• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jul 18, 2006; 103(29): 10947–10951.
Published online Jul 3, 2006. doi:  10.1073/pnas.0602712103
PMCID: PMC1544154
From the Cover

Tropical forests are both evolutionary cradles and museums of leaf beetle diversity


The high extant species diversity of tropical lineages of organisms is usually portrayed as a relatively recent and rapid development or as a consequence of the gradual accumulation or preservation of species over time. These explanations have led to alternative views of tropical forests as evolutionary “cradles” or “museums” of diversity, depending on the organisms under study. However, biogeographic and fossil evidence implies that the evolutionary histories of diversification among tropical organisms may be expected to exhibit characteristics of both cradle and museum models. This possibility has not been explored in detail for any group of terrestrial tropical organisms. From an extensively sampled molecular phylogeny of herbivorous Neotropical leaf beetles in the genus Cephaloleia, we present evidence for (i) comparatively ancient Paleocene–Eocene adaptive radiation associated with global warming and Cenozoic maximum global temperatures, (ii) moderately ancient lineage-specific diversification coincident with the Oligocene adaptive radiation of Cephaloleia host plants in the genus Heliconia, and (iii) relatively recent Miocene–Pliocene diversification coincident with the collision of the Panama arc with South America and subsequent bridging of the Isthmus of Panama. These results demonstrate that, for Cephaloleia and perhaps other lineages of organisms, tropical forests are at the same time both evolutionary cradles and museums of diversity.

Keywords: Cephaloleia, diversification, evolutionary radiation, phylogeny

The extraordinarily high species diversity of tropical forest floras and faunas is often attributed to the recent and rapid accumulation of species via high speciation rates (14) or the gradual accumulation and/or preservation of species over time via low extinction rates (58). These observations have led to the widespread belief that tropical forests are evolutionary “cradles” of diversity for some lineages and “museums” of diversity for others (6, 7).

Evidence in support of cradle models comes largely from geographic patterns of distribution, species richness, and endemism of extant tropical organisms. These patterns are often ascribed to evolutionary radiation in response to relatively recent climatic, tectonic, or biotic events; e.g., Pleistocene [≈1.8–0.01 mega-annum (Ma) ago] glaciation or Pliocene (≈5.3–1.8 Ma ago) bridging of the Isthmus of Panama (14). However, paleontological evidence implies that many of the evolutionary radiations that account for the present diversity (“crown diversification”) of taxonomically disparate groups of tropical organisms may have occurred comparatively early, during the late Paleocene and early to middle Eocene (Thanetian, Ypresian, and Lutetian Ages, ≈58.7–40.4 Ma ago) (917), associated with global warming, Cenozoic maximum global temperatures (18), and the latitudinal expansion (16, 19) and taxonomic diversification (9, 11, 14, 16, 17) of characteristically tropical lineages of plants, consistent with museum models of diversification. Although cradle and museum models are often presented as temporal alternatives (4, 7), their predictions are not mutually exclusive, and, as the aforementioned observations suggest, the evolutionary histories of tropical lineages of organisms may be expected to exhibit features of both kinds of models. However, this possibility has not been explored in detail for any group of terrestrial tropical organisms.

Using molecular genetic, paleontologic, and biogeographic evidence, we investigated timing and tempo in the diversification of Cephaloleia, a species-rich genus of herbivorous Neotropical leaf beetles (Chrysomelidae: Cassidinae). Our goals were to determine (i) whether the evolutionary history of Cephaloleia diversification (speciation–extinction) departs significantly from a constant rate model, (ii) whether Cephaloleia has experienced unusually rapid shifts in diversification rate during its evolutionary history, indicative of adaptive radiation(s), and (iii) whether, taken together, these data support temporal patterns of diversification consistent with cradle models (recent and rapid diversification), museum models (slow accumulation of diversity over time and/or preservation of comparatively ancient diversity), or some combination of timing and rate components from both kinds of models.

The genus Cephaloleia presents a remarkable opportunity to study timing and tempo in the taxonomic diversification of a demonstrably ancient lineage of herbivorous Neotropical insects. Cephaloleia is species-rich, with ≈202 extant species (20), and has figured prominently in the development of community ecological theory (e.g., ref. 21). All life stages feed on monocots, mostly in the order Zingiberales (22). A subset of species feed only in the juvenile rolled leaves of Zingiberales (Fig. 1), earning Cephaloleia the common name of rolled-leaf “hispine” beetles. Cephaloleia-like feeding damage on the leaves of latest Cretaceous (66.2 Ma) and early Eocene Zingiberales (20) documents the antiquity of the Cephaloleia/Zingiberales interaction and demonstrates that Cephaloleia-like beetles diverged from their most recent common ancestor before the end of the Cretaceous (65.5 Ma ago).

Fig. 1.
Cephaloleia variabilis at the open tip of an immature rolled leaf of Heliconia metallica (Darien Province, Panama).


Parsimony, maximum likelihood (ML), and Bayesian analyses of the combined mitochondrial DNA (≈1,800 bp) and nuclear DNA (≈400 bp) sequence data representing 95 a priori designated ingroup species, six outgroups, and 133 total specimens generated highly resolved, well supported, and compatible phylogenetic trees (see Supporting Results and Discussion in Supporting Text and Fig. 4, which are published as supporting information on the PNAS web site). Cephaloleia was monophyletic only with the inclusion of eight a priori designated Cephaloleia-like taxa from other genera, in agreement with ref. 22. Interspecific ingroup p distances were high (mean: 16.07%), consistent with deep divergences between taxa. Intraspecific “replicates” and other identical or nearly identical sequences were excluded from further analyses (see Table 2, which is published as supporting information on the PNAS web site), leaving 83 “exemplars.”

Among lineage diversification rate variation was concentrated at relatively deep nodes in the tree, as evidenced by an overall increase in P values for six tree balance statistics known to vary in the phylogenetic depth at which they are most sensitive (listed deepest to shallowest; B1 = 0.002, MΣ = 0.003, MΣ* = 0.006, MΠ = 0.003, MΠ* = 0.012, Ic = 0.062) (23) (see Supporting Materials and Methods in Supporting Text). Multiple iterations of sequential deletion of taxa from rapidly evolving lineages and reanalysis did not result in significantly more balanced trees.

Based on the Monte Carlo constant rates (MCCR) test, internal nodes in the Cephaloleia exemplar phylogeny were significantly closer to the root than expected, indicative of a decrease in net diversification rate (speciation–extinction) over time [species (exemplars) sampled, x = 83; total number of species, y ≈ 202; γ = −7.82; critical value of γ = −3.77 at P = 0.05]. This result was robust to taxonomic underestimation; in sensitivity analyses, values of extant species richness up to 1,523 returned a significant result. A semilogarithmic plot of lineages through time (LTT) (Fig. 2b) deviated significantly from a simulated curve generated under a birth–death model with incomplete taxon sampling and a constant diversification rate, corroborating evidence from the MCCR test for a decrease in diversification rate through time. Further patterns apparent in the empirical LTT plot included a trend toward reduced diversification rates beginning ≈38 Ma ago (late Eocene) and a Miocene–Pliocene increase in diversification rate that peaked ≈4–5 Ma ago. Overall, a model of diversification specifying a gradual decrease in rate through time best fit the empirical LTT plot (Figs. 2b and and3),3), consistent with results from the MCCR test.

Fig. 2.
Timing and tempo in the evolutionary radiation of Cephaloleia. (a) Chronogram showing significant diversification rate shifts (numbers correspond to clades identified in the text) and source(s) of support: purple (shift 6) indicates Δ1, Δ ...
Fig. 3.
Testing the fit of alternative models of diversification to the empirical LTT plot. (a) LTT plots illustrating the approximate pattern of diversification predicted under each of three proposed models. Model A exhibits a constant diversification rate, ...

We located unusually rapid shifts in diversification rate along branches subtending seven clades (Fig. 2a and Table 1). These shifts were confined to the Eocene (55.8–33.9 Ma ago; shifts 1, 2, 6, and 7) and Oligocene (≈33.9–23.0 Ma ago; shifts 3–5). No significant shifts in diversification rate were recovered in the last ≈23 Ma, perhaps because of the lack of power for these tests at shallow nodes (23). These patterns were robust to differences among alternative tree topologies. Analyses of 100 alternative ML trees and 100 randomly sampled trees from the Bayesian posterior distribution identified diversification rate shifts of comparable age and placement in the tree.

Table 1.
Fossil calibrated mean estimated ages and bootstrap estimates of standard error, probabilities for the Δ1, Δ2, and SG statistics, and significance levels from the RC test for clades identified in the text


We propose that massive alteration of ecosystems associated with the Cretaceous–Paleogene boundary event ≈65.5 Ma ago (24), including extinctions of plants and insects (25), “primed the phylogenetic fuse” for the adaptive radiation of Cephaloleia in the latest Paleocene–early/middle Eocene ≈54.97–43.47 Ma ago, triggered by global warming (26) and the latitudinal expansion (16, 19) and taxonomic diversification (9, 11, 14, 16, 17) of characteristically tropical lineages of plants, including important Cephaloleia hosts in the order Zingiberales (20, 27). The Cretaceous–Paleogene boundary event may have precipitated the extinction of many specialized plant–insect associations (25), setting the stage for later Paleocene–Eocene adaptive radiation from surviving Cretaceous lineages [such as documented here for Cephaloleia and elsewhere for many other groups of organisms (26)]. The proposed early Cenozoic diversification of angiosperm feeding beetle genera (28), Phyllonorycter moths (29), and herbivorous insects in general (10, 13) are consistent with this hypothesis. The robustness of our age estimates to relaxation of the maximum constraint on ingroup age (Table 1) lends further support to the timing of Cenozoic crown diversification reported here.

A slowdown in Cephaloleia diversification during the late Eocene to Oligocene (Fig. 2b) may have resulted from decreased origination rates [e.g., from the filling of ecological niches (30)] and/or increased extinction rates [e.g., because of the retreat of characteristically tropical lineages of plants to lower latitudes (19)]. The closely nested diversification rate shifts during the Oligocene (shifts 3–5; Fig. 2a; see also Supporting Results and Discussion) occur in a lineage composed entirely of host specialists on immature rolled leaves of plants in the genus Heliconia (Heliconiaceae) (22). These shifts (beginning 33.25 ± 2.91 Ma ago; Table 1) closely coincide with crown diversification of Heliconia beginning ≈32 Ma ago (27), possibly indicating an effect of increased niche availability. Ensuing cycles of global warming and cooling, e.g., late Oligocene and middle Miocene (see Fig. 2b), show little overall influence on diversification in Cephaloleia, perhaps because of the aforementioned prior filling of ecological niches.

We observed a slight increase in diversification rate in the LTT plot from ≈12.6 to 4 Ma ago (middle/late Miocene–early Pliocene), peaking ≈4–5 Ma ago (Figs. 2b and and3).3). This increase is coincident with the late Miocene–Pliocene collision of the Panama arc with northwestern South America ≈12.8–7.1 Ma ago (31) and subsequent bridging of the Isthmus of Panama ≈3.5 Ma ago (32). All of the cladogenetic events contributing to this peak in diversification rate (4–5 Ma ago) involve allopatric or parapatric sister species pairs, one in lower Central America and the other in extreme lower Central America and/or northwestern South America (Fig. 2a; see also Table 2). The near absolute allopatry or parapatry of post-Eocene (<33.9 Ma) sister species further suggests that recent speciation has occurred chiefly by vicariance, peripatry, and/or parapatry. The reconstructed evolutionary history of Cephaloleia diversification therefore exhibits timing and rate components characteristic of cradle and museum models, suggesting that, for Cephaloleia and perhaps other groups of organisms, tropical forests are at the same time both evolutionary cradles and museums of species diversity.

The relative stability of tropical environments over geological time may favor the preservation of comparatively ancient lineages resulting from adaptive radiation and facilitate the continued accumulation of species diversity, predominantly via vicariance, peripatry, and/or parapatry. Viewed in this light, major climatic, tectonic, and biotic events may be expected to leave a telltale signature on diversification histories, appropriate to the geographic/taxonomic (e.g., global, regional, or lineage-specific) and temporal scales at which they most influence diversification.

Materials and Methods

See Supporting Materials and Methods for more detailed methods.

Sampled Species.

The species included in this study comprise a broad sample of the ecological and morphological variation known in Cephaloleia and potentially allied taxa in other genera (see Table 2) (33). Taking into account undescribed, cryptic, and over-split species, and with the inclusion of Cephaloleia-like taxa currently ascribed to other genera, we estimate that the total number of extant species in the genus is close to the 202 species reported by Wilf et al. (20). Outgroups included representatives of six cassidine tribes: Alurnus ornatus (Alurnini), Chalepus sp. (Chalepini), Chelobasis bicolor (Arescini), Imatidium rufiventre (Imatidiini), Prosopodonta limbata (Prosopodontini), Stenispa vespertina (Cephaloleiini), and one criocerine, Crioceris duodecimpunctata (tribe Criocerini). Voucher specimens are deposited with the Harvard University Museum of Comparative Zoology.

Phylogenetic Analyses.

We sequenced ≈1,800 bp of mitochondrial DNA, including parts of cytochrome oxidase I and cytochrome oxidase II, tRNA-leucine, and an ≈400-bp fragment of the nuclear gene elongation factor 1-α (EF 1-α). We selected substitution models (modeltest 3.6) for use in separate and combined analyses using hierarchical likelihood ratio tests and Akaike information criteria. We used the incongruence length difference test (paup* 4.03b10) to assess combinability of data partitions. We used parsimony (paup), ML (phyml 2.4.4), and Bayesian inference (mrbayes 3.01) to reconstruct Cephaloleia phylogeny. Support for branches recovered under parsimony and ML was estimated with nonparametric bootstrap values. Decay indices (parsimony) were generated by using a command file of constraint trees (paup/macclade 4.05). All trees were rooted with Crioceris duodecimpunctata (GenBank accession no. AF467886). We calculated uncorrected pairwise distances (p distances) (paup) between ingroup taxa based on the combined mitochondrial DNA and EF 1-α.

Tree Calibration and Dating.

We used penalized likelihood (PL) (r8s 1.7) to generate an ultrametric tree (Fig. 2a). We used fossils to calibrate the tree and to date internal nodes, including the following. (i) Cephaloleichnites strongi (Cassidinae) feeding damage on Zingiberales leaves mapped as a minimum constraint (66.2 Ma) on ingroup age. [These fossils exhibit feeding damage from an extinct ancestor of Cephaloleia on the leaves of an ancient ginger, based on modern observations (20). Ginger leaves occur in even older fossil strata but have not been systematically examined for Cephaloleia-like feeding damage.] (ii) A fossil specimen of Chalepus sp. (Cassidinae) mapped as a minimum age constraint (44.1 Ma) to the root of Chalepus sp. (iii) A fossil specimen of Crioceris sp. (Criocerinae) mapped as a minimum age constraint (44.1 Ma) to the root of Crioceris duodecimpunctata. Monocot feeding is thought to be the pleisiomorphic condition in Cassidinae (20, 34), so we constrained the maximum age of the outgroup to 120 Ma, the approximate age of the earliest unequivocal fossil monocot (35). We separately applied a 145.5-Ma (Jurassic–Cretaceous boundary) maximum constraint on outgroup age (Table 1) because molecular clock studies suggest that monocots originated between ≈130 and 160 Ma ago (36). We used the 120-Ma constraint in all subsequent analyses. We obtained the central 95% of the distribution of age estimates for each node from the distribution of estimated ML branch lengths on the tree (generated from 100 bootstrapped data sets) under PL criteria (seqboot/phylip 3.5c; r8s).

Timing and Tempo of Diversification.

We obtained an estimate for γ (37) from the ingroup PL chronogram (genie 3.0). The statistic γ indicates whether internal nodes are closer to the root or to the tips of the tree than expected under a CR model (γ = 0). A significant P value for a negative value of γ indicates a decrease in diversification rate over time. To account for incomplete taxon sampling, we adjusted the critical value for γ using the MCCR test (37) (mccrtest 1.1).

Semilogarithmic Plot of LTT.

We generated an LTT plot from the ingroup PL chronogram (genie 3.0). Our sample of Cephaloleia (and allied taxa in other genera) consisted of ≈41% (83 of 202) of extant species. To evaluate the effects of incomplete taxon sampling on the slope of the empirical LTT plot, we generated 1,000 replicate phylogenetic trees with 202 extant taxa and randomly pruned 119 taxa from each tree (phylogen 1.1). These 1,000 subsampled trees with 83 taxa were then used to construct a mean LTT curve and 95% confidence interval for comparison with the empirical LTT curve. We evaluated the fit of the empirical LTT curve to each of three generalized models of diversification (38) using hierarchical likelihood ratio tests and Akaike information criteria (ape 1.8) (see Fig. 3).

Diversification Rate Shifts.

To detect and locate unusually rapid diversification rate shifts and to identify putative correlates, we used four “shift” statistics differing in power and bias, two likelihood ratio-based statistics (Δ1 and Δ2), the Slowinski and Guyer statistic (39) (symmetree 1.1), and the relative cladogenesis statistic (40) (implemented in end-epi 1.0 as the relative cladogenesis test).

Supplementary Material

Supporting Information:


We thank K. McKenna for field assistance and D. Windsor (Smithsonian Tropical Research Institute, Panama) for specimens and helpful discussions about cassidine phylogeny. Logistical support for fieldwork was provided by the Organization for Tropical Studies, the Arthropods of La Selva Project, the Costa Rican National Biodiversity Institute, and the Smithsonian Tropical Research Institute. Bruce Archibald, R. Burnham, L. Harmon, J. Kress, J. Losos, B. Moore, C. Specht, and P. Wilf commented on early versions of the manuscript. Financial support was provided by the Harvard University Department of Organismic and Evolutionary Biology, the Harvard University David Rockefeller Center for Latin American Studies, the Putnam Fund for Harvard University Museum of Comparative Zoology Expeditions, the Organization for Tropical Studies, the Smithsonian Tropical Research Institute, and the U.S. Environmental Protection Agency STAR Fellowship Program.


maximum likelihood
penalized likelihood
Monte Carlo constant rate
lineages through time.


Conflict of interest statement: No conflicts declared.

This paper was submitted directly (Track II) to the PNAS office.

Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. DQ026066DQ026082, DQ026084DQ026114, DQ026116DQ026120, DQ026122DQ026124, DQ026126DQ026133, DQ026135DQ026149, DQ026151DQ026153, DQ026155DQ026157, DQ026159DQ026162, DQ026165, DQ026167, and DQ538137DQ538308).

See Commentary on page 10827.


1. Haffer J. Science. 1969;165:131–137. [PubMed]
2. Prance G. T. Acta Amazonica. 1974;3:5–28.
3. Gentry A. H. Ann. Mo. Bot. Gard. 1982;69:557–593.
4. Richardson J. E., Pennington R. T., Pennington T. D., Hollingsworth P. M. Science. 2001;293:2242–2245. [PubMed]
5. Wallace A. R. Tropical Nature, and Other Essays. London: Macmillan; 1878.
6. Stebbins G. L. Flowering Plants: Evolution Above the Species Level. Cambridge, MA: Harvard Univ. Press; 1974.
7. Gaston K. J., Blackburn T. M. Proc. R. Soc. London Ser. B; 1996. pp. 63–68.
8. Fischer A. G. Evolution (Lawrence, Kans.) 1960;14:64–81.
9. Rull V. Rev. Palaeobot. Palynol. 1999;107:83–95.
10. Wilf P., Labandeira C. C. Science. 1999;284:2153–2156. [PubMed]
11. Jaramillo C. A., Dilcher D. L. Geology. 2000;28:815–818.
12. Bermingham E., Dick C. W. Science. 2001;293:2214–2216. [PubMed]
13. Wilf P., Labandeira C. C., Johnson K. R., Coley P. D., Cutter A. D. Proc. Natl. Acad. Sci. USA. 2001;98:6221–6226. [PMC free article] [PubMed]
14. Jaramillo C. A. Paleobiology. 2002;28:222–243.
15. Archibald S. B., Farrell B. D. Acta Zool. Cracov. 2003;46:17–23.
16. Wilf P., Cúneo N. R., Johnson K. R., Hicks J. F., Wing S. L., Obradovich J. D. Science. 2003;300:122–125. [PubMed]
17. Jaramillo C. A. Science. 2006;311:1893–1896. [PubMed]
18. Zachos J., Pagani M., Sloan L., Thomas E., Billups K. Science. 2001;292:686–693. [PubMed]
19. Collinson M. E. In: Palaeobiology II. Briggs D. E. G., Crowther P. R., editors. Oxford: Blackwell; 2001. pp. 112–115.
20. Wilf P., Labandeira C. C., Kress W. J., Staines C. L., Windsor D. M., Allen A. L., Johnson K. R. Science. 2000;289:291–294. [PubMed]
21. Strong D. R. Ecology. 1982;63:1039–1049.
22. McKenna D. D., Farrell B. D. Mol. Phylogenet. Evol. 2005;37:117–131. [PubMed]
23. Moore B. R., Chan K. M. A., Donoghue M. J. In: Phylogenetic Supertrees: Combining Information to Reveal the Tree of Life. Bininda-Emonds O. R. P., editor. Dordrecht, The Netherlands: Kluwer; 2004. pp. 487–533.
24. Alvarez L. W., Alvarez W., Asaro F., Michel H. V. Science. 1980;208:1095–1108. [PubMed]
25. Labandeira C. C., Johnson K. R., Wilf P. Proc. Natl. Acad. Sci. USA. 2002;99:2061–2066. [PMC free article] [PubMed]
26. Berggren W. A., Lucas S. G., Aubry M.-P. In: Late Paleocene–Early Eocene Climatic and Biotic Events in the Marine and Terrestrial Records. Aubry M.-P., Berggren W. A., Lucas S. G., editors. New York: Columbia Univ. Press; 1998. pp. 1–17.
27. Kress W. J., Specht C. D. In: Monocots: Comparative Biology and Evolution (Excluding Poales) Columbus J. T., Friar E. A., Hamilton C. W., Porter J. M., Prince L. M., Simpson M. G., editors. Claremont, CA: Rancho Santa Ana Botanic Garden; 2006. pp. 621–632.
28. Farrell B. D. Science. 1998;281:555–559. [PubMed]
29. Lopez-Vaamonde C., Wikström N., Labandeira C., Godfray C. J., Goodman S. J., Cook J. M. J. Evol. Biol. 2006 in press.
30. Harmon L. J., Schulte J. A., II, Larson A., Losos J. B. Science. 2003;301:961–964. [PubMed]
31. Coates A. G., Collins L. S., Aubry M., Berggren W. A. Geol. Soc. Am. Bull. 2004;116:1327–1344.
32. Coates A. G., Obando J. A. In: Evolution and Environment in Tropical America. Jackson J. B. C., Budd A. F., Coates A. G., editors. Chicago: Univ. of Chicago; 1996. pp. 21–56.
33. Staines C. L. Rev. Biol. Trop. 1996;3:3–87. [PubMed]
34. Hsiao T. H., Windsor D. In: Advances in Chrysomelidae Biology. Cox M., editor. Leiden, The Netherlands: Backhuys; 1999. pp. 39–50.
35. Friis E. M., Pedersen K. R., Crane P. R. Proc. Natl. Acad. Sci. USA. 2004;101:16565–16570. [PMC free article] [PubMed]
36. Chase M. W. Am. J. Bot. 2004;91:1645–1655. [PubMed]
37. Pybus O. G., Harvey P. H. Proc. R. Soc. London Ser. B; 2000. pp. 2267–2272.
38. Paradis E. Mol. Biol. Evol. 1998;15:476–479.
39. Chan K. M. A., Moore B. R. Bioinformatics. 2005;21:1709–1710. [PubMed]
40. Nee S., Mooers A. Ø., Harvey P. H. Proc. Natl. Acad. Sci. USA. 1992;89:8322–8326. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence links
  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • PopSet
    Published population set
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles
  • Taxonomy
    Related taxonomy entry
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...