Functional Plant Traits and Species Assemblage in Pyrenean Snowbeds

Abstract

In mid-latitude mountains, snowbeds often consist of small, scattered alpine belt units that host many plants of high biogeographic interest. Because most snowbed species are weak competitors, it is important to study the dissemination and persistence of their seeds to better understand their population dynamics. This study analyzed the snowbed flora of the Central Pyrenees using 11 morpho-functional traits, mostly related to seed function. The seeds of most plants found in snowbeds are small or very small, they have ovoid to elliptical shapes, and have no attributes related to dispersal. When only snowbed specialists (i.e., with a phytosociological optimum in snowbed habitats) are considered, three strategy groups become apparent: i) annuals or pauciennials producing abundant small seeds prone to accumulate in the soil; ii) chamaephytes or hemicryptophytes that produce anemochorous seeds; and iii) other perennials – mainly hemicryptophytes – with no specific seed traits. In the first two groups, the extant populations are maintained either by permanent soil seed banks or by means of vegetative persistence and dispersal. The lack of specific traits in the third group suggests that these plants could be more sensitive to direct competitive exclusion from non-chionophilous species under a changing climatic scenario in which snowbeds tend to disappear.

Full text

Functional Plant Traits and Species Assemblage
in Pyrenean Snowbeds
Josep M. Ninot &Oriol Grau &Empar Carrillo &
Roser Guàrdia &Artur Lluent &Estela Illa
#Institute of Botany, Academy of Sciences of the Czech Republic 2012
Abstract In mid-latitude mountains, snowbeds often consist of small, scattered
alpine belt units that host many plants of high biogeographic interest. Because most
snowbed species are weak competitors, it is important to study the dissemination and
persistence of their seeds to better understand their population dynamics. This study
analyzed the snowbed flora of the Central Pyrenees using 11 morpho-functional
traits, mostly related to seed function. The seeds of most plants found in snowbeds
are small or very small, they have ovoid to elliptical shapes, and have no attributes
related to dispersal. When only snowbed specialists (i.e., with a phytosociological
optimum in snowbed habitats) are considered, three strategy groups become apparent:
i) annuals or pauciennials producing abundant small seeds prone to accumulate in the
soil; ii) chamaephytes or hemicryptophytes that produce anemochorous seeds; and iii)
other perennials –mainly hemicryptophytes –with no specific seed traits. In the first
two groups, the extant populations are maintained either by permanent soil seed
banks or by means of vegetative persistence and dispersal. The lack of specific traits
in the third group suggests that these plants could be more sensitive to direct
competitive exclusion from non-chionophilous species under a changing climatic
scenario in which snowbeds tend to disappear.
Keywords Alpine vegetation .Dispersal .Plant strategies .Salicetea herbaceae .Soil
seed bank
Plant nomenclature Bolòs et al. (2005)
Folia Geobot (2013) 48:23–38
DOI 10.1007/s12224-012-9138-9
Electronic supplementary material The online version of this article (doi:10.1007/s12224-012-9138-9)
contains supplementary material, which is available to authorized users.
J. M. Ninot (*):O. Grau :E. Carrillo :A. Lluent :E. Illa
Group of Geobotany and Vegetation Mapping, Department of Plant Biology, University of Barcelona,
Av. Diagonal 643, E-08028 Barcelona, Catalonia
e-mail: jninot@ub.edu
R. Guàrdia
Centre de Documentació de Biodiversitat Vegetal, University of Barcelona, C/ Baldiri Reixac 2,
E-08028 Barcelona, Catalonia
Author's
personal
copy

Introduction
High-altitude mountains, typically the alpine belt, are a suitable environment to
investigate the relationships between environmental constraints, species pools and
adaptive plant traits (Körner 2003: 2). Snowbeds are a particularly interesting habitat
in the alpine belt, because long-lasting snow cover adds specific constraints over the
general restrictions associated with altitude. Snowbeds are mainly characterized by a
very short growing period. Plants living there undergo marked physical disturbance
as a consequence of snow accumulation (on the plant carpet and soil) and intense soil
leaching during the long snowmelt period (Komárková 1993; Körner 2003: 47–62).
Notwithstanding, plants settling on snowbeds benefit from efficient thermal insula-
tion during winter and spring, and from an abundant water supply during the growing
season (Björk and Molau 2007; Lluent 2007).
Several plant communities in the snowbeds of alpine mountains have been de-
scribed in relation to the environmental constraints of these habitats (e.g., Braun-
Blanquet 1948; Onipchenko 2004; Choler 2005). Such communities are clearly
different in terms of species composition and their functional structure and dynamics
from those in neighbouring grasslands. Moreover, different snowbed communities
can be found within a short transect (just a few meters) from an alpine grassland to the
almost bare ground in the center of the snowbed, because the length of the growing
season causes a steep gradient that affects snowbed communities. The substrate type
(e.g., lime rich vs siliceous, soil texture) is also a determinant of the plant commu-
nities’composition.
The most common plants growing in snowbeds correspond to Arctic-Alpine taxa,
which occur in rather isolated localities across the Pyrenees, where most of them have
their southernmost populations in Europe (Bolòs et al. 2005). Most snowbed special-
ists are poor competitors (Kudo et al. 1999; Onipchenko and Semenova 2004; Schöb
et al. 2010). They may occur in small gaps in alpine grasslands with abundant snow
cover in winter, but they only develop optimally where long-lasting snow cover (i.e.,
snowbeds) hampers the establishment of grassland species. Also, a number of
grassland plants settle on snowbeds from neighbouring mother-plants, and form
scattered populations (Galen and Stanton 1995; Volkova et al. 2005).
The seed phase is important in a plant’s life cycle because colonizing new sites and
regenerating extant populations promotes gene flux. In relation to seed morphology,
production and functioning in the particular snowbed environment, we hypothesize
that seed dispersal across the landscape and persistence in the soil seed bank could
have shaped to some extent the composition of snowbed communities. According to
this, some snowbed species have been identified as opportunistic in other alpine
ecosystems because of their regenerative strategies (Scherff et al. 1994; Semenova
2004), mostly in relation to their seed traits that ensure long persistence in the soil
(Thompson 1993b; Cerabolini et al. 2003). However, other plant traits related to life
history could also be responsible for the particular plant assemblage occurring in
snowbeds.
Our aim is to assess the extentto which the species pool found in Pyrenean snowbeds
supports this hypothesis. Thus, we analyze these species in terms of morpho-functional
traits, mostly related to the seed function in this particular environment; we also
highlight and discuss the main trends found in snowbed specialists –plants with their
24 J.M. Ninot et al.
Author's
personal
copy

optimum in snowbed communities –compared to the remaining taxa –plants more
frequent in pastures and rocky places, but also found in snowbeds. This comparison may
help to explain the present and future fate of both groups, in the context of present
interannual snow-cover variation and of future hypothetical reduction and disappear-
ance of most Pyrenean snowbeds.
Material and Methods
Study Area and Vegetation
The study was performed in the Aigüestortes i Estany de Sant Maurici National Park,
in the Central Pyrenees, in Catalonia (NE Iberian Peninsula). The alpine landscape of
the area is highly representative of the Central Pyrenean high mountains. It consists of
varied plant formations covering from exposed rocky surfaces to gentle slopes and
valley bottoms with fens and lakes. The core of the study area is formed by a massive
granodiorite batholith, although a few calcareous and slate outcrops are also found
surrounding it.
Snowbed vegetation is widespread in the alpine belt, from (2,000)2,300 m a.s.l. to
2,700(2,900) m a.s.l., and includes most of the Pyrenean plant communities of the
class Salicetea herbaceae (Carrillo and Ninot 1992). The abrupt relief given by
granodiorite and lime leads to the formation of snowbeds mainly on footslopes of
cliffs and other glacial geoforms. As such, chionophilous vegetation is restricted to
small, irregular patches where concavities or gentle landforms favor some soil
formation. In the most developed examples, distinct plant communities make clear
graded catenas; these are modulated by substrate characteristics and by the length of
the snowmelt period, which varies from the edges to the center of the snowbeds.
On siliceous soils the vegetation gradient goes from bryophyte carpets in the center
of the snowbed, where snow lasts until mid-July and summer radiation is at its
minimum, to Gnaphalio-Sedetum and Salix herbacea formations at the edges, which
give way to grasslands with moderate snow cover and higher radiation (Trifolio-
Phleetum gerardii Br.-Bl. 1948, Selino-Festucetum eskiae Nègre 1969). In calcareous
areas, snowbeds are usually found at lower altitudes in the alpine belt, coinciding
with particular, rocky north-facing slopes. In most cases, chionophilous vegetation
consists of a dwarf-shrub community of Salix reticulata and S. retusa, which covers
skeletal soils with protruding bedrock. The calcicole Potentillo-Gnaphalietum hop-
peani and a moderately calcifugue Salix herbacea formation (Salici-Anthelietum
thalictretosum) are found only where fine-textured soil covers gentle slopes.
Outside the snowbed, this chionophilous vegetation contacts with the calcicole grass-
lands –Festuco-Trifolietum thalii Br.-Bl. 1948 in slight-snowpack zones, and
Oxytropido-Elynetum myosuroidis Chouard 1943 in early-melting areas. More pre-
cise information on the snowbed vegetation is given in Table S1 in Electronic
Supplementary Material and in Carrillo and Ninot (1992).
Given the small area covered by snowbeds in most cases and the marked changes
in topography and snow conditions of these surfaces, the whole set of snowbed plant
communities may occur within a few meters. In addition, snowbed communities
frequently include species more characteristic of neighbouring grasslands and of
Plant Traits in Pyrenean Snowbeds 25
Author's
personal
copy

rocky places. Also, these grasslands offer good conditions for some plants typical of
snowbeds (Gnaphalium supinum,Sagina saginoides,Sibbaldia procumbens), which
frequently settle in small gaps. Moreover, interannual variability of snow cover
causes small changes in the location and nature of the boundaries between plant
communities (Lluent et al. 2006; Lluent 2007).
Analyses
We investigated the Spermatophyta occurring in the snowbeds of the Aigüestortes i
Estany de Sant Maurici National Park, namely the species or subspecies found in at
least 5 % of 233 phytosociological relevés of Salicetea herbacea taken in the area, for
a particular monitoring program (reported in Lluent et al. 2006). Then, we added to
this list a few other taxa considered characteristic of this vegetation class that were
less frequent or were not sampled due to their local or regional rarity. This provided a
list of 81 taxa, including species or subspecies that were strictly chionophilous, or
typical of alpine habitats other than snowbeds, or even ubiquitous. The nomenclature
used follows that given in Bolòs et al. (2005).
We selected a set of seed traits corresponding to morpho-functional aspects, five
referring to seed characteristics (following the proposal of Thompson 1993a), and
two measuring seed production and size of the soil seed bank (Table 1). Seed is here
understood as the diaspore unit or dispersule, i.e., including the walls of indehiscent
fruits and the appendages (awns, pappi) or bracts attached to it during dispersal. Only
in the case of seed mass do we refer to the germinule (i.e., excluding bracts, pappi or
other attachments), because this trait is examined with the aim of evaluating the
chance of emerging and settling on a new site.
The seed traits were analyzed from a specific collection of seeds and fruits
sampled in the study area from 2006 to 2010, and kept in the Herbarium of the
University of Barcelona (BCN). Morphological aspects were observed under a
stereoscopic microscope at 10–40-fold magnification. The measurements corre-
spond to the mean of ten seeds, and seed mass was calculated from five
samples of at least 50 air-dried seeds each, weighed to the nearest mg.
Exceptionally, for the few taxa that were not well documented from this
collection, data were obtained or complemented from a few sources (particu-
larly Bojňanský and Fargašová 2007; Castroviejo et al. 1986–2009; and Liu et al.
2008), under the assumption that these morphological data are relatively constant
among populations of the same species living in different areas.
We evaluated the dispersal mode from seed morphology, assuming barochory for
the seeds with no apparent dispersal attributes, short-distance anemochory (or zoo-
chory) for seeds with awns or small wings (or small elaiosomes), and long-distance
anemochory for seeds with pappi. Seed production was estimated after multiplying
mean values of the number of seeds per fruit, and then fruits per infructescence, and
infructescences per ramet, which were assessed from a minimum of ten cases during
the specific sampling mentioned above.
The estimation of the seed-bank size was based on greenhouse germination
experiments of soil samples obtained from the study area. The data on seed germi-
nation were recorded from more than 100 samples of different snowbed communities,
and synthesized as the average densities for surface area from the samples where each
26 J.M. Ninot et al.
Author's
personal
copy

species occurred in the standing vegetation (methodology and results reported in
Lluent 2007). Only in the few cases in which we could not collect enough reliable
Table 1 Plant traits considered and data structure
Traits Data structure Units or categories
Seed length, width and breadth
(including appendages)
quantitative mm
Weight of germinule quantitative mg
Number of seeds per ramet quantitative –
Seed surface categorized 1, smooth
2, rugose, muricate
3, striate
4, hairy
Seed appendices categorized 1, absent
2, beak or hook
3, wing(s)
4, (hygroscopic) awn
5, elaiosome
6, persistent pappus
Seed morphology categorized 1, sphaerical
2, ovoid, rhomboidal
3, tigonous, triquetrous
4, lenticular, reniform, elliptical,
subulate
5, cylindrical, fusiform, ligulate
6, conical, clavate
Density of seeds in the soil bank categorized 1, <3 per m
2
2, from 3 to 29 per m
2
3, from 30 to 299 per m
2
4, >300 per m
2
Lateral expansion categorized none
few tillers, to short distance (<2.5 cm)
many tillers, to short distance
(dense turf)
few tillers, to long distance
(>2.5 cm)
Life form categorized therophyte
non-gaminoid hemicryptophyte
gramidoid hemicryptophyte
diffuse chamaephyte
pulvinular chamaephyte
(small cushions)
creeping chamaephyte
Plant Traits in Pyrenean Snowbeds 27
Author's
personal
copy

data from our experiments did we take density data from literature specific for
snowbeds or alpine vegetation (Cerabolini et al. 2003; Semenova 2004; Welling et
al. 2004; Marcante et al. 2009).
The seed analysis was complemented with the evaluation of two traits related to
plant performance, namely lateral expansion and life-form (Table 1). We sorted the
taxa into six life-forms and into four categories of lateral expansion, as stated in Illa et
al. (2006).
To define the distinct ecological groups and classify the plants analyzed, we
synthesized the information about species ecological preferences assessed in the
phytosociological studies by Carrillo and Ninot (1992) and in Bolòs et al. (2005)
and created three distinct categories, i.e., three main habitats, in relation to the
duration of snow cover: i) snowbeds (species characteristic of Salicetea herbaceae
and included syntaxa, or typical of grasslands of late-melting sites); ii) grass-
lands (taxa from sites with medium snow duration, mainly typical of the alpine
grassland classes Juncetea trifidi or Elyno-Seslerietea); and iii) rocky sites (plants
from rock crevices, screes or fellfields, mainly thriving on early-melting surfaces). In
the case of the first group, we distinguished the snowbed specialists by taking into
account the phytosociological information from the BDBC data bank (Font 2009);
from a pool of 1,600 relevés taken in the alpine belt of the Catalan Pyrenees, we
considered to be snowbed specialists those taxa with at least 30 % of their occurrence
in relevés of the class Salicetea herbaceae (i.e., more than 28 occurrences over 96
relevés). Although this threshold may seem a bit low, it allowed us to define a broader
ecological group, which included locally rare snowbed species and also weakly
chionophylous species. Among the non-specialist taxa, we defined the grassland
species by using the same threshold (at least 30 % of occurrence in relevés of
Caricetea curvulae or Elyno-Seslerietea). The inclusion of species into the third
group, i.e., ecological preference for rocky sites, was based only on literature sources
(Carrillo and Ninot 1992; Bolòs et al. 2005), given that most of them have a narrow
ecological niche (e.g., humid crevices, lime-rich scree) although making a rather
diverse group.
The data were organized in a rectangular table (81 species × 11 traits) to
summarize the general spectra of the species found in the Pyrenean snowbeds.
From this table we categorized the distribution of seed measurements or trait
categories in the species pool investigated, and drew comparative analyses
between the three ecological groups defined, based on species percentages of
each trait category.
A general assessment of the functional diversification of the seeds was
derived from a multivariate ordination of the species according to the traits
stated in Table 1. Data were analyzed in a Principal Component Analysis (PCA)
using R.2.8.0 (R Development Core Team 2011). In the qualitative traits the catego-
ries were arranged in order of their functional significance and scored accordingly as
qualitative traits (as ordered in Table 1). The quantitative data were transformed
logarithmically to normalize the distribution of the observations. Then, all variables
were standardized so as to facilitate the comparison of different scales and units. To
test the significance of the multivariate statistics, we ran a multivariate-ANOVA with
999 permutations (PERMANOVA test, Anderson 2001), using the ‘vegan’package
in R (Oksanen 2009).
28 J.M. Ninot et al.
Author's
personal
copy

Results
Trait Spectra in the Species Pool
Seeds of snowbed species are mostly very small: in 25 % of species the germinule
weighs less than 0.1 mg, and in 75 % less than 0.6 mg (Table 2; Fig. S1a in Electronic
Supplementary Material). The number of seeds produced per ramet is in general
moderate or low. For instance, in 36 % of species this value is lower than 40, and in
60 % it is lower than 80 (Fig. S1bin Electronic Supplementary Material). However,
because most of the plants considered are small and have ramets that occupy an area
of a few cm
2
, these seed yields are relatively high at population or community levels.
As for the soil seed bank, the obtained values of viable seed density also gave a
power distribution among species (more detailed data in Lluent 2007), with a few
species (Sagina saginoides,Murbeckiella pinnatifida,Cardamine bellidifolia) con-
centrating the vast majority of viable seeds. 57 % of the species were apparently
absent or almost absent in the soil seed bank, whereas the rest of the species were
evenly distributed among the other density categories established (Table 2).
68 % of the species had slightly longer than wide seeds, i.e., varying from ovoid to
elliptical (Fig. S1cin Electronic Supplementary Material), with a smooth surface
(78 %) and without appendages (58 %). However, 14 % of the species had pappi
(Asteraceae, Salicaceae) that are more or less efficient in wind dispersion; and 17 %
had awns –mostly short –or narrow wings, which may play some role in seed
dispersal. No species had clear adaptations to zoochory (i.e., fleshy fruits, hooked
dispersule, etc.), except for the small elaiosomes present in three Luzula species.
Of the life-forms considered, the most commonly observed in the snowbeds were
hemicryptophytes (72 %), most of which were non-graminoid (Fig. S1din Electronic
Supplementary Material), as is generally the case in alpine vegetation (Illa et al.
2006). Chamaephytes made up the second group (22 %), including three woody
creeping species of Salix. Therophytes were rare and geophytes were absent. In terms
of ecological preference, we included 23 species in the group of snowbed specialists,
and distributed the remaining species into the groups of grasslands (45 taxa) and
rocky sites (13 taxa), as described in the methods section.
Relationships between Seed Traits and Species
The first PCA based on the seed traits revealed strong correlation between width,
breadth, length and weight, and between these traits and the type of appendix. To
avoid an over-biased ordination, we performed a second analysis rejecting some of
the most correlated traits (seed breadth, mostly correlated with seed width, and seed
shape, also dependent on the three seed dimensions). Among the remaining traits,
correlation within pairs in absolute values ranged from almost zero (weight and
surface) to about 0.71 (length and width) as shown in Table 3.
The species ordination shown in Fig. 1ahighlights the trade-off between seed mass
and seed production, i.e., from taxa producing a few big seeds (in the upper central
part: e.g., Trifolium alpinum,Androsace carnea,Arenaria purpurascens or Galium
pyrenaicum) to those bearing large numbers of tiny seeds (in the lower right corner:
e.g., Gentiana nivalis,Saxifraga moschata,S. aizoides or Veronica alpina). Most of
Plant Traits in Pyrenean Snowbeds 29
Author's
personal
copy

Table 2 Characterization of the taxa analyzed according to nine traits (see Table 1) and to habitat
preference: Longitude and Width of the dispersule (mm), Weight of the germinule (mg), Number of seeds
per ramet, Surface type (1–4, from smooth to hairy), Appendices of the dispersule (1 –6, from none to
pappus), Seed Bank density in the soil (1–4, from <3 to >300 of seeds * m
–2
), Lateral Expansion (1 –none;
2–few tillers to short distance; 3 –many to short distance; 4 –few to long distance), Life Form (Th –
therophyte; Hng –non-graminoid hemicryptophyte; Hg –graminoid hemicryptophyte; Chd –difuse
chamaephyte; Chp –pulvinular chamaephyte; Chr –creeping chamaephyte), and Habitat preference
relating snow duration (Sb –snowbeds; Gld –grasslands, Rk –rocky places)
Long Width Weight Nmb Srf App S-B L-E L-F Hab
Agrostis alpina 6.4 0.5 0.070 30 3 4 1 3 Hg Gld
Agrostis rupestris 3.6 0.6 0.080 40 3 4 3 3 Hg Gld
Alchemilla fissa 1.5 0.9 0.400 18 1 1 1 4 Hng Sb
Alchemilla pentaphyllea 1.3 0.8 0.336 6 1 1 1 4 Hng Sb
Alopecurus alpinus 5.2 0.8 0.260 60 4 4 1 3 Hg Sb
Androsace carnea 2.5 1.5 1.532 10 2 1 1 2 Hng Gld
Antennaria carpatica 4.5 2.0 0.080 100 1 6 1 2 Hng Gld
Arenaria biflora 0.8 0.7 0.133 144 2 1 3 4 Chr Sb
Arenaria purpurascens 1.7 1.2 1.056 12 1 1 1 3 Chp Gld
Armeria alpina 5.5 1.2 0.928 15 3 4 1 3 Hng Gld
Astrantia minor 4.0 1.6 0.796 120 3 3 1 2 Hng Rk
Campanula scheuchzeri 0.8 0.4 0.043 75 1 1 1 2 Hng Gld
Cardamine bellidifolia subsp. alpina 1.3 1.0 0.198 40 1 1 4 2 Hng Sb
Cardamine resedifolia 1.3 1.0 0.200 184 1 3 1 2 Hng Rk
Carex atrata subsp. nigra 3.3 1.6 0.490 96 1 2 3 3 Hg Gld
Carex curvula subsp. curvula 4.8 1.6 1.304 20 1 2 3 3 Hg Gld
Carex pyrenaica 3.8 0.9 0.336 16 1 2 3 3 Hg Sb
Cerastium cerastoides 1.1 0.9 0.166 26 2 1 4 2 Chr Sb
Epilobium anagallidifolium 8.2 3.5 0.075 130 1 6 3 2 Hng Sb
Euphrasia minima 1.3 0.5 0.110 70 1 1 1 1 Th Gld
Festuca eskia 7.4 1.4 1.292 140 3 4 1 3 Hg Gld
Festuca glacialis 5.0 0.8 0.510 18 3 4 1 3 Hg Rk
Festuca nigrescens 7.0 1.0 0.980 80 3 4 3 3 Hg Gld
Galium pyrenaicum 1.5 1.2 1.400 12 1 1 1 3 Chp Rk
Gentiana alpina 0.9 0.3 0.200 50 2 1 1 2 Hng Gld
Gentiana nivalis 0.9 0.5 0.078 1250 2 1 1 1 Th Gld
Gentiana verna 0.9 0.5 0.090 625 2 1 1 2 Hng Gld
Gnaphalium hoppeanum 4.9 2.5 0.090 96 4 6 2 2 Hng Sb
Gnaphalium supinum 5.0 2.7 0.107 96 4 6 4 2 Hng Sb
Helictotrichon sedenense 18.0 1.3 1.800 56 1 4 3 4 Hg Gld
Hieracium lactucella 4.8 1.7 0.093 25 3 6 1 2 Hng Gld
Kobresia myosuroides 2.5 1.0 0.690 14 1 1 3 3 Hg Gld
Leontodon pyrenaicus 9.0 1.0 0.768 30 1 6 2 2 Hg Gld
Leucanthemopsis alpina 3.0 1.0 0.400 30 3 3 1 2 Hng Gld
Linaria alpina 1.8 1.6 0.160 140 1 1 1 2 Chr Rk
Lotus corniculatus subsp. alpinus 1.9 3.1 1.608 60 1 1 1 2 Hng Gld
30 J.M. Ninot et al.
Author's
personal
copy

Table 2 (continued)
Long Width Weight Nmb Srf App S-B L-E L-F Hab
Luzula alpinopilosa 1.2 0.6 0.210 108 1 5 1 2 Hg Sb
Luzula lutea 1.4 0.7 0.360 108 1 5 2 2 Hg Gld
Luzula spicata 1.1 0.7 0.356 57 1 5 2 2 Hg Gld
Minuartia sedoides 0.9 0.8 0.260 75 2 1 1 3 Chp Rk
Minuartia verna 1.2 1.0 0.312 85 2 1 1 2 Chp Gld
Mucizonia sedoides 0.5 0.2 0.030 40 1 1 3 1 Th Sb
Murbeckiella pinnatifida 1.1 0.5 0.122 280 1 3 4 2 Hng Rk
Myosotis alpestris 1.8 1.1 0.540 192 1 1 1 2 Hng Gld
Nardus stricta 10.4 0.9 0.380 15 3 4 3 3 Hg Gld
Oreochloa disticha subsp. blanka 3.9 1.3 0.340 11 4 1 1 3 Hg Gld
Oxytropis pyrenaica 2.2 1.7 2.000 48 1 1 3 2 Hng Gld
Pedicularis kerneri 2.2 1.0 0.202 120 3 1 1 2 Hng Gld
Phyteuma hemisphaericum 1.1 0.4 0.150 500 1 1 1 2 Hng Gld
Plantago alpina 2.1 0.9 0.544 75 1 1 2 2 Hng Gld
Plantago monosperma 3.3 1.6 2.990 15 2 1 3 2 Hng Gld
Poa alpina 3.2 0.8 0.350 100 3 1 3 3 Hg Gld
Polygonum viviparum 2.6 1.5 2.500 20 2 1 3 2 Hng Gld
Potentilla brauneana 1.1 0.9 0,320 12 1 1 1 2 Hng Sb
Primula elatior subsp. intricata 1.6 1.1 0.850 288 2 1 1 2 Hng Gld
Primula integrifolia 1.5 1.0 0.200 159 1 1 2 2 Hng Gld
Pritzelago alpina 1.7 0.7 0.236 30 3 3 4 2 Chd Rk
Ranunculus alpestris 2.2 0.9 0.408 40 1 2 4 2 Hng Sb
Ranunculus pyrenaeus 2.7 1.5 1.040 25 1 2 1 2 Hng Gld
Sagina saginoides 0.4 0.3 0.020 450 1 1 4 2 Chd Sb
Salix herbacea 4.0 3.0 0.510 20 1 6 1 2 Chr Sb
Salix reticulata 3.2 2.7 0.050 96 1 6 1 4 Chr Sb
Salix retusa 5.5 2.7 0.050 70 1 6 1 4 Chr Sb
Saxifraga aizoides 0.7 0.4 0.050 600 2 1 2 2 Chd Rk
Saxifraga androsacea 0.6 0.3 0.050 80 1 1 3 2 Hng Sb
Saxifraga moschata 0.6 0.3 0.030 432 2 1 3 3 Chp Rk
Saxifraga oppositifolia 1.1 0.5 0.110 258 3 1 2 4 Chr Rk
Sedum alpestre 0.7 0.3 0.030 225 1 1 3 2 Chd Sb
Sedum atratum 0.8 0.3 0.040 50 3 1 1 1 Th Rk
Selinum pyrenaeum 4.5 3.0 1.580 300 3 3 1 2 Hng Gld
Sibbaldia procumbens 1.2 0.9 0.470 48 1 1 3 2 Hng Sb
Silene acaulis 1.2 1.0 0.316 21 2 1 1 3 Chp Rk
Soldanella alpina 1.0 0.8 0.228 60 1 1 1 2 Hng Gld
Taraxacum alpinum 11.0 4.0 0.450 30 3 6 1 2 Hng Gld
Taraxacum dissectum 10.3 4.0 0.416 30 3 6 1 2 Hng Gld
Thalictrum alpinum 3.1 1.1 0.432 30 3 1 1 2 Hng Gld
Thymus nervosus 0.7 0.5 0.130 48 1 1 1 4 Chr Gld
Plant Traits in Pyrenean Snowbeds 31
Author's
personal
copy

the species studied are spread along the area between these two extremes, with those
with more rounded, unappendaged, smoother seeds located near the upper right edge. In
contrast, dispersules with more conspicuous appendages and higher dimensions appear
from the central part of the field to the lower left corner. This extreme consists of
Asteraceae, Salicaceae and analogous types (e.g., Gnaphalium supinum,G. hoppea-
num,Salix retusa,S. reticulata,Taraxacum alpinum or Epilobium anagallidifolium).
Species Groups
Life-forms show no obvious pattern in relation to seed traits. The most abundant type,
hemicryptophytes, occupies almost all the area defined by the PCA (Fig. 1a). Among
these, while non-graminoids are widespread, graminoids remain mostly reduced to
the central-left part of the cloud. Therophytes and short-lived chamaephytes tend to
be found at the lower right part, and pulvinules are mainly placed in the upper central
part.
The ecological groups also overlap in the area defined by the PCA, although
snowbed specialists are mostly concentrated in peripheral parts of the species cloud
(Fig. 1b). The PERMANOVA test gives statistical support (P-value 0.018) to separate
the two ecological groups shown in the figure (snowbed specialists vs non-
specialists). Besides, two sub-groups of snowbed specialists may be observed in
rather extreme positions (right and lower left), whereas the others are spread within
the main core (upper central part).
Table 2 (continued)
Long Width Weight Nmb Srf App S-B L-E L-F Hab
Trifolium alpinum 2.7 2.3 5.300 8 1 1 1 2 Hng Gld
Trifolium thalii 1.4 1.2 0.840 30 1 1 1 2 Hng Gld
Veronica alpina 0.9 0.6 0.040 170 1 1 4 2 Hng Sb
Veronica aphylla 1.1 0.9 0.090 63 1 1 1 2 Hng Sb
Table 3 Pearson’s product moment correlations between the variables measured: Numb –number of seeds
per ramet; Width & Long –dimensions of the dispersule; Weight –weight of the germinule; Append –
appendage type of the dispersule; Soil –density of seeds in the soil bank; Surf –surface type of the
dispersule. * –statistically significant at P<0.05
Numb Width Long Weight Append Soil
Width −0.3082*
Long −0.3442* 0.7071*
Weight −0.4961* 0.5247* 0.455*
Append −0.0497 0.4919* 0.6766* −0.0492
Soil 0.0918 −0.1812 −0.0689 −0.1503 −0.0484
Surf −0.0255 0.1342 0.4067* 0.0086 0.2699* −0.0861
32 J.M. Ninot et al.
Author's
personal
copy

The distinct distribution of traits in each ecological group is illustrated in Fig. 2,
which refers to three crucial aspects of plant life: vegetative expansion, soil seed
bank, and dispersal. The group of snowbed specialists is similar to the grassland
group in terms of general spectra, but is differentiated by the noticeable presence of
species forming dense soil seed banks, by fewer turf-forming species and by more
stoloniferous or long-rhizomatous plants. The snowbed group includes most of the long-
distance disseminators; short-distance anemochory is less frequent than in grasslands,
whereas barochory is more common, although clearly less so than in rocky places.
Fig. 1 Ordering of the taxa on the first two components (variance: 40.21 % and 20.15 %, respectively)
given by the PCA, with the continuous variables transformed to log
10
:aDirection and relative weight of the
traits in the analysis (abbreviated as in Table 3) indicated by the direction and size of the arrows, and life-
form of the taxa; bPosition of the snowbed specialists within the total pool considered, with the main
species (Table S1 in Electronic Supplementary Material) abbreviated as follows: Ab –Arenaria biflora;
Cb –Cardamine bellidifolia; Cp –Carex pyrenaica; Gh –Gnaphalium hoppeanum; Gs –Gnaphalium
supinum; Ms –Mucizonia sedoides; Pb –Potentilla brauneana; Ra –Ranunculus alpestris; Sb –Salix
herbacea; Sru –Salix retusa; Sri –Salix reticulata; Sp –Sibbaldia procumbens; Ss –Sagina saginoides
Plant Traits in Pyrenean Snowbeds 33
Author's
personal
copy

Discussion
The Effect of Snowbed Constraints on the Flora
Snowbeds are particular habitats within alpine landscapes, in which plants are limited
by various environmental constraints. Our results, together with other studies
(Cherednichenko 2004; Choler 2005; Illa et al. 2006) show that snowbeds harbor
less diversity of plant traits than do less restrictive habitats, and that they promote
greater frequency of specific types or traits, such as deciduous woody chamaephytes,
long stolons or rhizomes, or higher N content in leaves (Choler 2005). The snowbed
species pool studied here includes no zoochorous species (producing fleshy fruits, or
hooked or adherent seeds) nor geophytes. The seeds are mostly reduced to low-
weight values and to ovoid or elliptical shapes (Fig. S1 in Electronic Supplementary
Material). A similar distribution of mass values has been found in other alpine
assemblages (Zhang et al. 2004; Navarro et al. 2009), whereas in temperate floras
the modal seed weight is one order of magnitude higher (Cerabolini et al. 2003).
A few plant types acquire particular relevance in snowbeds, more in terms of
dominance than of species numbers. This is the case of deciduous chamaephytes (i.e.,
Salix spp.), and of therophytes and pauciennials, which produce large numbers of
persistent seeds (Mucizonia sedoides, Sagina saginoides). In fact, the density of seeds
found in the soil has been identified as a good surrogate for seed persistence in some
snowbed ecosystems (Cerabolini et al. 2003; Semenova 2004; Lluent 2007).
These weak tendencies become stronger if the 23 snowbed specialists are considered
as a unit (Fig. 2). This is partly related to the particular taxonomical composition of
this group, in which the richer families in the alpine flora (i.e., Asteraceae, Poaceae,
Brassicaceae, Caryophyllaceae and Fabaceae; Körner 2003: 14) are poorly repre-
sented (Table S2 in Electronic Supplementary Material). The rarity of Poaceae among
snowbed specialists may be related to seed traits, because most grasses produce
moderate numbers of seeds, which disseminate over a short distance and have low
persistence in soil. Although these limitations are of minor importance in grasslands,
they seem to become disadvantageous in snowbeds, where most species produce
Fig. 2 Species percentages for three crucial traits in the three ecological groups considered (Sb –Snowbed
specialists; Gld –Grasslands; Rk –Rocky places). The categories in each trait are indicated as growing grey
intensities, in the following order: i) Lateral expansion: none, few tillers to short distance, dense turf, few
tillers to long distance; ii) Dispersal mode: barochory, anemochory to short distance, anemochory to long
distance; and iii) Density in the soil seed bank: from absent or very low to very high
34 J.M. Ninot et al.
Author's
personal
copy

seeds that are persistent in the soil bank or are easily air-dispersed (Onipchenko and
Semenova 2004). The relative decline of the other families mentioned above seems to
be related to other biological features, because their seeds show no clear trends in
relation to snowbed function. In fact, the relative increase of other taxa in the group of
snowbed specialists (Salix, Crassulaceae) could explain the increase of co-occurring
plant traits derived from their phylogenetic signal, which do not necessarily enhance
their adaptation to snowbeds. Thus, the ecological filtering of plant traits in snowbeds
seems to have shaped the chionophilous assemblage at the family level, because seed
traits are rather dependent on these taxonomic entities (Šerá and Šerý 2004).
Diversification into Snowbed Microhabitats
It should be taken into account that snowbed specialists thrive in rather different
microhabitats. Not only the duration of the period with snow cover, but also substrate
properties affect them. Thus, chionophilous plants may be included in a few con-
trasting types and strategies, some of which involve specific seed traits. The morpho-
functional analyses presented in this study (Fig. 1) reveal three main strategy groups:
i) annuals or short-lived perennials producing relatively small and abundant seeds,
mostly without any dispersal aptitudes, which tend to accumulate in the soil bank up
to high densities; ii) woody chamaephytes (ex. Salix spp.) or hemicryptophytes
producing medium-to-high numbers of anemochorous seeds, thus ensuring vegetative
persistence and wide wind dispersion; and iii) other perennials with no specific seed
traits (medium production, low or no seed bank, barochory). In the first group, the
maintenance of populations is based on the occupancy of patches with lower com-
petition, which may generally develop in the late-melting parts of snowbeds or in
small gaps of other snowbed parts or of neighbouring pastures (Semenova 2004;
Welling et al. 2004; Lluent et al. 2006; Schöb et al. 2010). Favourable years would
lead to a strong increase in the soil seed bank of these species in the areas where they
normally or exceptionally thrive. In the second group, population maintenance is
more ensured through perennial habit, mostly in the case of woody chamaephytes,
and through annual seed rain facilitating occasional recolonization of favorable sites
(Choler 2005; Lluent et al. 2006). Because the climatic conditions affecting snowbeds
follow a very irregular pattern between years, species of the first two groups will
presumably persist better if the decrease in areas occupied by snowbeds in the
Pyrenees persists in the future (López-Moreno et al. 2009) under the expected warmer
climate at regional scale (Esteban et al. 2010). Comparatively, the third group seems
to face a more uncertain future; dispersal and persistence in soil are very limited
(Scherff et al. 1994; Welling et al. 2004) and, in the cases of poorer tillering,
vegetative performance may not ensure strong persistence.
Long-Term Persistence of Snowbed Plants
When estimating the nature of the soil seed bank in snowbeds, it should be taken into
account that the persistence of viable seeds in natural conditions is only partially
known. The few specific studies done on this subject (Molau and Larsson 2000;
Semenova 2004) and our data neglect the presence of viable seeds of taxa with
specific germination requirements, which cannot be reproduced in non-natural
Plant Traits in Pyrenean Snowbeds 35
personal
copy

conditions. Most of the species pool investigated fulfils the morphological template
proposed by Thompson (1993b; reviewed in Fenner and Thompson 2005: 82–86),
which predicts that small seeds (<1 mg) with a fairly isometric shape (variance among
dimensions <0.18) would persist easily in the soil (Fig. 3). This hypothesis has been
supported by case studies on distinct species pools. However, these morphological
traits do not ensure seed viability (Cerabolini et al. 2003). In our study, most of the
species considered, and most of the snowbed specialists, fall in the graphic area
(Fig. 3) that is characteristic of persistent seed banks. The only exceptions within
these specialists are Alopecurus alpinus and Carex pyrenaica. However, a few non-
chionophilous species with variance values higher than 0.18 were found in the soil
bank (i.e., Poa alpina,Helictotrichon sedenense,Nardus stricta,Kobresia myosur-
oides and Carex atrata subsp. nigra). All these exceptions belong to Poaceae and
Cyperaceae, whose bare fruits fulfil the shape conditions and may thus incorporate
into the soil after losing the attached bracts or utricles. Moreover, the coarse soils
found in snowbeds facilitate seed trapping, even for grasses and sedges, as experi-
mentally found in sandy alpine soils (Chambers et al. 1991; Marcante et al. 2009).
Several snowbed specialists combine medium-to-high seed production with medium-
to-high density of their soil seed bank, a trend also found in other snowbed ecosystems
(Semenova 2004). Even some species producing only moderate seed numbers form
large soil banks (Sibbaldia procumbens,Sagina saginoides,Mucizonia sedoides),
which may progressively accumulate through time (Semenova 2004; Lluent 2007).
It has been hypothesized that snowbed specialists will be negatively affected under
a scenario of rising temperatures because of the greater competitiveness of grassland
species, which are favored by longer growing seasons (Schöb et al. 2009). More
precisely, Björk and Molau (2007) envisage a stronger effect on chionophilous
grasses and sedges than on chionophilous chamaephytes or forbs. In our case, only
two species of these taxonomic groups are snowbed specialists, while others are
weakly chionophilous. The seeds of these graminoids cannot ensure persistence in the
Fig. 3 Ordering of the taxa according to seed weight and seed shape coefficient (variance between
dispersule length, width and breadth). The area hypothesized for species with persistent soil seed bank
(seeds slighter than 1 mg, and with shape coefficient lower than 0.18) includes most of the snowbed
specialists, and all the species well represented in the soil seed bank
36 J.M. Ninot et al.
personal
copy

soil or re-colonization from distant sites if they are displaced by more competitive
alpine graminoids. More generally, snowbed specialists belonging to the third strat-
egy group mentioned above seem to be prone to suffer more directly from compet-
itive exclusion, in a scenario of a warmer climate (Esteban et al. 2010) and fluctuating
or vanishing snowbeds. Their competitive disadvantage in respect to grassland
species may not then be compensated by seed traits that favor their success in finding
spatial or temporal ecological windows.
Concluding Remarks
As hypothesized, seed function seems to have played a noticeable role in the
ecological selection of the extant snowbed plant assemblage in the Pyrenees. The
group of snowbed specialists clearly shows some trait convergence in terms of seed
types compared to the whole alpine flora. However, it still includes a few contrasting
strategy sub-groups, whose persistence and regeneration in snowbeds depend on
specific seed morphology and function combined with other plant traits.
The ecological filtering occurring in snowbeds has positively or negatively select-
ed some taxonomic groups (families, genera) through the selection of some functional
plant traits inherent to phylogeny (e.g., fruit type and vegetative tillering in grasses, or
seed and habit in Salix).
The ongoing climate warming in the Pyrenees may reduce the chionophilous flora
at regional scale, and we predict that it will have a stronger effect on those species that
lack the specific morpho-functional traits associated with snowbeds.
References
Anderson M (2001) A new method for non-parametric multivariate analysis of variance. Austral Ecol
26:32–46
Björk RG, Molau U (2007) Ecology of alpine snowbeds and the impact of global change. Arctic Antarctic
Alpine Res 37:444–453
Bojňanský V, Fargašová A (2007) Atlas of seeds and fruits of central and east-European flora. The
Carpathian mountains region. Springer Verlag, Dordrecht
Bolòs O, Vigo J, Masalles, Ninot JM (2005) Flora manual dels Països Catalans. Pòrtic, Barcelona
Braun-Blanquet J (1948) La végétation alpine des Pyrenées Orientales. CSIC, Barcelona
Carrillo E, Ninot JM (1992) Flora i vegetació de les valls d’Espot i de Boí. Arxius Sec Cièn 99(1):1–474;
(2):1–350
Castroviejo S et al. (eds) (1986–2009) Flora iberica, vols. 1–8, 10, 13–15, 18, 21. CSIC, Madrid
Cerabolini B, Ceriani RM, Caccianiga M, De Andreis R, Raimond B (2003) Seed size, shape and
persistence in soil: a test on Italian flora from Alps to Mediterranean coasts. Seed Sci Res 13:75–85
Chambers JC, MacMahon JA, Haefner JH (1991) Seed entrapment in alpine ecosystems: effects of soil
particle size and diaspore morphology. Ecology 72:1668–1677
Cherednichenko OV (2004) Plant life forms. In Onipchenko VG (ed) Alpine ecosystems in the northwest
Caucasus. Geobotany 29, Kluwer Academic Publishers: Dordrecht, pp 77–91
Choler P (2005) Consistent shifts in alpine plant traits along a mesotopographical gradient. Arctic Antarctic
Alpine Res 39:34–43
Esteban P, Prohom MJ, Aguilar E, Mestre O (2010) Evolució recent de la temperatura i la precipitació a
Andorra (1934–2008): resultats anuals i estacionals. La revista del CENMA (Andorra) 5:22–33
Fenner M, Thompson K (2005) The ecology of seeds. Cambridge University Press, Cambridge
Font X (2009) Mòdul Flora, Banc de Dades de Biodiversitat de Catalunya (BDBC). Available at: http://
biodiver.bio.ub.es/biocat/homepage.html
Plant Traits in Pyrenean Snowbeds 37
Author's
personal
copy

Galen C, Stanton ML (1995) Responses of snowbed plant species to changes in growing-season length.
Ecology 76:1546–1557
Illa E, Carrillo E, Ninot JM (2006) Patterns of plant traits in Pyrenean alpine vegetation. Flora 201:528–
546
Komárková V (1993) Vegetation type hierarchies and landform disturbance in Arctic Alaska and alpine
Colorado with emphasis in snowpatches. Vegetatio 106:109–120
Körner C (2003) Alpine plant life. Functional plant ecology of high mountain ecosystems. Springer Verlag,
Berlin
Kudo G, Nordenhäll U, Molau U (1999) Effects of snowmelt timing on leaf traits, leaf production, and
shoot growth of alpine plants: comparisons along a snowmelt gradient in northern Sweden. Ecoscience
6:439–450
Liu K, Eastwood RJ, Flynn S, Turner RM, Stuppy WH (2008) Seed Information Database (release 7.1,
May 2008). Available at: http://www.kew.org/data/sid
Lluent A (2007) Estudi de l’estructura i funcionament de les comunitats quionòfiles als Pirineus en relació
a la variació dels factors ambientals. PhD Thesis, University of Barcelona: Barcelona
Lluent A, Illa E, Carrillo E (2006) Inventario, cartografía y monitorización de la vegetación de los neveros
del Parque Nacional de Aigüestortes i Estany de Sant Maurici. Bull Soc Hist Nat Toulouse 141:131–
137
López-Moreno JI, Goyette S, Beniston M (2009) Impact of climate change on snowpack in the Pyrenees:
Horizontal spatial variability and vertical gradients. J Hydrol 374:384–396
Marcante S, Schwienbacher E, Erschbamer B (2009) Genesis of a soil seed bank on a primary succession in
the Central Alps (Ötztal, Austria). Flora 204:434–444
Molau U, Larsson EL (2000) Seed rain and seed bank along an alpine altitudinal gradient in Swedish
Lapland. Canad J Bot 78:728–747
Navarro T, El Oualidi J, Taleb MS, Pascual V, Cabezudo B (2009) Dispersal traits and dispersal patterns in
an oro-Mediterranean thorn cushion plant formation of the eastern High Atlas, Morocco. Flora
204:658–672
Oksanen J (2009) Vegan package: R functions for vegetation ecologists. R development core team.
Available at: http://vegan.r-forge.r-project.org/
Onipchenko VG (ed) (2004) Alpine ecosystems in the northwest Caucasus. Geobotany 29, Kluwer
Academic Publishers: Dordrecht
Onipchenko VG, Semenova GV (2004) Population strategies. In Onipchenko VG (ed) Alpine ecosystems in
the northwest Caucasus. Geobotany 29, Kluwer Academic Publishers, Dordrecht, pp 212–222
R Development Core Team (2011) R: a language and environment for statistical computing. R Foundation
for Statistical Computing, Vienna
Scherff EJ, Galen C, Stanton ML (1994) Seed dispersal, seedling survival and habitat affinity in a snowbed
plant: limits to the distribution of the snow buttercup, Ranunculus adoneus. Oikos 69:405–413
Schöb C, Kammer PM, Choler P, Veit H (2009) Small-scale plant species distribution in snowbeds and its
sensitivity to climate change. Pl Ecol 200:91–104
Schöb C, Kammer PM, Kikvidze Z, Choler P, von Felten S, Veit H (2010) Counterbalancing effects of
competition for resources and facilitation against grazing in alpine snowbed communities. Oikos
119:1571–1580
Semenova GV (2004) Soil seed banks. In Onipchenko VG (ed) Alpine ecosystems in the northwest
Caucasus. Geobotany 29, Kluwer Academic Publishers, Dordrecht, pp 194–212
Šerá B, Šerý M (2004) Number and weight of seeds and reproductive strategies of herbaceous plants. Folia
Geobot 39:27–40
Thompson K (1993a) Morphology and colour. In Hendry GAF, Grime JP (eds) Methods in comparative
ecology. A laboratory manual. Chapman & Hall, London, pp 194–196
Thompson K (1993b) Persistence in soil. In Hendry GAF, Grime JP (eds) Methods in comparative ecology.
A laboratory manual. Chapman & Hall, London, pp 199–202
Volkova EV, Onipchenko VG, Blinnikov MS (2005) Reciprocal transplantations in alpine plant commu-
nities. In Onipchenko VG (ed) Alpine ecosystems in the northwest Caucasus. Geobotany 29, Kluwer
Academic Publishers, Dordrecht, pp 223–236
Welling P, Tolvanen A, Laine K (2004) The alpine soil seed bank in relation to field seedlings and standing
vegetation in subarctic Finland. Arctic Antarctic Alpine Res 36:229–238
Zhang ST, Du GZ, Chen JK (2004) Seed size in relation to phylogeny, growth form and longevity in a
subalpine meadow on the east of the Tibetan plateau. Folia Geobot 39:129–142
Received: 17 March 2011 / Revised: 3 March 2012 / Accepted: 19 April 2012 /
Published online: 11 October 2012
38 J.M. Ninot et al.
Author's
personal
copy

1
Electronic Supplementary Material
Fig. S1 Characterisation of the species pool found in the snowbeds, in terms of number of species per
category in four traits: a Seed weight; b Number of seeds per ramet; c Seed morphology (sphaerical,
ovoid, trigonous, lenticular, cylindrical, conical); d Life-form (therophytes, non-graminoid
hemicryptophytes, graminoid hemicryptophytes, diffuse chamaephytes, pulvinular chamaephytes,
creeping chamaephytes)

2
Table S1 Main characteristics of the snowbed communities (class Salicetea herbaceae Br.-Bl. 1948)
found in the study area. Data synthesised from Braun-Blanquet 1948; Carrillo and Ninot 1992; and Lluent
et al. 2006
Plant communities Main species
(* dominant) Physiognomy Substrate Location in the
snowbed
Polytrichetum sexangularis
Br.-Bl. 1948 Polytrichum sexangulare* Bryophyte
carpet Siliceous;
Sand and gravel with a
thin humic layer
The innermost,
latest-melting zone
Gnaphalio supini-Sedetum
candollei Br.-Bl. 1948
(incl. Carici pyrenaicae-
Cardaminetum alpinae Rivas
Mart. et al. 1991)
Gnaphalium supinum
Cardamine bellidifolia ssp.
alpina
Mucizonia sedoides
Arenaria biflora
Carex pyrenaica
Patched
herbaceous
carpet
Siliceous;
Skeletal soils, rich in
gravel or sand
Late-melting zone
Salici-Anthelietum juratzkanae
Br.-Bl. 1948 (incl. subass.
thalictretosum alpini E.
Carrillo et Vigo 1992)
Salix herbacea*
Gnaphalium supinum
Sibbaldia procumbens
Sagina saginoides
Dense
herbaceous
carpet
Siliceous or lime
leached;
Dense, humus-rich,
fine soils
Medium-melting
zone
Potentillo dubiae-
Gnaphalietum hoppeani Br.-
Bl. 1948
Ranunculus alpestris
Potentilla brauneana
Gnaphalium hoppeanum
Patched
herbaceous
carpet
Calcareous;
Fine-textured, deep
soil
Late-melting zone
Carici parviflorae-Salicetum
retusae Rivas Mart. 1969 Salix reticulata*
Salix retusa* Dense dwarf-
shrub formation Calcareous;
Gravel-rich, irregular
soils on rocky slopes
Medium-melting
zone

3
Table S2 Main families found in the snowbeds, evaluated by the number of species in
the total pool, and by the number of specialists
Family total pool snowbed
specialists
Poaceae 10 1
Asteraceae 8 2
Caryophyllaceae 7 3
Scrophulariaceae 5 2
Brassicaceae 4 1
Cyperaceae 4 1
Fabaceae 4 0
Primulaceae 4 0
Rosaceae 4 4
Saxifragaceae 4 1
Crassulaceae 3 2
Gentianaceae 3 0
Juncaceae 3 1
Ranunculaceae 3 1
Salicaceae 3 3
Other 12 1
Total 81 23