Content uploaded by Jean-François Ponge
Author content
All content in this area was uploaded by Jean-François Ponge on Nov 08, 2018
Content may be subject to copyright.
1
Type of contribution: Short communication 1
2
Date of preparation of the revised version: 2005-09-02 3
4
Number of text pages: 8 5
6
Number of tables: 2 7
8
Number of figures: 1 9
10
Title: Decreased biodiversity in soil springtail communities: the importance of dispersal 11
and landuse history in heterogeneous landscapes 12
13
Jean-François Ponge1, Florence Dubs2, Servane Gillet1, Jose Paulo Sousa3, Patrick Lavelle2 14
15
1Muséum National d’Histoire Naturelle, CNRS UMR 5176, 4 avenue du Petit-Chateau, 91800 16
Brunoy, France 17
2Institut de Recherche pour le Développement, UMR 137 BioSol, 32 rue Henri Varagnat, 93143 18
Bondy Cédex, France 19
3Universidade de Coimbra, Instituto do Ambiente e Vida, Lg. Marquês de Pombal, 3004-517 20
Coimbra, Portugal 21
22
Correspondence: Jean-François Ponge, Muséum National d’Histoire Naturelle, CNRS UMR 5176, 4
avenue du Petit-Chateau, 91800 Brunoy, France. E-mail: jean-francois.ponge@wanadoo.fr
2
Abstract 1
2
In previously published papers it had been demonstrated that at the local level the species 3
richness of soil springtail communities was negatively influenced by landuse diversity. When the 4
dispersal rate of soil animals was taken into account, quite opposite trends were displayed by 5
species having poor or high dispersal capabilities. At the local level, species with short legs, non 6
functional jumping aparatus and reduction of visual organs were distinguished against by landuse 7
diversity, while species with long legs, functional jumping apparatus (furcula) and complete eyes, 8
thus able to disperse at the soil surface, were not. It was verified, through aerial photographs taken 9
fifty years ago, that landuse changes, expected to be more frequent in heterogeneous landscapes, 10
may contribute to explain this phenomenon. 11
12
Keywords: Landuse intensification, Springtails, Landuse diversity, Dispersal rate 13
14
In a previous study conducted within the BioAssess EC program, we have shown that the 15
springtail species richness of core samples (local biodiversity) was inversely related to landuse 16
diversity along a gradient of landuse intensification (Ponge et al., 2003), while an opposite trend 17
was displayed by plant species (Fédoroff et al., 2005). We hypothesized that, in the studied region, 18
heterogeneous landscapes were most subject to changes in landuse history, to which soil animals 19
were lesser adapted than plant forms. To test this hypothesis we decided to revisit our data set, by 20
taking into account the dispersal abilities of the different springtail species, and the landuse 21
changes that occurred over the last half century. 22
23
Sampling took place in the Morvan Regional Nature Park (western Burgundy, Centre of 24
France). Six landuse units (LUUs), one square kilometer each, have been chosen on the basis of 25
aerial photographs, taking into account the distribution of forested areas (coniferous, deciduous), 26
meadows and agricultural crops. LLUs 1 to 6 depicted a gradient of increasing influence of human 27
3
activities. Soil and climate conditions, as well as landuses, were described in two previously 1
published papers (Ponge et al., 2003; Fédoroff et al., 2005). The distribution of landuse types in the 2
six LUUs is shown in Table 1. In each LUU the diversity of landuse types was expressed by the 3
Shannon Index. 4
5
Using aerial photographs, a grid of 16 regularly spaced plots (200 m) was identified in each 6
of the six LUUs. Sampling of Collembola took place in June 2001. Methods used for sampling, 7
extraction, sorting and identification of Collembola at the species level were detailed in Ponge et al. 8
(2003). One sample was discarded for extraction, because of waterlogging at the time of sampling. 9
10
Collembolan species were classified in two groups, according to their ability to disperse 11
actively or not (Table 2). Species with long legs and antenna, a developed jumping apparatus 12
(furcula) and complete visual apparatus (8 ocella per eye spot) were considered able to disperse 13
rapidly by their own means (Hopkin, 1997). All other species, because of a reduction in motion or 14
vision organs, were considered as having poor dispersal capabilities. It has been demonstrated that 15
fully functional visual organs allow springtails to move directionally over long distances (Hågvar, 16
1995). Conversely, springtails showing a reduction in their eye number, even when fully motile, 17
exhibit negative phototaxis and thus cannot disperse easily (Salmon and Ponge, 1998). 18
19
Ecological requirements of Collembolan species were derived from the distribution of 20
species over all samples (n = 95), which was studied by multivariate analysis (Ponge et al., 2003). 21
Axis 1 coordinates of correspondence analysis (CA) in Ponge et al. (2003) were used to separate 22
forest from agricultural species (Table 2). It should be noted that the environmental gradient 23
depicted by the first axis of CA was a combination of all factors which contrasted woodland and 24
agricultural land, humus type included. 25
26
4
Ancient aerial photographs (1948 IGN campaign) were examined for each LUU, in order to 1
identify the landuse type which prevailed at the place where sampling took place in 2001. Black 2
and white photographs easily distinguished woodland, agricultural land, and hedgerows, but could 3
not be used for finer resolution. Thus landuse types were gathered into two gross categories, 4
namely woodland (deciduous and coniferous forests) and agricultural land (hay meadows, pastures, 5
arable crops, recent fallows). Hedgerows (one sample) and clear-cuts (two samples) were not taken 6
into account in the census. Afforested land was comprised of 10- to 50-yr-old coniferous 7
plantations (Douglas fir, Norway spruce) and old fields (wooded fallows). 8
9
When landuse units were ordered according to increasing landuse diversity (Table 1), the 10
balance between the two springtail categories changed markedly (Fig. 1). Slow-dispersal species 11
were largely dominant in forested areas, decreasing from LUU 1 to LUU 4, while the fast-dispersal 12
species increased. In the most diverse landscape (LUU 4) fast-dispersal and slow-dispersal species 13
were at the same richness level. In areas dominated by agriculture (LUU 5 and LUU 6), slow-14
dispersal species were also dominant, but to a lower extent than in areas dominated by forests 15
(LUU1 and LUU2). The mean species richness of slow-dispersal Collembola was negatively 16
correlated with landuse diversity (r = -0.93, P = 0.003), while fast-dispersal species were positively, 17
but poorly correlated with landuse diversity (r = 0.68, P = 0.07). 18
19
Examination of past landuse revealed that some changes took place over the last half 20
century. Eight samples were taken in places where there was a shift from agricultural land to 21
woodland (afforestation), while two samples were taken in places where woodland was replaced by 22
agricultural land (deforestation). Calculation of the impact of landuse change on local species 23
richness of springtail communities was only possible in afforested sites. It revealed a deficit of 24
species richness in sites where agricultural land was afforested (8.4±1.1), compared to stable 25
woodland (13.2±0.6, Mann-Whitney U = 54.5, P = 0.001). When springtail species were separated 26
in two dispersal groups, still clearer features appeared. In afforested land, the richness of slow-27
5
dispersal species was equal to that of the original agricultural land (i.e. near half that of woodland), 1
while the richness of fast-dispersal species decreased to the level typical of woodland, i.e. near half 2
that of agricultural land. While dominance of slow- over fast-dispersal species was prominent in 3
woodland as well as agricultural land, afforested land displayed the reverse phenomenon. The 4
examination of ecological requirements of species (forest vs agricultural species) showed that in 5
afforested land fast-dispersal species were shared between forest (57%) and agricultural species 6
(43%), while nearly all slow-dispersal species were still those typical of agricultural soils (88%). 7
8
We found that (i) the heterogeneity of the landscape exerted a negative influence on slow-9
dispersal collembolan species, (ii) at least part of these effects could be explained by time-related 10
processes, acting differentially on organisms with distinct life habits. 11
12
Soil collembolan communities are negatively affected by deforestation as well as 13
afforestation and this impact was shown to be durable (Jordana et al., 1987; Deharveng, 1996). The 14
contrast between closed and open habitats is one of the chief complex of factors which govern the 15
species composition of most soil animal groups (Nordström and Rundgren, 1974; Ponge, 1993). 16
Soil and climate factors are in play in the influence of landuse change on soil animal communities, 17
more especially when agricultural land shifts to woodland, or the reverse. However, in the present 18
state of our knowledge of ecological requirements and dispersal abilities of Collembola, only 19
cursory explanation can be found for the observed changes in species composition and diversity, 20
which are summarized below. 21
22
In the present study, the passage from agricultural land to woodland was accompanied by 23
soil acidification (Ponge et al., 2003). In mixed landscapes, the higher acidity of the soil in woody 24
areas is not solely due to the acidifying influence of forest growth (Nilsson et al. 1982) but also to 25
(i) the choice of more fertile (thus less acidic) soils for agriculture and (ii) the fertilization 26
associated with the permanent use of land for agriculture (Brady and Weil, 1999). Soil acidity and 27
6
associated factors influence the species composition of most soil animal communities (Nordström 1
and Rundgren, 1974; Wauthy, 1982; Ponge, 1993), but they affect primarily species in permanent 2
contact with humified organic matter (Ponge, 1993). 3
4
Micro-climate changes affect primarily soil animal species living at the soil surface or not 5
far from it, with agricultural land exhibiting more contrasting thermic and hygric conditions than 6
woodland at the ground surface (Coffin and Urban, 1993). In springtails, it has been demonstrated 7
that the first instar, i.e. the first stage of development following egg hatching, was more sensitive to 8
desiccation that further instars and that this phenomenon was more pronounced in woodland than in 9
agricultural land species (Betsch and Vannier, 1977). 10
11
Slow-dispersal Collembola, contrary to fast-dispersal species, are favoured by forest 12
environments. Most of them need a protection towards desiccation, because they are badly 13
equipped for jumping rapidly from a micro-site to another in a changing environment (Bauer and 14
Christian, 1987). Most species classified in the slow-dispersal category are soil-dwelling species, to 15
the exception of small surface species, such as Xenylla spp. and Brachystomella parvula, which 16
live under the protection of mosses and lichens or exhibit anhydrobiosis (André, 1976; Poinsot-17
Balaguer, 1976). 18
19
In a complex landscape where forest and agricultural land are intimately mixed, two 20
categories of underground biodiversity should be considered apart: those organisms which are able 21
to disperse at the scale of landscape heterogeneity (these communities will be positively or not 22
affected by heterogeneity, here fast-dispersal springtails), and those unable to do that, which will be 23
negatively affected except those tolerant of landuse changes. Examination of literature shows that 24
most springtail species which are considered of particular interest, because they are endemic or 25
rare, fall into the slow-dispersal category, thus are threatened by landuse changes (Barrocas et al., 26
1998; Deharveng, 1996; Lauga-Reyrel and Deconchat, 1999). 27
7
1
Acknowledgements 2
3
The authors thank the staff of the Morvan Regional Nature Park and numerous private 4
owners for facilities during choice of sites and sampling operations. This study was part of the 5
European Community program BioAssess EVK2-CT-1999-00041, coordinated by A. Watt (UK), 6
who is gratefully acknowledged for financial support and fruitful exchange of ideas between 7
participants. 8
9
References 10
11
André, H., 1976. Introduction à l’étude écologique des communautés de microarthropodes 12
corticoles soumises à la pollution atmosphérique. Bulletin d’Écologie 7, 431-444. 13
14
Barrocas, H.M., da Gama, M.M., Sousa, J.P., Ferreira, C.S., 1998. Impact of reafforestation with 15
Eucalyptus globulus Labill. on the edaphic collembolan fauna of Serra de Monchique 16
(Algarve, Portugal). Miscellània Zoològica 21.2, 9-23. 17
18
Bauer, T., Christian, E., 1987. Habitat dependent differences in the flight behaviour of Collembola. 19
Pedobiologia 30, 233-239. 20
21
Betsch, J.M., Vannier, G., 1977. Caractérisation de deux phases juvéniles d’Allacma fusca 22
(Collembole, Symphypleona) par leur morphologie et leur écophysiologie. Zeitschrift für 23
Zoologische Systematik und Evolutionsforschung 15, 124-141. 24
25
Brady, N.C., Weil, R.R., 1999. The Nature and Properties of Soils, 12th edn. Prentice Hall, Upper 26
Saddle River. 27
8
1
Coffin, D.P., and Urban, D.L., 1993. Implications of natural history traits to system-level 2
dynamics: comparisons of a grassland and a forest. Ecological Modelling 67, 147-178. 3
4
Deharveng, L., 1996. Soil Collembola diversity, endemism, and reforestation: a case study in the 5
Pyrenees (France). Conservation Biology 10, 74-84. 6
7
Fédoroff, E., Ponge, J.F., Dubs, F., Fernández-González, F., Lavelle, P., 2005. Small-scale 8
response of plant species to land-use intensification. Agriculture, Ecosystems and 9
Environment 105, 283-290. 10
11
Hågvar, S., 1995. Long distance, directional migration on snow in a forest collembolan, 12
Hypogastrura socialis (Uzel). Acta Zoologica Fennica 196, 200-205. 13
14
Hopkin, S.P., 1997. Biology of the Springtails. Oxford University Press, Oxford. 15
16
Jordana, R., Arbea, J.I., Moraza, L., Montenegro, E., Mateo, M.D., Hernandez, M.A., Herrera, L., 17
1987. Effect of reafforestation by conifers in natural biotopes of middle and South Navarra 18
(Northern Spain). Revue Suisse de Zoologie 94, 491-502. 19
20
Lauga-Reyrel, F., Deconchat, M., 1999. Diversity within the Collembola community in fragmented 21
coppice forests in south-western France. European Journal of Soil Biology 35, 177-187. 22
23
Nilsson, S.I., Miller, H.G., Miller, J.D., 1982. Forest growth as a possible cause of soil and water 24
acidification: an examination of the concepts. Oikos 39, 40-49. 25
26
9
Nordström, S., Rundgren, S., 1974. Environmental factors and lumbricid associations in southern 1
Sweden. Pedobiologia 14, 1-27. 2
3
Poinsot-Balaguer, N., 1976. Dynamique des communautés de Collemboles en milieu xérique 4
méditerranéen. Pedobiologia 16, 1-17. 5
6
Ponge, J.F., 1993. Biocenoses of Collembola in atlantic grass-woodland ecosystems. Pedobiologia 7
37, 223-244. 8
9
Ponge, J.F., Gillet, S., Dubs, F., Fédoroff, E., Haese, L., Sousa, J.P., Lavelle, P., 2003. Collembolan 10
communities as bioindicators of landuse intensification. Soil Biology and Biochemistry 35, 11
813-826. 12
13
Salmon, S., Ponge, J.F., 1998. Responses to light in a soil-dwelling springtail. European Journal of 14
Soil Biology 34, 199-201. 15
16
Wauthy, G., 1982. Synecology of forest soil oribatid mites of Belgium (Acari, Oribatida). III. 17
Ecological groups. Acta Oecologica, Oecologia Generalis 3, 469-494. 18
19
10
LU U 1 LU U 2 LU U 6 LU U 3 LU U 5 LU U 4
Dec iduo us forest 16 1 0 8 3 0
Con ifer ous forest 0 14 0 2 2 3
Clear-cut 0 1 0 0 0 1
Hedgerow 0 0 0 0 1 0
Hay m eadow 0 0 0 4 4 4
Past u re 0 0 2 1 6 3
Fallow 0 0 5 0 0 1
A rable crop 0 0 9 1 0 3
Not sam pled 0 0 0 0 0 1
Shannon Index 0 0 .6 7 1.3 7 1.8 8 2.1 1 2.4 2
Table 1. D istributio n of land use types am ong the six land use unit s (LUUs), ordered acco rdin g to
in creasing landuse div ers ity. Sixteen sam ples w ere tak en in each LU U, ex cept LUU 4 w it h f ifteen
sam ples only
1
2
3
11
AA nu rid a granaria FA llacm a fusc a
AA nu rid a unif orm is ADeu teros mint hurus sulp hu reu s
FA rrhopalit es bifid us ADic yrtom a fusca
FA rrhopalit es sp. ADic yrtom ina minut a
ABrachy st om ella parvula ADic yrtom ina ornat a
FCerat op hy sella ar mat a AEntom ob rya m ult ifas ciat a
ACerat op hy sella den ticulata FEntom ob rya n ivalis
FCerat op hy sella lu teospina AFasc iosm in th urus q uinquef asc iat us
ACyphoderus alb inus AIsotom a an tennalis
AFolso mia c andid a AIsotom a tig rin a
AFolso mia f im etaria AIsotom a vir idis
FFolso mia q uadrioculata AIso to muru s palustris
FFriesea c lav iset a ALepid oc yrt us cyan eus
FFriesea m irabilis FLep idocyrtus lan ugin osus
FFriesea t ru ncata ALepid ocy rtus lig no rum
AHet eromurus nitid us FLipothrix lu bb ocki
FIsotom iella m inor FOrc hesella c incta
FIsotom od es p roductus AOrc hesella q uinquef asc iat a
AKalap horura burm eisteri AO rches ella v illos a
FM egalot horax m inim us FPogonognat hellus flav escen s
FM esap ho rura bet sc hi FPseu diso to ma sen sibilis
FM esap ho rura jevanica ASmin th urid es parvulu s
FM esap ho rura leit zaensis ASm inthurides schoet ti
FM esap ho rura m acro ch aet a ASmin th urin us aur eus
FM esap ho rura yosii ASmin th urin us nig er
FM icran urid a pygmaea FSmin th urin us sig natus
FM icran urid a sensillat a ASm inthuru s nig rom aculatus
FM icrap horura ab solo ni ASm inthurus viridis
FNeanura m uscorum ASphaeridia pu milis
FNeelides m inutus ASten acidia vio lac ea
ANeo tu llberg ia ram ic uspis FTo mocerus m in or
FOncopodu ra cr ass ico rnis FV ert agopus arboreus
FOnychiu roid es pseu do gran ulosu s AW illow sia nig rom aculata
FOnychiu rus c eben narius
AOnychiu rus jubilariu s
FParatullbergia callipygos
AParisotoma n ot abilis
FProisotom a m inim a
AProtap ho rura ar mat a
AProtap ho rura m eridiata
AProtap ho rura prolata
FPseu dachoru tes parvulu s
APseu danuro ph orus binoculatus
APseu do sin ella alba
APseu do sin ella illiciens
FPseu do sin ella mauli
ASpinonychiu rus edin ensis
ASten aphor ura denisi
ASten aphor ura quadrisp ina
FW illemia an op ht halm a
FW illemia den isi
FW illemia in termedia
FX enylla grisea
FX enylla t ullberg i
FX enyllodes arm atus
Slow -dispersal
Fast-dispersal
Table 2. Lis t of collem bolan s pecies fo und in t he st ud y sites, classified
ac cordin g to their abilit y ot d isperse (estim at ed by append age and ey e
dev elo pm ent ). F = forest species. A = agricult ural species
1
12
Figure captions 1
2
Fig. 1. Local species richness of fast- and slow-dispersal springtail communities along a gradient of 3
increasing landuse diversity. The two springtail categories were compared at each LUU by 4
Mann-Whitney non-parametric tests. N.S. = not significant, ** = P<0.01, *** = P<0.001. 5
Bars are standard errors of the means. Within each category, LUUs were compared by 6
repeated Mann-Whitney non-parametric tests, the significance of differences between 7
groups (at P 0.05) being indicated by common letters (a, b for fast-dispersal species, x, y, z 8
for slow-dispersal species) 9
10
13
0
1
2
3
4
5
6
7
8
9
10
11
12
13
LUU 1
LUU 2
LUU 6
LUU 3
LUU 5
LUU 4
Species richness of a Collembolan sample
Slow-dispersal species
Fast-dispersal species
***
***
**
N.S.
***
***
a
bb
ab ab
a
x
xy
y
z
y
y
1
Fig. 1 2