in common. This vegetation, quite uniform in appearance, was named

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 25

the Madro-Tethyan v egetation after its westernmost limit, Sierra Madre, and the Tethy s Sea,
along the northern coast of which this veget a tion typ e was spread. It was widely sprea d, too
in the Upper Miocene of Central Europ e. For example, the Sarm atian flora describ ed from
Hun gary (ANDREANSZKY 1956, l962a, b) is also designated by AXELROD as b eing parts of
t he Madro-Tethyan vegetation. This sclerophyllous broa d-leaved vegetation, which was formed
by plants highly a dapted to the increasing summer droughts in the Tertiary, h as now b een
splitted into remote , isolated stands living under very different ecological conditions. In
California and Europe the summer is dry and the winter is wet, in Arizona and New Mex ico
there are two dry seasons. This veget ation typ e occurs under the dry winter a n d rainy sum mer
climate of Mexico and the West Indies. Moreover, extensive stands are also found in places
where the clim ate is wet throughout the year, but ther e are some sufficiently dry or nutrient
. !f
poor habitats thanks t o edaphic effects (e.g., white sand, serpentine, grani te-gneiss dom es ,
lava, lim estone dog-tooth areas, etc.).

The southern link

AXELROD p ointed out correctly that the laurel fore st s of the Canary I slands b elong to
the Madro-Tethyan vegeta tion, representing its southernmost portion. However, he did not
think of the sclerophyllous communities of the Antilles whi ch are also a part of that section.
In the Miocene Honduras, the Yucatan p eninsula and the Greater Antilles, which w ere con-
nected to the first t wo, formed the southern coast of the Tethys Sea in America. These areas
were similar in effect to the northern coast of Afric a and the Canary Islands on the other
side of the Tethys Sea. Accordingly, a special t yp e of the sclerophyllous Madro-Tethyan
v egetation, which should preferably be t ermed the Peri-Tethyan sclerophyllous vegetation,
develop ed in the Greater Antilles. I n addition to the Madro-Tethya n elements, this Antillean
t y p e is characterized by its richness in Neotropi cal sclerophylls and particularly in endemics,
and also by its isola tion. No doubt that t he serp entin e charrascals (dry evergreen shrublands),
pine-oak woodlands, pine-laurel forests, and pine-Dracaena for ests of Cuba are relicts left of
the sout h ern p ortion of the Madro-Tethyan veget ation.

The 111adro- T ethyanflora in Cuba

The sclerophyllous broa d-leaved Madro-Tethyan flora immi grated into the Antilles
presumably in the Lower a nd Middle Miocene. It was probably a dire ct continuation of the
form er invasion of the Honduras flora . In thi s p eriod an increasingly arid sub tropical clim ate
was predominant not only in the tropical zone but al so in mo st parts of the warm t emperate
zone. More recently, this unusual expansion of the subtrop ical zone is a ttributed to a t em-
porarily formed ring around the Earth, similar to that of the Saturn. This dry period dis-
played a strong influence on the evolution of the Cuban flora , in which the sclerophylls are
still dominant. 75% of the endemic sp ecies are micro- or nanophanerophytes (2265 sp ecies),
most of them (86 o/o) b eing micro- , n a no- and leptophylls. Many genera typical of the Madro- \
Tethyan vegetation immigrated into Cuba and b ecam e important elements of the flora sup-
posedly in this period. Th ese genera are: Pinus, Juniperus, Quercus, ]uglans, Buxus, fl ex,
Lyonia, V accinium, Pieris, Kalmiella , Berberis, Celtis, Pnmus, M y rica, Acacia, Calliandra,
Erythrina, Karwinskia , Zizyphus, R ey nosia, Colubrina, Thouinia, Dodonaea, Myrtus, Psidium,
Euphorbia, Amyris, Helietta, Auerodendron, S arcomphalus , Rhaco ma, Forestiera, Saba/, Cocco-
thrinax, Maytenus, Paepalanthus, Lachnocaulon, Syngonanthus, Eriocaulon, Chaetolepis, B e/a-
ria, and Garrya.

Acta Botan ica Hungarica 31 , 1985

26 A. BORHIDI

N cotropical sclerophylls in the Madro-Tethyan vegetation of Cuba

The southern part of the Madro-Tethyan vegetation has its own typical features.
N umerous species of the earlier established tropical broadleaf Honduras flora became adapted
to the n ew arid climate, and several secondary evolutionary centres developed. Different
Neotropical fa milies and genera, represented by membranous leaved macrophy lls and meso-
phylls in the continent, suffered a change so that new species groups or entire sections con-
taining only sclerophylls and coriaceous leaved microphylls app eared in Cuba. Such genera
are, for example, Croton, Phy llanthus, Harp alyce, Miconia, Ossaea, Plumeria, Jacquinia
Plinia, Eugenia, Diospyros, Antirhea, Exostema, Rondeletia, Guettarda, Machaonia, Melocactus,
and Gesneria, etc. The endemic genera of the Antilles, Hisp aniola and Cuba, that are typical
of the sclerophyllous vegetation t ypes, may also be considered as results of this climatic
change. Examples are Acrosynanthus, Phialanthus, Neomazaea, Ariadne, Phyllacanthus, Sar-
comphalt£s, Doerpfeldia, Pachyanthlts, Scolosanthus, Catesbaea, Notodon, N ashia, Sauvallella,
Leucocroton, Phidiasia, L escaillea, Harnackia, Phania, Shafera, Moacroton, Hemithrinax, Ceutho-
carpus, Schmidtottia, Roigella, Suberanthus, Kodalyodendron, H enoonia, Linodendron, Adenoa,
and Tetralix.

Drought resistant ecotypes

The numerous newly occurred sclerophyllous species and genera w ere only one indica-
tion of the adap tation to the dry climate of the Miocene. The membranous l eaved trees of
the Honduras flora became also adapted to aridity by d eveloping drought resistant ecotypes.
Although there are ver y little morphological differ en ces b etween these ecotypes and the
corresponding Central American p opulations the habitats of these species are apparently
drier in Cuba t han in the continent. Examples are: Ceiba pentandra, Trichilia glabra , T. hirta,
Cupania macrophylla, Bursera simaruba, etc.

Archipelago phase

The third period is termed the archipelago phase. It came about in different points of
time over the Greater Antilles. In Cuba it started at the end of the Miocene, when the Bartlett
Trench, that separat~s Cuba from Jamaica and Hisp aniola, was fo rmed. Of these islands
Jamaica had been conne cted t o the continent for the longest time, as clearly indicated b y the
floristic composition of this island (AsPREY and RoBBINS 1953, ADAMS 1972 ). Changes t ypical
of thi s period are the ecological specialization of floras isolated from one another and, as a
consequence of this , the internal migration of species. In this phase the flora and fauna of
Cuba were subjected to severe climatic and geological changes. In the Pliocene and the
Pleistocene the wet climate fr equently alternated w ith d ry periods. The effect s of this climat e
are clearly indicated by the stratification of ferritic soils in the Nip e Mts. Cooler p eriods, to o,
were frequent in the Pleistocene. Large scale tectonic events took place simultaneously with
the climatic chan ges. The mo st effective of them was that the six, formerly separate land
blocks emerged and joined together to form a single island with extensive bare are as. Tht'
colonization of t hese lands required further differentiation and migration of the flo ra and the
adaptation of species to the new environmental conditions. Other external effects, inde-
pendently of the geological changes, also influ enced t he flora of Cuba.

Acta Bolanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 27

The broad-leaved rainforest flora of Guyana

The direct land connections between Cuba and Central America disappeared by the
end of the Miocene. Later, in the early Pliocene, the climate became wetter. In the second
half of the Pliocene the arch of many young volcanoes, i.e., the Lesser Antilles, emerged and
the islands of the Greater Antilles also reached the maximum elevation. Presumably, this is
the period when the South American elements immigrated into Cuba via the Lesser Antilles,
although at that time land strips no longer existed. Yet, the genera Carapa, Ochroma, Coussa-
rea, Tocoyena, Paratheria, Phinaea, Proustia, Guarea, and the Myrsine guianensis group must
have been originated in this way.

The broad-leaved semi-deciduous Yucatan flora

It strikes one that there are numerous species common to the semi-deciduous fore sts
of lowlands and hill-countries in western Cuba and Yucatan. Some of them occur solely in the
Guanahacabibes peninsula and Yucatan. As examples the genera Hirtella, Poiretia, Eriosema,
Belotia, Luehea, Schwenckia, Chimarrhis, Calycophyllum, Elaeagia, D eherainia, Ateleia, Forch-
hammeria, and Neoma cfadya are worth mentioning. Although the presence of these elements
suggests a direct relationship between the Miocene flora of Mexico and Cuba, there are neither
geological nor biogeographical evid ences supporting the existence of land connections in
that period.

Elements from the temperate North America

The next most important region influencing the flora of Cuba, particularly western
Cuba, is Florida and the southeastern United States. Several elements, such as Pinus, Quercus,
Fraxinus , and many species of Ericaceae, Gramineae, and Cyperaceae, etc., of the temperate
zone immigrated from the north and became predominant in some parts of western Cuba.
ALAIN (1958 p. 19) presumes the existence of a former land connection between western Cuba
and Florida, although according to geologist s this possibility has been excluded since the
Eocene Epo ch. In my view, the flora from the temperate zone in western Cuba d eveloped
during the glacial periods in the Pleistocene, that took place simultaneously in the whole
continent as pointed out by VANDER HAMMEN (1961) and VEILLEUMIER (1971). Consequently,
instead of assuming a land strip it is more re asonable to trace ba ck the Florida- Western
Cuba relationship to the Madro-Tethyan flora. As already seen, this sclerophyllous evergreen
flora originally contained several genera from the families of the temperate zone. In addition ,
there was sufficient amount of time during the Miocene available for a fairly uniform vegeta-
tion to d evelop, first on the coast of the Tethys Sea and lat er, following th e emergence of
Central America, along the Gulf of Mexico. This flom was exposed to the cooler climatic
periods occurring many times in the Pliocene and the Pleistocene. In particular, the coastal
region of the Gulf of Mexico was affect ed, since the cold Labrador Current often rea che d this
zone during the Pleistocene. The new environmental conditions, e.g., modified so il type s,
an:l the ad aptive responses of plant populations generated by this cooling process were
probably similar on both sides of the gulf. The standardization of the flora may have been
furthere d, too, by the strong northerly and westerly winds during the glacial periods and by
the more intensive bird migration. In Cuba, the expansion of this flora was also promoted by
the newly emerged lands and white sand s that offer ed extensive bare areas for plants. Thus,
the r ecent flora elements common to western Cuba and Florida are remnants of a Pleistocene

Acta Botanica Hungarica 31, 1985

28 A. BORHIDI

flora sprea d along the entire coast of the Gulf of Mexico. The continuity of this flora was
broken in the postglacial period, when the Central American subtropical flora retreated to the
western coast of the gulf, forcing back the remnants of the Madro-Tethyan flora to a north-
westem and northeastern direction.

R emarks on the driving for ces of migration

In connection with the above discussion, it has to he emphasized that very little is
known about the water and a erial dispersal of seeds. Seed dispersal is likely to be more im-
portant than generally thought, and the explanation of flori stic relationship s doe s not neces-
sarily require a search for disapp eared land strip s. Only confirmed geological re sults may
s-erve as a basis for hypothese s on the distribution, development and migration of the flora.
Moreover, even undoubtful land connections may prove to he of no help in promoting dis-
persal. For instance, Befaria cubensis, Microcycas calocoma, and Rhus copallina, did not reach
I sla de Pinos despite the direct terrestrial contact to Cuba in the Pleistocen e. Although Fraxi-
nus caroliniana ssp. cubensis produces easily flying seeds, it could not become established on
m oorlands outside the Zapata peninsula. Many gen era (e.g., Victorinia) and species (e.g.,
Omphalea commutata) are still restricted to an area along the former contact line between
Cuba and Hispaniola (Fig. 21) although the ecological conditions would allow a larger geo -
graphic range on both islands. Many elements of the Bahamas, that m anage d to get across
the wide and deep Bahama Trench and became established on the reefs around northern
Camagiiey, Tunas and Holguin, were unabl e to penetrate into Cuba through the narrow and
shallow bays and even via the former land strips. Consequently, a terrestrial contact may not
necessarily be enough by itself and cannot he a single explanation for migration. The migratory
activity of taxa does not always coincide in time with the possibilities. ln other cases in-
favourabl e ecological and cenological conditions, su ch as the saturation of the flora, may
prevent otherwise feasible migration.

Flora migration in the interior of Cuba

The chorological groups reflecting the origin of the Cuban flora, the
approximate time of their manifestation and the potential migratory routes
have already been discussed. These are, however, only the initial steps toward

"' Omph oleo comm ut oto

• Pe t it io Urbon ii

~
Fig. 21. Geographical distribution of Omphalea commutata Mu ell. Arg. and P etitia Urba nii
Ekm. (after ALAIN 1972 and BoRHIDI 1973)

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, L 29

a complete flora development. The actual floristic composition is a result of

further differentiation, adaptation and internal migration. Two fundamental
principles should he considered in this regard.

Two principles of internal floristic changes in Cuba

A. Oriente is the earliest cradle of the Cuban flora, a starting point for the most im-
porta nt migrations.
B. The floras of mountains are older than those in the lowlands. The flora and vegeta-
tion of plains are originated from the mountainous r egions.

Three stages of internal migration

According to time, starting point and direction, three stages of internal migration can
be distinguished.
a. M igrations in the early T ertiary (Eocen e to Miocene Epochs). The centre of these move-
ments, that were directed from the east to the west, was the mountains of Oriente. The low
differentiation and specialization level of the flora is typical of this stage. Several migratory
waves may be r ecognized , these are associated with different periods. The radius of action w as
determined in each case by geological events, transgression and the emergence of lands, that
actually happened . The serp entine and karst flor a of western Cuba and then the montane
flora of the E scambray Mts were originated in this way. Later, the submontane tropical flora,
which subsequently colonized the rising plains, took r efuge in the valleys .
b. liiigrations at the end of the Tertiary and in the Quaternary (Pliocene and Pleistocene
Epochs). Starting from both tip s of the island the migrations were directed toward the interior.
As a result, a new flora in the central plains of Cuba became established at the end of the
Pliocene. The migr ations of alternate direction in the Pleistocene are also mentioned here.
During cooler periods these migratory waves swept over Cuba from the west to the east,
but in the warm p eriods the migration was of the opposite direction. This is the time of eco-
logical an d cenological adaptation and specialization. The vegetation zones and , among
others, the formatio n of the endemic montane flora of Turquino may also b e dated back to
thi s stage. Furthermore, the homogenization of the flora and v egetation, that were originated
from different mountains, m ay have b een caused by the repeated " shake up" of the flora
of plains.
c. Migrations driven by anthropogenic forces during the historical ages. The starting point
for these migrations is usually wes tern Cuba, so the direction is toward the east. The r eason
for this is that west ern Cuba was the sanctuary of many sava nna and prairie elements in the
Pleistocen e, that found refuge particularly in the herb layer of coniferous for ests. The migra-
tion of these plants was made possible by tropical deforest ation, which is mainly responsible
for secondary savannas and the recent landscape of Cuba. In addition to the migration of native
plants, many species were introduced most recently by man. These elements, too, migrate
usually from the west to th e east .

Evolutionary centres and the migration of ecological groups

The evolutionary centres a~_d the migratory characteristics of genera,
families and sections may h e explored by comprehensive morphological and
microevolutionary studies at the given taxonomic level. However, only a few
relevant works have been published, for instance, WEBSTER (1958) on Phyllan-

Acta Botanica Hungarica 31, 1985

30 A. BORHIDI

thus, J unn (1982) on Lyonia, DAHLGREN and GLASSMAN (1963) on Copernicia,

and HowARD (1949) on Coccoloba. Of course, besides the analysis of taxonomic
categories, the evolutionary centres and migratory pathways of life forms or
v egetation t y pes may also be revealed by thorough studies . For example, the
mangrove communities reach the highest diversity in the Indonesian archi-
pelago, which may be considered the evolutionary centre of the mangrove
vegetation types (ANDREANSZKY 1954). The localization of the richest and the
poorest stands of each vegetation typ e or plant community in Cuba may be
readily determined bas ed on phytosociological relev es and vegetation maps.
In the present study releves taken by BoRHIDI, MuNIZ, DEL Risco, CAPOTE
and SAMEK are used together with published (LEON 1936, 1939, MARIE·
VICTORIN 1942, 1944, 1956, WEBSTER 1956-58, BoRHIDI and MuNIZ 1977,
KLOTZ 1979, and BoRHIDI 1972, 1981) and unpublished (BoRHIDI) distribu-
tion maps.

Evolutionary centres and migratory 1·outes of the serpentine vegetation

The distribution of some typical elements of the serp entine vegetation is shown in Figs
3, 9, 11 , 22. The evolutionary centres and principal migratory routes are illustr a ted in Fig. 23.
The section Orbicularia of the genus Phyllanthus is a typical serpentinophilous group that in -
cludes many species occurring on the old ferritic soils of northeastern Oriente. Of thi s section
Ph. orbicularis is the only species that is found on all the serpentine outcrops of the island
(Fig. 8). Like Neobra cea valenzuelana (Fig. 9), thi s sp ecies must have reache d its suitable
habitats alon g the route indicated by arrows in the figure. LE6N (1946) and WEBSTER (1958)
assumed tha t this migration took place along a continuous serp entine "axis" ("eje serpen-
tino so" ). Such a serpentine strip could not have existe d later than the Oligocene, due to th e
long transgression of central Cuba . It is doubtful , h owever, w hether such a l arg e continuous
serpen tine area could have been denuded by that time, although the di stributions of v icarious
gen era (Fig. 10) and species (Figs 11- 12) of the Nip e Mts and Cajalbana sug gest that a direct
relationship may have exist ed. Yet, according to our recent geolo gical knowledge, it is more
likely that th e serpentine blo ck s were more or less separated all the time. The serpentinophilous
speci es were presumably distributed either by wind (Neobra cea) or by animals (Phyllanthus).
They ma y h ave had wider di stribution earlier , but due to the depleti on of h abitats (KRUCKE-
BERG 1954) they survived only on serpentines and became subjects of intens ive speciation.
It see ms th at there were several stage s of the east-west migration of the serpentine flor a.
Taxa w ith strongly disjunct geographical range, e.g. , Anemia, 1~1oacroton , L escaillea, and
Harna ckia, etc., were distributed during the earli er stages. Later, so me serpentinophilous
species, e.g. , Jacaranda cowellii, Coccoloba geniculata, Zanthoxylum nannophyllum (Fi g. 22),
a nd Platygyne parviflora (Fig. 14), migrated replacing the a bove m entioned taxa in the plains
of eastern Cuba. In these st ages the sp ecies with continuous distribution, such as Ne obracea
valenzuelana and Phyllanthus orbicularis, b ecame also widespread. On the h asis of all thes e,
three evolutionary centres of the serpentine flora may b e distinguished in Cuba (F ig. 23).
The f irst and the oldest is Moa from which the serpentine flora of the entire island, and the
flor a of other ferriti c soil areas in northeastern Orie nte, the Baracoa-Jauco zone a nd the
Crista! and Nipe Mts were originated. The second is Nipe, a migratory centre of a more xero-
p hilous serpentine flora which was capable of being established in the arid lowlands and of
migrating to the west. The third centre is Caj albana in western Cuba. A part of its flora retreated

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHI C SURVEY OF CUBA, I. 31

to the serpentine blocks risen later in central Cuba, especially to Habana and Matanzas.
Many latosol elements of this flora became a dapted to the acidic slatey habitats of Pinar del
Rio and I sla d e Pinos. In Fig. 23 the dotted lin e directed to the east indicates the secondary
migration of serpentine elements from areas which subsequently became secondary savannas
under th e influ ence of man.

Evolutionary centres and migratory routes of the flora of limestone cliffs

The flor a of limestone karsts ha s two primary evolutionary centres (Fi g. 24). One of
them is in th e conical karsts of Sierra de los Organos in western Cuba, which is consid ered to
be the oldest geological formation on the surface of that area. The majority of species found
here r emained endemics, although a portion of the flora spread to the younger T ertiary karst s
in the Hab an a a nd Matanzas Heights, e.g., Bombacopsis cubensis, Thrina.'t morrisii (Fig. 14)
and Phania matricarioides (Fig. 25). Certain element s managed to get dispersed as far as to
the hi gh-altitude mogotes and to the southern coast at the Escambray Mts. During the trans-
gression, Cuba itself was an old coastal zon e serving as a refuge of the vegetation of coastal
karsts. This v egetation started to spread over the r ecent coastal zone at the end of the Tertiary.
The other primary evolutionary centre is the mountains of Oriente. Here, the development of
the karstic flora is not as clear-cut as in western Cuba, since the karsts of Oriente are younger
than the serpentine formations and, also, a part of the limestone flora is originated from
lJrimary so ils d erived from serpentine ( e.g., Nipe Mts, Monte Libano). Of the many isolated

[]]] Serpentine areas + Zanthoxy lum nannophyllum

• Coccoloba geniculate Limit line of t he serpentine flora
.A Jacaranda Cowellii districts within the Central Cuba
sector
Fig. 22. Geographical distribution of Coccoloba geniculata Lindau, Jacaranda cowellii Britt. et
Wils. and Zanthoxylum nannophy llum (Urb.) Alain (after BoRHIDI 1973)

Fig. 23. Evolution centres and migratory routes of the serpentine flora in Cuba
(after BORHIDl 1973)

Acta Botanica Hurtgarica 31, 1985

32 A. BORHIDI

Fig. 24. Evolution centres and presumable miratory routes of the flora of limestone karsts
in Cuba (after BoRHIDI 1973)

• • ••

Phania

• Ph. matricarioides
.A Ph. cajalbanica

Fig. 25. Geographical distribution of the genus Phania (after BORHIDI 1973)

Central and S. Hispaniola

Fig. 26. Possible centres of evolution and migratory routes of the subdesertic, xero-mega-
thermophilous elements of Cuba (after BORHIDI 1973)

karstic regions, the group of Monte Libano and Monte Verde appears the most peculiar evolu-
tionary centre, so much the more because this area was crossed by many migratory paths of
varying nature and direction. From this old coastal karstic zone spread the karst flora over
the country, first along the coastline. In the westerly direction two secondary evolutionary
centres developed: the karsts southwest of the Nipe Mts, and at the northern border of Sierra
Maestra. In both areas the montane elements descending from the mountains were mixed
with the karstic flora spread along the coast. The influence of this flora was extended to the
east as far as the Yumuri karsts and Yunque de Baracoa, with a secondary evolutionary
centre developed in the latter place. The rise of the terraces in southern Baracoa made possible

Acta Botanica Hungarica 31, 1985

 PHYTOGEOGRAPHIC SURVEY OF CUBA, I. 33

lsidorea

~
Fig. 27. Geographical distribution of the genus Isidorea A. Rich. (after BORHIDI 1973 modified)

for the flora of Monte Libano to spread over the semi-desert coastal zone. As a result of the
migration, xero-megatherm species evolved. Some elements of the Oriente karstic flora migrated
along the southern coast at the Sierra Maestra. The others spread to the west from the Nipe
Mts via the form er limestone block mountains along the northwestern coast of central Cuba,
forming the flora of the Cubitas and Najasa Mts and, too, the flora of coastal shelves. The
influence of this migration may be seen even in the mogotes of Caguagas (cf. the distribution
of Hemithrinax, Fig. 13). This bidirectional distribution exhibited by karstic elements is an
important characteristic of the recent flora, too, so it was considered in the delimitation of the
flora region s of central Cuba.

Evolutionary centres and migratory routes of the xero-megatherm

elements
Some semi-desert xero-megatherro elements are originated from central and southern
Hispaniola (Fig. 26), to which the southeastern coast of Cuba was formerly connected. This
relationship is proved by the distribution of several genera and species, e.g., Omphalea com-
mutata, P etitia urbani (Fig. 21), Victorini, and Isidorea (Fig. 27), now restri cted to the
coastline. These elements, thanks to the gradual rising of the southern Baracoa coast, were
intermixed with the endemic flora of karstic slopes. As a result, several centres of speciation
developed, the richest of them being the Guantanamo Basin. The flora of Macambo-lmias
and Maisi, and the Daiquiri-Siboney-Santia go area in the west are almost as rich as that of
the Guantanamo Basin, so they can be considered as secondary and tertiary evolutionary
centres, r espe ctively. Th e number of xero-megatherm elements on the southern coast at Sierra
'"". Maestra, on the coasts of northwestern Oriente and northern Camagiiey, in the southern
., .• ..- foothills of Escambray Mts, on the northern rocky coast between Habana and Matanzas and
on the southern limestone shores of the Guanahacabibes peninsula descreases in that order.

Evolutionary centres and migratory routes of the montane rainforest

elements

The evolutionary centres and migratory routes of the flora in the montane rainforests
are shown in Fig. 28. The distribution of Hedyosmum serves as an example (Fig. 7). Again,
the main cen tre is the lower and old mountains in the Sagua-Baracoa Massif, espe cially the
Moa-Toa area. According to my hypothesis, this region received its montane elements via
Hisp aniola , and served as a new centre for their further migration to Crista!, Nipe, Purial,

3 Acta Botanica Hungarica 31, 1985

34 A. BORHID!

~~aniola
~
Jamaica -·--::::::
Fig. 28. Centres of evolution and possible migratory routes of the montana rain forest elements
in Cuba (after BORHIDI 1973) , .. ..
.~
lmias mountains, and, in particular, to Sierra Maestra. The first stage of thi s process took place
presumably in the mountain-range connecting Oriente and E scambray. The more permanent
--
changes, however, are results of migrations in the cooler p eriods during the Quaternary
between Sierra del Crista! and Gran Piedra through the Central Valley area whi ch subsided
la t er. At that time the lower border of montane rainforests may have run at a 2-300 m altitude.
This montane flora was mixed with Jamaican and southern Hi sp aniolan montane elements
a nd, during the cool stages in the Pleistocene, may hav e repeatedly retreated to the subalpine-
alpine zone in the Turquino group. In this place a secondary centre of sp eciation developed.
It is likely that even at that time some montane elements of Sierra Maestra got across to
the Escambray Mts, which received elements from the montane flora of Jamaica, too.

Distribution centres and migratory routes of the semi-deciduous elements

The origin and migration of the elements of semi-d eciduous forests and seasonal rain-
forests in the lowlands pose different problems. ALAIN (1958) traces the origin of some elements
back to Mexico, assuming a west-east migration of these forests. It is my conviction that this
flora survived in the. refuge of valleys at the time of transgression, and subsequently spread
over to every direction. Especially the mountains of Oriente, northern Camagiiey, Pinar d e
Rio and the Escambray Mts should be mentioned as potential r efuges. The high similarity
-'""'.,
between the characteristic composition of lowland and submontane for est formations also
refers to this fact.

For cited references see the second part.

Ac&a Botanica Hungarica 31, 1985