Spatial congruence between biotic history and species richness of Muscidae (Diptera, Insecta) in the Andean and Neotropical regions

Introduction

It is believed that the same ecological factors, or at least different combinations of these same factors, can explain species distribution patterns in any region on Earth (Hawkins et al. 2003). However, the functional ecological approach (Hubbel 2001) usually treats different species equally, rejecting particular information on taxon history and macroevolutionary dynamics. Indeed, this information could be used to reconstruct area history (Humphries and Parenti 1986) and to speculate on the historical scenario that leaded the areas to be species rich or poor.

Species richness has been rarely associated to area history. This may happen because of three factors: (i) species richness studies began under ecological paradigms (e.g. Dobzhansky 1950; Pianka 1966), (ii) the link between Earth and life (Croizat 1964) is neglected and (iii) a methodological impediment (see Crisci et al. 2003 for methods compilation). Due to this ‘distance’ between historical approach and species richness analyses during the development of these research fields, no historical method has been developed to handle many taxa simultaneously to compare and associate species richness to biotic history (Crisci et al. 2003). In traditional historical biogeography, methods were employed to rescue past spatial arrangements and related processes of few taxa of interest (Crisci et al. 2006) instead of dealing with large collections of species, or species richness itself. On the other hand, ecological analyses usually deal simultaneously with large number of taxa and information about their distribution patterns are merged into a single parameter (i.e., species richness), creating difficulties in dealing with particular historical information at the supraspecific level. By considering the above mentioned facts, the knowledge of factors driving species richness may be improved by taking into account vicariant paradigms (Croizat 1964).

Thus, to compare species richness and area history it is necessary to have an evolutionary interface, and the recently proposed niche conservatism model may be a feasible background hypothesis (see Wiens and Donoghue 2004). Niche conservatism is an intrinsic species trend in keeping the ancestral characteristics along time. The derived forms are related to an evolutionary answer (niche evolution) due to environmental changes (Wiens and Graham 2005). This evolutionary trend has been seen in many studies of animals and plants (Losos et al. 2003; Graham et al. 2004) and was pointed out as an important issue in distribution modelling, with implications to biogeographical and richness gradient studies (Peterson et al. 1999).

The species richness association with phylogeny exposes its evolving structure and there is a recent trend towards incorporating evolutionary elements that may put the issue in a more explicit historical framework (see Mittelbach et al. 2007 for a recent review). By considering that it is possible to recognize areas richer in basal and derived taxa (Hawkins et al. 2005, 2006), these richness patterns could be associated to results from cladistic biogeography. The possible confluence of results may evidence that species richness patterns may be evolving structures that could be linked to the area history.

In the Neotropical region, the Caribbean (Ca) formation was complex, and resulted from many geological events (Rosen 1975; Grehan 2001). Biotic information supports Ca as sister of the continental, which is compounded by Northwest (NW) and Southeast (SE) (Amorim and Pires 1996). The region had been extensively studied in detail for several organisms and with many methods (e.g. Cracraft 1985; Amorim and Pires 1996; Morrone 2006). Although different hypotheses were proposed for the relationship among endemic areas within the NW and SW components (Camargo and Pedro 2003; Nihei and de Carvalho 2007a), the general relationship of [Ca (NW, SE)] was recurrently corroborated (Mermudes 2005; Espinosa et al. 2006; Paula et al. 2007). On the other hand, for the Andean region different hypotheses are available for the relationship among its components. The first hypothesis was drawn based on Curculionidae species (Coleoptera) and it suggests that the subregions are monophyletic (Morrone 1994). However, subsequent studies diverged about the relationships among provinces and no general pattern was yet recognized (Amorim and Pires 1996; Posadas and Morrone 2001; Soares and de Carvalho 2005a).

In this paper, we used Muscidae as model organisms to verify whether the species richness evolving as a ‘structure’ is spatially congruent to biotic history, using niche conservatism as a background mechanistic hypothesis underlying speciation and extinction processes. If the relationship between organism and space can be established for a single taxon, this could also be done for several taxa, so that a parallel analysis between species richness and cladistic biogeography can be drawn. Muscidae, as many insect groups, are a sensible and precise model for understanding a scenario of evolutionary changes because they have a short life cycle than vertebrates. Generally speaking, insects have more generations per year than vertebrates and this fact allows faster and diversified responses to environmental pressures, such as climatic changes.

We believe that there is sufficient information about muscid systematics and biogeography to perform an analysis. Systematic knowledge includes at least four phylogenetic hypotheses for the family (Hennig 1965; de Carvalho 1989; Couri and de Carvalho 2003; Schuehli et al. 2007), for three tribes – Azeliini, Coenosiini and Muscini – (Couri and Pont 2000; Savage and Wheeler 2004; Nihei and de Carvalho 2007b; respectively) and for 11 Neotropical genera. In the Neotropical region, areas of endemism and diversification were identified by historical methods by de Carvalho et al. (2003) and Nihei and de Carvalho (2005).

A previous study on Muscidae also suggested that geological events in the Andean region had a paramount role in species spatial arrangement. An analysis of Apsil, Palpibracus and Reynoldsia revealed the diversification areas and a recent event of dispersion (Soares and de Carvalho 2005a). Ocean transgressions in the Andean southwest coast during Late Oligocene and Early Miocene (Donato et al. 2003) seem to be responsible for the diversification of these genera (Soares and de Carvalho 2005a; de Carvalho and Pont 2006) in the region. More recently, Pleistocenic glaciations had important role in the southern occupation area which was covered by an ice sheet until 42°S. This event affected species distributions that were spatially shifted northwards or locally got extinct, creating then patterns in species richness. At the same time, all modern Muscidae occurrences southern 42°S could be potentially explained by postglaciation dispersion events (Soares and de Carvalho 2005a).

Methods

Eleven Neotropical and Andean genera of Muscidae were chosen for this analysis, based on the available systematic reviews and phylogenetic hypotheses. The geographic information was gathered from the literature listed in the Neotropical Muscidae Catalog (de Carvalho et al. 2005). The genera analysed and the phylogeny hypotheses references were: Apsil (de Carvalho and Couri 2002), Bithoracochaeta (Couri and Motta 2000), Brachygasterina (de Carvalho and Pont 2006), Cyrtoneurina (Pamplona 1999), Cyrtoneuropsis (Pamplona 1999), Palpibracus (Soares and de Carvalho 2005b), Philornis (Couri et al. 2007), Polietina (Nihei and de Carvalho 2005), Pseudoptilolepis (Schuehli and de Carvalho 2005), Reynoldsia (de Carvalho and Couri 2002) and Souzalopesmyia (de Carvalho 1999). These genera contain a total of 171 species, about 20% of the total number of described species for the biogeographical regions studied. However, species with less than three known occurrence points were excluded from the dataset, so that the final analyses were performed on 100 species.

The phylogenetic information on each genus was employed to split the species in two groups: ‘basal’ and ‘derived’. Terminal taxon root distances (RD) were calculated for each genus phylogeny. This value is the number of nodes between the cladogram root and a given terminal taxon. It quantifies the lineage evolutionary changes. This value was divided by the total number of nodes in the genus cladogram and the coefficient was associated to each terminal taxon. Species were arranged in decreasing order of coefficient value and the upper third were classified as ‘derived’ species, the lower third as ‘basal’ (Hawkins et al. 2006) (see Table S1).

Richness maps were constructed using the species’ geographic range defined by minimum convex polygons based on occurrences points, and were mapped on a 220 × 220-km grid cell, approximately 2 × 2° at the equator. All species considered ‘basal’ from different genera were grouped (n = 31) and a richness map constructed. For the ‘derived’ group (n = 34) the same was done and a total richness map was constructed with all species (n = 100) as well, including taxa with intermediate RD values.

A difference richness map was drawn by subtracting basal from derived richness (=derived−basal), so that grid cells with positive values contain more derived species and the negative values in grid cells contain more basal species. This map was compared to biogeographical components relationship proposed by Amorim and Pires (1996) for the Neotropical region.

A component analysis (Nelson and Platnick 1981) was performed when the difference between basal and derived richness within a region could not be detected (values close to zero). This procedure was employed to verify, through a cladistic methodology, if there is a congruent taxon-area history. The results may help establishing why the area has no richness evolving structure under vicariant paradigms. The terminal taxa were replaced by province name of occurrence (following the scheme of Morrone 2006) for each genus phylogeny and all cladograms were compared using component 2.0 (Page 1993). The following options were assumed: absent associate, heuristic search, nearest-neighbour interchanges and criterion to minimize: leaves added. A strict consensus cladogram was obtained, and this analysis was performed for the species of Apsil, Brachygasterina, Palpibracus and Reynoldsia in the Andean region. The resulting area cladogram was compared to previous area relationship hypotheses recognized by other studies (Morrone 1994; Amorim and Pires 1996; Posadas and Morrone 2001).

Results

For the 11 genera analysed, the total richness map showed high richness in a NW-SE diagonal zone throughout South America. These areas encompass the Panama isthmus, Amazonia, South American NW countries up to Atlantic forest, in southeastern Brazil. Also, Central Chile in southern South America shelters a high Muscidae species richness (Fig. 1). On the other hand, basal richness was higher in northwestern areas when compared with total richness, whereas derived richness took place in central-southeastern Brazil, being then complementary when compared with the total richness (Fig. 1). Difference between basal and derived richness mapped remarkable well over the components of Neotropical history biotic proposed by Amorim and Pires (1996) (Fig. 2).

In the Andean region, basal and derived richness did not generate any pattern and the difference between them was not detected (1, 2). The component analysis performed with the Andean genera Apsil, Brachygasterina, Palpibracus and Reynoldsia resulted in six cladograms and a strict consensus cladogram was generated. The area relationship based on Muscidae species grouped Santiago, Maule + Valdivian Forest, and Magellanic Forest in a polytomy. The Central Chilean subregion, constituted by Coquimbo and Santiago provinces, was arranged as a paraphyletic group (Fig. 3). These results were divergent from the monophyletic subregion proposal by Morrone (1994) and they were not congruent with previous relationships established based on Curculionidae studies (Morrone 1993; Amorim and Pires 1996; Posadas and Morrone 2001).

Discussion

Basal and derived Muscidae richness showed a very explicit area pattern in the Neotropical region. However, in contrast to contemporary climate hypothesis (Wright et al. 1993; Hawkins et al. 2003), these results evidence that the Muscidae richness spatial pattern is evolutionarily based and that historical events had a paramount role in the pattern establishment (Ricklefs 1987, 2006). Although not presented here, due to low representativeness of the studied species in relation to overall species pool, a visual inspection of the richness maps shown in Fig. 1 clearly shows that contemporary climate hypothesis has relatively low power in explaining spatial patterns in Muscidae genera studied here (see patterns in Hawkins et al. 2006).

It is striking that the separation in basal and derived groups generates richness patterns that are directly comparable with biogeographical components. Amorim and Pires (1996) argued about the basal position of the Caribbean component related to the sister group (NW and SE). The NW and SE component split was explained by the formation of the Amazon Lake, through Mamore, Madeira and Amazon rivers in Late Cretaceous. The components relationship was hypothesized based on callitrichines monkeys (non-aged information) and non-Brachycera dipterans phylogeny and geographic distribution. The Muscidae gradient richness in the Neotropical region corroborates this biogeographical components relationship, since basal species richness was more related to the Caribbean and NW components, whereas derived species richness was related to SE component (Fig. 2). This relationship was tested by Polietina species under BPA cladistic methodology in the Neotropical region. The results corroborated the Amazonian composite hypothesis (Amorim and Pires 1996) and advanced in the Southeastern component detail (Nihei and de Carvalho 2007a).

From the association between species richness and phylogeny it is impossible to understand which geological or climatic event was responsible for a specific genus distribution pattern. However, as richness is a cumulative parameter through time, the method may clarify patterns that would be obscure in cladistic biogeography analysis. The replication in cladistic methods may get results confused and not objective. The general pattern found in richness analysis is very robust and even with some interference of sympatric speciation, for example, it can be recognized. In some cladistic biogeography analyses, mainly when morphological phylogenies are employed, some of the results may be due the pseudo-congruence. Temporal pseudo-congruence is found when the spatial pattern and the cladogram topology are coincident, but the cladogenetic events differ in time (Donoghue and Moore 2003; Bortolanza et al. 2006). It is possible that when the data is treated as richness, the temporal pseudo-congruence is ignored. The parameter is cumulative and it may evidence a spatially wide temporal range.

The age of Muscidae was postulated to the Middle-Late Cretaceous (de Carvalho et al. 2003), based on the congruence of species distribution patterns of six Muscidae genera and previous models of evolution of the South American continent under vicariant paradigms (Amorim and Pires 1996). However, it is believed that Muscidae diversification occurred more recently. If the minimum age of its sister group Anthomyiidae (Michelsen 1991) is considered, Muscidae may have appeared until Late Eocene, 40 Ma (Michelsen 2000) with high diversification during Miocene–Oligocene. Additionally, the species Phaonia succini, Phaonia electra (Phaoniinae) and Archaepolietes tertiaria (Muscinae) were found in the Republic Dominican amber from 15 to 25 Ma (Pont and de Carvalho 1997).

Assuming niche conservatism, basal species occupied favourable areas of high temperature that, in the Paleogene, virtually extended to the whole continent. With the progressive cooling occurred after Paleocene–Eocene maximum thermal event, species distributed at high latitude started to be affected by climate change (Hawkins et al. 2006). Evolutionary responses to this event could be extinction, dispersal and diversification. Therefore, in low latitude areas, basal richness of flies can be found where environments today are very similar to past climate conditions, while derived richness was found at higher latitudes.

The Paleogene climatic scenario began with ascendant temperatures upwards to the maximum in Paleocene–Eocene. It was followed by a progressive cooling, until reaching it’s minimal in Oligocene. Based on this picture, it is possible to hypothesize that the establishment of biogeographical components was congruent to this climate history, as an alternative to the first hypothesis (Amorim and Pires 1996). Admitting that Paleogene climate change drove Muscidae pattern of richness and it is spatially congruent to NW and SE dichotomy, it is possible to speculate that the split of NW and SE components occurred at the same epoch. The components may have been established later than Cretaceous (see Paula et al. 2007) as alternative to Late Cretaceous split hypothesis (Amorim and Pires 1996) and probably influenced by Miocene/Late Miocene marine transgression (Donato et al. 2003). Additionally, the Platyrrhini species, which was used in the first hypothesis (Amorim and Pires 1996), no longer corroborate the components split age. Platyrrhini was aged from Eocene with ascendant diversification in Oligocene by molecular clock (Kay et al. 1997; Schneider 2000; Schneider et al. 2001).

In the Andean region, no richness gradient was found. Neither basal nor derived richness arranged an explicit pattern. The difference values (derived minus basal) in the region were about +1, 0 and −1 (Fig. 2). The absence of a richness pattern spatially based on niche conservatism hypothesis evidences that richness was not based on clear evolving elements.

Indeed, the component analysis results evidenced a complex area relationship. Muscidae component analysis showed a polytomy among provinces from different subregions and provinces from the same subregion were grouped as a paraphyletic group. No clear area relationship pattern could be recognized in the region (see Fig. 3). This may suggest that the establishment of biotic components was driven by more than a single vicariant event or that they were blurred by dispersal events. Recent glaciation events in the region may explain the absence of a richness pattern based on niche conservatism as also seen for the cladistic methodology.

Previous studies had produced different hypotheses of components relationship for this region (Morrone 1993; Amorim and Pires 1996; Posadas and Morrone 2001). No common pattern was evidenced when the area relationship established by Muscidae species was compared to previous hypotheses neither among them (Fig. 3). The establishment of the biotic components in the Andean region was very complex and lead to different responses for different organisms (Morrone 1994; Posadas and Morrone 2001; Donato et al. 2003). Despite the cladistic impossibility to resolve the components relationship under a vicariant paradigm in the Andean region, it is believed that the absence of spatially evolved richness structure in the region may be due dispersion and/or extinction events caused by recent glaciations (Soares and de Carvalho 2005a).

Species richness analyses, associated to phylogenetic information, can be compared to cladistic biogeography as for results and methodology. In one hand, species richness assessment evidence whether it was congruent to biotic history or not. It may have advantages on delimiting general patterns where a reticulate and sympatric history took place. Thus, it is impossible to understand a single taxon occupation history from richness data analysis. On the other hand, cladistic biogeography may help in understanding the evolutionary and biotic history elements in species richness gradient establishment. It is believed that even species richness can be linked to deep or recent area history. Geological events may be an important process in species richness input.