Phytocoenologia, 39 (2), 157–174
Berlin – Stuttgart, June 16, 2009
Old growth mature forest types and their floristic composition
along the altitudinal gradient on Silhouette Island (Seychelles) –
the telescoping effect on a continental mid-oceanic island
by Bruno SENTERRE, Bruxelles, Justin GERLACH, Cambridge, James MOUGAL and Denis MATATIKEN,
Seychelles
with 4 figures, 2 tables and 3 photos
Abstract. The granitic Seychelles are the only mid-oceanic islands of continental origin. Botanists have long focused
on taxonomy, and plant communities were described in a qualitative way, based on simple observation. Therefore the
altitudinal belts, their floristic characteristics and distribution are still poorly understood and conservation efforts
focus mainly on species-centred actions. Here we describe a quantitative study of plant communities and indicator
species differentiated along the altitudinal gradient on Silhouette, the most pristine and second highest (740 m) island
of this archipelago.
Twelve plots were sampled from 80 m to 640 m above sea level. Each plot contains three nested subplots corresponding to three forest strata: 50 x 10 m for all individuals of the tree layer (i.e. trunk diameter > 5 cm), 50 x 4 m
for the shrub layer (ligneous > 90 cm height), and 50 x 4 m for estimation of abundance-dominance coefficient in the
herbaceous stratum (all herbaceous plants, plus ligneous plants < 90 cm height).
The results are summarized in a two-way table. Both indicator species and vegetation types are discussed in
relation to previous studies in Seychelles. Three altitudinal belts are distinguished: lowland, submontane and lower
montane rain forests. Most of the best indicator species are found in the understory, especially ferns. The submontane belt develops at about 350 m and turns into a typical lower montane belt at ca. 550 m. Such transition zones occur respectively at about 900 and 1500 m in most of the tropical mountains, illustrating here a perfect example of the
“telescoping effect”. Although the flora of Seychelles is relatively species poor, the strong characterisation of these
altitudinal belts is unusual compared to younger islands in the Pacific, and may be a result of the longer evolution
of its flora.
eschweizerbartxxx author
Keywords: altitudinal zonation, cloud forest, IndVal, Massenerhebung effect, submontane, understory.
Introduction
Our capability to protect species is ultimately dependent on the conservation of representative and
functionally sustainable examples of the plant communities, i.e. vegetation types that support these
species. Conserving biological diversity at the level
of natural communities is an important complementary approach to single-species conservation efforts,
not only because they are home for such species but
also because natural communities contain important
assemblages of both plant and animal species, both
known and still undescribed. Why are some rare species limited to small areas? Which are the factors associated with such special habitats and where should
we expect to find new populations of these species?
Which other species have an important place in the
functionality of such ecosystems? Which species can
be used as indicators for recognizing and mapping
such plant communities?
On the Seychelles islands, botanists have long focused on taxonomy (Baker 1877, Christensen 1912,
Friedmann 1994, Robertson 1989), and plant communities have been described in a qualitative way,
DOI: 10.1127/0340 – 269X/2009/0039– 0157
based on simple observation (Carlström 1996, Gerlach 1993, Jeffrey 1962, Procter 1984b, VeseyFitzgerald 1940), or focussing on invasive species
(Fleischmann et al. 2003, 2005), or not considering
all strata (Gerlach 1997). As far as we know, quantitative studies on plant communities considering all
strata and all vascular plant groups (i.e. including the
poorly known ferns) have not been carried out in
Seychelles (see also Gerlach 1993, p.19). The main
forest types distinguished by most authors correspond to the three main classes already described by
Vesey-Fitzgerald: the “lowland forest” (0–300 m), the
“intermediate forest” (300–550 m) and the “mountain
moss forest” (from 550 m to the top, at 905 m). However, the location of these intervals and the definition
of vegetation types can vary widely between authors.
Such vegetation belts have been reported only for the
two higher islands of the Seychelles, i.e. Mahé and
Silhouette. During more than two centuries of human
presence, both islands have suffered extensive human
impact due to forestry, agriculture and cinnamon
plantations. The lowland and intermediate levels are
the most highly degraded, especially on Mahé where
nearly no pristine or even old growth humid forest
0340 – 269X/09/0039 – 0157 $ 08.10
© 2009 Gebrüder Borntraeger, D-14129 Berlin · D-70176 Stuttgart
158
B. Senterre et al.
still occurs below 550 m. In contrast, Silhouette is undoubtedly the most pristine island for humid forest
types. There, human disturbance is mainly restricted
to areas below 300 m or only in limited areas above
that altitude (see Vesey-Fitzgerald 1940, p.466,
for detailed explanations). In the lowland belt, many
areas are now covered by an old growth secondary
forest, characterized by many introduced species in
the upper layer but retaining many native species in
the understory. In the present study, we propose an
original description of the floristic variation along the
altitudinal gradient in the humid forests of Silhouette.
The forest types are described using both the most
abundant and the most characteristic species. Reference is made to other plant communities and other
names given in the literature.
Study area
The Seychelles archipelago comprises a group of
115 islands located in the western Indian Ocean. Of
these, 41 are granitic and constitute most of the socalled inner islands, lying between 4–5° South and
55–56° East. These are the only mid-oceanic granitic
islands in the world and are located about 1200 km
of the African coasts, 1000 km north of Madagascar
and 2800 km south-east of India. Their granitic nature denotes their continental origin as fragments of
Gondwanaland that were separated millions of years
ago when India drifted north toward Asia. Isolated
for about 75 million years, the Seychelles now hosts
a unique biota, many species being extremely primitive (Procter 1984a, WWF 2007). The majority of
those species are restricted to the larger granitic islands. The main granitic islands have a combined area
of about 230 km2, of which Mahé (154 km2), Praslin
(37 km2), Silhouette (20 km2) and La Digue (10 km2)
are the most important in terms of size.
According to Senterre (2009), the climate is an
intermediate between tropical wet (Af) and tropical
monsoon (Am), following the Köppen’s classification
(Kottek et al. 2006). The average annual rainfall is
2200 mm at sea level, but exceeding 4000 mm on the
summits. Mean temperature ranges from 24 °C to
30 °C (Walsh 1984). The prevailing winds are NW
monsoon (December to March) and SE trade winds
(May to October) and the inner islands lie outside the
cyclone belt.
eschweizerbartxxx author
Fig. 1. Location of the 12 plots inventoried for the present study of plant communities on Silhouette Island.
Forest types on Silhouette Island
159
considered all ligneous plant individuals with a dbh t
5 cm (i.e. diameter at breast height, measured at 1.3 m
above ground level, or following international standards, White & Edwards 2001) and included in a 50
x 10 m area (i.e. 500 m2). For each individual, dbh, total
height and structural position (i.e. emergent, canopy
or under-canopy tree; see Oldeman 1990) were recorded. Hemi-epiphytes like Ficus spp. or Schefflera
procumbens were also counted in this subplot if they
reached the canopy. The second subplot included all
ligneous individuals belonging to the shrub stratum,
i.e. small understory lianas (if any), palms, pandans,
tree ferns and mainly true shrubs, all higher than
90 cm and with dbh < 5 cm. This second subplot was
sampled along the same 50 m line but only 4 m wide
(i.e. 200 m2) and all individuals were counted. The
third subplot, sampled in the same area as subplot 2,
recorded herbaceous species and included all plants <
90 cm height thus also accounting for the regeneration
of species from the upper layers. Taller herbaceous
species (e.g. Mapania spp.) and understory epiphytes
were also included in this subplot. Individuals were
not counted but the abundance coefficient was estimated using van der Maarel’s (1979) categories (Table 1). For each plot, environmental and geographical
parameters were recorded: GPS coordinates, altitude,
slope inclination and orientation, percentage rock
cover on the ground, bryophyte cover on the ground,
trunks and branches, thickness of bryophyte cover,
and percentage cover of each of the strata (i.e. canopy
trees, under-canopy trees, shrubs and herbs) and for
palms specifically.
An important variety of habitat types exists on
these islands, mostly related to altitude (up to 905 m),
coastal influence, soil structure, topography and perturbation. The tropical wet climate and small size
of these islands do not allow the presence of either
mesophilous or dry forests or related formations,
except in special habitats such as “glacis”. These glacis are granite outcrops and have been considered as
inselbergs (Fleischmann et al. 1996). They are very
common on Silhouette and are found from sea level
up to the highest point. Silhouette, the third largest
island, has the second highest point of the archipelago (Mont Dauban at about 740 m) and is covered by
some of the most pristine forests of Seychelles. This
island is the least developed of the granitic islands and
has a population of fewer than 300 people. Its flora
is characterised by several island and archipelago endemics, some of which are already extinct from Mahé
(Friedmann 1994). Although the geology of Silhouette differs from that of the other large islands of Seychelles (Piggott 1968), the vegetation is similar to
that found on Mahé.
Materials and methods
The field sampling was carried out between the 11th
of February and the 23rd of March 2008. Considering the local logistic and climatic constraints, this
period included 23 days in the forest. Plot sampling
design followed the basic principles of quantitative
plant community studies. The plot disposition was
chosen to study the variation of the floristic composition along the altitudinal gradient, all other ecological variants being as constant as possible, especially
stand maturity and the proximity of ravine/valley
areas. Each plot was within an area which was considered ecologically and structurally homogeneous.
Twelve vegetation plots were studied on Silhouette
Island (Fig. 1). To include all strata, each plot was
subdivided into three subplots. The tree layer subplot
eschweizerbartxxx author
Results and discussion
Within the 12 studied plots (comprising 36 subplots),
we recorded 108 taxa in 95 genera and 56 families.
The tree layer contained 1068 individuals of 45 taxa.
Shrub subplots included 1756 individuals in 41 taxa,
Table 1. Abundance-dominance coefficients, based mainly on van der Maarel (1979) and Braun-Blanquet (1932). If a species is
observed only out of the plot but still in the same stand, its presence is quoted by 0.5. The categories are defined with correspondence
between semi-quantitative and quantitative scales, where “f” is respectively the estimated abundance-dominance coefficient or the
relative abundance of individuals.
Coefficient value
0.5
Definition
Average cover
out of the plot
0.5
Braun-Blanquet
van der Maarel
1
r
r
+
1
1 individual
2
2 individuals
2
+
3
f d 5%
3.5
1
1
4
5< f d 10 %
7.5
2
2m
5
10< f d 15 %
12.5
2
2a
6
15< f d 25 %
20
2
2b
7
25< f d 50 %
37.5
3
3
8
50< f d 75 %
62.5
4
4
9
f > 75 %
87.5
5
5
160
B. Senterre et al.
and the herb layer included 92 taxa, 29 of which were
ferns and fern allies.
To identify floristic groups in the dataset, we
performed a Detrended Correspondence Analysis (DCA, Legendre & Legendre 1998) based on
data synthesized from all strata (Fig. 2). The results
indicate that floristic variability is highly correlated
with the first ordination axis, which explains already
23.4 % of the total floristic variability. A Canonical
Correspondence Analysis (CCA, Legendre & Legendre 1998) carried out on the same dataset, with altitude as an environmental variable, showed that this
factor is highly correlated with the first ordination
axis, explaining 23.2 % of the floristic variability (p =
0.001). We conclude that altitudinal belts correspond
to distinct plant communities and that four groups
of plots could be considered for the indicator species
analysis. The relatively isolated position of plot 9 on
the second DCA axis, could indicate the relatively
low level of disturbance and invasion of alien species
in this plot within the lowland group.
What is the floristic relationship between the different floristic groups, and which species are peculiar
to some groups, i.e. which can be regarded as indicator species? To investigate the first question, we used
a classification method based on the Bray-Curtis similarity index and UPGMA (Unweighted Pair Group
Method with Arithmetic mean, Fig. 3). The results
were not totally concordant for the distinct strata, i.e.
using the herb layer submontane (“intermediate forests”) and lower montane (“mountain moss forest”)
belts were more clearly distinct than when the upper
tree layer was used. In fact, the difference between
these two belts was marked mainly by the presence of
Glionnetia sericea within an otherwise impoverished
tree flora, while the understory flora was much more
differentiated between these two altitudinal belts (see
Table 2). This observation underlines the importance
of considering the herb layer for forest type classification. Both results show that the submontane
forests have a higher floristic affinity with the lower
montane forests than with the lowland. Therefore,
we considered four groups in the analysis of indicator
species, using IndVal (Dufrêne & Legendre 1997)
with a 4-levels typology synthesised from the ordination and classification results. The justification of
each group at different levels of the hierarchy was estimated by the number and/or goodness of indicator
species and a two-way table is produced emphasizing
such indicator species (Table 2).
Results of the IndVal analysis, detailed in the Table
2, confirm the differentiation of three altitudinal belts
in the rain forests of Silhouette. The two submontane
variants share only few species, including typically
submontane species and excluding the most important lower montane ones. Nevertheless, the upper
variant of this submontane group (plots near Mon
Plaisir) included some lower montane species probably because of its somewhat transitional position at
500–540 m altitude. At Mon Plaisir, these lower montane species are mainly present as fallen epiphytes.
eschweizerbartxxx author
We therefore consider these two submontane variants as belonging to the same plant community. More
vegetation plots are required to clarify the variability
in the submontane belt. Hereafter, we describe forest types distinguished in the present study and make
reference to the literature on Seychelles and to worldwide synthesis of mountain vegetation habitats.
The nomenclature used here for the three forest
types distinguished is based on review of worldwide
studies of vegetation types and their nomenclature
(Areces-Mallea et al. 1999, Aubréville 1965,
Camirand & Evelyn 2003, Grossman et al. 1998,
Mueller-Dombois & Ellenberg 1974) and implemented using more detailed studies of altitudinal belts
in the tropics (Ashton 2003, Boughey 1955a, 1955b,
1965, Bruijnzeel & Hamilton 2000, Bussmann
2006, Churchill et al. 1995, Grimshaw 2001, Hedberg 1951, Letouzey 1968, Ohsawa 1993, Schnell
1976, 1977, Trochain 1980, Van Steenis 1935). In
addition, one of us recently revised the definitions,
homologies and nomenclature of forest types in the
tropical mountains worldwide (Senterre 2005b)
focussing on the lower belts. Altitudinal belts have
often been misinterpreted, especially with respect to
the submontane belt (Ashton 2003, White 1978),
for reasons related to levels of perturbation (Tchouto et al. 1999, Thomas & Achoundong 1994), heterogeneous transect data, and a lack of consideration
of ecological transgressors, i.e. species narrowly associated with more than one habitat, and ecological
equalizations (Senterre 2005b). The vegetation belts
and their definition, as understood here, are synthesized in our previous work (Senterre 2005b): lowland, submontane, lower montane (may include two
mostly physiognomic subdivisions), upper montane,
subalpine and alpine. From these, only the lowland,
submontane and lower montane belts are reported
from Silhouette and the Seychelles islands overall.
For each forest type, a conservation rank was estimated using categories proposed in the international vegetation classification system (Grossman et
al. 1998). We also accounted for information on the
quantitative data from our plots, field observations
and GIS (Geographic Information System) data.
Our personal observations on the transition from
one altitudinal belt to another in Seychelles are based
on extensive field explorations on Silhouette, mainly
from La Passe to Mon Plaisir, Mont Pot à Eau, Mont
Dauban, from Grand Barbe to Rende D’Avance,
Gratte Fesse, Mont Cocos Marrons, on the northern
slopes of Mont Dauban and on the eastern and northern slopes of Grand Congoman. We also explored the
Congo Rouge – Morne Seychellois – Pérard area and
Montagne Brûlée area, on Mahé. Throughout these
explorations, we focussed on the identified indicator
species. The transitions appeared to be always marked
at the same altitude and within an altitudinal interval
of less than 50 m.
Forest types on Silhouette Island
161
Fig. 2. Ordination of the floristic data (DCA, Detrended Correspondance Analysis) compiling all the strata (12 plots, 108 taxa), with
van der Maarel coefficient indexes. The first two axes of the ordination explain 30.7 % of the total variance of floristic data (sum of all
eigenvalues = 2.565). The first axis alone explains already 23.4 % of the total variance. Samples names are: L#, lowland; S#, submontane;
M#, lower montane.
eschweizerbartxxx author
Fig. 3. Classification using Bray-Curtis similarity index and UPGMA (Unweighted Pair Group Method with Arithmetic mean) (a) for
the tree layer and (b) for herb layer (using van der Maarel coefficient indexes).
162
B. Senterre et al.
Table 2. Two-way table produced with IndVal (Dufrêne & Legendre 1997) for all strata, i.e. 108 species in 12 plots (made of 36 subplots). Species are ordered following their IndVal value for the floristic groups identified. The most relevant species are emphasized
in bold. The values in the table are, for a given species, on the left, the sum of van der Maarel index values for all plots of a plant community and, on the right, the number of plots with presence of the species in the concerned plant community. The “Lowland” plots
are the number 1 (180 m a.s.l.), 6 (90 m), 8 (300 m), 9 (80 m); “Submont1” plots are the number 7 (400 m) and 10 (380 m); “Submont2”
plots are the number 2 (540 m), 3 (500 m), 4 (580 m); “Montane” plots are the number 5 (640 m), 11 (590 m) and 12 (590 m). The origin
of the species (“Orig.”) is abbreviated as follows: end. = endemic to Seychelles; ind. = indigenous; inv. = invasive; inv.p. = potentially
invasive; nat. = naturalized but non invasive. The stratum (“Str.”) to which the species belongs to is mentioned: A = canopy tree; Ad =
under-canopy tree; ar = shrub; H = herbaceous understory.
Species
Family
Orig. Str. IndVal
Lowland
Submont1
Submont2
Lower
Montane
1. Characteristic species of the Calophyllum inophyllum / Deckenia nobilis / Psilotum nudum lowland rainforests
Hypoxidia rhizophylla (Baker) F.Friedmann
Nephrosperma vanhoutteanum (H.Wendl. ex vanHoutt.) Balf.f.
Paraserianthes falcataria (L.) I.C.Nielsen
Cocos nucifera L.
Pyrostria bibracteata (Baker) Cavaco
Tabebuia pallida (Lindl.) Miers
Memecylon elaeagni Bl.
Allophyllus pervillei Blume
Dracaena reflexa Lam. var. angustifolia Baker
Adenanthera pavonina L.
Anacardium occidentale L.
Deckenia nobilis H.Wendl. ex Seem.
Flagellaria indica L.
Morinda citrifolia L.
Wielandia elegans Baill.
Calophyllum inophyllum L.
Oplismenus compositus (L.) P.Beauv.
Phymatosorus scolopendria (Burm.f) Pic.Serm.
Abrus precatorius L. subsp. africanus (Vatke) Verdc.
Albizia lebbeck (L.) Benth.
Artocarpus heterophyllus Lam.
Averrhoa bilimbi L.
Cananga odorata (Lam.) Hook.f. & Thomson
Casuarina equisetifolia L.
Citrus reticulata Blanco
Hevea brasiliensis (Kunth) Müll.Arg.
Melia dubia Cav.
Psilotum nudum (L.) P.Beauv.
Pterocarpus indicus Willd.
Sandoricum koetjape (Burm.f.) Merr.
Tacca leontopetaloides (L.) Kuntze
Tarenna sechellensis (Baker) Summerh.
Selaginella fissidentoides (Hook. & Grev.) Spring
Hypoxidaceae
Arecaceae
end.
end.
H
Ad
100
86
9./ 4
16./ 4
0./ 0
4./ 2
0./ 0
0./ 0
0./ 0
1./ 1
Mimosaceae
Arecaceae
Rubiaceae
Bignoniaceae
Melastomataceae
Sapindaceae
Dracaenaceae
Mimosaceae
Anacardiaceae
Arecaceae
Flagellariaceae
Rubiaceae
Euphorbiaceae
Clusiaceae
Poaceae
Polypodiaceae
Fabaceae
Mimosaceae
Moraceae
Oxalidaceae
Annonaceae
Casuarinaceae
Rutaceae
Euphorbiaceae
Meliaceae
Psilotaceae
Fabaceae
Meliaceae
Taccaceae
Rubiaceae
Selaginellaceae
inv.
ind.
ind.
inv.
end.
ind.
ind.
nat.
nat.
end.
ind.
nat.
ind.
ind.
ind.
ind.
ind.
nat.
nat.
nat.
nat.
nat.
nat.
inv.p.
nat.
ind.
nat.
inv.p.
nat.
end.
ind.
A
Ad
ar
A
ar
ar
ar
A
Ad
A
H
ar
ar
A
H
H
H
A
A
Ad
A
A
Ad
A
A
H
A
A
H
ar
H
76
75
75
75
72
70
69
59
50
50
50
50
50
47
39
35
25
25
25
25
25
25
25
25
25
25
25
25
25
25
19
8./
12./
12./
14./
12./
7./
11./
11./
4./
4./
3./
5./
2./
7./
7./
6./
2./
1./
3./
1./
1./
1./
2./
9./
1./
3./
1./
8./
3./
3./
3./
4./
0./
0./
0./
1./
1./
6./
6./
0./
0./
0./
0./
0./
1./
3./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
2./
0./
0./
0./
0./
0./
0./
0./
1./
2./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
2./
1./
0./
0./
0./
0./
0./
2./
0./
0./
0./
0./
0./
0./
0./
0./
3./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
eschweizerbartxxx author
4
3
3
3
3
3
4
3
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
0
0
0
1
1
2
2
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2. Characteristic species of the Northea hornei / Roscheria melanochaetes / Haplopteris ensiformis
cloud forests (submontane + lower montane)
Northea hornei (M.M.Hartog) Pierre
Dillenia ferruginea (Baill.) Gilg
Lindsaea kirkii Hook.
Mapania floribundum (Nees ex Steud) Koyama
Roscheria melanochaetes (H.Wendl.) H.Wendl. ex
Balf.f.
Elaphoglossum hornei C.Chr.
Protarum sechellarum Engl.
Trichomanes cupressoides Desv.
Elaphoglossum lancifolium (Desv.) C.V.Morton
Haplopteris ensiformis (Sw.) E.H.Crane
Canthium carinatum (Baker) Summerh.
Platylepis occulta (Thouars) Rchb.f.
Sapotaceae
Dilleniaceae
Dennstaedtiaceae
Cyperaceae
Arecaceae
end.
end.
end.
end.
end.
A
A
H
H
Ad
100
88
88
88
85
0./
0./
0./
0./
3./
0
0
0
0
1
7./
1./
4./
5./
9./
2
1
1
2
2
11./
9./
11./
12./
14./
3
3
3
3
3
15./
8./
10./
7./
12./
3
3
3
2
3
Lomariopsidaceae
Araceae
Hymenophyllaceae
Lomariopsidaceae
Vittariaceae
Rubiaceae
Orchidaceae
end.
end.
ind.
ind.
ind.
end.
ind.
H
H
H
H
H
ar
H
75
75
74
63
50
25
25
0./
0./
2./
0./
0./
0./
0./
0
0
1
0
0
0
0
0./
2./
3./
0./
2./
0./
0./
0
1
1
0
1
0
0
9./
6./
11./
7./
5./
1./
3./
3
2
3
3
2
1
1
9./
9./
8./
5./
3./
1./
2./
3
3
3
2
1
1
1
2.1 Characteristic species of the Drypetes riseleyi / Lomariopsis pervillei - Trichomanes fulgens submontane forests
Pandanus hornei Balf.f.
Lomariopsis pervillei Mett. ex Kuhn
Cynanchum callialatum Buch.-Ham. ex Wight
Drypetes riseleyi Airy Shaw
Ixora pudica Baker
Scleria sieberi Ness ex Kunth
Pandanaceae
Lomariopsidaceae
Asclepiadaceae
Euphorbiaceae
Rubiaceae
Cyperaceae
end.
end.
ind.
end.
end.
ind.
A
H
H
A
Ad
H
87
79
50
50
50
35
0./
0./
0./
0./
0./
0./
0
0
0
0
0
0
9./
5./
1./
1./
3./
3./
2
2
1
1
1
1
2./
2./
0./
0./
0./
0./
1
1
0
0
0
0
0./
0./
0./
0./
0./
2./
0
0
0
0
0
1
Forest types on Silhouette Island
163
Species
Family
Orig. Str. IndVal
Ludia mauritiana J.F.Gmel. var. sechellensis F.Friedmann
Verschaffeltia splendida H.Wendl.
Trichomanes fulgens C.Chr.
Syzygium jambos (L.) Alston
Bolbitis bipinnatifida (Kuhn) Ching
Cyathea sechellarum Mett.
Haplopteris scolopendrina (Bory) C.Presl
Phaius tetragonus (Thouars) Reichb.f.
Polystichopsis wardii (Bak. in Hook. & Baker) Tardieu
Procris insularis H.Schröter
Schefflera procumbens (Hemsl.) F.Friedmann
Sphaerostephanos subtruncatus Holttum
Merremia peltata (L.) Merr.
Psychotria pervillei Baker
Angiopteris evecta (G.Forst.) Hoffm.
Asplenium protensum Schrad.
Syzygium aromaticum (L.) Merr. & L.M.Perry
Trema orientalis (L.) Blume
Flacourtiaceae
Arecaceae
Hymenophyllaceae
Myrtaceae
Lomariopsidaceae
Cyatheaceae
Vittariaceae
Orchidaceae
Dryopteridaceae
Urticaceae
Araliaceae
Thelypteridaceae
Convolvulaceae
Rubiaceae
Marattiaceae
Aspleniaceae
Myrtaceae
Ulmaceae
end.
end.
end.
inv.
ind.
end.
ind.
ind.
end.
end.
end.
ind.
inv.
end.
ind.
ind.
nat.
ind.
Ad
A
H
Ad
H
ar
H
H
H
H
A
H
A
ar
H
H
Ad
Ad
33
30
80
40
67
67
67
67
67
67
67
60
53
36
33
33
33
33
Lowland
Submont1
Submont2
1./
4./
0./
0./
0./
0./
0./
0./
0./
0./
0./
1./
1./
1./
0./
0./
0./
0./
1./
4./
2./
1./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
9./
1./
7./
7./
6./
3./
8./
3./
3./
7./
3./
2./
4./
4./
2./
1./
1
1
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
1
2
2
2
2
2
2
2
2
2
2
1
1
1
1
Lower
Montane
0./ 0
1./ 1
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
0./ 0
1./ 1
0./ 0
0./ 0
0./ 0
0./ 0
2.2 Characteristic species of the Glionnetia sericea / Gastonia crassa / Elaphoglossum macropodium
lower montane forests
Ctenopteris albobrunnea (Baker) Tardieu
Gastonia crassa (Hemsl.) F.Friedmann
Glionnetia sericea (Baker) Tirv.
Nepenthes pervillei Bl.
Timonius sechellensis Summerh.
Pandanus sechellarum Balf.f.
Hymenophyllum polyanthos (Sw.) Sw.
Hymenophyllum hygrometricum (Poir.) Desv.
Agrostophyllum occidentale Schltr.
Colea seychellarum Seem.
Elaphoglossum macropodium (Fée) T.Moore
Gynura sechellensis (Baker) Hemsl.
Syzygium wrightii (Baker) A.J.Scott
Xiphopteris serrulata (Sw.) Kaulf.
Paragenipa wrightii (Baker) F.Friedmann
Erythroxylum sechellarum O.E.Schulz
Rumohra adiantiformis (G.Forst.) Ching
Gleichenia linearis (Burm.f.) C.B.Clarke
Humata repens (L.f.) J.Small ex Diels
Aphloia theiformis (Vahl) Benn. subsp. madagascariensis
(Clos) H.Perr. var. seychellensis (Clos) F.Friedmann
Curculigo sechellensis Boj.
Asplenium caudatum G.Forst.
Bulbophyllum intertextum Lindl.
Craterispermum microdon Baker
Hederorkis seychellensis Bosser
Mapania seychellaria D.A.Simpson
Microsorum punctatum (L.) Copel.
Polystachya sp.1
Psilotum complanatum Sw.
Psychotria sp.1 Senterre
Psychotria sp.2 Senterre cf. dupontiae
Seychellaria thomassetii Hemsl.
Grammitidaceae
Araliaceae
Rubiaceae
Nepenthaceae
Rubiaceae
Pandanaceae
Hymenophyllaceae
Hymenophyllaceae
Orchidaceae
Bignoniaceae
Lomariopsidaceae
Asteraceae
Myrtaceae
Grammitidaceae
Rubiaceae
Erythroxylaceae
Davalliaceae
Gleicheniaceae
Davalliaceae
Flacourtiaceae
end.
end.
end.
end.
end.
end.
ind.
ind.
end.
end.
ind.
end.
end.
ind.
end.
end.
ind.
inv.
ind.
end.
H
Ad
A
H
Ad
A
H
H
H
Ad
H
H
Ad
H
ar
ar
H
H
H
Ad
100
100
100
100
100
90
88
87
67
67
67
67
67
67
66
60
53
51
51
48
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
4./
0./
0./
0./
0./
0./
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0./
0./
0./
0./
0./
2./
0./
0./
0./
0./
0./
0./
0./
0./
2./
1./
0./
0./
0./
0./
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0./
0./
0./
0./
0./
0./
3./
3./
0./
0./
0./
0./
0./
0./
0./
0./
3./
2./
3./
4./
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
1
1
1
1
9./
7./
12./
10./
9./
11./
13./
12./
5./
2./
7./
3./
5./
5./
8./
5./
7./
4./
6./
6./
3
3
3
3
3
3
3
3
2
2
2
2
2
2
3
2
2
2
2
2
Hypoxidaceae
Aspleniaceae
Orchidaceae
Rubiaceae
Orchidaceae
Cyperaceae
Polypodiaceae
Orchidaceae
Psilotaceae
Rubiaceae
Rubiaceae
Triuridaceae
end.
ind.
ind.
end.
end.
end.
ind.
H
H
H
A
H
H
H
ind.
end.
end.
ind.
H
ar
ar
H
45
33
33
33
33
33
33
33
33
33
33
33
2./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
1
0
0
0
0
0
0
0
0
0
0
0
3./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
2
0
0
0
0
0
0
0
0
0
0
0
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0./
0
0
0
0
0
0
0
0
0
0
0
0
7./
3./
2./
1./
4./
2./
2./
2./
1./
4./
2./
3./
2
1
1
1
1
1
1
1
1
1
1
1
Cinnamomum verum J.Presl.
Clidemia hirta (L.) D.Don
Phoenicophorium borsigianum (K.Koch.) Stuntz
Nephrolepis biserrata (Sw.) Schott
Planchonella obovata (R.Br.) Pierre
Psidium cattleianum Sabine
Grisollea thomassetii Hemsl.
Lauraceae
Melastomataceae
Arecaceae
Oleandraceae
Sapotaceae
Myrtaceae
Icacinaceae
100
92
92
75
67
67
25
21./
14./
24./
5./
6./
15./
1./
4
4
4
2
3
4
1
17./
4./
13./
4./
5./
0./
1./
2
1
2
1
2
0
1
25./
21./
13./
11./
6./
11./
0./
3
3
3
3
2
2
0
24./
17./
7./
8./
1./
7./
2./
3
3
2
3
1
2
1
eschweizerbartxxx author
3. Ubiquist species
inv.
inv.
end.
ind.
ind.
inv.
end.
A
ar
Ad
H
A
Ad
Ad
164
B. Senterre et al.
Tropical ombrophilous lowland forest
(Lowland rain forest)
Synonyms: Lowland forest (Fleischmann et al. 2003,
Gerlach 1993, Vesey-Fitzgerald 1940); Natural
intermediate forest, pro parte (Gerlach 1993, within
the Mid-altitude forest); forêt de basse altitude, facies
à Mimusops sechellarum et Vateriopsis seychellarum
(Friedmann 1994).
Physiognomy: The lowland forests are characterised
by a high canopy, at about 25–30 m, and commonly
emergent trees up to 30–50 m, giving the canopy an
undulated aspect. The main four strata are typically
well differentiated: (1) the upper tree layer (including
sub-canopy trees, canopy trees and emergent trees);
(2) the under canopy trees; (3) the shrub layer; and (4)
the herb layer (Photo 1). Epiphytes and lithophytes
are rare or not developed, or localized in wet microhabitats (e.g. in the shade of big granite boulders).
Floristics: The lowland forests are well differentiated not only by their remarkable abundance of nonindigenous species (Adenanthera pavonina, Hevea
brasiliensis, Paraserianthes falcataria, Sandoricum
koetjape, Tabebuia pallida) and more or less scattered
coconut trees (Cocos nucifera), but also by some na-
tive indicator species such as Calophyllum inophyllum, Casuarina equisetifolia, Deckenia nobilis (in the
upper tree layer), Nephrosperma vanhoutteanum (in
the under-canopy tree layer), Allophyllus pervillei,
Memecylon elaeagni, Pyrostria bibracteata, Wielandia
elegans (in the shrub layer), Hypoxidia rhizophylla,
Psilotum nudum, Tacca leontopetaloides (in the herb
layer). The main diagnostic species are Calophyllum
inophyllum, Deckenia nobilis, Wielandia elegans, Tacca leontopetaloides, and most of the non-indigenous
timber trees. In addition, lowland rain forests are easily recognized from higher belts by the absence of the
most characteristic species of both submontane and
lower montane belts, e.g. Northea hornei, Roscheria
melanochaetes, and Haplopteris spp. Nevertheless,
if some of these (sub)montane species are present,
they are marginal and most probably associated with
wetter microhabitat (ravine/valley, rock shelter, etc.)
revealing ecological equalizations (Budowski 1965,
Rübel 1935, Senterre 2005b), or are relict populations from past climate changes.
Ecology: On all the areas explored, both windward or
leeward the lowland belt extended from about 20 m
up to about 300–350 m altitude. It is well developed
on non-superficial soils, with 0–50(-75) % cover of
granite blocks.
eschweizerbartxxx author
Photo 1. Lowland rain forest at plot 1, with many Tabebuia pallida trees, some Cocos nucifera, and a quite diverse understory with still
a lot of native species.
Forest types on Silhouette Island
165
Distribution: It has been reported, in the Seychelles
inner islands, mainly on Mahé, Praslin, Silhouette,
La Digue, Marianne, Curieuse and Félicité. On Silhouette, a more or less pristine patch of this forest
type was observed on the hills between Grand Barbe and Mont Cocos Marrons (plot 9, see DCA results). Some other well conserved areas are expected
on the northern slopes of Mare aux Cochons, the
south-western slope of main Mont Dauban (probably never explored), and near Pointe Civine (Fig. 4).
At all other sites the lowland forest was semi-natural,
mainly dominated by non native timber tree species.
On Silhouette most areas are old growth secondary
forests, i.e. closed canopy with late secondary species
(Oldeman 1990), still hosting many native species.
The lowland rain forests are mixed with other vegetation types like the open glacis formations, the ravine/
valley forests variants, and secondary eco-units (secondary woodlands and pioneer scrub habitats).
Variability and related vegetation types: Occasionally
one or two non-indigenous species can become dominant providing a floristically and physiognomically
distinct appearance. In such places, the conservation
value, vulnerability and management of lowland forest may differ with more pristine parts of the forest.
Therefore, some authors have decided to consider
highly invaded forests as distinct forest types. Gerlach (1993) described three main types within his
“mid-altitude forests”: 1-”tree plantations”, characterised by non native timber species (on Silhouette,
mainly Sandoricum koetjape and Tabebuia pallida) or
rubber trees (Hevea brasiliensis); 2-Psidium littorale
forest; and 3-Cinnamomum verum forest (= “Cinnamon coppice”, Vesey-Fitzgerald 1940). Off course,
all intermediates could exist from a non-invaded forest to a highly invaded one, but such distinctions are
useful in a conservation perspective, for the reason already mentioned. On Silhouette, ancient tree plantations and Psidium littorale forests are more common
at low altitude while Cinnamomum verum forests are
mainly found in the submontane belt. Nevertheless,
apart from tree plantations, these forest types can be
observed from sea level to the higher summits.
Where perturbations have occurred in the past (tree
fall gaps, forest fire, logging), early to late secondary
forests are found in relatively localized stands, sharing some species from the old growth mature lowland
rain forests, mainly in the understory, mixed with
successional species. These successional eco-units
are still poorly understood in Seychelles and have
caused much confusion in previous classifications.
Nevertheless, it seems that two main variants could
be distinguished: the “secondary palmetum” and the
“Agati woodland”.
The “secondary palmetum” of Vesey-Fitzgerald
(1940) is mainly characterised by the abundance and
size of three palm species (Nephrosperma vanhoutteanum, Phoenicophorium borsigianum and especially Deckenia nobilis) occasionally associated with
Pandanus hornei. Because of the common organisa-
eschweizerbartxxx author
tion of these palm species into a single layer or a few
strata, such secondary eco-units appear to be associated with fire events or rare tropical storms. This forest type is quite localized on Mahé and Silhouette but
can cover more extensive areas on Praslin. It can also
be observed in the submontane belt, i.e. on Montagne
Brûlée on Mahé. Several species of this community
are shared with quite distinct plant communities, associated with distinct environmental conditions and
characterised by a distinct physiognomy, i.e. the “dry
land type of Palmetum”, or Deckenia-Memecylon society (Vesey-Fitzgerald 1940). The dry land type
of Palmetum occurs on unequal sized blocks divided
from each other by “Gleichenia-brake” or bare slabs
of rock, where rapid drainage through a coarse, gritty
soil caused conditions of physiological dryness. This
vegetation type includes some of the glacis vegetation units defined by Procter (1984b) and Gerlach
(1993). It corresponds mainly to what Friedmann
(1994) called “forêt de basse altitude, facies à Dillenia
ferruginea, Deckenia, Phoenicophorium”. The ecological transgression (Senterre 2005b, White 1978,
1979) of species from such dry habitats to secondary
eco-units of humid forest, especially fire-disturbed,
is a phenomenon often observed in other tropical regions (Budowski 1965, Senterre et al. 2004). Note
that the term “palm forest” have been used for different plant communities in Seychelles, often without
making any distinctions between the range of variants, i.e. Verschaffeltia ravine/valley forests, Deckenia secondary Palmetum, Deckenia-Memecylon dry
land Palmetum, etc.
The other secondary variant distinguished is
dominated by fast-growing naturalised trees, i.e. Adenanthera pavonina and Paraserianthes falcataria
(the “Agati woodland” of Vesey-Fitzgerald). Here
the term woodland gives a very good picture of this
forest type, characterized by an irregular canopy
layer, not entirely closed, with high emergent trees
and an understory with much light penetration. This
forest type covers extensive areas in the lowland but
also in the submontane belts. It seems related to local
perturbations such as tree fall gaps (more frequent on
steep slopes) or logging.
Along the ravines and valleys that run through the
humid forests, with or without water on the ground
surface, a distinct lowland forest type develops. This
ravine/valley forest is mainly characterised by the
palm Verschaffletia splendida and the screw palm
Pandanus hornei, both of which have stilt roots as a
physiognomic adaptation to such habitats (Richards
1952, Schnell 1976, Senterre 2005b, Senterre &
Lejoly 2001). This community is also developed in
the submontane and lower montane ravine/valley
rain forests, though associated with a different suit of
understory species (see next section). In addition, the
lowland ravine/valley rain forests are often characterised by Barringtonia racemosa (L.) Spreng. (Friedmann 1994: 189).
At the coast between 0 and 20 m altitude, the lowland rain forest turns into another plant community:
166
B. Senterre et al.
the coastal rain forest. This forest type, shaped by the
presence of the sea can be considered as a distinct entity, which is characterised by many species of the inland lowland rain forest (as in other tropical islands:
Keppel 2002) coupled with the natural presence of
Cocos nucifera, with more abundant Casuarina equisetifolia, and some more characteristic species like
Barringtonia asiatica (L.) Kurz and Terminalia catappa L. (Friedmann 1994: 187, 337). Pisonia grandis
R.Br. is another coastal forest species but may be more
associated with the presence of bird populations (see
Friedmann 1994: 110) and could therefore be characteristic of what Gerlach (1993) called the “seabird
island coastal forest”. The coastal forests were called
the “coastal plateau forest” by Procter (1984b), the
“typical coastal forest” by Gerlach (1993), or the “littoral forest” (Fleischmann et al. 2003). These coastal forests should not be confused with the coastal
swamp forests characterised mainly by Heritiera littoralis Ait. (see “lowland marsh” of Gerlach 1993).
Such coastal swamp forests are localised in the coastal
swamps, at the back of the mangrove, and sometimes
penetrate inland along the lower course of rivers.
Mangrove forest types constitute a totally distinct
floristic unit, including several distinct forms (Chapman 1976, Macnae 1971, Pascal et al. 2001).
Conservation rank: G1-GH (critically imperilled or
historically eliminated: Grossman et al. 1998). Within the four lowland forest plots, we recorded 49 taxa
of which 29 are indigenous (59 %). About 45 % of
these indigenous taxa are endemic to Seychelles. Although the diversity of indigenous taxa is higher than
that of exotic species, native trees comprised only
25 % of the total tree abundance. Undisturbed lowland rain forest on Silhouette is considered extremely
rare. Restoration efforts are deemed vital to maintain
this forest type, although detailed information on the
original forest composition is absent. We estimate the
original area of lowland rain forests to about 1309 ha
(66 % of Silhouette island, Fig. 4), nowadays reduced
to 5–20 % of more or less undisturbed fragmented
stands.
Tropical ombrophilous submontane forest
(Submontane rain forests)
Synonyms: Natural intermediate forest, pro parte
(Gerlach 1993, within the Mid-altitude forest);
High-altitude forest, pro parte but mainly (Gerlach
1993); forêt de basse altitude (Friedmann 1994); Moist
eschweizerbartxxx author
Photo 2. Submontane rain forest at plot 10 with a lot of Northea hornei in the tree layer, next to many cinnamon trees. We can also see
the typical undercanopy tree Roscheria melanochaetes.
Forest types on Silhouette Island
167
forest slopes and valleys habitats (Procter 1984b: type
1d); Intermediate forest (Vesey-Fitzgerald 1940:
the typical form); slopes and valleys of the intermediate forests, pro parte (Fleischmann et al. 2003). In
other tropical regions, homologous vegetation types
have received multiple, sometimes very confusing
names: lower montane forest (Bruijnzeel & Hamilton 2000, Whitmore 1998), premontane rain forest
or wet forest (many authors in central America, e.g.
Holdridge 1967), transition or intermediate forests
(in most tropical studies), cloud forest (mostly by authors not differentiating submontane and lower montane belts). However, the identity of submontane forest as a distinct altitudinal belt rather than a transition
zone is clear, at least in most tropical areas (Senterre
2005b) and in Seychelles as demonstrated here.
Physiognomy: The physiognomy (Photo 2) is similar
to that of lowland rain forest, except that emergent
trees are less common. The four strata are still well
differentiated and the understory is marked by the
presence of relatively abundant vascular epiphytes,
on the trunk base and extending to rock surfaces, i.e.
lithophytes. The tree ferns may appear very locally,
especially near ravines, but are not abundant and their
size is quite small. Bryophyte patches are localized on
some tree trunks and never form a continuous cover.
They may be more abundant in ravine/valley variants
of the submontane forest, but then their thickness is
still less than 1 cm.
Floristics: The upper tree stratum is dominated by
Northea hornei and Dillenia ferruginea, more or less
mixed with cinnamon trees. The under-canopy layer
is mainly dominated by Roscheria melanochaetes,
Phoenicophorium borsigianum and cinnamon trees
and, locally, by Pandanus hornei and Verschaffeltia
splendida. The main species of the shrub stratum includes cinnamon with some scattered Psychotria pervillei and locally abundant Ixora pudica. Non shrub
species of the understory include most of the characteristic submontane species: Lomariopsis pervillei,
Trichomanes fulgens, Haplopteris scolopendrina and
Grammitis pervillei (Mett. ex Kuhn) Tardieu (the last
one only observed out of the plots). Therefore, the
submontane rain forests can be distinguished from
the lowland rain forests by the abundance of Northea
hornei, Roscheria melanochaetes and Haplopteris spp.
They can be distinguished from the lower montane
belt, on the one hand, by some preferential submontane species such as those cited here for the understory and, on the other hand, by the absence of the
characteristic lower montane species. Within the tree
strata, we consider only two species as characteristic
submontane species, i.e. Craterispermum microdon
and Drypetes riseleyi, even though our data did not
emphasize that observation, these species being quite
rare and also marginally present in lower montane
rain forests. On Silhouette, they are more abundant
on the southern part of Jardin Marron and Grand
Congoman, which are probably the best remnants of
submontane rain forests in Seychelles.
eschweizerbartxxx author
Ecology: The lower limit of the submontane rain forest has mostly been removed, especially on Mahé.
We had the opportunity to observe it very clearly
on the following sites on Silhouette: on the eastern
slopes of Grand Congoman, on the slopes of Rende
D’Avance near Gratte Fesse, on the northern slope of
Mont Cocos Marrons, and on the northern slope of
Mont Dauban. In all sites, the first species appearing,
at about 300 m, are Northea hornei, Roscheria melanochaetes, Haplopteris ensiformis, Lindsaea kirkii
and Mapania floribundum, thus species common to
submontane and lower montane forests. The submontane belt is generally fully developed at 350 m
and is very homogeneous up to about 550 m, where
the transition to the lower montane belt occurs.
Distribution: This vegetation type once occupied extensive areas on Mahé and Silhouette. On Silhouette,
the best representatives are found, from the most
pristine to the least pristine, at: Grand Congoman,
the southern part of Jardin Marron, the eastern slopes
(unexplored) and northern slopes of Mont Dauban
and the northern slope of Mont Cocos Marrons. On
Mahé, we do not know of any locality for a well conserved submontane rain forest. The submontane rain
forest has not been recorded for Praslin (summit at
367 m) or La Digue (333 m). Nevertheless, some submontane indicators are cited for the flora of both islands, especially Praslin, and, therefore, further studies are needed in order to examine the exact position
and ecology of such submontane indicators, focussing on fern species (not considered by previous authors).
Variability and related vegetation types: The variants invaded by alien species, and distinguished for
the lowland belt, can occur also in the submontane
belt, having very similar floristic and physiognomic
patterns. Nevertheless, understory species, including
ferns and herbs, will differ from lowland to submontane invaded variants. In addition, soils and ecology
will differ. Therefore, even though lowland and submontane invaded forests presents very similar characteristics, they should be differentiated based on
their understory flora and/or ecological parameters
(mainly altitude) and considered as distinct units for
conservation/restoration actions. In addition, another type of invaded forest has been observed in
the submontane belt only and is based on Syzygium
jambos. This species already invaded quite large areas on Mahé and is now also becoming a problem on
Silhouette (especially in the eastern part of the Mont
Dauban ridge).
The successional stages in the submontane belt
present mainly common characteristics with homologous eco-units in the lowland belt. We found, for
example at Montagne Brûlée (Mahé), a forest type
related to the “secondary palmetum” cited for the
lowland variants, but differing clearly by the presence of Roscheria melanochaetes and most of the understory (sub)montane indicators. On Silhouette, the
submontane forests have been degraded mainly on
168
B. Senterre et al.
the slopes of Jardin Marron, where we now can see
a submontane form of the “Agati woodland”. This
forest type is also well developed on the northern and
southern faces of the Mont Dauban ridge, where the
steep slopes promote tree fall gaps and natural succession to take place.
Even more common than in the lowland belt, submontane ravine/valley rain forests alternate with the
typical submontane rain forests. This submontane
ravine/valley variant differs from the lowland one
by the absence of Barringtonia racemosa, disappearing at about 300 m, and by the simultaneous appearance of abundant populations of Begonia seychellensis Hemsl., Cyathea sechellarum, Sphaerostephanos
subtruncatus, Angiopteris evecta. The tree layer of
these ravine/valley forests is often irregularly closed.
Pandanus hornei and Verschaffeltia splendida are especially characteristic, accompanied by other casual
(sub)montane species like Northea hornei. This forest
type has been highly degraded and has a very high
conservation value. Indeed, we collected several unknown species in the submontane ravine/valley rain
forests of the northern face of Mont Dauban. Some
fauna elements also seem limited to this forest type,
i.e. between Gratte Fesse and Mont Corgat (Silhouette, Gerlach 2003: 27). This forest type includes
the “Pandanus hornei-Verschaffeltia splendida society” of Vesey-Fitzgerald, the “Cyathea seychellarum
patches” of Gerlach (1993), and the “moist forest
ravines habitat” of Procter (1984b: type 1c).
In only a few places, the topography allows the
development of submontane swamp forests, see
“high altitude marsh habitat” of Gerlach (1993).
This habitat is still poorly understood. We only have
few data on its flora, based on observations from the
Mahé’s Mare aux Cochons (Carlström 1996). On
Silhouette, such a forest type probably once occurred
on its homonymous locality, i.e. Mare aux Cochons,
but this area has been totally remodelled before any
detailed botanical observation could have been done.
Based on the Carlström qualitative observations, this
habitat is not floristically different from the submontane ravine/valley rain forests. The forest ecotone of
this swamp habitat is not detailed by the previous authors who were more interested by the general flora.
eschweizerbartxxx author
Conservation rank: G1(-G2) (critically imperilled).
Within the 4 most typical submontane forest plots
(excluding plot 4), we recorded 59 taxa including 50
indigenous ones (85 %). About 56 % of these indigenous taxa are endemic to Seychelles. Here the indigenous taxa are still clearly more important than
exotics in terms of species, and about one tree out of
two (47 %) belongs to a native species. This vegetation type is nearly extinct on Mahé, and imperilled
on Silhouette. Its status on Praslin, if it exists, is totally unknown. We estimate the original area covered
on Silhouette up to about 365 ha (18 % of the island,
Fig. 4) of which 60 % remain undisturbed. Restoration is still possible because the original undisturbed
composition still exists on Silhouette, and the pres-
ent paper gives a new contribution to our knowledge
of it. But here the expansion of invasive species is
of great concern and additional plots are needed to
understand the variability of this plant community
more precisely. Permanent plots would also be useful
to follow up the invasion of non-native species into
the very well conserved areas mentioned here.
Tropical ombrophilous lower montane forest (Lower montane rain forest)
Synonyms: Mossy montane forest, not High-altitude
forest (Gerlach 1993); forêt hygrophile d’altitude,
faciès à Northea hornei et Glionnetia sericea (Friedmann 1994); Moist forest boulder peat habitat
(Procter 1984b: type 1a); Mountain moss forest,
Randia-Nepenthes society (Vesey-Fitzgerald 1940);
Mountain mist forest (Fleischmann et al. 2003: type
5c). In other tropical regions, homologous vegetation
types have received a diversity of names: cloud forest,
lower montane cloud forest (Bruijnzeel & Hamilton 2000), moist montane forest (Trochain 1957),
dwarf cloud forest, mist montane forest (Boughey
1955a).
Physiognomy: At first sight, the most striking difference with the lowland and submontane belts is the
spectacular physiognomy of these lower montane
forests (a characteristic shared worldwide). Here the
canopy becomes much lower, i.e. about 8–16 m (depending on the soil structure), and emergent trees are
nearly absent, giving to the canopy a very flat aspect.
The reduction of the canopy height is clearly associated with the reduction of the trunks, nearly undifferentiated in the extreme examples, where the soil is
less developed. The lower canopy height causes a lack
of differentiation of the under-canopy tree layer. As
with the submontane rain forests, vascular epiphytes
are abundant (Photo 3), but in this case very abundant and mixed with an impressive covering of bryophytes on the “trunks”, on the main branches and on
the ground. The bryophytes cover can easily reach
more than 5 cm and up to 15 cm of thickness.
Floristics: Here the invasive species become clearly
less dominant and the main tree is now Northea hornei. It is associated in the upper tree layer with Glionnetia sericea, Timonius sechellensis (a common species
often abundant in such habitat). Pandanus sechellarum and Roscheria melanochaetes are sub-canopy
trees. Dillenia ferruginea, Cinnamomum verum and
Psidium littorale may replace most of the Northea
and Glionnetia trees in more degraded areas. The
shrub layer is mainly composed of Gastonia crassa,
Rapanea seychellarum Mez, Syzygium wrightii,
Erythroxylum sechellarum, Paragenipa wrightii, Psychotria pervillei and its related taxa or forms. Most of
the herbaceous species found in the understory are
typically lower montane, i.e. not seen in the submontane belt: Ctenopteris albobrunnea, Nepenthes pervillei, Hymenophyllum polyanthos, H. hygrometricum,
Forest types on Silhouette Island
169
Agrostophyllum occidentale, Elaphoglossum macropodium, Gynura sechellensis, Xiphopteris serrulata, Rumohra adiantiformis, Humata repens, Bulbophyllum
intertextum, etc. (Table 2). Among the species already
present in the submontane belt, the most important
ones include Lindsaea kirkii, Mapania floribundum,
Protarum sechellarum and Haplopteris ensiformis.
For the shrub layer, the best indicator species are
Gastonia crassa and Rapanea seychellarum. The main
diagnostic tree species is Glionnetia sericea. Within
the herbaceous characteristic species cited here, the
first ones appearing at the transition level and best
diagnostic species are the ferns Elaphoglossum macropodium, Humata repens, Ctenopteris albobrunnea
and Xiphopteris serrulata. Another understory species not mentioned in our plots but an excellent indicator of the lower montane belt is Oleandra distenta
Kunze.
Ecology: In all visited sites, on Mahé and Silhouette,
the transition from submontane to lower montane
forests can still be observed quite easily, especially
if we focus on the understory indicator species. We
found generally the first lower montane indicators
at about 550 m, and then the typical lower montane
belt become well developed before the 600 m (ca.
580 m), except where secondary eco-units or ravine/
valley variants cover too large an area, i.e. the Casse
Dent entry to Congo Rouge. On steep or precipitous mountain ridges, several lower montane species
can often be observed much below the 550 m transition: in some limited areas of Montagne Brulée, and
on the north-eastern side of Mont Dauban, at as low
altitude as 450 m. This example of ecological equalization is very well explained by Vesey-Fitzgerald
(1940): when “the prevailing wind strikes these spots
on the ridge and the strong up-draught created by the
steepness of the terrain causes a great expansion and
cooling of the air with a consequent excessive development of mist” at a lower altitude than in areas less
exposed to winds or without precipitous steep slopes.
This phenomenon has probably not been considered
by most of the botanists in Seychelles while observing montane species at unusual lower altitude.
Distribution: The lower montane rain forest is mainly found on the higher ridges of Mahé and Silhouette,
above 550 m. On Mahé, it could be present at Mont
Varigault (550 m), Montagne Planneau (680 m), Copolia (580 m), Trois Frères (690–760 m), Congo Rouge
(740–830 m), Morne Seychellois (905 m), Morne
Blanc (655 m) and Pérard (845 m). On Silhouette, this
forest type occurs only on Mont Pot à Eau (620 m)
and along the Mont Dauban ridge (740 m), where it
eschweizerbartxxx author
Photo. 3. Lower montane rain forest at plot 5 with a very important cover of epiphytes, including many fern species. Most trees are
Northea hornei, mixed with Glionnetia sericea, Timonius sechellensis and some cinnamon trees.
170
B. Senterre et al.
is very well preserved. On Mahé, the best preserved
examples of this forest type are found on the slopes of
Congo Rouge, Morne Seychellois and Pérard.
Variability and related vegetation types: Here, the invaded variants distinguished for the lowland and submontane belts could in theory still be distinguished,
for the same reasons. Nevertheless, monodominant
forest of non-indigenous species are not so developed
above the transition of 550 m, although cinnamon and
Syzygium jambos can locally constitute monodominant forest eco-units on Mahé. On Madagascar, the
lower montane forest belt can be totally replaced by
Psidium forests (Carrière et al. 2008) but such a situation has not yet been attained in Seychelles.
As already mentioned, secondary lower montane
forests may be those areas with more Dillenia ferruginea trees, another native species replacing the
Northea hornei and Glionnetia sericea individuals.
Therefore, the floristic composition of the tree layer
is no longer very distinct from the submontane forests and this emphasizes again the importance of the
understory flora for the precise recognition of these
two mountain vegetation belts. In addition, on the
higher slopes of both northern and eastern sides of
the main summit of Mont Dauban, the forest above
550 m altitude is on somewhat deeper soils and takes
the appearance of the Agati woodlands, but detailed
data or observations have not been collected. Getting
to this site with time for observation is especially difficult because of the incredible density of Clidemia
hirta, both in the understory of these Agati woodlands (northern slope) and in the open patches of the
eastern slope (also invaded by Gleichenia linearis and
Rubus rosifolius J.E. Smith). It is clear that this area
of Mont Dauban needs urgent detailed studies. This
is also true for the south-western side, which is probably the most pristine and least prospected (probably
never explored), and which is the highest well conserved lower montane forest of Silhouette.
Apart from these invaded and secondary variants,
we also find ravine/valley variants within the lower
montane belt. Such lower montane ravine/valley rain
forests are floristically very similar to the submontane
ravine/valley forests, especially for the herbaceous
understory. But several species are, as far as we know,
endemic to such habitat type: Pisonia sechellarum
F.Friedmann and Psychotria silhouettae F.Friedmann
(see the “Pisonia forest” of Gerlach, 1993). This
variant has been observed only in one small area of
Silhouette, but it is probable that similar plant communities, or other micro-endemics, will be found
when other high altitude exposed ravines have been
prospected in detail, especially in the north of Mont
Dauban (in the higher part of the rivière Machabée
valley) and on the northern parts of Morne Seychellois and Pérard (on Mahé).
Finally, when climbing the highest peaks of Seychelles (on Mahé), it appears that further transitions
could be recognized at higher altitude. At about 780 m
altitude, both on the main summit of Congo Rouge
eschweizerbartxxx author
and on the north-eastern slope of Morne Seychellois,
the tree ferns become extraordinary abundant and
attain huge sizes, even far from the ravines and valleys. Rare and/or small species from the typical lower
montane rain forest become at the same time both
more common and larger, i.e. Canthium sechellense
Summerh., Syzygium wrightii, Gastonia crassa. This
higher altitude variant of the lower montane rain forest corresponds to what has also been described on
central African mountains (Letouzey 1968, Monod
1960). Such a tree fern lower montane rain forest is
only present on the two mentioned mountains and
on Pérard. In addition, the very top of Morne Seychellois, above ca. 880 m, is marked in the upper tree
layer by the monodominance of Pandanus sechellarum (usually an uncommon shrub or under-canopy
tree in lower montane rain forests) and the absence
of most of the typical lower montane understory
species, and could therefore also constitute another
vegetation belt. These very high altitude variants have
never been studied in detail.
Conservation rank: G2(-G3) (imperilled to vulnerable). Within the 3 lower montane forest plots, we recorded 58 taxa including 52 indigenous ones (90 %).
About 62 % of these indigenous taxa are endemic to
Seychelles. Here the indigenous taxa are still clearly
more important than exotics in terms of species, and
about one tree out of two (46 %) belongs to a native
species. This vegetation type is still well represented
on Mahé, and imperilled on Silhouette. We estimate
the potential area of lower montane rain forests at
about 52 ha (2.6 % of Silhouette island, Fig. 4), of
which 60–75 % are undisturbed stands. As for submontane forests, restoration will be problematic. The
unstudied uppermost altitudinal belts (the tree fern
variant of the lower montane belt and the summit
Pandanus forest) are only present on Mahé and must
be considered as critically imperilled due to their very
limited surface areas.
Conclusions
In the literature on Seychelles, the lower montane
and submontane rain forests have been distinguished
mainly based on tree species and moss abundance. In
the present study, however, we identified a number of
indicator species within all forest strata and biological types. We conclude that most of the best indicator
species are to be found in the understory, i.e. ferns,
and that the lack of detailed consideration of this
stratum is responsible for the incomplete description
of the altitudinal belts and partial misunderstanding
in previous literature on Seychelles (see variation in
given altitudinal intervals, mixing of cited indicator
species belonging to different ecological groups, etc.).
The importance of the understory species in forest
types studies has been underlined by other authors
in other tropical regions (Hemp 2002, Poulsen et al.
2005, Senterre 2005a).
Forest types on Silhouette Island
171
eschweizerbartxxx author
Fig. 4. Potential distribution of the main forest types of Silhouette. Light grey = lowland rain forest; dark grey = submontane rain
forest; black = lower montane rain forest; white = ravine/valley forests; crosshatch = mangroves, swamp forests, marshes; dotted =
glacis.
The altitudinal zonation of forest types on Seychelles is characterised by a marked telescoping effect
(Bruijnzeel & Hamilton 2000, Senterre 2005b).
On most of the larger, or drier, tropical islands, and
humid continental mountains, the submontane belt
develops above ca. 700–900 m and turns into the lower montane belt above ca. 1500 m. Where the level of
humidity is very high (tropical islands like Seychelles,
coasts like some humid parts in the Gulf of Guinea)
and combined with a small land mass resulting in
lower temperatures (only on small high islands), the
altitude needed for the air to reach condensation is
lower. Such mountains with a marked telescoping effect are rare throughout the tropics: Mont Doudou in
Gabon (Sosef et al. 2004, associated with cold sea surface currents), the Monte Mitra in Equatorial Guinea
(littoral mountain), Korup and Rumpi mounts in
Cameroon, Mayotte (Pascal et al. 2001), Mount
Payung near the western tip of Java, Mount Finkol
on Kosrae island (Micronesia), Gau island in the Fiji
archipelago, Mt. Tinggi (Bawean island). As far as we
know, few of these regions have been studied in detail for the complete gradient, from lowland to lower
montane, passing through submontane belt (Pendry
& Proctor 1997). In many areas, the reason is that
the two lowermost belts (lowland and submontane)
have already been removed (FAO 2000, MuellerDombois & Fosberg 1998). In Seychelles, we still
have well preserved submontane forests, and further
explorations may reveal more undisturbed lowland
forests on Silhouette (i.e. south-western slopes of
Mont Dauban). In addition, islands like Silhouette or
Mahé are both relatively small but high, which reduces the land mass elevation effect (“Massenerhebung
effect”). Further studies are needed on the factors
influencing the telescoping effect, such as comparing
the mountains with a telescoping effect on distinct
islands: floristic patterns, diversity, phytogeographical and phylogenetical affinities, island size, climate,
isolation and geographical context.
The better understanding of the submontane belt
resulting from the present study also underlines the
importance of Silhouette as the last island of Seychelles where enough old growth forests exist to be
able to identify suitable indicator species. Indeed, on
Mahé, most of the submontane forests are early sec-
172
B. Senterre et al.
ondary forests with a very irregular canopy and many
more individuals of invasive species in all strata. Actually we have not seen or heard of any area of submontane forest on Mahé that could be compared to
areas such as Grand Congoman on Silhouette. Once
the indicator species have been identified based on
well preserved areas like those of Silhouette, then it is
possible to use indicator species from the understory
and thus still be able to recognize the altitudinal belts
even in secondary woodlands as on Mahé, where all
the tree indicator species have disappeared.
Nevertheless, Mahé also has its originality. The
area of Congo Rouge, Pérard and Morne Seychellois includes considerable large areas of undisturbed
lower montane rain forests and are therefore useful
for a more precise understanding of this forest type
(not so well represented on Silhouette). When comparing the altitudinal zonation in Seychelles to other
tropical regions, and the typical telescoping effect
observed here, it seems very probable that some further altitudinal belts should be distinguished between
the interval of 550 m (beginning of the lower montane belt) and 905 m (top of Morne Seychellois). The
very top of the still pristine Morne Seychellois, dominated by Pandanus sechellarum and characterised by
a clear reduction of the diversity of vascular plants,
could correspond to the upper montane belt, which
develops mainly above ca. 2300 m in most tropical
areas and is characterised by such a reduction of the
vascular epiphytes diversity. Such habitat is unique to
the summit of Morne Seychellois and has never been
studied in detail (Carlström 1996). Therefore, and
despite a reduced native flora, the altitudinal zonation on Seychelles seems to be strongly marked, i.e.
presence of diagnostic species and constancy of the
transition intervals. This is unusual when compared
to younger, species poor islands (i.e. islands east of
the Wallace line: Ashton 2003), and may result from
the longer evolution of the flora of Seychelles allowing more developed niche differentiation. In addition,
glacis from the higher summits of Mahé need to be
explored since they could be refuges for some relict
species of a subalpine belt. In Africa, we know that
montane species can be seen at very low altitude on
inselbergs (Parmentier & Maley 2001, Senterre
2005b: 75), due to a complex combination of historical and ecological parameters (glaciations, steep rock
faces, dryer habitat but periodically humid, etc.).
This study, and especially the more detailed configuration of the altitudinal belts, is important for conservation in Seychelles. A map of these forest types
has been prepared (Fig. 4) and will be the base for
decisions in terms of conservation and further studies
or explorations. We identified priority areas on Silhouette for further studies, especially the three main
areas of primary lowland rain forests. Their study
and conservation must be considered as a high priority. They may be the very last examples of lowland
rain forests in the archipelago and some of them have
never been studied nor explored. Studying such areas
will help us to better understand how to restore de-
eschweizerbartxxx author
graded habitats. We also note that most of the high altitude ravine/valley forests have not been prospected
and should also be considered as a high priority since
such habitat has revealed the presence of microendemic species, e.g. Pisonia sechellarum, and may play
an important role as refugia for species from upper
altitudinal belts.
Following the observations on Silhouette and comparisons on Mahé, we expect to find the remnants of a
submontane belt on Praslin (possibly at Fond Azore,
still not described), where the highest point is about
367 m. Indeed, some submontane species are listed for
this island, i.e. Northea hornei, Roscheria melanochaetes, Protarum sechellarum, Mapania floribundum,
Scleria sieberi, Lindsaea kirkii, and even the more
montane specialist Gastonia crassa. So, where are
these species occurring, and in which habitats in Praslin? We expect most of them to occur in humid microhabitats of boulders of big granite blocks, like at La
Réserve (Mahé), but what about the top of Praslin?
Finally, to understand how these ecosystems are
functioning, their history, their evolution on a short
to long term, we need more detailed study of their
phytogeographical (Procter 1984a) and phylogenetic patterns (Hardy & Senterre 2007), as well as
plant functional types (Senterre 2005a). Indeed, we
know that such characteristics may be very different
from one vegetation type to another. For example, in
Hawai’i, the lowland flora has a typically southeast
Asian affinity while the alpine flora (on the top of the
Haleakalā, Maui) has the strongest affinity with the
Americas, where homologous belts are much more
frequent (Medeiros & Loope 1994: p.33). In the
light of such considerations, what is the phytogeographical pattern of the flora of Seychelles when vegetation types are considered separately: e.g. glacis vs.
forest, lowland vs. montane? Such questions can be
investigated when considering the total flora of such
vegetation types (based on plant community plots) or
just the indicator species (ecological groups, Tanghe
1995). Therefore, in future, the ecology of native species needs to be re-evaluated when ecological equalizations are considered.
Acknowledgements. We are grateful to the Island Development
Company (IDC) and Seychelles Bureau of Standards (SBS) for
supporting the present study. The travel expense and local accommodation were supported by the FNRS (Fonds National pour
la Recherche Scientifique), in Belgium, with the support of the
Laboratoire de Botanique systématique et de Phytosociologie of
the Free University of Brussels, i.e. Professor Jean Lejoly. Finally, the set-up of this project was in collaboration with the Nature
Protection Trust of Seychelles (NPTS) and the Botanical Garden
of Mahé. We are also grateful to Mr. Charles Morel for giving his
support for the collection of plant vouchers and for providing
access to the national collections (SEY), and to Dr. Christopher
Kaiser for his contribution to the writing of the manuscript. The
two referees also provided very useful comments and corrections
of the paper. Finally, we thank people having contributed to the
field data collection or drying of the specimens: Abdula Jumaye,
Danny Malbrook and Frances Seilor.
Forest types on Silhouette Island
173
References
Areces-Mallea, A.E., Weakley, A.S., Li, X., Sayre, R.G., Parrish,
J.D., Tipton, C.V. & Boucher, T. (1999): A Guide to Caribbean Vegetation Types: Preliminary Classification System
and Descriptions – N. Panagopoulos (ed.) The Nature Conservancy.
Ashton, P.S. (2003): Floristic zonation of tree communities on
wet tropical mountains revisited. – Perspect. in Plant Ecol.
Evol. and Syst. 6 (1–2): 87–104.
Aubréville, A. (1965): Principes d’une systématique des formations végétales tropicales. – Adansonia 5 (2): 153–197.
Baker, J.G. (1877): Flora of Mauritius and the Seychelles. –
Reeve & Co., London.
Boughey, A.S. (1955a): The nomenclature of the vegetation
zones on the mountains of the Tropical Africa. – Webbia 11:
413–423.
– (1955b): The vegetation of the mountains of Biafra. – Proc.
Linn. Soc. (Bot.) 165 (2): 144–150.
– (1965): Comparisons between the montane forest floras of
America, Africa and Asia. – Webbia 19: 507–517.
Braun-Blanquet, J. (1932): Plant sociology: the study of plant
communities. – McGraw-Hill Book Company, Inc., New
York.
Bruijnzeel, L.A. & Hamilton, L.S. (2000): Decision time for
cloud forests. – IHP Humid Tropics Programme Series, Vol.
13. UNESCO.
Budowski, G. (1965): Distribution of tropical American rain
forest species in the light of successional processes. – Turrialba 15 (1): 40–42.
Bussmann, R.W. (2006): Vegetation zonation and nomenclature of African mountains – An overview. – Lyonia 11 (1):
41–66.
Camirand, R. & Evelyn, O.B. (2003): Ecological Land Classification for Forest Management and Conservation in Jamaica. – Tecsult International, Québec, Canada.
Carlström, A. (1996): Areas of Special Conservation Value for
the Plants of the Granitic Islands of Seychelles. – Unpublished report, Seychelles Government. Ministry of Foreign
Affairs, Planning and Environment. Conservation & National Parks Section, Victoria, Seychelles.
Carrière, S.M., Randrianasolo, E. & Hennenfent, J. (2008): Aires protégées et lutte contre les bioinvasions: des objectifs
antagonistes? Le cas de Psidium cattleianum Sabine (Myrtaceae) autour du parc national de Ranomafana à Madagascar. – VertigO 8 (1): 1–14.
Chapman, V.J. (1976): Mangrove vegetation. – J. Cramer, Vaduz.
Christensen, C. (1912): On the ferns of the Seychelles and the
Aldabra group. – Trans. Linn. Soc. Lond. 2nd Ser. Zool. 15:
407–422.
Churchill, S.P., Balslev, H., Forero, E. & Luteyn, J.L. (1995):
Biodiversity and conservation of neotropical montane forests: Proceedings of the neotropical montane forest biodiversity and conservation symposium. – The New York Botanical Garden, USA.
Dufrêne, M. & Legendre, P. (1997): Species assemblages and indicator species definition: the need of an asymmetrical and
flexible approach. – Ecol. Monogr. 67: 345–366.
FAO (2000): FAO Workshop – Data Collection for the Pacific
Region. – Forest Resources Assessment Working Paper –
051 Forestry Department, Apia, Samoa.
Fleischmann, K., Porembski, S., Biedinger, N. & Barthlott, W.
(1996): Inselbergs in the sea: vegetation of granite outcrops
on the island of Mahé, Praslin and Silhouette (Seychelles). –
Bulletin of the Geobotanical Institute ETH 62: 61–74.
eschweizerbartxxx author
Fleischmann, K., Héritier, P., Meuwly, C., Küffer, C. & Edwards, P.J. (2003): Virtual gallery of the vegetation and flora
of the Seychelles. – Bulletin of the Geobotanical Institute
ETH 69: 57–64.
Fleischmann, K., Edwards, P.J., Ramseier, D. & Kollmann, J.
(2005): Stand structure, species diversity and regeneration of
an endemic palm forest on the Seychelles. – Afr. J. Ecol. 43
(4): 291–301.
Friedmann, F. (1994): Flore des Seychelles: Dicotylédones. –
Orstom, Paris.
Gerlach, J. (1993): The conservation of Silhouette island, Seychelles. I. Plants. – Phelsuma 1: 18–29.
– (1997): A study of habitat structure and vegetation in Seychelles. – Phelsuma 6: 53.
– (2003): The biodiversity of the granitic islands of Seychelles. – Phelsuma 11 (B):1–47.
Grimshaw, J.M. (2001): What do we really know about the Afromontane Archipelago? – Syst. Geogr. Plants 71: 949–957.
Grossman, D.H., Faber-Langendoen, D., Weakley, A.S., Anderson, M., Bourgeron, P., Crawford, R., Goodin, K., Landaal,
S., Metzler, K., Patterson, K.D., Pyne, M., Reid, M. & Sneddon, L. (1998): International classification of ecological communities: terrestrial vegetation of the United States. Volume
I. The National Vegetation Classification System: development, status, and applications. – The Nature Conservancy,
Arlington, Virginia, USA.
Hardy, O. & Senterre, B. (2007): Characterizing the phylogenetic structure of communities by an additive partitioning of
phylogenetic diversity. – J. Ecol. 95: 493–506.
Hedberg, O. (1951): Vegetation belts of the East African mountains. – Svensk Botanisk Tidskrift 45 (1): 141–202.
Hemp, A. (2002): Ecology of the Pteridophytes on the southern slopes of Mt. Kilimanjaro – I. Altitudinal distribution. –
Plant Ecol. 159 (2): 211–239.
Holdridge, L.R. (1967): Life zone ecology. – Tropical Science
Center, San Jose, Costa Rica.
Jeffrey, C. (1962): The Botany of the Seychelles (Report by the
visiting botanist of the Department of Technical Cooperation
Seychelles Botanical Survey, 1961–62). – London. 37 pp.
Keppel, G. (2002): Coastal Vegetation of Taunovo Bay, Pacific
Harbour, Viti Levu, Fiji – A Proposed Development Site. –
S. Pac. J. Nat. Sci. 20 (A): 25–29.
Kottek, M., Grieser, J., Beck, C., Rudolf, B. & Rubel, F. (2006):
World Map of the Köppen-Geiger climate classification updated. – Meteorol. Z. 15: 259–263.
Legendre, P. & Legendre, L. (1998): Numerical Ecology. – Developments in Environmental Modelling, Vol. 20. Elsevier
Science B.V., Amsterdan.
Letouzey, R. (1968): Etude phytogéographique du Cameroun. –
P. Lechevalier, Paris.
Macnae, W. (1971): Mangroves on Aldabra. A Discussion on the
Results of the Royal Society Expedition to Aldabra 1967–
68. – Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 260 (836):
237–247.
Medeiros, A.C. & Loope, L.L. (1994): Rare animals and plants
of Haleakalā National Park. – Hawai’i Natural History Association.
Monod, T. (1960): Notes botaniques sur les îles de São Tomé et
Príncipe. – Bulletin de L’IFAN T. XXII, sér.A, n°1: 19–79.
Mueller-Dombois, D. & Ellenberg, H. (1974): Vegetation types:
a consideration of available methods and their suitability for
various purposes. – Technical Report 49, Island Ecosystems
IRP, U.S. International Biological Program. 50 pp.
Mueller-Dombois, D. & Fosberg, F.R. (1998): Vegetation of
the Tropical Pacific Islands. – Ecological Studies, Vol. 132.
Springer, New York.
174
B. Senterre et al.
Ohsawa, M. (1993): Latitudinal pattern of mountain vegetation
zonation in southern and eastern Asia. – J. Veg. Sci. 4 (1):
13–18.
Oldeman, R.A.A. (1990): Forests: elements of silvology. –
Springer-Verlag, Berlin, New York.
Parmentier, I. & Maley, J. (2001): L’arbre et le pigeon, ou le pigeon et l’arbre? – Canopée 19: 12–14.
Pascal, O., Labat, J.-N., Pignal, M. & Soumille, O. (2001): Diversité, affinités phytogéographiques et origines présumées
de la flore de Mayotte (Archipel des Comores). – Syst.
Geogr. Plants 71: 1101–1123.
Pendry, C.A. & Proctor, J. (1997): Altitudinal zonation of rain
forest on Bukit Belalong, Brunei: Soils, forest structure and
floristics. – J. Trop. Ecol. 13: 221–241.
Piggott, C.J. (1968): A soil survey of the Seychelles. – Min.
Overseas Dev. Tech. Bull. 2. Overseas Surveys, England.
Poulsen, A.D., Hafashimana, D., Eilu, G., Liengola, I.B., Ewango, C.E.N. & Hart, T.B. (2005): Composition and species
richness of forest plants along the Albertine Rift, Africa. –
Biol. Skr. 55: 129–143.
Procter, J. (1984a): Floristics of the granitic islands of the Seychelles. – In: D.R. Stoddard (ed.): Biogeography and ecology
of the Seychelles islands, Pp. 209–220. – Junk Publishers,
The Hague.
– (1984b): Vegetation of the granitic islands of the Seychelles. –
In: D.R. Stoddard (ed.): Biogeography and ecology of the
Seychelles islands, Pp. 193–208. – Junk Publishers, The
Hague.
Richards, P.W. (1952): The tropical rain forest. – Cambridge
University Press, Cambridge, UK.
Robertson, S.A. (1989): Flowering plants of Seychelles (An annotated checklist of angiosperms and gymnosperms with
line drawings). – Royal Botanic Gardens, Kew.
Rübel, E. (1935): The replaceability of ecological factors and the
law of the minimum. – Ecology 16 (1): 336–341.
Schnell, R. (1976): Flore et Végétation de l’Afrique Tropicale. –
Vol. 1. Gauthier-villars, Bordas, Paris.
– (1977): Flore et Végétation de l’Afrique Tropicale. – Vol. 2.
Gauthier-villars, Bordas, Paris.
Senterre, B. (2005a): Recherches méthodologiques pour la typologie de la végétation et la phytogéographie des forêts denses
d’Afrique tropicale. – Acta Bot. Gall. 152 (3): 409–419.
– (2005b): Recherches méthodologiques pour la typologie de la
végétation et la phytogéographie des forêts denses d’Afrique
tropicale. – Ph.D. thesis, Université Libre de Bruxelles,
Bruxelles. 345 pp.
– (2009): Distribution and determinants of forest fires and
land degradation on Praslin, Seychelles. – Consultancy report, Plan Conservation Action Group, Victoria, Mahé, Seychelles. 72 pp.
Senterre, B. & Lejoly, J. (2001): Los grandes tipos de hábitat
forestales de Río Muni. – In: C. Aedo, R. Morales, M.T. Tellería & M. Velayos (eds.): Botánica y Botánicos en Guinea
Ecuatorial, Pp. 171–199. – Real Jardín Botánico – Agencia
Española de Cooperación Internacional, Madrid.
Senterre, B., Lejoly, J. & Sonké, B. (2004): Analyse du gradient
de continentalité et identification de communautés végétales
en forêts denses d’Afrique centrale par la méthode du mégatransect. – Phytocoenologia 34 (3): 491–516.
Sosef, M.S.M., Issembe, Y.A., Bourobou, H.P. & Koopman,
W.J.M. (2004): Botanical biodiversity of the Pleistocene forest refuge Monts Doudou (Gabon). – In: B.L. Fisher (ed.):
Monts Doudou, Gabon: A floral and faunal inventory with
references to elevational variation, 28, Pp. 17–91. – Memoirs
of the California Academy of Sciences, California Academy
of Sciences, San Francisco.
eschweizerbartxxx author
Tanghe, M. (1995): Groupes écologiques d’espèces végétales et
application à la végétation forestière de la Wallonie. – In: J.F. Dulière, M. Tanghe & F. Malaisse (eds.): Répertoire des
groupes écologiques du fichier écologique des essences. –
Ministère de la Région Wallonne, Direction générale des
Ressources naturelles et de l’Environnement, Belgique.
Tchouto, M.G.P., Edwards, I., Cheek, M., Ndam, N. & Ackworth, J. (1999): Mount Cameroon cloud forest. – In: J.
Timberlake & S. Kativu (eds.): African Plants: Biodiversity,
Taxonomy and Uses, Pp. 263–277. – Royal Botanic Gardens,
Kew.
Thomas, D.W. & Achoundong, G. (1994): Montane forest of
western Africa. – In: J.H. Seyani & A.C. Chikuni (eds.):
Proc. XIIIth Plenary Meeting AETFAT. Pp. 1015–1024, Malawi.
Trochain, J.-L. (1957): Accord interafricain sur la définition
des types de végétation de l’Afrique Tropicale. – Bulletin
de l’Institut d’Etudes centrafricaines, nouvelle série 13–14:
55–93.
– (1980): Écologie végétale de la zone intertropicale non désertique. – Université Paul Sabatier, Toulouse.
van der Maarel, E. (1979): Transformation of cover-abundance
values in phytosociology and its effects on community similarity. – Vegetatio 39 (2): 97–114.
Van Steenis, C.G.G.J. (1935): On the origin of the Malaysian
mountain flora. Part 2. Altitudinal zones, general considerations and renewed statement of the problem. – Bull. Jard.
bot. Buitenz. 3 (13): 289–417.
Vesey-Fitzgerald, D. (1940): On the vegetation of the Seychelles. – J. Ecol. 28: 465–484.
Walsh, R.P.D. (1984): Climate of the Seychelles. – In: D.R. Stoddard (ed.): Biogeography and ecology of the Seychelles islands, Pp. 39–62. – Junk Publishers, The Hague.
White, F. (1978): The afromontane region. – In: M.J.A. Werger
& A.C. Van Bruggen (eds.): Biogeography and ecology of
southern Africa, Pp. 463–513. – Dr. W. Junk b.v. Publishers,
The Hague.
White, F. (1979): The Guineo-Congolian Region and its relationships to other phytochoria. – Bull. Jard. Bot. Nat. Belg.
49 (1/2): 11–55.
White, L. & Edwards, A. (2001): Conservation en forêt pluviale
africaine – Méthodes de recherche. – Wildlife Conservation
Society, New York, USA.
Whitmore, T.C. (1998): An introduction to tropical rain forests. – Oxford University Press, Oxford.
WWF (2007): Granitic Seychelles forests (AT0113). -. URL
http://www.nationalgeographic.com/wildworld/profiles/
terrestrial/at/at0113.html
Address of authors: Bruno Senterre*, 1, 4, Justin Gerlach2, 3,
James Mougal4 & Denis Matatiken4
1
Laboratoire de Botanique systématique et de Phytosociologie, Université Libre de Bruxelles (ULB), CP 169, 50 av. F.
Roosevelt, 1050 Bruxelles, Belgium.
2
133 Cherry Hinton Road, Cambridge, CB1 7BX, England.
3
Nature Protection Trust of Seychelles, PO Box 207, Victoria,
Mahé, Seychelles
4
Present address: Ministry of Environment, Natural Resources & Transport, Division of Nature and Conservation, Terrestrial & Ecological Research Centre, Botanical Gardens,
Mont Fleuri, P.O. Box 445, Republic of Seychelles.
Corresponding author, e-mail: bsenterre@gmail.com