Effect of Low-Intensity Grazing on the Species-Rich Vegetation of Traditionally Mown Subalpine Meadows

Abstract

Subalpine meadows, which traditionally were mown every other year, are particularly rich in plant species, especially forbs. Near Davos (Switzerland) we compared the vegetation of mown sites with that of sites grazed for up to 50 years by non-lactating cows. We recorded an overall mean of 51.5 plant species per 4×4 m plot. Among grazed sites, evenness decreased with time since conversion to grazing (−0.11 in 50 years; P<0.05), suggesting progressive vegetation change, which may eventually result in the loss of species. Ground cover by forbs tended to be higher in mown than in grazed sites (by 7.2%; P<0.1). The proportion of not-clonally colonising perennial species decreased after conversion to grazing (−7.72%; in 50 years, P<0.05), while the cover by graminoid species increased (+14.2% in 50 years; P<0.1). More intensively grazed sites had a lower cover of dwarf shrubs and higher cover of legume species (P<0.05). Because grazing negatively affects both botanical richness and agricultural quality, mowing of traditionally mown subalpine meadows should be maintained, and recently grazed meadows should be reconverted to mowing.

Full text

Effect of low-intensity grazing on the species-rich vegetation
of traditionally mown subalpine meadows
Markus Fischer*, Sonja Wipf
Institut fu
¨r Umweltwissenschaften, Universita
¨tZu
¨rich, Winterthurerstr. 190, CH-8057 Zu
¨rich, Switzerland
Received 23 June 2000; received in revised form 23 April 2001; accepted 8 May 2001
Abstract
Subalpine meadows, which traditionally were mown every other year, are particularly rich in plant species, especially forbs. Near
Davos (Switzerland) we compared the vegetation of mown sites with that of sites grazed for up to 50 years by non-lactating cows.
We recorded an overall mean of 51.5 plant species per 44 m plot. Among grazed sites, evenness decreased with time since con-
version to grazing (0.11 in 50 years; P<0.05), suggesting progressive vegetation change, which may eventually result in the loss of
species. Ground cover by forbs tended to be higher in mown than in grazed sites (by 7.2%; P<0.1). The proportion of not-clonally
colonising perennial species decreased after conversion to grazing (7.72%; in 50 years, P<0.05), while the cover by graminoid
species increased (+14.2% in 50 years; P<0.1). More intensively grazed sites had a lower cover of dwarf shrubs and higher cover of
legume species (P<0.05). Because grazing negatively affects both botanical richness and agricultural quality, mowing of tradition-
ally mown subalpine meadows should be maintained, and recently grazed meadows should be reconverted to mowing. #2002
Elsevier Science Ltd. All rights reserved.
Keywords: Biodiversity; Conservation biology; Ecological compensation measures; Low-intensity land use; Vegetation change
1. Introduction
In Europe, many species depend on semi-natural
agricultural landscapes (Landolt, 1991), which have
been created and maintained by long-term constant
low-intensity land use (Bignal and McCracken, 1996;
Ellenberg, 1996). Nutrient-poor grasslands of the mon-
tane and subalpine zones originally created by clear
cutting of forest and maintained by low-intensity mow-
ing or grazing exemplify such man-made ecosystems.
They constitute the vegetation type richest in plant spe-
cies in central Europe (Wolkinger and Plank, 1981) and
are very important habitats for many animal species
(Erhardt, 1995). For example, half of the European
butterfly species live in nutrient-poor grasslands
(Erhardt and Thomas, 1991).
Nutrient-poor grasslands in the upper subalpine zone
at altitudes between 1800 and 2300 masl, which are
mown every other year, are called ‘Ma
¨hder’. They have
been created in regions where hay production in the
valley and in the zone at altitudes between 1200 and
1700 m asl, which is inhabited only in summer, was not
sufficient to feed cattle through the winter (Niederer,
1996). This was the case in very high valleys and in
narrow lower ones with steep slopes. Ma
¨hder were
probably created with the first settlements in the early
middle ages (Niederer, 1996). Since then Ma
¨hder, which
are far away from farms and difficult to reach, have
remained unfertilised and used at very low intensity by
late mowing every other year. Their vegetation is char-
acterised by high species richness and a high proportion
of forbs. Eighty to 90 plant species per 1010 m have
been reported (Bischof, 1981, 1984). Because most of
the grasslands around the current tree line are used as
pastures (so-called Alpweiden), Ma
¨hder represent a
relatively rare habitat type ( <5% of the total area of
subalpine grasslands in 1980; Zoller and Bischof, 1980).
Recently, the area of traditionally used Ma
¨hder is
declining rapidly (Lenzin, 1995). For example, in the
region of Davos in Switzerland, it decreased by 70%
between 1945 and 1984 (Gu
¨nter, 1985). The reasons for
this decline are fertilisation and the intensification of
mowing, abandonment, or conversion to grazing,
mainly with non-lactating cows (Zumbu
¨hl, 1983).
0006-3207/02/$ - see front matter #2002 Elsevier Science Ltd. All rights reserved.
PII: S0006-3207(01)00149-5
Biological Conservation 104 (2002) 1–11
www.elsevier.com/locate/biocon
* Corresponding author. Tel.: +41-1-635-4805; fax: +41-1-635-
5711.
E-mail address: fischerm@uwinst.unizh.ch (M. Fischer).

The consequences of abandonment and fertilisation
for the floristic species composition and structure of
Ma
¨hder are quite well-known. Fertilisation increases
biomass production and decreases diversity (Zoller and
Bischof, 1980; Willems, 1985; DiTommaso and Aarssen,
1989) because grasses and legumes may increase, while
characteristic species, such as many forbs and orchids,
which are weak competititors, are excluded (Da
¨hler,
1993; Mountford et al., 1993). Such changes are pro-
found and can still be detected even decades after ferti-
lisation has ceased (Da
¨hler, 1993). Abandonment causes
slower vegetation changes than fertilisation, which
however, are also far-reaching (Schiefer, 1981; Schreiber
and Schiefer, 1985; Willems, 1985). Secondary succes-
sion in unused Ma
¨hder leads to dwarf shrub commu-
nities, and the number of (characteristic) plant species
declines (Bischof, 1984; Stampfli and Ha
¨felfinger, 1995).
The probability of soil erosion, land slides and ava-
lanches increases (Surber et al., 1973).
In Switzerland, a prerequisite for agronomic subsidies
for so-called integrated production is that 7% of the
area of a farm is used for ecological compensation. Both
unfertilised meadows and pastures are accepted as eco-
logical compensation areas. Traditional late mowing of
unfertilised Ma
¨hder parcels is subsidised, and subsidies
have also been suggested for low-intensity grazing,
because it supposedly does not affect the species com-
position of formerly mown nutrient-poor grasslands as
drastically as intensification or abandonment (Baur et
al., 1997; Lu
¨thy and Wiedemeier, 1998). However, the
effects of low-intensity grazing on subalpine Ma
¨hder
have not yet been studied.
After conversion to grazing, a spatially heterogeneous
structure with both nutrient-rich and nutrient-poor
patches, hoof prints, bushes and a corresponding species
composition may form (Van den Bos and Bakker, 1990;
WallisDeVries et al., 1998). With increasing hetero-
geneity, species diversity may also increase, e.g. because
the creation of vegetation gaps favours the establish-
ment of species of low abundance (Grubb, 1977). Cor-
respondingly, at low altitudes grazed nutrient-poor
grasslands were found to be richer in species than mown
ones (Schla
¨pfer et al., 1998). In contrast, however, sub-
alpine pastures which have never been mown have a
vegetation of lower species diversity than the Ma
¨hder
(Marschall and Dietl, 1974). This may be due to a
competitive advantage of graminoids with high regen-
eration ability after browsing and low sensibility to
trampling (Wilmanns, 1998) and of species avoided by
grazers (such as woody, thorny, or impalatable species;
Schla
¨pfer et al., 1998). These species may suppress weak
competitors such as orchids (Lu
¨thy and Wiedemeier,
1998).
We studied the effect of grazing on the vegetation of
formerly traditionally used Ma
¨hder in the upper sub-
alpine zone in the area of Davos. There, traditionally
>400-ha meadows in >30 regions were used as Ma
¨h-
der. We compared sites that had been grazed by non-
lactating cows for different periods of time and with
different intensity with sites that are still traditionally
mown. We asked the following specific questions: (1)
Are there differences in species richness, diversity, and
composition of the vegetation of mown and grazed
Ma
¨hder sites? (2) What is the effect of time since con-
version to grazing by non-lactating cows, and of the
intensity of grazing on these measures? We discuss the
consequences of our findings for agronomic policy and
nature protection.
2. Methods
2.1. Study area
Davos is in the district of Graubu
¨nden in Switzerland
at 1560 m asl, (481102400 N, 71405600 E). The climate is
relatively dry and warm: with a mean annual rainfall at
1500 m of 1007 mm, while nearby, at 2560 m (Weiss-
fluhjoch), it is 1163 mm. The mean annual temperature
at Davos is 2.6 C, and at Weissfluhjoch 3.2 C.
In this area, hay production in the valley is not suffi-
cient for the long period of winter feeding, which lasts
up to 260 days per year (Senn, 1952). Therefore, many
meadows in the upper subalpine zone above the current
tree line have traditionally been used as Ma
¨hder. This
gave us the opportunity to study the Ma
¨hder vegetation
on different types of bedrock, at different altitudes, and
at different aspects.
These Ma
¨hder range in size from two to a few hun-
dred hectares and are subdivided into parcels ranging
from 0.03 to 0.5 ha. Each parcel consists of two parts
which have traditionally been mown in alternate years.
Mowing and hay removal, formerly done manually by
scythes and rakes, is now done with scythe mowers and
hay loaders. Depending on the weather and the progress
of the hay harvest at lower altitudes, the mowing date
may vary between mid-July and early September. In
years with early mowing dates, some species in the mown
halves of parcels may be prevented from setting seed.
On the other hand, in exceptional years mowing may
even be omitted. Some former Ma
¨hder parcels are now
grazed annually from July until the end of September.
We interviewed the users about the management his-
tory of candidate areas. We excluded all areas that had
been fertilised, mown more than once every other year,
or had been fallow. We also excluded parcels for which
we had only contradictory or unreliable information.
Altogether, we found six Ma
¨hder areas with parcels that
were still traditionally used, seven Ma
¨hder areas with
parcels where traditional mowing had been replaced by
low-intensity grazing by non-lactating cows, and four
areas where grazed and mown parcels were directly
2M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11

adjacent to each other. (Table 1). In these areas grazing
and browsing by other species such as deer, chamois,
and marmots is negligible (S. Wipf, personal observa-
tion). In four Ma
¨hder areas, grazed and mown parcels
are directly adjacent to each other. Grazed sites in our
study are grazed at different intensities (Section 2.2).
Field work took place between 21 June and 26 August
1998. In each parcel, we randomly selected two plots of
44 m, but taking care that the two plots were situated
at distances of about 20 m from each other, this plot
size had been selected based on species-area curves for a
phytosociological description of the vegetation of mown
meadows around Davos, including three of our study
sites (Zumbu
¨hl, 1983). Two plots had to be excluded
from our analyses, because they turned out to have been
fallow for several years, or were unusually heavily
damaged by trampling because of a close fence.
2.2. Site characteristics
We classified the bedrock of the Ma
¨hder areas into
three classes as calcareous sediment (three parcels),
gneiss (14 parcels), or ‘Aroser Schuppenzone’ (i.e. a
small-scale mosaic of calcareous and acidic sediment
and crystalline rock, which consists of moraine material
from the ice-age Parsenn glacier; four parcels; Table 1;
Geologische Karte der Schweiz, Paul Haupt, Bern,
Switzerland). The type of land use was independent of
bedrock. For grazed parcels, we noted the time since
conversion to grazing, which ranged from 3 to 50 years
(mean 24.14.3 years). The users could not give very
detailed information about the intensity of grazing,
because the duration of grazing and the stocking rate
varied between years. Therefore, the information only
allowed to classify the intensity of grazing coarsely into
two categories, i.e. very low intensity (in all years low ani-
mal density for a short period of time, seven parcels) and
low intensity (higher density for a short period of time,
or low density for a long period of time, four parcels).
The time since conversion to grazing did not differ sig-
nificantly between parcels of different grazing intensity.
We recorded the aspect, altitude, and slope of each
plot. Slope (mean 23%) and aspect (50mean deviation
from south) did not differ significantly between mown
and grazed parcels. The average altitude of grazed (G)
parcels was 100 m lower than that of mown (M) parcels
(M: 2116 m34 m. G: 2016 m33 m). We classifed
plots into three classes of structural heterogeneity, i.e.
low (flat and completely covered by vegetation), inter-
mediate (few cow prints, no cow paths, proportion of
bare ground <5%), and high (both cow prints and
paths, proportion of bare ground >5%).
2.3. Vegetation records
We noted the presence of each species of vascular
plant and estimated its abundance on the Braun-
Blanquet scale (1964) converted by Zumbu
¨hl (1983).
Table 1
Location and features of 17 traditionally mown study sites at Davos
a
Site Latitude
(m)
Longitude
(m)
Years grazed by
non-lactating cows
Mean plot
altitude (masl)
Aspect Type of
bedrock
Number
of plots
Currently mown plots
Erber Ma
¨hder 779800 183250 – 1980 SW Gneis 1
Fideriser Heuberge 773600 192600 – 2160 E–SE Sediment 2
Peister Ma
¨hder 770700 192000 – 2140 SW–SE Sediment 2
Salezer Ma
¨hder 783000 188850 – 2090 E–SE Gneis 2
Weng 780250 173900 – 2310 SW Sediment 2
Witibergma
¨hder 784000 180900 – 2145 W Gneis 2
Currently mown plots and grazed plots
Parsennma
¨hder 783500 191500 – / 45 2030 S–SE AS 2+2
Sa
¨ltenu
¨ebma
¨hder 782000 179950 – / 25 2110 E Gneis 2+2
Stafler Bergma
¨hder 779300 183300 – / 10 2010 W–SW Gneis 2+2
Studenma
¨hder 781100 180650 – / 20 2030 W–SW Gneis 2+2
Currently grazed plots
Bu
¨elenbergma
¨hder 786000 185400 30 2130 SW Gneis 2
Fanezma
¨hder 780650 174850 3 2225 S–SW Gneis 2
Flu
¨elama
¨hder 787750 187350 30 2080 SE Gneis 2
Jatzma
¨hder (Dischma) 785000 183150 20 1820 E Gneis 2
Leidbachma
¨hder 780300 178250 20 1970 NW Gneis 2
Mittelalp 783600 191200 50 1835 S–SE AS 1
Steinenma
¨hder 780100 173400 12 2090 SW Sediment 2
a
Longitude and latitude are recorded as in Swiss topographical maps. The longitude and latitude of the site ‘Stafler Bergma
¨hder’ corresponds to
N46
4604200 and E 94701400. ‘Sediment’ denotes calcareous sediments, and AS denotes ‘Aroser Schuppenzone’ (Section 2).
M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11 3

Moreover, we estimated the relative cover percentages
for the five functional groups graminoids (Poaceae,
Cyperaceae, Juncaceae), woody plants, legumes, orch-
ids, and remaining forbs. We visited nine plots recorded
before 7 July 1998 again at the end of July to check for
species that flower late. Nomenclature of species follows
Flora Helvetica (Lauber and Wagner, 1998).
For each plot and parcel, we counted the number of
species S and calculated Shannon’s index of diversity
as H=P
i
ln P
i
(where P
i
is the relative abun-
dance of species i) and evenness Jas J=H/ln S. For
each plot, we counted the number of species per func-
tional group.
We classified species into three categories according to
their life form: short-lived (annuals and biennials), per-
ennials (without pronounced clonal growth) and clon-
ally colonising (species with pronounced clonal growth;
after Sto
¨cklin, 1992; Fischer and Sto
¨cklin, 1997; Ju
¨rg
Sto
¨cklin, University of Basel, personal communication).
We further noted whether species are listed in the Swiss
red data book (Landolt, 1991).
To assess productivity, we randomly selected a 0.3 x
0.3 m area within each plot (in the case of grazed plots
randomly in ungrazed patches) and harvested biomass 3
cm above ground in late July or mid-August. Three
plots were mown or heavily grazed by the time of sam-
pling and had to be omitted from biomass sampling.
Biomass samples were dried at room temperature and
sorted into woody and non-woody plants. Samples were
then dried for another 24 h at 70C and weighed to the
nearest 0.1 g. Mean biomass >3 cm above ground was
29127 g m
2
(without woody plants it was 221 17 g
m
2
). Productivity (biomass with or without woody
plants) was independent of sampling date (r
2
<0.03;
P>0.25) and of the land use parameters.
2.4. Statistical analysis
Because the parcel is the unit of land use in our study,
we analyzed average values of the two plots per parcel.
We investigated the effects of land use with hierarchical
ANOVA-models with sequential sums of squares. Fac-
tors in the model are current land use (mown versus
grazed), time since conversion to grazing, and intensity
of grazing. Because the two last factors only apply to
grazed parcels, they were nested within current land use.
Effects of all factors were tested against variation due to
remaining differences among parcels.
Multivariate species abundance data were analysed
after log-transformation with DCA (Detrended Corre-
spondence Analysis; Ter Braak, 1987) to identify gra-
dients in vegetation composition. To test whether
gradients were associated with land use parameters, we
analysed site scores along the four ordination gradients
with our ANOVA-model (indirect gradient analysis, in
contrast to direct gradient analysis; Ter Braak, 1986).
For multivariate analysis (ordination), we used the
program CANOCO 4.0 (Ter Braak and Smilauer,
1998). All other analyses were done with the statistical
package JMP 3.1 (SAS Institute, Cary, North Carolina,
USA).
3. Results
3.1. Relationships of land use with structure of the
Ma
¨hder
The soil surface of grazed parcels was more hetero-
geneous than the surface of mown parcels (Mown: eight
parcels with low and two with intermediate hetero-
geneity; Grazed: one with low, five with intermediate,
and five with high heterogeneity; 
2
2
=11.7; P<0.003).
Among grazed parcels, heterogeneity did not depend on
the intensity of grazing (
2
2
=2.36; P>0.30), and the
mean time since conversion to grazing did not differ
between plots of different heterogeneity (F
2, 8
=0.04; P
>0.95). The proportion of ground not covered by
vegetation differed between grazed and mown parcels
(Table 2; grazed: 5.68%, mown: 2.45%; P>0.05). In
parcels grazed at very low intensity cover was 3.93%,
while in parcels grazed at low intensity it was 8.75%
(P<0.05). In summary, grazed parcels were more het-
erogeneous and had a higher proportion of bare ground
than mown parcels.
3.2. Relationships of land use with diversity
In mown parcels, we recorded on average 50.5 3.08
species per 44 m plot (n=10) and in grazed parcels
52.52.94 species (n=11). The Shannon index of plots
in mown parcels was 2.78 0.066, while that in grazed
parcels was 2.830.063. Evenness in mown parcels was
0.7130.014, while that in grazed parcels was
0.7220.014. None of these three diversity measures
differed between mown and grazed parcels, nor did they
differ between more and less intensively grazed sites.
However, evenness (Table 2; Fig. 1a, P<0.05) and
Shannon index (Table 2; Fig. 1b, P<0.1) were lower
among grazed sites that had been converted to grazing
for longer periods.
When we analysed combined species lists of the two
44 m plots per parcel, we found the same patterns for
combined per-parcel diversity as reported above. Mown
parcels on average contained 68.8 3.55 species, and
grazed parcels 67.8 3.37 species. The Shannon index of
mown parcels was 3.05 0.070, and of grazed parcels
3.070.066. The evenness of mown parcels was
0.7200.014, and of grazed parcels 0.7310.013. These
differences were not significant. All three per-parcel
diversity measures were also independent of the inten-
sity of grazing although grazed parcels that had been
4M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11

converted to grazing earlier tended to have a lower
Shannon index and lower evenness than parcels that
were converted later (P<0.1 for both cases).
3.3. Relationship of land use with vegetation composition
Eigenvalues of the four gradients resulting from plot-
DCA (Detrended Correspondence Analysis) were 0.40,
0.22, 0.12, and 0.08. These gradients cumulatively
explained 12, 18.6, 22.1, and 24.6% of variation in spe-
cies data. Marginally significant variation in the first
two gradients was explained by land use (P<0.1) and
significant variation in the first gradient by the intensity
of grazing (P<0.05; Fig. 2). Significant variation in
DCA gradients was explained by bedrock, altitude,
slope, and aspect (Table 3). These results indicate that
land use is a determinant of species composition, but
that other environmental characteristics explain the
variation in species composition even more.
The proportions of the number of species per plot of
the different functional groups (graminoids, woody
plants, legumes, orchids, and other forbs) were not sig-
nificantly different between mown and grazed sites, and
they were independent of the time since conversion to
grazing (see Appendix for the effects of land use on
occurrence and abundance of those 98 (out of 203)
Table 2
The effect of land use on parameters describing Ma
¨hder vegetation in the study area near Davos
a
Parameter Mean Mown Grazed Very low-intensity
grazing
Low-intensity
grazing
Change in 50 years
of grazing
Percentage of bare ground 4.14 2.45 5.68 3.93 8.75 3.47
Number of species 51.5 50.5 52.5 52.8 51.9 2.14
Shannon index 2.81 2.78 2.83 2.85 2.81 0.417
Evenness 0.717 0.713 0.722 0.725 0.716 0.108
Number of species per parcel 68.3 68.8 67.8 68.6 66.0 6.48
Per-parcel Shannon index 3.06 3.05 3.07 3.05 3.12 0.448
Per-parcel Evenness 0.726 0.720 0.731 0.725 0.745 -0.085
Productivity (gm
2
) 291 261 318 354 235 20.8
Productivity (non-woody species, gm
2
) 220 219 222 217 233 3.73
Percentage of graminoid species 20.7 20.6 20.8 20.4 21.4 0.864
Percentage of woody species 7.73 7.20 8.21 9.80 5.43 2.49
Percentage of legume species 5.63 5.20 6.01 5.09 7.61 0.205
Percentage of orchid species 2.84 2.96 2.73 3.35 1.64 0.441
Percentage of other forb species 63.1 64.1 62.3 61.4 63.9 4.00
Percentage of graminoid cover 48.8 47.3 50.2 51.1 48.8 +14.2
Percentage of woody cover 11.3 10.3 12.2 16.7 4.38 5.64
Percentage of legume cover 5.41 6.05 4.82 3.86 6.50 0.726
Percentage of other forb cover 31.2 36.0 26.8 24.1 31.6 15.2
Percentage of clonally colonising species 16.3 15.0 17.5 18.3 16.0 +6.77
Percentage of perennial species 79.6 81.1 78.3 77.3 80.0 7.72
Percentage of high-turnover species 4.01 3.80 4.19 4.16 4.25 0.371
a
Significant differences between 10 still traditionally mown and 11 recently grazed sites, between seven more and four less intensively grazed sites,
and significant changes with time since conversion to grazing (P<0.05) are given in bold face, and in bold italics for marginal significance (P<0.01).
Table 3
Correlation coefficients of the relationships between mean parcel scores on four gradients detected with Detrended Correspondence Analysis of the
vegetation of traditionally mown subalpine meadows near Davos and environmental variables
a
Bedrock Altitude Slope Aspect
Gradient1 0.48 + 0.34* 0.50*** 0.46**
Gradient2 0.57* 0.63*** 0.37* 0.14
Gradient3 0.32 0.13 0.23 0.03
Gradient4 0.18 0.53*** 0.22 0.48**
a
Aspect was measured as deviation from south. Relationships with type of bedrock were analysed with ANOVA.
*
P<0.01.
**
P<0.05.
***
P<0.001.
M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11 5

individual species, that occurred in at least six parcels).
The proportion of legume species was significantly
lower and that of woody species was higher for less
intensively than for more intensively grazed parcels
(Table 2).
The proportion of ground covered by non-legume
forbs was higher in mown (36.0%) than in grazed plots
(26.8%; P<0.1, Table 2). Among grazed plots, the
proportion covered by graminoids tended to increase
with increasing time since conversion to grazing (by
14.2% in 50 years, P<0.1; Table 2, Fig. 1c). The cover
of woody plants was higher in less intensively grazed
parcels (16.7%) than in more intensively grazed parcels
(4.38%; P<0.05).
The proportions of the number of clonally colonising
species and of high turnover species did not differ sig-
nificantly between grazed and mown plots. However,
the proportion of perennial species tended to be higher
in mown than in grazed parcels (P<0.1; Table 2).
Moreover, the proportion of perennial species decreased
with the length of time since conversion (7.72% in 50
years; P<0.5, Table 2; Fig. 1d).
4. Discussion
4.1. Vegetation and conservation value of Ma
¨hder
The highest number of species per 44 m plot in our
study was 78 and the mean was 51.5, confirming that
Ma
¨hder in general are very rich in species, and that our
study sites are among the Ma
¨hder sites richest in plant
species (Zoller and Bischof, 1980; Bischof, 1981, 1984;
Hegg et al., 1993). In our study, graminoids made up
less than half of the vegetation cover. Compared with
Fig. 2. Detrended correspondence analysis of vegetation records of 10
mown parcels (filled circles), four grazed at low intensity (open
squares), and seven at very low intensity (open circles). Gradients are
given in arbitrary units.
Fig. 1. Relationship between vegetation diversity and composition and the time since conversion of traditionally mown subalpine meadows (Ma
¨h-
der) to low-intensity grazing by non-lactating cows. (a) Evenness; (b) Shannon index; (c) proportion of ground covered by graminoid species; and (d)
proportion of not clonally colonising perennial species.
6M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11

other grassland types used at low intensity, this pro-
portion of graminoids appears small. For example, in
the species-rich calcareous grasslands of the Swiss Jura
and in pre-alpine fens used at low intensity, graminoids
make up about two-thirds of total biomass (Thomet et
al., 1989; Pauli, 1998). Even though graminoids may
contribute somewhat more to biomass than to ground
cover, this large difference indicates that Ma
¨hder vege-
tation is especially rich in forbs. Fourteen species in our
study, including the 12 orchid species, are legally pro-
tected in Switzerland. Nine of them are listed in the
Swiss red list (Landolt, 1991) because of their attrac-
tiveness. Moreover, Nigritella rubra, which is listed as
endangered, grew next to several of our study plots.
Thus, the quality and worthiness of protection of Ma
¨h-
der vegetation is due to their high diversity, to their high
proportion of forbs, and to the attractiveness of their
species.
4.2. Effect of change to grazing on Ma
¨hder vegetation
Grazed and mown parcels in our study each covered
an altitudinal range of about 300 m, and grazed sites
were situated on average at altitudes 100 m lower than
mown sites. However, because altitude was not sig-
nificantly correlated with species richness, Shannon
index, or evenness among mown plots alone, or among
grazed plots alone (data not shown), we conclude that
altitudinal effects did not affect our conclusions on
effects of land use.
The most immediate effect of cows is the damage to
plants and soil by trampling and grazing. The effects of
footprints were reflected in the higher heterogeneity of
plots and the higher proportion of bare ground on
grazed versus mown sites. Similar effects have been
reported for other grassland habitats (Ellenberg, 1996).
In our study, the increased heterogeneity of grazed
compared with mown sites was not accompanied by a
significant grazing effect on species richness, Shannon
index or evenness, either at the 4 x 4m plot scale or at
the parcel scale. However, among grazed sites, evenness
and Shannon index decreased with increasing time since
conversion to grazing (Fig. 1), and this increased varia-
tion may, therefore, have decreased the likelihood of
detecting a significant effect of grazing itself.
Decreased evenness with increasing time since con-
version to grazing (Fig. 1a) indicates, firstly, that domi-
nant species became more dominant, while less
abundant species further decreased in abundance. In the
long run, i.e. >50 years, a continued decrease of even-
ness would result in the loss of less abundant species
and a decrease of species richness. This would be in line
with the observation that mown Nardetum sites are
more species-rich than sites which had always been
grazed (Marschall and Dietl, 1974). Secondly, decreas-
ing evenness corresponds well with the observation that
graminoid cover increased with time since conversion to
grazing (Fig. 1c). This suggests that continued grazing
will further reduce the relative abundance of non-gra-
minoids, which are of particular concern from a con-
servation point of view, in favour of graminoid species.
The lower forb cover of grazed sites also corresponds
with the visual impression that grazed sites appear more
uniformly green and less flower-rich than mown sites
(Zumbu
¨hl, 1983; S. Wipf, personal observation).
Thirdly, while the reduction in diversity with grazing
is in line with the reduced species diversity in grazed
compared with mown subalpine grasslands reported by
Marschall and Dietl (1974), it contrasts with reports
from grasslands at lower altitudes. There, higher species
diversity was reported for grazed than for mown nutri-
ent-poor grasslands (Schla
¨pfer et al., 1998). Generally,
alpine plants grow more slowly, have a higher degree of
vegetative reproduction, and live longer than plants at
lower altitudes (Ko
¨rner, 1999). This may imply that in
the Alps regeneration niches created by grazing may not
be as easily exploited by species of low abundance as at
lower altitudes, but will rather be used by dominant
species, possibly via vegetative reproduction. However,
comparisons of the effects of grazing on different types
of grassland vegetation are difficult, because they may
depend on climate and geology, on grassland history, on
the intensity of grazing, and on the type of grazer
(WallisDeVries et al., 1998).
The effects of land use on the composition of Ma
¨hder
vegetation were partly attributable to changes in the
proportions of functional groups. Forbs potentially
respond more sensitively to grazing than graminoids,
because they have a lower ability to regenerate (Wil-
manns, 1998). This could be especially pronounced in
the studied Ma
¨hder sites where grazing takes place
every year while traditional mowing only takes place
every other year. Indeed, in our study, the proportion of
non-clonal perennials decreased with the length of time
since conversion to grazing (Fig. 1d). The effects of
grazing on short-lived monocarpic plants could be det-
rimental if plants are trampled or eaten before seeds are
mature, which is quite possible because grazing takes
place at earlier dates in the year than mowing. On the
other hand, grazing animals may create favourable
microsites for germination (Grubb, 1977). The indiffer-
ence of short-lived forbs to land use observed in our
study may reflect a balance between these mechanisms.
The decreased occurrence of woody plants, i.e. of
dwarf shrubs mostly of the Ericaceae, in more inten-
sively grazed plots (Table 2) is most likely due to direct
damage by grazing cows (Schla
¨pfer et al., 1998). The
higher cover of dwarf shrubs in less intensively grazed
sites reflects the fact that these sites are more similar to
abandoned sites than to more intensively grazed sites.
Nardus stricta is avoided by grazers; it was more
abundant in grazed than mown plots, and strongly
M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11 7

increased with increasing time since conversion to graz-
ing (Appendix). These pronounced (but not statistically
significant) effects are in line with the idea that the
development of Nardus grasslands was greatly furthered
by grazing (Ellenberg, 1996). Higher abundance of
Anthoxanthum alpinum and Ligusticum mutellina in
mown than in grazed sites suggest a higher fodder
quality of the vegetation in these plots. Moreover,
among grazed plots, A. alpinum was more abundant in
parcels grazed at very low density, and L. mutellina
decreased in abundance with increasing time since
conversion.
4.3. Conservation implications
Grazing, and especially more intensive grazing, nega-
tively affected the quality of Ma
¨hder vegetation, both
from a conservation and from an agricultural point of
view. To conserve Ma
¨hder vegetation, the primary need
is to maintain or re-establish the traditional manage-
ment of mowing every other year. Attractive subsidies
should be made available to persuade farmers to adopt
this management, while grazing of Ma
¨hder vegetation
should not generally attract subsidies. However, of the
three alternatives, i.e. abandonment, fertilisation, and
extensive grazing, the last appears to be the least harm-
ful and is to be preferred when mowing is not possible.
However, even then, occasional cutting will be needed
to avoid the spread of dwarf shrubs and Nardus stricta.
Acknowledgements
We thank the land users in Davos, especially Ursula
Hofer and the Wegunterhaltsgenossenschaft Parsenn,
for very helpful information and the permission to work
on their land, Veronika Sto
¨ckli and Walter Ammann
for the use of the facilities of the Swiss Federal Institute
for Snow and Avalanche Research Davos, Georg Zum-
bu
¨hl for old vegetation records, Hannes Schlumpf for
maps, Ju
¨rg Sto
¨cklin for help with the classification of
species according to life forms and for helpful com-
ments, Bernhard Schmid for his advice, Penelope
Oertli-Barnett for her most helpful proof-reading, and
the editor and three anonymous referees for fruitful
comments.
8M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11

Appendix. The effect of land use on the abundance of 98 species recorded in at least six out of 21 study parcels in
subalpine ‘Ma
¨hder’ vegetation near Davos (a list of further 105 species recorded in fewer parcels is available from
the authors)
a
Species N AB ABM ABG ABVLI ABLI Change in 50 years
Anthoxanthum alpinum 21 1.97 3.82** 1.08 2.73*** 0.21 1.43
Lotus alpinus 20 0.44 0.33 0.58 0.64 0.48 1.71
Campanula scheuchzeri 20 0.33 0.53 0.21 0.14 0.40 1.19+
Potentilla aurea 20 0.54 0.97 0.31 0.34 0.26 1.09
Festuca rubra s.l. 20 2.24 1.56 3.11 5.42 1.17 5.42+
Nardus stricta 20 5.93 4.48 7.66 13.80 2.73 26.92
Carex sempervirens 20 6.38 10.32 4.12 9.30* 0.99 13.97
Homogyne alpina 19 0.28 0.42 0.19 0.22 0.15 1.16
Soldanella alpina 19 0.10 0.12 0.09 0.10 0.08 0.62+
Ligusticum mutellina 19 1.47 4.67* 0.51 0.36 0.95 20.88**
Helictotrichon versicolor 19 0.39 1.07+ 0.15 0.14 0.16 0.51
Crocus albiflorus 19 0.08 0.07 0.10 0.10 0.10 0.32
Arnica montana 19 0.86 1.27 0.60 1.67* 0.09 5.26
Gentiana acaulis 19 0.39 0.85+ 0.19 0.33 0.07 1.02
Hieracium hoppeanum 18 0.45 0.35 0.57 0.38 1.14 1.89
Geum montanum 18 0.24 0.31 0.19 0.55* 0.03 0.02
Carlina acaulis 17 0.08 0.05 0.13 0.11 0.17 0.88
Anthyllis alpestris 17 0.18 0.21 0.15 0.06* 0.69 0.35
Plantago alpina 17 0.29 0.48 0.18 0.34 0.05 0.16
Campanula barbata 17 0.13 0.21 0.08 0.14 0.03 1.44
Vaccinium gaultherioides 17 1.16 2.17 0.66 3.59** 0.03 0.91
Pulsatilla alpina ssp. apiifolia 17 0.18 0.09 0.33 0.62 0.11 1.21+
Leontodon helveticus 17 0.49 1.51+ 0.17 0.22 0.11 2.55
Vaccinium myrtillus 17 1.16 1.91 0.73 3.83** 0.03 2.27
Galium anisophyllon 16 0.13 0.03** 0.43 0.35 0.60 1.15
Hypochaeris uniflora 16 0.16 0.23 0.11 0.31+ 0.01 0.07
Gentiana campestris 16 0.12 0.20 0.07 0.06 0.11 0.58
Luzula multiflora 16 0.21 0.34 0.14 0.32+ 0.03 1.09
Festuca violacea 16 0.26 0.34 0.20 0.13 0.43 8.44+
Selaginella selaginoides 16 0.12 0.28 0.06 0.06 0.04 0.04
Potentilla erecta 16 0.59 1.05 0.35 0.48 0.20 0.63
Gymnadenia conopsea 15 0.04 0.05 0.04 0.06 0.01 0.71
Bartsia alpina 15 0.08 0.21* 0.03 0.02 0.06 1.62*
Trollius europaeus 15 0.15 0.18 0.13 0.13 0.14 3.16**
Polygala alpestris 13 0.03 0.01 0.05 0.05 0.05 0.49
Trifolium pratense ssp. pratense 13 0.15 0.27 0.09 0.04 0.30 1.73
Poa alpina 13 0.05 0.03 0.07 0.04 0.18 0.89+
Calluna vulgaris 13 0.25 0.32 0.20 1.16** 0.00 0.62
Thesium alpinum 13 0.03 0.02 0.03 0.03 0.03 0.06
Luzula sylvatica ssp. sieberi 13 0.07 0.05 0.08 0.15 0.03 0.31
Solidago virgaurea ssp. minuta 13 0.12 0.08 0.17 0.13 0.26 1.43
Ranunculus villarsii 12 0.06 0.04 0.10 0.04 0.34 1.87*
Silene vulgaris s.l. 12 0.03 0.02 0.04 0.02 0.11 1.48
Leucanthemum adustum 12 0.04 0.03 0.05 0.03 0.11 0.85
Euphrasia minima 12 0.09 0.18 0.04 0.10 0.01 0.11
Polygonum viviparum 12 0.04 0.03 0.04 0.02 0.15 0.78
Hieracium pilosella 11 0.03 0.04 0.02 0.02 0.03 0.06
Vaccinium vitisidaea 11 0.12 0.11 0.13 0.66* 0.00 0.10
Crepis aurea 11 0.04 0.10+ 0.01 0.02 0.01 0.07
Pedicularis tuberosa 11 0.04 0.05 0.02 0.06+ 0.00 0.32
(continued on next page)
M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11 9

Appendix (continued)
Species N AB ABM ABG ABVLI ABLI Change in 50 years
Ranunculus montanus 10 0.02 0.05 0.01 0.02 0.00 0.02
Phyteuma orbiculare 10 0.02 0.01+ 0.04 0.02+ 0.15 1.89*
Pulsatilla vernalis 10 0.02 0.02 0.03 0.08* 0.00 0.63
Trifolium badium 10 0.04 0.05 0.03 0.00* 0.18 1.25*
Agrostis capillaris 10 0.04 0.01* 0.14 0.19 0.08 1.20
Briza media 10 0.09 0.09 0.09 0.11 0.07 0.04
Scabiosa lucida 10 0.05 0.05 0.05 0.01 0.26 1.97
Aster bellidiastrum 10 0.03 0.03 0.03 0.03 0.02 0.28+
Alchemilla glabra 9 0.02 0.03 0.01 0.01 0.03 0.44
Gentiana verna 9 0.02 0.01 0.02 0.01 0.03 1.90+
Euphrasia montana 9 0.03 0.06 0.02 0.03 0.01 0.47
Deschampsia caespitosa 9 0.04 0.07 0.03 0.02 0.04 0.63
Polygala chamaebuxus 9 0.04 0.03 0.06 0.06 0.05 0.13
Ranunculus tuberosus 9 0.03 0.02 0.03 0.06 0.01 0.25
Phleum alpinum 9 0.02 0.01 0.05 0.04 0.07 0.50
Thymus polytrichus 9 0.06 0.00*** 0.34 0.21 0.80 2.23
Carex ornithopoda 8 0.01 0.02 0.01 0.00 0.03 0.05
Geranium silvaticum 8 0.01 0.02 0.01 0.02 0.00 0.20
Botrychium lunaria 8 0.01 0.01 0.02 0.05* 0.00 0.47
Plantago atrata 8 0.02 0.02 0.01 0.00* 0.06 1.92*
Carduus defloratus 8 0.02 0.01 0.02 0.01 0.05 0.39
Coeloglossum viride 8 0.01 0.01 0.01 0.01 0.02 0.06
Dianthus superbus 8 0.01 0.02 0.01 0.02 0.01 0.04
Rhinanthus glacialis 8 0.05 0.09 0.02 0.06 0.00 0.63
Chaerophyllum villarsii 8 0.05 0.02 0.13 0.09 0.27 4.65+
Phyteuma betonicifolium 8 0.02 0.02 0.01 0.02 0.01 0.01
Antennaria dioica 8 0.02 0.00** 0.06 0.14* 0.01 0.69*
Androsace obtusifolia 8 0.03 0.06 0.01 0.02 0.01 0.00
Myosotis alpestris 7 0.01 0.01 0.00 0.00 0.01 0.04
Alchemilla hybrida 7 0.03 0.03 0.02 0.03 0.01 0.26
Alchemilla sp. 7 0.01 0.01 0.01 0.00 0.07 0.68+
Prunella vulgaris 7 0.01 0.01 0.02 0.01+ 0.07 0.22
Anemone narcissiflora 7 0.02 0.02 0.02 0.03 0.02 2.17
Agrostis alpina 7 0.03 0.04 0.01 0.03 0.00 0.12
Nigritella nigra 7 0.01 0.01 0.01 0.01 0.00 0.05
Phleum hirsutum 7 0.01 0.01 0.02 0.01 0.07 0.82
Sesleria caerulea 7 0.02 0.01 0.06 0.03 0.13 1.69
Trifolium alpinum 7 0.04 0.06 0.03 0.06 0.01 1.39
Pseudorchis albida 7 0.01 0.02 0.00 0.01 0.00 0.04
Trifolium pratense ssp.nivale 7 0.03 0.03 0.03 0.02 0.05 1.02
Helianthemum grandiflorum 7 0.05 0.01* 0.16 0.06 0.73 2.34
Knautia dipsacifolia 6 0.01 0.01 0.01 0.01 0.03 0.43
Pedicularis foliosa 6 0.01 0.01 0.01 0.01 0.01 0.02
Erica carnea 6 0.02 0.01+ 0.05 0.03+ 0.14 7.43**
Leontodon hispidus 6 0.01 0.02 0.01 0.01 0.03 0.46
Veratrum album ssp.lobelianum 6 0.01 0.01 0.01 0.01 0.01 0.03
Cerastium fontanum 6 0.01 0.01 0.01 0.01 0.01 0.03
Laserpitium halleri 6 0.02 0.01 0.05 0.16* 0.00 6.90
a
Ndenotes the number of parcels with the species, AB, ABM and ABG the backtransformed mean abundance (in % ground cover) for all, for
mown and for grazed parcels, and ABLI and ABVLI for parcels grazed at low and very-low intensity. Significant differences between 10 still tradi-
tionally mown and 11 recently grazed sites, between seven more and four less intensively grazed sites, and significant changes with time since con-
version to grazing are given in bold face, and in bold italics for marginal significance. +, P<0.1; *, P<0.05; **, P<0.01; ***, P<0.001.
10 M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11

References
Baur, B., Ewald, K.C., Freyer, B., Erhardt, A., 1997. O
¨kologischer
Ausgleich und Biodiversita
¨t. Grundlagen zur Beurteilung des Nat-
urschutzwertes ausgewa
¨hlter landwirtschaftlicher Nutzfla
¨chen. Bir-
kha
¨user, Basel.
Bignal, E., McCracken, D., 1996. Low-intensity farming systems in the
conservation of the countryside. Journal of Applied Ecology 33,
413–424.
Bischof, N., 1981. Gema
¨hte Magerrasen in der subalpinen Stufe der
Zentralalpen. Bauhinia 7, 81–128.
Bischof, N., 1984. Pflanzensoziologische Untersuchungen von Sukzes-
sionen aus gema
¨hten Magerrasen in der subalpinen Stufe der Zen-
tralalpen. Beitra
¨ge zur geobotanischen Landesaufnahme der
Schweiz 60, 1–128.
Braun-Blanquet, J., 1964. Pflanzensoziologie, 3rd Edition. Springer
Verlag, Wien.
Da
¨hler, W., 1993. Langfristige Auswirkungen menschlicher Eingriffe
in alpine Vegetation. Beitra
¨ge zur geobotanischen Landesaufnahme
der Schweiz 69, 1–139.
DiTommaso, A., Aarssen, L.W., 1989. Resource manipulation in nat-
ural vegetation: a review. Vegetatio 84, 9–29.
Ellenberg, H., 1996. Vegetation Mitteleuropas mit den Alpen in
oekologischer, dynamischer und historischer Sicht, 5th edn. Ulmer,
Stuttgart.
Erhardt, A., 1995. Ecology and conservation of alpine Lepidoptera.
In: Pullin, A.S. (Ed.), Ecology and Conservation of Butterflies.
Chapman and Hall, London, pp. 258–276.
Erhardt, A., Thomas, J.A., 1991. Lepidoptera as indicators of change
in the semi-natural grasslands of lowland and upland Europe. In:
Collins, N.M., Thomas, J.A. (Eds.), The Conservation of Insects
and their Habitats. 15th Symposium of the Royal Entomological
Society of London. Academic Press, London, pp. 213–236.
Fischer, M., Sto
¨cklin, J., 1997. Local extinctions of plants in remnants
of extensively used calcareous grasslands 1950–1985. Conservation
Biology 11, 727–737.
Grubb, P.J., 1977. The maintenance of species-richness in plant com-
munities. the importance of the regeneration niche. Biological
reviews 52, 107–145.
Gu
¨nter, T., 1985. Landnutzungsa
¨nderungen in einem alpinen Tour-
ismusort. Ein integraler Ansatz zur Erfassung von Wechselbezie-
hungen zwischen raumwirksamen sozio-o
¨konomischen Prozessen
und dem Naturhaushalt, dargestellt am Beispiel Davos. MAB-
Schlussbericht Nr. 13, Bundesamt fu
¨r Umwelt, Wald und Land-
schaft BUWAL, Bern, Switzerland.
Hegg, O., Be
´guin, C., Zoller, H., 1993. Atlas schutzwu
¨rdiger Vegeta-
tionstypen der Schweiz. Bundesamt fu
¨r Umwelt, Wald und Land-
schaft BUWAL, Bern, Switzerland.
Ko
¨rner, C., 1999. Alpine Plant Life. Springer, Berlin.
Landolt, E., 1991. Rote Liste. Gefa
¨hrdung der Farn- und Blu
¨tenp-
flanzen in der Schweiz. Bundesamt fu
¨r Umwelt, Wald und Land-
schaft BUWAL, Bern, Switzerland.
Lauber, K., Wagner, G., 1998. Flora Helvetica, 2nd Edition. Haupt,
Bern, Switzerland.
Lenzin, H., 1995. Vegetationsa
¨nderungen durch Nutzungsaufgabe und
ihre a
¨sthetischen Konsequenzen. PhD thesis, University of Basel.
Lu
¨thy, W., Wiedemeier, P., 1998. Extensive Weiden und Artenvielfalt.
Fallstudien 1997. Bern, Graubu
¨nden, St. Gallen. Agrofutura,
Brugg, Switzerland.
Marschall, F., Dietl, W., 1974. Beitra
¨ge zur Kenntnis der Borst-
grasrasen der Schweiz. Schweizerische landwirtschaftliche For-
schung 13, 115–127.
Mountford, J.O., Lakhani, K.H., Kirkham, F.W., 1993. Experimental
assessment of the effects of nitrogen addition under hay-cutting and
aftermath grazing on the vegetation of meadows on a Somerset peat
moor. Journal of Applied Ecology 30, 321–332.
Niederer, A., 1996. Alpine Alltagskultur zwischen Beharrung und
Wandel: ausgewa
¨hlte Arbeiten aus den Jahren 1956 bis 1991.
Haupt, Bern, Switzerland.
Pauli, D., 1998. Plant species diversity and productivity in wetland
communities: patterns and processes. PhD thesis, University of
Zu
¨rich.
Schiefer, J., 1981. Bracheversuche in Baden-Wu
¨rttemberg. Beihefte zu
den Vero
¨ffentlichungen u
¨ber Naturschutz und Landschaftspflege
Baden-Wu
¨rttemberg 22, 1–325.
Schla
¨pfer, M., Zoller, H., Ko
¨rner, C., 1998. Influences of mowing and
grazing on plant species composition in calcareous grassland. Bota-
nica Helvetica 108, 57–67.
Schreiber K.F., Schiefer, J., 1985. Vegetations- und Stoffdynamik in
Gru
¨nlandbrachen—10 Jahre Bracheversuche in Baden-Wu
¨rttem-
berg. In: Schreiber K.F. (Ed.), Sukzession auf Gru
¨nlandbrachen.
Mu
¨nstersche Geographische Arbeiten, Muenster, Germany 20, 111–
153.
Senn, U., 1952. Die Alpwirtschaft der Landschaft Davos. PhD thesis,
University of Zu
¨rich.
Stampfli, A., Ha
¨felfinger, S., 1995. Tessiner Magerwiesen im Wandel.
Schriftenreihe Umwelt, 246, BUWAL, Bern, 1–134.
Sto
¨cklin, J., 1992. Environment, morphology and growth of clonal
plants—an overview. Botanica Helvetica 102, 3–21.
Surber, E., Amiet, R., Kobert, H., 1973. Das Brachlandproblem in der
Schweiz. EAFV Bericht Nr. 112, Swiss Federal Institute for Forest,
Snow, and Landscape Research, Birmensdorf, Switzerland.
Ter Braak, C.J.F., 1986. Canonical correspondence analysis. A new
eigenvector technique for multivariate direct gradient analysis.
Ecology 67, 1167–1179.
Ter Braak, C.J.F., 1987. Ordination. In: Jongman, R.H.G., Ter Braak,
C.J.F., van Tongeren, O.F.R. (Eds.), Data Analysis in Community
and Landscape Ecology. Pudoc, Wageningen, pp. 91–173.
Ter Braak, C.J.F., Smilauer, P., 1998. CANOCO Reference Manual
and User’s Guide to Canoco for Windows: software for Canoni-
cal Community Ordination (version 4). Microcomputer Power,
Ithaca.
Thomet, P., Schmid, W., Daccord, R., 1989. Erhaltung von artenrei-
chen Wiesen. Nationales Forschungsprogramm ‘‘Boden’’, Bericht
37, Bern-Liebefeld, Switzerland.
Van den Bos, J., Bakker, J.P., 1990. The development of vegetation
patterns by cattle grazing at low stocking density in the Nether-
lands. Biological Conservation 51, 263–272.
WallisDeVries, M.F., Bakker, J.P., van Wieren, S.E. (eds.) 1998.
Grazing and Conservation Management. Kluwer Academic Pub-
lishers, Dordrecht.
Willems, J.H., 1985. Growth form spectra and species diversity in
permanent grassland plots with different management. In: Schreiber
K.F. (Ed.), Sukzession auf Gru
¨nlandbrachen. Mu
¨nstersche Geo-
graphische Arbeiten, Muenster, Germany 20, 35–43.
Wilmanns, O., 1998. O
¨kologische Pflanzensoziologie, 6th Edition.
Quelle and Meyer, Heidelberg.
Wolkinger, F., Plank, S., 1981. Dry grasslands of Europe. Council of
Europe, Nature and Environment Series no. 21, Strasbourg.
Zoller, H., Bischof, N., 1980. Stufen der Kulturintensita
¨t und ihr Ein-
fluss auf Artenzahl und Artengefu
¨ge der Vegetation. Phytocoenolo-
gia 7, 35–51. (Festband Tu
¨xen).
Zumbu
¨hl, G., 1983. Pflanzensoziologisch-o
¨kologische Untersuchun-
gen von gema
¨hten Magerrasen bei Davos. PhD thesis, ETH Zu
¨rich.
M. Fischer, S. Wipf / Biological Conservation 104 (2002) 1–11 11