Spatial variation in the phylogenetic structure of flea assemblages across geographic ranges of small mammalian hosts in the Palearctic Boris R. Krasnov a,⇑ , Shai Pilosof a , Georgy I. Shenbrot a , Irina S. Khokhlova b a Mitrani Department of Desert Ecology, The Swiss Institute for Dryland Environmental and Energy Research, Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede-Boqer Campus, 84990 Midreshet Ben-Gurion, Israel b Wyler Department of Dryland Agriculture, French Associates Institute for Agriculture and Biotechnology of Drylands, Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede-Boqer Campus, 84990 Midreshet Ben-Gurion, Israel article info Article history: Received 19 March 2013 Received in revised form 9 May 2013 Accepted 10 May 2013 Available online 4 June 2013 Keywords: Fleas Geographic range Phylogenetic structure Small mammals abstract We investigated spatial variation in the phylogenetic structure (measured as a degree of phylogenetic clustering) of flea assemblages across the geographic ranges of 11 Palearctic species of small mammalian hosts and asked whether the phylogenetic structure of the flea assemblage of a host in a locality is affected by (i) distance of this locality from the centre of the host’s geographic range, (ii) geographic posi- tion of the locality (distance to the equator) and (iii) phylogenetic structure of the entire flea assemblage of the locality. Our results demonstrated that the key factor underlying spatial variation of the phyloge- netic structure of the flea assemblage of a host was the distance from the centre of the host’s geographic range. However, the pattern of this spatial variation differed between host species and might be explained by their species-specific immunogenetic and/or distributional patterns. Local flea assemblages may also, to some extent, be shaped by environmental filtering coupled with historical events. In addi- tion, the phylogenetic structure of a local within-host flea assemblage may mirror the phylogenetic struc- ture of the entire across-host flea assemblage in that locality and, thus, be affected by the availability of certain phylogenetic lineages. Ó 2013 Australian Society for Parasitology Inc. Published by Elsevier Ltd. All rights reserved. 1. Introduction Spatial variation in the composition of plant and animal com- munities is a central theme in ecological biogeography. Thousands of publications have been dedicated to patterns of spatial variation in species richness and diversity such as latitudinal gradients (e.g., Rohde, 1992), distance decay of similarity (e.g., Nekola and White, 1999) or species–area relationships (e.g., Rosenzweig, 1995). The vast majority of these studies have dealt with free-living species, while parasites have received less attention despite forming a large if not the largest proportion of the diversity of life (Windsor, 1998; Poulin and Morand, 2000, 2004). Nevertheless, the last decade and a half has witnessed a burst of studies on spatial patterns in species diversity and composition in parasite communities (e.g., Poulin and Morand, 1999; Carney and Dick, 2000; Rohde, 2002; Poulin and Valtonen, 2002; Poulin, 2003; Calvete et al., 2004; Krasnov et al., 2005; Oliva and González, 2005; Vinarski et al., 2007; Pérez-del-Olmo et al., 2009), although some pioneering studies were carried out earlier (e.g., Kisielewska, 1970; Kennedy and Bush, 1994). Many spatial patterns found initially for free-living species have been supported by data on parasites. However, some parasite-specific patterns have also been revealed due to the inti- macy of their relationships with their hosts (e.g., Poulin, 2010; Krasnov et al., 2004, 2012). Recently, phylogenetic information has started to be introduced into community ecology and biogeography and has proven to be a powerful tool allowing better understanding of evolutionary pro- cesses involved in the assembly of plant and animal communities and their spatial variation (Webb et al., 2002; Cavender-Bares et al., 2009; Morlon et al., 2011). Unsurprisingly, the studies com- bining phylogenetic data with community ecology and biogeogra- phy have been carried out on free-living species, while the role of phylogeny in determining spatial variation of parasite assemblages remains to be studied, although some initial steps have already been taken (Poulin, 2010; Krasnov et al., 2012). The species composition of a community in a locality is shaped by a variety of ecological and evolutionary factors (Vuilleumier and Simberloff, 1980; Ricklefs, 1987). The parasite assemblage of a par- ticular host in a particular locality is determined by two main com- ponents. One part of an assemblage is due to host identity, while another part is due to the host’s biotic and abiotic environments (Kennedy and Bush, 1994). Some of the parasite species on a host may be inherited from its ancestors, whereas other parasites can switch from other hosts that occupy the same habitat as the focal host (e.g., Paterson and Gray, 1997). In addition, the abiotic envi- 0020-7519/$36.00 Ó 2013 Australian Society for Parasitology Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijpara.2013.05.001 ⇑ Corresponding author. Tel.: +972 8 6596841; fax: +972 8 6596772. E-mail address: [email protected](B.R. Krasnov). International Journal for Parasitology 43 (2013) 763–770 Contents lists available at SciVerse ScienceDirect International Journal for Parasitology journal homepage: www.elsevier.com/locate/ijpara
Transcript
International Journal for Parasitology 43 (2013) 763–770
Contents lists available at SciVerse ScienceDirect
International Journal for Parasitology
journal homepage: www.elsevier .com/locate / i jpara
Spatial variation in the phylogenetic structure of flea assemblages acrossgeographic ranges of small mammalian hosts in the Palearctic
0020-7519/$36.00 � 2013 Australian Society for Parasitology Inc. Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.ijpara.2013.05.001
Boris R. Krasnov a,⇑, Shai Pilosof a, Georgy I. Shenbrot a, Irina S. Khokhlova b
a Mitrani Department of Desert Ecology, The Swiss Institute for Dryland Environmental and Energy Research, Jacob Blaustein Institutes for Desert Research, Ben-Gurion Universityof the Negev, Sede-Boqer Campus, 84990 Midreshet Ben-Gurion, Israelb Wyler Department of Dryland Agriculture, French Associates Institute for Agriculture and Biotechnology of Drylands, Jacob Blaustein Institutes for Desert Research,Ben-Gurion University of the Negev, Sede-Boqer Campus, 84990 Midreshet Ben-Gurion, Israel
a r t i c l e i n f o a b s t r a c t
Article history:Received 19 March 2013Received in revised form 9 May 2013Accepted 10 May 2013Available online 4 June 2013
We investigated spatial variation in the phylogenetic structure (measured as a degree of phylogeneticclustering) of flea assemblages across the geographic ranges of 11 Palearctic species of small mammalianhosts and asked whether the phylogenetic structure of the flea assemblage of a host in a locality isaffected by (i) distance of this locality from the centre of the host’s geographic range, (ii) geographic posi-tion of the locality (distance to the equator) and (iii) phylogenetic structure of the entire flea assemblageof the locality. Our results demonstrated that the key factor underlying spatial variation of the phyloge-netic structure of the flea assemblage of a host was the distance from the centre of the host’s geographicrange. However, the pattern of this spatial variation differed between host species and might beexplained by their species-specific immunogenetic and/or distributional patterns. Local flea assemblagesmay also, to some extent, be shaped by environmental filtering coupled with historical events. In addi-tion, the phylogenetic structure of a local within-host flea assemblage may mirror the phylogenetic struc-ture of the entire across-host flea assemblage in that locality and, thus, be affected by the availability ofcertain phylogenetic lineages.
� 2013 Australian Society for Parasitology Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Spatial variation in the composition of plant and animal com-munities is a central theme in ecological biogeography. Thousandsof publications have been dedicated to patterns of spatial variationin species richness and diversity such as latitudinal gradients (e.g.,Rohde, 1992), distance decay of similarity (e.g., Nekola and White,1999) or species–area relationships (e.g., Rosenzweig, 1995). Thevast majority of these studies have dealt with free-living species,while parasites have received less attention despite forming a largeif not the largest proportion of the diversity of life (Windsor, 1998;Poulin and Morand, 2000, 2004). Nevertheless, the last decade anda half has witnessed a burst of studies on spatial patterns in speciesdiversity and composition in parasite communities (e.g., Poulinand Morand, 1999; Carney and Dick, 2000; Rohde, 2002; Poulinand Valtonen, 2002; Poulin, 2003; Calvete et al., 2004; Krasnovet al., 2005; Oliva and González, 2005; Vinarski et al., 2007;Pérez-del-Olmo et al., 2009), although some pioneering studieswere carried out earlier (e.g., Kisielewska, 1970; Kennedy andBush, 1994). Many spatial patterns found initially for free-livingspecies have been supported by data on parasites. However, some
parasite-specific patterns have also been revealed due to the inti-macy of their relationships with their hosts (e.g., Poulin, 2010;Krasnov et al., 2004, 2012).
Recently, phylogenetic information has started to be introducedinto community ecology and biogeography and has proven to be apowerful tool allowing better understanding of evolutionary pro-cesses involved in the assembly of plant and animal communitiesand their spatial variation (Webb et al., 2002; Cavender-Bareset al., 2009; Morlon et al., 2011). Unsurprisingly, the studies com-bining phylogenetic data with community ecology and biogeogra-phy have been carried out on free-living species, while the role ofphylogeny in determining spatial variation of parasite assemblagesremains to be studied, although some initial steps have alreadybeen taken (Poulin, 2010; Krasnov et al., 2012).
The species composition of a community in a locality is shapedby a variety of ecological and evolutionary factors (Vuilleumier andSimberloff, 1980; Ricklefs, 1987). The parasite assemblage of a par-ticular host in a particular locality is determined by two main com-ponents. One part of an assemblage is due to host identity, whileanother part is due to the host’s biotic and abiotic environments(Kennedy and Bush, 1994). Some of the parasite species on a hostmay be inherited from its ancestors, whereas other parasites canswitch from other hosts that occupy the same habitat as the focalhost (e.g., Paterson and Gray, 1997). In addition, the abiotic envi-
764 B.R. Krasnov et al. / International Journal for Parasitology 43 (2013) 763–770
ronment may act as a filter that excludes some species from a par-asite assemblage due to their environmental requirements (Lebri-ja-Trejos et al., 2010). This is especially important for parasites thatspend some part of their life cycle as free-living (e.g., Krasnov et al.,2001). Thus, processes that shape the parasite assemblage of a hostin a locality have a historical component (associated with parasitesinherited from ancestors), an abiotic environmental component(associated with parasites for which the environment offered bya host is favourable) and a biotic environmental component (asso-ciated with parasites that switch from co-occurring hosts). Consid-ering the variation in phylogenetic structure of parasiteassemblages may allow us to disentangle these components andto identify the predominant force behind assemblage composition.
Here, we investigated spatial variation in the phylogeneticstructure of flea assemblages across the geographic ranges of 11Palearctic species of small mammalian hosts. Fleas are characteris-tic insect ectoparasites of small mammals. Imagos of these insectsare holometabolous obligatory haematophages. Their larvae areusually not parasitic, feed on various kinds of organic matter andreside in the host’s burrow or nest. Abiotic conditions (tempera-ture, humidity, and substrate texture) strongly affect the survival,longevity and reproductive performance of fleas (Krasnov et al.,2001, 2002a,b). Furthermore, there is a substantial difference inabiotic environmental preferences among flea species (see Kras-nov, 2008 for review).
We used a recently proposed index of phylogenetic speciesclustering (Helmus et al., 2007; see details below; see Section 2.4)and asked whether the phylogenetic structure of the flea assem-blage of a host in a locality is affected by (i) distance of this localityfrom the centre of the host’s geographic range, (ii) geographic po-sition of the locality (that is, latitude; measured as distance to theequator) and (iii) phylogenetic structure of the entire flea assem-blage of the locality (that is, flea species recorded on all host spe-cies inhabiting the locality). The relationships between thephylogenetic structure of a host’s flea assemblage in a localityand the centre of the host’s geographic range are likely to reflecthistorical processes involved in the shaping of flea assemblages.In many species, the centre of a geographic range is an area wherea species attains its highest abundance, while abundance decreasestoward the periphery of the range (Hengeveld and Haeck, 1982;Hengeveld, 1990). Although this pattern is not universal (Sagarinand Gaines, 2002; Gaston, 2003; Sagarin, 2006), it is rather wide-spread (Hengeveld, 1990). Among several explanations of the‘‘abundant-centre’’ hypothesis (e.g., Carson, 1959; Brown, 1984;Kirkpatrick and Barton, 1997), the most parsimonious one is thatconditions for survival and reproduction are most favourable atthe centre of the range, and become gradually poorer toward theperiphery (Hengeveld, 1990). The decline in abundance away fromthe centre of the range is often accompanied with increased patch-iness and isolation in peripheral populations (Lawton, 1993). Smalland isolated populations of both hosts and parasites may be sub-jected to random evolutionary forces such as inbreeding and drift(Holt, 1990), experience genetic bottlenecks (Brussard, 1984) andthus be characterised by low genetic diversity. We expected thatflea assemblages of peripheral host populations would be morephylogenetically diverse than those of the central populations be-cause (i) a host may acquire new parasites from different phyloge-netic lineages at the periphery of its range (Hoberg and Brooks,2008a,b), (ii) hosts in isolated populations may be less immuno-competent than those in the core populations (Whiteman et al.,2006), and (iii) parasites in the isolated populations may eventu-ally speciate (Banks and Paterson, 2005).
Relationships between the phylogenetic structure of a host’sflea assemblage in a locality and its geographic position (distanceto the equator) may mirror environmental processes affecting fleaassemblages. Successful development of pre-imaginal fleas takes
place at air temperatures greater than 10–15 �C but lower than30 �C, and relative humidities greater than 60% (Marshall, 1981;Krasnov, 2008). As a result, their geographic distribution in thePalearctic is characterised by peaks of species richness in the tem-perate and steppe zones, with a decrease to the north (tundra andboreal forests) and south (deserts) (Yudin et al., 1976; Medvedev,1996). Given that the southernmost localities in our study didnot include hyperarid areas and desert host species (Krasnovet al., 2010; see Section 2), we expected an increase in phylogeneticclustering of flea assemblages with increasing latitudes because (i)the occurrence of phylogenetically distant lineages is more proba-ble in richer assemblages and (ii) environmental filtering may re-strict flea assemblages in the coldest localities to a certainphylogenetic subset. The association between phylogenetic struc-ture of the local flea assemblage of a host with that of the entireflea community on all flea-supporting host species may be ex-pected if a host’s flea assemblage represents a random sample fromthe surrounding species pool (Krasnov et al., 2004), so that phylo-genetic structure of within-host assemblages correlates positivelywith that of across-host assemblages.
2. Materials and methods
2.1. Selection of data on fleas and small mammals
We used data from our database compiled from published sur-veys of fleas parasitic on small mammals (Soricomorpha, Erinace-omorpha, Rodentia and Lagomorpha) across the Palearctic (60surveys in 52 localities). These surveys reported the number offleas of each individual species found on a given number of individ-uals of each mammalian species. The complete list and geographiclocation of surveys can be found elsewhere (Krasnov et al., 2010;see also Supplementary Fig. S1). We selected host species thatwere recorded in at least six localities, harboured at least six fleaspecies per locality, and for which at least 10 individuals per local-ity were examined. This resulted in datasets of local flea assem-blages for 11 host species (10 rodents and one shrew; seeSupplementary Fig. S1) occurring in six to 20 localities situatedat latitudes from 38�N to 68�N.
2.2. Phylogenetic information
The phylogenetic tree of fleas was based on the only availablemolecular phylogeny of fleas recently constructed by Whitinget al. (2008). This tree includes 128 flea species (ca. 6% of the globalfauna) belonging to 83 genera (ca. 34% of the entire number of fleagenera). Most genera in our dataset were represented by the treepublished by Whiting et al. (2008), but this was not the case forspecies. Consequently, the positions of the species which werenot represented in the original tree of Whiting et al. (2008) weredetermined using their morphologically-derived taxonomy (seedetails in Krasnov et al., 2011). All branch lengths were set equalto 1.0. The tree was ultrametrised using the option ‘‘chronopl’’ inthe package ‘‘ape’’ (2.8) (Paradis et al., 2004) implemented in theR 2.15 statistical environment (R Development Core Team, 2011,http://www.R-project.org).
2.3. Geographic information
To estimate the geographic range of a host species, we appliedspecies distribution modelling based on occurrence records andenvironmental data (see details in Shenbrot and Krasnov, 2005;Shenbrot, in press). In brief, records of occurrences of a specieswere obtained from the Global Biodiversity Information Facility(GBIF; http://data.gbif.org), museum collection databases and pub-
B.R. Krasnov et al. / International Journal for Parasitology 43 (2013) 763–770 765
lished sources. All localities that had no original geo-referencinginformation in data sources were geo-referenced using GeographicNames Gazetteers (http://earth-info.nga.mil/gns/html/cntry_fi-les.html). Data that could not be geo-referenced precisely(±5 km), were excluded. Environmental data for modelling wereused as 30 arc-second grids (approximately 1 km resolution) andwere represented by climate, relief, substrate and vegetation vari-ables. The climate variables were obtained as a part of theWORLDCLIM Version 1.4 (BIOCLIM) package (Hijmans et al.,2005) available at http://www.worldclim.org. Slope data were de-rived from altitude (extracted from a GOTOPO30 data set distrib-uted with an ArcGIS) using the Spatial Analyst module of ArcMapsoftware. The data on the Normalised Difference Vegetation Index(NDVI) were obtained from the VEGETATION Programme (http://www.spot-vegetation.com; http://free.vgt.vito.be; data for 1998–2007, estimations every 10 days) and averaged by seasons (winter,spring, summer and fall) across all available years. The models ofspecies distribution were built with the MAXENT 3.3.3 k software(Phillips et al., 2006).
To delineate the areas of real species occurrence, the originalmodel values, ranging continuously from 0 to 1, were transformedto binary 0 or 1 using a threshold value equal to the smallest prob-ability value that contains all of the observed presences in a region(the ‘‘lowest presence threshold’’; Pearson et al., 2007; McCormacket al., 2009), and the raster was transformed to polygons. Onlypolygons containing occurrence records were considered as areasof occurrence. The area of each polygon included in a geographicrange was calculated in square km using the Calculate Geometrytool of ArcMap 10.1. The centre of a geographic range and its dis-tance to the equator were calculated using Mean Center utilitywith weighting by area (Measuring Geographic Distributions, Spa-tial Statistics Tools, ArcMap 10.1). Both distance measures werelog-transformed prior to analyses.
2.4. Phylogenetic community structure
Various metrics have been proposed to estimate the phyloge-netic structure of a species assemblage (Webb et al., 2002). In gen-eral, these metrics indicate whether species in a community aremore or less phylogenetically related than expected by chance(phylogenetic clustering and phylogenetic overdispersion, respec-tively) or else a community is randomly assembled from a phylo-genetic viewpoint. Initially, we calculated two indices ofphylogenetic structure for each regional flea assemblage of eachhost. These indices were the phylogenetic species variability(PSV) and phylogenetic species clustering (PSC) (Helmus et al.,2007). Both indices compare the expected variance of a neutraltrait that evolves under Brownian motion along the real phyloge-netic tree of species in a community with the variance of a neutraltrait expected if these species would evolve simultaneously fromthe same ancestor, so that their pairwise phylogenetic distanceswould be equal (e.g., star phylogeny). The main differences be-tween the two indices is that PSV considers all species, while PSCtakes into account close relatives only and is thus similar to theNearest Taxon Index (NTI) proposed earlier (Webb, 2000; Webbet al., 2002). In other words, the two indices capture different com-ponents of phylogenetic structure. Values of both indices vary be-tween 0 and 1 with values close to 0 indicating high phylogeneticrelatedness, while maximal values of 1 may be observed only in acommunity of phylogenetically independent species (Helmus et al.,2007). We calculated PSV and PSC using the package ‘‘picante’’(1.5–2) (Kembel et al., 2010) implemented in R 2.15 (R Develop-ment Core Team, 2011; http://www.R-project.org). Hereafter, wefocus on PSC because spatial variation in PSV was rather low (seeSection 3).
2.5. Data analyses
To test whether the phylogenetic structure of each local fleaassemblage of a host species differed significantly from that ex-pected by chance, we compared the observed index (PSC) withthe average of the indices calculated for 500 randomly generatedassemblages. In each replicate, the null assemblage was generatedby random sampling of the observed number of species from thetotal pool of flea species harboured by a given host across its geo-graphic range. For the sake of biological reality, for each given hostspecies we restricted the pool of fleas from which null communi-ties were assembled only to those flea species that were recordedon this host throughout its geographic range. To characterise thephylogenetic structure of each across-hosts flea assemblage, wecalculated its PSC as described above but the null communitieswere generated using the entire pool of flea species recorded inall 52 regions. In subsequent analyses, we used only those fleaassemblages in which phylogenetic structure differed significantlyfrom that expected by chance.
To test for the relationships between the phylogenetic structureof flea assemblages of a given host in a given locality and (i) dis-tance of that locality from the centre of the host’s geographicrange, (ii) distance of the locality from the equator and (iii) phylo-genetic structure of the entire across-hosts flea assemblage, we ap-plied Generalised Linear Models (GLMs) with normal distributionand log-link separately for each host species. The phylogeneticstructure of the across-host flea assemblages in a locality corre-lated with the distance of this locality from the equator (see Sec-tion 3). Consequently, prior to analysis, we substituted theoriginal values of PSC for these flea assemblages with their residualdeviations from the linear regression on log-transformed distanceto the equator. We selected the best model using Akaike’s Informa-tion Criterion (AIC). Then, we further investigated the best modelsand tested for significance of coefficients using Wald statistics.P < 0.05 was considered significant.
3. Results
The phylogenetic structure in terms of both PSV and PSC of theabsolute majority of within-host local flea assemblages differedsignificantly from that expected by chance (PSV: from 62% ofassemblages in Microtus arvalis to 100% of assemblages in Arvicolaamphibius; P < 0.05 for all; and PSC: from 71% of assemblages inMicrotus agrestis to 100% of assemblages in A. amphibius and Crice-tulus migratorius; P < 0.05 for all). The phylogenetic structure of 45(in terms of PSC) or 49 (in terms of PSV) from 52 across-hosts fleaassemblages proved to be significantly different from those ex-pected by chance. Spatial variation of PSV for regional flea assem-blages within a host species was extremely low (coefficient ofvariation within species ranged from 0.84% in Apodemus uralensisto 2.58% in Apodemus agrarius). The same was true for the across-host regional assemblages (coefficient of variation across regionswas 0.84%). Spatial variation of PSC was much higher (within hostspecies: coefficient of variation ranged from 21.99% in Microtusoeconomus to 40.84% in M. agrestis; and across-regions: coefficientof variation 36.23%).
The phylogenetic structure in terms of PSC varied from a mini-mum of 0.08 for fleas on M. agrestis of the middle Ural Mountains(Russia) to a high value of 0.46 for fleas on C. migratorius of TerskeyAlatau Mountains (Tien Shan; Kyrgyzstan). The phylogenetic struc-ture of across-hosts flea assemblages varied from a low of 0.08 forfleas of Kamchatka peninsula (Russia) to a high of 0.43 for fleas ofTarbagatai Mountains (Kazakhstan). PSC of the across-hosts fleaassemblages in a locality correlated negatively with the distanceof that locality to the equator (r2 = 0.16, F1,51 = 8.8, P = 0.005;slope = �0.40 ± 0.13).
766 B.R. Krasnov et al. / International Journal for Parasitology 43 (2013) 763–770
The results of the GLM of the relationships between PSC of thewithin-host flea assemblages and distance of the region from thecentre of the host’s geographic range, distance of the region fromthe equator and residual PSC of the across-hosts flea assemblageare presented in Table 1 (see all tested models and their respectiveAIC values in Supplementary Table S1). In eight of 11 host species,the phylogenetic structure of flea assemblages was associated withthe distance from the centre of the geographic range. Furthermore,in six species, the degree of phylogenetic relatedness of flea assem-blages decreased (as PSC increased) from the centre toward theperiphery of the geographic range (Fig. 1A), while the oppositewas true for two host species (both belonging to the same genus,Myodes) (Fig. 1B). Removal of the data point from the upper rightcorner of the graph of Fig. 1B (flea assemblage of the populationfrom Central Yakutia (Russia)) did not change the relationship be-tween phylogenetic structure of flea assemblages and the centre ofgeographic range of Myodes rutilus (Table 1). Significant relation-ships between PSC of local flea assemblages and latitude (mea-sured as distance to the equator) were found in four hosts.Higher latitudes were associated with a higher degree of phyloge-netic clustering (lower PSC) of fleas harboured (see illustrativeexample with M. rutilus in Fig. 2). PSC of the within-host fleaassemblages were positively correlated with PSC of the entire poolof flea species in a locality (after controlling for the effect of lati-tude) in five host species. In two of these hosts, the phylogeneticstructure of the across-hosts flea assemblages was the only factorassociated with the phylogenetic structure of their local assem-blages (see illustrative example with Myodes glareolus in Fig. 3).
4. Discussion
Our study demonstrated that factors affecting spatial variationin the phylogenetic structure of flea assemblages of a host speciesdiffered among host species. The distance of a local populationfrom the centre of the host’s geographic range appeared to bethe main factor explaining variation in the phylogenetic structureof fleas harboured by this population as its effect was found inthe majority of species. However, the direction of the relationshipwas positive in some species and negative in others. The geo-graphic position of a locality and the phylogenetic structure ofthe entire flea species pool affected the phylogenetic structure ofthe flea assemblage of a host population to a lesser extent.
Spatial variation in the phylogenetic structure of flea assem-blages was manifested mainly in the PSC rather than in the PSV in-dex. This suggests that flea assemblages in different populations of
Table 1Best models explaining the relationships between the phylogenetic structure (measured asand distance of the region from the centre of the host’s geographic range (DC), position offlea assemblage (AHPSC; controlled for confounding effect of distance to Equator; see Sec
AIC, Akaike Information Criterion; L-l v2, log-likelihood v2.a Terms with non-significant coefficients. Only significant intercept terms are shown.b After removal of the data from populations of Central Yakutia (see Section 3 for exp
the same host differed at shallow rather than at deep phylogeneticlevels. The main difference between these two indices is that PSV isrelated to mean phylogenetic distance among all species, while PSCmeasures distance among nearest neighbours (Helmus et al., 2007;Kembel et al., 2010; Gonzalez-Caro et al., 2012). Consequently, thephylogenetic structure of within-host flea assemblages variedacross space due to differences in closely-related species, whilevariation among these assemblages in the presence or absence ofbasal flea lineages was much weaker (if at all).
Lower phylogenetic clustering (higher PSC) in flea assemblagesfarther from the centre of the host’s geographic range may arisedue to the small size, isolation and substantial density fluctuationsin host populations in suboptimal conditions (Lesica and Allendorf,1995; Williams et al., 2003). All of these may facilitate interspecificcontacts and acquisition of ectoparasites from other host species.Fleas can easily be acquired via direct contact (Krasnov andKhokhlova, 2001) or during visits to burrows occupied by otherhosts (Krasnov, 2008). Another reason behind the differences inthe phylogenetic structure of flea assemblages between centraland peripheral host populations might be decreasing genetic diver-sity towards the range periphery, especially if these populationshave experienced genetic bottlenecks (Brussard, 1984; Eckertet al., 2008). In particular, the decreased genetic diversity can bemanifested in decreased polymorphism in the Major Histocompat-ibility Complex (MHC; responsible for acquired immunity) and/orToll-like receptor (TLR; responsible for innate immunity) genes. In-deed, a recent meta-analysis of studies on five fish, one amphibian,four reptile, seven bird and five mammalian species has demon-strated that bottlenecks have resulted in greater loss of MHC diver-sity than neutral genetic diversity (Sutton et al., 2011). However,polymorphism at TLR genes was found to be relatively high in abottleneck population of a New Zealand bird (Grueber et al., 2012).
Interestingly, patterns of geographic variation of diversity inimmune genes appear to be different among species even if theyare ecologically similar and their geographic distributions overlap.Recently, Tschirren et al. (2011a) studied the diversity and popula-tion differentiation in TLR2 genes across several populations ofApodemus flavicollis and M. glareolus. In A. flavicollis, the across-population diversity at TLR was low and one haplotype was pre-dominant in all populations, so that there was no population differ-entiation. On the contrary, populations of M. glareolus differedsubstantially in predominant haplotypes. Moreover, genetic differ-ences in TLR increased with increasing geographic distances be-tween populations in M. glareolus, while this was not the case forA. flavicollis. Species-specific immunogenetic patterns mightunderlie the spatial patterns of variation in the phylogenetic struc-
phylogenetic species clustering (PSC)) of flea assemblages of a given host in a regionthe region relative to the equator (DE) and phylogenetic structure of the across-hoststion 2.5).
Fig. 1. Relationships between the phylogenetic structure of local flea assemblages (measured as phylogenetic species clustering (PSC)) of Apodemus agrarius (A) and Myodesrutilus (B) and distances from the centre of their geographic ranges.
B.R. Krasnov et al. / International Journal for Parasitology 43 (2013) 763–770 767
ture of parasite assemblages found in our study. The only two spe-cies for which an increase of phylogenetic clustering (that is, a de-crease in PSC) with increasing distance from the centre of the rangewas found were congenerics (M. rutilus and Myodes rufocanus). Theexact mechanisms behind a particular species-specific spatial pat-tern of immunogenetic variation and its effect on the structure ofparasite communities are unknown and warrant further investiga-tion. One explanation can be that parasite-mediated selection hasacted differently in different host species. For example, Tschirrenet al. (2011b) reported results suggesting that past positive selec-tion has shaped parts of the TLR genes differently in Myodes voles,on the one hand, and Apodemus mice and Microtus voles, on theother hand.
Contrasting spatial patterns of phylogenetic structure of fleacommunities between M. rufocanus/M. rutilus and other speciesin this study might also be related to the positions of their geo-graphic ranges. Among our study species, M. rufocanus and M. ruti-lus are the only true boreal species with geographic ranges situatedalmost entirely between 50�N and 70�N (Shenbrot and Krasnov,
2005). The majority of the regions where these species were sam-pled and their ectoparasites collected are to the north and to theeast of the centres of their ranges. These regions are characterisedby depauperate flea faunas (e.g., Yudin et al., 1976), so that popu-lations of M. rufocanus and M. rutilus at the northern and easternperiphery of their ranges simply do not have a source pool to diver-sify their flea assemblages. In contrast, such a source may be pres-ent for the southern populations of these hosts.
Although the effect of distance to the equator on the phyloge-netic clustering of flea assemblages was found in four host speciesonly, it was consistently negative and conformed to our expecta-tion. The pattern of this relationship conformed to the above expla-nation, namely the PSC of flea assemblages decreases (that is,phylogenetic clustering increases) to the north. The mechanismsunderlying this pattern could be both ecological and historical.From an ecological perspective, the increasing phylogenetic clus-tering of flea assemblages harboured by the northernmost hostpopulations may be due to environmental filtering that allowsthe occurrence of only those flea species that are resistant to low
Fig. 2. Relationships between the phylogenetic structures of local flea assemblages of Myodes rutilus (measured as phylogenetic species clustering (PSC)) and distance to theequator.
Fig. 3. Relationships between the phylogenetic structure of local flea assemblages (measured as phylogenetic species clustering (PSC)) of Myodes glareolus and thephylogenetic structures of the across-host flea assemblages in a locality (controlled for the distance to the equator; see Section 2.5 for explanations).
768 B.R. Krasnov et al. / International Journal for Parasitology 43 (2013) 763–770
temperatures. For example, among fleas parasitic on rodents innorthern regions, representatives of the ceratophyllid generaMegabothris and Amalaraeus are very common (Yudin et al.,1976; Holland, 1985). However, the temperature preferences ofthese species are unknown, making this explanation rather specu-lative. Nevertheless, earlier we demonstrated that the taxonomicdiversity of flea assemblages of a host often, albeit not always, cor-relates with environmental variables such as air temperature andprecipitation (Krasnov et al., 2005). The results of this study showthat this may also be the case for the phylogenetic structure.
From a historical perspective, the phylogenetic structure of par-asites may follow the phylogenetic structure of their hosts (Kras-
nov et al., 2012). In general, northern assemblages of thePalearctic mammals are less phylogenetically diverse than south-ern assemblages (e.g., Davis and Buckley, 2011). In particular thismay be related to Quaternary climate fluctuations and periodic gla-ciation of the northern part of the Palearctic that have led to thehigh level of diversification of many taxa in southern refugia, whilenorthern species are mainly derived from these refugium popula-tions, expanding their ranges in late glacial and early post-glacialperiods (Hewitt, 1996). Higher phylogenetic clustering of fleas(that is, lower PSC) in both within-host and across-host assem-blages of the northern localities might be a response to the spatialpattern of variation in phylogenetic structure of their hosts. In
B.R. Krasnov et al. / International Journal for Parasitology 43 (2013) 763–770 769
addition, the geographic distribution of extant flea taxa suggeststhat glaciation/post-glaciation cycles could affect flea diversity ina fashion similar to that of their hosts but independently of thehosts’ phylogeographic processes (Medvedev, 1996).
The effect of the phylogenetic structure of the surrounding fleaassemblage on that of the assemblages harboured by individualhost species was consistently positive in all species for which thiseffect was found. In other words, the phylogenetic affinities of theset of flea species harboured by some hosts reflected the availabil-ity of certain flea lineages among all fleas inhabiting the locality.Host species for which the phylogenetic structure of the across-host assemblages was the only factor affecting the phylogeneticstructure of their local assemblages (M. arvalis and M. glareolus)seemed thus to be random samplers of flea lineages from the sur-rounding pools. However, for other host species, the surroundingpool of flea species affected the phylogenetic structure of within-host assemblages together with other factors.
In conclusion, our results demonstrated that the key factorunderlying spatial variation of the phylogenetic structure of theflea assemblage of a host was the distance from the centre of thehost’s geographic range. However, the pattern of this spatial varia-tion differed between host species and might be explained by theirspecies-specific immunogenetic and/or distributional patterns. Lo-cal flea assemblages may also, to some extent, be shaped by envi-ronmental filtering coupled with historical events. In addition, thephylogenetic structure of a local within-host flea assemblage maymirror the phylogenetic structure of the entire across-host fleaassemblage in that locality and, thus, be affected by the availabilityof certain phylogenetic lineages.
Acknowledgements
We thank Robert Poulin for helpful comments on an earlier ver-sion of the manuscript. This study was partly supported by the Is-rael Science Foundation (Grant No. 26/12 to B.R.K. and I.S.K.). S.P.was supported by a fellowship from the Kreitman Foundation, (Is-rael). This is Publication No. 803 of the Mitrani Department of Des-ert Ecology, (Israel).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ijpara.2013.05.001.
References
Banks, J.C., Paterson, A.M., 2005. Multi-host parasite species in cophylogeneticstudies. Int. J. Parasitol. 35, 741–746.
Brown, J.H., 1984. On the relationship between abundance and distribution ofspecies. Am. Nat. 124, 255–279.
Brussard, P.F., 1984. Geographic patterns and environmental gradients: the central–marginal model in Drosophila revisited. Ann. Rev. Ecol. Syst. 15, 25–64.
Calvete, C., Blanco-Aguiar, J.A., Virgós, E., Cabezas-Díaz, S., Villafuerte, R., 2004.Spatial variation in helminth community structure in the red-legged partridge(Alectoris rufa L.): effects of definitive host density. Parasitology 129, 101–113.
Carney, J.P., Dick, T.A., 2000. Helminth community in yellow perch (Perca flavescens(Mitchill)): determinants of pattern. Can. J. Zool. 78, 538–555.
Carson, H.L., 1959. Genetic conditions that promote or retard the formation ofspecies. Cold Spring Harbor Symp. Quant. Biol. 24, 87–103.
Cavender-Bares, J., Kozak, K.H., Fine, P.V.A., Kembel, S.W., 2009. The merging ofcommunity ecology and phylogenetic biology. Ecol. Lett. 12, 693–715.
Davis, T.J., Buckley, L.B., 2011. Phylogenetic diversity as a window into theevolutionary and biogeographic histories of present-day richness gradients formammals. Phil. Trans. R. Soc. Lond. B 366, 2414–2425.
Eckert, C.G., Samis, K.E., Lougheed, S.C., 2008. Genetic variation across species’geographical ranges: the central–marginal hypothesis and beyond. Mol. Ecol.17, 1170–1188.
Gaston, K.J., 2003. The Structure and Dynamics of Geographic Ranges. OxfordUniversity Press, Oxford.
Gonzalez-Caro, S., Parra, J.L., Graham, C.H., McGuire, J.A., Cadena, D.C., 2012.Sensitivity of metrics of phylogenetic structure to scale, source of data andspecies pool of hummingbird assemblages along elevational gradients. PLoSOne 7, e35472.
Grueber, C.E., Wallis, G.P., King, T.M., Jamieson, I.G., 2012. Variation at innateimmunity Toll-like receptor genes in a bottlenecked population of a NewZealand robin. PLos One 7, e45011.
Hengeveld, R., 1990. Dynamic Biogeography. Cambridge University Press,Cambridge.
Hengeveld, R., Haeck, J., 1982. The distribution of abundance. I. Measurements. J.Biogeogr. 5, 303–316.
Hewitt, G.M., 1996. Some genetic consequences of ice ages, and their role indivergence and speciation. Biol. J. Linn. Soc. 58, 247–276.
Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very highresolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25,1965–1978.
Hoberg, E.P., Brooks, D.R., 2008a. A macroevolutionary mosaic: episodic host-switching, geographical colonization and diversification in complex host–parasite systems. Ecol. Lett. 35, 1533–1550.
Hoberg, E.P., Brooks, D.R., 2008b. Beyond vicariance: integrating taxon pulses,ecological fitting, and oscillation in evolution and historical biogeography. In:Morand, S., Krasnov, B.R. (Eds.), The Biogeography of Host–Parasite Interactions.Oxford University Press, Oxford, pp. 7–20.
Holland, G.P., 1985. The fleas of Canada, Alaska and Greenland (Siphonaptera).Mem. Entomol. Soc. Can. 130, 1–631.
Kennedy, C.R., Bush, A.O., 1994. The relationship between pattern and scale inparasite communities: a stranger in strange land. Parasitology 109, 187–196.
Kirkpatrick, M., Barton, N.H., 1997. Evolution of a species’ range. Am. Nat. 150, 1–23.Kisielewska, K., 1970. Ecological organization of intestinal helminth groupings in
Clethrionomys glareolus (Schreb.) (Rodentia). III. Structure of helminth groupingsin C. glareolus populations of various forest biocoenoses in Poland. ActaParasitol. 18, 163–176.
Krasnov, B.R., 2008. Functional and evolutionary ecology of fleas: a model forecological parasitology. Cambridge University Press, Cambridge.
Krasnov, B.R., Khokhlova, I.S., 2001. The effect of behavioural interactions on theexchange of flea (Siphonaptera) between two rodent species. J. Vector Ecol. 26,181–190.
Krasnov, B.R., Khokhlova, I.S., Fielden, L.J., Burdelova, N.V., 2001. The effect of airtemperature and humidity on the survival of pre-imaginal stages of two fleaspecies (Siphonaptera: Pulicidae). J. Med. Entomol. 38, 629–637.
Krasnov, B.R., Khokhlova, I.S., Fielden, L.J., Burdelova, N.V., 2002a. The effect ofsubstrate on survival and development of two species of desert fleas(Siphonaptera: Pulicidae). Parasite 9, 135–142.
Krasnov, B.R., Khokhlova, I.S., Fielden, L.J., Burdelova, N.V., 2002b. Time to survivalunder starvation in two flea species (Siphonaptera: Pulicidae) at different airtemperatures and relative humidities. J. Vector Ecol. 27, 70–81.
Krasnov, B.R., Shenbrot, G.I., Khokhlova, I.S., Degen, A.A., 2004. Flea species richnessand parameters of host body, host geography and host ‘‘milieu’’. J. Anim. Ecol.73, 1121–1128.
Krasnov, B.R., Shenbrot, G.I., Mouillot, D., Khokhlova, I.S., Poulin, R., 2005. Spatialvariation in species diversity and composition of flea assemblages in smallmammalian hosts: geographic distance or faunal similarity? J. Biogeogr. 32,633–644.
Krasnov, B.R., Mouillot, D., Shenbrot, G.I., Khokhlova, I.S., Vinarski, M.V., Korallo-Vinarskaya, N.P., Poulin, R., 2010. Similarity in ectoparasite faunas of Palaearcticrodents as a function of host phylogenetic, geographic, or environmentaldistances: which matters the most? Int. J. Parasitol. 40, 807–817.
Krasnov, B.R., Stanko, M., Khokhlova, I.S., Shenbrot, G.I., Morand, S., Korallo-Vinarskaya, N.P., Vinarski, M.V., 2011. Nestedness and beta-diversity inectoparasite assemblages of small mammalian hosts: effects of parasiteaffinity, host biology and scale. Oikos 120, 630–639.
Krasnov, B.R., Mouillot, D., Khokhlova, I.S., Shenbrot, G.I., Poulin, R., 2012.Compositional and phylogenetic dissimilarity of host communities drivescompositional and phylogenetic dissimilarity of ectoparasite assemblages:geographic variation and scale-dependence. Parasitology 139, 338–347.
Lawton, J.H., 1993. Range, population abundance and conservation. Trends Ecol.Evol. 8, 409–413.
Lebrija-Trejos, E., Pérez-García, E.A., Meave, J.A., Bongers, F., Poorter, L., 2010.Functional traits and environmental filtering drive community assembly in aspecies-rich tropical system. Ecology 91, 386–398.
Lesica, P., Allendorf, F.W., 1995. When are peripheral populations valuable forconservation. Conserv. Biol. 9, 753–760.
Marshall, A.G., 1981. The Ecology of Ectoparasitic Insects. Academic Press, London.McCormack, J.E., Zellmer, A.J., Knowles, L.L., 2009. Does niche divergence
accompany allopatric divergence in Aphelocoma jays as predicted underecological speciation? Insights from tests with niche models. Evolution 64,1231–1244.
Medvedev, S.G., 1996. Geographical distribution of families of fleas (Siphonaptera).Entomol. Rev. 76, 978–992.
Nekola, J.C., White, P.S., 1999. The distance decay of similarity in biogeography andecology. J. Biogeogr. 26, 867–878.
Oliva, M.E., González, M.T., 2005. The decay of similarity over geographical distancein parasite communities of marine fishes. J. Biogeogr. 32, 1327–1332.
Paradis, E., Claude, J., Strimmer, K., 2004. APE: analyses of phylogenetics andevolution in R language. Bioinformatics 20, 289–290.
Paterson, A.M., Gray, R.D., 1997. Host–parasite co-speciation, host switching, andmissing the boat. In: Clayton, D.H., Moore, J. (Eds.), Host–Parasite Evolution:General Principles and Avian Models. Oxford University Press, Oxford, pp. 236–250.
Pearson, R.G., Raxworthy, C.J., Nakamura, M., Peterson, A.T., 2007. Predicting speciesdistributions from small numbers of occurrence records: a test case usingcryptic geckos in Madagascar. J. Biogeogr. 34, 102–117.
Pérez-del-Olmo, A., Fernández, M., Raga, J.A., Kostadinova, A., Morand, S., 2009. Noteverything is everywhere: the distance decay of similarity in a marine host–parasite system. J. Biogeogr. 36, 200–209.
Poulin, R., 2003. The decay of similarity with geographical distance in parasitecommunities of vertebrate hosts. J. Biogeogr. 30, 1609–1615.
Poulin, R., 2010. Decay of similarity with host phylogenetic distance in parasitefaunas. Parasitology 137, 733–741.
Poulin, R., Morand, S., 1999. Geographic distances and the similarity among parasitecommunities of conspecific host populations. Parasitology 119, 369–374.
Poulin, R., Morand, S., 2000. The diversity of parasites. Quart. Rev. Biol. 75, 277–293.Poulin, R., Morand, S., 2004. Parasite Biodiversity. Smithsonian Books, Washington.Poulin, R., Valtonen, T., 2002. The predictability of helminth community structure in
space: a comparison of fish populations from adjacent lakes. Int. J. Parasitol. 32,1235–1243.
Ricklefs, R.E., 1987. Community diversity: relative roles of local and regionalprocesses. Science 235, 167–171.
Rohde, K., 1992. Latitudinal gradients in species diversity: the search for theprimary cause. Oikos 65, 514–527.
Rohde, K., 2002. Ecology and biogeography of marine parasites. Adv. Marine Biol.43, 1–86.
Rosenzweig, M.L., 1995. Species Diversity in Space and Time. Cambridge UniversityPress, Cambridge.
Sagarin, R.D., 2006. Recent studies improve understanding of population dynamicsacross species ranges. Oikos 115, 386–388.
Sagarin, R.D., Gaines, S.D., 2002. The ‘‘abundant centre’’ distribution: to what extentis it a biogeographical rule? Ecol. Lett. 5, 137–147.
Shenbrot, G.I., in press. Geographic range, morphometric variation and nichedifferentiation in greater Egyptian gerboa, Jaculus orientalis. Mammalia http://dx.doi.org/10.1515/mammalia-2012-0035.
Shenbrot, G.I., Krasnov, B.R., 2005. An Atlas of the Geographic Distribution of theArvicoline Rodents of the World (Rodentia, Muridae: Arvicolinae). PensoftPublishers, Sofia, Bulgaria.
Sutton, J.T., Nakagawa, S., Robertson, B.C., Jamieson, I.G., 2011. Disentangling theroles of natural selection and genetic drift in shaping variation at MHCimmunity genes. Mol. Ecol. 20, 4408–4420.
Tschirren, B., Andersson, M., Scherman, K., Westerdahl, H., Råberg, L., 2011a.Contrasting patterns of diversity and population differentiation at the innateimmunity gene Toll-like receptor 2 (TLR2) in two sympatric rodent species.Evolution 66, 720–731.
Tschirren, B., Råberg, L., Westerdahl, H., 2011b. Signatures of selection acting on theinnate immunity gene Toll-like receptor 2 (TLR2) during the evolutionaryhistory of rodents. J. Evol. Biol. 24, 1232–1240.
Vinarski, M.V., Korallo, N.P., Krasnov, B.R., Shenbrot, G.I., Poulin, R., 2007. Decay ofsimilarity of gamasid mite assemblages parasitic on Palaearctic smallmammals: geographic distance, host species composition or environment? J.Biogeogr. 34, 1691–1700.
Vuilleumier, F., Simberloff, D., 1980. Ecology versus history as determinants ofpatchy and insular distributions in high Andean birds. Evol. Biol. 12, 235–379.
Webb, C.O., 2000. Exploring the phylogenetic structure of ecological communities:an example for rain forest trees. Am. Nat. 156, 145–155.
Whiteman, N.K., Matson, K.D., Bollmer, J.L., Parker, P.G., 2006. Disease ecology in theGalápagos Hawk (Buteo galapagoensis): host genetic diversity, parasite load andnatural antibodies. Proc. R. Soc. Lond. B 273, 797–804.
Whiting, M.F., Whiting, A.S., Hastriter, M.W., Dittmar, K., 2008. A molecularphylogeny of fleas (Insecta: Siphonaptera): origins and host associations.Cladistics 24, 677–707.
Williams, C.K., Ives, A.R., Applegate, R.D., 2003. Population dynamics acrossgeographical ranges: time-series analyses of three small game species.Ecology 84, 2654–2667.
Windsor, D.A., 1998. Most of the species on Earth are parasites. Int. J. Parasitol. 28,1939–1941.
Yudin, B.S., Krivosheev, V.G., Belyaev, V.G., 1976. Small Mammals of the NorthernFar East. Nauka Siberian Branch, USSR, Novosibirsk (in Russian).
