Content uploaded by Eduardo Mateos
Author content
All content in this area was uploaded by Eduardo Mateos on Jul 03, 2018
Content may be subject to copyright.
Ecology and Evolution . 201 8 ;1–1 5 .
|
1
www.ecolevol.org
1 | INTRODUCTION
Soil fauna communities generally present a structure that is caused
by different factors depending on the spatial scales (Ettema &
Wardle, 2002). Principal biological factors are types of vegetation,
food resources availability, and interactions of animal species with
other organisms, especially microorganisms (L avelle & Spain, 2001).
Abiotic factors, such as bedrock composition, microsite humidity,
Received:12Januar y2018
|
Revised:3A pril2018
|
Accepted:20April2018
DOI:10.1002/ece3.4178
ORIGINAL RESEARCH
Hidden diversity in forest soils: Characterization and
comparison of terrestrial flatworm’s communities in two
national parks in Spain
Marta Álvarez-Presas1 | Eduardo Mateos2 | Marta Riutort1
This is an op en access article under t he terms of t he Creat ive Commons Attr ibutio n License , which pe rmits u se, dist ributi on and rep roduc tion in any m edium,
provide d the orig inal work is proper ly cited.
© 2018 The Aut hors. Ecology and Evolution pu blished by John Wiley & Sons Ltd .
1Departament de Genètica, Microbiologia
i Estadístic a, Institut de Recerca de la
Biodiversitat (IRBio), Universitat de
Barcelona, Barcelona, Spain
2Departament de Biologia Evolutiva,
Ecologia i Ciències Ambientals, Universitat
de Barcelona, Barcelona, Spain
Correspondence
Marta Riutort, Departament de Genètica,
Microbiologia i Estadística, Institut
de Recerc a de la Biodiversitat (IRBio),
Universitat de Barcelona, Avinguda
Diagonal,643,Barcelona08028,Spain.
Email: mriutort@ub.edu
Funding information
Ministerio de Agricultura, Alimentación y
Medio Am biente (Spain) program “Ayudas
a la investigación en Parques Nacion ales”,
Grant /Award Numbe r: ref. 589, 2012
Abstract
Terrestrial flatworms (Platyhelminthes, Tricladida, and Geoplanidae) belong to what
is known as cryptic soil fauna of humid forests and are animals not easily found or
captured in traps. Nonetheless, they have been demonstrated to be good indicators
of the conservation status of their habitat as well as a good model to reconstruct the
recent and old events affecting biodiversity. This is mainly due to their delicate con-
stitution, their dependence on the integrity of their habitat, and their very low dis-
persal capacity. At present, little is known about their communities, except for some
studies performed in Brazil. In this work, we analyze for the first time in Europe ter-
restrial flatworm communities. We have selected two protected areas belonging to
the Red Española de Parques Nacionales. Our aims include performing a first study of
the species richness and community structure for European terrestrial planarian spe-
cies at regional and local scale. We evaluate the effect of type of forests in the com-
munity composition and flatworms’ abundance, but also have into account the
phylogenetic framework (never considered in previous studies) analyzed based on
molecular data. We find differences in the species composition among parks, with an
astonishingly high diversity of endemic species in the Parque Nacional de Picos de
Europa and an extremely low diversity of species in the Parque Nacional de Ordesa y
Monte Perdido. These divergent patterns cannot be attributed to differences in physi-
cal variables, and in addition, the analyses of their phylogenetic relationships and, for
a few species, their genetic structure, point to a more probable historical
explanation.
KEYWORDS
Last Glacial Maximum, Microplana, molecular phylogenetics, refugia, soil ecology, species
diversity
2
|
ÁLVAREZ- PRESA S Et AL.
mean annual precipitation, or type of forest cover, can result in
variations of abundance and species composition in the soils at dif-
ferent scales, from microsite, site, local to regional levels (Melguizo-
Ruiz, Verdeny- Vilalta, Arnedo, & Moya- Laraño, 2012). On the other
hand, taxa from soil communities can exhibit also the genetic im-
print of ancient climatic and geographic event s that may have been
lost in other organisms with higher dispersal capacity (P fenninger
& Posada, 2002; Sunnucks et al., 2006). As a consequence, these
groups of organisms allow the reconstruction of old events affecting
the generation and maintenance of biodiversity and become excel-
lent indic ators of the conservation st atus of forest soils. However,
to be used for such aim, an extensive knowledge about the state
and functioning of their communities is necessary. Land planarians
(Platyhelminthes, Tricladida, Geoplanidae) belong to this group of
animals; they are inhabitants of humid forests soils and top pred-
ators of other invertebrates. They are simple animals that do not
possess mechanisms for water retention; therefore, they are depen-
dent on soil moisture to maintain their water requirements and use
vertical migration through soil, lit ter, and vegetation to keep their
humidit y (Winsor, Johns, & Yeates, 1998). Land planarians are in gen-
eral sensible to disturbed habitats, although some are reported to
be adapted to inhabit them (Carbayo, Leal- Zanchet, & Vieira, 2002;
Oliveiraetal.,2014;Álvarez-Presas,Amaral,Carbayo,Leal-Zanchet,
& Riutort, 2015). Based on these features, some studies have high-
lighted the value of this group of organisms as bioindicators in rela-
tion to the habitat perturbations caused by human activities (Sluys,
1998).
The highest species richness of autochthonous land flatworms
worldwide has been documented in the southern hemisphere
(Winsor et al., 1998), especially in areas originally covered by the
south- eastern Brazilian Atlantic Rain Forest (Carbayo et al., 2002;
Fick, Leal- Zanchet, & Vieira, 2006; Fonseca et al., 2009; Leal-
Zanchet & Baptista, 2009; Sluys, 1998, 1999). This could, of course,
be in part related to the existence of research teams interested in
the group. Probably due to this bias also most studies on commu-
nities for this group have been per formed in a restricted number
of very specific areas in South America: ombrophilous forests, de-
ciduous, and semideciduous forests in Southern Brazil (Antunes,
2008; Baptista, de Matos, Fick, & Leal- Zanchet, 20 06; Baptista &
Leal- Zanchet, 2010; Carbayo, Leal- Zanchet, & Vieira, 2001; Carbayo
et al., 2002; De Castro & Leal- zanchet, 2005; Fick et al., 20 06; Leal-
Zanchet & Baptista, 2009; Leal- Zanchet, Baptista, Campos, & Raffo,
2011; Leal- Zanchet & Carbayo, 200 0, 2001; Palacios, Baptista, &
Leal- Zanchet, 2006) and in the Atlantic Forest of nor thern Argentina
(Negrete, Colpo, &Brusa,2014). Other studies have analyzedthe
ecology of introduced terrestrial planarian communities in Europe
in relation to their invasive capacity (Boag, Yeates, & Johns, 1998;
Boag, Jones et al., 1998; Christensen & Mather, 1998; Jones, Green,
&Pali n,1998;Yeates ,B oa g,&Johns,1997),b utnost ud yoncommu-
nity composition or ecology of autochthonous European terrestrial
planarians has been performed.
Studies conducted in Brazil have shown that the commu-
nity structure for terrestrial flatworms can be influenced by the
vegetation type and by the degree of anthropic alteration (Carbayo
et al., 2002; Fick et al., 2006; Fonseca et al., 2009). However, studies
analyzing the effect of environmental factors have not found any
of them as driver of the abundance or species composition of ter-
restrial planarians communities (Antunes, Leal- Zanchet, & Fonseca,
2012; Baptista & Leal- Zanchet, 2010; Boag, Jones et al., 1998; Boag,
Yeated et al., 1998; Fick et al., 2006; Johns, Boag, & Yeates, 1998;
Sluys, 1998; Winsor et al., 1998) with the only exception of pH and
organic matter (see Baptista & Leal- Zanchet, 2010). At last, none of
the community studies conducted so far has taken into account the
phylogenetic relationships between the species found, nor whether
these relationships or the climatic and geological history of the area
can explain communities’ composition differences among areas.
Moreover, none has used molecular data in conjunc tion with phy-
logenetic inference methods to delineate genetic lineages, which
also provides a more accurate delimitation of molecular operational
taxonomic units (MOTUs) or species and avoids the problem of iden-
tifying morphologically cryptic or pseudocryptic species (common in
terrestrialplanarians,Álvarez-Presasetal.,2015;Carbayo,Álvarez-
Presas, Jones, & Riutort, 2016; Amaral et al., 2018) that may take
much time and lengthen the process of community study.
In Europe, although the diversity of species is still much lower
than in tropical regions, recent publications have shown that the
species richness of this group in this temperate region is higher than
previously suspected (Mateos, Sluys, Riutort, & Álvarez-Presas,
2017; Sluys, Mateos, Riutort, & Álvarez-Presas, 2016; and refer-
ences therein). In this continent, all the native species belong to the
subfamily Microplaninae and to a single genus, Microplana; although
there are some doubts about whether the genus Rhyncodemus can
also be autochthonous (Jones, 1988, 1998; Ogren & Kawakatsu,
1998). The European species are in general much smaller than the
tropical ones and less color ful, which renders them less prone to
be found and identified in any soil community study. Moreover, the
fact that this group of animals belong to cryptic soil fauna, due to
the difficulty in sampling them by the usual methodologies as using
traps or soil fauna extrac tors, makes difficult their inclusion in any
study of communities, nonetheless, they can contribute important
information.
In Spain, the forests suitable for terrestrial planarians are few
and are loc ated mainly in the nor th of the Peninsula. In that region,
three national parks bear forests with the characteristics needed
to host terrestrial planarians: in the P yrenees, the national parks
Aigüestortes i Sant Maurici and Ordesa y Monte Perdido, and in the
Cantabrian Mountains Picos de Europa. The two latter present a
broader extension of these types of forests and are where we
have focused our study. The Red Española de Parques Nacionales is
an integrated system intended to protect and manage a selection
of the best samples of the Spanish Natural Patrimony (http://www.
mapama.gob.es/es/red-parques-nacionales/ last visited January
2018). Among it s objectives, one of the most important is the pro-
tection and management of its biodiversity in order to ensure the
proper functioning of ecosystems. It is desirable therefore that
they house a biological diversity that is representative of it s original
|
3
ÁLVAREZ- PRESAS E t AL.
biodiversity(Araújo,Lobo,&Moreno,2007),thatincludesthelevels
of genetic diversity needed for the maintenance of its populations
and that it is also representative of the diversity in the region. But
for this it is necessar y to know what was the original situation of the
fauna inhabiting them, knowledge that is currently inexistent for the
terrestrial planarian communities, as well as for other cryptic forest
soil dwellers.
In this work, we performed a study on the diversity of terrestrial
planarian communities focusing on two national parks of the Red
Española de Parques Nacionales: Ordesa y Monte Perdido (hereafter
referred as Ordesa) and Picos de Europa (hereaf ter referred as Picos).
These parks have been selected because they are located in the area
of the Iberian Peninsula with the highest probabilit y of housing ter-
restrial planarians, as explained above. The broader extent of Picos
and its higher diversity of forests may influence the genetic diver-
sity distribution between and within the parks, predicting finding a
higher diversity and species richness in Picos. In addition, the two
parks are situated in regions that have gone through different an-
cient climatic events (Hewitt, 1999; Petit, Brewer, Bordács, & Burg,
2002;Petitetal., 2003)alsopointingtoan expectationofahigher
genetic diversity in Picos than in Ordesa, which could be reflected in
a higher number of species and/or within species diversity. Thus, the
specific aims of the study are as follows: (i) performing a first analysis
of the species richness and the community structure for European
terrestrial planarian species at a regional scale; (ii) analyze the effect
of the forest type in their communities (local scale); and (iii) explore
the drivers of the communities composition in the parks under a phy-
logenetic framework.
2 | MATERIALS AND METHODS
2.1 | Study area
The study area comprises two national parks in Northern Spain
(Figure 1): Ordesa y Monte Perdido (42°40′N, 0°3′E) and Picos de
Europa(43°30′N,4°55′W).Ordesa,withintheAragonesePyrenees,
in the Hue sca province, extends acro ss 15,636ha and exhib its a
mixture of climates, with both Mediterranean and Oceanic influ-
ences, and with a mean annual rainfall range between 1,129 and
1,690 mm/year. The highlands of the park (above 2,000 m altitude)
are extremely arid, as all water from rainfall is quickly picked up by
the karstic system. On the contrar y, valley bottoms are covered with
lush vegetation (forested area occupies 21% of the park range) domi-
nated by beech (Fagus sylvatica,7.8%),pineforestofPinus sylvestris
(6.6%), and fir trees giving way to the mountain pine (Pinus uncinata)
as altitude increases (Benito Alonso, 2010). Our sites were located
at elevations from 991 to 1618 m a.s.l. and equally spread across
the four main valleys composing the national park (Añisclo, Escuaín,
Ordesa, and Pinet a valleys).
Picos extends across Asturias, León, and Cant abria provinces, in
theCantabrian mountain range.Ithasasurfacearea of 67,127ha,
with a pronounced influence of Oceanic climate (Atlantic climate),
with cool summers and comparatively warm winters (Felicísimo,
1994).IthasthehighestlimestoneformationinAtlanticEurope,with
important karstic processes, chasms reaching more than 1,000 m,
very clear glacial erosion and presence of lakes. It is characterized by
a narrower range of annual temperatures than those encountered in
Ordesa at comparable latitudes, lacking, for instance, the extremely
dry summers typical of the other park, which is more influenced by
Mediterranean climate, and with a mean annual precipitation be-
tween 1,109 and 1,968 mm/year. Unlike Ordesa, most rainfall on
Picos comes as drizzle, and mountain fogs are very frequent due
totheOceanicinfluence(Felicísimo, 1994).Forestedareaoccupies
25% of the park ’s range, being beech forests of Fagus sylvatica the
most abu ndant forest ty pe (18.4% of the park area), fol lowed by
mixed and oak forest s (<0.5%). Atlantic mixed forests of Picos, relics
difficult to find in Spain, appear on the lower part of the mountain
and are interspersed with meadows areas. Common and cornish
oaks (Quercus robur or Quercus petraea) and hazels (Corylus avellana)
are intermingled with birch (Betula celtiberica) as the main tree spe-
cies (5%), maples (Acer sp.), linden (Tillia sp.), ash (Fraxinus excelsior),
chestnut (Castanea sp.), and walnut (Juglans regia) trees; at its feet,
an undergrowth of brambles, briars, and thorn. Our collection sites
wereestablishedatelevationsrangingfrom115to1,353ma.s.l.dis-
tributed in the three main areas composing the park (western, cen-
tral, and eastern massifs).
2.2 | Collection sites and sampling protocol
When working with cr yptic soil fauna, one important question is the
establishment of a standard sampling unit and the sampling method-
ology used in order to get comparable data between these sampling
FIGURE1 Map showing the situation within Spain of the two
National Parks, and a detail of the distribution of plots within each
park. Picos and Ordesa map at the same scale. Squares, oak forest;
triangles, mixed forest; circles, beech forest; asterisks, pine forest
4
|
ÁLVAREZ- PRESA S Et AL.
units. The authors working on terrestrial flatworm’s ecology (see
Baptist a & Leal- Zanchet, 2010; Baptista et al., 2006; Carbayo et al.,
2002; De C astro & Leal- zanchet, 2005; Fick et al., 2006; Leal-
Zanchet et al., 2011) used plots of fixed sizes (50 × 2 m, 2 × 2 m,
7×7m)ortransectsoffixedlength(30to50m)asasamplingunit.
On each plot or transect, the sampling methodology used was al-
ways direct searching by a fixed number of researchers (between
one and five) during a standardized lapse of time (between 15 min
and 1 hr). In our study, each sampling unit consisted of a forest plot
(of approximately 50 × 50 m) where two people searched for terres-
trial flatworms actively during one hour under logs and stones. Each
planarian detected was captured alive and deposited in a little plastic
container with humid substrate.
Atfirst,westablishedin24thenumberofplotstobesampledon
each forest type of each national park. However, it was not always
possibletosampleinthe24plotsselectedapriori.Inallcases,the
selection of the plots had into account a visual inspection of the hu-
midity of the site, selecting always the wetter sites maximizing the
probability of encountering terrestrial planarians. In Picos (Figure 1),
threeforesttypeswereselectedasfollows:beechforest(24plots),
oak forest (nine plots, main tree species Quercus robur in two plots,
and Quercus petraeainsevenplots),andmixedforest(14plots,main
tree species Quercus robur mixed with Fraxinus excelsior and Betula
celtiberica in five plots, and Quercus petraea mixed with Fraxinus ex-
celsior and Betula celtiberica in nine plots). In Picos, only nine plot s of
oakforestsand14plotsofmixedforestweresampledbecauseofthe
limited availability of forest s patches that met the appropriate condi-
tions for sampling. In Ordesa (Figure 1), samplings were per formed
in two forest types: beech forest (22 plots) and pine forest, (24
plots). In this park, t wo plots of beech forest were not sampled due
to the deterioration status of the forest patch selected. All sampling
protocol was performed twice (dates in format day/month/year),
from2/10/2013to13/10/2013andfrom24/5/2014to4/6/2014in
Ordesa,andfrom 21/10/2013to01/11/2013andfrom22/6/2014
to03/7/2014inPicos.Everysamplingday,plotsofdifferentforests
type were sampled in order to avoid temporary self- replication. The
data of the t wo sampling campaigns were pooled.
In the same day of the collections, the animals were visualized
under a stereomicroscope, their morphological external appearance
recorded and photographed, and finally fixed. When the specimens
were big enough, a small anterior section was fixed in absolute eth-
anol for DNA extraction, and the rest (including the parts necessar y
for histological studies) were fixed with Steinmann fluid and stored
in70%alcoholinordertostudythecopulatoryapparatusandother
structures that allow the diagnosis of the species.
2.3 | Environmental parameters
Pluviometry data were obtained from two meteorological stations
located at the east and west ends of each national park. Total an-
nualrainfallaswell as accumulated rainfallofthe 3monthspre-
vious to each sampling campaign was recorded. For calculation
purposes, the mean value of the two stations of each park was
considered. In Ordesa, the t wo meteorological stations were lo-
cated in Bielsa municipality (at 1,100 m asl) and Torla municipal-
ity (1,000 m asl). In Picos, the two meteorological stations were
located in Soto de Sajambre municipality (1,500 m asl) and Sotres
municipality (1,200 m asl). On each sampling campaign soil tem-
perature, pH and water content were measured. On each plot,
soil temperature was taken in three points and the mean value
was used for calculations. Also on each plot, two samples of soil
substrate were taken to measure hydric content and two other
samples were taken to measure the pH in the laboratory; the mean
values of each variable were used for calculations. Measures of pH
were performed the same day with a portable pH meter (PH25,
Crison) by diluting the soil substrate samples in 5 volumes of dis-
tilled water. Also the same day, the two soil samples for hydric con-
tent were weighted to obtain their fresh weight (Fw). Once back
in the University laborator y, the samples were dried in a stove at
105°Cfor48hrand thenwereweighted againtohavetheir dry
weight (Dw). The hydric content for each sample was expressed in
percentage of water content and calculated gravimetrically by the
equation 100*[(Fw- Dw)/Fw].
2.4 | Species assignation and phylogenetic inference
Samples preser ved in 100% ethanol were used for DNA extraction
with the Wizard® Genomic purification kit (Promega, Madison, WI,
USA) following the same protocol as in Álvarez-Presas, C arbayo,
Rozas, and Riutort (2011). A fragment of the gene encoding the
mitochondrial cy tochrome oxidase I (Cox1) was analyzed by poly-
merase chain reaction (PCR) together with three nuclear genes:
genes encoding the 18S t ype II rRNA (18S) and 28S rRNA (28S) and
a fragment of Elongation Factor 1- alpha (EF). For 18S and 28S, we
usedprimersandPCRconditionsasinÁlvarez-Presas,Baguñà,and
Riutor t (2008) for Cox1 a s in Álvarez-Presa s etal. (2011) and for
theEFas inCarbayo etal.(2013).Thesame primerswereusedfor
PCR amplification and sequencing. The amplification products were
purified directly with a vacuum pump (Multiscreen®HTS Vacuum
Manifold, Millipore Corporation, Billerica, MA 01821, USA). DNA
sequences were determined from both strands by Sanger sequenc-
ing in Macrogen (Amsterdam, Europe). Chromatograms were revised
andcontigs constructedin Geneiousv8.1.7.software(Biomatters;
available from http://www.geneious.com last visited Januar y 2018).
Genes Cox1 and EF were aligned based on the amino acid se-
quences using the Clustal W plugin included in the BioEdit software.
7.0.9.0.(Hall, 1999).Ribosomal RNAgene sequenceswere aligned
usingtheonlineversionofthesoftwareMafftv.7(Katoh&Standley,
2013)applying theG-INS-i iterative refinementmethodand,sub-
sequently, checked the alignments by eye with Bioedit. Positions
that could not be unambiguously aligned were subsequently ex-
cluded from the analyses by applying GBlocks v 0.91b (Talavera &
Castresana, 2007), with halfallowed gap positions and a minimum
length of a block of 10 nucleotides, to obtain the maximum num-
ber of nucleotides. Based on these alignments, we estimated the
DNA sequence evolution model that better fits the data by using
|
5
ÁLVAREZ- PRESAS E t AL.
jModelTestv.2.1.4.(Darriba,Taboada,Doallo,&Posada,2012),ap-
plying the Akaike information criterion (AIC).
For the Cox1 gene, a saturation test was run using the soft ware
DAMBE6(Xia,2017)byplottingobservedtransitionsandtransver-
sions vs. gene divergences under the GTR model.
Two datasets have been used for different analyses: (i) Cox1
dataset, included all Cox1 mitochondrial gene sequences to assign
individuals from Picos and Ordesa to MOTUs; (ii) concatenated data-
set, included the information of the three nuclear genes to infer a
general phylogeny. Both alignments included sequences down-
loaded from GenBank (see Supplementary Information Table S1 for
the individuals included in each dataset).
Phylogenies of the two datasets were inferred applying the
maximum- likelihood method (ML) with the software RaxML v.8
(Stamatakis,2014)estimatingbootstrapsupportvaluesfrom10,000
replicates. We also use d the Bayesian inferen ce method (BI) as impl e-
mented inthesoftwareMrBayesv3.2.2.(Ronquist etal.,2012)for
the concatenated dataset. Two runs were applied producing 5 mil-
lion generations for each and of these a tree each 1,000 was stored.
A 25% default burn- in was used, likelihood values (log- likelihood) of
cold chains were checked to have reached stationarity, and the con-
vergence of the two runs was verified by the average standard de-
viation of split frequencies being ≪0.01. A consensus tree from the
remaining trees was obtained. For the Cox1 dataset, we used BEAST
softwarev2.3.2(Bouckaertetal.,2014)sincewithMrBayes,itwas
impossible to reach the convergence of the two runs. A relaxed log
normal clock was applied, and 100 million generations were run stor-
ing a tree of every 10,000.
For the MOTUS shared by the t wo parks and with a higher abun-
dan ce(M02andM78),haplotypenetwor kswereconstructedusingthe
median- joining met hod (Bandelt, Fo rster, & Röhl, 1999) as implemente d
in the Popart program (available at http://popart.otago.ac.nz, last vis-
itedNovember2017).Toperformthesehaplotypenetworks,informa-
tion from t he Cox1 dataset was us ed, adding seque nces from a previou s
analysis of Microplana terrestrisintheIberianpeninsula(Álvarez-Presas,
Mateos, Vila- Farré, Sluys, & Riutort, 2012) in the case of M02, and add-
ing known sequences from M. aixandrei from the GenBank database, in
thecaseofM78(seeSupportinginformationTableS1).
2.5 | Numerical methods
On each plot, the measure unit in numerical analyses was the ter-
restrial flatworm abundance, defined as the number of specimens
collected by two researchers during one hour in the two sampling
campaigns pooled. We also quantified the species richness (number
of species or MOTUS) per sampling plot. These two variables were
compared among the parks and the forest types by ANOVA or t-
Student tests (with Tukey test post hoc comparisons when neces-
sary) af ter checking the homogeneity of variances by the Levene
test (Levene, 1960). A Pearson correlation test between soil water
content, temperature, and pH with respect to flatworm abundance
and richness per plot was performed. These tests were performed
usingR-languagepackage(Team,2003).
From the matrix of abundance data of flatworm species (elimi-
nating species with overall abundance inferior to three specimens),
we carried out an analysis of similarity in flatworm species compo-
sition bet ween national parks and among forest t ypes. To do this
analysis, the abundance data were pooled by forest t ype and trans-
formed using log(x + 1) procedure. Differences in flatworm species
composition between pairs of sampling points were quantified by
the Bray–Curtis similarity index. From the similarity matrix, we per-
formed a similarity analysis (ANOSIM, PRIMER- E 2001), which gives
a general R- value and allows pairwise comparisons between the
areas compared (parks or forest types).
For each national park, we estimated the three basic compo-
nents of the diversity (sensu Whittaker, 1960). We define α- diversity
as the diversity measured on each forest type, β- diversity as the spe-
cies turnover between the forest types of each national park, and
γ- diversity as the diversity value measured on each national park as
a whole (pooling the forest types). Estimates of α- and γ- diversity
were done using Shannon–Weiner diversit y index (H), and species
richness. β diversity was estimated using the Sørensen dissimilarity
index bet ween pairs of forests type on each national park; this index,
defined as βtinWilsonandShmida(1984),isboundbetween0and
1, where 0 means that the two sites have the same composition (i.e.,
they share all the species), and 1 means that the t wo sites do not
share any species. Commands “diversity” (with base 2 logarithms)
and “betadiver,” from R Vegan package, were used for diversity
calculations.
We have used species accumulation cur ves to model the species
increase in relation to sample size at two different scale levels. First,
analyzing the two national parks as a whole and second analyzing
only beech forest s of the two parks (as this is the shared forest type
in the two parks). Due to the disparity in the sampled area (Picos is
4.3timeshigherthanOrdesa),inordertomakecomparablethespe-
cies accumulation curves bet ween beech forests in the t wo parks,
the second analysis has been performed constructing four different
plot data matrices in Picos, each of them with surface area equiv-
alent to that found in Ordesa (see Supporting information Figure
S1): north (12 plots, A + B), south (12 plots, C + D), east (eight plots,
B + D), and west (16 plots, A + C). For the two species accumulation
curves analyses we used the “method = random” from the command
“specaccum” in R Vegan package, calculating the mean species accu-
mulation curves and its standard deviation from random permuta-
tions of the sampling units (plots).
In general, in sampling protocols, not all species are detected in
any site, and these unseen species also belong to the species pool
(Oksanen, 2016). In order to estimate the number of unseen species
on each forest type and on each national park as a whole, command
“specpool” of R Vegan package was used. Command “specpool”
studies a collection of sites and assumes that the number of unseen
species is related to the number of rare species, or species seen
only once or twice (see Oksanen, 2016 for a detailed discussion).
Command “specpool” implements several nonparametric estimators
ofwhichweselectedtheChao1estimator(Chao,1987;Chiu,Wang,
Walther,&Chao,2014),thefirstorderjackknifeestimator,andthe
6
|
ÁLVAREZ- PRESA S Et AL.
bootstrapestimator(Smith&vanBelle,1984),calculatingthemean
estimate values, and associate variances.
3 | RESULTS
3.1 | Species- MOTUs
Atotalof350individualsbelongingtothesubfamilyMicroplaninae
were collected from across all the sampling plots (Supporting infor-
mationTablesS1andS2),202fromOrdesaand148fromPicos.
The phylogenetic tree obtained with Cox1 gene sequences
(Figure 2) defines 15 monophyletic groups or entities (MOTUs), four
ofthempresentinOrdesaand 13inPicos(twoarepresentin both
parks). Five of the clades correspond to already described species,
M. terrestris (M02), M. nana (M01), M. fuscomaculosa (M25), M. ner-
vosa (M22), and M. cf. aixandrei-2 (M78; Mateos eta l., 2017). The
rest here designated are under study and are going to be described
as new species elsewhere (Álvarez-Presas etal. in preparation).
Nonetheless, in this work, we will designate all of them with their
MOTU codes (Mxx).
Of the total of 15 MOTUs found, M02 (M. terrestris) an d M78
(M. cf. aixandrei- 2) are the only two species shared in the two parks.
In Ordesa, we have found four MOTUs all of them having a known
distribution through Europe or wide in the Iberian Peninsula, not
being in any case endemic from the Park or the area (Table 1). On
theotherhand,inPicos,sevenofthe13MOTUsfoundareendemic
from the Park, three are found only in wet forests in North- Western
Spain, and three have a wide distribution through Europe (Table 1).
3.2 | Phylogenetic relationships
The analyses of the Cox1 gene phylogeny showed no resolution for
the relationships among MOTUs, and the saturation test shows this
molecule to be effectively saturated (Suppor ting information Figure
S2); for this reason, only nuclear genes were used to infer the re-
lationships among MOTUs. Figure3 and Supporting Information
Figure S3 s how the phylogenet ic trees inferre d from the concat-
enated dataset by maximum likelihood and Bayesian inference, re-
spectively; there are small differences between both trees affecting
nodes with low support but some clear relationships appear. Both
trees show a clade constituted exclusively by MOTUS coming from
PicosthatishighlysupportedinBIbutnotinML(0.93PPand50%
BP)(AinFigure3).Thisgroupissister toaclade(B) constitutedby
M. terrestris (M02) and M. cf. aixandrei-2(M78);thetwospeciespre-
sentinthetwoparksandalsoinotherregionsaroundEurope.M35,
a species exclusive from Picos, in the BI analyses is sister to clade B,
while in the ML tree is sister to clades A and B but with low support
in both analyses.
FIGURE2 Maximum- likelihood (ML) tree inferred from Cox1 dataset. Monophyletic groups comprising MOTUs have been collapsed.
Values at nodes correspond to boot strap values (above, ML analysis) and posterior probability (below, BI analysis); only values over 0.85 PP
and75%bootstraparedisplayed,respectively.Scalebar=numberofsubstitutionspersite
|
7
ÁLVAREZ- PRESAS E t AL.
Fortherestofspecies,M71(fromPicosandotherEuropean
regions) is sister group to all other MOTUS included in the
analysis with maximum suppor t. Then, two MOTUS from
Ordesa and other regions split , M01 (M. nana) followed by M25
(M. fuscomaculosa).
For Microplana terrestris (M02), species shared between parks,
the haplotype network (Figure4a) shows that individuals from
Ordesa share haplotypes found in the eastern clade defined in a
previous study (Álvarez-Presasetal., 2012)orhaveafew differ-
ences from those; the haplotypes of the individuals from Picos
coincide with haplot ypes found in the western clade defined in
the same publication. Also as in the previous study, the patterns
of diversit y between the two clades (east and west) are different,
showing a star pattern in the east and a more structured pattern
in the west.
Asregardstheotherspeciessharedbetweenparks(Figure4b),
M. cf. aixandrei-2(M78),th ereisahighlyfrequenthaplotypes ha re d
in both par ks and also by spec imens from other r egions of the Iber ian
Peninsula (Barcelona and Málaga), showing no correlation between
genetic structure and geographical distribution at this level. There
is an haplot ype composed of sequences only from Ordesa highly
differentiated from this frequent common haplotype. However, the
limited representation of individuals of this species in Picos (only
three animals) prevents us from stablishing a good comparison of
its haplotype distribution with that of M. terrestris.
3.3 | Abundance, diversity, and community
composition at regional scale (parks)
Given the low general abundance of the MOTUs, we have pooled
the data from spring and autumn samplings for the following analy-
ses. The proportion of plots without any terrestrial planarian was
higherinPicos(34%,16of47plots)thaninOrdesa(11%,fiveof46
plots) (Supporting information Table S2). The abundance per MOTU
in each plot was in general low with only a few MOTUs reaching val-
ues close or over 10 animals in some plots, resulting in mean abun-
dances per type of forest below 1 (Table 1) with a few exceptions
(M02, M28 , and M78). The distr ibution of MOTUs is not u niform
through parks (Table 1) as demonstrated by the ANOSIM analysis
that shows significant differences in species composition between
parks (p level < 0.05, Table 2).
Mean abundances and richness per plot for each park were not
signific antly different (Table3). Nonet heless, the tota lnumber of
species and its diversity was different among parks, presenting Picos
a higher ri chness (13 vs. 4 specie s, Table4). The species a ccumu-
lation models performed with all plot s of the two parks (Figure 5)
showed that , with around 12 accumulated plot s, the predicted spe-
cies number is higher in Picos than in Ordesa. Also in Ordesa, the
predicted species number stabilizedquickly (flat line)and with 35
accumulated samples the deviation of the mean is zero, while in
Picos, the predicted species number increases continuously with
the accumulated plots. Moreover, all three nonparametric total spe-
cies estimators (Chao- 1, Jack- 1 and Boot) showed a little increment
TABLE1 Mean abundance of terrestrial flatworm MOTUs by forest type on each National Park
Park Forest n n- t p M01 M02 M19 M20 M22 M23 M 24 M25 M28 M35 M37 M71 M73 M77 M78
Ordesa Beech 22 20 0.09 2.73 –––––0.23 – – – – – – 1.00
Ordesa Pine 24 21 0.17 2.96 –––––0.17 – – – – – – 1.42
Picos Beech 24 13 –0.54 0.21 0.08 0.08 0.08 – – 0.67 0.04 0.08 –0.17 0.04 0.08
Picos Mixed 14 13 –3.43 – – 0.29 0.14 0.14 –1.64 0.07 0.14 0.14 0.14 –0.07
Picos Oak 94–0.22 0.22 0.22 –0 .11 0.11 –0.33 – – – – – –
Distrib NE S Eur. Picos Picos NWS Picos Picos Orde Picos Picos NW S Picos NW S Picos S
F F F
UK BUK
Note. n, number of plots s ampled; n- tp, number of plots with terrestrial planarians; Distrib, Species geographical distribution; B, Bulgary; Eur, Europe; F, France; NE, northeast; NW, nor thwest; Orde, Ordesa
y Monte Perdido National Park; Picos, Picos de Europa National Park; S, Spain; UK, United Kingdom.
8
|
ÁLVAREZ- PRESA S Et AL.
in predicted species number in Picos, but no increment in Ordesa
(Table4). The species accumulation models performed only with
beech forests plots of the t wo parks also show that beech forest
in Picos host more species than Ordesa irrespective of the area se-
lected (Supporting information Figure S1). At last, Gamma diversity
measured with the Shannon diversity index (H) was higher in Picos
thaninOrdesa(Table4).
3.4 | Abundance, diversity, and community
composition by forest type
In the two forest types analyzed in Ordesa, beech, and pine forests,
the proportion of plots with terrestrial planarians was high (Table 1).
In both forests, the same four species and in the same proportion
were found (Figure 6a), and M02 was the most common species,
FIGURE3 Phylogenetic tree obtained with the concatenated dataset (18S, 28S, and EF). Maximum- likelihood topology is shown.
Numbersatnodesindicatebootstrap(when>75%)andposteriorprobability(when>0.85)values.Blackdotsatnodescorrespondto
maximum support values both in bootstrap and posterior probability. A: clade corresponding to MOTUS exclusively from Picos; B: clade
corresponding to the species found both in Picos and Ordesa. Scale bar = number of substitutions per site
FIGURE4 Median- joining network reconstructed with PopART for the mitochondrial (Cox1) haplotypes: (a) for Microplana terrestris
(MOTU02),eastandwestcorrespondtohaplotypesfoundintheeasternandwesterncladesinÁlvarez-Presasetal.(2012);(b)forM. cf.
aixandrei-2(MOTU78).Circlesizeisproportionaltosamplesize;crossinglinesbetweenhaplotypesindicatemutationalsteps
(a) (b)
|
9
ÁLVAREZ- PRESAS E t AL.
foll owedbyM78and,fur theraway,byM2 5a ndM 01.AN OSIManal-
ysis showed no significant difference in terrestrial flatworm species
composition between the two forests (Table 2). In this park, mean
terrestrial flat worm’s abundance per plot was similar in the two for-
est types, while mean richness per plot was significantly higher in
pine than i n beech forest ( Table3). Total speci es number (S) was
identical in the two forest t ypes (four species), and α- diversity values
measuredwith Shannon index (H) werealso very similar(Table4).
The Sørensen dissimilarity index value (βt) measured between beech
and pine forests in Ordesa was 0.00, meaning that the same species
were found in both forest types. As a result, the nonparametric total
species estimators showed the same values as actual total species
number for beech and pine forests (four species, with only a lit tle
deviationforbootstrapestimator,Table4).
In Picos, three forest types were analyzed, beech forest (11
species of terrestrial flatworms), oak forest (six species), and mixed
forest (10 species), the two- first having a lower proportion of plot s
with terrestrial planarians (Table 1). In beech and oak forests, the
abundance of different species was similar, but in mixed forest, M02
was by far the most abundant species (Figure 6b). ANOSIM analysis
(Table 2) showed no significant difference in species composition
between beech and oak forests (p value > 0.05), but significant dif-
ferences were found between terrestrial flatworm communities of
mixed forest with respect to beech and oak forest s (p values < 0.05).
Mean terrestrial flatworm’s abundance per plot was significantly
higher in mixed forest than in beech and oak forests, while mean
species richness per plot was not significantly dif ferent between
thesethreeforesttypes(Table3).Shannondiversityvalues(H)were
higher in beech and oak forests than in mixed forest due to a higher
abundanceofM02 and M28 inthelast forest (Table4,Figure6b).
The mean Sørensen dissimilarity index value (βt) measured between
thethree forest types was 0.38(βt between beech and mixed for-
ests=0.24,betweenbeechand oakforests=0.41,between mixed
and oak forests = 0.50). The three nonparametric species number
estimators (Chao1, Jack1 and Boot) showed an increment in total
species number in the three forest types compared with actual total
speciesnumber,especiallyforoakforestandmixedforest(Table4).
3.5 | Environmental parameters
Total annual rainfall and accumulated rainfall of the three previous
months to each sampling campaign are shown in Suppor ting infor-
mation Table S3. In Ordesa, the meantotal annual rainfall for the
years of this study was within the normal range of historical data
(mean annual rainfall range bet ween 1,129 and 1,690 mm/year).
In Picos, however, rainfall was very low during the sampling period
(especiallyin2013), far belowthemeanannual precipitationrange
normal for the park (between 1,109 and 1,968 mm/year) and, which
is more important to soil humidity, the accumulated rainfall during
the three previous months to sampling dates was also ver y low (also
especiallyin2013).
The mean soil water content, temperature, and pH values mea-
sured in the two sampling campaigns on ever y forest type of each
TABLE2 Differences in species composition (ANOSIM results)
between parks (regional scale) and between forest types (local
scale). Data from fall and spring pooled
Comparisons R p
Regional scale
Ordesaallplots(41)vsPicosallplots
(30)
0.256 0.0001*
Local scale by forest type
Ordesa
Beech forest (20) vs. Pine forest (21) 0.039 0.1070ns
Picos
Beechforest(13)vs.Oakforest(4)
vs.Mixedforest(13)
0.194 0.0040*
Beechforest(13)vs.Oakforest(4) −0.040 0.6210ns
Mixedforest(13)vs.Beechforest
(13)
0.192 0.0030*
Mixedforest(13)vs.Oakforest(4) 0.4 49 0.0200*
Notes. ns, not significant probability. *Significant probability.
TABLE3 Species abundance and richness per plot on each park
(regional scale) and forest type within parks (local scale) and
Student’s t test or ANOVA for the comparisons of abundance and
richness
Area nAb (SE) Ri (SE)
Regional scale
Ordesa all forests
(Ord)
46 4.39(0.48) 1.57(0.12)
Picos all forests (Pic) 47 3.15(0.67) 1.32(0.18)
Student’s t- test t=1.499 t=1.135
p=0.137ns p = 0.260ns
Ord = Pic Ord = Pic
Local scale by forest type
Ordesa
Beech forest (Fo) 22 4.05(0.72) 1.32(0.14)
Pine forest (P) 24 4.71(0.66) 1.79(0.18)
Student’s t- test t=−0680 t=−0.473
p = 0.500ns p=0.046*
Fo = PFo < P
Picos
Beech forest (Fp) 24 2.08 (0.65) 1.17(0.26)
Oak fores t (Q) 9 1.22(0.57) 0.78(0.36)
Mixed forest (M) 14 6.21 (1.68 ) 1.93(0.29)
One- way ANOVA
test
F=5.451 F = 2.988
p = 0.008*p = 0.061ns
M > Fp = Q Fp = Q = M
Notes. Dat a from fall and spring pooled. n, number of plots; Ab, mean
number of individuals per plot; Ri, mean species richness per plot; SE,
standard error; ns, not significant difference between the above values
in the statistic test. *Signific ant difference between values above.
10
|
ÁLVAREZ- PRESA S Et AL.
national park are shown in Supportinginformation Table S4. Asthe
same plot s were visited during the two sampling campaigns, altitude
measures were the same in the two dates (Supporting information
TableS4).InOrdesa,ingener al,beechfore stsoilsshowedhighermean
water content and pH values and lower temperature than pine forest
soils. Comparing the two sampling campaign’ values, the main differ-
ences were the higher mean soil water content and the lower mean soil
temperaturevaluesin2014(especiallyinbeechforest).InPicos,gen-
erally beech forest soils showed higher mean water content and lower
temperature than oak and mixed soil forests, while mean pH value was
higher in mixed forest than in beech and oak forests. Comparing the
two sampling campaigns values, the main differences were the higher
soilwatercontentandtemperature,andlowerpHvaluesin2014.
In the two parks, no significant correlation was obtained be-
tween flatworm abundance and richness per plot with respect to soil
water content and temperature. In Picos, total terrestrial flat worm
abundance and richness per plot were significantly correlated with
soil pH values (Supporting information Table S2), presenting higher
values of abundance and richness those plots with moderate- to- high
pHvalues(SupportinginformationFigureS4),whileinOrdesa,these
correlations were not significant.
4 | DISCUSSION
This study presents the first community analysis of terrestrial pla-
narians in Europe. The analysis, performed in two national parks in
northern Spain, has revealed an unexpected diversity for this group
of animals. However, this diversity is not equivalent in both parks;
community composition and total amount of species significantly
differ. Moreover, within parks, only Picos shows some signific ant dif-
ferences among forest types. In the following sections, we analyze
the reasons to explain such differences.
4.1 | Differences in flatworm’s abundance and
richness between forests within parks
Using arthropods as model organism in three national parks in Spain
(Aigües Tortes, Ordesa and Picos de Europa), Melguizo- Ruiz et al.
(2012) found higher densities of edaphic arthropod fauna in micro-
sites located at the base of the hillsides, which were more humid
and rich in litter. Moreover, these authors point that calcareous
soils (with higher pH values) present higher quantities of arthropod
edaphic fauna than silicic soils. However, in the case of land planar-
ians, several publications have indicated that some environmental
charac teristics may af fect the presen ce and abundance of th is animal
group in forest soils, such as pH, depth, temperature, soil moisture,
Area n S H Chao1 (SD) Jack1 (SD) Boot (SD)
Regional scale γ- H
Ordesa all forests 46 41 .26 4.00(0.00) 4.00(0.00) 4.00(0.05)
Picos all forests 47 13 2. 51 13.17(0.54) 13.98(0.98) 13.91(0.83)
Local scale by forest
type
α- H
Ordesa
Beech forest (Fo) 22 41.22 4.00(0.00) 4.00(0.00) 4.14(0.34)
Pine forest (P) 24 41.28 4.00(0.00) 4.00(0.00) 4.05(0.23)
Picos
Beech forest (Fp) 24 11 2.81 11.40(0.87) 12.92(1.36) 12.41(1.10)
Oak fores t (Q) 9 6 2.48 18.50(17.14) 10.44(2.39) 7.84(1.33)
Mixed forest (M) 14 10 1.96 14.17(4.88) 14.64(2.48) 12.13(1.42)
Notes. Dat a from fall and spring sampling campaigns pooled. n, number of plots; S, total species
number; H, Shannon diversity index; γ- H, γ- diversit y measured on each National Park; α- H, α-
diversity measured in each forest t ype; Chao1, Chao- 1 total species number estimator; Jack1, first-
order jackknife total species number estimator; Boot, boot strap total species number estimator; SD,
standard deviation.
TABLE4 Species diversity per parks
(regional scale) and per forest t ype within
parks (local scale)
FIGURE5 Flatworm species accumulation curves per park.
Vertical lines indicate standard deviations
|
11
ÁLVAREZ- PRESAS E t AL.
and prey abundance (Boag, Jones et al., 1998; Boag, Yeated et al.,
1998; Fick et al., 2006; Johns et al., 1998; Winsor et al., 1998), but
none of them point to any of these factors as the main driver of the
presence/absence of these terrestrial invertebrates. Antunes et al.
(2012), working in an undisturbed area of Araucaria forest, found
that terrestrial flatworms were not significantly associated with
any particular microhabitat condition, which included environmen-
tal parameters as soil humidity, but also leaf- litter and other fauna
composition. Comparing land flatworm communities in two types of
forests in Southern Brazil, Fick et al. (2006) found that pH differ-
ences together with thermal amplitudes may explain dissimilarities
in composition between the communities due to disparities in pH
preferences for different species.
In the present study, flatworm abundance and richness per plot
showed a significant correlation only with pH in Picos (Supporting
informationFigure S4),withhighervalues of theseparameters for
neutraltobasicsoils(pHvaluesbetween6.5and7.5).Thiscouldbe
the explanation of the significant differences found for one of the
forest types in Picos, the mixed forest, that seems to harbor higher
density of planarians per square meter (individuals per plot), and
shows a significantly different species composition with respect to
beech and oak forests.
In Ordesa, although the total abundance per plot was equiva-
lent between forests, we found a higher species richness in pine
forests. Both forest types in Ordesa showed mean pH values over
6 (Suppor ting information Table S4), around what appears to be
optimal for Microplana species attending to the result s in Picos
(SupportinginformationFigureS3),whichmaybeoneofthefactors
to explain the lack of differences between beech and pine forests
in this park.
4.2 | Differences in flatworm’s abundance and
richness at a regional scale
Our results showed the existence of significant differences between
parks in terms of species composition and richness ( Table 2) that
cannot be explained by differences in the carrying capacity be-
tween parks(Table3), nor by the inequalityinthearea of the two
parks (Supporting information Figure S1). The environmental factors
analyzed also fail to explain such differences. However, the histori-
cal analysis based on the phylogeny of the species and the genetic
structure within the shared species provides some clues about the
differences observed.
Of the two shared species bet ween parks, M02, (M. terres-
tris,Figure4a),presentedahighlystructuredhaplotypenetwork
with no shared haplot ypes between parks. In a previous work
(Álvarez-Presas etal.,2012), we foundthat M. terrestris was di-
vided into t wo highly differentiated groups in the north of the
Iberian peninsula, one east and one west, and hypothesized that
the species had followed the forests history of refugia and re-
colonization during and after the Pleistocene. The western clade
(highly genetically structured) may have remained in refugia in
the Cant abrian and Basque regions while the eastern clade (pre-
senting only a few very similar haplotypes) may have remained in
a very small refuge in the Pyrenees or surrounding area, or even
may have arrived from the east or nor th of Europe after the Last
Glacial Maximum (LGM), as it has been hypothesized for some
plant species (Hewit t, 1999; Petit et al., 2002). As could be ex-
pected, in the present work, the M. terrestris haplotypes of the
two parks corresponded to the respective geographic clades
(Ordesa with the east, west for Picos). In addition, the observed
FIGURE6 Abundance of flatworm
species per plot on each forest type. (a)
Ordesa; (b) Picos Forest types: Fo, beech
Ordesa; Fp, beech Picos; M, mixed; P,
pine; Q, oak.
(a)
(b)
12
|
ÁLVAREZ- PRESA S Et AL.
patterns were different, and Picos was highly structured while
Ordesa shared haplotypes with populations from more northern
localities (UK, France) in a star pattern that is expected for popu-
lations that have recently expanded from a small number of indi-
viduals(Figure4a).Thepresentdatashowthesamepatternata
higher taxonomic level: Picos harbored a higher species diversity
than Ordesa. Most of the species in Picos are phylogenetically
closelyrelated(Figure3,cladeA)andatthesametimequitege-
netically divergent among them. In fac t, of all the MOTUs found
in this park, only those that are common bet ween parks and with
other regionsofEurope (M02, M78 and M71) do not belong to
the same lineage that apparently diversified in the region. Thus,
although we cannot determine exactly the phylogenetic relation-
ships between the endemic MOTUs found in Picos due to the low
support in some basal nodes, we can infer an ancient common
ancestor for all of them. This scenario could be a consequence of
a long process of diversification, not necessarily occurring in the
area occupied now by the park. Picos Valleys may have acted as
refugia for them during the Pleistocene glaciations, as occurred
with M. terrestris diversity, and after that they would have re-
mained restricted there, resulting in their present distribution.
Factors that may have influenced the fact that these species did
not expand from the Picos valleys are probably the inadequacy of
the surrounding forests as habitat for them, or their small popu-
lations. A deeper analysis will have to be under taken to test the
different hypotheses.
On the other hand, the fact that in Ordesa only four species
have been detected, all with a wide distribution range, including the
northernmost populations, may reflect, as in the case of M. terres-
tris eastern clade, that this region would have provided a single mi-
crorefugefromwhich M78 andM02have recolonized Europe.Or,
on the contrary, it could have been recently colonized from eastern
or northern Europe. Unlike M. terrestris, M. cf. aixandrei-2 is also dis-
tributed in the south of the Iberian Peninsula (in Cádiz and Málaga,
Mateosetal., 2017) and in this case,thedifferentregions share at
leastonecommonhaplotype(Figure4b).Thiscouldbe indicatinga
relatively recent expansion, although we cannot know, based on the
available data, what was the origin and direction of this colonization,
whether from north to south or from south to north, and also if the
origin is in the rest of Europe or in the peninsula. We cannot discard,
also, the possibilit y of human transpor t, although the characteristics
of the species and the habitat it occupies (they are very small, white
animals that hide under rocks and rotten trunks in humid forests)
make it unlikely.
Nonetheless, in both parks, we find species with high genetic
diversit y and occupying basal positions in the phylogenetic tree, as
M71inPicosandM01andM25inOrdesa.Thesespeciesmayrep-
resent old lineages with a wide distribution through Europe before
the LGM that may have recovered from different refugia throughout
the continent.
In summar y, historical factors offer plausible explanations to be
considered the drivers of the present differences among regions in
the Iberian Peninsula.
4.3 | Final remarks
The species richness found in Ordesa and Picos may seem low if
compared with similar studies performed in different Brazilian for-
ests, where total flatworm species number ranges between 18 in
areas of deciduous forests (De Castro & Leal- zanchet, 2005) and
51 in areas of Araucaria forest (Leal- Zanchet et al., 2011). Although
these Brazilian species richness result from studies with different
sampling effort with respect to our study, both in the number of
samplingcampaigns(between14and15inBrazil,twointhepre-
sent study) and in the number of people involved in the samplings
(five collectors in Brazil, two collectors in the present study),
the species accumulation models showed both parks are close
to a saturation point, indicating there is a realistic species rich-
ness difference between the areas involved (temperate Europe
vs. neotropical Brazil). Nonetheless, our figures seem extremely
highwhenanalyzedonaEuropeancontext.Till1998,only17spe-
cies of Microplana were known from Europe. Recent works have
lately increasedthe number of species toa total of 43(Mateos
etal.,2017;Sluysetal.,2016;Vila-farré,Mateos,Sluys,&Romero,
2008;Vila-Farré,Sluys,Mateos,Jones,&Romero,2011;Álvarez-
Presas et al. in preparation); nonetheless, this is still poor when
compared to the more than 119 and 98 recorded species from São
Paulo and Florianopolis only in Brazil (Sluys, 1999). These figures
simply confirm the known rule that species richness is much higher
inthetropicsthan in temperateand cold regions(Brown,2014),
being terrestrial planarians one more example to certif y this fact.
Nevertheless, our result s confirm another rule stating that there is
a very large bias in species descriptions according to the number
of taxonomists dedicated to the group: Since we started working
on species identification in Europe, we have seen that the number
of species has increased rapidly.
We aimed to analyze the terrestrial planarian community com-
position in two protected areas to have a view of the original fauna
situation in front of what can be found out of those areas and to
have a first approximation of the value of protec ted areas for soil
communities. Our results allow us to certify that in both parks, ter-
restrial planarians fauna is representative of the diversity found in
their area, and in the case of Picos, it certainly protects a highly en-
demic and at the same time genetically diverse and genealogically
old group of species, which completely fulfills the objectives of the
park. Our results also point to the impor tance of the study of soil
fauna, generally poorly considered when planning protected areas
management or new protected areas.
ACKNOWLEDGMENTS
We thank Amparo Mora Cabello de Alba (Parque Nacional Picos de
Europa), Elena Villagrasa, and Ramón Castillo (Parque Nacional Ordesa
y Monte Perdido) for providing us with the infrastructure and logisti-
cal support necessary to carr y out the sampling campaigns. We also
acknowledge all the guards of both national parks that have accompa-
nied us during our field work days. We thank Laia Leria, Paula Escuer,
|
13
ÁLVAREZ- PRESAS E t AL.
and Arnau Poch for help with fieldwork. Two reviewers (Ana Maria
Leal- Zanchet and Lisandro Negrete) are gratefully acknowledged for
their hel pful comments. This resear ch was financed by the Ministerio de
Agricultura, Alimentación y Medio Ambiente (Spain), within the program
“Ayudas a la investigación en Parques Nacionales” (ref. 589, 2012).
CONFLICT OF INTEREST
None declared.
AUTHOR CONTRIBUTIONS
The three authors contributed to the design of the project. MAP
and EM performed the samplings. MAP was in charge of the labora-
tory work and analysis of molecular data. EM performed numeric
and statistical analyses on communities. The three authors contrib-
uted to final analyses, discussion of results, and writing of the ms.
ORCID
Marta Álvarez-Presas http://orcid.org/0000-0002-4825-9965
Eduardo Mateos http://orcid.org/0000-0001-9741-5744
Marta Riutort http://orcid.org/0000-0002-2134-7674
REFERENCES
Álvare z-Presas, M., A maral, S. V., Carba yo, F., Leal-Zanc het, A. M., &
Riutor t, M. (2015). Focus on the details: Morphological evidence
suppor ts new cryptic land flatworm (Platyhelminthes) species re-
vealed with molecules. Organisms, Diversity, and Evolution, 15,379–
403,https://doi.org/10.1007/s13127-014-0197-z
Álvarez-Presas,M.,Baguñà, J., &Riutort,M.(2008).Molecularphylog-
eny of land and freshwater planarians (Tricladida, Platyhelminthes):
From freshwater to land and back. Molecular Phylogenetics and
Evolution, 47, 555–568.
Álvare z-Pres as, M., Carbayo, F., Rozas , J., & Riutort, M . (2011). Land
planarians (Platyhelminthes) as a model organism for fine- scale phy-
logeographic studies: Understanding patterns of biodiversity in the
Brazilian Atlantic Forest hotspot. Journal of Evolutionary Biology, 24,
887–896.https://doi.org/10.1111/j.1420-9101.2010.02220.x
Ál v are z-P res a s,M . ,Ma teo s,E . ,Vi la- Far r é,M .,S luys ,R., &Ri uto r t ,M. (20 12) .
Evidence for the persistence of the land planarian species Microplana
terrestris (Müller, 1774) (Platyhelminthes, Tricladida) in microrefugia
during the Last G lacial Maximum in the northern section of the Iberian
Peninsula. Molecular Phylogenetics and Evolution, 64,491–499.
Amaral, S. V., Ribeiro, G. G., Valiati, V. H., & Leal-Zanchet, A. M. (2018).
Body doubles: An integrative taxonomic approach reveals new sib-
ling species of land planarians. Invertebra te Systematics, 32,533–550.
htt ps://doi.org/10.1071/IS17046
Antunes, M. B. (20 08). Composição das comunidades de planárias ter-
restres (Platyhelminthes, Tricladida, Terricola) em duas áreas de flo-
resta estacional semidecidual do sul do Br asil Composition in Two
Areas of Semi- Caducifolius Fores t in Southern Brazil. Neotropical
Biology and Conservation, 3,34–38.
Antunes, M. B., Leal-Zanchet, A. M ., & Fonseca, C. R . (2012). Habitat
struc ture, soil properties , and food availability do not predict ter-
restrial flatworms occurrence in A raucaria Forest sites. Pedobiologia
(Jena), 55,25–31.https://doi.org/10.1016/j.pedobi.2011.09.010
Araújo, M. B., Lobo, J. M., & Moreno, J. C. (2007). The effective-
ness of iberian protected areas in conserving terrestrial bio-
diversity. Conservation Biology, 21, 1423–1432 . https://doi.
org /10.1111/j.1523-1739.2007.0 0827.x
Bandelt, H.-J., Forster, P., & Röhl, A. (1999). Median- joining networks for
inferring intraspecific phylogenies. Molecular Biology and Evolution,
16,37–48.
Baptista, V. A., de Matos, L. B ., Fick, I. A., & Leal-Zanchet, A. M .
(2006). Composiç ao das comunidade s de planárias terrestres
(Platyhelminthes, Tricladida, Terricola) do Parque Nacional dos
Aparados da Serra, Brasil. Iheringia Série Zoologia, 96,293–297.
Baptista, V. A., & Leal-Zanchet, A. M. (2010). Land flatworm community
struc ture in a subtropical deciduous forest in Southern Brazil. Belgian
Journal of Zoology, 140,83–90.
Benito Alonso, J. L . (2010). Vegetación del Parque Nacional de
Ordesa y Monte Perdido (Sobrarbe, Pirineo Central Aragonés). In
Publicaciones del Consejo de Protección de la Naturaleza de Aragón,
Serie: Investigación Vol 50 Consejo de Protección de la Naturaleza de
Aragón(Ed.)(pp.324).Zaragoza:GobiernodeAragón.
Boag, B., Jones, H. D., Evans, K . A., Neilson, R., Yeates, G. W., & Johns, P.
M. (1998). The application of GIS techniques to estimate the estab-
lishment and potential spread of Artioposthia triangulata in Scotland.
Pedobiologia (Jena), 42,504–510.
Boag, B., Yeates, G. W., & Johns, P. M. (1998). Limitations to the distribu-
tion and spread of ter restrial flatworms with special reference to the
New Zealand flatworm (Ar tioposthia triangulat a). Pedobiologia (Jena),
42,495–503.
Bouckaert, R ., Heled, J., Kühnert, D., Vaughan, T., Wu, C.-H., Xie, D., …
Drummond, A.J.(2014).BEAST2: A software platformforbayes-
ian evolutionary analysis. PLoS Computational Biology, 10,e10 03537.
https://doi.org/10.1371/journal.pcbi.1003537
Brown,J.H.(2014).Whyaretheresomanyspeciesinthetropics?Journal
of Biogeography, 41, 8–22. https://doi.org/10.1111/jbi.12228
Carbayo,F.,Álvarez-Presas,M.,Jones,H.D., & Riutort, M. (2016). The
true identity of Obama (Platyhelminthes: Geoplanidae) flatworm
spreading across Europe. Zoological Journal of the Linnean Society,
177,5–28.https://doi.org/10.1111/zoj.12358
Carbayo, F., Álvarez-Presas, M., Olivares, C. T., Marques, F. P. L.,
Froehlich, E. M., & Riutort, M. (2013). Molecular phylogeny of
Geoplaninae (Platyhelminthes) challenges current classification:
Proposal of taxonomic actions. Zoologica Scripta, 42, 508–528.
https://doi. org/10.1111/zsc.12019
Carbayo, F., Leal-Zanchet , A. M., & Vieira, E. M. (20 01). Land planarians
(Platyhelminthes: Tricladida: Terricola) as indicators of man- induced
disturbance in a South Brazilian. Belgian Jou rnal of Zoology, 131(Suppl),
223–224.
Carbayo, F., Leal-Zanchet , A. M., & Vieira, E. M. (20 02). Terrestrial flat-
worm (Platyhelminthes: Tricladida: Terricola) diversity versus man-
induced disturbance in an ombrophilous forest in southern Brazil.
Biodiversity and Conservation, 11, 1091–1104. https://doi.org/10.102
3/A:101 58 65 005 604
Chao, A . (1987). Est imating the pop ulation size for c apture-rec apture
data with unequal catchability. Biometrics, 43,783–791.https://doi.
org/10.2307/2531532
Chiu,C.-H.,Wang,Y.-T.,Walther,B.A.,&Chao,A .(2014).AnImproved
nonparametric lower bound of species richness via a modified good-
Turing frequency formula. Biometrics, 70, 671–682. https://doi.
org /10.1111/biom .12200
Christensen, O. M., & Mather, J. G. (1998). The “New Zealand flat-
worm”, Artioposthia triangulata, in Europ e: The Faroese situation.
Pedobiologia (Jena), 42,532–540.
Darriba, D., Taboada, G . L., Doallo, R., & Posada, D. (2012). jModelT-
est 2: More models, new heuristics and parallel computing. Nature
Methods, 9,772–772.https://doi.org/10.1038/nmeth.2109
14
|
ÁLVAREZ- PRESA S Et AL.
De Cast ro, A., & Leal-zanchet, A. M . (2005). Composição de comu-
nidades de planárias terrestres (Platyhelminthes) em áreas de
floresta est acional decidual e de campo na região central do
Rio Grande do Sul, Brasil. Acta Biologica Leopoldensia, 27, 14 7–
150.
Ettema, C. H., & Wardle, D. A. (2002). Spatial soil ecology. Tren ds
in Ecology & Evolution, 17, 177–183. https://doi.org/10.1016/
S0169-5347(02) 02496-5
Felicísimo,A.(1994).El clima de Asturias. Geografía de Asturias (2nd ed.).
Oviedo: Prensa Ibérica, S.A ..
Fick, I . A., Leal-Za nchet, A . M., & Vieir a, E. M. (20 06). Communi ty struc ture
of land flatworms (Platyhelminthes, Ter ricola): Comparisons between
Arauc aria and Atlantic forest in Southern Brazil. Invertebrate Biology,
125,306–313.https://doi.org/10.1111/j.1744-7410.2006.00062.x
Fonseca, C. R., Ganade, G., Baldissera, R., Becker, C. G., Boelter, C. R .,
Brescovit, A. D., … Vieira, E. M. (20 09). Towards an ecologically-
sustainable forestr y in the Atlantic Forest. Biological Conservation,
142,1209–1219.https://doi.org/10.1016/j.biocon.2009.02.017
Hall, T. A. (1999). BioEdit: A user- friendly biological sequence alignment
editor and analysis program for Windows 95/98/NT. Nucleic Acids
Symposium Series, 41, 95–98.
Hewitt, G. M. (1999). Post- glacial re- colonization of European biota.
Biological Journal of the Linnean Society, 68,87–112.
Johns, P. M., Boag, B., & Yeates, G . W. (1998). Observations on the geo-
graphic distribution of flatworms (Turbellaria: Rhynchodemidae,
Bipaliidae, Geoplanidae) in New Zealand. Pedobiologia (Jena), 42,
469–476 .
Jones, H . D. (1988). The status and dis tribution of British terrestrial pla-
narians. Fortschritte der Zoology, 36, 511–516.
Jones, H . D. (1998). The Afric an and European land planarian faunas,
with an identific ation guide for field workers in Europe. Pedobiologia
(Jena), 42,477–489.
Jones, H . D., Green, J., & Pa lin, D. W. (1998). Monthl y abundance, size a nd
maturity in a population of the “Australian Flatworm”, Australoplana
sanguinea alba. Pedobiologia (Jena), 42,477–489.
Katoh, K ., & Standley, D. M. (2013). M AFFT multip le sequence al ign-
ment software version 7: Improvements in performance and us-
ability. Molecular Biology and Evolution, 30, 772–780. https://doi.
org/10.1093/molbev/mst010
Lavelle, P., & Spain, A. V. (2001). Soil ecology. Amsterdam: K luwer
Scientific.
Leal-Zanchet, A. M., & Baptista, V. A.(2009). Planárias terrestres
(Platyhelminthes, Tricladida) em remanescentes de Floresta com
Arauc ária. In C . Fonseca, A. S ouza, A . M. Leal-Zanchet , T. Dutra, A .
Backes, & G . Ganade (Eds.) , Floresta Com Araucária: Ecologia, Conservação
E Desenvolvimento Sustentável(pp.199–207).Holos:RibeirãoPreto.
Leal-Zanchet, A. M., Baptista, V. A., Campos, L. M., & Raffo, J. F. (2011).
Spatial and temporal patterns of land flatworm assemblages in
Brazilian Araucaria forests. Invertebrate Biology, 130,25–33.https://
doi.org/10.1111/j.1744-7410.2010.00215.x
Leal-Zanchet, A. M., & Carbayo, F. (2000). Fauna de planárias terrestres
(Platyhelminthes, Tricladida, Terricola) da Floresta Nacional de São
Francisco de Paula, RS, Brasil: Uma análise preliminar. Acta Biologica
Leopoldensia, 22, 19–25.
Leal-Zanchet, A. M., & Carbayo, F. (2001). Two new species of
Geoplanidae (Platyhelminthes, Tricladida, Terricola) of south
Brazil. Journal of Zoology, 253, 433–4 46. ht tps://doi.org/10.1017/
S0952836901000401
Levene, H. (1960). Robust tests for equality of variances. In I. Olkin, H.
Hotelling, et al. (Eds.), Contrib utions to probabili ty and statistics : Essays
in honor of Harold hotelling (pp. 278–292). Palo Alto, C A: Stanford
University Press.
Mateos, E.,Sluys,R., Riutort, M.,&Álvarez-Presas,M. (2017). Species
richness in the genus Microplana (Platyhelminthes, Tricladida,
Microplaninae) in Europe: As yet no asymptote in sight . Invertebrate
Systematics, 31,269–301.https://doi.org/10.1071/IS16038
Melguizo-Ruiz, N., Verdeny-Vilalt a, O., Arnedo, M. A ., & Moya-Laraño, J.
(2012). Potential drivers of spatial structure of leaf- lit ter food webs
in south- western European beech forests. Pedobiologia (Jena), 55,
311–319.https://doi.org/10.1016/j.pedobi.2012.06.003
Negrete,L.H.L.,Colpo,K.D.,&Brusa,F.(2014).Landplanarianassem-
blages in protected areas of the interior Atlantic forest: Implications
for conservation. PLoS One, 9,e90513.https://doi.org/10.1371/jour-
nal.pone.0090513
Ogren, R . E., & Kawakat su, M. (1998). American nearcti c and neotropic al land
planarian (Tricladida: Terricola) faunas. Pedobiologia (Jena), 42,441–451.
Oksanen, J. (2016). Vegan: Ecological diversity [W WW Document].
Community Ecology Package. Retrieved from http://cran.irsn.fr/
web/packages/vegan/vignettes/diversity-vegan.pdf
Oliveir a, S., Boll , P., dos Anjo s Baptist a, V.,& L eal-Zanchet , A. (2014).
Effects of pine invasion on land planarian communities in an area
covered by Araucaria mois t forest. Zoological Studies, 53(1), 19.
https://doi.org/10.1186/s40555-014-0019-1
Palacios, C. M., Baptista, V. A., & Leal-Zanchet , A. M. (2006). Land Flat worm
(Tricladida: Ter ricola) community structure and composition: comparisons
between dense and mixed ombrophilous forests in Southern Brazil. In
10th International Symposium on Flatworm Biology, Innsbruck. Berichte
DesNaturwissenschaftlichmedizinischenVerein16(Supp).p.72.
Petit, R . J., Aguinagalde, I., De Beaulieu, J.-L., Bittkau, C., Brewer, S.,
Cheddadi,R.,…Vendramin,G.G.(2003). GlacialRefugia: Hotspots
but not melting pots of genetic diversity. Science, 300,1563–1565.
https://doi.org/10.1126/science.1083264
Petit, R . J., Brewer, S., Bordác s, S., & Burg, K. (2002). Identification of
refugia and post- glacial colonisation routes of European white oaks
based on chloroplast DNA and fossil pollen evidence. Forest Ecology
and Management, 156,49–74.
Pfenninger, M., & Posada, D. (2002). Phyloge ographic histor y of the
land snail Candidula unifasciata (Helicellinae, Stylommatophora):
Fragmentation, corridor migration, and secondary contact. Evolution,
56,1776–1788.
Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D. L., Darling, A ., Höhna,
S. ,… Hue ls enb eck ,J.P.(2 012).Mr Bayes3.2 :Effi cie nt Bayes ia nphyl oge -
netic inference and model choice across a large model space. Systematic
Biology, 61,539–542.https://doi.org/10.1093/sysbio/sys029
Sluys, R. (1998). Land planarians (Platyhelminthes, Tricladida, Terricola)
in biodiversity and conservation studies. Pedobiologia (Jena), 42,
490–494.
Sluys, R . (1999). Global diversit y of land planarians (Platyhelminthes,
Tricladida, Terricola): A new indicator- taxon in biodiversity and con-
servation studies. Biodiversity and Conservation, 8,1663–1681.
Sluys, R.,Mateos,E.,Riutor t,M., &Álvarez-Presas,M. (2016).Towards
a comprehensive, integrative analysis of the diver sity of European
microplaninid land flatworms (Platyhelminthes, Tricladida,
Microplaninae), with the description of two peculiar new species.
Systematics and Biodiversity, 14,9–31. https://doi.org/10.1080/1477
2000.2015.1103323
S m i t h , E .P . , &v a n B e l l e , G . ( 19 8 4 ) . N o n p a r a m e t r ic e s t i m a t i o n o f s p e c i e s
richness. Biometrics, 40,119.https://doi.org/10.2307/2530750
Stamatakis, A. (2014).RA xMLversion8: Atool for phylogenetic analy-
sis and pos t- analysis of large phylogenies. Bioinformatics, 30, 1312–
1313.https://doi.org/10.1093/bioinformatics/btu033
Sunnuc ks, P., Blacket, M. J. , Taylor, J. M., S ands, C. J., C iavaglia, S. A ., Garri ck,
R. C. , … Pavlova, A. (20 06). A tale of tw o flatties: Di fferent resp onses of
two terrestrial flatworms to past environmental climatic fluctuations
at Tallaganda in montane southeastern Australia. Molecular Ecology,
15,4513–4531.https://doi.org/10.1111/j.1365-294X.2006.03107.x
Talavera,G.,& Castresana,J.(2007).Improvement ofphylogeniesafter
removing divergent and ambiguously aligned blocks from protein
|
15
ÁLVAREZ- PRESAS E t AL.
sequence alignments. Systematic Biology, 56, 564–577. http s://doi.
org /10.10 80/10635150701472164
Team,T.R . D. C.,2003.R: A language andenvironment for statistical
computing. [W WW Document]. R Found. Stat . Comput. Vienna,
Austria.https://doi.org/isbn3-900051-00-3
Vila-farré, M., Mateos, E., Sluys, R ., & Romero, R. (2008). Peninsula : New
records and description of three new species. Zootaxa, 1739, 1–20 .
Vila-Farré, M., Sluys, R., Mateos, E ., Jones , H. D., & Romero,
R. (2011). Land planarians (Platyhelminthes: Tricladida:
Geoplanidae) from the Iberian Peninsula: New records and
description of two new species, with a discussion on ecology.
Journal of Natural History, 45, 869–891. https://doi.org /10.1080
/00222933.2010.536267
Whitt aker, R. H. (1960). Vegetation of the Siskiyou mountains, Oregon
and California. Ecological Monographs, 30, 279–338. https://doi.
org /10. 2307/1943563
Wilson, M. V., & Shmida, A . (1984). Measuring beta diversity with
presence- absence data. Journal of Ecology, 72, 1055–1062. htt ps://
doi.org /10.2307/2259551
Winsor, L., Johns, P. M., & Yeates, G. W. (1998). Introduction, and eco-
logical and systematic background, to the Terricola (Tricladida).
Pedobiologia (Jena), 42,389–404.
Xia, X . (2017). DAMBE6: New tool s for microbial gen omics, phylo ge-
netics , and molecular evolution. Journal of Heredity, 108, 431–437.
https://doi.org/10.1093/jhered/esx033
Yeates,G.W.,Bo ag ,B .,&Joh ns ,P.M .(1997).Obser vationsonfeeding
and population structure of five New Zealand terrestrial planari-
ans which prey on lumbricid earthworms. Annals of Applied Biology,
131,351–358.
SUPPORTING INFORMATION
Additional suppor ting information may be found online in the
Suppor ting Information section at the end of the article.
How to cite this article:Álvarez-PresasM,MateosE,RiutortM.
Hidden diversity in forest soils: Characterization and
comparison of terrestrial flatworm’s communities in two
national parks in Spain. Ecol Evol. 2018;00:1–15. ht t p s : //d o i.
org /10.1002/ece3.4178
