Heterogeneity in connectivity of habitat networks saves stable coexistence of competing species

Naoki Masuda; Norio Konno

REVIEW

Heterogeneity in connectivity of habitat networks saves stable coexistence of competing species

Naoki Masuda, Norio Konno

Disordered Systems and Neural Networks · Submitted 6 Mar 2006 · 27 references

Pith's verdict: Clean analytic result for RSP coexistence on two-point degree networks with broad numerical support; the main caveat is the annealed mean-field scope, but the paper deserves serious refereeing.

T0 · AI review

How this review was made: An AI referee (deepseek-v4-flash) read the paper on August 28, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 5 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv IDcond-mat/0603114 (version 4)

Pith ID57DQNO3J, Pith's permanent record of this paper and its review

CitationNaoki Masuda, Norio Konno. "Heterogeneity in connectivity of habitat networks saves stable coexistence of competing species." arXiv:cond-mat/0603114 (2006). https://pith.science/paper/57DQNO3J

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

Short version: this is a credible paper with one genuinely new analytic result and solid numerical support. The RSP stability analysis for the two-point degree distribution is clean: the Routh-Hurwitz condition reduces to |H3| proportional to p(1-p)(k1-k2)^2, so any degree heterogeneity stabilizes the coexistence equilibrium in the annealed mean-field model. That is a real, testable mechanism, and it goes beyond the prior literature, which mostly treated homogeneous or Poisson degree distributions. The numerics on ER, BA, Gaussian, and two-point networks showing smaller density fluctuations with wider degree dispersion are consistent and make the qualitative point convincingly.

The May-Leonard section is the weaker half. The eigenvalue calculation is fine, but the stability region is illustrated for alpha=2/3, beta=2 and a narrow range of p and k1/k2. The paper does note that extreme heterogeneity destroys coexistence, so it does not overclaim, but the parametric scope is thin. A reader cannot tell how generic the ML result is beyond the shown slice.

The main soft spot, which the stress-test note correctly identifies, is the annealed mean-field closure. Equations (1)-(2) and (7) assume the neighbor state probability Theta_i is independent of the focal vertex's degree. That is exact for the annealed model but not for static networks with degree correlations, and BA networks do have these. So the analytic proof of stability is strictly for the annealed two-point model; the claim that it transfers to simulated static networks is supported only by numerics, with no quantified error. This is a scope limitation, not a fatal flaw, because the simulations are direct and the trend is clear. But the paper should state the analytic scope more carefully.

A second, minor issue: the fluctuation measurements in Fig. 1(d) and 4(c) appear to come from single runs with no error bars or ensemble statistics. Probably fine for a qualitative trend, but it would be better to see some repeat statistics.

Citation pattern is honest. The reliance on their own prior work for equilibrium invariance is legitimate; the new stability result is self-contained.

Who this is for: anyone working on cyclic competition, graph-structured population dynamics, or biodiversity mechanisms. It is a solid contribution to that subfield, not a rewrite of coexistence theory. It deserves a serious referee - send it to review, with the mean-field caveat and the ML parameter dependence as the main points the referee should probe.

Referee report

Significance

If the result holds, it identifies a purely structural mechanism for coexistence in cyclic competition: heterogeneity in connectivity alone, without spatial segregation or environmental heterogeneity, can convert neutral or unstable coexistence into stable coexistence. The RSP analytical derivation is self-contained and the numerical evidence spans several network families, which is a strength. The ML part is more limited and parameter-dependent. The main caveat is that the analytical proof is carried out in the annealed mean-field approximation, while the broad conclusion is extended to static correlated networks such as the Barabási-Albert model. The paper also gives a useful conceptual link to oscillator desynchronization. Overall, the contribution is significant for ecological and evolutionary dynamics on networks, but the scope of the central claim needs to be stated more carefully.

Major objections (3)

[II.C, Eq. (6)] The Routh-Hurwitz determinants as printed are algebraically incorrect. For the quartic characteristic equation (5), the first principal minor is H1 = 3 k1 k2 / ⟨k⟩, not 1, and the second is H2 = (9 k1 k2 / ⟨k⟩^3) times the displayed expression, not the displayed expression alone. The positivity conclusion and the final stability condition k1 ≠ k2, p ∈ (0,1) remain correct because the correct quantities are positive, but the formulas in Eq. (6) must be corrected before publication.
[II.B-II.D] The analytical stability proof is based on the annealed mean-field closure in which Θ_i is the global degree-weighted average, independent of the focal vertex's degree and of degree-degree correlations. The simulations in Section II.D, however, are run on static Erdős-Rényi and Barabási-Albert networks; the BA network in particular has degree correlations generated by preferential attachment. The paper does not quantify the approximation error of the mean-field closure or establish when the two-point annealed stability boundary transfers to static networks. The statement that heterogeneity 'promotes stable coexistence' is therefore strictly proved only for the annealed mean-field model; for the simulated static networks it is an unquantified assumption. The authors should add a discussion of this approximation and, ideally, a numerical comparison between annealed and static dynamics for the same degree sequence.
[III.B, Eq. (9) and Fig. 3] The May-Leonard stability claim is demonstrated only for the specific parameter values α = 2/3, β = 2, and for a narrow regime p = 0.9, 0.95 with k1/k2 ≈ 0.1. The paper itself finds that p < 0.5 and k1/k2 < 0.05 do not stabilize coexistence, so the ML response to degree heterogeneity is non-monotonic and parameter-dependent. No closed-form stability condition is derived for the general ML case; the conclusion rests on numerical evaluation of Eq. (9). The abstract and Section IV.A state the general conclusion 'heterogeneity in degree distributions promotes stable coexistence' without these qualifications. The ML result should be reported as conditional, and the general claim in the abstract should be adjusted accordingly.

Minor objections (5)

[II.C, after Eq. (5)] For the homogeneous limit k1 = k2 = k, the eigenvalues should read x = √3 k i and (-3 ± √3 i) k / 2; the imaginary unit is missing in the second pair as printed. In addition, the characteristic equation appears to be written for 3 times the eigenvalues of the linearized Jacobian, and this scaling should be stated explicitly to avoid confusion.
[II.B, Eq. (4)] The invariance of the equilibrium densities under the degree distribution is cited to Ref. [20] but not derived; stating the result explicitly or outlining its proof would make the two-point stability analysis self-contained.
[II.D, Fig. 1(d)] The BA network contributes a single point (a horizontal line at its fixed variance), so the claim that 'more dispersed degree distributions generally lead to more stable coexistence' across network types is based on comparing different network families at different variances rather than on a trend within the BA family. This comparison should be clarified in the text.
[III.C, Fig. 4(c)] Gaussian networks are absent from the ML coexistence results, and the main text notes that they do not yield coexistence, in contrast to the RSP case. This limitation should be stated before the general conclusion, as it reinforces the parameter-dependent nature of the ML mechanism.
[II.D and III.C] The 'Gaussian pk' networks are described only by a mean and standard deviation; since a Gaussian on degrees can assign nonzero probability to negative degrees, the precise truncation or discrete construction used in the simulations should be specified.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

Networks in which individuals differ in how many connections they have can keep three cyclically competing species coexisting, even when a well-mixed population would lose all but one.

The reading

This paper establishes that the very unevenness of a contact network—how differently connected different individuals are—can keep competing species coexisting when a well-mixed population would collapse to a single winner. For rock-paper-scissors competition on networks, the authors prove that if the degree distribution takes at least two distinct values in nontrivial proportions, the coexistence equilibrium changes from neutrally stable to linearly stable. They confirm numerically that population-density fluctuations shrink as degree dispersion grows, across random, scale-free, Gaussian, and two-point degree networks. The same qualitative result holds for the May–Leonard model with natural death, where stable coexistence requires a long-tailed degree distribution with a few high-degree hubs. Because the equilibrium densities themselves do not depend on the network, the mechanism is purely about stability, not about shifting the balance of species.

What carries the argument

The carrying object is the degree-weighted mean field $\Theta_i = \sum_k k p_k \rho_{i,k}/\langle k\rangle$, the probability that a randomly chosen neighbor of any vertex is in state $i$, which enters the rate equations $\dot{\rho}_{i,k}$ for both models. The analytic proof uses the two-point degree distribution to reduce the infinite-dimensional system to a finite-dimensional one, and the Routh–Hurwitz criterion (for rock-paper-scissors) or the eigenvalues of the block-circulant Jacobian (for May–Leonard) to decide stability. The numerics take the standard deviation of the degree, $(\langle k^2\rangle - \langle k\rangle^2)^{1/2}$, as the independent variable and the temporal standard deviation of species densities as the measure of coexistence stability.

Extended reading notes

Core claim

The paper shows that heterogeneous connectivity alone is a coexistence mechanism for cyclically competing species. On an annealed mean-field network, the rock-paper-scissors coexistence equilibrium has densities $\rho^*_{i,k} = \Theta^*_i$ that are independent of the degree distribution, but whose stability changes when degrees are dispersed. For a two-point degree distribution $p_k = p\delta_{k,k_1} + (1-p)\delta_{k,k_2}$, the Routh–Hurwitz conditions on the characteristic polynomial reduce to $k_1 \neq k_2$ and $p \neq 0,1$; hence coexistence is linearly stable whenever both degree classes are occupied. Numerically, the standard deviation of the population densities after transients decreases as the degree standard deviation grows on all network families tested, including networks of $n=200$ where survival probabilities are higher for broader degree distributions. The May–Leonard model, whose well-mixed coexistence is an unstable spiral, is stabilized only by a long-tailed degree distribution with a small subpopulation of hubs; excess heterogeneity destroys coexistence because the small-degree vertices become effectively isolated.

Load-bearing premise

The analytical stability proofs assume an annealed mean field in which every vertex sees the same global, degree-weighted average density $\Theta_i$, ignoring correlations between the degrees and states of neighboring vertices.

Editorial extensions

If this is right

Biodiversity can be maintained by the contact structure of a population alone, without spatial segregation or environmental variation.
Coexistence here is mixed and asynchronous: species densities oscillate with small amplitude rather than forming spatially separated clusters, so this network mechanism is distinct from lattice-based mechanisms.
Any population-dynamical model with neutral or unstable oscillations in a well-mixed population should be stabilized by the same degree-heterogeneity logic.
More heterogeneous (for example, scale-free) contact networks should show more stable coexistence, an effect testable in microbial or lizard systems with controlled interaction structures.
Conservation of habitat connectivity should attend to the variance of connectivity, not only the average number of contacts.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The mean-field projection suggests the stabilization is a form of oscillator desynchronization; one could look for a critical degree variance below which coexistence is lost, analogous to a Kuramoto transition.
Because the two-point calculation already shows linear stability for any $k_1 \neq k_2$, the mechanism likely extends beyond two-point distributions, and the fluctuation curves in the paper's Fig. 1(d) may be collapsible when rescaled by the degree variance.
For contagion dynamics, degree heterogeneity lowers epidemic thresholds; for cyclic competition it stabilizes coexistence—so the same network feature can have opposite qualitative effects depending on whether transitions are neighbor-driven.
A natural extension would replace the two-point degree distribution by an arbitrary $p_k$ and derive the stability boundary in terms of moments such as $\langle k^2\rangle/\langle k\rangle^2$.

Original abstract

Coexistence of individuals with different species or phenotypes is often found in nature in spite of competition between them. Stable coexistence of multiple types of individuals have implications for maintenance of ecological biodiversity and emergence of altruism in society, to name a few. Various mechanisms of coexistence including spatial structure of populations, heterogeneous individuals, and heterogeneous environments, have been proposed. In reality, individuals disperse and interact on complex networks. We examine how heterogeneous degree distributions of networks influence coexistence, focusing on models of cyclically competing species. We show analytically and numerically that heterogeneity in degree distributions promotes stable coexistence.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13-20260917 via OpenRouter, question bank abstract-p1-20260923, 2026-09-23: 1 request, 4,809 input tokens, $0.0002 reported.

What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.06 · human participants 0.02 · controlled comparison 0.13 · randomized assignment 0.01 · states quantitative result 0.05 · compares with prior results 0.12 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v4. Results and the main claim are underlined; show every role.

Coexistence of individuals with different species or phenotypes is often found in nature in spite of competition between them. Stable coexistence of multiple types of individuals have implications for maintenance of ecological biodiversity and emergence of altruism in society, to name a few. Various mechanisms of coexistence including spatial structure of populations, heterogeneous individuals, and heterogeneous environments, have been proposed. In reality, individuals disperse and interact on complex networks. We examine how heterogeneous degree distributions of networks influence coexistence, focusing on models of cyclically competing species. We show analytically and numerically that heterogeneity in degree distributions promotes stable coexistence.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	Coexistence of individuals with different species or phenotypes is often found in nature in spite of competition between them.	background 1.00	0.04	0.02	asserted 0.98
s1	Stable coexistence of multiple types of individuals have implications for maintenance of ecological biodiversity and emergence of altruism in society, to name a few.	background 0.84	0.04	0.02	asserted 0.94
s2	Various mechanisms of coexistence including spatial structure of populations, heterogeneous individuals, and heterogeneous environments, have been proposed.	background 1.00	0.04	0.02	asserted 0.91
s3	In reality, individuals disperse and interact on complex networks.	background 1.00	0.04	0.02	asserted 1.00
s4	We examine how heterogeneous degree distributions of networks influence coexistence, focusing on models of cyclically competing species.	contribution 0.97	0.82	0.05	unclear 0.77
s5	We show analytically and numerically that heterogeneity in degree distributions promotes stable coexistence.	result 1.00	0.97	0.96	asserted 1.00

What kind of paper the classifier reports this is: theoretical result 1.00 · evidence from proof or derivation 0.67.

What was examined, and what was not

sources read

arXiv listing abstract v4

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** RSP dynamics on networks with n = 5000 and hki = 10. The initial condition is given by the Bernoulli distribution with ρ0 = 0.7 and ρ1 = ρ2 = 0.15, where ρi is the proportion of vertices that take state i. The densities ρ0 (thin lines), ρ1 (moderate lines), and ρ2 (thick lines) are shown for (a) the R random graph. For (b) the ER random graph and (c) the BA model of the same size, only ρ0 is shown for clarity. (d) Fluctuation of population density as a function of the standard deviation of the vertex degree (ER, triangle; BA, horizontal line [ hk 2 i − hki 2 1/2 = 157.8]; Gaussian…

**Figure 2.** Survival probabilities for the R random graph, the ER rand

**Figure 3.** Stability of the coexistence solution of the ML model. Real p

**Figure 4.** ML dynamics on networks with n = 5000, hki = 10, α = 2/3, and β = 2. The initial condition is given by ρ0 = 0, ρ1 = ρ2 = 0.25, and ρ3 = 0.5. The R random graph (a) and the ER random graph (b) do not allow stable coexistence (ρ0, dotted lines; ρ1, thin solid lines; ρ2, moderate solid lines; ρ3, thick solid lines). (c) Fluctuation of population density (BA, horizontal line [ hk 2 i − hki 2 1/2 = 157.8]; two-point pk, circles), defined by the variance of ρi from time 150 through 300 averaged over i = 1, 2, and 3

Figures reproduced from arXiv:cond-mat/0603114 by the authors.

References

Works cited (27)

[1]

J. H. Connell, Science 199, 1302 (1978)

View on Pith 1978
[2]

+ λµ Θ ∗ 1Θ ∗ 2   1 − Θ ∗ 1 − Θ ∗ 2 µΘ ∗ 2   . (3) 4 The coexistence solution to Eq. (3) is given by   Θ ∗ 1 Θ ∗ 2   =   ρ∗ 1,k ρ∗ 2,k   = 1 λ + µ + 1   1 λ   , (4) for any k. The degree distribution does not aﬀect the equilibrium population de nsities [20]. C. Stability of coexistence equilibrium When pk = δk, ⟨k⟩, each vertex has the sam...

View on Pith
[3]

When k1 = k2 = k, we turn back to the ordinary mean ﬁeld case with neutrally stable osc illations: x = √ 3ki, (−3 ± √ 3)k/ 2. More generally, the Routh-Hurwitz criteria for Eq. (5) is |H1| = 1 , |H2| = pk2 2 (⟨k⟩ − k1/ 2)2 + (1 − p)k2 1 (⟨k⟩ − k2/ 2)2 + 3k2 1k2 2/ 4 > 0, |H3| = 81 k2 1k2 2p(1 − p)(k2 − k1)2 ( ⟨ k2⟩2 + 2k1k2 ⟨k⟩2 ) / ⟨k⟩4 , |H4| = 9 k2 1k2...

View on Pith
[4]

P. Chesson, Annu. Rev. Ecol. Syst. 31, 343 (2000)

View on Pith 2000
[5]

R. Axelrod, Evolution of Cooperation (Basic Books, New York, 1984). 11

View on Pith 1984
[6]

R. M. Anderson and R. M. May, Infectious Diseases of Humans (Oxford University Press, Oxford, 1991)

View on Pith 1991
[7]

R. M. May, Nature (London) 238, 413 (1972)

View on Pith 1972
[8]

D. Tilman, Ecology 75, 2 (1994)

View on Pith 1994
[9]

J. Hofbauer and K. Sigmund, Evolutionary Games and Population Dynamics (Cambridge University Press, Cambridge, 1998)

View on Pith 1998
[10]

W. W. Murdoch, C. J. Briggs, R. M. Nisbet, W. S. C. Gurney, a nd A. Stewart-Oaten, Am. Nat. 140, 41 (1992)

View on Pith 1992
[11]

M. P. Hassell, H. N. Comins, and R. M. May, Nature (London) 353, 255 (1991)

View on Pith 1991
[12]

K. Tainaka, Phys. Lett. A 176, 303 (1993)

View on Pith 1993
[13]

M. Frean and E. R. Abraham, Proc. R. Soc. London Ser. B 268, 1323 (2001)

View on Pith 2001
[14]

R. Durrett and S. Levin, Theor. Popul. Biol. 53, 30 (1998)

View on Pith 1998
[15]

A. -L. Barab´ asi and R. Albert, Science 286, 509 (1999)

View on Pith 1999
[16]

M. E. J. Newman, SIAM Rev. 45, 167 (2003)

View on Pith 2003
[17]

L. W. Buss, Proc. Natl. Acad. Sci. U.S.A. 77, 5355 (1980)

View on Pith 1980
[18]

O. Gilg, I. Hanski, and B. Sittler, Science 302 866 (2003)

View on Pith 2003
[19]

B. Kerr et al. , Nature (London) 418, 171 (2002)

View on Pith 2002
[20]

B. Sinervo and C. M. Lively, Nature (London) 380, 240 (1996)

View on Pith 1996
[21]

C. Hauert, S. De Monte, J. Hofbauer, and K. Sigmund, Scie nce 296, 1129 (2002); G. Szabo and C. Hauert, Phys. Rev. Lett. 89, 118101 (2002); D. Semmann, H.-J. Krambeck, and M. Milinski, Nature (London) 425, 390 (2003)

View on Pith 2002
[22]

N. Masuda and N. Konno, J. Theor. Biol. 243, 64 (2006)

View on Pith 2006
[23]

R. M. May and W. J. Leonard, SIAM J. Appl. Math. 29, 243 (1975)

View on Pith 1975
[24]

P. D. Taylor and L. B. Jonker, Math. Biosci. 40, 145 (1978)

View on Pith 1978
[25]

H. W. Hethcote and J. A. Yorke, Lect. Notes Biomath. 56, 1 (1984); R. Pastor-Satorras and A. Vespignani, Phys. Rev. Lett. 86, 3200 (2001)

View on Pith 1984
[26]

Y. Kuramoto, Chemical Oscillations, Waves, and Turbulence (Springer-Verlag, Berlin, 1984)

View on Pith 1984
[27]

M. Barth´ elemy, A. Barrat, R. Pastor-Satorras, and A. Vespignani, Phys. Rev. Lett. 92, 178701 (2004). 12 Figure captions Figure 1: RSP dynamics on networks with n = 5000 and ⟨k⟩ = 10. The initial condition is given by the Bernoulli distribution with ρ0 = 0 . 7 and ρ1 = ρ2 = 0 . 15, where ρi is the proportion of vertices that take state i. The densities ρ...

View on Pith 2004

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.

[1] [1]

J. H. Connell, Science 199, 1302 (1978)

View on Pith 1978

[2] [2]

+ λµ Θ ∗ 1Θ ∗ 2   1 − Θ ∗ 1 − Θ ∗ 2 µΘ ∗ 2   . (3) 4 The coexistence solution to Eq. (3) is given by   Θ ∗ 1 Θ ∗ 2   =   ρ∗ 1,k ρ∗ 2,k   = 1 λ + µ + 1   1 λ   , (4) for any k. The degree distribution does not aﬀect the equilibrium population de nsities [20]. C. Stability of coexistence equilibrium When pk = δk, ⟨k⟩, each vertex has the sam...

View on Pith

[3] [3]

When k1 = k2 = k, we turn back to the ordinary mean ﬁeld case with neutrally stable osc illations: x = √ 3ki, (−3 ± √ 3)k/ 2. More generally, the Routh-Hurwitz criteria for Eq. (5) is |H1| = 1 , |H2| = pk2 2 (⟨k⟩ − k1/ 2)2 + (1 − p)k2 1 (⟨k⟩ − k2/ 2)2 + 3k2 1k2 2/ 4 > 0, |H3| = 81 k2 1k2 2p(1 − p)(k2 − k1)2 ( ⟨ k2⟩2 + 2k1k2 ⟨k⟩2 ) / ⟨k⟩4 , |H4| = 9 k2 1k2...

View on Pith

[4] [4]

P. Chesson, Annu. Rev. Ecol. Syst. 31, 343 (2000)

View on Pith 2000

[5] [5]

R. Axelrod, Evolution of Cooperation (Basic Books, New York, 1984). 11

View on Pith 1984

[6] [6]

R. M. Anderson and R. M. May, Infectious Diseases of Humans (Oxford University Press, Oxford, 1991)

View on Pith 1991

[7] [7]

R. M. May, Nature (London) 238, 413 (1972)

View on Pith 1972

[8] [8]

D. Tilman, Ecology 75, 2 (1994)

View on Pith 1994

[9] [9]

J. Hofbauer and K. Sigmund, Evolutionary Games and Population Dynamics (Cambridge University Press, Cambridge, 1998)

View on Pith 1998

[10] [10]

W. W. Murdoch, C. J. Briggs, R. M. Nisbet, W. S. C. Gurney, a nd A. Stewart-Oaten, Am. Nat. 140, 41 (1992)

View on Pith 1992

[11] [11]

M. P. Hassell, H. N. Comins, and R. M. May, Nature (London) 353, 255 (1991)

View on Pith 1991

[12] [12]

K. Tainaka, Phys. Lett. A 176, 303 (1993)

View on Pith 1993

[13] [13]

M. Frean and E. R. Abraham, Proc. R. Soc. London Ser. B 268, 1323 (2001)

View on Pith 2001

[14] [14]

R. Durrett and S. Levin, Theor. Popul. Biol. 53, 30 (1998)

View on Pith 1998

[15] [15]

A. -L. Barab´ asi and R. Albert, Science 286, 509 (1999)

View on Pith 1999

[16] [16]

M. E. J. Newman, SIAM Rev. 45, 167 (2003)

View on Pith 2003

[17] [17]

L. W. Buss, Proc. Natl. Acad. Sci. U.S.A. 77, 5355 (1980)

View on Pith 1980

[18] [18]

O. Gilg, I. Hanski, and B. Sittler, Science 302 866 (2003)

View on Pith 2003

[19] [19]

B. Kerr et al. , Nature (London) 418, 171 (2002)

View on Pith 2002

[20] [20]

B. Sinervo and C. M. Lively, Nature (London) 380, 240 (1996)

View on Pith 1996

[21] [21]

C. Hauert, S. De Monte, J. Hofbauer, and K. Sigmund, Scie nce 296, 1129 (2002); G. Szabo and C. Hauert, Phys. Rev. Lett. 89, 118101 (2002); D. Semmann, H.-J. Krambeck, and M. Milinski, Nature (London) 425, 390 (2003)

View on Pith 2002

[22] [22]

N. Masuda and N. Konno, J. Theor. Biol. 243, 64 (2006)

View on Pith 2006

[23] [23]

R. M. May and W. J. Leonard, SIAM J. Appl. Math. 29, 243 (1975)

View on Pith 1975

[24] [24]

P. D. Taylor and L. B. Jonker, Math. Biosci. 40, 145 (1978)

View on Pith 1978

[25] [25]

H. W. Hethcote and J. A. Yorke, Lect. Notes Biomath. 56, 1 (1984); R. Pastor-Satorras and A. Vespignani, Phys. Rev. Lett. 86, 3200 (2001)

View on Pith 1984

[26] [26]

Y. Kuramoto, Chemical Oscillations, Waves, and Turbulence (Springer-Verlag, Berlin, 1984)

View on Pith 1984

[27] [27]

M. Barth´ elemy, A. Barrat, R. Pastor-Satorras, and A. Vespignani, Phys. Rev. Lett. 92, 178701 (2004). 12 Figure captions Figure 1: RSP dynamics on networks with n = 5000 and ⟨k⟩ = 10. The initial condition is given by the Bernoulli distribution with ρ0 = 0 . 7 and ρ1 = ρ2 = 0 . 15, where ρi is the proportion of vertices that take state i. The densities ρ...

View on Pith 2004