ARTICLE IN PRESS
Flora 200 (2005) 416–433
www.elsevier.de/flora
Superpáramo plant species diversity and phytogeography in Ecuador
Petr Sklenářa,, Henrik Balslevb
a
Department of Botany, Charles University, Benátská 2, 128 01 Prague, Czech Republic
Department of Systematic Botany, Aarhus University, Universitetsparken Bygn. 540, DK–8000 Aarhus C., Denmark
b
Received 14 April 2004; accepted 21 December 2004
Abstract
In 18 superpáramo sites in Ecuador we found 388 species of vascular plants belonging to 146 genera and 52 families,
making the Ecuadorian superpáramo flora richer in species than that of Venezuela and that of Colombia which
appears to have fewer species although the number remains uncertain. The most species rich families were Asteraceae
(83 species) and Poaceae (49) which also dominate the grasspáramo that surrounds the superpáramo. Otherwise the
superpáramo is dominated by families that are mostly herbaceous such as Cyperaceae, Brassicaceae, Caryophyllaceae
and Valerianaceae, whereas shrubby families that dominate the subpáramo and grasspáramo, such as
Melastomataceae, Ericaceae, and Solanaceae, have only few species or are absent in the superpáramo. The generic
spectrum was dominated by a suite of species-rich genera (Lachemilla, Gentianella, Valeriana, Calamagrostis, Draba)
with many páramo-endemic species reflecting a high level of autochthonous speciation in the (super)páramo. Species
richness varied from 71 to 149 in the individual superpáramos surveyed, but species richness was only weakly
correlated to their area and log-transformed area, and negatively correlated to the vertical range of each superpáramo.
b-diversity was significantly correlated to the vertical range (i.e., the number of surveyed altitudinal levels) but it was
not correlated to the area or log-transformed area of each superpáramo. Most of the species were narrowly distributed
and 112 (29%) of them were found in a single superpáramo, while eight (2%) occurred in all 18 superpáramo sites.
Floristic similarity was not correlated to the geographical distance between the sites. Redundance Analysis suggested
that geological origin of the substrate (metamorphic versus volcanic bedrock) was important for the floristic
composition. Occurrence of mountains built from metamorphic rocks, is however, correlated to areas with high
rainfall and the amount of rainfall may be a stronger determinant for species distribution than the presence of volcanic
versus metamorphic bedrocks. TWINSPAN cluster analysis divided the 18 sites into three groups, which corresponded
to dry, humid and very humid superpáramos. The groups were also separated along the first ordination axis in
Correspondence Analysis, while the second axis may correlate to the volcanic history of the area.
r 2005 Elsevier GmbH. All rights reserved.
Keywords: Andes; Habitat islands; Páramo; Phytogeography; Species distribution; Tropical alpine
Introduction
Corresponding author.
E-mail addresses: petr@natur.cuni.cz (P. Sklenář),
henrik.balslev@biology.au.dk (H. Balslev).
0367-2530/$ - see front matter r 2005 Elsevier GmbH. All rights reserved.
doi:10.1016/j.flora.2004.12.006
The superpáramo is the upper vegetation belt of the
páramo – an essentially aseasonal high elevation grassand shrub-vegetation of the humid tropical Andes with
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
outliers in Costa Rica and Panama. Where more
seasonal climate reigns further north and south, the
páramo gives way to other vegetation types above the
timberline; in Guatemala and Mexico to herbaceous and
tussock-grass communities called ‘‘zacatonales’’ and in
central Peru and Bolivia to dry, barren grass- and bushsteppe called ‘‘puna’’ (Almeida-Lenero et al., 2004;
Balslev and Luteyn, 1992; Islebe and Cleef, 1995;
Luteyn, 1999; Simpson, 1983; Troll, 1959, 1968). Like
other tropical mountains, the Andes are floristically rich
and often have high endemism, particularly in their
upper zones such as the tropical alpine belts above the
timber line (Balslev and Luteyn, 1992; Hedberg, 1957;
Smith, 1975). Among the tropical alpine ecosystems, the
páramo is the most diverse, mainly because of its large
geographic extent (Hedberg, 1992; Smith and Cleef,
1988; Vuilleumier and Monasterio, 1986). About 3500
species of vascular plants have been reported from the
páramo and possibly as many as 60% of them are
endemic (Luteyn, 1992, 1999). The high concentration
of endemic species is usually attributed to changes of
horizontal and vertical extent of the páramo during
Pleistocene climatic oscillations and associated pulses of
geographic isolation (Cuatrecasas, 1986; Simpson and
Todzia, 1990; Van der Hammen and Cleef, 1986; Van
der Hammen et al., 1973).
Above 4100–4300 m elevation, the so-called superpáramo, which is the subject of this study, replaces the
páramo grasslands (or the bamboo páramo in humid
areas) and extends to the permanent snow-line at
74800 m. In the superpáramo, vegetation cover is
broken up and patches of open soil appear, the tussock
grasses that are characteristic of the grass páramo are
less abundant, and the vegetation is dominated by other
life-forms such as small herbs, cushion plants, and
sclerophyllous dwarf shrubs. Superpáramo vegetation
has been reported from Venezuela, Colombia, and
Ecuador (Berg, 1998; Cleef, 1981; Cuatrecasas, 1968;
Luteyn, 1992; Monasterio, 1979), whereas the summits
in Panama and Costa Rica do not reach elevations that
are high enough to support superpáramo. In northern
Peru, several summits of the ‘‘jalca,’’ which is the local
name for páramo, reach elevations at which superpáramo may have developed, but so far there are no
descriptions in the literature that document the existence
of superpáramo vegetation in Peru. On a few occasions
the name superpáramo has been used for high-elevation
vegetation as far south as Bolivia (Cuatrecasas, 1968),
but since it replaces puna there, super-páramo is an
inappropriate name.
Since the superpáramo is confined to the highest
mountains, it occurs scattered as ‘‘islands’’ separated by
grass páramo, mountain forest, and deep Andean
valleys (Cuatrecasas, 1958; Simpson, 1974, 1975; Van
der Hammen and Cleef, 1986). In Ecuador, superpáramo is particularly well represented and occurs
417
isolated on 25 or more mountains in both the western
and eastern cordilleras along a stretch of ca. 450 km
from the Colombian border in the north to the province
of Azuay in the south. On most of these mountains the
superpáramo is developed on volcanic bedrocks but on
a few mountains in the eastern cordillera it has
developed on metamorphic bedrocks. Depending on
local climatic conditions and orientation relative to the
trade winds, the superpáramo may be dry or humid.
Some areas of superpáramo are found on mountains
that are high enough to have permanent glaciers
whereas others do not reach such heights. This setting
is an attractive one for comparative studies of species
diversity, taxonomic composition, and distribution of
vascular plant species, and phytogeographic relations
among the superpáramo ‘‘islands’’. In this paper, we
address the following questions: (1) How species rich is
the Ecuadorian superpáramo compared to Colombian
and Venezuelan superpáramos? (2) Is the superpáramo
flora dominated by the same taxa as the grass páramo
that surrounds it, and are some genera particularly
diverse, indicating high levels of autochthonous speciation? (3) How does species richness and b-diversity vary
among Ecuadorian superpáramos and is the variation
related to their areas or their vertical ranges? (4) How
are individual species distributed among the superpáramos and how is endemism structured? (5) How
floristically similar are the superpáramos and are the
observed differences related to ecological or other
factors?
Study sites
We selected 16 representative areas of superpáramo in
the eastern and western cordilleras of the Andes in
Ecuador (Fig. 1) stretching over 430 km from Volcán
Chiles (01480 N, 771570 W) on the border with Colombia
to Cajas National Park (21520 S, 791180 W) in southern
Ecuador. The superpáramos on Chimborazo and
Antisana were studied on two opposite sides of the
mountains. These data were treated as independent sets,
giving a total of 18 study sites, of which 14 were located
on active or extinct volcanoes, and four on mountains of
non-volcanic origin with metamorphic bedrocks. Nine
of the studied mountains are topped by glaciers today,
but all of them were glaciated at some point during
Pleistocene (Hastenrath, 1981). The basic characteristics
of each site are summarized in Table 1 and further
details can be found in Sklenář (2000).
In Ecuador, the superpáramo forms two distinct
altitudinal belts. The lower superpáramo from 4100 to
(4400–)4500 m is usually rich in species. It is characterized by sclerophyllous dwarf shrubs (Loricaria
spp., Chuquiraga jussieui, Valeriana microphylla,
ARTICLE IN PRESS
418
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
78 W
1
Ecuador
South
America
2
3
4
5
0
6
8 7
9
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
11
13 14
12
15
16
17
18
0
50
100
species
richness
Chiles
103
Cotacachi
102
Imbabura
114
Cayambe
138
Saraurcu (M)
95
Pichincha
131
Antisana - east
129
Antisana - west
143
Iliniza
143
Cotopaxi
91
Hermoso (M)
122
Tungurahua
71
Chimborazo - west 107
Chimborazo - east 131
El Altar
132
Quilimas (M)
137
Yanaurcu (M)
106
Cajas
149
beta
diversity
3.2
1.8
1.9
3
1.8
2.2
2.4
2.3
3.9
3.6
1.4
1.2
1.6
2.1
1.9
1.9
0.7
1.4
150 km
Fig. 1. Distribution of 18 vegetation study sites in the
superpáramo in Ecuador along a 430 km long stretch of the
Andes in both the western and eastern Cordilleras. Squares
correspond to Group 1 and circles to Group 2 in Fig. 4. (M)
after the name of site indicates metamorphic bedrock.
Diplostephium rupestre, Pentacalia spp.), and cushion
plants (Plantago rigida, Xenophyllum humile, Azorella
spp.), but tussock grasses (Calamagrostis intermedia,
Festuca asplundii) are usually also important. The upper
superpáramo above (4400–)4500 m is characterized by
the presence of Senecio nivalis, S. canescens, Arenaria
dicranoides, and Cerastium floccosum. The vegetation is
poor in species and confined to favorable habitats.
Tussock grasses and sclerophyllous shrubs are mostly
absent, and the most common growth-forms are shortstem grasses, prostrate subshrubs and herbs, acaulescent
rosettes, and cushion plants (Sklenář, 2000).
Methods
Field work was carried out from May through
December 1995, June through September 1997, and
June through July 1999. The vegetation was sampled in
a stratified random design (Mueller-Dombois and
Ellenberg, 1974). This sampling was limited to the zonal
vegetation, avoiding azonal patches of cushions mires,
stream-sides, etc. The lower limit of the superpáramo
was here defined as the transition to the grass páramo,
i.e., where a continuous grass cover begins, while the
upper limit was defined as where the plant cover
dropped below 5%. At 100 m altitudinal (stratified)
levels we randomly set three plots along a 100 m long
transect running parallel with the contours. Most plots
were 25 m2, but in the upper belt (generally above
4400 m) they were 100 m2 because there we found
increased spatial variability in species composition. In
each plot we scored all species using an 8-grade semiquantitative scale combining abundance and cover
(Mueller-Dombois and Ellenberg, 1974). Data were
converted to their midpoint values (r ¼ 0.05%,
+ ¼ 0.5%,
1 ¼ 2.5%,
2a ¼ 10%,
2b ¼ 20%,
3 ¼ 37.5%, 4 ¼ 62.5%, 5 ¼ 87.5%) for further analysis.
In addition to the quantitative sampling we searched
the study sites for additional species outside the sample
plots but this time including also azonal habitats, such
as rocky outcrops, cushion mires, and lake or stream
shores, and they were added to the species lists for each
site. Species planted occasionally around mountain
refuges or occurring on ruderal sites around huts were
excluded from the phytogeographic analyses. Introduced species that were naturalized, i.e., they reproduce
and survive independently in natural plant communities,
were included in all analyses. Identification was backed
up with voucher specimens and usually verified by
specialists for each taxonomic group. Information about
species endemism in Ecuador was extracted from
Jørgensen and León-Y. (1999) and Valencia et al.
(2000) and was updated from our own knowledge.
b-diversity was calculated from the vegetation samples for each site as: bT ¼ s=a 1, where s is the total
number of species present in all samples at a site and a is
the average number of species in the samples from that
site (Whittaker, 1960; Wilson and Shmida, 1984).
Correlation of species richness and b-diversity to
environmental and spatial variables were calculated.
The explanatory variables were superpáramo area
(measured as plane area estimates from the IGM
1:50,000 maps; if present, the extension of glaciers was
excluded), vertical range of the superpáramo (i.e., for
species richness vertical distance between the base and
the top of the superpáramo or the position of the
permanent snow-line, but for b-diversity number of
surveyed altitudinal levels), presence/absence of glacier,
volcanic/metamorphic bedrock, volcanic activity yes/no,
position in western or eastern cordilleras.
Floristic relationship between pairs of sites was
measured as the difference between observed and
expected numbers of shared species (Connor and
Simberloff, 1978). Expected numbers were estimated
by a random pair-wise sampling from the entire species
list in which each species had an adjusted probability of
being sampled, for species found in the vegetation plots
the expected numbers were the frequency of occurrence
and for species found only outside the vegetation
plots the expected numbers were the number of site
Table 1.
Baseline data for the 18 study sites.
Maximum
elevation (m)
Coordinates
latitude longitude
Estimated superpáramo
area (km2)
Vertical
range (m)
Recent
glacier
Geological
origin
Recent volcanic
activity
Chiles
Cotacachi
Imbabura
Cayambe
Saraurcu
Pichincha
Antisana-west
Antisana-east
Iliniza
Cotopaxi
Hermoso
Tungurahua
Chimborazo-west
Chimborazo-east
El Altar
Quilimas
Yanaurcu
Cajas
Carchi
Imbabura
Imbabura
Pichincha
Pichincha
Pichincha
Napo
4768
4939
4609
5790
4676
4698
5704
No
No
No
No
—
Yes
Yes
Yes
Yes
No
Yes
Yes
Volcanic
Volcanic
Metamorphic
Volcanic
Volcanic
No
Yes
—
Yes
No
Chimborazo
Chimborazo
Chimborazo
Azuay
5319
4730
4580
4451
564
737
330
600
576
398
500
500
600
600
539
400
700
500
500
630
380
151
Volcanic
Volcanic
Volcanic
Volcanic
Metamorphic
Volcanic
Volcanic
5263
5897
4639
5016
6310
6.4
6.7
1.6
43.3
12.6
6.5
29
29
26.9
65.5
7.2
11.1
62.2
62.2
44.3
47.2
15.4
17.7
No
Yes
No
Yes
Yes
No
Yes
Cotopaxi
Cotopaxi
Tungurahua
Tungurahua
Chimborazo
01480 N 771570 W
01220 N 781200 W
01160 N 781100 W
01020 N 771590 W
01060 S 771550 W
01100 S 781330 W
01300 S 781100 W
01270 S 781070 W
01400 S 781420 W
01400 S 781260 W
11100 S 781120 W
11270 S 781260 W
11280 S 781520 W
11280 S 781460 W
11410 S 781240 W
11460 S 781240 W
21150 S 781300 W
21520 S 791180 W
Yes
No
No
No
Volcanic
Metamorphic
Metamorphic
Volcanic
No
—
—
No
The vertical range of the superpáramo is the vertical distance in meters between the lowest sampling plot and the top of the mountain or the transition to the permanent snow-line, coordinates
indicate the approximate location of the transects, estimated superpáramo area is a plane estimate obtained from the IGM maps (excluding glaciers where present).
ARTICLE IN PRESS
Province
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
Name of site
419
ARTICLE IN PRESS
420
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
occurrences (null hypothesis II of Connor and Simberloff,
1978). The size of the random sample was equal to the
actual number of species for any given site. The expected
numbers of shared species, independent of sample size
(Connor and Simberloff, 1978; Simberloff et al., 1981),
were estimated repeatedly 999 times via computer
simulation (Sklenář, 2000). The chi-square (w2 ) distance
between observed and expected numbers was used as a
standardized measure of floristic relationships between
the mountain pairs. The w2 values were tested against
geographical distances between the sites by means of
Mantel test, part of the R-Package (Legendre and
Vaudor, 1991); the null hypothesis of the test was that
there would be no linear relationship between the
geographical distance and the floristic similarity.
Two-way indicator species analysis (TWINSPAN;
Hill, 1979) was used to group the 18 sites according to
floristic and vegetation similarities with presence–absence data for all species and semi-quantitative abundance estimates for species present in the plots based on
average values of each species in the plots at each site.
For the abundance data we used CA (Correspondence
Analysis, part of CANOCO package; Ter Braak and
Šmilauer, 1998) to demonstrate the relative position of
study sites in the multivariate space; data were squareroot transformed and rare species were down-weighted
prior the analysis. For the presence–absence data we
used RDA (Redundancy Analysis) to test the correlation between species distribution among the sites and six
environmental and spatial variables.
Results
Species richness and taxonomic composition
There were 383 species, 146 genera and 52 families of
vascular plants in the 18 Ecuadorian superpáramo sites
surveyed. There were two subspecies in Eudema nubigena
and Valeriana alypifolia, two varieties in Potentilla
dombeyi, and two hybrids in Gentianella, which we, for
simplicity treated in the same way as species in the
analyses, so the number of taxa (subsequently referred to
as ‘‘species’’) was 388. Of these, 374 were recorded within
the continuous superpáramo, i.e., above the transition with
the grass- or bamboo páramo at (4000–) 4100–4200 m. The
remaining 14 species were found within 100 m below this
limit, in isolated patches of superpáramo vegetation, such
as rock crevices, surroundings of boulders, lake shores, or
eroded land, and we therefore expect that they will
eventually be found in the continuous superpáramo. Of
the 388 species, five (Aira caryophyllea, Anthoxanthum
odoratum, Poa annua, Rumex acetosella, and Sagina
procumbens) are introduced to Ecuador. A full listing of
the species is provided by Sklenář (2000).
Species diversity was unevenly distributed among
taxonomic groups (Fig. 2). Asteraceae and Poaceae
entirely dominated the flora and together they comprised one-third of the total species diversity. Other
families were conspicuously less diverse and there were
16 families with a single species. There were a suite of
species rich genera, but the majority of the 146 genera
present in the Ecuadorian superpáramo, i.e., 84 genera,
were represented by a single species.
Variation in species richness and b-diversity among
superpáramos
Species richness differed considerably among the
superpáramo sites (Fig. 1). Cajas was richest in species
(149 species) even if it is one of the lowest mountains
(Table 1). The lowest number of species was found in
Tungurahua (71) and Cotopaxi (91), which are the
mountains that have the most recent volcanic activity.
Species richness correlated weakly with the estimated
area of superpáramo at each site (Spearman rank
correlation coefficient r ¼ 0:173, p ¼ 0:49), or the logtransformed area (r ¼ 0:264, p ¼ 0:29), and correlation
with the vertical range of the superpáramo was negative
(r ¼ 0:191, p ¼ 0:45; n ¼ 18 in all cases). When the
correlation analyses were performed independently for
each of the two groups of mountains resulting from the
cluster analysis (see below), the results became significant for the area and log-area of the 13 mountains of
Group 2 (r ¼ 0:664, p ¼ 0:01) but not for the five
mountains of Group 1 (r ¼ 0:667, p ¼ 0:22). The
correlation did not change for the vertical range of the
superpáramo (Group 1: r ¼ 0:395, p ¼ 0:51; Group 2:
r ¼ 0:088, p ¼ 0:77).
A considerable spatial pattern of b-diversity in the
superpáramo vegetation (Fig. 1) was significantly
correlated with the number of altitudinal levels surveyed
at each site (r ¼ 0:59, p ¼ 0:01), whereas its correlation
with other variables (i.e., area, vertical range of superpáramo) was weak and not significant (p40:05). This
significant correlation to the altitudinal levels remained
within the cluster Group 2 (r ¼ 0:605, p ¼ 0:03) but not
within the Group 1 (r ¼ 0:051, p ¼ 0:93), whereas the
other correlations remained not significant.
Species distributions
The majority of species were narrowly distributed; 112
species (29% of the total) were found only at one single
site, and 274 or over two-thirds of the species occurred
at fewer than seven sites (Fig. 3). Only eight species (e.g.,
Calamagrostis intermedia, Gentiana sedifolia, Luzula
racemosa, Poa cucullata) were found at all 18 sites.
Species restricted to either dry or humid sites (Groups 1
and 2 in Fig. 4, resulting from the first step of the cluster
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
421
90
Asteraceae (26 endemic)
80
Families
Number of Species
70
60
50
Poaceae (12 endemic)
40
Brassicaceae (12 endemic)
30
Gentianaceae (11 endemic), Rosaceae
20
Scrophulariaceae, Caryophyllaceae, Cyperaceae
10
0
16
Genera
14
Lachemilla (2 endemic), Gentianella (10 endemic)
Valeriana (5 endemic)
Number of Species
12
Calamagrostis (3 endemic)
Draba (8 endemic)
10
Lupinus
Festuca (5 endemic), Geranium (3 endemic), S enecio
8
Cerastium,Huperzia
6
genera 63-146
4
2
0
1
3
5
7
9
11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63
Fig. 2. Ranked species numbers of families and genera encountered in 18 study sites in the superpáramo in Ecuador.
analyses, see below) were also very narrowly distributed
within these groups, and only few of them occurred at
many sites (Fig. 3). The five dry sites had a total of 241
species, of which 64 were restricted to them, and only
Festuca vaginalis and Plantago nubigena were present in
all of them (Fig. 5A). The 13 humid sites had a total of
323 species, of which 137 were restricted to them, and
there were no species occurring in all of them. The most
widespread were Geranium sibbaldioides (Fig. 5H) and
Luzula gigantea (Fig. 5L) which occurred in 11, and F.
ARTICLE IN PRESS
422
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
120
100
All superpáramo sites
Number of Species
80
Humid superpáramo sites
60
Dry superpáramo sites
40
20
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Number of sites (N)
Fig. 3. Pattern of geographic distribution of all species among the 18 study sites in the superpáramo in Ecuador (open bars), and
pattern of distribution of species restricted to the 13 humid superpáramos (hatched bars) and to the five dry superpáramos (black
bars.).
asplundii (Fig. 5J) and Lachemilla holosericea (Fig. 5I)
which occurred in 10 humid sites, respectively.
Floristic similarity among superpáramos
Grouping of the superpáramo sites was similar in
TWINSPAN clustering using either presence–absence
data of species from the superpáramo sites (from both
zonal and azonal habitats) or abundance data of the
zonal vegetation, and the differences were essentially
only in the final divisions. This shows that adding the
species from the azonal vegetation patches (presence–absence) does not dramatically change the picture
obtained on the basis of the survey that included only
the zonal vegetation (abundance). The analysis based on
abundance generally had higher eigenvalues and only
this result is presented (Fig. 4B).
The first division step separated Chimborazo-west,
Antisana-west, Iliniza, Cotopaxi, and Pichincha (Group
1), due to the occurrence of Plantago nubigena and
Festuca vaginalis (Fig. 5A). In the second group resulting
from the first division (Group 2), Imbabura, Cajas,
Cayambe, and Chimborazo-east were separated next due
to occurrence of Pentacalia peruviana; Cotacachi was
added to this group in the analysis using presence–ab-
sence data. Tungurahua was consistently separated from
the remaining nine mountains because of absence of
Azorella aretioides. The rest formed two sub-groups; the
southern superpáramo sites (Quilimas, Yanaurcu, Altar,
Hermoso) were separated from the north-central ones
(Antisana-east, Saraurcu, Chiles, Cotacachi), due to the
occurrence of Calamagrostis guamanensis (Fig. 5O) in the
latter. This species was recorded also on Hermoso but
only outside vegetation plots and therefore this site
appeared in either the southern or the north-central subgroup in the two analyses.
The first CA ordination axis using the abundance data
captured the gradient expressed by the first division step
of TWINSPAN clustering; sites from the Group 1 occur
on the right side while sites from the Group 2 on the left
side of the ordination diagram (Fig. 4A). Cajas appears
a very distinct site separated from the remainder by the
second ordination axis.
A total of 374 species formed the common species
pool, which was used for estimations of the expected
numbers of shared species (Appendix 1). Six sites
(Chiles, Cotacachi, Imbabura, Cayambe, Saraurcu,
Antisana-east) had positive w2 values between each pair
of them (the only exception being Imbabura–Antisanaeast), mostly positive or low negative values with
Chimborazo-east, Hermoso, Altar, Quilimas, and Ya-
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
2.5
423
Chimborazo - west
Group 1
(A)
Cajas
(B)
! 0.473
Antisana - west
0.457
Cotopaxi
0.395
Iliniza
Pichincha
0.421
0.403
Imbabura
! 0.325
Cajas
0.526
Cayambe
Chimborazo - east
0.447
0.276
Hermoso (M)
Cotopaxi
Iliniza
Altar
Cotacachi
Yanaurcu (M)
Antisana-west
Saraurcu (M)
Chimborazo-east
Imbabura
Pichincha
Chiles Quilimas (M)
Cayambe
Antisana-east
Tungurahua
Tungurahua
Cotacachi
Chiles
Antisana - east
Saraurcu (M)
Hermoso (M)
El Altar
Yanaurcu (M)
0.495
0.353
0.304
0.398
0.311
0.445
0.423
! 0.328
Quilimas (M)
-1.0
-1.0
Group 2
Chimborazo-west
0
1.5
Fig. 4. Correspondence Analysis of floristic abundance data from 18 superpáramo sites in Ecuador (l1 ¼ 0:316, l2 ¼ 0:138, total
inertia 1.537); squares (&) indicate sites of Group 1 and circles (J) indicate sites of Group 2 resulting from the cluster analysis (A)
and cluster (TWINSPAN) analysis of abundance data from the 18 sites with the first division separating two groups based on the
presence (Group 1) or absence (Group 2) of Festuca vaginalis and Plantago nubigena; eigenvalues for each division step provided, !
indicate divisions with higher residual than the tolerance given by the algorithm (B); (M) indicates site with metamorphic bedrock.
naurcu, and high negative values with Chimborazowest, Cotopaxi, Antisana-west, and Iliniza (Appendix
2). The latter four sites had positive values between each
pair of them. Three sites (Pichincha, Tungurahua,
Cajas), did not appear to demonstrate a clear pattern.
Pichincha showed mostly small values (negative or
positive), the highest in combination with Cotacachi,
Cayambe, and Iliniza, and the lowest with Hermoso and
Cajas. Tungurahua had small negative values with most
of the mountains, only with Chimborazo-west the value
is remarkably low (w2 ¼ 24:56). Cajas was unusual since
all but one of its values were negative, the lowest one
with Chimborazo-west.
The overall spatial pattern of floristic similarities
between the mountains was not correlated with the
geographic distance between the sites; a Mantel test
between the two matrices (Appendices 1 and 2) gives
r ¼ 0:034, p ¼ 0:38, using 999 permutations. The
RDA showed that of the six variables tested (estimated
area, vertical range of superpáramo, presence/absence of
a recent glacier, position in western/eastern cordillera,
presence/absence of recent volcanic activity, and volcanic/metamorphic bedrocks), only the latter (bedrock
volcanic/metamorphic) was significantly correlated to
the observed species composition of the sites (forward
manual selection, po0:05, 199 permutations). However,
the bedrock variable accounted for only a small portion
(10.7%) of total data variability.
Discussion
Species richness
Most of the sites were visited only once, but Iliniza
and Chimborazo were explored repeatedly. The first 5days survey of Iliniza yielded 127 species, but 16
additional species (13% of the total) were found during
the second 7-days intensive research of the mountain
three months later. In 1999, the transect area on
Chimborazo-west was revisited during a 1-day trip.
Four additional species were found (3.5% of the total),
but three of them occurred above the limit surveyed in
1995. Some ‘‘gaps’’ in our species distributions may
therefore result from sampling error. However, if our
experience at Iliniza is representative, we have recorded
80–90% of all species at each superpáramo site.
The remarkable diversity of neotropical páramos,
previously highlighted by Smith and Cleef (1988) and
Luteyn (1999), is underlined by our record of 388 species
from the superpáramo in Ecuador. This number
ARTICLE IN PRESS
424
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
Fig. 5. Examples of distributions of species among 18 study sites in the superpáramo in Ecuador ( species not found at study site;
J species found at study site); A–F species characteristic of dry sites, G–L species characteristic of humid sites, M–R species
characteristic of very humid sites, S–T species negatively characterizing very humid sites, U–V narrowly distributed species with
trans-Andean distribution, W–X narrowly distributed species with and cis-Andean distribution.
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
includes some species that typically occur in the grass
páramo at lower elevations (e.g., Hieracium frigidum,
Stachys elliptica) or bamboo páramo (e.g., Gentianella
splendens), and only marginally enter the lower parts of
the superpáramo belt. On the other hand, we are aware
of several species, such as Azorella ecuadorensis, Eudema
rupestris, and Geranium chimborazense, previously collected in the superpáramo in Ecuador, that were not
encountered in this survey. Those species may either be
very rare, or restricted to sites that were not covered by
our study (e.g., Xenophyllum roseum from the central
part of the Cajas National Park), or possibly they were
not recognized in the field (e.g., sterile specimens of
Carex are very hard to identify to species level).
With 388 species of vascular plants the Ecuadorian
superpáramo flora seems to be richer in species than
other superpáramo floras. In the superpáramo in
Venezuela Ricardi et al. (1997) reported 234 species in
119 genera and Berg (1998) found 168 species in 147
vegetation samples. In the Colombian eastern cordillera
Van der Hammen and Cleef (1986) found 112 species in
69 genera, mostly in the lower superpáramo belt (Cleef,
1981). A transect study in the Parque Los Nevados in
the central cordillera of Colombia, revealed 170 species
growing above 4000 m (Cleef et al., 1983; Rangel et al.,
1983). In this study we surveyed representative superpáramo sites throughout Ecuador on 16 mountains, but
at least ten distinct superpáramo areas were not
explored (e.g., Carihuairazo, Quilindaña, Sincholagua).
In Venezuela, superpáramo is confined to only three
areas in the surroundings of Mérida (Monasterio, 1979),
and the low number of superpáramo species reported
from that country may therefore reflect the actual
situation well. In Colombia, in contrast, there are many
areas that reach the superpáramo belt (Cleef, 1981), but
their total area may not surpass that of the Ecuadorian
superpáramo (Hofstede et al., 2003). Different total
areas of superpáramo in Ecuador and Colombia may
determine the differences in species richness between
their floras. Nevertheless the number of known Colombian superpáramo species will certainly grow when more
sites are explored (A. M. Cleef, pers. comm.; P. Pedraza,
pers. comm.) so definitive conclusions concerning the
relative number of species in the three countries cannot
be made until the species lists are more complete.
Taxonomic composition
It is well known that Asteraceae and Poaceae dominate
the páramo flora (Luteyn, 1999; Luteyn et al., 1992;
Ramsay, 1992; Ricardi et al., 1997) and this study
confirms that these two families are the most species rich
also in the superpáramo flora. Together they account for
one third of all superpáramo species in Ecuador, which is
roughly the same proportion as in the páramo in general
425
(Luteyn, 1999). Their importance is further underlined by
the fact that nearly half of the 40 most important
(combining frequency and abundance) species of the
Ecuadorian superpáramo belong to these two families
(Sklenář, 2000). Orchidaceae, especially epiphytic ones,
and ‘‘shrubby’’ families, such as Melastomataceae,
Ericaceae, and Solanaceae rank among the most important páramo groups (Luteyn, 1999), but they are poor
in species or even absent in the superpáramo flora. In
contrast, several ‘‘herbaceous’’ families (e.g., Cyperaceae,
Brassicaceae, Caryophyllaceae, Valerianaceae) are
among the most species-rich groups in the superpáramo
flora whereas they are (relatively) less important when the
whole páramo belt is considered.
The most species-rich genera, usually with numerous
country-endemic species, are Lachemilla, Gentianella,
Valeriana, and Draba (Fig. 2), and they are diverse also
in the páramos of Colombia and Venezuela (Gaviria,
1997; Santana Castañeda, 1994; van der Hammen and
Cleef, 1986; Xena de Enrech, 1992). In Ecuador,
especially Gentianella and Draba are very conspicuous
superpáramo genera, contributing probably with the
highest number of endemic species (Jørgensen and
León-Y, 1999; León-Y, 2000; Pringle, 1995). However,
if Senecio were considered in a broad sense, including
Dorobaea and Pentacalia, its total number of species
would be 17 and the genus would then be the richest in
the superpáramo. At the other extreme, 57% of the
genera have only one species. Such unequal distribution
of species richness among genera was also found in the
páramo flora of the eastern cordillera in Colombia by
Simpson and Todzia (1990), who argued that presence
in the páramo of many species-rich genera reflected a
high degree of autochthonous speciation. The skewed
taxonomic composition towards a few speciose genera
documented here for the Ecuadorian (super)páramo
would agree with that scenario.
Variation in species richness among superpáramos
We expected to find a strong positive correlation
between the area of the surveyed superpáramo and its
species richness (Simpson, 1974), but instead we found
that the correlation was weak and not significant.
However, when the correlations were calculated for the
ecologically more homogeneous groups of mountains
separately, the expected positive correlation pattern did
appear, at least for the group of humid mountain sites
(Group 2 in Fig. 4). We also expected to find a positive
correlation between the vertical range of the superpáramo
and its species richness (Simpson, 1974), but instead we
found a negative correlation. These unexpected correlation
patterns reflect the great ecological variability among the
superpáramo sites. Small mountains with only species-rich
lower superpáramo, such as Cajas, are more diverse than
ARTICLE IN PRESS
426
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
some of the bigger mountains with an overall impoverished flora, such as Cotopaxi, Cotacachi, or Chimborazowest, despite their larger superpáramo area and/or larger
vertical range. Apparently, the exceptional superpáramos
of Cotopaxi and Chimborazo-west influenced the overall
correlation pattern of species richness with area and also
produced the strong negative albeit not significant
correlation among the five dry sites (Group 1 in Fig. 4B).
Variation in b-diversity among superpáramos
We expected, and found a positive correlation
between b-diversity and the vertical range of the
superpáramo (measured as number of surveyed altitudinal levels), i.e., the longer the gradient of the superpáramo, the greater its b-diversity. As previously
demonstrated, species composition and vegetation
structure change considerably from the lower to the
upper superpáramo belts (Cleef, 1981; Sklenář, 2000;
Sklenář and Ramsay, 2001), which would give greater bdiversity on higher mountains (e.g., Cayambe and
Iliniza mts.). The Cotopaxi superpáramo is among the
most b-diverse in our sample despite the fact that it is
among the least species rich (Fig. 1). This paradox may
be related to the occurrence of a pioneer vegetation in
the Cotopaxi superpáramo in addition to grass páramo
as documented by Stern and Guerrero (1997) and
Sklenář (2000). Lower mountains, where only lower
superpáramo is developed (e.g., Yanaurcu, Cajas mts.),
have lower b-diversity values, even though the overall
species richness may be high.
Among the highest Ecuadorian mountains (Antisana,
Cayambe, Cotopaxi, Iliniza, and Chimborazo) and
consequently those with the largest superpáramos
(Table 1), the Chimborazo-west has rather low bdiversity. This superpáramo is among the least diverse
in Ecuador, but it is conspicuous because of its
vegetation structure and presence of several central/
southern Andean floristic elements (e.g., Werneria
apiculata, Stipa spp., Viola polycephala) and due to this
it is sometimes compared to ‘‘puna’’ which is the dry
páramo-equivalent in Peru (Acosta-Solı́s, 1985).
Through the combined effects of climate (rain-shadow
of the mountain) and human activities (burning and
grazing), this area has been altered into patchy, desertlike superpáramo (Sklenář and Lægaard, 2003) with
reduced species richness and b-diversity. Smaller mountains seem to exhibit consistently lower b-diversity, with
the exception of Chiles. Plant communities on Chiles in
general have lower a-diversity (less species per samples)
than communities at other sites of similar size but there
is a high species turnover along the altitudinal gradient
in the rocky habitats (Sklenář, 2001a) which produce
high b-diversity at this particular site.
Species distributions
The superpáramo flora is distinct from the grass
páramo and subpáramo belts below (Cleef, 1981;
Cuatrecasas, 1968; Ramsay and Oxley, 1997). Species
endemism is generally high in the superpáramo (Berg,
1998; Luteyn, 1992) although it may be even higher in
the montane forest (Balslev, 1988; Sklenář and Jørgensen, 1999; but cf. Young et al., 2002). The proportion of
species endemic to Ecuador in our sample was estimated
at 23%, which is consistent with previous reports for the
Ecuadorian páramo (León-Y, 2000; Sklenář and Jørgensen, 1999). Two genera, Floscaldasia and Raouliopsis
(Asteraceae), seem to be restricted to superpáramo
vegetation, but only the former reaches to Ecuador
(Cuatrecasas, 1979; Luteyn, 1999; Sklenář and Robinson, 2000; Van der Hammen and Cleef, 1986).
Almost one-third (29%) of the Ecuadorian superpáramo species are restricted to a single mountain and
almost one-half of them (47%) are distributed in less
than four superpáramo sites. This is consistent with
findings of Ramsay (1992) and Sklenář and Jørgensen
(1999), who reported a high number of narrowly
distributed species in the Ecuadorian páramo. The
number of very narrowly distributed superpáramo
species may be partly inflated by the occasional
occurrence of grass páramo and bamboo páramo
species in the superpáramo belt at one or a few
mountains. In addition some species reach Ecuador
only marginally. That is true for Valeriana henrici and
Lysipomia multiflora that are found in the Cajas region
but are otherwise distributed in Peru (Ayers, 1999;
Eriksen, 1989). Nevertheless, many Ecuadorian superpáramo species have a truly limited geographical
distribution, e.g., Gentianella sulphurea (Fig. 5W) and
Loricaria antisanensis (Fig. 5X), and some are restricted
to a single mountain. Other examples of such stenoendemic species are Senecio ferrugineus at Chimborazoeast (Sklenář, 2001b), Aphanactis antisanensis from
Antisana (Robinson, 1997), Oritrophium llanganatense
from Hermoso (Sklenář and Robinson, 2000), and
several species from the Cajas region, e.g., Gentianella
longibarbata, G. hirculus (Pringle, 1995), Valeriana
secunda (Eriksen, 1989), and Draba steyermarkii.
Mountain ridges and deep valleys acted as major
barriers to migrations of high-Andean plants during
glacial and interglacial periods (Jørgensen and Ulloa,
1994; Jørgensen et al., 1995). Many páramo species are
wind-dispersed or epizoochorous (Frantzen and Bouman, 1989; Melcher et al., 2000, 2004; Simpson and
Todzia, 1990), which would make them capable of
dispersal over long distances. Superpáramo species do
tend to occur over broad vertical ranges which generally
correlate with the species’ geographic distribution
(Sklenář and Jørgensen, 1999). In our data, about
60% of the species have a trans-Andean distribution,
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
i.e., they occur in both cordilleras and are distributed
across the deep inter-Andean valley. This is somewhat
less than 72% of trans-Andean species previously
reported for the highlands of Ecuador (Balslev, 1988),
but our number is certainly affected by the singlemountain species. Trans-Andean distributions are found
in many species, e.g., Senecio nivalis, Nototriche spp.,
Pentacalia microdon (Fig. 5V), Saxifraga magellanica,
Arenaria dicranoides, and Xenophyllum rigidum (Fig.
5U), from the highest elevations, that should be the
most isolated from each other by the inter-Andean
valley. This may be a ‘‘relictual’’ occurrence of species
which were widespread during glacial periods or it may
have resulted from a post-glacial over-land dispersal of
propagules along and across the Andes. Nevertheless,
the pattern suggests that the distribution of (super)páramo plants is less affected by topographical barriers
than are the distributions of montane forest species
where only about half of species have trans-Andean
distribution (Balslev, 1988).
Although the species richness of separate superpáramo areas seems to be higher in Ecuador, onemountain endemics are much more common in the
eastern cordillera of Colombia. In the Sierra Nevada del
Cocuy, for instance, 20% of the 107 superpáramo
species are endemic (Van der Hammen and Cleef, 1986).
We found that only about 3% and 2% of the species
found at Cajas and Chimborazo, respectively, were
endemic, and most other mountains have no stenoendemic species at all. Unlike the metamorphic eastern
cordillera of Colombia, the major Ecuadorian superpáramos occur on volcanoes, several of which are still
active. Repeated destruction of vegetation through
volcanic events followed by recolonization of the new
substrate may not provide sufficient time between
successive eruptions for a speciation. It is interesting
to note in this context that the páramo flora of the
volcanic central cordillera of Colombia is poorer in
species than that of metamorphic eastern cordillera
(Salamanca, 1992; Van der Hammen and Cleef, 1986).
An additional factor may be the size of individual
superpáramos; possibly the larger area of the Sierra
Nevada del Cocuy compared to Ecuadorian páramo
areas (Hofstede et al., 2003) also contributed to the high
proportion of endemic species found there. An additional contributing factor could be the varying degree of
isolation of the areas today and in the past. While
downward shifts of vegetation belts by 1000–1500 m
during glacial periods (Simpson, 1974; Van der Hammen, 1974; Van der Hammen and Cleef, 1986) would
merge most areas of Ecuadorian superpáramos and
facilitate species migrations between the now separated
superpáramos, the Colombian Cocuy superpáramo area
would remain isolated from other areas of superpáramo
allowing longer time for speciation. Phylogeographic
studies of wide-spread species could test this hypothesis.
427
Due to its scattered occurrence along the Andes, the
páramo has often been treated as an insular system
(Simpson, 1974, 1975; Vuilleumier, 1970). In one
classical study the extent of the páramo during glacial
periods was found to be more important for explaining
the patterns of páramo plant species diversity than the
current extent of the páramo (Simpson, 1974). In
Ecuador, species richness is only weakly correlated to
the present size of the individual superpáramos. On the
other side the extent of superpáramo during glacial
periods is not meaningful given the fact that the
assumed downward movement of vegetation belts would
merge almost all Ecuadorian areas of superpáramo to
one large confluent area. Although geographic distance
at various scales is (negatively) correlated to floristic
similarity in the Andes (e.g., Luteyn, 2002; Ricardi et al.,
1997; Simpson and Todzia, 1990) it does not explain the
observed general pattern of floristic similarity in the
Ecuadorian superpáramos. This finding of poor explanatory power of geographic distance confirms previous
results obtained from a sample of the high-páramo flora
(Sklenář and Jørgensen, 1999).
The presence of volcanic versus metamorphic bedrock
is the only significant variable in the RDA, but it
accounts for only a small portion of the variability.
Nevertheless, in three of the sites (Hermoso, Saraurcu,
Quilimas), we thoroughly examined the boundary
between volcanic and metamorphic basements, and we
failed to discern any obvious changes in species
composition or vegetation structure when crossing the
boundary, although Vargas et al. (2000) reported
distinct changes in species composition for the Llanganatis region where Hermoso is located. The cluster and
indirect ordination analyses (Fig. 4) give a similar result;
the superpáramos of mountains built from metamorphic
rocks are not clustered together but are grouped with
superpáramos of nearby volcanoes.
The four superpáramo sites developed on metamorphic bedrocks all occur in the eastern cordillera
where they face the Amazon lowlands and therefore
receive the highest amounts of precipitation among the
studied superpáramo sites (Bendix and Lauer, 1992;
Hastenrath, 1981). We believe that the significance of
geology (volcanic versus metamorphic bedrocks) for the
species distributions (Fig. 5) in the RDA is an artifact
and that this variable acts as a substitute for climate,
especially humidity on the sites. Unfortunately, there are
no climatic data of the necessary spatial resolution
available for the highest elevations of the Ecuadorian
Andes for a rigorous test of this hypothesis. Some
support for it, however, comes from data concerning the
effect of a rain-shadow on species composition of two
Ecuadorian superpáramos, Antisana and Chimborazo
(Sklenář and Lægaard, 2003), that was based on
published precipitation measurements. Other circumstantial evidence for the effect of humidity on the
ARTICLE IN PRESS
428
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
distribution of superpáramo species comes from the
observation that some species that are mostly restricted
to the ‘‘humid’’ superpáramos from the eastern slopes
(e.g., Distichia muscoides, Caltha sagittata, Cuatrecasasiella isernii), can be found in wet sites along springs or
small depressions in the terrain that occasionally occur
in the ‘‘dry’’ superpáramos (Sklenář, 2000; Cerón, 1994).
Floristic similarity among superpáramos
Based on the distribution of species from the montane
forests and páramos, two floristic regions have been
proposed for the high Andes of northern and central
Ecuador, with the divide being the inter-Andean valley
(Jørgensen and Ulloa, 1994). Our data suggest two
somewhat different regions. Although the climatic data
is very limited, we attribute this pattern to differences in
humidity based on our field observations of vegetation
physiognomy. Most mountains of the eastern cordillera,
northern Ecuador (Chiles, Cotacachi, Imbabura), Chimborazo-east (in the western cordillera), and Cajas in the
south belong to humid superpáramos (Group 2 in Fig.
4B), which are characterized by the presence of
Jamesonia spp., Geranium sibbaldioides, Lachemilla
holosericea, F. asplundii, Carex sect. Acicularis, and
Luzula gigantea among others (Figs. 5G–L). The dry
superpáramos (Group 1 in Fig. 4B), are distinguished by
the occurrence of Festuca vaginalis, Plantago nubigena,
Astragalus geminiflorus, Bidens andicola, Conyza cardaminifolia, Calamagrostis mollis, Cerastium imbricatum,
and Silene thysanodes (Figs. 5A–F; although the latter
two were also found in some more humid superpáramos), and comprise Iliniza, Pichincha, Cotopaxi,
and the Chimborazo-west and Antisana-west.
Among the dry superpáramo sites, Chimborazo-west
stands out from the other sites (see above). Among the
humid superpáramo sites, two groups are apparent.
Sites from the eastern cordillera (except Cayambe),
Chiles and possibly also Cotacachi from the western
cordillera represent very humid superpáramos that
receive high amounts of precipitation originating in
the Amazon basin. Among them, Tungurahua is
exceptional, being an active volcano on which the
vegetation is constantly recovering from the influence
of volcanic events (Sklenář, 2000). These very humid
superpáramo sites are characterized by the occurrence
or abundance of species, such as Huperzia rufescens,
Nertera granadensis, Loricaria complanata, Calamagrostis guamanensis, C. ecuadoriensis, Draba spruceana, and
Xenophyllum sotarense (Figs. 5M–R). These species
often have a strictly cis-Andean distribution, but usually
do not occur on all of those mountains. These very
humid sites are also characterized by the absence of a
group of species that occur only in the dry or moderately
humid superpáramo sites, such as Perezia pungens and
Trisetum spicatum (Figs. 5S and T). The remaining four
mountains (Cayambe, Imbabura, Chimborazo-east, and
Cajas) basically differ from the very humid superpáramo
by the absence of the above species rather than by
forming a homogeneous group among themselves.
However, in agreement with previous studies that
pointed to the peculiarity of the Cajas region (e.g.,
Sklenář and Jørgensen, 1999) this site occurred well
separated from the other sites in the Correspondence
Analysis. Volcanic activity terminated during the
Tertiary in the southern Ecuadorian Andes, including
Cajas (Hall, 1977), whereas the northern páramos all
experienced volcanic activity up till present time. The
different volcanic histories may be reflected in the
second CA ordination axis.
Simpson (1983) pointed to the importance of the
ecological similarity between areas to explain the
composition of different Andean floras. At the geographic scale studied here of tens to hundreds of
kilometers, we suggest that ecological similarity, considered mainly as humidity, is more important for
distribution of superpáramo species than other factors,
such as geographic distance or topographic separation
by valleys. Our findings support previous works that
stressed the importance of precipitation for vegetation
composition by distinguishing humid and dry superpáramos in the Parque Los Nevados, Cordillera Central
of Colombia (Pérez and Van der Hammen, 1983), and
others who recognized rain-shadow desert páramo on
the western side of Chimborazo (Ramsay, 1992).
Several interesting questions remain for further exploration in the high páramo. Some of these are: How
can we explain the smaller number of one-mountain
endemic species in Ecuador compared to Colombia? At
what geographic scale do island biogeographical factors
become more important than ecological factors? Do grass
páramo and subpáramo species conform to the pattern as
observed in the superpáramo flora?
Acknowledgements
Petr Sklenář thanks the Danish Research Academy
and the Grant Agency of the Czech Republic (Grant
No. 206/97/1198) and Henrik Balslev thanks the Danish
Natural Science Research Council (Grant No. 21-010617) for support. Two anonymous reviewers are
thanked for suggestions to improve the quality of the
paper. Both authors acknowledge Renato Valencia,
director of QCA herbarium, P.U.C.E., Quito, for
research facilities, INEFAN, Quito, for research and
export permits. Veronika Kostečková is thanked for
assistance in the field and Flemming Nørgaard for
producing figures.
Appendix 1
Matrix of OBSERVED (above diagonal) and EXPECTED numbers of species held in common (below diagonal) between pairs of sites, diagonal indicates
the number of species for each site used in the analysis.
Cota
Imba
Caya
Sara
EAnti WAnti Pich
Ilin
Coto
WChim EChim Herm Tung Altar Quil
56
64
34
69
67
55
64
62
43
72
67
59
71
59
42
68
64
54
93
72
46
82
77
63
60
74
39
76
76
63
77
84
48
84
84
66
87
53
39
62
72
49
83
54
45
68
67
55
83
48
42
67
67
53
46
30
29
34
34
30
53
29
27
44
41
35
129
58
45
73
76
61
71.6*** 117
48
74
83
65
47.1
44.3 71
45
46
43
77
71.4 47.1 130 92
87
80
73.9* 48.5 80*** 135
80
66.1
61.9 41.5 66*** 68.3** 106
85
78.6** 51*** 84.7 87.7 72.4*
61
71
69
83
60
73
77
71
79
47
53
78
68
39
82
83
65
147
Species pool ¼ 374, significant differences marked with asterisks (0.05Xp40.01*, 0.01Xp40.001**, pp0.001***).
ARTICLE IN PRESS
Chiles
100
62
60
67
67
70
44
51
49
26
30
Cotacachi
53.8* 102
77
78
59
69
61
76
69
39
41
Imbabura
57.9
59.4*** 114
92
55
66
66
74
72
42
43
Cayambe
66
67.4** 73.4** 137
69
84
85
92
91
50
53
Saraurcu
49.6*** 50.8* 54.6
61.7* 93
77
45
49
47
22
26
Antisana-east
63.2* 64.5
70.1
80.6
59.4*** 129
67
66
62
35
33
Antisana-west
66.7*** 68.4* 74.4* 85.8
63.1*** 82.1*** 140
87
102
66
69
Pichincha
63.5*** 65.1** 70.7
81.1** 59.7** 77.8** 82.4
130
95
62
61
Iliniza
68.5*** 70.2
76.2
87.9
64.5*** 84.1*** 89.4** 84.4** 143
74
73
Cotopaxi
48.7*** 49.6*** 53.6** 60.7** 46*** 58.5*** 62
58.8
63.3** 91
54
Chimborazo-west 53.6*** 55*** 59.3*** 67.4*** 50.8*** 64.9*** 68.5
65.1
69.8
49.9
103
Chimborazo-east 63.3* 64.9
70.2
80.8** 59.6
77.3
81.9
77.7
84.1
58.6*** 64.6**
Hermoso
59.1
60.5
65.2
74.8
55.6*** 71.5*** 76*** 72.1*** 77.7*** 54.7*** 60.3***
Tungurahua
39.6* 40.6
43.7
49.1
37.6
47
49.8*** 47.5
50.6** 37.1** 40.6***
El Altar
63.2
64.7* 70.1
80.7
59.6*** 77.3
82.2** 77.7** 83.9*** 58.6*** 64.6***
Quilimas
65.2
66.9
72.5
83.5
61.3*** 79.9
84.8** 80.2*** 87.1*** 60.3*** 66.8***
Yanaurcu
54.9
56.2
60.6
69.2
51.8*** 66.3
70.2*** 66.5*** 71.9*** 50.8*** 56***
Cajas
69*
70.7
77*
88.8
64.9
84.8** 90.4** 85.6*** 92.8*** 63.6*** 70.9***
Yana Cajas
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
Chil
429
430
Appendix 2
Matrix of geographic distances between the sites in kilometers (above diagonal) and matrix of floristic dissimilarities measured by w2 distance (below
diagonal) which were used in the Mantel test.
Imba Caya Sara
E Anti WAnti Pich
Ilin
Coto
WChim EChim Herm Tung Altar Quil
Yana Cajas
0
1.25
0.076
0.015
6.104
0.732
7.73
2.46
5.55
10.6
10.4
0.84
.406
0.79
0.532
0.05
0.0002
0.93
64
0
5.215
1.667
1.324
0.314
0.8
1.825
0.02
2.27
3.56
0.01
0.037
0.142
0.824
0.0001
0.14
0.001
64
22
0
4.713
0.003
0.24
0.95
0.154
0.23
2.51
4.48
0.009
0.59
0.07
0.06
1
0.72
0.83
140
94
80
56
45
0
2.78
1.79
5.81
9.44
15.7
0.001
2.185
0.021
0.581
0.21
0.001
1.64
183
122
120
111
107
69
62
58
0
1.809
0.147
0.01
11.4
1.46
3.4
4.64
4.97
2.05
172
115
108
93
85
43
35
57
30
0
0.337
2.71
11.2
1.77
10.3
11.5
8.52
4.33
272
212
208
193
185
140
133
149
91
101
0
2.08
16.2
4.56
6.57
9.96
7.88
4.52
345
292
282
260
248
205
198
232
177
176
96
92
125
89
64
55
0
0.76
85
54
33
0
0.864
0.143
0.01
1.465
0.109
1.89
3.08
1.842
0.1
0.2
0.021
0.51
0.56
0.38
100
70
49
17
0
5.215
5.19
1.92
4.75
12.5
12.1
0.003
6.089
0.052
4.513
3.525
2.422
0.37
147
98
85
63
52
8
0
0.257
1.776
0.258
0.004
0.318
6.96
2.34
4.96
1.93
6.4
1.99
126
64
64
67
71
58
56
0
1.331
0.174
0.26
0.362
4.54
0.13
1.21
2.17
1.99
2.49
268
209
204
188
179
134
126
147
89
96
11
0
2.58
0.09
0.21
0.2
0.39
0.58
220
171
159
136
123
80
74
118
79
61
81
71
0
0.309
0.095
1.121
0.155
1.43
256
202
193
172
161
117
110
143
92
87
48
37
41
0
0.09
0.13
0.054
2.82
281
228
218
196
184
141
134
169
118
113
57
47
62
26
0
1.8
6.682
0.09
290
237
228
205
193
150
143
179
127
122
62
53
70
35
9
0
2.004
0.25
The positive or negative signs with the w2 values indicate respectively whether or not the observed number of shared species is higher or lower than the expected number.
434
375
370
354
344
299
292
312
254
263
163
167
225
185
165
158
112
0
ARTICLE IN PRESS
Cota
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
Chiles
Cotacachi
Imbabura
Cayambe
Saraurcu
Antisana-east
Antisana-west
Pichincha
Iliniza
Cotopaxi
Chimborazo-west
Chimborazo-east
Hermoso
Tungurahua
El Altar
Quilimas
Yanaurcu
Cajas
Chil
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
References
Acosta-Solı́s, M., 1985. El Arenal del Chimborazo, ejemplo de
puna en el Ecuador. Rev. Geográfica (Quito) 22, 115–122.
Almeida-Lenero, L., de Azcarate, J.G., Cleef, A.M., Trapaga,
A.G., 2004. Plant communities of the zacatonal alpino area
of the Popocatepetl and Nevado de Toluca volcanoes in
Central Mexico. Phytocoenologia 34, 91–132.
Ayers, T., 1999. Biogeography of Lysipomia (Campanulaceae),
a high elevation endemic: an illustration of species richness
at the Huancabamba Depression, Peru. Arnaldoa 6, 13–27.
Balslev, H., 1988. Distribution patterns of Ecuadorian plant
species. Taxon 37, 567–577.
Balslev, H., Luteyn, J.L., 1992. Páramo: an Andean Ecosystem
under Human Influence. Academic Press, London.
Bendix, J., Lauer, W., 1992. Die Niederschlagsjahreszeiten in
Ecuador und ihre klimadynamische Interpretation. Erdkunde 46, 118–134.
Berg, A., 1998. Pflanzengesellschaften und Lebensformen des
Superpáramo des Parque Nacional Sierra Nevada de
Mérida in Venezuela. Phytocoenologia 28, 157–203.
Cerón, C.E.M., 1994. Vegetación y diversidad en la Reserva de
Producción Faunı́stica del Chimborazo–Ecuador. Rev.
Geográfica (Quito) 33, 19–42.
Cleef, A.M., 1981. The vegetation of the páramos of the
Colombian Cordillera Oriental. Diss. Bot. 61, 1–320 J.
Cramer, Vaduz.
Cleef, A.M., Rangel, C.O.J., Salamanca, S.V., 1983. Reconocimiento de la vegetación de la parte alta del transecto
Parque Los Nevados. In: Van der Hammen, T., Pérez, P.A.,
Pinto, P.E. (Eds.), La Cordillera Central Colombiana.
Transecto Parque Los Nevados. Introducción y datos
iniciales, vol. 1. J. Cramer, Vaduz, pp. 150–173.
Connor, E.F., Simberloff, D., 1978. Species number and
compositional similarity of the Galápagos flora and
avifauna. Ecol. Monogr. 48, 219–248.
Cuatrecasas, J., 1958. Aspectos de la vegetación natural de
Colombia. Rev. Acad. Colomb. Ci. Exact. 10, 221–268.
Cuatrecasas, J., 1968. Páramo vegetation and its life forms. In:
Troll, C. (Ed.), Geo-ecology of the tropical mountainous
regions of the tropical Americas. Colloq. Geograph., vol. 9.
Bonn, pp. 163–186.
Cuatrecasas, J., 1979. Comparación fitogeográfica de páramos
entre varias cordilleras. In: Salgado-Labouriau, M.L. (Ed.),
El Medio Ambiente Páramo. Centro de Estudios Avanzados, Caracas, pp. 89–99.
Cuatrecasas, J., 1986. Speciation and radiation of the
Espeletiinae in the Andes. In: Vuilleumier, F., Monasterio,
M. (Eds.), High altitude Tropical Biogeography. Oxford
University Press, New York, pp. 267–303.
Eriksen, B., 1989. 186. Valerianaceae. In: Harling, G.,
Andersson, L. (Eds.), Flora of Ecuador, vol. 34. Department of Systematic Botany, University of Göteborg, and
the Section for Botany, Riksmuseum, Stockholm, pp. 1–60.
Frantzen, N.M.L.H.F., Bouman, F., 1989. Dispersal and
growth form pattern of some zonal páramo vegetation
types. Acta Bot. Neerl. 38, 449–465.
Gaviria, J., 1997. Sinopsis del género Lachemilla (Focke)
Rydberg (Rosaceae) para Venezuela. Plantula 1, 189–212.
431
Hall, M., 1977. El Volcanismo en el Ecuador. Biblioteca
Ecuador, Quito.
Hastenrath, S., 1981. The Glaciation of the Ecuadorian Andes.
Balkema, Rotterdam.
Hedberg, O., 1957. Afroalpine vascular plants. A taxonomic
revision. Symb. Bot. Upsal. 15, 1–411.
Hedberg, O., 1992. Afroalpine vegetation compared to
páramo: convergent adaptations and divergent differentiation. In: Balslev, H., Luteyn, J.L. (Eds.), Páramo: an
Andean Ecosystem under Human Influence. Academic
Press, London, pp. 15–29.
Hill, M.O., 1979. TWINSPAN: a FORTRAN Program for
Arranging Multivariate Data in an Ordered Two-Way
Table by Classification of the Individuals and Attributes.
Cornell University, Ithaca.
Hofstede, R., Segarra, P., Mena, P.V. (Eds.), 2003. Los
Páramos del Mundo. Proyecto Atlas Mundial de los
Páramos. Global Peatland Initiative/NC-IUCN/EcoCiencia, Quito.
Islebe, G.A., Cleef, A.M., 1995. Alpine plant communities of
Guatemala. Flora 190, 79–87.
Jørgensen, P.M., León-Y, S. (Eds.), 1999. Catalogue of the
vascular plants of Ecuador, Monogr. Syst. Bot. Missouri
Bot. Gard. 75, 1–1181.
Jørgensen, P.M., Ulloa, C.U., 1994. Seed plants of the high
Andes of Ecuador – a checklist. AAU Reports 34,
1–443.
Jørgensen, P.M., Ulloa, C.U., Madsen, J.E., Valencia, R.R.,
1995. A floristic analysis of the high Andes of Ecuador. In:
Churchill, S.P., Balslev, H., Forero, E., Luteyn, J.L. (Eds.),
Biodiversity and Conservation of Neotropical Montane
Forests. The New York Botanical Garden, Bronx,
pp. 221–237.
Legendre, P., Vaudor, A., 1991. The R Package: Multidimensional Analysis, Spatial Analysis. Département de
Sciences Biologiques, Université de Montréal.
León-Y, S., 2000. La flora de los páramos ecuatorianos. In:
Josse, C., Mena, P.A., Medina, G. (Eds.), La Biodiversidad
de los Páramos. Serie Páramo 7, GTP/Abya Yala, Quito,
pp. 5–21.
Luteyn, J.L., 1992. Páramos: why study them. In: Balslev, H.,
Luteyn, J.L. (Eds.), Páramo: an Andean Ecosystem Under
Human Influence. Academic Press, London, pp. 1–14.
Luteyn, J.L., 1999. Páramos: a checklist of plant diversity,
geographical distribution, and botanical literature. Mem.
New York Bot. Gard. 84, 1–278.
Luteyn, J.L., 2002. Diversity, adaptations, and endemism in
neotropical Ericaceae: biogeographical patterns in the
Vaccinieae. Bot. Rev. 68, 55–87.
Luteyn, J.L., Cleef, A.M., Rangel, C.J.O., 1992. Plant diversity
in páramo: towards a checklist of páramo plants and a
generic flora. In: Balslev, H., Luteyn, J.L. (Eds.), Páramo:
an Andean Ecosystem Under Human Influence. Academic
Press, London, pp. 71–84.
Melcher, I.M., Bouman, F., Cleef, A.M., 2000. Seed dispersal
in páramo plants: epizoochorous and hydrochorous taxa.
Plant Biol. 2, 40–52.
Melcher, I.M., Bouman, F., Cleef, A.M., 2004. Seed atlas of
the monocotyledonous genera of the páramo. Flora 199,
286–308.
ARTICLE IN PRESS
432
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
Monasterio, M., 1979. El páramo desertico en el altiandino de
Venezuela. In: Salgado-Labouriau, M.L. (Ed.), El Medio
Ambiente Páramo. Centro de Estudios Avanzados, Caracas, pp. 117–146.
Mueller-Dombois, D., Ellenberg, H., 1974. Aims and Methods
of Vegetation Ecology. Wiley, New York.
Pérez, P.A., Van der Hammen, T., 1983. Unidades ecogeográficas y ecosistemas en el Parque Natural Los
Nevados: una sintesis inicial. In: Van der Hammen, T.,
Pérez, P.A., Pinto, E.P. (Eds.), La Cordillera Central
Colombiana. Transecto Parque Los Nevados. Introducción
y datos iniciales, vol. 1. J. Cramer, Vaduz, pp. 277–300.
Pringle, J., 1995. 159A. Gentianaceae. In: Harling, G.,
Andersson, L. (Eds.), Flora of Ecuador, vol. 53. Department of Systematic Botany, University of Göteborg, and
the Section for Botany, Riksmuseum, Stockholm,
pp. 1–131.
Ramsay, P.M., 1992. The páramo vegetation of Ecuador: The
community ecology dynamics and productivity of tropical
grasslands in the Andes. Ph.D. thesis, University of Wales,
Bangor.
Ramsay, P.M., Oxley, E.R.B., 1997. The growth form
composition of plant communities in the Ecuadorian
páramos. Plant Ecol. 131, 173–192.
Rangel, C.O.J., Diaz, S.P., Jaramillo, R.M., Salamanca, S.V.,
1983. Lista del material herborizado en el transecto del
Parque Los Nevados (Pteridophyta–Spermatophyta). In:
Van der Hammen, T., Pérez, P.A., Pinto, P.E. (Eds.), La
Cordillera Central Colombiana. Transecto Parque Los
Nevados. Introducción y Datos Iniciales, vol. 1. J. Cramer,
Vaduz, pp. 174–205.
Ricardi, H.M.S., Gaviria, J., Estrada, J., 1997. La flora del
superpáramo venezolano y sus relaciones fitogeográficas a
lo largo de los Andes. Plantula 1, 171–187.
Robinson, H., 1997. New species of Aphanactis in Ecuador
and Bolivia and new combinations in Selloa (Heliantheae:
Asteraceae). Brittonia 49, 71–78.
Salamanca, 1992. La vegetación del páramo y su dinámica en
el macizo volcanico Ruiz-Tolima (Cordillera Central,
Colombia). Análisis Geográficos (Bogotá) 21, 1–155.
Santana Castañeda, E., 1994. Las especies de Draba de la
Cordillera Oriental, Colombiana. Trabajo de grado, Universidad Nacional de Colombia, Bogotá.
Simberloff, D., Heck, K.L., McCoy, E.D., Connor, E.F., 1981.
There have been no statistical tests of cladistic biogeographical hypotheses. In: Nelson, G., Rosen, D.E. (Eds.),
Vicariance Biogeography. A Critique. Columbia University
Press, New York, pp. 40–93.
Simpson, B.B., 1974. Glacial migrations of plants: island
biogeographic evidence. Science 185, 698–700.
Simpson, B.B., 1975. Pleistocene changes in the flora of the
high tropical Andes. Paleobiol 1, 273–294.
Simpson, B.B., 1983. An historical phytogeography of the high
Andean flora. Rev. Chilena Hist. Nat. 56, 109–122.
Simpson, B.B., Todzia, C.A., 1990. Patterns and processes in
the development of the high Andean flora. Am. J. Bot. 77,
1419–1432.
Sklenář, P., 2000. Vegetation ecology and phytogeography of
Ecuadorian superpáramos. Ph.D. thesis, Charles University, Prague.
Sklenář,, P., 2001a. Superpáramo flora and vegetation of
Volcán Chiles. In: Ramsay, P.M. (Ed.), The Ecology of
Volcán Chiles: High Altitude Ecosystems on the Ecuador–Colombia Border. Pebble & Shell, Plymouth, pp.
39–45.
Sklenář, P., 2001b. Notes on taxonomy, distribution and
ecology of Ecuadorian high páramo Asteraceae, tribe
Senecioneae. Compositae Newsletter 36, 1–8.
Sklenář, P., Jørgensen, P.M., 1999. Distribution patterns of
páramo plants in Ecuador. J. Biogeogr. 26, 681–691.
Sklenář, P., Lægaard, S., 2003. Rain-shadow in the high Andes
of Ecuador evidenced by páramo vegetation. Arctic
Antarctic Alpine Res. 35, 8–17.
Sklenář, P., Ramsay, P.M., 2001. Diversity of páramo
plant communities in Ecuador. Diversity Distribution 7,
113–124.
Sklenář, P., Robinson, H., 2000. Two new species of
Oritrophium
and
Floscaldasia
(Asteraceae,
tribe
Astereae) from the superpáramos of Ecuador. Novon 10,
144–148.
Smith, J.M.B., 1975. Notes on the distribution of herbaceous
angiosperm species in the mountains of New Guinea. J.
Biogeogr. 2, 87–101.
Smith, J.M.B., Cleef, A.M., 1988. Composition and origins of
the world’s tropicalpine floras. J. Biogeogr. 15, 631–645.
Stern, M.J., Guerrero, M.G., 1997. Sucesión primaria en el
Volcán Cotopaxi y sugerencias para el manejo de hábitats
frágiles dentro del Parque Nacional. In: Valencia, R.,
Balslev, H. (Eds.), Estudios Sobre Diversidad y Ecologı́a de
Plantas. Memorias del II Congreso Ecuatoriano de
Botánica, Pontificı́a Universidad Católica del Ecuador,
Quito. Centro de Publicaciones, P.U.C.E., Quito,
pp. 217–229.
Ter Braak, C.J.F., Šmilauer, P., 1998. CANOCO Reference
Manual and User’s Guide to Canoco for Windows.
Wageningen, Centre of Biometry.
Troll, C., 1959. Die tropische Gebirge. Bonner Geogr.
Abhandl. 25, 1–93.
Troll, C., 1968. The cordilleras of tropical Americas. Aspects
of climatic, phytogeographical and agrarian ecology. In:
Troll, C. (Ed.), Geo-ecology of the Tropical Mountainous
Regions of the Tropical Americas. Colloq. Geogr, vol. 9.
Bonn, pp. 15–56.
Valencia, R., Pitman, N., León-Y., S., Jørgensen, P.M. (Eds.),
2000. Libro Rojo de las Plantas Endémicas del Ecuador
2000. Herbario QCA, P.U.C.E., Quito.
Van der Hammen, T., 1974. The Pleistocene changes of
vegetation and climate in tropical South America.
J. Biogeogr. 1, 3–26.
Van der Hammen, T., Cleef, A.M., 1986. Development of the
high Andean páramo flora and vegetation. In: Vuilleumier,
F., Monasterio, M. (Eds.), High Altitude Tropical
Biogeography. Oxford University Press, New York, pp.
153–201.
Van der Hammen, T., Werner, J.H., van Dommelen, H., 1973.
Palynological record of the upheaval of the Northern
Andes: a study of the Pliocene and Lower Quaternary of
the Colombian Eastern Cordillera and the early evolution
of its high-Andean biota. Palaeogeogr. Palaeoclim. Palaeoecol. 16, 1–24.
ARTICLE IN PRESS
P. Sklenář, H. Balslev / Flora 200 (2005) 416–433
Vargas, H., Neill, D., Asanza, M., Freire-Fierro, A., Narvaez,
E., 2000. Vegetación y flora del Parque Nacional
Llanganates. In: Vazquez, M.A., Larrea, M., Suarez, L.
(Eds.), Biodiversidad en el Parque Nacional Llanganates.
EcoCiencia, Quito, pp. 13–66.
Vuilleumier, B.S., 1970. Insular biogeography in continental
regions. I. The northern Andes of South America. Am.
Naturalist 104, 373–388.
Vuilleumier, F., Monasterio, M., 1986. High Altitude Tropical
Biogeography. Oxford University Press, New York.
433
Whittaker, R.H., 1960. Vegetation of the Siskiyou Mountains,
Oregon and California. Ecol. Monogr. 30, 279–338.
Wilson, M.V., Shmida, A., 1984. Measuring beta diversity
with presence–absence data. J. Ecol. 72, 1055–1064.
Xena de Enrech, N., 1992. Valerianaceae. In: Morillo, G.
(Ed.), Flora de Venezuela, vol. 5(1) Acta Cientı́fica
Venezolana, Caracas. pp. 221–267.
Young, K.R., Ulloa, C.U., Luteyn, J.L., Knapp, S., 2002.
Plant evolution and endemism in Andean South America:
an introduction. Bot. Rev. 68, 4–21.