SEARCH

SEARCH BY CITATION

Keywords:

  • extinction;
  • Gastropoda;
  • mark–recapture;
  • multi-species approach;
  • recolonization

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • 1
    The fragmentation of natural habitats is generally considered to be a major threat to biodiversity. Different species may respond differently to habitat fragmentation, depending on species-specific traits such as body size, dispersal ability, mating system, and habitat requirement.
  • 2
    The population sizes, extinction and recolonization frequencies of six naturally occurring land snail species (Cochlicopa lubrica, Vertigo pygmaea, Pupilla muscorum, Punctum pygmaeum, Helicella itala, and Trichia plebeia) were examined over 3 years in an experimentally fragmented nutrient-poor, calcareous grassland in the northern Swiss Jura mountains using a mark–recapture technique. Fragments of different size (0·25 m2, 2·25 m2, and 20·25 m2) were isolated by a 5-m wide strip of frequently mown vegetation. Control plots of corresponding size were situated in adjacent undisturbed grassland.
  • 3
    Experimental grassland fragmentation influenced the population size in all snail species except H. itala, which is the species with the biggest shell and it is also active under mild conditions in winter. However, fragmentation affected different species to a different extent.
  • 4
    Extinction (= disappearance from a plot) frequency increased with time, decreasing population size and decreasing plot size in all species. Large populations had a lower extinction probability than small populations. Fragmentation increased the probability of extinction, which also differed among snail species. The effect of plot size on extinction probability was still significant even after the effect of population size had been taken into account.
  • 5
    Fragments and control plots did not differ in recolonization frequencies when all six species were considered. However, fragmentation influenced recolonization frequency when the two species with large shells (H. itala and T. plebeia) were excluded from the analysis.
  • 6
    Our study shows that small-scale grassland fragmentation affects different land snail species to a different extent. This finding strengthens the claim for multi-species approaches to obtain general predictions of fragmentation impact.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

One of the most dramatic landscape changes during the 20th century in Europe has been the reduction and fragmentation of habitats such as semi-natural grasslands (WallisDeVries et al. 2002; Cremene et al. 2005). Fragmentation reduces the area suitable for organisms and leads to isolation and decreased size of remnant populations in plants and animals, which are exposed to an increased risk of local extinction (Saunders, Hobbs & Margules 1991; Saccheri et al. 1998). The disadvantages suffered by small populations involve greater sensitivity to demographic stochasticity (Harrison 1991; Holsinger 2000) and reduced genetic variation (Krauss et al. 2004). Because of isolation of fragments, recolonization following local extinction may be low (Hanski 1994). Furthermore, effects of habitat fragmentation on individual species or populations may lead to the disruption of biotic interactions such as pollination, seed dispersal or predation and hence, can affect species with previously stable populations (Groppe et al. 2001; Goverde et al. 2002; Lennartsson 2002; Stoll et al. 2006). Finally, habitat characteristics may change in fragments through the increased impact of external factors such as physical disturbances, and changed temperature and moisture conditions (Dolt, Goverde & Baur 2005; Aguilar et al. 2006). Thus, habitat fragmentation can influence an entire suite of processes, ranging from individual behaviour through population dynamics to ecosystem fluxes.

Due to isolation, colonization by new species or recolonization by previously present species is expected to be lower in fragments than in sections of continuous habitat of the same area (MacArthur & Wilson 1967; Hanski 1999). Altered extinction and colonization rates in fragments may subsequently lead to shifts in community composition and result in a net decline of species richness (Robinson et al. 1992). The fate of remnant populations in a fragment will depend on the size of the fragment, the degree of isolation, the quality of the matrix habitat, the size of the remnant population, and on species characteristics such as body size, mating system, dispersal ability and longevity (Shaffer 1981). Hence, the response to habitat fragmentation is expected to be species specific (Wolff, Schauber & Edge 1997).

Multi-species approaches allow a more general assessment of fragmentation impact to the community involved but may also yield more equivocal results (Saunders et al. 1991). It has been suggested that further research should be dedicated to the identification of traits that allow to predict the species’ response to fragmentation (Davies, Margules & Lawrence 2000). Theory predicts that generalists should be less influenced by habitat fragmentation than specialists (Gibb & Hochuli 2002; Joshi et al. 2006). Species of low trophic rank should be less affected than species of high trophic rank (Davies et al. 2000). Similarly, species with a pronounced dispersal ability should be less affected than species with a low dispersal ability (Davies & Margules 1998). Most of the common species have broad niches and can exist in disturbed habitat or matrix habitat. Consequently, these species are less affected by the isolation following fragmentation.

Most of the studies that examined the response of individual species to habitat fragmentation were dealing with birds in forests (for reviews see Opdam 1991; Turner 1996). However, very few studies have simultaneously tested species-specific responses to habitat fragmentation in several species in experiments with proper replication and controls (Debinski & Holt 2000; Braschler & Baur 2003, 2005). Furthermore, herbivores and decomposers have received little attention with respect to habitat fragmentation, although they influence primary production in most habitats and thus control the nutrient supply to a large extent (Margules, Milkovits & Smith 1994; Didham et al. 1996). In the present paper, we examine the effects of experimental small-scale grassland fragmentation on population size, extinction and recolonization frequencies in six land snail species, which are typical inhabitants of nutrient-poor, dry calcareous grasslands (Boschi & Baur 2007a,b). As the snail species differ in shell size, breeding system, dispersal ability and seasonal activity, we expected species-specific extinction and recolonization frequencies.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

study site and experimental design

The fragmentation experiment was carried out in a nutrient-poor, dry calcareous grassland on a southwest-facing slope situated near Nenzlingen, 13 km south of Basel, in the northern Swiss Jura mountains (47°27′N, 7°34′E; altitude 500 m a.s.l.). Originally covered by beech forest, the grassland has been grazed by cattle for many centuries, leading to the characteristic Mesobromion alliance (Ellenberg 1988). Site descriptions can be found in Baur et al. (1996) and Zschokke et al. (2000).

The experimental fragmentation of the grassland was created in April 1993 by mowing the vegetation around the experimental fragments. One experimental unit, called block, measured 32 × 29 m and contained two small (0·5 × 0·5 m), one medium (1·5 × 1·5 m) and one large fragment (4·5 × 4·5 m), all of them separated by a 5-m wide strip of mown vegetation, as well as the corresponding control plots, which were mirror-symmetrically arranged and surrounded by undisturbed vegetation (Fig. 1). Within each block, the positions of the different sizes of fragment-control plot pairs as well as the fragment and control halves were randomized. We used five blocks with a total of 20 fragments (5 large, 5 medium and 10 small fragments) and 20 control plots of the corresponding sizes. The blocks were part of a larger study area (2 ha) that was enclosed by a fence to exclude large herbivores. The experimental fragmentation was maintained from April 1993 to November 1999 by frequently mowing (6–12 times per year) the area between the fragments in the period from March to October. The entire experimental area was mown in November every year to prevent succession. This type of fragmentation is reversible, but reduces dispersal in numerous invertebrate species (Zschokke et al. 2000; Goverde et al. 2002). In December–February, the grassland is covered by snow during short periods.

image

Figure 1. Layout of one block (out of five) set up in the field in April 1993. Each block contained four small (0·25 m2), two medium (2·25 m2) and two large plots (20·25 m2). Half of each block served as control, the other half was experimentally fragmented by frequent mowing between plots. Isolation distance between plots and between plots and continuous vegetation was 5 m. The position of each pair of plots of a given size as well as the position of control and treatment side was randomly chosen in each of the five blocks.

Download figure to PowerPoint

snail community

A total of 16 snail species was recorded in this grassland. Here we focus on six snail species that are typical inhabitants of dry grasslands and occurred in moderate to high numbers [Cochlicopa lubrica (O.F. Müller 1774), Vertigo pygmaea (Draparnaud 1801), Pupilla muscorum (Linneaus 1758), Punctum pygmaeum (Draparnaud 1801), Helicella itala (Linnaeus 1758) and Trichia plebeia (Draparnaud 1805)]. The remaining 10 species were recorded in very small numbers (total abundance < 5%) and therefore not further considered.

C. lubrica has an elongate oval shell which measures 6·5 mm in height. Individuals reach sexual maturity after 2–3 years and live up to 4 years. C. lubrica is able to reproduce by self-fertilization (Armbruster & Schlegel 1994). V. pygmaea has an ovoid 1·7–2·2 mm high shell. P. muscorum has a cylindrical ovoid 3·3 mm high shell and is ovoviviparous. Punc. pygmaeum has a discoidal shell with a diameter of 1·2–1·5 mm. It reaches sexual maturity at an age of 0·8 years and is able to reproduce by self-fertilization (Baur 1989). H. itala has a flat white shell with a diameter of 12–15 mm (Oggier 1998). Individuals become sexually mature at an age of 1·5 years, and adults live another 1–1·5 years (Frömming 1954). H. itala reproduces by outcrossing, and is active under favourable conditions during winter (Oggier 1998). T. plebeia has a slightly globular shell with a diameter of 7–10 mm. H. itala and T. plebeia belong to the Helicidae, whereas the other four species belong to different families.

mark–recapture procedure

We estimated the population sizes of the six snail species in each of the 20 fragments and 20 control plots in May 1994 (= 1 year after initiation of the experimental fragmentation), 1995 and 1996. Searches for snails were conducted only under conditions favourable for snail activity (moist vegetation and between 06.00 h and 11.00 h). We standardized the searching effort to 16 min per square metre (4 min. in small fragments and control plots, 36 min. in medium plots, and 324 min. in large plots). To avoid trampling of snails and vegetation in the plots, we used a 5-m long ladder with a 50-cm high wooden support at each end as working platform.

All snails found (juveniles and adults) were determined to the species, counted, marked on their shells with a minute dot of car lacquer and released in the same plot within 3–4 h. Seven and 14 days after marking, the plots were searched again for snails with the same searching effort. All snails found were determined to the species, checked for colour marks, counted and released in the same plot. Population size of the six snail species in each fragment and control plot and corresponding standard errors were estimated with Bailey's modification of Petersen's estimate (Seber 1982).

A total of 3111 individual snails were marked in the course of the 3-year study. V. pygmaea was by far the most frequent species (59·8%), followed in decreasing frequency by Punc. pygmaeum (10·6%), T. plebeia (10·5%), H. itala (7·8%), P. muscorum (6·8%) and C. lubrica (4·5%).

statistical analyses

We used r (R Development Core Team 2007, version 1·7·1) for all analyses. Population sizes from mark–recapture estimates were analysed as count data using an appropriate generalized linear model (GLM with poisson errors and log-link function). The GLM ensures that fitted values are positive and considers the fact that count data are integers with variances equal to their means (Crawley 2005). Reciprocal standard errors of the population size estimates were used as weights. This minimized the problem of inflated error probabilities due to overdispersion. To take the remaining overdispersion into account, mean deviances (i.e. deviance changes divided by the degree of freedom) caused by removing terms were corrected using mean deviance of the residual as scale parameter (Crawley 2005). Mean deviances were divided by the scale parameter to obtain approximately F-distributed ratios to assess error probabilities. Three-way interactions between period, plot size and fragmentation were in no case significant and therefore excluded from the analyses.

Year-to-year variation was accounted by treating year as categorical variable. Treating year as continuous variable did not substantially change any of the reported P values. Finally, if density (i.e. number of individuals per square metre) was analysed using analyses of variance with log(density+1) as dependent variable (Table S1), results for the four small snail species were almost identical to the results obtained from the GLM models. However, plot size was less significant for most species and never significant if only small- and medium-sized plots were analysed. Fragmentation effects were significant for all four smaller species and not significant for the two bigger species.

To examine possible competitive interactions among and within species, we correlated population size (log-transformed) in 1994 with growth rate from 1995 to 1996 for all plots in which the corresponding species occurred.

Extinctions (i.e., disappearance of a population from a fragment or control plot) and recolonizations were analysed as binary response variables using GLM's with binomial error distributions and logit link functions. Actual extinctions can only occur in previously occupied plots. Consequently, the number of potential extinctions is usually smaller than the number of experimental plots. Similarly, actual colonizations can only occur in previously unoccupied plots. This approach was justified because we estimated the probability of species detection given presence to be 1·0, thus precluding the need to use methods accounting for detection probability (e.g. MacKenzie et al. 2003).

For extinctions in the two periods (i.e. 1994–95 and 1995–96), population size in the previous year (1994 and 1995, respectively) was used as a unique measurement of the experimental unit (fragment and control plot). This leads to a logistic analysis of covariance with extinction events as binary response, population size as continuous and period, species, plot size and fragmentation as categorical explanatory variables. In these models, the issue of overdispersion does not arise. No unique measurements of the experimental units were available for recolonizations of previously unoccupied plots. Therefore, recolonization events were analysed as binomial variables (i.e. potential events giving the binomial total and actual events giving observed recolonizations) using GLM's with binomial error distributions and logit link functions. All higher order interactions with P > 0·1 were excluded from the analyses.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

population size

Experimental fragmentation influenced population size in all species except H. itala (Table 1, Fig. 2). Furthermore, plot size affected population size in all six species (Table 1). Not surprisingly, population size increased with plot size (data not shown). Moreover, the population sizes of four species differed significantly among years, irrespective of experimental fragmentation and plot size (Table 1). Population sizes decreased from 1994 to 1996 in C. lubrica, P. muscorum, H. itala and T. plebeia (Fig. 2).

Table 1.  Analyses of deviance (GLM with Poisson-distributed errors and log link) for the effects of year (1994, 1995, 1996), plot size (large, medium and small) and experimental fragmentation on population size of six land snail species. Population size was estimated using Bailey's modification of Petersen's mark–recapture index and reciprocal standard errors of the estimates were used as weights
SpeciesSource (in order of fitting)d.f.Deviance*Mean devianceRatioP
  • *

    Change in deviance by removing source term.

  • Ratios are corrected for overdispersion by dividing mean deviances through an empirical scale parameter (i.e. mean deviance of the residual term) and using the F-distribution with degrees of freedom (d.f.) of the source term and the residual degrees of freedom (e.g. 2, 102 for year, respectively) to find the error probabilities (P).

Cochlicopa lubricaBlock411·22·82·10·086
Year227·713·810·4< 0·001
Plot size2145·272·654·5< 0·001
Fragmentation110·410·47·80·006
Year × plot size43·70·90·70·600
Year × fragmentation22·61·31·00·380
Plot size × fragmentation27·23·62·70·071
Residual1021361·3  
Vertigo pygmaeaBlock415·73·92·20·079
Year24·52·21·20·296
Plot size2571·4285·7157·3< 0·001
Fragmentation173·773·740·6< 0·001
Year × plot size426·86·73·70·008
Year × fragmentation20·90·40·20·783
Plot size × fragmentation21·30·60·40·705
Residual102185·21·8  
Pupilla muscorumBlock420·75·22·70·033
Year211·75·93·10·049
Plot size264·532·217·1< 0·001
Fragmentation121·321·311·30·001
Year × plot size40·70·20·10·983
Year × fragmentation20·10·10·00·967
Plot size × fragmentation24·22·11·10·333
Residual102192·51·9  
Punctum pygmaeumBlock4 28·97·24·10·004
Year27·53·72·10·126
Plot size2292·9146·483·1< 0·001
Fragmentation113·813·87·80·006
Year × plot size419·84·92·80·029
Year × fragmentation22·51·20·70·499
Plot size × fragmentation22·51·30·70·493
Residual102179·71·8  
Helicella italaBlock414·73·72·60·038
Year272·836·426·3< 0·001
Plot size2204·1102·173·6< 0·001
Fragmentation11·01·00·70·397
Year × plot size412·43·12·20·070
Year × fragmentation2 0·80·40·30·752
Plot size × fragmentation21·60·80·60·565
Residual102141·41·4  
Trichia plebeiaBlock426·26·53·80·006
Year219·29·65·60·005
Plot size2214·3107·262·4< 0·001
Fragmentation110·210·26·00·016
Year × plot size43·50·90·50·727
Year × fragmentation21·40·70·40·671
Plot size × fragmentation24·82·41·40·249
Residual102175·21·7  
image

Figure 2. Population size (mean ± 1 SE, log-scale) of six snail species in small (inline image), medium (inline image) and large (inline image) fragments and corresponding control plots (open symbols) over 3 years (i.e. 1994, 1995, and 1996).

Download figure to PowerPoint

Considering single species, C. lubrica had smaller populations in fragments than in control plots (Table 1). Populations were smaller in large fragments in 1996 and in all 3 years in small fragments. In 1996, it was absent in small fragments. V. pygmaea had smaller populations in small and medium fragments in all 3 years. P. muscorum had smaller populations in large fragments than in the corresponding control plots in 1994, but the differences in population size between fragments and control plots decreased in the succeeding years. In small plots, differences between fragments and control plots did not change across years. Similar to C. lubrica, P. muscorum was not found in small fragments in 1996. Punc. pygmaeum populations were smaller in large fragments than in control plots in 1996. In 1994, Punc. pygmaeum was only present in small fragments but not in control plots. In 1995, however, smaller populations were observed in small fragments than in control plots. In H. itala, fragmentation effects were not significant (Table 1). Populations of T. plebeia were bigger in medium fragments compared to control plots in all 3 years. In 1996, this was also the case in small plots.

There was no correlation between the population size of a given snail species and the growth rate of any other species, suggesting no interspecific interactions among species. However, in H. itala, growth rate from 1995 to 1996 was negatively correlated (r = –0·75, t = 3·0, d.f. = 7, P = 0·021) with population size in 1994, indicating intraspecific competition.

extinctions

Extinction events occurred less frequently in the first (1994–95) than in the second period of the experiment (1995–96; Table 2). Extinction frequencies were higher in fragments than in control plots (Table 3), especially in small plots in the first period (64·3% vs. 14·3%, respectively), but also in large plots in the second period (17·9% vs. 6·9%, respectively). In contrast, extinction frequencies in large fragments were almost identical to those in control plots in the first period. Furthermore, extinction frequency decreased with increasing plot size in both periods (Table 2).

Table 2.  Extinction events (indicated by number of actual/number of potential extinctions) in six land snail species over two periods (1994–95 and 1995–96) in experimental fragments and control plots of various sizes (small, medium and large)
PeriodSpeciesFragmentsControl plots
SmallMediumLargeSmallMediumLarge
  1. Note: actual and potential events (see Material and methods) are separated by a slash.

1994–95Cochlicopa lubrica1/10/30/50/30/40/5
Vertigo pygmaea1/21/40/50/90/50/5
Pupilla muscorum3/30/11/41/51/50/5
Punctum pygmaeum1/10/10/50/00/20/5
Helicella itala1/42/40/53/51/31/5
Trichia plebeia2/30/50/50/61/50/5
Total9/143/181/294/283/241/30
Frequency (percentage)64·316·73·414·312·53·3
1995–96C. lubrica1/13/42/54/44/51/5
V. pygmaea1/60/40/50/100/50/5
P. muscorum1/13/41/34/52/40/5
P. pygmaeum2/21/50/53/81/50/5
H. itala4/42/32/52/31/21/4
T. plebeia2/41/50/57/90/40/5
Total11/1810/255/2820/398/252/29
Frequency (percentage)61·140·017·951·332·06·9
Table 3.  Analyses of deviance (logistic ancova with binomial errors and logit link) for the effects of period (1994–95 and 1995–96), species, plot size (small, medium and large) and experimental fragmentation on extinction probabilities of six land snail species. Population sizes in 1994 and 1995 were used as covariable for extinction events in the subsequent period
Source (in order of fitting)d.f.Deviance changeP
  1. Notes: Population size might be fitted before year. This does, however, not change the significances of these terms (inline image = 89·4, P < 0·001 and inline image = 8·1, P < 0·001, for population size and year, respectively) and deviance changes of subsequent terms are not affected.

Block34·00·400
Period116·2< 0·001
Population size181·7< 0·001
Species529·7< 0·001
Plot size214·30·001
Fragmentation18·40·004
Period × species515·10·010
Population size × fragmentation21·80·407

Fragmentation significantly increased the probability of extinction (Table 3). Furthermore, population size in a given year significantly influenced the probability of extinction in the succeeding year. The extinction probability differed among species and between periods (Table 3). The effect of plot size on extinction probability was still significant after the effect of population size had been taken into account.

In all six species, the extinction probability depended on population size in both periods (Fig. 3). Large populations had lower extinction probabilities than small populations irrespective of period and plot size. Extinction probabilities were lower in the first than in the second period, except for V. pygmaea which generally had a low extinction probability. Extinction probability of Punc. pygmaeum was most affected by experimental fragmentation, whereas differences between periods were most pronounced in C. lubrica. Overall, 75 (97·4%) of the 77 populations that had disappeared from a fragment or control plot consisted of 10 or fewer individuals in the preceding year. The effect of plot size on extinction probability was significant even after population size had been taken into account (cf. Table 3). Consequently, extinction probabilities at a given population size were higher in small than in medium plots (Fig. 3). For example, a population of Punc. pygmaeum with two individuals in 1995 went extinct from 1995 to 1996 in small fragments with a probability of 0·69 (± 0·13) compared to 0·35 (± 0·14) in medium-sized fragments. Corresponding extinction probabilities in control plots were 0·38 (± 0·13) in small control plots but 0·13 (± 0·07) in medium ones. Fitted extinction probabilities for all species in fragments and control plots in both periods are given in the Table S2. Finally, if we pooled extinction events at the level of half blocks (i.e. control vs. fragmented half), one extinction event occurred for H. itala in a control and one for P. muscorum in a fragmented half from 1994–95. In the second period (1995–96), no extinctions occurred in control halves. However, C. lubrica and H. itala went extinct in two and P. muscorum in one of the five fragmented half blocks.

image

Figure 3. Extinction probabilities of six land snail species from 1994–95 (squares, dashed lines) and 1995–96 (circles, continuous lines) in fragments (closed symbols, thick lines) and control plots (open symbols, thin lines) of small, medium and large size as a function of population size in the previous year. The lines are fitted values (± 1 SE at population size 3 in 1994 and 2 in 1995, respectively) from the full logistic ancova (cf. Table 3) with population size in the previous year as covariable. Symbols and error bars are slightly displaced in order to avoid cluttering. Numbers next to symbols indicate the number of coinciding data points.

Download figure to PowerPoint

recolonization

Recolonization events occurred more frequently in the first than in the second period (Table 4). The recolonization frequency also differed among species and was related to plot size (Table 5). More recolonizations were observed in medium fragments than control plots in both periods (Table 4). Because almost all large plots were occupied in both periods, the potential to observe recolonizations in large plots was low and no event was observed. In addition, there was a significant interaction between period and plot size. Frequency of recolonizations was lower in small fragments compared to control plots in the first but higher in the second period. Fragments and control plots did not differ in colonization frequencies when all six species were considered (Table 5). However, when the two helicid species with large shells (H. itala and T. plebeia) were excluded from the analysis, the fragmentation effect became significant (c1 = 5·7, P = 0·017). Recolonization events for the four non-helicid species were less frequent in small and medium fragments (39% and 82%, respectively) than in control plots (48% and 100%, respectively) in the first period. In the second period, recolonization frequencies did not differ between small fragments and control plots (23% each) but were higher in medium fragments (33%) than in the corresponding control plots (0%).

Table 4.  Recolonization events (indicated by number of actual/number of potential recolonizations) in six land snail species over two periods (1994–95 and 1995–96) in experimental fragments and control plots of various sizes (small, medium and large)
PeriodSpeciesFragmentsControl plots
SmallMediumLargeSmallMediumLarge
  1. Note: actual and potential events (see Material and methods) are separated by a slash.

1994–95Cochlicopa lubrica1/91/20/01/71/10/0
Vertigo pygmaea5/81/10/01/10/00/0
Pupilla muscorum1/73/40/11/50/00/0
Punctum pygmaeum2/94/40/08/103/30/0
Helicella itala1/61/10/01/50/20/0
Trichia plebeia3/70/00/03/40/00/0
Total13/4610/120/115/324/60/0
Frequency (percentage)28·383·30·046·966·70·0
1995–96C. lubrica0/90/10/02/60/00/0
V. pygmaea3/41/10/00/00/00/0
P. muscorum0/90/10/20/50/10/0
P. pygmaeum4/80/00/01/20/00/0
H. itala0/60/20/01/70/30/1
T. plebeia4/60/00/00/10/10/0
Total11/421/50/24/210/50/1
Frequency (percentage)26·220·00·019·00·00·0
Table 5.  Analyses of deviance (GLM with binomial errors and logit link) for the effects of period (1994–95 and 1995–96), species, plot size (small, medium and large) and experimental fragmentation on recolonization probabilities of six land snail species
Source (in order of fitting)d.f.Deviance changeP
Block33·10·541
Period19·30·002
Species537·7< 0·001
Plot size29·60·008
Fragmentation12·20·141
Period × plot size210·30·006

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Habitat fragmentation is expected to result in increased rates of local extinction and in decreased patch colonization, both of which can result in population declines greater than expected from habitat loss alone (Hanski & Ovaskainen 2000). Our study showed that experimental habitat fragmentation increased the extinction risk in land snails. However, the six land snail species examined differed in their response to the experimental grassland fragmentation. Recolonization was influenced by fragmentation in four of the six species. Furthermore, fragmentation decreased the population size in all species, except H. itala. The surrounding matrix, the frequently mown grassland, provides less favourable conditions for most snail species than the undisturbed vegetation (Bruno Baur, unpublished data). Similarly, snail abundance and species richness were found to decrease with increasing grazing intensity by livestock in these grasslands (Boschi & Baur 2007a,b). In the present study, the matrix may have functioned as partial barrier for some snail species. In general, larger animals are able to move larger distances than smaller ones and thus should be better colonizers. Field experiments showed that H. itala is able to move more than 6 m in 1 week or 10 m in 2 weeks (Oggier 1994), and thus can travel much farther than the isolation distance (5 m) of the present study. Most important, however, H. itala is active during wet periods in winter (Oggier 1998), whereas individuals of the other species hibernate in soil. We maintained the artificial fragmentation from March to October but not from December to February. During this period, individuals of H. itala might have crossed the isolation area.

The small-scale fragmentation in our field experiment significantly changed the above- and below-ground plant biomass and plant species composition (Dolt et al. 2005; Joshi et al. 2006). Fragments contained more above- and below-ground plant biomass than control plots. The increase in above-ground plant biomass was partly due to an increased density of plants, which in turn was a result of reduced competition for light at the edge of fragments, and partly due to altered plant species composition in fragments (Dolt et al. 2005). Land snail activity is constrained by air humidity and thus temporally limited in dry grassland (Oggier, Zschokke & Baur 1998). Because of the increased plant biomass, more dew occurs in fragments than in control plots and accordingly, high air humidity is retained longer in fragments. This may allow the snails to be longer active in fragments when the temperature rises in the morning. Fragments are, however, influenced by the slightly higher average air temperature and the more pronounced temperature fluctuations measured in the surrounding mown area (Zschokke et al. 2000). These external influences alter the micro-climatic conditions in the fragments, especially in the edge zone. Thus, apart of the partial isolation, fragments also differed in habitat quality from the control plots.

Independent of fragmentation, population size increased with plot size in all species examined. This is not surprising, as larger areas usually harbour larger populations than smaller areas (MacArthur & Wilson 1967; Hanski 1999). Furthermore, population size increased from 1994 to 1996 in V. pygmaea and Punc. pygmaeum, but decreased in the same period in C. lubrica, P. muscorum, H. itala and T. plebeia. Species-specific differences in temporal changes of population size could be explained by a combination of different factors. V. pygmaea and Punc. pygmaeum, the species with an increase in population size, have both tiny shells (1·5–2·2 mm in adults) and short generation time (< 1 year), whereas the remaining four species showing a general decline in population size have larger shells (3·3–15 mm) and longer generation time (> 1 year). Individuals of snail species with tiny shells may experience the size of a fragment in relation to their own body size in another way than snails with large shells. Moreover, the short generation time of these species allow them to react more rapidly to altered environmental conditions than species with long generation time.

Our study covered the first three years of the fragmentation experiment. Due to the recent fragmentation immigration from the surrounding, now less suitable habitat (mown area), could increase the size of the populations in fragments over a short period, which decreases the risk of immediate extinction. Similar effects were observed in remnants of Amazonian rainforest shortly after logging (Lovejoy et al. 1986). This could partly explain the population size decline in the four species with large shells and long generation time. However, the population size of H. itala, C. lubrica, and – to a minor extent – of P. muscorum also decreased in control plots. Summer and autumn 1994 (the first year of our study) were extraordinary warm, while the winter 1994/95 was rather cold (Bruno Baur, unpublished data). The prevailing weather conditions are of key importance for land snail reproduction and survival (Baur & Raboud 1988). Thus, the extreme weather conditions in the first year could contribute to the observed differences in the population dynamics of the snail species examined.

All species except V. pygmaea were found more frequently in large plots than in smaller ones. V. pygmaea was the most abundant species and thus present in almost all plots. This is in agreement with the findings from other studies on abundant species (Hanski 1994; Vos & Chardon 1998).

Our study showed that the extinction probability highly depended on the population size in all species examined. Large populations had a lower extinction probability than small populations. The generally smaller population size in fragments than in control plots of corresponding size contributed to the higher extinction probability in fragments. Furthermore, the extinction probability was negatively related to plot size, even after population size had been accounted for. This suggests that – in addition to population size – plot size-related factors like the extent of edge effects negatively affected population survival in fragments.

The most mobile species (H. itala and T. plebeia) occurred at relatively low densities in the grassland. It is therefore possible that the few individuals found in a small fragment had left this plot between two surveys. This would be recorded as an extinction event. In general, small populations are expected to be exposed to a high risk of extinction due to increased demographic and genetic stochasticity. However, pulmonate land snails possess traits that reduce the risk of extinction due to a low number of available mating partners: they are hermaphrodites, have long-term sperm storage, and self-fertilization may occur in some species (Baur 1998). Thus, species-specific differences in extinction rate could partly be explained by a differential expression of these traits and the breeding system. For example, C. lubrica, Punc. pygmaeum and most probably V. pygmaeum are able to reproduce by self-fertilization which guarantees reproduction even under extremely low population densities.

Our study showed that the recolonization frequency was species-specific and related to plot size. Because almost all large plots were constantly occupied by all species, extremely few (re)colonization events could be observed in these plots. However, considering only the four non-helicid species, recolonizations occurred less frequently in small and medium fragments than in control plots, indicating that the frequently mown area functions as a partial barrier (see above).

Our results support the theory of metapopulation biology (Kareiva & Wennergren 1995; Hanski & Ovaskainen 2000). We found species-specific fragmentation effects on population size and probabilities of extinction and (re)colonization within a 3-year period of experimental grassland fragmentation. The review by Debinski and Holt (2000) showed that effects of fragmentation become stronger with time; a pattern also observed in our field experiment with perennial plants recorded over 7 years (Joshi et al. 2006). In the long term, fragmentation may lead to higher extinction not only due to increased demographic and genetic stochasticity of populations, but also via indirect effects. These include altered abiotic and biotic conditions and the disruption of biological interactions (Groppe et al. 2001; Braschler & Baur 2003; Braschler, Lampel & Baur 2003).

Interspecific competition could increase the extinction probability and reduce the rate of recolonization (Bengtsson 1993). Competitive interactions have commonly been assumed to be weak among herbivorous and decomposing terrestrial gastropod species. There is, however, experimental evidence demonstrating the importance of interference competition in helicid snails (Cowie & Jones 1987). Furthermore, both exploitation competition and interference by mucus trails have been demonstrated within and among specialized lichen-feeding land snail species (Baur & Baur 1990). In the present study, competitive interactions could be expected between H. itala and T. plebeia, two helicid species with an overlap in diet and slightly different periods of activity, but not among the small-sized decomposers (C. lubrica, V. pygmaea and Punc. pygmaeum) and the algae- and lichen-feeding P. muscorum. We could not find any evidence for interspecific competition in the snail species examined. Most interestingly, however, population growth of H. itala was negatively correlated with population size in the preceding year, indicating intraspecific competition in this species.

Our study provides experimental evidence for an increased extinction risk in small snail populations living in distinct grassland patches (fragments). Demographic stochasticity may negatively affect the growth rate of these populations and frequently lead to their disapperance from a plot (Shaffer 1981; Holsinger 2000). However, our results also indicate that demographic processes may be amplified both by changes in habitat structure in fragments which lead to a decreased habitat quality and by external effects influencing the edge zone of fragments. Our multi-species approach revealed that different snail species reacted differently to small-scale habitat fragmentation, depending on their dispersal ability, shell size, longevity, mating system and habitat specificity. To predict a species’ response to fragmentation remains a difficult task, because a variety of life-history traits, changing habitat characteristics and other abiotic factors all influence to a different extent the population dynamics of gastropod species.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank S. Schüpbach, G. Hofer and numerous students for field assistance, A. Baur, Hans-Peter Rusterholz and three anonymous reviewers for comments on the manuscript and M. Kery for statisitcal advice. Financial support was received from the Swiss National Science Foundation.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Aguilar, R., Ashworth, L., Galetto, L. & Aizen, M.A. (2006) Plant reproductive susceptibility to habitat fragmentation: review and synthesis through meta-analysis. Ecology Letters, 9, 968980.
  • Armbruster, G. & Schlegel, M. (1994) The land snail species of Cochlicopa (Gastropoda: Pulmonata: Cochlicopidae): presentation of taxon-specific allozyme patterns, and evidence for high level of self-fertilization. Journal of Zoological and Evolutionary Research, 32, 282296.
  • Baur, B. (1989) Growth and reproduction of the minute land snail Punctum pygmaeum (Draparnaud). Journal of Molluscan Studies, 55, 383387.
  • Baur, B. (1998) Sperm competition in molluscs. Sperm Competition and Sexual Selection (eds T.R.Birkhead & A.P.Møller), pp. 255305. Academic Press, London.
  • Baur, B. & Baur, A. (1990) Experimental evidence for intra- and interspecific competition in two species of rock-dwelling land snails. Journal of Animal Ecology, 59, 301315.
  • Baur, B. & Raboud, C. (1988) Life history of the land snail Arianta arbustorum along an altitudinal gradient. Journal of Animal Ecology, 57, 7187.
  • Baur, B., Joshi, J., Schmid, B., Hänggi, A., Borcard, D., Stary, J., Pedroli-Christen, A., Thommen, G.H., Luka, H., Rusterholz, H.-P., Oggier, P., Ledergerber, S. & Erhardt, A. (1996) Variation in species richness of plants and diverse groups of invertebrates in three calcareous grasslands of the Swiss Jura mountains. Revue suisse de Zoologie, 103, 801833.
  • Bengtsson, J. (1993) Interspecific competition and determinants of extinction in experimental populations of 3 rockpool Daphnia species. Oikos, 67, 451464.
  • Boschi, C. & Baur, B. (2007a) The effect of horse, cattle and sheep grazing on the diversity and abundance of land snails in nutrient-poor calcareous grasslands. Basic and Applied Ecology, 8, 5565.
  • Boschi, C. & Baur B. (2007b) Effects of management intensity on land snails in Swiss nutrient-poor pastures. Agriculture, Ecosystems and Environment, 120, 243249.
  • Braschler, B. & Baur, B. (2003) Effects of experimental small-scale grassland fragmentation on spatial distribution, density and persistence of ant nests. Ecological Entomology, 28, 651658.
  • Braschler, B. & Baur, B. (2005) Experimental small-scale grassland fragmentation alters competitive interactions among ant species. Oecologia, 143, 291300.
  • Braschler, B., Lampel, G. & Baur, B. (2003) Effects of experimental small-scale grassland fragmentation on the population dynamics of aphids. Oikos, 100, 581591.
  • Cowie, R.H. & Jones, J.S. (1987). Ecological interactions between Cepaea nemoralis and Cepaea hortensis: competition, invasion, but no niche displacement. Functional Ecology, 1, 9197.
  • Crawley, M.J. (2005) Statistics: An Introduction Using r. John Wiley & Sons, Chichester, England.
  • Cremene, C., Groza, G., Rakosy, L., Schileyko, A.A., Baur, A., Erhardt, A. & Baur, B. (2005) Alterations of steppe-like grasslands in Eastern Europe: threat to regional biodiversity hotspots. Conservation Biology, 19, 16061618.
  • Davies, K.F. & Margules, C.R. (1998) Effects of habitat fragmentation on carabid beetles: experimental evidence. Journal of Animal Ecology, 67, 460471.
  • Davies, K.F., Margules, C.R. & Lawrence, J.F. (2000) Which traits of species predict populations declines in experimental forest fragments? Ecology, 81, 14501461.
  • Debinski, D.M. & Holt, R.D. (2000) A survey and overview of habitat fragmentation experiments. Conservation Biology, 14, 342355.
  • Didham, R.K., Ghazoul, J., Stork, N.E. & Davis, A.J. (1996) Insects in fragmented forests: functional approach. Trends in Ecology & Evolution, 11, 255260.
  • Dolt, C., Goverde, M. & Baur, B. (2005) Effects of experimental small-scale fragmentation on the above- and below-ground plant biomass in calcareous grasslands. Acta Oecologica, 27, 4956.
  • Ellenberg, H. (1988) Vegetation Ecology of Central Europe. Cambridge University Press, Cambridge, UK.
  • Frömming, E. (1954) Biologie der mitteleuropäischen Landgastropoden. Duncker & Humblot, Berlin, Germany.
  • Gibb, H. & Hochuli, D.F. (2002) Habitat fragmentation in an urban environment: large and small fragments support different arthropod assemblages. Biological Conservation, 106, 91100.
  • Goverde, M., Schweizer, K., Baur, B. & Erhardt, A. (2002) Small-scale fragmentation affects pollinator behaviour: experimental evidence from the bumblebee Bombus veteranus on calcareous grasslands. Biological Conservation, 104, 293299.
  • Groppe, K., Steinger, T., Schmid, B., Baur, B. & Boller, T. (2001) Effects of habitat fragmentation on choke disease (Epichloë bromicola) in the grass Bromus erectus. Journal of Ecology, 89, 247255.
  • Hanski, I. (1994) Patch-occupancy dynamics in fragmented landscapes. Trends in Ecology & Evolution, 9, 131135.
  • Hanski, I. (1999) Metapopulation Ecology. Oxford University Press, Oxford, UK.
  • Hanski, I. & Ovaskainen, O. (2000) The metapopulation capacity of fragmented landscape. Nature, 404, 755758.
  • Harrison, S. (1991) Local extinction in metapopulation context: an empirical evaluation. Biological Journal of the Linnean Society, 42, 7388.
  • Holsinger, K.E. (2000) Demography and extinction in small populations. Genetics, Demography and Viability of Fragmented Populations (eds A.G.Young & G.M.Clarke), pp. 5574. Cambridge University Press, Cambridge, UK.
  • Joshi, J., Stoll, P., Rusterholz, H.-P., Schmid, B., Dolt, C. & Baur, B. (2006) Small-scale experimental habitat fragmentation reduces colonization rates in species-rich grasslands. Oecologia, 148, 144152.
  • Kareiva, P. & Wennergren, U. (1995) Connecting landscape patterns to ecosystem and population processes. Nature, 373, 299302.
  • Krauss, J., Schmitt, T., Seitz, A., Steffan-Dewenter, I. & Tscharntke, T. (2004) Effects of habitat fragmentation on the genetic structure of the monophagous butterfly Polyommatus coridon along its northern range margin. Molecular Ecology, 13, 311320.
  • Lennartsson, T. (2002) Extinction thresholds and disrupted plant-pollinator interactions in fragmented plant populations. Ecology, 83, 30603072.
  • Lovejoy, T.E., Bierregaard Jr., R.O., Rylands, A.B., Malcolm, J.R., Quintela, C.E., Harper, L.H., Brown Jr., K.S., Powell, A.H., Powell, G.V.N., Schubart, H.O.R. & Hays, M.B. (1986) Edge and other effects of isolation on Amazon forest fragments. Conservation Biology: The Science of Scarcity and Diversity (ed. M.E.Soulé), pp. 257285. Sinauer & Associates, Sunderland, Massachusetts.
  • MacArthur, R.H. & Wilson, E.O. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton, New Jersey.
  • MacKenzie, D.I., Nichols, J.D., Hines, J.E., Knutson, M.G. & Franklin, A.D. (2003) Estimating site occupancy, colonization and local extinction when species is detected imperfectly. Ecology, 84, 22002207.
  • Margules, C., Milkovits, G.A. & Smith, G.T. (1994) Contrasting effects of habitat fragmentation on the scorpion Cercophonius squama and an Amphipod. Ecology, 75, 20332042.
  • Oggier, P. (1994) Das Ausbreitungsverhalten der Heideschnecke (Helicella itala) bei erhöhter Dichte. Taxonomie und Oekologie alpiner Landschnecken am Beispiel von Arianta und Cylindrus (eds H.Sattmann, A.Bisenberger & H.Kothbauer), pp. 4043. Naturhistorisches Museum, Wien, Austria.
  • Oggier, P. (1998) Circadian and year-round activity of the land snails Candidula unifasciata and Helicella itala in grasslands of the Swiss Jura mountains. Malakologische Abhandlungen, 19, 89101.
  • Oggier, P., Zschokke, S. & Baur, B. (1998) A comparison of three methods for assessing the gastropod community in dry grasslands. Pedobiologia, 42, 348357.
  • Opdam, P. (1991) Metapopulation theory and habitat fragmentation: review of holarctic breeding bird studies. Landscape Ecology, 5, 93106.
  • R Development Core Team (2007) r: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
  • Robinson, G.R., Holt, R.D., Gaines, M.S., Hamburg, S.P., Johnson, M.L., Fitch, H.S. & Martinko, E.A. (1992) Diverse and contrasting effects of habitat fragmentation. Science, 257, 524526.
  • Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius, W. & Hanski I. (1998) Inbreeding and extinction in butterfly metapopulation. Nature, 392, 491494.
  • Saunders, D.A., Hobbs, R.J. & Margules, C.R. (1991) Biological consequences of ecosystem fragmentation: review. Conservation Biology, 5, 1832.
  • Seber, G.A.F. (1982) The Estimation of Animal Abundance. Charles Griffin and Company Limited, Bucks, England.
  • Shaffer, M.L. (1981) Minimum population sizes for species conservation. Bioscience 31, 131134.
  • Stoll, P., Dolt, C., Goverde, M. & Baur, B. (2006) Experimental habitat fragmentation and invertebrate grazing in herbaceous grassland species. Basic and Applied Ecology, 7, 307319.
  • Turner, I.M. (1996) Species loss in fragments of tropical rain forest: review of the evidence. Journal of Applied Ecology, 33, 200209.
  • Vos, C.C. & Chardon, J.P. (1998) Effects of habitat fragmentation and road density on the distribution pattern of the moor frog Rana arvalis. Journal of Applied Ecology, 35, 4456.
  • WallisDeVries, M.F., Poschlod, P. & Willems, J.H. (2002) Challenges for the conservation of calcareous grasslands in northwestern Europe: integrating the requirements of flora and fauna. Biological Conservation, 104, 265273.
  • Wolff, J.O., Schauber, E.M. & Edge, W.D. (1997) Effects of habitat loss and fragmentation on the behavior and demography of gray-tailed voles. Conservation Biology, 11, 945956.
  • Zschokke, S., Dolt, C., Rusterholz, H.-P., Oggier, P., Braschler, B., Thommen, G.H., Lüdin, E., Erhardt, A. & Baur, B. (2000) Short-term responses of plants and invertebrates to experimental small-scale grassland fragmentation. Oecologia, 125, 559572.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1. Analyses of variance for the effects of year (1994, 1995, 1996), plot size (small, medium and large) and experimental fragmentation on population size expressed as density [individuals per square metre, log(density+1)] of six land snail species

Table S2. Fitted extinction probabilities (P ± SE) from a logistic analysis of covariance (Table 3 in the main text) for the periods 1994–95 and 1995–96 in small, medium and large fragments and control plots as a function of decreasing population size from 10 to 1

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
JANE_1478_sm_TableS1-S2.doc264KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.