Ecological interactions are evolutionarily conserved across the entire tree of life.
ABSTRACT Ecological interactions are crucial to understanding both the ecology and the evolution of organisms. Because the phenotypic traits regulating species interactions are largely a legacy of their ancestors, it is widely assumed that ecological interactions are phylogenetically conserved, with closely related species interacting with similar partners. However, the existing empirical evidence is inadequate to appropriately evaluate the hypothesis of phylogenetic conservatism in ecological interactions, because it is both ecologically and taxonomically biased. In fact, most studies on the evolution of ecological interactions have focused on specialized organisms, such as some parasites or insect herbivores, belonging to a limited subset of the overall tree of life. Here we study the evolution of host use in a large and diverse group of interactions comprising both specialist and generalist acellular, unicellular and multicellular organisms. We show that, as previously found for specialized interactions, generalized interactions can be evolutionarily conserved. Significant phylogenetic conservatism of interaction patterns was equally likely to occur in symbiotic and non-symbiotic interactions, as well as in mutualistic and antagonistic interactions. Host-use differentiation among species was higher in phylogenetically conserved clades, irrespective of their generalization degree and taxonomic position within the tree of life. Our findings strongly suggest a shared pattern in the organization of biological systems through evolutionary time, mediated by marked conservatism of ecological interactions among taxa.
- SourceAvailable from: Bo Dalsgaard[Show abstract] [Hide abstract]
ABSTRACT: Modularity is a recurrent and important property of bipartite ecological networks. Although well-resolved ecological networks describe interaction frequencies between species pairs, modularity of bipartite networks has been analysed only on the basis of binary presence-absence data. We employ a new algorithm to detect modularity in weighted bipartite networks in a global analysis of avian seed-dispersal networks. We define roles of species, such as connector values, for weighted and binary networks and associate them with avian species traits and phylogeny. The weighted, but not binary, analysis identified a positive relationship between climatic seasonality and modularity, whereas past climate stability and phylogenetic signal were only weakly related to modularity. Connector values were associated with foraging behaviour and were phylogenetically conserved. The weighted modularity analysis demonstrates the dominating impact of ecological factors on the structure of seed-dispersal networks, but also underscores the relevance of evolutionary history in shaping species roles in ecological communities.Ecology Letters 01/2014; · 17.95 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Understanding the evolution of specialization in host plant use by pollinators is often complicated by variability in the ecological context of specialization. Flowering communities offer their pollinators varying numbers and proportions of floral resources, and the uniformity observed in these floral resources is, to some degree, due to shared ancestry. Here, we find that pollinators visit related plant species more so than expected by chance throughout 29 plant–pollinator networks of varying sizes, with “clade specialization” increasing with community size. As predicted, less versatile pollinators showed more clade specialization overall. We then asked whether this clade specialization varied with the ratio of pollinator species to plant species such that pollinators were changing their behavior when there was increased competition (and presumably a forced narrowing of the realized niche) by examining pollinators that were present in at least three of the networks. Surprisingly, we found little evidence that variation in clade specialization is caused by pollinator species changing their behavior in different community contexts, suggesting that clade specialization is observed when pollinators are either restricted in their floral choices due to morphological constraints or innate preferences. The resulting pollinator sharing between closely related plant species could result in selection for greater pollinator specialization.Ecology and Evolution 04/2014; · 1.66 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Abstract We discuss approaches to agent-based model visualization. Agent-based modeling has its own requirements for visualization, some shared with other forms of simulation software, and some unique to this approach. In particular, agent-based models are typified by complexity, dynamism, nonequilibrium and transient behavior, heterogeneity, and a researcher's interest in both individual- and aggregate-level behavior. These are all traits requiring careful consideration in the design, experimentation, and communication of results. In the case of all but final communication for dissemination, researchers may not make their visualizations public. Hence, the knowledge of how to visualize during these earlier stages is unavailable to the research community in a readily accessible form. Here we explore means by which all phases of agent-based modeling can benefit from visualization, and we provide examples from the available literature and online sources to illustrate key stages and techniques.Artificial Life 02/2014; · 1.59 Impact Factor
Ecological interactions are evolutionarily conserved
across the entire tree of life
Jose ´ M. Go ´mez1, Miguel Verdu ´2& Francisco Perfectti3
Ecological interactions are crucial to understanding both the eco-
logy and the evolution of organisms1,2. Because the phenotypic
traits regulating species interactions are largely a legacy of their
logenetically conserved, with closely related species interacting
with similar partners2. However, the existing empirical evidence
is inadequate to appropriately evaluate the hypothesis of phylo-
genetic conservatism in ecological interactions, because it is both
ecologically and taxonomically biased. In fact, most studies on the
evolution of ecological interactions have focused on specialized
to a limited subset of the overall tree of life. Here we study the
evolution of host use in a large and diverse group of interactions
comprising both specialist and generalist acellular, unicellular and
multicellular organisms. We show that, as previously found for
specialized interactions, generalized interactions can be evolutio-
narily conserved. Significant phylogenetic conservatism of inter-
action patterns was equally likely to occur in symbiotic and
non-symbiotic interactions, as well as in mutualistic and antagon-
istic interactions. Host-use differentiation among species was
higher in phylogenetically conserved clades, irrespective of their
generalization degree and taxonomic position within the tree of
tion of biological systems through evolutionary time, mediated by
marked conservatism of ecological interactions among taxa.
Shared ancestry may produce ecological similarity, with closely
related species having similar ecological niches8,9. This idea may be
is most severe among related species because they have similar pheno-
types and niche requirements10. Interspecific interactions comprise a
substantial part of the niche of most species11. Conventional wisdom
with similar organisms than would species that are remotely related,
because the phenotypic traits that regulate the interactions are often
other niche components, ecological interactions are evolutionarily
We explored this idea by compiling information from 116 clades
belonging to seven kingdoms (Euryarchaeota, Bacteria, Excavata,
Chromalveolata, Fungi, Plantae and Animalia) from the three cel-
lular domains (Archaea, Bacteria and Eukarya) and RNA and DNA
viruses (Supplementary Data, appendix 1). We chose these systems
because (1) they contain all types of ecological interaction, from
antagonism (for example, endophytic herbivory, folivory and para-
sitism) to mutualism (for example, pollination, mycorrhiza, seed
dispersal and nitrogen fixing); (2) by exploring organisms from dis-
parate portions of the tree of life, our data set avoids taxonomic and
systematic biases; (3) they comprise a wide range of generalization/
specialization degree; (4) reliable records of interacting organisms
(hereafter referred to as hosts for the sake of simplicity) are available
tary Data, appendix 1). We have used genus as our target clade level
because it is the taxonomic level at which interaction-mediated spe-
ciation mostly manifests2,7(Methods).
The host range of the studied clades, calculated as the average
number of organisms interacting with each species of that clade,
varied from 1 (extreme specialization) to 11.2. Nevertheless, within
most systems there were species interacting with many hosts (up to
50) coexisting with species interacting with very few hosts (Sup-
plementary Data, appendix 1). Of the studied clades, 58% (N567)
were generalist (host range, $1.5 hosts per species). Specialization
depended significantly on taxonomic affiliation: 95% of the viruses
but only 53% of the eukaryotes and 48% of the prokaryotes were
specialist. No other system characteristic affected specialization
degree (Supplementary Table 1), suggesting that our distribution
of host range across genera was not biased by the sampling intensity
of the original data set.
Tracking theevolutionaryhistory ofspecialized interactions isnot
difficult, and has been performed for different kinds of interaction.
Becauseinextremely specialized cladesthehostrangeisvery narrow,
it is easy to identify the host for each species in the phylogeny and to
quantify host shifts and host conservatism. Specialized interactions
are conserved when there is non-independence in host use among
species within a clade owing to their phylogenetic relatedness. This
the tendency for related species to resemble each other in interaction
patterns more than species randomly drawn from the phylogenetic
tree do12. However, the use of this approach becomes increasingly
difficult, to the point of becoming unfeasible, as the diversity of
organisms interacting with the focal clade increases.
Generalist species interact with many other species, and therefore
form networks of interacting organisms. Network analysis has been
successfully used to analyse complex ecological interactions13,14. On
the basis of the pattern of shared interactions, species can be grouped
in compartments or modules. Species are tightly linked if they share a
high proportion of interactions, that is, if they are ecologically similar.
Groups of species interacting with similar organisms form modules
within the general network15,16. Significant modularity emerges in a
network when distinct groups of species closely share links with each
other more than with species in other modules17. Using a network
approach, we explore the evolution of ecological interactions by track-
ing the changes in module affiliation across the phylogenies (Fig. 1).
dynamics of ecological communities18,19. However, rather than taking
the standard perspective on building networks, we use clade-oriented
1Departamento de Ecologı ´a, Universidad de Granada, E-18071 Granada, Spain.2Centro de Investigaciones sobre Desertificacio ´n, Consejo Superior de Investigaciones Cientı ´ficas-
Universidad de Valencia-Generalitat Valenciana, E-46470 Valencia, Spain.3Departamento de Gene ´tica, Universidad de Granada, E-18071 Granada, Spain.
Vol 465|17 June 2010|doi:10.1038/nature09113
Macmillan Publishers Limited. All rights reserved
networks (that is, groups of phylogenetically related species sharing a
For extremely specialized clades, those showing high host specificity,
each module contains the group of species interacting with the same
host (Fig. 1a). Exploring phylogenetic conservatism in module ascrip-
tion is analogous in these types of system to exploring phylogenetic
conservatism in host use using the standard methodology. In clades
in which species interact with more than one host, detecting modules
of species sharing similar hosts allows for the exploration of phylo-
genetic conservatism even when it is not possible to group species
according to their exact equivalence in host use (Fig. 1b). From this
occurs in module affiliation.
For each clade we built a network (Fig. 1), including as nodes only
species were linked when they shared at least one host. We then used
simulated annealing to establish significant modules within each
network15,16. A modular network is one in which the clade is made
up of species that can be grouped according to their affinity in host
use. All but six clades were significantly modular (Supplementary
Table 2). The modules of all studied clades, both specialist and
generalist, differed significantly in identity and composition of the
host assemblage (P,0.01 for all systems, from multivariate analyses
tary Table 3). Furthermore, modules within a network did not differ
among themselves in number of hosts (P.0.05 for most systems;
Supplementary Table 3), meaning that the emergence of modules in
generalized clades was not due to differences in host range across
species. Together, these results show that modules describe distinct
and discrete interactive niches both in specialized and in generalized
The number of modules in a given network may be considered a
measure of the diversity of interactive niches occupied by the genus.
The number of modules per network ranged from 2 to 20. It was
affected by the interaction intimacy, with symbiotic genera having
more modules (6.260.5, N575 clades) than did non-symbiotic
ones (4.260.4, N540 clades). The number of modules was nega-
tively related to host range, even after controlling for number of
species in the clade (Supplementary Table 4). This means that the
number of distinct interacting niches is higher in specialized genera
than in generalized ones. Domain also affected the number of
modules per system, with viruses and prokaryotes having more
modules per clade than did eukaryotes (Supplementary Tables 4
and 5). This may reflect a trend towards greater diversification of
ecological niches and specialization in microorganisms.
For each system, we calculated the modularity index, M, which
estimates how clearly delimited the modules are16. This index
decreases when the fraction of between-module links increases in
different modules but share hosts. Consequently, it can be used as an
estimate of the between-module differentiation in host use. Low
between different modules, and high values of M indicate high dif-
ferentiation because the modules do not use common hosts. The
extreme situation is exemplified by those genera in which most
modules are completely isolated, without any links with the remain-
ing modules (Supplementary Fig. 1). In our data set, modularity
ranged from 0 to 0.833 (Supplementary Appendix 1). The value of
M was higher in specialist clades (0.39060.026) than in generalist
ones (0.23260.030). There was indeed a significant negative rela-
tionship between modularity and host range across clades (Sup-
plementary Table 4). This means that between-module differenti-
ation in host use decreases with the generalization of the clades. In
fact, in community networks modularity is expected to increase
with host specificity17. Similarly, M was higher in symbiotic
(0.36360.030) clades than in non-symbiotic ones (0.23960.034;
Supplementary Table 3), probably because symbionts tend to be
more specialized (host range, 1.7760.17) than do non-symbionts
(host range, 2.9560.40) (F59.63, d.f.51,108, P50.002, from
To explore how evolutionarily conserved ecological interactions
are, we statistically tested whether phylogenetically related species
were more prone to belonging to the same module than would be
expected randomly21(that is, we tested for phylogenetic signal of
ecological interactions). We found that over 83% of the specialist
clades showed a significant phylogenetic signal for host use (Sup-
plementary Data, appendix 1). Furthermore, 52% of the generalist
host range to have no effect on the probability of there being signifi-
cant phylogenetic conservatism in ecological interaction (Fig. 2 and
ecological interactions did not depend on the sign of the interaction
(Supplementary Table 6),as 69%ofantagonistic systemsand59%of
mutualistic systems had a significant phylogenetic signal. That is,
parasites and predators have the same probability of having phylo-
genetically conserved ecological interactions as do pollinators, seed
dispersers or mycorrhizae.
There was a slight tendency of symbiotic systems to have more-
conserved interactions (71% of the symbiotic systems had a signifi-
cantphylogenetic signal)thandonon-symbiotic systems(57%hada
significant phylogenetic signal), although this difference was not
statistically significant (Supplementary Table 6). Similarly, there
was a tendency for phylogenetic conservatism to be more frequent
in viruses (85%) and prokaryotes (80%) than in eukaryotes (59%),
although this difference was also nonsignificant. Finally, the occur-
rence of a phylogenetic signal in our data set was significantly related
Infact, if weremovefrom ourdata set thoseclades withfewer than20
of generalist clades (N533 systems) had a significant phylogenetic
a Specialized clade
b Generalized clade
Figure 1 | How to study the evolution of both specialized and generalized
interactions. a, In specialized clades, each species (numbered tips in the
host used. Phylogenetic conservatism is determined by mapping such
groups onto the phylogeny. b, In contrast, grouping species according to
host use is more complex in generalized clades. Network analysis allows the
themselves than they do with species in other modules. Phylogenetic
conservatism of host use is determined by mapping such modules onto the
phylogeny. Consequently, this method allows the exploration of the
evolutionary conservatism of ecological interactions in all types of system,
irrespective of their degree of generalization or host specificity.
NATURE|Vol 465|17 June 2010
Macmillan Publishers Limited. All rights reserved
RNA viruses to herbivorous insects3–7,22–24. In agreement with this
traditional view, several studies have pointed out that co-cladogenesis
and phylogenetic conservatism in ecological interaction disappears
when generalist species are included in the analyses7,24. Our study
indicates, however, that ecological interactions are also conserved in
generalist (both symbiotic and non-symbiotic) clades. Evolutionary
conservatism in ecological interactions is a recurrent phenomenon
across the entire tree of life (Fig. 3).
Clades with a significant phylogenetic signal in ecological interac-
tions also had higher values of modularity, and this occurred in all
kinds of organism (virus, prokaryote and eukaryote) and interaction
ary conservatism in their ecological interactions also have stronger
differentiation in host use among modules. This means that species
modules in conserved systems, whereas in non-conserved systems
species belonging to different modules tend to share some hosts.
This probably occurs because the use of a specific host assemblage
requires particular adaptations. In clades in which modules are con-
served, species retain ancestral traits that influence their ecological
interactions7,23, constraining the present and future capacity to use
alternative hosts from other modules1. In contrast, in non-conserved
systems most traits involved in host use are likely to represent new
using alternative and disparate hosts. It is remarkable that this rela-
tionship was also found for generalist clades (Fig. 2f), despite the fact
that modularity is negatively related to host range. That is, although
generalist species usually share some hosts among different modules,
among-module differentiation in the composition of their host
assemblages is higher in evolutionarily conserved interactions.
Conservatism in ecological interactions is associated with high host-
use differentiation both in generalist and specialist organisms.
Our study has demonstrated that phylogenetic conservatism in
ecological interactions is a general pattern occurring in many taxa
belonging to very separate branches of the entire tree of life, from
viruses to animals, and in most types of interaction, from specialized
symbiotic antagonisms to generalized non-symbiotic mutualisms.
The same rules seem to drive the evolution of most ecological inter-
We constructed bipartite networks (N5116 systems) of species belonging to the
same genus and their known hosts. Species were then connected through the co-
number of modules per network were determined using an algorithm based on
simulated annealing15,16. This algorithm identifies modules, which are groups of
species having most of their links within their own module, with an accuracy of
90% (ref. 16). The modules were validated statistically by permutational multi-
variate analyses of variance based on species composition dissimilarity (using the
function ‘adonis’ in the R package VEGAN)20. We determined phylogenetic con-
servatism in host use in each system by estimating the significance of the phylo-
genetic signal following ref. 25. The character ‘host use’ was the module to which
the species was ascribed by the annealing algorithm. We mapped the evolution of
host use onto published phylogenetic trees.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 2 February; accepted 22 April 2010.
Published online 2 June 2010.
1.Price, P. W. Macroevolutionary Theory on Macroecological Patterns (Cambridge
Univ. Press, 2003).
Sasal, P. Desdevises, Y. & Morand, S. Host-specialization and species diversity in
fish parasites: phylogenetic conservatism? Ecography 21, 639–643 (1998).
Figure 3 | Ecological interactions are evolutionarily conserved across the
entire tree of life. The phylogenetic tree shows the evolutionary
of the ecological interactions mapped onto the tips. Red names represent
genera with conserved ecological interactions (that is, with a significant
phylogenetic signal) and blue names represent genera with non-conserved
ecological interactions (that is, with a nonsignificant phylogenetic signal).
Black names represent genera in which the phylogenetic signal could not be
tested because modularity was not significant or an accurate phylogeny was
not available. Viruses were excluded because they are polyphyletic and
cannot be accurately located within the tree.
Figure 2 | Differences between phylogenetically conserved and non-
occurrence of phylogenetic conservatism was not related to host range,
of generalization (quantitatively estimated as host range) was independent
of the phylogenetic conservatism of the interactions. e–h, In contrast, the
modularity index was consistently higher in phylogenetically conserved
interactions, indicating that in these systems there was a decrease in the
sharing of hosts among modules, irrespective of the type of organism.
(average6s.e.m., n5110, *P,0.01; NS, not significant).
NATURE|Vol 465|17 June 2010
Macmillan Publishers Limited. All rights reserved
4. Jackson, A. P. & Charleston, M. A. A cophylogenetic perspective of RNA–virus
evolution. Mol. Biol. Evol. 21, 45–57 (2004).
Gilbert, G. S. & Webb, C. O. Phylogenetic signal in plant pathogen-host range.
Proc. Natl Acad. Sci. USA 104, 4979–4983 (2007).
Winkler, I. S. & Mitter, C. in Specialization, Speciation and Radiation: The
Evolutionary Biology of Herbivorous Insects (ed. Tilmon, K. J.) 240–263 (Univ.
California Press, 2007).
Freckleton, R.P., Harvey, P.H.&Pagel,M.Phylogeneticanalysis andcomparative
data: a test and review of evidence. Am. Nat. 160, 712–726 (2002).
Wiens, J. J. & Graham, C. H. Niche conservatism: integrating evolution, ecology,
and conservation biology. Annu. Rev. Ecol. Evol. Syst. 36, 519–539 (2005).
10. Darwin, C. On the Origin of Species 78–79 (Murray, 1859).
11.Chase, J. M. & Leibold, M. A. Ecological Niches: Linking Classical and Contemporary
Approaches 19–45 (Univ. Chicago Press, 2003).
and rate. Syst. Biol. 57, 591–601 (2008).
13. Dunne, J. A., Williams, R. J. & Martinez, N. D. Food-web structure and network
biodiversity. Annu. Rev. Ecol. Evol. Syst. 38, 567–593 (2007).
15. Guimera `, R. & Amaral, L. A. N. Functional cartography of complex metabolic
networks. Nature 433, 895–900 (2005).
16. Guimera `, R. & Amaral, L. A. N. Cartography of complex networks: modules and
universal roles. J. Stat. Mech. P02001 (2005).
17. Olesen, J. M., Bascompte, J., Dupont, Y. L. & Jordano, P. The modularity of
pollination networks. Proc. Natl Acad. Sci. USA 104, 19891–19896 (2007).
18. Rezende, E. L., Lavabre, J. E., Guimara ˜es, P. R.Jr, Jordano, P. & Bascompte, J. Non-
448, 925–928 (2007).
19. Rezende, E. L., Albert, E. M., Fortuna, M. A. & Bascompte, J. Compartments in a
marine food web associated with phylogeny, body mass, and habitat structure.
Ecol. Lett. 12, 779–788 (2009).
20. Oksanen, J. Multivariate Analysis of Ecological Communities in R: Vegan Tutorial (R
Project for Statistical Computing, 2008).
21. Blomberg, S. P., Garland, T. & Ives, A. R. Testing for phylogenetic signal in
22. Pride,D.T.,Wassenaar,T.M.,Ghose, C.&Blaser,M.J.Evidenceofhost-virusco-
evolution in tetranucleotide usage patterns of bacteriophages and eukaryotic
viruses. BMC Genomics 7, 8 (2006).
23. Price, P. W. in Specialization, Speciation and Radiation: The Evolutionary Biology of
Herbivorous Insects (ed. Tilmon, K. J.) 174–187 (Univ. California Press, 2007).
24. Refre ´gier, G. et al. Cophylogeny of the anther smut fungi and their
caryophyllaceous hosts: prevalence of host shifts and importance of delimiting
parasite species for inferring cospeciation. BMC Evol. Biol. 8, 100 (2008).
a character on a phylogenetic tree. Evolution 45, 1184–1197 (1991).
Supplementary Information is linked to the online version of the paper at
Acknowledgements We thank J. Bascompte, J. Bosch, A. Gonza ´lez-Megı ´as,
P. Jordano, M. Lineham, M. Me ´ndez, I. Reche, E.W. Schupp and S. Strauss for
comments on a previous draft, R. Guimera ` for kindly providing NETCARTO
software, and B. Krasnov, C. Mitter, L. Navarro, J. Ollerton and J. M. Pleguezuelos
of Science (J.M.G., M.V. and F.P.) and by the Junta de Andalucı ´a (J.M.G. and F.P.)
Author Contributions J.M.G., M.V. and F. P. designed the study, J.M.G. compiled
the data set and performed the analysis of host use, M.V. performed the
phylogenetic analyses, J.M.G. wrote a first version of the manuscript and all
authors contributed to the final draft.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Readers are welcome to comment on the online version of this article at
www.nature.com/nature. Correspondence and requests for materials should be
addressed to J.M.G. (firstname.lastname@example.org).
NATURE|Vol 465|17 June 2010
Macmillan Publishers Limited. All rights reserved
The data set. Our data set includes 116 genera belonging to seven kingdoms, as
described. Because genus is the taxonomic level at which ecological interaction-
mediated speciation mostly manifests, most macro-coevolutionary theories
predict that coevolution generates the appearance of new species (diversifying
antagonistic interactions (escape-and-radiate coevolution) or that some lineages
track the evolution of other lineages (sequential evolution). In all cases, the evolu-
tion of interactions is most apparent between related species, usually belonging to
the same genus. In addition, using taxonomic levels above genus would make it
(Onthophagus). The list of these clades and their source references are included in
63.364.3% of the species belonging tothe studied genera, with 36% of the clades
R250.68, N5116 systems, t512.03, P,0.0001, from log-linear regression),
becausespecies-rich genera havetraditionally beenlessintensely studied thanhave
those with low numbers of species.
each studied system. It was family in 57 systems, genus in 26 systems, species in
12 systems, order in 7 systems, section in 5 systems, subfamily in 4 systems, class
taxonomic resolution of the hosts, we performed all the statistical analyses
including host taxonomic resolution as covariate. This inclusion of taxonomic
resolution did not change the results of the analyses.
Clade-oriented network analysis. Network analysis has been successfully
appliedtothestudyofecologicalcommunities inrecentdecades14.In thisstudy,
we have extended the application of network tools to the study of phylogeneti-
cally related clades. We constructed bipartite networks of species belonging to
the same genus and their known hosts throughout their distribution ranges.
Species were then connected through the co-occurrence of interactions. We
subsequently converted the bipartite networks into unipartite networks accord-
ing to shared interactions. Consequently, we obtained networks that connected
species used. All network analyses were done with PAJEK26.
Modularity analysis. The modularity level and the number of modules per
network was determined using an algorithm (in the software NETCARTO)
based on simulated annealing and provided by R. Guimera `15,16. This algorithm
identifies modules with an accuracy of 90% (ref. 16). For each network, we
calculated the index of modularity, M (a measure of the extent to which species
have more links within their modules than would be expected if linkage was
where r is the number of modules in the network, L is the number of links in the
degrees of the nodes in module s (refs 15, 16). Because random networks also
may have strong modularity27, we explored whether our networks were signifi-
cantly more modular than random networks by running the same simulated
annealing algorithm in 100 random networks with the same species degree
distribution as the empirical one15. This method produces a modularity index
that is a measure of the degree to which the network is organized into clearly
permutational multivariate analyses of variance using distance matrices (using
the function ‘adonis’ in VEGAN20), which test whether element similarity (that
taxa as a function of their interacting organisms) was significantly higher within
than between modules. The function ‘adonis’ partitions dissimilarities for the
sources of variation, and uses permutation tests to inspect the significances of
those partitions. Dissimilarity was calculated as a Bray–Curtis distance.
Phylogenetic conservatism. Phylogenetic conservatism in host use was deter-
mined in each system by estimating the significance of the phylogenetic signal
following ref. 25. This test estimates whether the minimum number of evolu-
This was determined under a null model in which data were reshuffled 1,000
as an unordered, multistate factor. We mapped the evolution of host use onto
were available in GenBank (Timarcha and Alnicola), we inferred the tree on the
We recovered the topology but not the branch lengths of these 111 phylogenetic
trees because the phylogenetic signal test we used is based on parsimony and
branch lengths are therefore not necessary.We did not use likelihood approaches
to take advantage of the information on branch lengths because of the limitation
associatedwiththe number ofstatesofthe characterbeing high28. This limitation
a discrete character makes transitions among its possible states as it evolves
through time27. With the high number of states encountered in most of our
phylogenies (mean, 6; range, [2,18]), the number of transition rates to estimate
is extremely large. All tests were performed using MESQUITE 2.7129.
Statistical analyses. We used generalized linear models to test the effect of
several characteristics of the studied system on their host range, modularity
and phylogenetic conservatism. In these models, we included as explanatory
variables the type of interaction (mutualistic versus antagonistic), intimacy
(symbioticversus non-symbiotic), classicaltaxonomic domain (eukaryote, pro-
karyote and virus), host range (when not used as a dependent variable) and
sample size (the number of species appearing in the phylogenies).
26. de Nooy, W., Mrvar, A. & Batagelj, V. Exploratory Social Network Analysis with
Pajek: Structural Analysis in the Social Sciences (Cambridge Univ. Press, 2005).
27. Guimera `, R., Sales-Pardo, M. & Amaral, L. A. N. Modularity from fluctuations in
random graphs and complex networks. Phys. Rev. E 70, 025101 (2004).
28. Pagel, M. Inferring the historical patterns of biological evolution. Nature 401,
29. Maddison, W. P. & Maddison, D. R. Mesquite: a modular system for evolutionary
analysis. Mesquite Æhttp://mesquiteproject.orgæ (2009).
Macmillan Publishers Limited. All rights reserved