Keywords

1 Toward a Unified Species Concept

The diversity of life is measured essentially in terms of number of species, even though there is an ongoing debate focusing on what a species is and how organisms speciate. The last three decades have seen prominent challenges to the current views of species concepts and species delimitation due to the advances in molecular biology and genetics (Mallet 1995). The definition of a species will depend on which species concept you choose among the 27 available definitions (Mayden 1997; Wilkins 2011). By far, the most used and widespread definition is the biological species concept (BSC) , which considers species as groups of interbreeding individuals, with boundaries defined by intrinsic barriers to gene flow that have a genetic basis (Mayr 1963). The main limitation of the BSC is that populations of the same species found at a distance from each other (allopatric populations) that could not be suitably treated, because they are not in contact to randomly mate. Not even the successful crossing of allopatric populations under laboratory conditions will prove conspecificity since all ecological/geographical barriers are being removed (Claridge et al. 1985; Mallet 1995). Moreover, different cases of bona fide species hybridizing at secondary contact zones [i.e., lineages that occur at least partially in a same geographical area (sympatry) after the speciation process] are well-known. For example, although the malaria vectors Anopheles gambiae, A. coluzzii, and A. fontenillei (Diptera: Culicidae) are valid species, introgressed genomic regions are found that encompass genes associated with detoxification, desiccation tolerance, and olfactory perception, which are the characteristics that can alter their ability as malaria vectors (Barron et al. 2018).

Beyond the practical use of the BSC, the debate over what a species is and how it should be defined has been a matter of a long theoretical dispute among biologists. Personal expertise with respect to a particular research model or taxonomic group of interest has contributed to a “divergent radiation” in the proposal of species concepts. It is now clear that this “species definition competition” has generated more heat than knowledge. Recent countercurrent attempts have been made toward the proposal of a “unified species concept” (Table 1). Most species concepts agree in treating existence as separately evolving metapopulation lineage (i.e., “an inclusive population made up of connected subpopulations extended through time”) as the primary defining property of the species category, but they disagree in adopting different properties acquired by lineages during the course of divergence (e.g., intrinsic reproductive isolation, diagnosability, and monophyly) as secondary defining properties (secondary species criteria). In other words, lineages do not have to be morphologically distinguishable, diagnosable, monophyletic, intrinsically reproductively isolated, ecologically divergent, or anything else to be considered species, but only to be evolving separately from other lineages (for more information, see de Queiroz 2007). It is time to put aside disagreements about species definition and focus on empirical data that can be used as evidence of lineage separation and species boundaries. Taxonomists have to agree that the definition of robust species concepts depends upon several lines of evidence, including morphological traits and ecological and molecular data.

Table 1 Often-used contemporary species concepts and the properties upon which they are based (modified from de Queiroz 2007). de Queiroz’s (2007) proposal is an attempt toward the unification of all species concepts presented here. Properties marked with an asterisk should be viewed as operational criteria (lines of evidence) relevant to assessing lineage separation of a single general concept that defines species as separately evolving metapopulation lineages
Full size table

2 Insect Diversity and Speciation

Insects are one of the most diverse group of multicellular organisms, being represented by at least 10–30 million species (Erwin 1982), which accounts for 60–65% of all living eukaryotic biodiversity (Hammond 1992). The high diversity of insect taxa is partially explained by their compact size, which allows for the occupation of small and different portions of habitats and the specialization on the use of resources that larger animals are unable to exploit (Bush and Butlin 2004). Insects are often used as model organisms in evolution research due to their relatively short generation time and the practical advantages of laboratory rearing, enabling to test speciation hypotheses with proper sample sizes (Mullen and Shaw 2014).

Speciation is a subject that has intrigued investigators for centuries. The term was coined by the American biologist Orator F. Cook in 1906, as the process by which new species arise from existing ones (Cook 1906). However, knowledge advancement on this issue has been hampered by two main limiting factors: (1) the impossibility of witnessing the phenomenon unravels in real-time (with the exception of fast-evolving viruses; Meyer et al. 2016) and (2) the difficulty in reaching a consensus regarding the understanding of what a species is and how it should be delimited.

Although alternative methods to categorize the speciation process have been proposed (cf. Butlin et al. 2008), the most used concepts rely on the geographical context of speciation, which can be assigned to three broad categories: allopatric, parapatric, and sympatric speciation methods.

Allopatric speciation occurs when an ancestral population is divided into at least two daughter populations geographically isolated; in this context, gene flow between populations is absent or, if present, largely irrelevant. Thus, these populations accumulate mutations independently, develop some degree of genetic divergence, and might become genetically isolated. A complete allopatric speciation can occur if populations of incipient species develop pre- or postzygotic barriers for reproduction. In the case of a possible secondary contact zone, selection against hybrids (reinforcement) can occur and bimodal populations (admixed local populations with a deficit of hybrid genotypes) are observed. If sexual barriers are not complete and a secondary contact zone exists between species, hybridization events occur and thus the allopatric speciation is considered incomplete, with unimodal populations (intermediate hybrid genotypes predominating).

The grasshoppers Chorthippus brunneus and C. jacobsi (Orthoptera: Acrididae) are found in Spain at a narrow band along the north coast and south of the Cantabrian Mountain, respectively. These species possibly speciated in allopatry, but have been in contact since the Pleistocenic post-glacial range expansion (Bridle et al. 2002). They can be distinguished by the number of stridulatory pegs (although there is a small degree of overlap) and different male-calling songs (Bailey et al. 2004). In the contact zone, populations with bimodal distribution are observed, with strong assortative mating, based on spatial (probably associated with habitat specialization), seasonal, and behavioral isolation (Bailey et al. 2004). Other examples on insects illustrate hybrid zones with binomial distribution, such as observed in Heliconius butterflies (Lepidoptera: Nymphalidae), and ground crickets of the Allonemobius (Orthoptera: Gryllidae), which show strong prezygotic isolation due to assortative mating and homogamic fertilization (gamete recognition evolves faster than mate recognition), respectively (Howard et al. 1998). On the other hand, populations with unimodal distributions were observed in pine and larch budmoth host races of Zeiraphera diniana (Lepidoptera: Tortricidae), defined by Bush and Diehl (1982) as “populations of a species that are partially reproductively isolated from other conspecific populations as a direct consequence of adaptation to a specific host.” Behavioral and molecular studies indicate that the probability of hybridization between sympatric host races is around 2–3.5% (Emelianov et al. 2003, 2004). When in sympatry, a strong genomic heterogeneity between host races in areas where hybridization occurs was observed, but no genomic heterogeneity in divergent geographical populations of the same host race. These results suggested that the divergence with gene flow is driven by selection in sympatric regions and also that low hybridization rates are sufficient to homogenize much of the genetic variation in neutral genomic regions in terms of host adaptation.

Parapatric and sympatric modes of speciation are much more controversial among molecular biologists, since considerable interspecific gene flow hampers population divergence (cf. Jiggins 2006). Because there are no clear geographical barriers, levels of assortative mating, habitat preferences, local adaptation, and hybrid fitness reduction must overcome genetic homogenization mechanisms in order to achieve speciation. Simulation models and theoretical studies proposed that high population divergence indeed requires little or no gene flow (Orr 1995; Tang and Presgraves 2009; Nosil and Flaxman 2010). In a low gene flow scenario, it is possible for populations to diverge through the fixation of adaptive mutations via positive selection (Barrett et al. 2008; Nosil and Flaxman 2010), or simply through genetic drift in small populations. In those cases, natural selection can overcome genome homogenization (through gene flow and recombination) by maintaining isolated gene pools without the intervention of geographic barriers (Turelli et al. 2001).

Parapatric speciation can be explained as an ancestral population that becomes two daughter species occupying contiguous ranges (while sympatric speciation occurs when the geographical ranges of the daughter species overlap). In both cases, speciation seems to be shaped by disruptive selection, as a consequence of favoring the evolution of specialist over generalist species through niche-partitioning or microhabitat preference. The stick insects Timema cristinae (Phasmatodea: Timematidae) is a great example of parapatric speciation on its course. This species inhabits southwestern North America, feeding and mating on two different host plant species that differ in foliage and general morphology. Host-specific populations have differences in morphology and can live in parapatry (Nosil 2007). Surprisingly, significantly stronger sexual isolation mechanisms seem to occur in parapatry, which means that there is a sign of ecological reinforcement (Nosil 2007). Next-generation sequencing (NGS) analysis based on thousands of Single-nucleotide polymorphism (SNPs) revealed that host adaptation leaves subtle differentiation patterns across the genome. Moreover, divergent selection on traits not related to host use (i.e., genes not related to reproductive isolation) seems to be more relevant for generating genomic divergence between the populations. Under greater geographical separation, gradual reductions in gene flow facilitate speciation (Nosil et al. 2012).

Probably the most recognized example of sympatric speciation was observed in the apple maggot, the tephritid fruit flies sibling species complex Rhagoletis pomonella (Diptera: Tephritidae). Many researchers believed that the colonization of a new host in a sympatric environment and the further host preferences had started the reproductive isolation between host races based on different diapause and eclosion periods (Bush 1969; Filchak et al. 2000; Dambroski et al. 2005). From DNA sequence data of three nuclear loci and mtDNA, Feder et al. (2003) concluded that the host races became geographically isolated ~1.5 million years ago (Ma), and rare episodes of gene flow with inversion polymorphisms (restricting recombination) might have affected key diapause traits and formed adaptive clines. Therefore, these populations must have experienced a past allopatry in order to accumulate molecular changes (Xie et al. 2007) before became sympatric species. Nowadays, it is known that the barrier for gene flow remains incomplete (4–6% gene flow/generation), but most genome regions show significant geographic and host-associated variation that can account for by initial diapause intensity and eclosion time, which cause a temporal isolation between populations (Doellman et al. 2019). It is worth mentioning that sympatric populations of different host races are genetically more divergent in comparison to geographic populations within the races, which suggest that host races are being recognized as different genotypic entities in this region (Doellman et al. 2019).

The advances of molecular biology and mathematical models unveil that the geographical contextualized categories of speciation (allopatric, parapatric, and sympatric) are actually interconnected and depend on the time-frame in which they have been analyzed. As stated by Butlin et al. (2008), “At each stage of speciation, there is a spatial context on the sympatry to allopatry continuum which determines the extent of the extrinsic isolation between diverging populations.” Geographical isolation reduces homogenizing gene flow and facilitates speciation events, but the evolutionary forces that shape variability are also tightly linked to the ecological factors and the mating interactions in speciation events (Fitzpatrick et al. 2009; Nosil et al. 2009).

3 Overconservative Systematics and the Paraphyly of Triatoma

The subfamily Triatominae is composed exclusively by hematophagous insects and seems to have evolved from predaceous Reduviidae bugs ~40 Ma (Hwang and Weirauch 2012; Ibarra-Cerdeña et al. 2014; Justi et al. 2016), which coincides with the invasion and diversification of caviomorph rodents and small marsupials (Flynn and Wyss 1998; Poux et al. 2006; Antoine et al. 2012), and birds (Burns 1997) in South America.

This subfamily includes 150 extant and two extinct recognized species, which are classified in 16 genera and five tribes (Monteiro et al. 2018). These species occur mainly in the Americas including the Caribbean, but can also be found in southeast Australasia (Lent and Wygodzinsky 1979).

Triatoma is the most species-rich genus in the subfamily Triatominae and includes 73 species within the tribe Triatomini (Galvão and Paula 2014). Most of this diversification could be associated with cladogenetic events caused by climatic and geological changes occurred during the formation of the Americas (Hwang and Weirauch 2012; Justi et al. 2016; Monteiro et al. 2018) and can be well explained by vicariance.

Species of Triatoma have been clustered by several authors in different groups and complexes based on their external morphology and geographical distributions (Usinger 1944; Ryckman 1962; Usinger et al. 1966; Lent and Wygodzinsky 1979; Carcavallo et al. 2000). Since the beginning of the use of molecular markers to test evolutionary hypotheses in Triatominae, some authors have proposed rearrangements for this original classification (Schofield and Galvão 2009; de la Rúa et al. 2014; Pita et al. 2016). Some of these studies also included morphometry (de la Rúa et al. 2014) and chromosomal analysis by Fluorescence in situ hybridization (FISH) (Pita et al. 2016).

Based on new cytogenetic and morphometric data and phylogenetic results of the very important work by Hypša et al. (2002) (see below), Schofield and Galvão (2009) proposed the currently most accepted Triatomini assemblage, which subdivides species in three groups, eight complexes, and eight subcomplexes.

Historically, the use of molecular markers to study the phylogeny of Triatoma (Table 2) started with one or few mitochondrial genes and few representatives of these species groups and complexes (Lyman et al. 1999; García et al. 2001; Monteiro et al. 2001), advancing over time to analyze with more markers, including nuclear markers (Marcilla et al. 2001, 2002), and a growing addition of more Triatomini species (Hypša et al. 2002; de Paula et al. 2005; Hwang and Weirauch 2012; de la Rúa et al. 2014; Ibarra-Cerdeña et al. 2014; Justi et al. 2014, 2016; Pita et al. 2016). Those first studies with limited species representing the groups and complexes (Lyman et al. 1999; García et al. 2001; Monteiro et al. 2001); however, either showed weak support for the original classification based on morphological characters (Lent and Wygodzinsky 1979; Carcavallo et al. 2000) or were inconclusive (Table 2).

Table 2 Molecular studies presenting results that can be used to reject or not the arrangements of “species groups and complexes” in Triatomini tribe. Studies were based on different markers and are showed in a progressive order of number of species (N) representing those Triatomini groupings plus markers (with a few exceptions)
Full size table

It was only with the analysis of a larger and taxonomically more comprehensive set of triatomine specimens that it became clearly demonstrated that the proposed species groups and complexes did not comprise reciprocally monophyletic assemblages (Hypša et al. 2002). Phylogenetic analyses based on 12S and 16S mtDNA sequencing rejected the monophyly of Triatomini rearrangements and indicated the paraphyly of Triatoma with respect to Linshcosteus, Dipetalogaster, Eratyrus, and Panstrongylus (Hypša et al. 2002). Table 2 shows that the number of species used in phylogenetic studies (more than the chosen markers) was decisive to establish that the morphological classification of groups and complexes of Triatomini was not correct (“weak support/inconclusive”: 9–17 species; “rejection”: 18–56 species; Table 2). In fact, it is known that the addition of molecular markers and taxa in phylogenetic analyzes should increase its accuracy (Wiens and Tiu 2012).

What followed after the important work of Hypša et al. (2002) were phylogenetic studies continuing to demonstrate the fragility of the initial morphological grouping hypotheses, but with discussions still considering at least their partial validity (de Paula et al. 2005; Hwang and Weirauch 2012; de la Rúa et al. 2014; Ibarra-Cerdeña et al. 2014; Justi et al. 2014). Further research aimed at revealing new lines of evidence to help understand relationships within Triatomini.

The addition of biogeography analyzes brought a new light to an already promising integrative taxonomic scenario (Justi et al. 2016; Monteiro et al. 2018). In a recent review, Monteiro et al. (2018) presented a new taxonomic arrangement hypothesis to represent the relationships between species groups and complexes of Triatoma. The hypothesis aimed to incorporate current evolutionary theories into the traditional classification scheme based on morphology (e.g., Schofield and Galvão 2009), by including new molecular, cytogenetic, morphometric, and biogeographical data published ever since (Monteiro et al. 2018).

In addition to the presentation of a rigorous and updated classification based on literature data, the authors proposed a new nomenclature consistent with the evolutionary scenario that relied on two main observations: (1) studies that reinforced the paraphyly of Triatoma also clearly supported the existence of three lineages in Triatomini (Justi et al. 2014, 2016; Monteiro et al. 2018) and (2) the meaning of the term “species complex” in triatomine systematic studies varies depending on the context from “subgeneric assemblages defined by morphological similarity” (e.g., Lent and Wygodzinsky 1979; Schofield and Galvão 2009) to “cryptic species” (i.e., morphologically indistinguishable species; e.g., Monteiro et al. 2003).

Therefore, Monteiro et al. (2018) proposed an arrangement for the Triatomini that followed an hierarchy of: (1) three major evolutionary “lineages” composed by Triatoma dispar, “North American,” and South American; (2) 11 “clades” within lineages defined by common ancestry and broad biogeographic correspondences; and (3) 19 “species groups” within clades, with some of these groups matching “species complexes” defined as closely related, morphologically similar or even indistinguishable species usually disclosed as a result of molecular investigations (Table 3 and Fig. 3 of Monteiro et al. 2018). The meaning of the term “species complex” in triatomine systematic studies varies depending on the context from “subgeneric assemblages defined by morphological similarity” (e.g., Lent and Wygodzinsky 1979; Schofield and Galvão 2009) to “cryptic species” (i.e., morphologically indistinguishable species; e.g., Monteiro et al. 2003).

Of the three lineages designation proposed by Monteiro et al. (2018), the “North American” lineage has the greatest morphological diversity and comprises most nominal genera (nine). In comparison, the South American lineage has only two genera: Triatoma and Eratyrus. The high morphological plasticity of Triatominae (Dujardin et al. 1999, 2009) can lead to misidentification and taxonomic uncertainties (Pita et al. 2016). However, most of the diversification seen in the “North American” lineage seems consistent with phylogenetic evidences (Galvão et al. 2003).

Although there are still many issues within Triatomini to be clarified, the accumulation of data in the literature has already shown that Triatoma is not monophyletic. Is it time to discuss the suitability of a taxonomic revision? Should the “North American” lineage retain the generic epithet “Triatoma” as it includes the type species of the genus, Triatoma rubrofasciata?

4 Phenotypic Plasticity and Classical Taxonomy

Phenotypic variability affects traits often used in classical taxonomy including color patterns (Abad-Franch et al. 2009; Pavan et al. 2015) or the size and shape of bodies, heads, wings (Schachter-Broide et al. 2004; Hernández et al. 2011; Nattero et al. 2013; Sandoval et al. 2015), and genital structures (Schofield and Galvão 2009). Chromatic variations of single or near-sibling species can result from adaptive plasticity, which may confound taxonomic classification. Indeed, the discovery of cryptic lineages with different vector capacity(ies) and also chromatic variants of single species from different micro-environments were some of the greatest achievements on the taxonomy of triatomines in the early 2000s.

Rhodnius neglectus is the most abundant Rhodnius species in the Cerrado biome. It inhabits various palm tree species, including those of the genera Attalea, Acrocomia, Mauritia, Oenocarpus, and Syagrus (Gurgel-Gonçalves et al. 2004; Abad-Franch et al. 2009). This species can be misidentified as R. nasutus, particularly in the nearby Caatinga biome and in Caatinga-Cerrado transitional areas (Dias et al. 2008; Lima and Sarquis 2008). Rhinacanthus nasutus is found predominantly in the Caatinga inhabiting Copernicia prunifera palms (Sarquis et al. 2004), but may also be found in other palms and trees in this region (Dias et al. 2008; Lima et al. 2012). The high similarity between those two vector species and the lack of reliable diagnostic characters leads naturally to an uncertainty regarding proper taxonomic identification and determination of their geographical boundaries.

The identification of these species is based on morphological characters as chromatic patterns of body and antennae, overall body size, and male genitalia (Lent and Wygodzinsky 1979). However, Harry (1993b) detected no clear-cut differences in the male genitalia structures; in addition, important chromatic variation has been described in both species (Barrett 1995).

Abad-Franch et al. (2009) applied a geometric morphometrics aiming to differentiate R. neglectus from R. nasutus and found wing and head shape differences between these species. Some specimens from Curaçá, Bahia (Brazil) collected in C. prunifera palms, although phenotypically similar to R. nasutus, were clustered within the R. neglectus group, while others from the same location were clustered within R. nasutus. If morphometry is able to correctly assign both species, these results showed that R. neglectus and R. nasutus are sympatric in the Cerrado-Caatinga transitional area and the former species may have chromatic forms similar to those observed in the latter.

Recent observations of Rhodnius insects at Caatinga and Caatinga-Cerrado revealed specimens with dubious chromatic patterns. Individuals collected in M. flexuosa palms had a dark phenotype, a similar color to the palm fibers and base of fronds, and with coloration and diagnostic traits of R. neglectus. However, those collected in C. prunifera palms displayed a lighter chromatic pattern more similar to that of R. nasutus (Pavan et al. 2015). Since R. neglectus and R. nasutus may occur in sympatry (Abad-Franch et al. 2009), it raises the possibility of natural hybridization.

An alternative explanation for this observation is that R. neglectus would exhibit one chromatic phenotype similar to R. nasutus and different from the pattern described by Lent and Wygodzinsky (1979). If correct, the lighter coloration of R. neglectus from C. prunifera may be naturally selected. This coloration might have improved its chances of survival and reproduction, since they would be camouflaged with the light substrate of C. prunifera fibers. Therefore, populations with light phenotype increased in frequency in these palm trees, as they would be less conspicuous and thus less predated than the typical phenotypes.

Phenotypic variation was also observed for R. nasutus, and it seems to be governed by the microhabitat it lives in. In Ceará, Brazil, this species was collected in five different palm tree species (Dias et al. 2008). The holotype of R. nasutus has a pale brownish-yellow coloring, with a red-like appearance and dark brown dots in certain regions of the body and appendices (Lent and Wygodzinsky 1979). Although populations inhabiting C. prunifera palms presented a reddish color, according to the original species description, other populations from A. intumescens, A. speciosa, M. flexuosa, and S. oleracea palms were chestnut-colored (Dias et al. 2008). As observed for R. neglectus, body coloration of R. nasutus specimens corresponded exactly to the fibers and base of fronds, strengthening the hypothesis that Rhodnius species have genes, which provide a menu of different phenotype possibilities, and the environment determines the phenotypic outcome by natural selection.

5 Tempo and Mode of Triatomine Speciation

As summarized in a recent review paper on the evolution and biogeography of the Triatominae (Monteiro et al. 2018), most studies focusing on the existence of cryptic triatomine taxa using molecular markers have often relied on two species concepts: the Biological Species Concept (BSC , Mayr 1963) and the Phylogenetic Species Concept (PSC, Cracraft 1989). We have learned that allopatric speciation seems to be the rule for most Reduviids (Monteiro et al. 2018). Here, we present three examples of triatomine speciation that probably involved vicariance and diversification with low/no gene flow among ancestral lineages. It is widely accepted that speciation is a process that requires very long time intervals to take place, usually hundreds of thousands of years (Butlin et al. 2008). With regard to the time needed for triatomines to speciate, two hypotheses were put forth that clearly challenged the tempo required for traditional insect speciation to occur (see below).

5.1 Fast or Slow Diversification?

5.1.1 Triatoma rubrofasciata and Old World Triatominae

The first hypothesis was advanced in an attempt to account for the occurrence of the six Linshcosteus and seven Triatoma species found in the Old World. It was suggested that they all descend from T. rubrofasciata , as a result of merchant shipping between the Americas and Asia during the sixteenth to seventeenth centuries, perhaps as a consequence of very fast (300 years) adaptive radiation processes (Schofield 1988; Gorla et al. 1997; Patterson et al. 2001; Schofield and Galvão 2009; Dujardin et al. 2015a, b). It is now well established that all Old World Triatominae are monophyletic and likely derive from a successful founding event that occurred approximately 20 Ma, with ancestral triatomine populations crossing the Bering land bridge, likely benefiting from the association with rodents, ultimately reaching Eurasia (Hypša et al. (2002); Patterson and Gaunt 2010; Justi et al. 2016).

5.1.2 The Origin of Rhodnius prolixus

The second hypothesis suggested that R. prolixus , the most important Chagas vector in Venezuela, Colombia and parts of Central America, is a domestically adapted “derivative” of a sylvatic R. robustus lineage, and that speciation was the consequence of a “discrete event in Venezuela at some time after the establishment of European settlements in the 16th century” (Schofield and Dujardin 1999). This hypothesis was proposed based on morphometric (Dujardin et al. 1998, 1999) and genetic evidence (allozymes and mtDNA; Harry et al. 1992, Stothard et al. 1998, respectively), available at the time, which pointed to a lack of phenotypic and genetic variability in R. prolixus populations. Further research relying on better sampling of both wild and domestic R. prolixus populations collected from six Venezuelan states and analyzed for mtDNA and microsatellites have challenged this view by revealing high levels of genetic variation (Fitzpatrick et al. 2008).

5.2 Vicariance and Allopatric Speciation of Triatomines

5.2.1 Rhodnius robustus and the Refugium Theory

For many years , the taxonomic status of R. robustus was questioned due to a combination of three factors: morphological similarity, loose diagnosis, and poor sampling (cf. Monteiro et al. 2003; Pavan and Monteiro 2007). Although indistinguishable according to morphological and isozymic analyses (Harry 1993a, 1994), these species play very different epidemiological roles—R. prolixus is an efficient domestic vector, whereas R. robustus populations are entirely sylvatic. Monteiro et al. (2003) put an end to the controversy of the validity of R. robustus as a bona fide species through the analysis of DNA sequences of mitochondrial and nuclear markers (663-bp fragment of cytochrome b (cytb) and the D2 variable region of the 28S nuclear RNA), revealing that R. robustus is not only a valid species separated from R. prolixus, but also represents a paraphyletic complex of at least four cryptic lineages (R. robustus I, II, III, IV). Pavan et al. (2013) further confirmed the paraphyletic assemblage of R. robustus with respect to R. prolixus through the analysis of another nuclear marker, the fourth intron of the transmembrane protein 165 (TP165) gene. The separation of R. prolixus and R. robustus was further corroborated by a behavior study showing that nymphs of R. prolixus and R. robustus II display different locomotor activity patterns on an automated recording system (Pavan et al. 2016).

The first attempt to associate triatomine phylogeographic patterns with possible vicariant events based on molecular clock time-estimates was published in 2003 by Monteiro and collaborators. A particular and notorious example of vicariant speciation is that of the refugium theory, advanced to account for the pattern of diversification seen in the Amazon region. The view that diversification of the Amazonian biota was caused by glaciation cycles during the Pleistocene was first introduced by Haffer (1969). The theory attempts to explain the latest of the series of differentiation events beginning in the Cenozoic that contributed to the development of the modern biota of the Amazon basin . In short, it is based on the premise that climatic changes during the Pleistocene caused rain forests to contract into isolated pockets separated by savannah. This would have confined small populations and favored their divergence by genetic drift, which would have facilitated allopatric speciation (Monteiro et al. 2003). The authors used in their phylogeographic inferences the value of 2.3% of sequence divergence per million years estimated for recently diverged arthropod taxa (Brower 1994). They concluded that all estimates between the clades within both Amazon and Orinoco regions are compatible with a Pleistocene origin and are consisted with the refugium theory (Monteiro et al. 2003).

5.2.2 Triatoma rubida and the Baja California Peninsula

Triatoma rubida was initially described as five morphologically distinguishably allopatric subspecies based mainly on chromatic differences in markings along the conexivum, distributed in Mexico and the USA: T. rubida rubida from the Cape region, Baja California Sur, T. rubida cochimiensis from Central Baja California peninsula, T. rubida jaegeri from Pond Island, Gulf of California, and T. rubida sonoriana from Sonora (all strictly Mexican subspecies); and T. rubida uhleri from Veracruz, Mexico, and Southwestern USA (Usinger 1944; Ryckman 1967). The “five subspecies” proposition was, however, later challenged in the 1979 revision of Lent and Wygodzinsky, who stated: “Although specimens seem to cluster around the phenotypes mentioned, not all fall easily into the categories listed above; there does seem to be a prevalence of comparatively light-colored, large-sized forms in the north and of smaller, more intensely pigmented forms in the southern part of the total range of the species. Much more abundant material than that examined by us, especially from Mexico, combined with rearing experiments, is needed for an understanding of the biosystematics of Triatoma rubida.”

The separation of the Baja California Peninsula from mainland Mexico during the formation of the Gulf of California 5–8 Ma is believed to be the vicariant event that caused the geographic isolation of ancestral T. rubida populations and gave rise to T. rubida cochimiensis (Baja peninsula) and T. rubida sonoriana (Sonora). Pfeiler et al. (2006) used this geological event to calibrate the first mtDNA molecular clock for triatomines: 1.1–1.8% pairwise sequence divergence per million years (lower than the 2.3% divergence for mtDNA generally applied to insects; Brower 1994).

5.2.3 Triatoma dimidiata and the Isthmus of Tehuantepec

The T. dimidiata cryptic species complex was first recognized by Marcilla et al. (2001) based on ITS-2 sequence divergence between bugs from Yucatan and specimens from elsewhere in Mexico and from Central and South America. Following studies based on cytogenetics and genome size (Panzera et al. 2006) and mtDNA (Dorn et al. 2009; Monteiro et al. 2013) corroborated these observations. Relying on mtDNA markers (cytb and ND4), Monteiro et al. (2013) described five genetically well-differentiated, monophyletic groups (named groups I–IV plus T. hegneri). Their results revealed that mtDNA groups I, II, and III match, respectively, ITS-2 groups 1, 2, and 3. Group IV represented cave-dwelling Belize specimens. As pointed out by Bargues et al. (2008) and Monteiro et al. (2013), some of these genetically divergent groups clearly deserved specific status. In accordance with these orientations, the two genetically most divergent groups III and IV were recently raised to the specific level and formally described as T. mopan and T. huehuetenanguensis, respectively (Dorn et al. 2018; Lima-Cordón et al. 2019).

With regard to Groups I and II (and based on their present distribution), as the Isthmus of Tehuantepec is known to represent an important recent geological barrier for a number of sister taxa of birds, mammals, and butterflies, Monteiro et al. (2013) have suggested that the Isthmus of Tehuantepec orogeny (15–5 Ma) might have been the vicariant event responsible for the splitting of the ancestral population that led to their origin. Although groups I and II still have a subspecies status, we argue that they merit specific status.

5.3 Parapatric/Sympatric Triatomine Speciation

5.3.1 Triatoma brasiliensis Complex and the Homoploid Hybridization Hypothesis

Organisms may also speciate quite rapidly via polyploidy (Lukhtanov et al. 2015). Polyploidy (or hybrid speciation) is the term given to a set of processes whereby two species hybridize and instantly generate a third new species. They can be classified as allopolyploidy (i.e., involving a genome-doubling event that provides reproductively isolation); or homoploid hybrid speciation that occurs without an increase in ploidy (Coyne and Orr 2004).

The northeastern Brazil Chagas species complex Triatoma brasiliensis was comprised of three subspecies—T. b. brasiliensis, T. b. macromelasoma, and T. b. melanica—defined based on chromatic differences of the pronotum, legs, and hemelytra (Galvão 1956). These subspecies were, however, synonymized by Lent and Wygodzinsky (1979), who argued that intermediate forms could be found in nature. Further allozyme-based analyses showed the three subspecies were real evolutionary lineages, and yet another form was later discovered (juazeiro form; Costa et al. 1997). Monteiro et al. (2004) confirmed the existence of the four forms based on mtDNA cytb phylogenetic analysis of specimens collected from the whole distribution area of the species. Juazeiro and melanica forms were raised to the specific level and formally described as T. melanica (Costa et al. 2006) and T. juazeirensis (Costa and Felix 2007). Kimura-2-parameter distances based on mtDNA evidence that bona fide sister Triatoma species diverge in more than 7.5% (K2P > 0.075) (cf. Monteiro et al. 2004), while intraspecific variation does not exceed 2%. Genetically less divergent sister forms brasiliensis and macromelasoma diverged in more than 2% (K2P = 0.027), and thus, were given subspecific ranks (Costa et al. 2013).

Costa et al. (2009) have analyzed morphometric, morphological, ecological, and geographic distribution data to advance the hypothesis that T. brasiliensis macromelasoma is a product of hybridization between the subspecies T. b. brasiliensis and T. juazeirensis. Authors acknowledge, however, that the evidence presented is not yet conclusive and that further studies are required to strengthen their claim (Costa et al. 2009). The subject has been recently revisited in a study based on chromosomal analysis, and band sizes of an ITS-1 PCR-amplified fragment (Guerra et al. 2019). Those authors also sequenced DNA from the three taxa for a fragment of the ND1 mitochondrial gene, which gave unexpectedly low pairwise genetic distances (Tamura-Nei < 0.006), pointing to a possible problem with the taxonomic identification of the specimens themselves (Guerra et al. 2019). It is well established that this magnitude of differentiation characterizes within-population or, at most, within-species levels of variation (Monteiro et al. 2004). Not surprisingly, all species showed the same cytogenetic characteristics (Guerra et al. 2019).

The speciation process of the T. brasiliensis complex probably involves ecological and/or temporal barriers (sympatric areas in the present may represent secondary contact zones of parapatric/allopatric populations), since they still have not evolved either pre- or post-mating barriers, as revealed by successful hybridizations in laboratory conditions (Almeida et al. 2012). New studies on ecology and genomics focusing on possible ecological selection (which would prevent backcrossing), behavioral changes (e.g., different periods of activity), or even chromosomal arrangements are still needed to clarify this issue.

5.3.2 The Rhodnius pallescens—R. colombiensis: A Case of Sympatric Speciation?

The Rhodnius pallescens complex is composed by three recognized species that occur in the trans-Andean region (Pacific area of South and Central America)—R. pallescens, R. colombiensis , and R. ecuadoriensis. Rhodnius. ecuadoriensis is restricted to southern Ecuador and northern Peru, occupying Phytelephas aequatorialis palms (Abad-Franch et al. 2009). This species is isolated from the others of this complex by geographical barriers, the Andean mountains, and probably speciated through allopatry (Galvão et al. 2003; Abad-Franch et al. 2009). Rhodnius. pallescens is widely distributed across Central America and Colombia in different ecological zones, inhabiting Attalea butyracea and Cocos nucifera palm trees (Díaz et al. 2014). Rhodnius. colombiensis seems to be restricted to the Andean Valley of Magdalena River in central Colombia (Moreno et al. 1999). Although this species inhabits the same ecoregion and same palm tree species as R. pallescens, natural hybrids had not been reported (Díaz et al. 2014).

Laboratory crosses reveals the existence of both pre-zygotic and post-zygotic reproductive barriers . Female R. pallescens I and male R. colombiensis do not produce progeny, while female R. colombiensis and male R. pallescens I produce infertile F1 hybrids (Gómez-Palacio et al. 2012). Cytogenetics analyses reveal that R. colombiensis structural chromosomes suffered rearrangements and DNA loss in comparison to the other species of the complex (Panzera et al. 2007; Gómez-Palacio et al. 2012; Díaz et al. 2014). A clear demarcation of the biogeographical distribution of the four lineages of the complex and additional analyses within and between R. pallescens lineages using different molecular markers are still needed for a better knowledge on the evolutionary trends, geographical dispersion, and signs of possible adaptive radiation.

6 Toward an Integrative and Evolutionarily Sound Taxonomy

We emphasize the importance of integrating morphological, ecological, behavioral, and molecular tools to elucidate epidemiological and taxonomic unresolved questions in triatomines. Abad-Franch et al. (2013) performed an integrative taxonomic analysis to describe Rhodnius barretti as a new species of triatomine. They evaluated traditional morphological traits, morphometric data, and molecular phylogenetics using a fragment of cytb. This species is difficult to distinguish phenotypically from those of the R. robustus lineage, with the exception of the sympatric R. robustus II that presents chromatic (lighter coloration) and size (larger individuals) differences. However, R. barretti differs from R. robustus s.l. in the shape of both head and wings, and also in length ratios of certain anatomical structures. Moreover, phylogenetic reconstructions showed that this species is a basal member of the “R. robustus lineage,” which encompasses R. nasutus, R. neglectus, R. prolixus, and the other five members of the R. robustus complex (Abad-Franch et al. 2013).

Besides the R. prolixus genome (Mesquita et al. 2015), the mitochondrial genomes of T. dimidiata and T. infestans are already available (Dotson and Beard 2001; Pita et al. 2017). As genome-sequencing is increasingly employed for nonmodel organisms, the ability to evaluate the taxonomic identity or status of a particular triatomine species via transcriptomes, proteomes, or metabolomes is now possible. These approaches have recently begun to be applied to triatomines, including T. brasiliensis (Marchant et al. 2015), T. dimidiata (Kato et al. 2010), R. prolixus (Ribeiro et al. 2014), and T. infestans (Traverso et al. 2016; Gonçalves et al. 2017). Genomic data were also useful for the identification of parasite, vector, and the microbiota present in T. dimidiata (Orantes et al. 2018). In the context of near-sibling species and varieties of single species, Brito et al. (2019) recently synonymized Rhodnius montenegrensis as R. robustus. Most likely, the upcoming years of triatomine research will present us with the gathering of increasingly large datasets that contain separate lines of evidence from independent loci.

An alternative approach for generating genomic data at a lower cost for population genetics studies and phylogenetic analyses of closely related species is the double-digested restriction-site-associated DNA sequencing (ddRAD-seq) method. This technique was first employed as a population genomics study to infer the structuring of R. ecuadoriensis populations in Ecuador (Hernandez-Castro et al. 2017). This method increases the coverage of different regions of the genome and recovers reliable microsatellite and SNP data (Davey and Blaxter 2011; Kai et al. 2014).

An overlooked issue in population studies of triatomines is the difficulty of infestation foci detection, especially when colonies are small and occupy structurally complex ecotopes (Abad-Franch et al. 2010, 2014; Valença-Barbosa et al. 2014; Pavan et al. 2015). A comprehensive approach must include genetic and ecological data of triatomine species to better understand the adaptive nature of plasticity (whether is heritable and ontogenetic), and detailed frequencies of different chromatic variations in the environment (Murren et al. 2015).

New genomic tools can help explore adaptive plasticity and the following approaches deserve further attention: (1) the “omic” basis behind it, (2) comparative genomics of near-sibling species to understand its evolution, and (3) epigenetic components of inheritance that may influence plastic responses (Richards et al. 2010; Glastad et al. 2011; Zhang et al. 2013; Murren et al. 2015).