Systematics and biogeography of Rhodniini (Heteroptera: Reduviidae: Triatominae) based on 16S mitochondrial rDNA sequences

Journal of Biogeography (J. Biogeogr.) (2007) 34, 699–712

ORIGINAL ARTICLE

Systematics and biogeography of Rhodniini (Heteroptera: Reduviidae: Triatominae) based on 16S mitochondrial rDNA sequences Alexandre Silva de Paula1*, Lile´ia Diotaiuti1 and Cleber Galva˜o2

1

Laborato´rio de Triatomı´neos e Epidemiologia da Doenc¸a de Chagas, Centro de Pesquisas Rene´ Rachou/FIOCRUZ, Av. Augusto de Lima 1715, 30190-002 Belo Horizonte, MG and 2 Laborato´rio Nacional e Internacional de Refereˆncia em Taxonomia de Triatomı´neos, Departamento de Protozoologia, Instituto Oswaldo Cruz/FIOCRUZ, Av. Brasil 4365, 21040-900 Rio de Janeiro, RJ, Brazil

ABSTRACT

Aim The tribe Rhodniini is one of six comprising the subfamily Triatominae (Heteroptera: Reduviidae), notorious as blood-sucking household pests and vectors of Trypanosoma cruzi throughout Latin America. The human and economic cost of this disease in the American tropics is considerable, and these bugs are unquestionably of great importance to man. Studies of the evolution, phylogeny, biogeography, ecology, physiology and behaviour of the Rhodniini are needed to help improve existing Chagas’ disease control programmes. The objective of the study reported here was to propose biogeographical hypotheses to explain the modern geographical distribution of the species of Rhodniini. Location Neotropical region. Methods We employed mitochondrial rDNA sequences (16S) currently available in GenBank to align sequences of Rhodniini species using ClustalX. The analyses included 16S sequences from predatory reduviid subfamilies (Stenopodainae, Ectrichodiinae, Harpactorinae, Reduviinae and Salyavatinae) present in GenBank as an outgroup. Cladistic analysis used the program PAUP to derive trees based on maximum parsimony (MP) and maximum likelihood (ML). Known distribution data for Rhodniini species were obtained from reviews and plotted on maps of South and Central America using the program iMap. An area cladogram was derived from the cladistic result to show the historical connections among the studied taxa and the endemic areas. The program TreeMap (Jungle Edition) was used to deduce taxon–area associations where the optimal solutions to explain the biogeographical hypothesis of the Rhodniini in the Neotropics were those with lowest total cost.

*Correspondence: Alexandre Silva de Paula, Laborato´rio de Triatomı´neos e Epidemiologia da Doenc¸a de Chagas, Centro de Pesquisas Rene´ Rachou/FIOCRUZ, Av. Augusto de Lima 1715, 30190-002 Belo Horizonte, MG, Brazil. E-mail: [email protected]

Results Parsimony and maximum-likelihood analysis of 16S rDNA sequences included 14 species of Rhodniini, as well as five species of predatory Reduviidae representing five of the predatory subfamilies. Tanglegrams were used to show the relationship between the Neotropical areas of endemism and Rhodniini species. When TreeMap with codivergence (vicariance) events were weighted as 0 and duplication (sympatry), lineage losses (extinction) and host switching (dispersal) were all weighted as 1, 20 scenarios were found to explain the biogeographical history of Rhodniini in the Neotropical region. Twelve of the optimal solutions with the lowest total cost were used to explain the biogeography of the Rhodniini in the Neotropics. These optimal reconstructions require six vicariance events, 20 duplications (sympatry), at least three dispersals, and at least one extinction event. Main conclusions The Rhodniini have a complex biogeographical history that has involved vicariance, duplications (sympatry), dispersal and extinction events. The main geological events affecting the origin and diversification of the Rhodniini in the Neotropics were (1) uplift of the Central Andes in the Miocene

ª 2006 The Authors Journal compilation ª 2006 Blackwell Publishing Ltd

www.blackwellpublishing.com/jbi doi:10.1111/j.1365-2699.2006.01628.x

699

A. S. Paula, L. Diotaiuti and C. Galva˜o

or later, (2) break-up of the Andes into three separate cordilleras (Eastern, Central and Western) in the Plio-Pleistocene, (3) formation of a land corridor connecting South and North America in the Pliocene, and (4) uplift of the Serra do Mar and Serra da Mantiqueira mountain systems between the Oligocene and Pleistocene. The relationships and biogeographical history of the species of Rhodniini in the Neotropical region probably arose from the areas of endemism shown in our work. Keywords Chagas’ disease control, Hemiptera, historical biogeography, Neotropical, Psammolestes, rDNA mitochondrial gene, Rhodnius, Triatominae.

INTRODUCTION The tribe Rhodniini Pinto, 1926 is one of six comprising the subfamily Triatominae (Heteroptera: Reduviidae), notorious as blood-sucking household pests and vectors of Trypanosoma cruzi Chagas, 1909 throughout the Neotropics (Galva˜o et al., 2003). Their genera belong to a well defined monophyletic group (Lent & Wygodzinsky, 1979). Morphological characters can be used to distinguish Rhodnius Sta˚l, 1859 and Psammolestes Bergroth, 1911, the two genera of Rhodniini, particularly the apically inserted antennae and the presence of distinct callosities behind the eyes (Lent & Wygodzinsky, 1979). Species of Rhodnius are primarily arboreal, often occupying ecotopes in palm tree crowns or epiphytic bromeliads. The genus is widely distributed in South and Central America. In Central America and the northern Andean countries (Peru, Ecuador, Colombia and Venezuela), Rhodnius species are primary targets of Chagas’ disease vector control initiatives. This is particularly true for Rhodnius prolixus Sta˚l, 1872, as well as Rhodnius ecuadoriensis Lent & Leo´n, 1958 in parts of Ecuador and northern Peru and Rhodnius pallescens Barber, 1932 in Panama and parts of Colombia. Other Rhodnius species have local epidemiological importance, including Rhodnius neglectus Lent, 1954 and Rhodnius nasutus Sta˚l, 1859 in central and northeastern Brazil; Rhodnius stali Lent et al., 1993 in Bolivia; and Rhodnius brethesi Matta, 1919 in the Brazilian Amazon (Schofield & Dujardin, 1999). The genus Rhodnius was reviewed by Lent (1948), Lent & Jurberg (1969), Lent & Wygodzinsky (1979). Three additional species have since been described: R. stali (Lent et al., 1993), Rhodnius colombiensis (Moreno et al., 1999) and Rhodnius milesi (Valente et al., 2001). The genus Rhodnius currently has 16 recognized species, including Rhodnius dalessandroi Carcavallo & Barreto, 1976 and Rhodnius paraensis Sherlock et al., 1977, neither of which has been collected since its original description. The genus Psammolestes includes Psammolestes arturi (Pinto), 1926, Psammolestes coreodes Bergroth, 1911 and Psammolestes tertius Lent & Jurberg, 1965 (Galva˜o et al., 2003). The genus was reviewed by Lent & Jurberg (1965) and Lent & Wygodzinsky (1979). Species of Psammolestes live in birds’ nests. They do not associate with man, and only rarely 700

with other mammals; as such they are not important in T. cruzi transmission (Lent & Wygodzinsky, 1979). The importance of the Rhodniini lies in the fact that some of its members feed on humans and many of these transmit T. cruzi, the protozoan that causes Chagas’ disease. The human and economic costs of this disease in the American tropics are considerable (Schaefer, 2005). A wide variety of reasons have been proposed for the high biological diversity seen in the Neotropics (Amorim, in press). Accepted causes of disjunction include: (1) tectonic displacement, (2) sea-level fluctuations, (3) interspecific competition together with climate change, (4) parapatric speciation along environmental gradients, (5) pest pressure, and (6) fine-scale habitat heterogeneity (for details see Amorim, 2006). The first two of these causes are classed as palaeogeographical, being Mesozoic–Lower Tertiary events, while the latter four occurred mainly in the Quaternary. Some of them represent competing explanations for the same biological events. Most of the causes proposed for species diversification in these models were not inferred based on a given method of biogeographical reconstruction, but rather were chosen a priori based on other sources of evidence (Amorim, in press). Several Neotropical groups of organisms have species that are widely distributed throughout South and Central America (Amorim, in press). However, groups as divergent as mammals and insects also contain species with restricted and overlapping geographical distributions. The areas of endemism proposed by dispersionists, refuge theory biogeographers and vicariance biogeographers, based on studies of different groups such as insects, arachnids, mammals and plants, are largely congruent. Thus, despite disagreements about the causes of cladogeneses, different biogeographical schools largely concur regarding the boundaries of the main areas of endemism in the Neotropics (Fig. 1). This strongly suggests common causes for the origin of these patterns. Methods that allow for both dispersal and vicariance have been proposed to reconstruct biogeographical history (Ronquist, 1997). Hence there is a growing plurality in the theoretical and methodological tools of biogeography. Nevertheless, few empirical studies have documented the relative roles of vicariance and dispersal (Zink et al., 2000). The aim of the study reported here was to formulate biogeographical

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

Systematics and biogeography of Rhodniini

Figure 1 Simplified picture of main areas of endemism for Neotropical organisms based on vertebrates, insects and other groups. The mere existence and the limits of areas of endemism are always hypotheses that may be corrected with additional studies. Although there may be additional areas, there are insufficient data to attain a minimally reliable hypothesis. (Source: artwork provided by Dr Dalton de Souza Amorim – Faculdade de Filosofia, Cieˆncias e Letras de Ribeira˜o Preto/ USP; see Amorim & Pires, 1996).

hypotheses to explain the modern geographical distribution of Rhodniini species. Both systematic and biogeographical approaches were used to construct testable hypotheses, using area cladograms (Cracraft, 1994) and the program TreeMap 2.02 (Charleston & Page, 2001). The biogeographical hypothesis was formulated using Amorim’s (in press) historical reconstruction of the Neotropical region (Fig. 2). METHODS Systematics In the present study we used mitochondrial rDNA sequences (16S) currently available in the NCBI genetic database. Other genes currently available in the NCBI database (e.g. 12S, cytochrome oxidase 1, cytochrome b, and nuclear rDNA sequences 18S and ITS2) were not considered because of the methodological difficulties of combining sequence information from different genes (Kitching et al., 1998; Sanderson & Shaffer, 2002), and the fact that different genes were

represented by unequal taxon sets in the construction of the outgroup. Initial analyses were made by aligning groups of sequences using ClustalX 1.83 (Thompson et al., 1997) under gap-opening/gap-extension penalties 15/9, 15/6, 15/3, 9/6, 9/3, 6/3, and by treating the gaps as missing (?). The analyses included the available 16S sequences from predatory reduviid subfamilies present in GenBank as an outgroup: Stenopoda spinulosa Giacchi, 1969 (Stenopodainae); Ectrychotes andreae (Thunberg), 1784 (Ectrichodiinae); Sycanus croceus Hsiao, 1979 (Harpactorinae); Tiarodes venenatus Matsumura, 1913 (Reduviinae); Lisarda rhypara Sta˚l, 1858 (Salyavatinae) (Table 1). The outgroup was chosen based on the findings of Paula et al. (2005) and the fact that the ancestral form of Rhodnius was placed in the Stenopodainae by Schofield & Dujardin (1999). The species R. dalessandroi, R. paraensis, Rhodnius amazonicus, R. milesi and P. arturi were not included in this analysis because there were no gene sequences for them in GenBank. Cladistic analysis used the program PAUP 4.0b10 (Swofford, 2002) to derive trees based on maximum parsimony (MP) and

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

701

A. S. Paula, L. Diotaiuti and C. Galva˜o

Figure 2 General biogeographical pattern of the Neotropical region based on different groups of vertebrates, insects and plants. The first vicariant event corresponds to the separation of the Caribbean arc from the continental Neotropical region. The second event divides north-west South America, Central America and coastal Mexico (NW) from south-east South America (SE). The third event separates Central America and the Choco´ regions from the Amazonian forest in the NW Neotropical component, and southeast Amazonia from the Atlantic Forest in the SE Neotropical component. (Source: artwork provided by Dr Dalton de Souza Amorim – Faculdade de Filosofia, Cieˆncias e Letras de Ribeira˜o Preto/USP; Amorim, in press).

Taxa Outgroup Stenopodainae Stenopoda spinulosa Giacchi, 1969 Ectrichodiinae Ectrychotes andreae (Thunberg, 1784) Harpactorinae Sycanus croceus Hsiao, 1979 Reduviinae Tiarodes venenatus Matsumura, 1913 Salyavatinae Lisarda rhypara Sta˚l, 1858 Ingroup Rhodnius pallescens Barber, 1932 Rhodnius ecuadoriensis Lent & Leo´n, 1958 Rhodnius colombiensis Mejia, Galva˜o & Jurberg, 1999 Rhodnius pictipes Sta˚l, 1872 Rhodnius stali Lent, Jurberg & Galva˜o, 1993 Rhodnius prolixus Sta˚l, 1859 Rhodnius nasutus Sta˚l, 1859 Rhodnius neglectus Lent, 1954 Rhodnius robustus Larrousse, 1927 Rhodnius domesticus Neiva & Pinto, 1923 Rhodnius brethesi Matta, 1919 Rhodnius neivai Lent, 1953 Psammolestes coreodes Bergroth, 1911 Psammolestes tertius Lent & Jurberg, 1965

Accession no.

Length

%GC

AY252684

314

28.0

AY127035

508

27.0

AY127043

510

30.0

AY127045

509

32.0

AY127039

508

29.0

AF045706 AF028746 AY035438 AF045709 AY035437 AF045707 AF028749 AF045704 AF045705 AY035440 AF045710 AY035441 AF045708 AY035439

374 285 510 373 508 373 284 372 372 508 374 508 371 503

24.0 23.0 28.0 26.0 29.0 27.0 24.0 29.0 30.0 32.0 27.0 31.0 27.0 30.0

Table 1 Species and 16S ribosomal DNA gene (mitochondrial gene) sequences used in maximum parsimony and maximum likelihood analyses

Species of subfamilies Stenopodainae, Ectrichodiinae, Harpactorinae, Reduviinae and Salyavatinae were used as outgroups (see Methods). Length ¼ DNA sequence length; %GC ¼ guanine/ cytosine content.

maximum likelihood (ML). Parsimony branch-and-bound searches were performed on the alignments using the chosen outgroup. Characters were treated as unordered and of equal weight, and the trees were rooted at an internal node with basal polytomy. Strict consensus trees were then obtained for each 702

branch-and-bound search. Parsimony bootstrap analyses were conducted employing a heuristic search with 100 bootstrap replicates using 10 random stepwise addition (tree-bisectionreconnection, TBR). Strict consensus trees were obtained from all the retained trees in the branch-and-bound searches, and

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

Systematics and biogeography of Rhodniini the topology of each tree under individual gap-opening/gapextension penalties was tested with ML, using a model of estimated gamma distribution (discrete approximation), HKY85 variant to allow for transition/transversion bias, unequal base frequencies and different substitution rates (Page & Homes, 1998), empirical base frequencies and an estimated substitution model following heuristic stepwise addition using TBR branch-swapping. Biogeography Distributional data for Rhodnius species were obtained from reviews by Lent (1948), Lent & Jurberg (1969), Lent & Wygodzinsky (1979). Additional localities for R. stali, R. colombiensis, R. milesi and R. amazonicus were obtained from Lent et al. (1993), Moreno et al. (1999), Valente et al. (2001) and Be´renger & Pluot-Sigwalt (2002), respectively. Distributional data for Psammolestes species were obtained from Lent & Jurberg (1965) and Lent & Wygodzinsky (1979). Coordinates of the localities were obtained from Vanzolini & Papavero (1969) and Brown (1979). Species distributions were plotted on maps of South and Central America using the program iMap 3.1 for Apple Macintosh. Phylogenetic analysis of Rhodniini species was required to test biogeographical patterns, and the areas of endemism proposed by Amorim (in press) (Fig. 2) were used to produce a derived-area cladogram to show the historical connections among the taxa studied and the endemic areas. In the biogeographical context, the four events used in most of the models were (1) vicariance, allopatric speciation caused by the origin of a dispersal barrier affecting many organisms simultaneously; (2) duplication (speciation within an area), which is usually allopatric and associated with a local or temporary dispersal barrier within an area; (3) dispersal, occurring between isolated areas and associated with speciation; and (4) extinction, which leads to the disappearance of a lineage from an area where it was predicted to occur (Sanmartı´n & Ronquist, 2004). The reconstruction can best be illustrated by using a trackogram that displays the organisms’ phylogeny on top of the area cladogram, with symbols denoting the four kinds of event. Historical associations can be divided into three basic categories (Page & Charleston, 1998): genes and organisms; organisms and organisms; and organisms and areas. Similarities among the event categories for the different kinds of association need not imply close analogies among the processes; rather the analogy is among the patterns these processes produce. Page & Charleston (1998) acknowledged that equivalent processes among different associations could be applied to historical biogeography. Following their view, ‘host– associate’ can be accepted as ‘organism–area’; ‘codivergence’ as ‘vicariance’; ‘duplication’ as ‘sympatry’; ‘host transfer’ as ‘dispersal; and ‘sorting event’ as ‘extinction’. The reconciled trees used in the previous versions of TreeMap have some limitations, the most severe being that they do not accommodate horizontal transfer (dispersal).

Charleston (1998) developed a solution to this problem that employs a mathematical structure called ‘jungles’, which contains all possible ways in which an associate tree (¼ taxa) can be mapped into a host tree (¼ areas), given the four processes of codivergence, duplication, sorting and horizontal transfer. This was implemented in TreeMap (Jungle Edition) ver. 2.02 (Charleston & Page, 2001) and the program was used to deduce taxon–area associations in our study. The optimal solutions to explain the biogeographical hypothesis of the Rhodniini in the Neotropics were those with lowest total cost (Charleston, 1998). Information from the studies of van der Hammen (1974), Clapperton (1993), Hallam (1994), Lundberg et al. (1998), Aleman & Ramos (2000) and Ramos & Aleman (2000) were accessed to fit the phylogenetic hypothesis to the geological events related to the historical distribution of the species studied here. RESULTS Systematics Parsimony and ML analyses of 16S rDNA sequences included 14 species of Rhodniini and five species of predatory Reduviidae, representing five of the predatory subfamilies: Stenopodainae, Ectrichodiinae, Harpactorinae, Reduviinae and Salyavatinae. The branch-and-bound search under gap-opening/gapextension penalties 15/9, 15/6, 15/3, 9/6, 9/3, 6/3, and using the outgroup above, resulted in 12 optimal trees (Table 2). The strict consensus tree for these 12 trees is shown in Fig. 3a. All the retained trees had the same topology except for the clade including R. brethesi, R. colombiensis and R. pictipes, which was unsolved in the strict consensus. Maximum-likelihood analysis under the same gap penalties resulted in eight trees (Table 3): the strict consensus of these is shown in Fig. 3b. Unlike the MP analysis, the strict consensus from the trees retained in the ML did not show resolution for most of the Rhodniini species, except for the clade including R. brethesi, R. stali and R. pictipes. To compare both results of the strict consensus and combine their resolution, the topology from the

Table 2 Parsimony branch-and-bound search results GO/GE

BP

PBP

15/9 15/6 15/3 9/6 9/3 6/3

547 550 550 551 554 560

144 139 136 134 129 129

„ TREE 3 3 1 1 1 3

L

CI

RI

RC

HI

545 524 515 502 491 474

0.607 0.620 0.617 0.620 0.623 0.631

0.565 0.569 0.563 0.573 0.580 0.593

0.343 0.353 0.348 0.355 0.362 0.374

0.393 0.380 0.383 0.380 0.377 0.369

GO/GE, gap-opening/gap-extension penalties; BP, total characters; PBP, parsimony-informative characters; „ TREE, number of trees retained; L, length; CI, consistency index; RI, retention index; RC, rescaled consistency index; HI, homoplasy index.

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

703

A. S. Paula, L. Diotaiuti and C. Galva˜o

Figure 3 (a) Strict consensus tree from parsimony branch-and-bound searches resulting in 12 retained trees; (b) strict consensus tree from maximum-likelihood searches resulting in eight retained trees – in both cases, total number of trees retained in all alignments (see Tables 2 & 3). Table 3 Maximum-likelihood results using the strict consensus tree retained by parsimony branch-and-bound searches GO/GE 15/9 15/6 15/3 9/6 9/3 6/3

„ TREE 1 1 1 3 1 1

)ln L

T/T

Time

3053.47534 2970.87875 2936.86807 2891.43064 2864.50300 2811.05100

1.60428 1.68883 1.75882 1.86062 1.93612 2.05809

06:14.9 03:38.5 07:06.9 10:20.2 05:40.6 05:50.0

GO/GE, gap opening/gap extension penalties; „ TREE, number of trees retained; )ln L, likelihood scores; T/T, transition/transversion ratio; Time, time used (h).

retained trees was chosen using the alignments 15/3, 9/6 and 9/3 (Fig. 4). Parsimony bootstrap values were obtained for the alignments 9/6 and 9/3, the consistency indexes of which were 0.620 and 0.623, respectively (Fig. 4). Only Rhodnius domesticus and the clade including Psammolestes species did not show bootstrap values over 50%. The members of these genera are morphologically very distinct, and our study suggests that Psammolestes should be included in the genus Rhodnius. The outgroup species did not show any sister-group relationship with the Rhodniini, so that no hypothesis could be provided to explain the relationship between this tribe and the subfamilies of Reduviidae. The inclusion of additional subfamilies of Reduviidae as outgroups in future studies could resolve this question, although Paula et al. (2005) postulated an apparent link between Rhodniini, Salyavatinae and Harpactorinae. Biogeography The distributions of Rhodniini species in the Neotropical region are shown in Figs 5 and 6, with R. ecuadoriensis, 704

R. pallescens and R. colombiensis to the west, and R. brethesi, R. pictipes and R. stali to the east of the Andes (Fig. 5). The ranges of Rhodnius neivai and R. domesticus are widely separated, the former occurring in northern South America and the latter in Atlantic forest in the south-east of the continent (Fig. 6). Both P. tertius and P. coreodes are found in south-east South America (Fig. 6), while R. nasutus is restricted to arid regions in the north-east of the continent; R. prolixus occurs throughout South and Central America; R. neglectus appears to be restricted to the Serra do Mar and Serra da Mantiqueira; and R. robustus is widespread in the Amazon basin (Fig. 6). An area cladogram for the species of Rhodniini is shown in Fig. 7, as it is not possible to observe an unambiguous vicariant pattern for all the species. The first clade, including R. colombiensis, R. ecuadoriensis and R. pallescens, showed the latter two species to be sympatric in the Andean/Mesoamerican (AnMA) area (Amorim & Pires, 1996). The presence of R. colombiensis in north-western Amazonia (NWAm) is probably the result of a vicariance event in the north-west Neotropical region (Fig. 2), and suggests speciation by vicariance following the Andean and Central American uplifts. The next clade links R. brethesi, R. stali and R. pictipes, three species with wide geographical ranges overlapping more than one endemic area, and does not provide a robust explanation of the biogeographical history of these species in the Neotropics. Rhodnius neivai occurs in the NWAm area and R. domesticus in the Atlantic Forest (AtlFor). The species P. tertius, P. coreodes and R. nasutus are found in AtlFor and speciated by duplication (paralogy) in this region. Rhodnius prolixus and R. robustus appear to have dispersed from the AtlFor, while R. neglectus also appear to have arisen by duplication in the AtlFor (Fig. 7). The tanglegram in Fig. 8 shows the relationship between the areas of endemism proposed by Amorim (in press) and the phylogeny of the Rhodniini species studied.

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

Systematics and biogeography of Rhodniini This can be explained by the uplift of the Isthmus of Panama acting as a vicariance event that allowed the lineage, including R. ecuadoriensis and R. pallescens, to spread. R. colombiensis dispersed from NWAm (Fig. 9a–d) and became extinct in the SWAm area (Fig. 9a–l). The history of the lineage, including R. brethesi, R. pictipes and R. stali, is puzzling. TreeMap indicated vicariance of R. brethesi in NWAm and also of R. stali in SWAm, whereas R. pictipes could have arisen through vicariance in NWAm. This seems the most robust scenario to explain the presentday geographical distribution of these species. All the solutions showed R. neivai in the AnMA endemic area following dispersal of the lineage R. domesticus–R. neglectus from AnMA. This clade showed duplication (speciation by sympatry) in AtlFor, followed by dispersal of R. prolixus and R. robustus to NWAm, or dispersal of R. neglectus from NWAm to AtlFor. This last solution deserves more study to explain the presence of R. prolixus in AtlFor, which has been interpreted by several epidemiologists as being due to laboratory escapes. TreeMap could elucidate the biogeographical history of the Rhodniini more effectively if more taxa and areas were included to generate the ‘jungles’. DISCUSSION Systematics

Figure 4 Selected topology from parsimony branch-and-bound search to show the phylogenetic hypothesis for the relationship among Rhodniini species. Numbers above and below branches are bootstrap support; frequencies ‡ 50%. Gap-opening/gap-extension penalties were 9/6 and 9/3, respectively, and are shown above and below the branches.

TreeMap 2.02 (Charleston & Page, 2001), with codivergence (vicariance) events weighted as 0 and duplication (sympatry), lineage losses (extinction) and host switching (dispersal) all weighted as 1 found 20 scenarios to explain the biogeographical history of Rhodniini in the Neotropical region (Table 4). The 12 optimal solutions with the lowest total cost to explain the biogeographical hypothesis of the Rhodniini are shown in Fig. 9 (reconstructions 5–16 in Table 4). These optimal reconstructions require six vicariance events (black circles), 20 duplications (sympatry; squares), at least three dispersals (arrows) and at least one extinction event (grey circles). TreeMap provided several patterns to explain the species/ area relationships of Rhodniini; thus R. ecuadorensis showed a vicariance event in the AnMA (Fig. 9a–h) and became extinct in the NWAm + SWAm (Fig. 9i–l), while R. pallescens dispersed from NWAm to AnMA (Fig. 9a–d), or speciated by vicariance when the R. ecuadoriensis lineage disappeared from those areas (Fig. 9i–l).

We refute the idea of an ancestral triatomine similar to extant Stenopodainae, as well as R. pictipes being the species closest to the ancestor of Rhodnius, as proposed by Schofield & Dujardin (1999). Although the sister group of Rhodnius may be the Salyavatinae or Harpactorinae (Paula et al., 2005), there is still no conclusive evidence to support this. According to Schaefer (2005), the main problems to be resolved in triatomine systematics are whether the subfamily has a truly independent origin and how it is related to the other subfamilies of the Reduviidae. We currently have no idea which of these subfamilies is most closely related to the Triatominae. The surprisingly few studies of reduviid subfamilies have allied the Triatominae with the Harpactorinae, Peiratinae, Physoderinae, Reduviinae and Stenopodainae. Ambrose (1999) suggested that the reduviids could be broadly divided into two groups based on whether or not they possessed tibial pads (fossulae espongiosae, or tibiarola). Reduviids with tibial pads may have evolved in the following sequence: Holoptilinae, Emesinae, Tribelocephalinae, Saicinae, Stenopodainae, Harpactorinae. Those without tibial pads live in tropical forest ecosystems and are known as timid predators that do not use their forelegs to capture prey, instead impaling prey items with their long rostra (Ambrose, 1999). Rain forest reduviids may have developed tibial pads and other features that made them more efficient predators when they migrated to deciduous scrub forest and other semi-arid habitats. The most advanced, aggressive predators, such as members of the Peiratinae and Reduviinae, live in

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

705

A. S. Paula, L. Diotaiuti and C. Galva˜o

Figure 5 Known distribution of Rhodniini species in South and Central America. See text for data sources.

semi-arid, prey-scarce situations where such features would be most needed. The Salyavatinae possess the least developed tibial pads, which may be rudimentary, consist of mere apical projections or be distinctly formed. Ambrose (1999) considered the members of this subfamily to be the most primitive of the predatory reduviids, ancestral to the subfamilies Triatominae and Ectrichodiinae (see his Figure 54). Although the Rhodniini and Salyavatinae could have shared the same Neotropical ancestor, the results of our study do not provide sufficient evidence to corroborate this. An alternative, and possibly more 706

robust hypothesis, is that the Rhodniini and Harpactorinae are closely related. Biogeography Vicariance and dispersalist schools of biogeographical analysis are both compatible with the dominance of allopatric speciation, but differ in how they construe the interaction between dispersal and allopatry. In the vicariance paradigm, rare but extensive dispersal (range expansion) is followed by a series of allopatric isolation events, interrupted by occasional random

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

Systematics and biogeography of Rhodniini

Figure 6 Known distribution of Rhodniini species in South and Central America. See text for data sources.

dispersals (Zink et al., 2000). If the isolation events affect many organisms simultaneously, this process will generate congruent tree topologies. Dispersalists consider range expansion to be a more common and regularly occurring phenomenon. Both dispersal and vicariance processes are viewed as possibly resulting in predictable as well as unpredictable (random) events. Conflicting or incongruent trees can be explained by differential dispersal across pre-existing barriers. Trees may also appear to conflict if they have unequal numbers of terminal taxa, which can result from failure of differentiation in response to a barrier (widespread species), or because some lineages have experienced extinctions. However, such trees can

be compatible with vicariance. The strongest statements about dispersal events can be made when they are rare and mixed with vicariance between areas of endemism. Under such conditions, there will be strong phylogenetic constraints on distributional patterns. Humphries & Ebach (2004) discussed the current state of cladistic biogeography and highlighted two critical points that require investigation: the definition of endemic areas and geographical congruence. Many other authors have discussed the concepts of endemic areas (Nelson & Platnick, 1981; Platnick, 1991; Harold & Mooi, 1994; Morrone, 1994; Humphries & Parenti, 1999; Hausdorf, 2002) without reaching

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

707

A. S. Paula, L. Diotaiuti and C. Galva˜o

a

b

Figure 7 Area cladogram of Rhodniini species using areas of endemism proposed by Amorim (in press) for the Neotropical region.

Table 4 Twenty optimal reconstructions satisfying the following event costs constraints: codivergences, 0; host switches, 1; duplications, 1; losses, 1 No.

C

D

L

S

E

z

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

4 4 4 4 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6

22 22 22 22 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20

0 0 0 0 1 1 2 2 1 1 2 2 2 2 3 3 8 8 15 20

6 6 6 6 5 5 4 4 5 5 4 4 4 4 3 3 2 2 1 0

28 28 28 28 26 26 26 26 26 26 26 26 26 26 26 26 30 30 36 40

28 28 28 28 26 26 26 26 26 26 26 26 26 26 26 26 30 30 36 40

No., reconstruction number; C, number of codivergence events (vicariance); D, number of duplication events (sympatry); L, number of losses (extinction); S, number of host switch events (dispersal); E, total number of non-codivergence events; z, total cost.

Figure 8 Tanglegram showing relationship between areas of endemism and phylogeny of Rhodniini species. Areas of endemism as proposed by Amorim (in press).

a consensus. Cox & Moore (2005) pointed out that some plants and animals are confined to the areas in which they evolved and are said to be endemic to that region. Their confinement may be due to physical barriers to dispersal, as in the case of many island faunas and floras (palaeoendemics), or to the fact that they have evolved only recently and have not 708

yet had time to spread (neoendemics). The concept of endemic areas requires more investigation and discussion, although Amorim & Pires (1996) and Amorim (2001, in press) have published interesting papers on the delimitation of endemic areas in the Neotropics. Similar vicariance patterns have been postulated for Coleoptera (Morrone, 2002) and Diptera (Nihei & Carvalho, 2004). Vicariance-induced and dispersion-induced elements explain the present diversity of the Neotropical region (Amorim, in press). Congruence of the distributions of different groups of organisms and the Cretaceous–Tertiary

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

Systematics and biogeography of Rhodniini a

b

c

d

e

f

g

h

i

j

k

l

Figure 9 Twelve optimal reconstructions with the lowest cost for the tanglegram shown in Fig. 8. Vicariance events (d); duplications (sympatry) (j); dispersals (arrows); extinction events (d).

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

709

A. S. Paula, L. Diotaiuti and C. Galva˜o geological history of the Neotropical region point to a number of vicariance events having caused the disjunction patterns observed today (Fig. 2). Most events were associated with tectonic movements and inundations, with long-term and local dispersions also having some impact. A general pattern shows a separation between Caribbean– Antillean elements from a continental Neotropical component, followed by a division between the south-east Amazonia– Atlantic Forest and north-west South America–Central America components (Fig. 2). Other, more regional events follow. There is evidence of repeated inundation of the Neotropical region that may have resulted in vicariance events in the Cretaceous as well as the Eocene, Miocene and Pleistocene epochs of the Quaternary period (Amorim, in press). Amorim & Pires (1996) and Amorim (2001, in press) showed many more endemic areas in the Neotropical region, but lacked information to reconstruct their histories. According to these authors, additional studies are needed to add new areas of endemism; subdivide some existing areas into smaller units (e.g. AnMA, SWAm); and establish a sequence for area components that can be subdivided into polytomies (as for SWAm). Although the Neotropical region may conveniently be considered as a single biogeographical unit, it is geologically complex. The Neotropics include not only the South American continental plate, but also the southern portion of the North American and Caribbean plates (Clapperton, 1993). The complicated geological history of the region, in which these plates intermittently separated and collided throughout the Cretaceous and the Tertiary, provides the milieu within which interactions between organisms have occurred. South America has been an island continent for most of the evolutionary history of some organisms (e.g. angiosperms), whereas Central America constitutes one of the two tropical parts of the Laurasian ‘supercontinent’. The outstanding geological feature of South America is the Andes, the longest mountain range in the world. Andean tectonic history is extremely important in understanding biogeographical process and pattern. It is now known that the Andes were built by compressional tectonics during the last 90 Myr or even longer. It is, therefore, overly simplistic to view Andean vicariance as a singular event occurring with the Miocene uplift (Lundberg et al., 1998). The Andes essentially represent a classical tectonic upthrust of continental rock, the result of a collision between the leading edge of the westward-moving South American and oceanic Pacific Plates (Lundberg et al., 1998). The southern Andes are the oldest, with significant uplift already present in the early Cenozoic, prior to the Oligocene. Most of the uplift of the Central Andes was in the Miocene or later, whereas that of the northern portion of the range was mostly PlioPleistocene (van der Hammen, 1974). Rhodnius ecuadoriensis could have speciated following the Central Andean uplift. As they extend northwards the Andes become more geologically complex, breaking into three separate cordilleras (Aleman & Ramos, 2000). The Western and Central Cordilleras of the Andes are typical subduction-related mountain 710

chains developed along the continental margin. However, the Eastern Cordillera was formed as a result of the interaction between the Paleogene Caribbean thrusting and Neogene tectonic inversion during Andean compression. These structures were greatly affected by a complex system of strike–slip faults and folds. We think that the break-up of the Andes into three separate cordilleras was a geological event leading to the evolution of R. colombiensis, R. brethesi and R. neivai within their respective geographical ranges. The major geological events believed to have occurred at the intersection of South, Central and North America are described by Hallam (1994). In the Jurassic, North and South America were joined and Central America as we know it today did not exist. In the early Cretaceous, North and South America separated just to the south of the Yucatan peninsula. Volcanic islands subsequently appeared in the gap between southern Mexico and Colombia. These were pushed northeastwards by the Farallon Plate, which in the mid-Cretaceous began to form Cuba, the Greater Antilles and the islands off the Venezuelan coast. By the early Oligocene, another archipelago had been created between South and North America, the widest gap between islands being in the Panama region. The land corridor between South and North America was completed in the Pliocene with the emergence of the Isthmus of Panama and north-west Colombia. Rhodnius pallescens occurs only in Central America and could only have speciated after the isthmus was formed. The Serra do Mar and Serra da Mantiqueira mountain systems are younger than the Andes, having formed between the Oligocene and Pleistocene (Amorim & Pires, 1996). The results of our study indicate that many duplication events (speciations within an area) occurred in AtlFor. As these events are usually allopatric and associated with a local or temporary dispersal barrier within an area, the uplift of the Serra do Mar and Serra da Mantiqueira could have resulted in the speciation of R. domesticus, P. tertius, P. coreodes and R. nasutus. Uplift of these mountains may also explain the origin and dispersal of R. prolixus and R. robustus from AtlFor. Pinho et al. (1998) collected R. prolixus in Atlantic rain forest near Tereso´polis, in the Serra do Mar. The specimens (adults, nymphs and eggs) were found in the axils of Pteridophyta leaves, in foliage and on the trunks of palm trees. This was the first report of Rhodnius colonizing Pteridophyta, and some researchers have suggested that these insects were descended from escaped laboratory-bred specimens. Based on previous studies and our own findings (Fig. 9), R. prolixus could have speciated in the Atlantic Forest of the Serra do Mar, following dispersal to north-west South America and Central America. The distribution of this species in the Serra do Mar should be studied further as it is the main target of Chagas’ disease vector control initiatives. CONCLUSIONS The Rhodniini have a complex biogeographical history that has involved vicariance, duplications (sympatry), dispersal and

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

Systematics and biogeography of Rhodniini extinction events. The main geological events affecting the origin and diversification of the Rhodniini in the Neotropics were: (1) uplift of the Central Andes in the Miocene or later, (2) break-up of the Andes into three separate cordilleras (Eastern, Central and Western) in the Plio-Pleistocene, (3) formation of a land corridor connecting South and North America in the Pliocene, and (4) uplift of the Serra do Mar and Serra da Mantiqueira mountain systems between the Oligocene and Pleistocene. The relationships and biogeographical history of the species of Rhodniini to the Neotropical region probably arose from the areas of endemism proposed by Amorim (2001, in press). ACKNOWLEDGEMENTS We thank Dr Carl Schaefer (University of Connecticut) and Dr Thomas Henry (Smithsonian Institution) for comments on an early version of the manuscript. Dr Dalton de Souza Amorim (Faculdade de Filosofia, Cieˆncias e Letras de Ribeira˜o Preto/USP) provided us with figures from his studies (Figs 1 & 2). Dr Gustavo Graciolli (Universidade Federal de Mato Grosso do Sul) commented on the TreeMap results. Dr Malte C. Ebach, Dr Juan J. Morrone and Dr John Grehan made constructive criticisms in reviewing our manuscript. Dr Bruce Alexander (Liverpool School of Tropical Medicine) made the English revision and provided comments that improved our manuscript. The study was supported by grants from the Centro de Pesquisas Rene´ Rachou/FIOCRUZ, Fundac¸a˜o de Amparo a Pesquisa de Minas Gerais (FAPEMIG), and Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq). REFERENCES Aleman, A. & Ramos, V.A. (2000) Northern Andes. Tectonic evolution of South America (ed. by U.G. Cordani, E.J. Milani, A. Thomaz Filho and D.A. Campos), pp. 453–480. 31st International Geological Congress, CPRM/SGB, Rio de Janeiro. Ambrose, D.P. (1999) Assassin bugs. Science Publishers, Enfield, NH, USA. Amorim, D.S. (2001) Dos amazonias. Introduccio´n a la biogeografia en Latinoamerica: teorı´as, conceptos, me´todos y aplicaciones (ed. by J. Llorente-Bousquets and J.J. Morrone), pp. 245–255. Facultad de Ciencias, UNAM, Mexico, DF. Amorim, D.S. (in press) Neotropical Diptera biogeography. Diptera diversity: status, challenges and tools (ed. by R. Pape and D. Bickel). Brill, Leiden, the Netherlands/Boston, MA, USA. Amorim, D.S. & Pires, M.R.S. (1996) Neotropical biogeography and a method for maximum biodiversity estimation. Biodiversity in Brazil: a first approach (ed. by C.E.M. Bicudo and N.A. Menezes), pp. 183–219. Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico – CNPq, Sa˜o Paulo, Brazil.

Be´renger, J.M. & Pluot-Sigwalt, D. (2002) Rhodnius amazonicus Almeida, Santos & Sposina 1973, bona species, close to R. pictipes Sta˚l, 1872 (Heteroptera, Reduviidae, Triatominae). Memo´rias do Instituto Oswaldo Cruz, 97, 73–77. Brown, K.S., Jr (1979) Ecologia geogra´fica e evoluc¸a˜o nas florestas neotropicais. Universidade Estadual de Campinas, Campinas, Brazil. Charleston, M.A. (1998) Jungles: a new solution to the host/ parasite phylogeny reconciliation problem. Mathematical Biosciences, 149, 191–223. Charleston, M.A. & Page, R.D.M. (2001) TreeMap for Macintosh. Version 2.0. http://taxonomy.zoology.gla.ac.uk/rod/ treemap.html Clapperton, C.M. (1993) Quaternary geology and geomorphology of South America. Elsevier, Amsterdam. Cox, C.B. & Moore, D. (2005) Biogeography: an ecological and evolutionary approach, 7th edn. Blackwell Publishing, Oxford, UK. Cracraft, J. (1994) Species diversity, biogeography, and the evolution of biotas. American Zoologist, 34, 33–47. Galva˜o, C., Carcavallo, R.U., Rocha, D.S. & Jurberg, J. (2003) A checklist of the current valid species of the subfamily Triatominae Jeannel, 1919 (Hemiptera, Reduviidae) and their geographical distribution, with nomenclatural and taxonomic notes. Zootaxa, 202, 1–36. Hallam, A. (1994) An outline of Phanerozoic biogeography. Oxford Biogeography Series No. 10. Oxford University Press, Oxford, UK. van der Hammen, T. (1974) The Pleistocene changes of vegetation in tropical South America. Journal of Biogeography, 1, 3–26. Harold, A.S. & Mooi, R.D. (1994) Areas of endemism: definition and recognition criteria. Systematic Biology, 43, 261– 266. Hausdorf, B. (2002) Units in biogeography. Systematic Biology, 51, 648–652. Humphries, C.J. & Ebach, M. (2004) Biogeography on a dynamic Earth. Frontiers of biogeography: new directions in the geography of nature (ed. by M.V. Lomolino and L.R. Heaney), pp. 67–86. Sinauer, Sunderland, MA, USA. Humphries, C.J. & Parenti, L.R. (1999) Cladistic biogeography: interpreting patterns of plant and animal distributions, 2nd edn. Oxford University Press, Oxford, UK. Kitching, I.J., Forey, P.R., Humphries, C.J. & Williams, D.M. (1998) Cladistics: theory and practice of parsimony analysis, 2nd edn. Oxford University Press, Oxford, UK. Lent, H. (1948) O geˆnero Rhodnius Sta˚l, 1859 (Hemiptera, Reduviidae). Revista Brasileira de Biologia, 8, 297–339. Lent, H. & Jurberg, J. (1965) O geˆnero Psammolestes Bergroth, 1911, com um estudo sobe a genitalia das espe´cies (Hemiptera, Reduviidae, Triatominae). Revista Brasileira de Biologia, 24, 349–376. Lent, H. & Jurberg, J. (1969) O geˆnero Rhodnius Sta˚l, 1859, com um estudo soˆbre a genita´lia das espe´cies (Hemiptera, Reduviidae, Triatominae). Revista Brasileira de Biologia, 29, 487–560.

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd

711

A. S. Paula, L. Diotaiuti and C. Galva˜o Lent, H. & Wygodzinsky, P. (1979) Revision of the Triatominae (Hemiptera, Reduviidae) and their significance as vectors of Chagas disease. Bulletin of the American Museum of Natural History, 163, 123–520. Lent, H., Jurberg, J. & Galva˜o, C. (1993) Rhodnius stali n. sp., afim de Rhodnius pictipes Sta˚l, 1872 (Hemiptera, Reduviidae, Triatominae). Memo´rias do Instituto Oswaldo Cruz, 88, 605– 614. Lundberg, J.G., Marshall, L.G., Guerrero, J., Horton, B., Malabarba, C.S.L. & Wesslingh, F. (1998) The stage for Neotropical fish diversification: a history of tropical South American rivers. Phylogeny and classification of Neotropical fishes (ed. by L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M. Lucena and C.A.S. Lucena), pp. 13–48. EDIPUCRS, Porto Alegre, Brazil. Moreno, J., Galva˜o, C. & Jurberg, J. (1999) Rhodnius colombiensis sp. n. da Coloˆmbia, com quadros comparativos entre estruturas fa´licas do geˆnero Rhodnius Sta˚l, 1859 (Hemiptera, Reduviidae, Triatominae). Entomologia y Vectores, 6, 601–617. Morrone, J.J. (1994) On the identification of areas of endemism. Systematic Biology, 43, 438–441. Morrone, J.J. (2002) The Neotropical weevil genus Entimus (Coleoptera: Curculionidae: Entiminae): cladistics, biogeography, and modes of speciation. Coleopterist Bulletin, 56, 501–513. Nelson, G. & Platnick, N. (1981) Systematics and biogeography: cladistics and vicariance. Columbia University Press, New York, USA. Nihei, S.S. & Carvalho, C.J.B. (2004) Taxonomy, cladistics and biogeography of Coenosopsia Malloch (Diptera, Anthomyiidae) and its significance to the evolution of anthomyiids in the Neotropics. Systematic Entomology, 29, 260–275. Page, R.D.M. & Charleston, M.A. (1998) Trees within trees: phylogeny and historical associations. Trends in Ecology and Evolution, 13, 356–359. Page, R.D.M. & Homes, E.C. (1998) Molecular evolution: a phylogenetic approach. Blackwell Science, Oxford, UK. Paula, A.S., Diotaiuti, L. & Schofield, C.J. (2005) Testing the sister-group relationships of the Rhodniini and Triatomini (Insecta: Hemiptera: Reduviidae: Triatominae). Molecular Phylogenetics and Evolution, 35, 712–718. Pinho, A.P., Gonc¸alves, T.C.M., Russell, N.S. & Cansen, A.M. (1998) The occurrence of Rhodnius prolixus Sta˚l, 1859, infected by Trypanosoma cruzi in the state of Rio de Janeiro, Brazil (Hemiptera, Reduviidae, Triatominae). Memo´rias do Instituto Oswaldo Cruz, 93, 14–143. Platnick, N.I. (1991) Commentary. Areas of endemism. Australian Systematic Botany, 4, xi–xii. Ramos, V.A. & Aleman, A. (2000) Tectonic evolution of the Andes. Tectonic evolution of South America (ed. by U.G. Cordani, E.J. Milani, A. Thomaz Filho and D.A. Campos), pp. 635–685. 31st International Geological Congress, CPRM/SGB, Rio de Janeiro. Ronquist, F. (1997) Dispersal–vicariance analysis: a new approach to the quantification of historical biogeography. Systematic Biology, 43, 195–203. 712

Sanderson, M.J. & Shaffer, H.B. (2002) Troubleshooting molecular phylogenetic analysis. Annual Review of Ecology and Systematics, 23, 49–72. Sanmartı´n, I. & Ronquist, F. (2004) Southern Hemisphere biogeography inferred by event-based models: plant versus animal patterns. Systematic Biology, 53, 216–243. Schaefer, C.W. (2005) Why are the subfamily relationships of Triatominae (Hemiptera: Reduviidae) important? Entomologı´a y Vectores, 12, 19–35. Schofield, C.J. & Dujardin, J.P. (1999) Theories on the evolution of Rhodnius. Actualidades Biolo´gicas, 21, 183–197. Sherlock, I.A., Guitton, N. & Miles, M. (1977) Rhodnius paraensis espe´cie nova do Estado do Para´, Brasil (Hemiptera, Reduviidae, Triatominae). Acta Amazonica, 7, 71–74. Swofford, D.L. (2002) Paup*. Phylogenetic analysis using parsimony (*and other methods), ver. 4. Sinauer Associates, Sunderland, MA, USA. Thompson, J.D., Gibson, T.J., Plewmak, F., Jeanmougin, F. & Higgins, D.G. (1997) The Clustalx window interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research, 25, 4876–4882. Valente, V.C., Valente, S.A.S., Carcavallo, R.U., Rocha, D.S., Galva˜o, C. & Jurberg, J. (2001) Considerac¸o˜es sobre uma nova espe´cie do geˆneroRhodnius Sta˚l, do estado do Para´, Brasil (Hemiptera, Reduviidae, Triatominae). Entomologia y Vectores, 8, 65–80. Vanzolini, P.E. & Papavero, N. (1969) ´Indice dos topoˆnimos contidos na Carta do Brasil 1:1000000 do IBGE. FAPESP, Sa˜o Paulo. Zink, R.M., Blackwell–Rago, R.C. & Ronquist, F. (2000) The shifting roles of dispersal and vicariance in biogeography. Proceedings of the Royal Society of London Series B, Biological Sciences, 267, 497–503.

BIOSKETCHES Alexandre Silva de Paula has a DS in Entomology from Universidade Federal de Vic¸osa, Brazil. His research focuses on the systematics and biogeography of Triatominae. He teaches Systematics at Centro de Pesquisas Rene´ Rachou/FIOCRUZ. Lile´ia Diotaiuti has a DS in Parasitology from Universidade Federal de Minas Gerais, Brazil. Her research focuses on Chagas disease vectors control in Latin America. She teaches Biology and Control of Triatominae, and Scientific Methodology at Centro de Pesquisas Rene´ Rachou/FIOCRUZ. Cleber Galva˜o has a DS in Veterinary Science from Universidade Federal Rural do Rio de Janeiro, Brazil. His research focuses on biology, systematics and comparative morphology of Triatominae. He teaches Medical Entomology and Protozoology at Instituto Oswaldo Cruz/FIOCRUZ.

Editor: Malte C. Ebach

Journal of Biogeography 34, 699–712 ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd