Occurrence of Characeae in Switzerland over the last two centuries (1800–2000)

(1)

Article

Reference

Occurrence of Characeae in Switzerland over the last two centuries (1800–2000)

AUDERSET JOYE, Dominique, CASTELLA, Emmanuel, LACHAVANNE, Jean-Bernard

Abstract

Some 3100 specimens of Characeae from herbarium collections and recent sampling programmes in Switzerland were redetermined and their provenance analysed, in order to investigate historical changes in frequency, distribution and current status of the species.

Altitudinal and geographic distribution patterns of 27 taxa are described in 125 km2 grid squares covering Switzerland, and an analysis was made of species richness of Characeae in different grid squares. Data were available for approximately half (117 out of 262) of the grid squares covering Switzerland, showing that the recording effort had not been uniform. The lowland was better investigated than either the hills or the mountains. And 17 taxa were categorised as rare, on the basis that they were each known from

AUDERSET JOYE, Dominique, CASTELLA, Emmanuel, LACHAVANNE, Jean-Bernard.

Occurrence of Characeae in Switzerland over the last two centuries (1800–2000). Aquatic Botany , 2002, vol. 72, no. 3-4, p. 369-385

DOI : 10.1016/S0304-3770(01)00211-X

Available at:

http://archive-ouverte.unige.ch/unige:25575

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Occurrence of Characeae in Switzerland over the last two centuries (1800–2000)

D. Auderset Joye

^∗

, E. Castella, J.-B. Lachavanne

Laboratoire d’Ecologie et de Biologie Aquatique, University of Geneva, 18 Chemin des Clochettes, 1206 Geneve, Switzerland

Abstract

Some 3100 specimens of Characeae from herbarium collections and recent sampling programmes in Switzerland were redetermined and their provenance analysed, in order to investigate historical changes in frequency, distribution and current status of the species. Altitudinal and geographic distribution patterns of 27 taxa are described in 125 km²grid squares covering Switzerland, and an analysis was made of species richness of Characeae in different grid squares. Data were available for approximately half (117 out of 262) of the grid squares covering Switzerland, showing that the recording effort had not been uniform. The lowland was better investigated than either the hills or the mountains. And 17 taxa were categorised as rare, on the basis that they were each known from

<100 records. Ten taxa were classified as not uncommon, among which Chara vulgaris and C.

globularis are the most frequent and widespread species. The distribution of some species shows a significant relationship with altitude, but analysis of the species distributions shows only a weak relationship with altitude of the grid square and biogeographical region, revealing that these two variables are not the main factors explaining the distribution of the species. Over the last 200 years, regression has occurred in the species that were formerly the rarest and expansion has occurred in a few species that were previously the most common. Further investigations are needed to precise the distribution of the declining species, particularly in the under-prospected parts of the country and in poorly known habitats. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Characeae; Switzerland; Biogeography; Altitude; Temporal changes; Non-symmetric correspondence analysis (NSCA)

1. Introduction

Over the past decades, a general increase has been observed in nutrient levels in freshwater ecosystems in Switzerland. Simultaneously, a qualitative and quantitative reduction of many aquatic plant species has been observed in the country’s lakes (Lang, 1968; Lachavanne and Wattenhoffer, 1975; Schmieder, 1991). It is shown that Characeae were particularly affected.

Studies in Europe and other parts of the world indicate that charophytes, or at least some

PII: S 0 3 0 4 - 3 7 7 0 ( 0 1 ) 0 0 2 1 1 - X

(3)

370 D. Auderset Joye et al. / Aquatic Botany 72 (2002) 369–385

of them, decline under conditions of increasing water nutrient concentration and turbidity (Forsberg, 1965; Ozimek and Kowalcezewski, 1984; Corillion, 1986; Blindow, 1992; Moss et al., 1996; Van den Berg et al., 1998). Stoneworts have high ecological significance in many aquatic systems (Pereyra-Ramos, 1981; Castella and Amoros, 1984; Casanova and Brock, 1991, 1999; Schwarz and Hawes, 1997; Proctor, 1999). A large proportion of the group is regarded as threatened in Europe. The present study aimed at: (i) documenting the current and previous distribution of Characeae in Switzerland and associated patterns of species richness, to identify any changes that have occurred over the last two centuries;

(ii) testing how species distribution and richness are related to altitude and biogeographic regions within Switzerland.

The study was based upon two types of data: (i) the complete set of Swiss herbarium collections gathered mostly between 1830 and 1930, and (ii) site samples collected mostly between 1975 and 1999.

The only reliable information associated with herbarium collections is date, location and altitude. This paper is thus primarily concerned with these variables and with analysis of all the Swiss Characeae inventory data available from the two last centuries.

Located in central Europe, Switzerland is characterised by a steep altitudinal gradient (ca. 200–4000 m) occurring within a small area (40,000 km²), providing a good basis for investigation of the altitudinal distribution of Characeae.

2. Materials and methods

2.1. Data collection

Many different botanists contributed records to the data set, which spans the period ca.

1800–2000. Herbarium collections of the following institutions were used, the Universities of Zürich, Bern, Fribourg, Neuchâtel, Lausanne and Geneva, the Federal Institute of Tech- nology Zürich and the Botanical Garden of Geneva. Among the records from before 1930, 40% were recorded from lakes, and 60% from various other types of aquatic ecosystems (ponds, rivers, floodplain channels, wetlands, bogs, etc.). Recent data were derived from field surveys, conducted mainly in lakes and ponds with various eutrophication levels. All identifications were revised using Corillion (1975). More than 3100 records, for a total of 27 taxa, were identified, recorded and mapped. A record is considered as a reliable obser- vation of a species at a given date and location, and is backed by a voucher specimen. Given the nature of the data set, no “zero observations” (i.e. sites prospected that contained no Characeae) were included. No bias towards rare species would be suspected in recent data because the field surveys did not aim at the specific collection of Characeae. For the older data no such information was available.

As a basis to describe species distributions, Switzerland was split into 12.5 km×10 km rectangles (hereafter called “grid squares”) each corresponding to one 1:25,000 map of the Swiss territory. Each record was assigned to one of the 262 grid squares thus defined. This size of the grid squares, which is similar to that used by Gimaret-Carpentier et al. (1998), allowed a sufficient number of records per grid. Each grid square was assigned to one of the 16 Swiss biogeographic regions. The country can be divided into three main geographic

(4)

entities, the Jura, which occupies the northwestern part, the plateau in the centre and the Alps in the southeast. These three main entities can be further subdivided. The map symbol used here to indicate geographic distribution was positioned (Figs. 1 and 4) at the central co-ordinates of each grid square. For grid squares shared with France, Germany, Austria or Italy, the centre of the grid was sometimes located outside the Swiss border.

The frequency of species occurrence was calculated on the occupied grid squares only, i.e.

grids for which at least one record exists. In analyses in which grid squares were involved, the altitude of a grid square was defined as the average of the altitudes of all records contained in the square. Each grid square was then assigned to an altitude class. Information on whether the empty grid squares were really unoccupied or had simply not been prospected was not available.

2.2. Altitudinal distribution

In order to express the altitudinal profiles of the species, the altitudinal range covered by the 3100 records was divided into six altitudinal classes. The altitudinal data associated with each individual record is considered here. The altitudinal class boundaries were set so that each class comprised at least 5% (151) of the records. The six classes so defined are: (1)≤400 m, (2) 401–500 m, (3) 501–600 m, (4) 601–1100 m, (5) 1101–1700 m, (6) 1701–2500 m. The profile of a species was then expressed as the difference between its frequency within an altitudinal class and its frequency in the whole set of data (overall frequency). This approach to calculation of species profiles is based on the methodology of Daget and Godron (1982). The profiles were then subjected to aχ²-test, to establish the degree to which they differed significantly from what would be expected were there an even distribution of the species between altitudinal classes. This approach was selected because an alternative logistic regression procedure would have been impossible because of the absence of “zero values” for the species distribution. These profiles were calculated for the 10 most frequent species only.

2.3. Temporal changes

The temporal distribution of samples showed clearly two periods of investigation, mainly before 1930 and after 1975. The very few specimens collected between 1930 and 1975 showed that botanists had other interests during that period of time.

To examine the changes between the two periods, the frequency of each species was calculated per grid square on the total number of specimens collected at each period. Frequency changes between the two periods were then calculated for each grid square.

3. Data analysis

3.1. Non-symmetric correspondence analysis (NSCA) and the ordination of species occurrence

In order to give a more synthetic picture of the distribution of Characeae species, a NSCA was applied to the table containing 117 grid squares and the frequency of 30 taxa

(5)

(27 species plus Chara sp., Nitella sp. and N. anarthrodactylae). NSCA was presented by Lauro and D’Ambra (1984) and introduced in vegetation analysis by Chessel and Gimaret (1997) and Gimaret-Carpentier et al. (1998). The mathematical foundation of NSCA can be found in these three papers, which also discuss its differences from classical (i.e. symmetric) correspondence analysis. NSCA is an ordination method especially designed for the analysis of species occurrences gathered from different sources such as literature, atlases or field surveys. It enables the investigation of data structure by taking into account the asymmetry between the distribution of occurrences among grid squares (controlled) and the distribution among species (observed). The use of NSCA of grid square profiles (i.e. species composition per square) yields information about their floristic differentiation. In this form of ordination,

“abundant species are rather taken into account while rare species are located near the origin”

(Gimaret-Carpentier et al., 1998).

Species occurrence was suspected to have altered between the two time periods (before and after 1930). In order to assess the effect of change over time and space (altitude, biogeography) a second NSCA was performed on the grid squares for which occurrences were reported in both periods. The file contained 100 grid squares (50 squares×2 dates) and the frequency of 30 taxa (27 species plus Chara sp., Nitella sp. and N. anarthrodactylae). The interpretation of this analysis was restricted to the Swiss Oriental plateau, an homogeneous and well-documented region for the two periods.

4. Results

4.1. Distribution of the records in space and time

The complete database included more than 3100 records. Of these, 64% (2033) were collected before 1930 and 36% (1045) after 1930. The specimens were collected at altitudes varying from 193 to 2500 m. About two-thirds (63%) of them were harvested at altitudes lower than 500 m and one-third between 500 and 2500 m (Table 1).

The majority (48%) of the samples was collected on the Swiss plateau, mostly from the Leman basin and the Oriental plateau. A smaller proportion came from the foothills of the Alps (15%), Grisons (11%) and Jura (4.5%). Only a very few specimens (1%) were gathered

Table 1

Number of Characeae records (N) within six altitudinal classes

Altitude class (m) N %

>400 784 24.7

400–499 1208 38.0

500–599 357 11.2

600–1099 237 7.5

1100–1700 198 6.2

1700–2500 243 7.6

Missing information 151 4.8

Total 3178 100.0

(6)

Table 2

Numbers (N) and percentages of Characeae records within the Swiss biogeographic regions

Biogeographic regions N %

Internal high Alps 1 0.0

Internal low Alps 49 1.5

Engadine and high valleys 344 10.8

Northern Grisons 1 0.0

Jura 142 4.5

Rhine basin 68 2.1

Oriental plateau 737 23.2

Occidental plateau 202 6.4

Leman basin 517 16.3

Central foothills of the Alps (regions 7 and 9)^a 381 12.0

Occidental foothills of the Alps 105 3.3

Southern high valleys 11 0.3

Bregaglia 3 0.1

Southern Tessin 13 0.4

Missing information^b 604 19.0

Total 3178 100.0

aCentral foothills of the Alps are divided into two parts.

bMissing information: many samples from lake Geneva and lake Constance were located outside Switzerland and were not assigned to a Swiss biogeographic region.

in the southern Alps. The specimens collected in the foothills of the Alps came mainly from the central part and the samples from Grisons originated mainly from Engadine and the high valleys (Table 2). The geographic distribution of the specimens showed a dominance of harvesting next to large cities (Geneva and Zurich) and Grisons, a holiday or health resort for botanists or hydro-biologists. These records came from a total of 117 grid squares (out of 262) corresponding to 48% coverage of Switzerland. Grid squares from which no Characeae were recorded (129) are located mainly in the Alps and the southern part of the country (Valais, Tessin).

4.2. Species inventory and frequencies

Based upon the complete set of herbarium collections (ca. 3100 data) the Charophyta flora of Switzerland consists of 27 taxa (Table 3). A high proportion of the species were recorded infrequently; seven taxa were classed as very rare, being recorded on fewer than 10 occasions (<1% of the records): N. capillaris, N. confervacea, C. gymnophylla, N. flexilis, Tolypella intricata, C. tenuispina, C. polyacantha. Ten taxa were classed as rare, with 10–100 records (<3% of the records): N. gracilis, C. delicatula, C. intermedia, N. tenuissima, C. aculeolata, N. mucronata, Tolypella glomerata, Nitellopsis obtusa, C. strigosa and C. hispida. Only ten species were recorded on more than 100 occasions: C. aspera, C. contraria, C. tomentosa, N. opaca, N. hyalina, N. syncarpa, and C. denudata. The most frequent species were C.

globularis (16% of the total records) and C. vulgaris (23%). About a 100 specimens were not determined to species (Chara sp., Nitella sp.).

In terms of grid square occupancy (Table 3), the most frequent species were also C.

vulgaris (71% of the squares) and C. globularis (57%). Six species, C. aspera, C. contraria,

(7)

374D.AudersetJoyeetal./AquaticBotany72(2002)369–385

(8)

C. major, N. syncarpa, C. intermedia, N. opaca occurred in one-quarter to one-third of the squares. The number of grid squares occupied by a given species was significantly (R²=0.898) correlated with the number of specimens for that species. This indicated that assigning a specimen to grid squares did not introduce a bias into the observed distribution of the species. This also showed that there was no strong geographical pattern in the distribution of the species.

4.3. Geographic distribution of the species

Species richness per grid square varied from 1 to 19 species. About one-third of the squares were occupied by a single species and the average species richness per square was four. For a given taxon the number of records differed from 2 to 658 and the number of occupied squares from 1 to 83. The geographical distributions for the 10 most frequent species are given in Fig. 1.

The distribution of some species showed clearly a significant relationship with altitude (Fig. 2). The 10 species with more than 100 records all showed significant altitudinal profiles (χ²,P =0). Five of the species: N. hyalina, C. contraria, C. tomentosa, N. syncarpa and C. denudata were more frequent at lower altitudes (<400 m) (Fig. 2). C. major was more frequent between 400 and 500 m elevation. C. vulgaris and C. globularis were found in a wide range of altitude (400–1700 m) but were less frequent in the extreme lower and higher altitude classes. C. aspera and N. opaca were more frequent at the higher elevations (1100–2500 m). Although the graph (Fig. 3) suggests a tendency of decreasing species richness with increase in altitude, the relationship between species richness and altitude was not significant in term of regression (neither linear withR² = 0.048, nor quadratic withR²=0.055). As shown in Fig. 3, high altitudes were associated with lower species richness but the entire range of richness values was observed at lower altitudes.

The changes in species frequency between the two periods revealed several trends (Table 3). C. tenuispina, N. capillaris, N. confervacea, and Tolypella intricata, four species with very few records before 1930, were not recorded at all in the second period. The same is true of N. hyalina, a relatively abundant species in the samples collected in the first pe- riod. C. aspera, C. major, C. tomentosa and N. syncarpa regressed, unlike C. globularis, C. denudata and N. obtusa, which showed an increase in the frequency of their records.

The geographic distribution of species richness values between the two periods of collection showed that grid squares located in the Jura, the Plateau and Grisons had a greater richness than other parts of the country before 1930 (Fig. 4). In the second period, the squares situated in the eastern and southwestern parts of the plateau, the central Alps, and the southwestern part (Grisons), had more species collected than the mean for that period.

Comparison between species frequencies in the altitudinal classes for both periods of collection, showed that the proportion of specimens collected was higher at the low altitudes

(9)

Fig. 1. Geographic distribution in 12.5 km×10 km grid cells of the 10 most frequently occurring Characeae species in Switzerland. The circle surface area is proportional to the number of records in the corresponding grid squares.

(10)

Fig. 2. Altitudinal profiles of the most frequent Swiss Characeae species. For a given altitudinal class on the ordinate, each bar represents the difference between the frequency of the species among all records for that class, and the frequency of the species in the total set of records (this latter figure is given as percentage for each of the 10 species). The horizontal line indicates zero (i.e. no difference between the two frequencies). The vertical axes of the 10 profiles are identical (minimum−14%, maximum+23%). All profiles are significantly different from an even distribution between the altitude classes (χ²-test,P=0). The altitude classes 1–6 are, from left to right:≤400 m (n=838), 401–500 m (n=1174), 501–600 m (n=342), 601–1100 m (n=233), 1101–1700 m (n=197), 1701–2500 m (n=243).

for the first period (before 1930) and more evenly distributed among altitude classes during the second period (Fig. 5).

4.6. Ordination of Characeae distributions 4.6.1. Spatial and altitudinal patterns

The results of the NSCA ordination were interpreted by the variations of two main ecological factors available in the study, elevation and the biogeographic regions.

In NSCA, if the frequency of occurrence of a species in a given grid square is determined only by the overall frequency of that species (i.e. there are no spatial effects) the explained variance (between-class inertia to total inertia) is null (R² = 0). On the contrary, if the species profile allows a perfect separation of grid squares the variance is totally explained (R² = 1). In our data the species profile explained 15% of variance. The square species profiles were significantly different as revealed by the Catanova index (13241 withP <

0.00001, d.f.=3364). This difference corresponded to a strong structure mainly accounted for by the first factors of NSCA (the two first eigenvalues represented 27 and 15% of the total

(11)

Fig. 3. Species richness of grid squares in relation to the elevation (m) of each square.

inertia, respectively). The distribution of the grid square scores on the first NSCA factorial map (not shown) corresponded to gradients in species composition mainly determined by three species (C. globularis, C. vulgaris and N. hyalina), but these were neither related to altitude nor to geography.

The identification of the altitudinal classes, and of the Swiss regions on the first NSCA factorial map of the grid squares revealed a dense overlap between the categories of each of the variables. This suggests an absence of relationships between the grid square species composition and either of these two geographic variables.

4.6.2. Temporal changes in the Swiss Oriental Plateau

Between the two time periods, changes in the species composition of the grid squares were evident in all biogeographic regions (Fig. 6B a–c, e, f, h, i, k, m–o) but showed heterogeneity in the direction in which the changes occurred, except in Oriental plateau (Fig. 6B b). This northeastern part of Switzerland lies between 400 and 500 m altitude and contains 11 grid squares, for which Characeae were recorded within both time periods, and are comparatively well documented (341 records before 1930, 219 records after 1930).

On the first factorial map (Fig. 6B) all 11 grid squares scores evidenced changes, represented by the homogeneous right to left direction of all but one of the 11 arrows. In terms of species changes (Fig. 6A), the grid squares of the first period contained more species, such as C. major, C. aspera, C. tomentosa, N. syncarpa, N. opaca. In the second period, the square factorial scores moved to the negative side of F1, which means that the grid squares harbour proportionally more C. vulgaris, C. denudata, N. obtusa and C. globularis. The net

(12)

Fig. 4. Geographic distribution of species richness per grid square for the two periods of collection (before 1930 and after 1930). The surface area of the circles is proportional to the species richness per grid square.

Fig. 5. Numbers of Characeae records (N) per altitudinal class for two collection periods.

(13)

Fig. 6. Ordination of grid squares documented for both time periods. (A) Species scores on the first factorial map, (B) First factorial map of the grid squares scores split up according to the 11 biogeographic regions occupied by Characeae: (a) Rhine basin; (b) Oriental plateau; (c) Jura; (e) Central foothills of the Alps; (f) Occidental plateau;

(h) Occidental foothills of the Alps; (i) Engadine and high valleys; (k) Leman basin; (m) Southern high valleys;

(n) Internal low Alps; (o) Southern Tessin. The arrows represent changes in species composition for each grid square between the pre-1930 score (circle) and the post-1930 score (tip of the arrow). The bold-framed Oriental plateau is discussed with more details in the text.

(14)

picture given by this analysis is one of a replacement over time of the scarce species by the more frequent C. vulgaris or C. globularis.

5. Discussion

5.1. Geographical distribution

Distribution data were only available for charophytes for half of the total grid squares covering the country. It is most unlikely that Characeae were missing from all of these empty squares. A much more likely explanation is recording bias, resulting from unequal sampling effort in the different parts of the country. Areas in northeast (Zurich) and west (Geneva) of the country showed the highest species richness, which is correlated with a higher number of samples collected. The number of species identified in a given area is (non-linearly) related to the number of herbarium specimens that have been collected (Soberón and Llorente, 1993; Rosenzweig et al., 1998). When an area has been poorly prospected, new specimens will frequently incorporate new species, but once the collection is growing, most new specimens will belong to species already collected.

As shown by the distribution maps most species were widespread throughout the country but the species richness of a particular region is directly related to the number of records and thus to the activity of aquatic botanists. So we can expect to increase the species richness for some regions, and the quality of distribution data, with increasing inventory activities.

No distribution pattern related to Swiss biogeographical regions seems to emerge from the current data within Switzerland. The size of the country is probably too small to exhibit a clear pattern and most species have a wider range, as described by Corillion (1957) and Wood and Imahori (1959).

Some Characeae species showed a preference for a certain range of altitudes, but were found however along the whole gradient. This might mean that they are not sensitive to the altitude itself but to other variables related to elevation.

It is widely observed that the patterns of species richness decrease with both latitudinal and altitudinal gradients (Rahbek, 1995; Rosenzweig and Sandlin, 1997). This relation is valid for a large number of groups (trees, mammals, birds, reptiles, insects and amphibians) but there is no evidence that plant diversity declines with elevation, bryophytes and ferns reach peak diversities at mid elevations; (Scheiner and Rey-Benayas, 1994; Rosenzweig, 1995; Odland and Birks, 1999). In the case of Characeae, species richness values were maximal at low altitude (lower than 400 m) and minimal at higher elevation (Fig. 3). Unlike the species richness of ferns and bryophytes, the diversity of Characeae appears to decline with elevation, but the relation appeared not linear and we do not have enough homogeneous data to infer this relationship. If in our case the richness was suspected to be lower at high altitude, it may have different reasons: sampling bias, lower diversity of ecosystems (absence of large lakes) and lack of the species growing preferably at low altitude might be alternative explanations.

(15)

Usually, in addition to change in species richness there is also a qualitative compositional change with an increasing altitude and latitude (Odland and Birks, 1999). Even though in our data altitude showed some influence on some Characeae species, this parameter did not appear highly discriminating for the grid squares species content. In other words, the grid square species profiles were not explained by altitude. In our analysis, the effects of altitude were probably hidden by the heterogeneity of the aquatic ecosystems within the grid squares. A single grid square had a surface of 125 km²and might contain large lakes, small lakes, ponds, streams, pits, etc. which do not necessarily share the same species.

Some species, collected during the first period, were not recorded more recently. For the scarce species, it is difficult to separate the possible effects of under-collection and of actual disappearance. For more frequent species (like N. hyalina), the chances of under-collection are weaker and the degradation of ecosystem conditions seems to represent a more logical explanation. In addition to the disappearance and the regression of a number of species, the expansion of C. globularis and N. obtusa was obvious (Table 3).

Species composition changed for most of the grid squares between the two periods. These changes occurred in a variety of ways and no common pattern emerged for all the squares.

In the Oriental plateau, a common unidirectional change appears to have occured. In this homogeneous region, the variations in grid square species profiles showed a replacement of less frequent species (C. aspera, C. major, C. tomentosa, N. syncarpa) by more frequent species (C. globularis). The same trend was generally observed at the level of the whole country.

Neither altitude nor regional differences were the main factors explaining the occurrence and distribution of Characeae. However, the changes in species occurrence observed over the two periods showed that 52% of species regressed while 22% of them progressed.

Recent strong increases in records of the common species C. vulgaris and C. globularis were also noticed in The Netherlands (Simons and Nat, 1996). These two species appear to be relatively resistant to eutrophication (Hutchinson, 1975; Moore, 1986; Bornette et al., 1996). C. globularis tolerates phosphate enrichment (Wiegleb, 1978; Auderset Joye, 1993) and mesotrophic to eutrophic water with high ammonium and phosphate contents (Bornette et al., 1996). C. vulgaris appears resistant to an increase of the trophic level (Wattenhoffer, 1984) and tolerant to nutrient enrichment (Bornette et al., 1996).

In contrast, C. hispida, one of the species in regression, is usually associated with low ammonium (Carbiener et al., 1990) or low phosphate and ammonium content (Bornette et al., 1996). C. major, a species closely related to C. hispida, is also known to decline currently, but probably tolerates higher phosphorus contents than C. hispida (Bornette et al., 1996).

We may then presume that the extension of the nutrient resistant species and the regression of the more sensitive may be caused by the general, well-documented, increase of nutrient levels in freshwater ecosystems. This increase is the most serious threat that Characeae are facing. The threat related to nutrient levels seems to differ with altitude. When the relative frequency of two species (C. aspera, N. opaca) found more frequently at high altitude is compared for the two time periods, both species show the same trend, before 1930, both were more frequent at low altitudes and after 1930 they were more frequent at high altitudes.

(16)

This suggests that the two species no longer encounter favourable conditions in the lowland but still find refuge in the mountains, where human impact is weaker.

Our knowledge of the current distribution of the Characeae has advanced recently, but we should be able to define the degree of threat to which the species are subject, some of the less well-studied species may persist where they have been recorded previously, despite lack of recent records, and these sites require careful monitoring. Several areas require more general surveys, Jura, the Alps and the southern part of Switzerland (Tessin), especially the part of this area with a granite substratum and soft water. If the major lakes have been relatively well prospected since 1970, the small water bodies and the temporary pools are not well known. Investigations on the ecological needs of the species should also be carried out. The links between species and types of ecosystem require to be established more precisely, as well as the relationships between species and environmental parameters, which are only partially documented at present. Experimental studies are also needed to develop predictive models.

6. Conclusions

The distribution of Characeae within Switzerland does not appear to be determined by biogeographical constraints and is weakly related to altitude. Temporal changes in the aquatic ecosystems have had a notable impact on the distribution of Characeae: there is no confirmation of the continued presence of some of the species in the country since 1930, whereas others have declined and only a few have increased in frequency. Species whose range is expanding (C. globularis and C. vulgaris) are those more resistant to eutrophication.

Eutrophication is probably the most important threat facing the declining species. The degree of threat of the declining species should be evaluated. Further investigations are needed to determine the distribution of these species, particularly in under-prospected parts of the country (Alps, Tessin) and in poorly known habitats, such as temporary ponds.

Autecological studies of the species are also needed.

Acknowledgements

The authors thank Martin C.D. Speight (Dublin) for editing the English text. They are grateful for valuable suggestions from Kevin J. Murphy and Egbert van Nes who helped improving the manuscript. Most of this work was funded by the Swiss National Science Foundation Grant 31-8521.85.

References

Auderset Joye, D., 1993. Contribution à l’Écologie des Characées de Suisse. Thesis no. 2580, Université de Genève.

Blindow, I., 1992. Decline of charophytes during eutrophication: comparison with angiosperms. Freshwat. Biol.

28, 9–14.

(17)

Bornette, G., Guerlesquin, M., Henry, C.P., 1996. Are Characeae able to indicate the origin of groundwater in former river channels? Vegetation 125, 207–222.

Carbiener, R., Tremolières, M., Mercier, J.L., Ortscheit, A., 1990. Aquatic macrophyte communities as bioindicators of eutrophication in calcareous oligosaprobe stream waters (upper Rhine plain, Alsace).

Vegetation 86, 71–88.

Casanova, M.T., Brock, M.A., 1991. Charophyte germination and establishment from the seed bank of an Australian temporary lake. Aquat. Bot. 36, 247–254.

Casanova, M.T., Brock, M.A., 1999. Life histories of charophytes from permanent and temporary wetlands in eastern Australia. Aust. J. Bot. 3, 383–397.

Castella, C., Amoros, C., 1984. Distribution of Characeae in backwaters of the Rhone River and the Ain River and ecological signification. Cryptogam. Algol. 5, 127–139.

Chessel, D., Gimaret, C., 1997. Analyse Non symétrique des Correspondances. ADE-4/Topic documentation 2.8.

30p. http://pbil.unir-lyon1.fr/ADE-4/

Corillion, R., 1957. Les Charophycées de France et de l’Europe occidentale. Bull. Soc. Sci. Bretagne 32, 1–499.

Corillion, R., 1975. Flore et végétation du Massif Armoricain. Flore des Charophytes (Characées) du Massif Armoricain et des contrées voisines d’Europe occidentale. Jouve Publishers, Paris, 216 pp.

Corillion, R., 1986. Régressions floristiques dans le Bas-Maine: le cas des algues characées. Bull. Soc. May. Sci., Laval, 1982–84, 69–83.

Daget, P., Godron, M., 1982. Analyse fréquentielle de l’écologie des espèces dans les communautés. Masson, Paris, 163 pp.

Forsberg, C., 1965. Environmental conditions of Swedish Charophytes. Symp. Bot. Upsal. 18, 3–67.

Gimaret-Carpentier, C., Chessel, D., Pascal, J.P., 1998. Non-symmetric correspondence analysis: an alternative for species occurrences data. Plant Ecol. 138, 97–112.

Hutchinson, G.E., 1975. A Treatise on Limnology. 3. Limnological Botany. Wiley, New York, 660 pp.

Lachavanne, J.B., Wattenhoffer, R., 1975. Les Macrophytes du Léman. Conservatoire Botanique de Genève, Genève, 147 pp.

Lang, G., 1968. Vegetationsänderungen am Bodenseeufer in den letzten hundert Jahren. Schr. Ver. Gesch. Bodensee u. Umgeb. 86, 295–319.

Lauro, N., D’Ambra, L., 1984. L’analyse non-symétrique des correspondances. In: Diday, E. (Ed.), Data Analysis and Informatics III. Elsevier, Amsterdam, pp. 433–446.

Moore, J., 1986. Charophytes of Great Britain and Ireland, Handbook Number 5. Botanical Society of the British Isles, London, 114 pp.

Moss, B., Stansfield, J., Irvine, K., Perrow, M., Phillips, G., 1996. Progressive restoration of a shallow lake: a 12-years experimental in isolation, sediment removal and biomanipulation. J. Appl. Ecol. 33, 71–86.

Odland, A., Birks, H.J.B., 1999. The altitudinal gradient of vascular plant richness in Aurland, western Norway.

Ecography 22, 548–566.

Ozimek, T., Kowalcezewski, A., 1984. Long-term changes of the submerged macrophytes in eutrophic Lake Mikolajskie (north Poland). Aquat. Bot. 19, 1–11.

Pereyra-Ramos, E., 1981. The ecological role of Characeae in the lake littoral. Ekol. Pol. 29, 167–209.

Proctor, V.W., 1999. Charophytivory, playas y papalotes, a local paradigm of global relevance. Aust. J. Bot. 47, 399–406.

Rahbek, C., 1995. The elevational gradient of species richness: a uniform pattern? Ecography 18, 200–205.

Rosenzweig, M.L., 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge, 436 pp.

Rosenzweig, M.L., Sandlin, E.A., 1997. Species diversity and latitudes: listening to area’s signal. Oikos 80, 172–176.

Rosenzweig, M.L., Leitner, W.A., Gray, J.S., Collins, C.D., Aukeman, J., Ugland, K.I., Poore, G.C.B., 1998.

(1, 1). . .infinity: Estimating Species Diversity. University of Arizona, Tucson, USA, 39 p.

Scheiner, S.M., Rey-Benayas, J.M., 1994. Global patterns of plant diversity. Evol. Ecol. 8, 331–347.

Schmieder, K., 1991. Veränderungen der submersen Makrophytenvegetation des Bodensee-Untersees innerhalb der vergangen 20 Jahre als Spiegelbild der trophischen Entwicklung. Verh. Ges. Ökol. 20, 101–122.

Schwarz, A.-M., Hawes, I., 1997. Effects of changing water clarity on characean biomass and species composition in a large oligotrophic lake. Aquat. Bot. 56, 169–181.

Simons, J., Nat, E., 1996. Past and present distribution of stoneworts (Characeae) in The Netherlands.

Hydrobiologia 340, 127–135.

(18)

Soberón, J., Llorente, J., 1993. The use of species accumulation functions for the prediction of species richness.

Conserv. Biol. 7, 480–488.

Van den Berg, M., Coops, M.S., Meijer, M.-L., Scheffer, M., Simons, J., 1998. Clear water associated with a dense Chara vegetation in the shallow and turbid Lake Veluwemeer, The Netherlands. In: Jeppesen, E., Søndergaard, M., Christoffersen, K. (Eds.), The Structuring Role of Submerged Macrophytes in Lakes. Springer, New York, pp. 339–352.

Wattenhoffer, R., 1984. Macrophytes. In: Commission Internationale pour le Léman (Ed.), La Protection des Eaux du Léman contre la Pollution, Synthèse 1957–1982. Lausanne, Switzerland, pp. 337–353.

Wiegleb, G., 1978. Untersuchungen über den Zusammenhang zwischen hydrochemischen Umweltfaktoren und Makrophytenvegetation in stehenden Gewässern. Arch. Hydrobiol. 83, 443–484.

Wood, R.D., Imahori, K., 1959. Geographic distributions of Characeae. Bull. Torrey Bot. Club. 86, 172–183.