Abstract
Speciation rates vary greatly among taxa and regions and are shaped by both biotic and abiotic factors. However, the relative importance and interactions of these factors are not well understood. Here we investigate the potential drivers of speciation rates in South American freshwater fishes, the most diverse continental vertebrate fauna, by examining the roles of multiple biotic and abiotic factors. We integrate a dataset on species geographic distribution, phylogenetic, morphological, climatic, and habitat data. We find that Late Neogene-Quaternary speciation events are strongly associated with body-size evolution, particularly in lineages with small body sizes that inhabit higher elevations near the continental periphery. Conversely, the effects of temperature, area, and diversity-dependence, often thought to facilitate speciation, are negligible. By evaluating multiple factors simultaneously, we demonstrate that habitat characteristics associated with elevation, as well as body size evolution, correlate with rapid speciation in South American freshwater fishes. Our study emphasizes the importance of integrative approaches that consider the interplay of biotic and abiotic factors in generating macroecological patterns of species diversity.
Similar content being viewed by others
Introduction
Speciation, the process of lineage splitting, varies remarkably across the Tree of Life and contributes to uneven species diversity on Earth. Some groups, such as cichlid fishes in Lake Victoria, Africa, radiated into more than 500 species in only 15,000 years, whereas coelacanths in the Indo-West Pacific have produced only two species over approximately 80 million years ago (Mya)1,2. Scientists have long been fascinated by the differences in speciation rates across taxa and among regions3, yet the main drivers of these differences remain hotly debated4.
From a macroevolutionary perspective, speciation arises from the complex interplay of biotic and abiotic factors5,6. Biotic factors include mechanisms associated with morphological evolution and feedback loops resulting in diversity-dependent speciation7,8,9. Morphology is closely related to the use of resources and habitats10. When morphological traits evolve rapidly (i.e., high rates of trait change), speciation rates are expected to increase because of the exploration and partitioning of a broad spectrum of available resources11,12,13. Several studies have reported pulses of morphological evolution, followed by accelerated speciation rates7,8,14. In contrast, other studies demonstrate that morphological differentiation can also be decoupled from speciation15,16. In addition to rates of morphological evolution, diversity-dependent mechanisms posit that the number of species locally co-occurring influences the dynamics of speciation rates9. Species diversity can slow speciation rates due to interspecific competition for resources17 or increase speciation rates due to the greater potential for species interaction18.
Abiotic factors, such as climate and physical habitat characteristics, can also affect speciation rates5,6. Variations in climatic conditions may drive speciation through niche divergence19 (e.g., ecological speciation) and niche conservatism20 (e.g., the barrier of unsuitable climate). Similarly, abiotic characteristics such as habitat volume and elevation can positively influence speciation rates because larger habitats and habitats at higher elevations are likely to be isolated by geographic barriers, thus creating opportunities for allopatric speciation21. In addition, regions with structural habitat heterogeneity (e.g., with more habitat or other resource patchiness) can promote ecological opportunities and trigger divergent adaptations that enhance speciation3. These biotic and abiotic factors are not mutually exclusive—species morphology is likely to evolve as a response to climate and habitat, which might, in turn, promote speciation6—but the interplay between these factors has been poorly studied.
Studies evaluating the relative contributions of biotic and abiotic mechanisms to speciation rates have frequently focused on single predictor variables and overlooked geographic dimensions13,22. Here we evaluate the association of multiple biotic and abiotic mechanisms with the speciation rates of the megadiverse fish fauna of South America (Table 1), which comprises at least 5160 species with remarkable ecological, morphological, and behavioral diversity23. Previous research investigating the factors driving speciation in freshwater fish has primarily relied on lineage-scale studies. More broadly, historical and geological events coincide with major shifts in lineage diversification of freshwater fishes24,25,26. In addition, high rates in the evolution of traits, such as body size7,8 or reproductive mode27, are often associated with rapid species accumulation. However, the spatial dynamics of speciation rates in this fauna are poorly understood, particularly the extent to which species traits and abiotic factors interact to influence speciation in different regions and clades.
We aim to understand the extent to which biotic and abiotic factors influence speciation rates across geographic regions. Our study concentrates on recent speciation events primarily taking place during the Late Neogene-Quaternary (mean species age of 7.3 Mya). This period holds paramount importance because it coincides with a better understanding of the paleoclimates and paleoenvironments of northern South America28, 29, while also experiencing generally low extinction rates for most taxa of South American freshwater fishes23. We analyze speciation rates and eleven predictors in 460 drainage sub-basins, encompassing 2638 fish species. Here, we show how the evolution of several morphological traits, species diversity (diversity-dependence), climate, and characteristics of the physical habitat interact with spatial variation in speciation rates.
Results
Freshwater fish speciation across phylogeny and geography
Fish speciation in South America steadily increased over time, with a particularly notable uptick starting in the late Cretaceous period (Supplementary Fig. 1). Specifically, speciation rates varied over 18-fold across the tips of the phylogeny, with the highest rates being concentrated in certain groups such as the cyprinodontiform Orestias, the characiform Serrasalmus, and the siluriforms Hemiancistrus, Pterygoplichthys, and Hypostomus (Fig. 1a).
When mapped onto geography, recent speciation rates varied >17-fold among sub-basins (Fig. 1b). Lineages with the highest recent speciation rates were found in regions of the Altiplano and Pacific slope of Peru, mesic areas in the Atlantic Forest, and some seasonally dry upland regions (e.g., dry diagonal) (Fig. 1b). In contrast, the Southern Cone, Northeast Atlantic slope of Brazil, and some parts of the highly diverse Central and Eastern Amazon lowlands had the lowest speciation rates (Fig. 1b). We did not find any clear latitudinal (Pearson’s r = 0.161, p < 0.001) or longitudinal trends (r = 0.108, p = 0.020) in speciation rates across South America.
Correlates of spatial variation in speciation rates
Multiple regression analysis, including both biotic and abiotic variables, such as rates of morphological evolution, species diversity, climate, and habitat, explained 90% of the spatial variation in speciation rates (Fig. 1b, d; Fig. 2a; Supplementary Table 1).
Rates of morphological evolution accounted for 69% of the total variance in speciation rates across sub-basins (Fig. 2b). Among the various morphological traits examined, the rate of maximum body length evolution emerged as the most influential factor explaining spatial variation in speciation rates (R2 = 0.470; Fig. 1e and Fig. 2). This suggests that higher rates of speciation are associated with accelerated body-size evolution (Fig. 3a and Fig. 4a; Supplementary Fig. 2). Evolutionary rates of oral gape position and relative maxillary length exhibited a negative relationship with speciation rates (R2 = 0.137 and R2 = 0.031, respectively; Fig. 2 and Fig. 3b, c). The rate of body elongation evolution had no significant association with speciation rates (Fig. 2a). These findings are supported by rate-through-time plots, which reveal an acceleration of body size evolution over time, notably around 50 Mya, coinciding with an increase in speciation rates. In contrast, rates for oral gape position and relative maxillary length remained relatively constant or decreased during this period (Supplementary Fig. 3). Additionally, we observed a weak but significant, positive relationship between speciation rates and species diversity (R2 = 0.004; Fig. 2 and Fig. 3d).
Our model also uncovered the considerable importance of abiotic factors on speciation rates (R2 = 0.202; Fig. 2; Supplementary Table 1). The strongest contribution was from habitat-related variables (R2 = 0.172; Fig. 2b), with speciation rates increasing with elevation (second-highest effect size, Fig. 3f and Fig. 4b) but decreasing with soil diversity (Fig. 3g) and stream gradient (Fig. 3h). Contrary to expectations, sub-basins with larger areas did not exhibit higher speciation rates (Fig. 2a). The overall climate-related effects were weak and mainly represented by temperature (R2 = 0.025; Fig. 2), which showed a positive correlation with speciation rates (Fig. 3e). Speciation rates were not correlated with land surface runoff (Fig. 2a).
To examine the unique and joint effects of morphological evolution, diversity-dependence, climate, and habitat on speciation, we performed a variance partitioning analysis (Fig. 5). The results showed that the joint effect of biotic and abiotic factors explained most of the variation in speciation rates (Fig. 5a). Specifically, morphological evolution (biotic) uniquely explained 36% of the variance in speciation rates. However, most of the variation in speciation (52%) could be jointly explained by morphological evolution and habitat variables (Fig. 5b). All other portions of the variation were consistently weak, accounting for less than 6%.
To ensure the robustness of our findings, we addressed a range of technical and methodological artifacts (see Supplementary Note 1: Sensitivity Analyses for a comprehensive discussion). These sensitivity analyses primarily considered uncertainties in estimating rates of speciation and morphological evolution (Supplementary Figs. 4–13), the influence of paleoclimate on speciation rates (Supplementary Figs. 14 and 15), and the potential impact of biological outliers on our results (Supplementary Figs. 16–19). By considering these potential limitations, we can confidently assert that our findings remain robust. They emphasize the importance of body size evolution, particularly its accelerated pace in upland areas, as a key factor in explaining the observed increase in speciation rates.
Discussion
We evaluated the association of biotic and abiotic factors with spatial variation of recent speciation rates in South American freshwater fishes—the most species-rich continental vertebrate fauna. We found that the rates of morphological evolution, notably body size, are the primary correlates of spatial variation in speciation rates among fishes (Fig. 1e and Fig. 2). We also found that abiotic factors, particularly elevation, appeared to play a role in driving speciation rates, where higher-elevation lineages speciate at faster rates. Further, these two variables were coupled, such that regions in which fish underwent faster body size evolution and speciation are commonly found at high elevations (Fig. 4). These results are robust, as demonstrated by their consistent performance when tested against several alternative methods and sensitivity analyses (Supplementary Note 1). Taken together, these findings suggest that habitat characteristics can promote speciation by affecting how traits evolve. Therefore, to better understand the origins of present-day biodiversity patterns, it is essential to consider multiple types of traits, their rates of evolution, and the coupled effects with climate and habitat conditions.
The evolutionary rate of one key morphological trait, maximum body length, emerged as the strongest factor associated with differences in speciation rates across regions (Fig. 2). The sub-basins that concentrate species with the fastest speciation rates, such as the Altiplano and Pacific slope of Peru and mesic sub-basins in the Atlantic Forest, also showed an exceptional rate of change in body size (Fig. 1b and Fig. 4a). Particularly, this most recent rapid speciation occurred in clades of small-bodied fishes, such as Orestias (Cyprinodontiformes) restricted to high-elevation (>1000 m.a.s.l) lakes and rivers in the Andes30, and Hypostomus (Siluriformes) principally distributed in rivers draining the Brazilian Shield and La Plata basin31. While our goal was to evaluate spatial variation in potential drivers of speciation, we also confirmed that a similar outcome unfolded over geological timescales (for more comprehensive coverage, see25). For instance, we found that during the Paleogene period (66 Mya), an acceleration in speciation rates coincided with an increase in body size evolution, while rates of other traits declined (Supplementary Fig. 3). These findings provide evidence supporting the Paleogene radiations32. Following the Cretaceous-Paleogene mass extinction32, new species emerged to occupy newly available ecological niches33, with high-elevation sub-basins potentially playing a crucial role as a speciation hotspot for the modern ichthyofauna during the Neogene23.
For many vertebrates7,8 and plants34, a positive relationship exists between body size evolution and speciation rates. There are several possible explanations for this positive relationship. One possibility is that the evolution of morphological novelty (morphological evolvability) triggers speciation by allowing a clade to diversify and occupy spaces previously inaccessible (adaptive radiation theory11). Another possibility is that the speciation process itself leads to rapid morphological change (punctuated equilibrium theory35), in which clade evolution is characterized by long periods of evolutionary stasis followed by rapid bursts of morphological change around the time of speciation36. Many examples of rapid adaptive radiation are recognized within highly speciose communities37, such as the close association between body size evolution and elevation identified here. While it is difficult to distinguish between adaptive radiation and punctuated equilibrium theories36, morphological evolvability emerges as a more plausible and direct explanation for the increased speciation rates.
Rates of speciation and morphological evolution can also be uncoupled15,16 or even negative38,39. The results presented here provide additional evidence to support this possibility, as we found a negative association between speciation rates and rates of oral gape position and relative maxillary length evolution (Fig. 2a and Fig. 3b, c). These findings suggest that speciation is not related to high rates of change in traits associated with feeding behavior, echoing previous studies in snails40, damselfly41, lizards42, salamanders43, and birds44. The inverted relationship between rates of speciation and morphological evolution likely emerged in a non-adaptive scenario, whereby bursts of species formation occur in groups with minimal morphological disparity45. These findings suggest that different traits may vary in importance in speciation13, and considering only one axis of morphological variation, such as body size, may obscure important aspects of the relationship between speciation and morphological evolution8. Overall, we argue that considering multiple axes of morphological variation is essential for gaining a complete understanding of the mechanisms driving speciation.
Abiotic factors additionally influenced recent speciation rates, with habitat characteristics explaining more spatial variation in speciation than climate (Fig. 2 and Fig. 5b). Elevation, the best predictor among the abiotic factors, indicates that higher-elevation lineages speciate at faster rates (Fig. 2 and Fig. 4b). There are many possible explanations for why elevation can affect speciation rates6. Areas formed during active uplift have more geographic barriers, physiographic heterogeneity, and new habitats6,46, all of which can lead to reproductive isolation3,47 and ultimately spur speciation. Our results are consistent with many cross-taxa observations that demonstrate faster speciation in mountains46,48, although the link between speciation and morphological evolution along elevational gradients remains largely unexplored. We observed that body size evolves faster in high elevations (Fig. 4c), which in turn correlates with a higher speciation rate. Body size is a fine-tuned trait directly linked to nearly every aspect of organismal biology (e.g., habitat use, life history, and metabolism10), and changes in body size are associated with ecological differentiation and reproductive isolation10. Therefore, environmental conditions along elevation gradients are likely to have imposed evolutionary pressures on body size that are tightly coupled with speciation rates.
Other factors played little or negligible roles in the rates of fish speciation across South American sub-basins. Proxies of habitat conditions, namely stream gradient, and diversity of soil types, had a weak negative relationship with speciation rates (Fig. 2 and Fig. 3g, h), indicating that the highest speciation rates occur in regions with lower soil diversity than Amazonian lowlands and lower gradients. These patterns are atypical, as they are contrary to the usual increase in speciation rates with greater environmental heterogeneity and a number of available niches49. The area was found to be unimportant, suggesting that increases in sub-basin size are not sufficient to drive a significant increase in speciation rates (Fig. 2). The effects of species diversity were also minimal (Fig. 2), although we did find higher speciation rates in areas with higher species richness. Greater species diversity itself can promote speciation by increasing the probability of divergence between populations, structural complexity, or through a trade-off between competition and dispersal9,18 (but see17 for a contrary view). Finally, we found that temperature was positively related to speciation rates (Fig. 3e). These results support literature stating that warmer and productive regions increase demographic rates, metabolic rates, and evolutionary rates50, yet a recent wave of studies has shown the fastest rates of speciation in cold regions51.
Speciation rates vary by several orders of magnitude across regions51, but a comprehensive view of their causes and consequences continues to puzzle scientists4. To date, much of what we know about drivers of speciation comes from systems at smaller spatial and phylogenetic scales than our study or does not evaluate multiple factors simultaneously. Our study is notable for its scale, incorporating a vast dataset of the most diverse vertebrate groups worldwide and considering multiple potential drivers of spatial patterns in speciation. We revealed that the rate of morphological evolution is the strongest predictor of geographic variation in recent speciation rates, supporting a positive relationship between rates of body size change and speciation rates7,8. Our findings also support the idea that changes in body size are likely to be accelerated in upland regions, leading to an increase in speciation rates of living fishes. Moreover, we found that elevation has a more significant relationship with rapid speciation than do climate, area, and diversity-dependent mechanisms. By evaluating a baseline set of biotic vs. abiotic factors, our work suggests a synergy between morphological evolution and habitat characteristics in shaping large-scale patterns in speciation.
Methods
Geographic occurrences and phylogenetic data
We obtained presence/absence data for 4967 South American species across 460 sub-basins25 delimited by the HydroBASINS framework (level 552). This dataset is based on an extensive survey of web repositories (e.g., GBIF, specieslink) and literature sources25, with several procedures to increase data quality (e.g., removal of georeferencing errors, exotic and migratory species). In addition, simulation approaches were performed to reduce sampling effort heterogeneities in species distribution25 by calculating a completeness index in each drainage basin based on the probability that the next sampled record would add a new species to those observed in the focal basin.
We used a newly compiled, time-calibrated tree of Neotropical Freshwater Fishes25. The tree was built on 5984 terminal taxa, including 3169 species with available genetic data (51 independently aligned and trimmed markers), 31 fossil-constrained nodes, and 2815 species inserted by taxonomic imputation. This tree summarizes the current knowledge on phylogenetic relationships among Neotropical freshwater fishes.
Estimating speciation rates and assessing their reliability
To explore diversification patterns among lineages, we used Bayesian Analysis of Macroevolutionary Mixtures53 (BAMM), diversification rate (DR) statistic54, and Missing State Speciation and Extinction (MiSSE) metric55. We employed species-specific diversification rate (tip-based estimates), which are more accurate than deep-time estimates55,56. BAMM allows the modeling of complex speciation and extinction dynamics by detecting major shifts in evolutionary rates from a time-calibrated phylogeny53,57. We ran 20 million generations of Markov Chain Monte Carlo (MCMC) over four chains, sampling every 2000 generations and accounting for incomplete taxon sampling (global sampling fraction of 0.53). We discarded the first 25% of the MCMC samples as burn-in and ensured that the effective sample size of all parameters was above 200 (CODA R package58). The prior settings for speciation and extinction were obtained by the setBAMMpriors function in the BAMMtools R package59. The DR statistic is calculated as the inverse equal splits rates (i.e., the sum of the branch lengths separating a tip from the root) and has been shown to reflect speciation rates better than net diversification rate54,57. Finally, MiSSE is a trait-free version of the HiSSE framework, designed to accurately estimate tip diversification rate while accommodating a broad range of speciation and extinction scenarios55.
While our model estimated both speciation and extinction rates, we focused on reporting the speciation estimates, as inferring extinction rates from extant phylogenies is known to be biologically unrealistic.60,61. Therefore, we extracted tip-based (recent time) speciation rates from BAMM, DR, and MiSSE and explored spatial patterns by averaging these rates among co-occurring species in each subdrainage basin. While the use of mean values is a common practice, we recognize that it may not always be suitable, especially when analyzing communities with high levels of variation in the variable of interest. To address this issue, we calculated the median, minimum, and maximum values of speciation rates across sub-basins and assessed the impact of predictor variables on all of these summary statistics. Our analysis revealed that spatial patterns of speciation were largely consistent when using mean, median, minimum, or maximum values (Supplementary Fig. 20). This was particularly evident for regions with lower (e.g., Amazon basin) and higher (e.g., upland sub-basins) rates of speciation. We also found that the importance of the predictors remained qualitatively similar across different summary statistics (Supplementary Fig. 21), further highlighting the major role of body size evolution in driving speciation.
We found that speciation rates estimated from alternative methods were strongly correlated (Supplementary Fig. 4), indicating that these methods are likely to produce consistent results (Supplementary Fig. 5). Therefore, we focus further analyses on BAMM speciation rates (see8,57 for method comparison) and the results using DR and MiSSE estimates are available in the supplementary material. BAMM speciation estimates were highly correlated with net diversification rate (BAMM speciation minus BAMM extinction) (Supplementary Fig. 6), suggesting that our findings are likely to hold for diversification rate as well. We used the most up-to-date and comprehensive reconstruction of the evolutionary relationships among South American fish species25. Nonetheless, alternative phylogenies exist, and BAMM speciation estimates could vary across phylogenies and affect our interpretation (Supplementary Note 1). We, therefore, compared our estimates with (i) the corresponding tree based solely on genetic data25, (ii) a ray-finned fish supertree62, or (iii) taxon-specific trees such as those for poecillidae63, cichlid26, and characoid24 fishes. We obtained similar BAMM speciation estimates across phylogenies (Supplementary Fig. 7).
Morphological traits, their rates of evolution, and robustness
We selected five morphological traits from the Fishmorph database64 to capture relevant ecological, physiological, and behavioral dimensions of fish species: body elongation, relative eye size, oral gape position, relative maxillary length, and maximum body length. Body elongation, functionally related to swimming performance and habitat utilization, has been identified as a dominant axis of phenotypic disparity in many clades65,66. Relative eye size is related to visual acuity and diel activity patterns67,68. Oral gape position and relative maxillary length are related to feeding behaviors and trophic position69,70. Maximum body length has been correlated to many important evolutionary processes, like metabolic rate, mutation rate, generation time, organismal vagility, species geographic range, and population structure71,72. Body size estimates were log10-transformed prior to estimating evolutionary rates, following the suggestion for size data73.
We pruned the phylogeny of Neotropical freshwater fishes25 to match the South American species with morphological data, resulting in 2638 species with phylogenetic information, geographic occurrences, and morphological data (53% of the South American ichthyofauna). Species missing morphological data are not phylogenetically clustered (Supplementary Figs. 22-26; tested with miss.phylo.d function74), therefore, our results do not over or under-represent any particular fish clade. As trait values were unknown for a small number of species in the phylogeny (0.6–10% among the five traits; Supplementary Figs. 22–26), we imputed missing trait data using the Rphylopars function75. This approach uses phylogenetic information for imputation criteria75 and has been shown to outperform other approaches76.
Posteriorly, we estimated per-lineage rates of morphological evolution of the five traits by using the BAMM ‘trait’ module59. We ran three independent MCMC chains with 100 million generations each, sampling every 10,000 generations and removing the first 25% of samples as burn-in. We ensured that the effective sample size was above 200. Based on tip-based evolutionary rates, we calculated the mean rates of morphological evolution for each sub-basin using their corresponding species assemblages (Supplementary Figs. 27–31).
We further compared the trait evolution estimates from the BAMM trait model with another robust method, known as the BayesTrait77 metric (settings are described in Supplementary Note 2: Extended Methods). Our analyses revealed no evidence of bias in the patterns of morphological evolution that could be attributed to the choice of method, as the estimates of morphological evolution obtained using both BAMM and BayesTraits methods exhibited strong consistency (Supplementary Figs. 8–12). As the proportion of sampled species was unequal among genera (Supplementary Table 2) and an incomplete taxonomic sampling may influence estimates of evolutionary rates78, we also verified the sensitivity of our estimates to a higher species coverage. More precisely, we retrieved data from Fishbase79 on maximum body length and reanalyzed the correlation between the rate of speciation and body size evolution with 4228 species (85% of the South American ichthyofauna). We showed that an increased taxonomic sampling produced spatial patterns of speciation and body size evolution consistent with our pruned data set (Supplementary Fig. 13).
Species diversity
We determined species diversity as the total number of fish species co-occurring in each subdrainage basin (Supplementary Fig. 32). Considering species richness and area as predictors in the model assumes a slope of the species-area regression equal to one, although this scaling exponent has been shown to empirically range from 0.25-0.5080. To address this issue, we reanalyzed the patterns of species richness in the context of species-area relationship80, following the power function: SD = SR/Az, where SD is the species density (i.e., number of species per unit area), SR is the number of species in area A, and z is the species-area scaling exponent (i.e., the slope of the species-area regression). Overall, we found that species density and area predicted speciation rates similarly to when using raw species richness (Supplementary Figs. 33 and 34), and, therefore, we provide these results as supplementary.
We conducted additional analyses to assess the potential impact of the number of species and wide-ranging species on our conclusions by: (i) excluding all sub-basins with less than 10, 15, and 20 species and (ii) applying a weighting approach to account for the disproportionate influence of wide-ranging species on the mean estimates across sub-basins. The similarity of results from these analyses indicates that our findings are robust (Supplementary Figs. 35 and 36), unaffected by potential biases related to the number of species and the inclusion of wide-ranging species.
Climate variables
We obtained data on four variables directly related to climate: (i) annual mean temperature, (ii) annual mean precipitation, (iii) actual evapotranspiration, and (iv) land surface runoff. Annual mean temperature and precipitation were taken from the WorldClim Version 1 database at 2.5 arc-min resolution81 (Supplementary Table 3). Air temperature was used because it is strongly associated with water temperature82, which is not available for most of the sub-basins. Data on actual evapotranspiration and land surface runoff were extracted from the HydroATLAS database at 15 arc-s resolution83. We calculated the mean values of all the variables for each of the 460 sub-basins (Supplementary Fig. 37).
As past climate may leave imprints on present-day speciation rates, we examined the potential effect that past temperature had on speciation rates. Past temperature data corresponding to the Pliocene (ca. 3.3 My) and Last Glacial Maximum (LGM; ca. 21 ka) were extracted from the CHELSA database (at 2.5 arc-min resolution)84. Overall, past temperature explained speciation rates in a similar fashion to contemporary temperature and was treated as supplementary (Supplementary Figs. 14). We also explored the paleoclimate effects by employing a temperature-dependent speciation model85. This model incorporated temperature variation through the Cenozoic epoch86 (~67 Mya) and only clades with more than 50 species originating within this time frame (Supplementary Note 2). The results of this analysis reaffirmed our findings, indicating that 73% of the examined clades demonstrated a positive effect of temperature on speciation (Supplementary Fig. 15).
Habitat variables
To represent the habitat, we considered: (i) mean elevation, (ii) geographic area of the sub-basin (measured in square meters), (iii) stream gradient, and (iv) soil diversity. We also tested the impact of alternative measures related to elevation (i.e., topographic complexity), which included standard deviation in elevation, terrain slope, and basin relief. Importantly, mean elevation was a better predictor of speciation rates, and therefore, we considered topographic complexity measures as supplementary (Supplementary Fig. 38).
Data on elevation were taken from the WorldClim Version 1 database at 2.5 arc-min resolution81. The geographic area of the sub-basin was obtained from HydroBASINS (level 552) and stream gradient and soil types from HydroATLAS at 15 arc-second resolution83 (Supplementary Table 3). Soil diversity was measured using the Shannon diversity index based on substrate types and soil conditions within each sub-basin. We calculated the mean values of all the variables for each of the 460 sub-basins (Supplementary Fig. 39).
Statistical analysis
To examine the connections between geographic variation in recent speciation rates and various factors, including biotic factors (such as rates of trait evolution and species diversity) as well as abiotic factors (such as climate and habitat variables), we performed a multiple linear regression analysis. Prior to the regression analysis, we assessed multicollinearity among predictors by examining the variance inflation factor87 (VIF). Three variables with a high VIF were removed from the analyses (rate of evolution of relative eye size, annual precipitation, and actual evapotranspiration; Supplementary Fig. 40). In this way; we retained eleven variables with low multicollinearity (VIF < 5; Supplementary Table 1). To make the effect size comparable among predictors, we standardized response and predictor variables using z-score standardization (mean = 0; sd = 1), with the sub-basin area being log-transformed before z-score standardization.
To quantify the relative contribution of each predictor in explaining the total variance in speciation rates, we performed hierarchical partitioning (metric lmg) as implemented in the R package relaimpo88. This procedure decomposes the model-explained variance into non-negative contributions and evaluates the relative importance of each predictor variable in linear models88. We also used variance partitioning to investigate the unique and shared effects of four major mechanisms as predictors of speciation rates. The predictor variables were consistent with those used in the linear multiple regression and were grouped into (1) rates of morphological evolution, (2) diversity-dependent speciation, (3) climate-driven speciation, and (4) habitat-driven speciation (Table 1).
We further investigated the sensitivity of our results to spatial autocorrelation in the residuals and outliers. To examine spatial autocorrelation, we computed Moran’s I89 based on residual speciation rates (i.e., the residuals from a linear multiple regression of the 11 predictors on speciation rates). We found low spatial autocorrelation in the residuals from the non-spatial, multiple regression (Global Moran’s I = −0.003, P = 0.423; Supplementary Fig. 41). Finally, the genus Orestias (Cyprinodontiformes) is endemic to the Andean basin of Peru, Bolivia, and Chile and had a recent and localized burst in diversification30, which was also recovered in our speciation estimates (Fig. 1a). To confirm that our primary conclusions were not influenced by Orestias species, we (i) ran the BAMM model after removing all Orestias species (ii) excluded Orestias species from the sub-basins where they are present, and (iii) removed five sub-basins predominantly occupied by Orestias species and with exceptional rates of speciation. The removal scenarios involving Orestias had no impact on our ability to detect high rates within other clades (Supplementary Fig. 16) or introduce bias to the mean speciation rates across sub-basins (Supplementary Figs. 16–19).
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Fish species occurrence records and phylogeny are from Cassemiro et al.25. Morphological traits are available at FISHMORPH database (Brosse et al.64). Climate and habitat data are from WorldClim (www.worldclim.org) and HydroATLAS database (www.hydrosheds.org/hydroatlas). All data supporting the findings of this study are available in the Zenodo database under Attribution 4.0 International license and accession code: https://doi.org/10.5281/zenodo.8301082. Source data are provided in this paper.
Code availability
The R codes employed for the analyses undertaken in this study can be accessed in the Zenodo database under the Attribution 4.0 International license. The corresponding accession code is https://doi.org/10.5281/zenodo.8301082.
References
Seehausen, O. et al. Nuclear markers reveal unexpected genetic variation and a Congolese-Nilotic origin of the Lake Victoria cichlid species flock. Proc. R. Soc. Lond. B Biol. Sci. 270, 129–137 (2003).
Amemiya, C. T. et al. The African coelacanth genome provides insights into tetrapod evolution. Nature 496, 311–316 (2013).
Turelli, M., Barton, N. H. & Coyne, J. A. Theory and speciation. Trends Ecol. Evol. 16, 330–343 (2001).
Hernández-Hernández, T., Miller, E. C., Román-Palacios, C. & Wiens, J. J. Speciation across the Tree of Life. Biol. Rev. 96, 1205–1242 (2021).
Mittelbach, G. G. et al. Evolution and the latitudinal diversity gradient: speciation, extinction and biogeography. Ecol. Lett. 10, 315–331 (2007).
Schluter, D. & Pennell, M. W. Speciation gradients and the distribution of biodiversity. Nature 546, 48–55 (2017).
Rabosky, D. L. et al. Rates of speciation and morphological evolution are correlated across the largest vertebrate radiation. Nat. Commun. 4, 1958 (2013).
Cooney, C. R. & Thomas, G. H. Heterogeneous relationships between rates of speciation and body size evolution across vertebrate clades. Nat. Ecol. Evol. 5, 101–110 (2020).
Rabosky, D. L. Diversity-dependence, ecological speciation, and the role of competition in macroevolution. Annu. Rev. Ecol. Evol. Syst. 44, 481–502 (2013).
LaBarbera, M. Analyzing body size as a factor in ecology and evolution. Annu. Rev. Ecol. Syst. 20, 97–117 (1989).
Losos, J. B. Adaptive radiation, ecological opportunity, and evolutionary determinism: American Society of Naturalists E. O. Wilson Award Address. Am. Nat. 175, 623–639 (2010).
Crouch, N. M. A. & Tobias, J. A. The causes and ecological context of rapid morphological evolution in birds. Ecol. Lett. 25, 611–623 (2022).
Wiens, J. J. What explains patterns of biodiversity across the Tree of Life? BioEssays 39, 1600128 (2017).
Price-Waldman, R. M., Shultz, A. J. & Burns, K. J. Speciation rates are correlated with changes in plumage color complexity in the largest family of songbirds. Evolution 74, 1155–1169 (2020).
Adams, D. C., Berns, C. M., Kozak, K. H. & Wiens, J. J. Are rates of species diversification correlated with rates of morphological evolution? Proc. R. Soc. B Biol. Sci. 276, 2729–2738 (2009).
Crouch, N. M. A. & Ricklefs, R. E. Speciation rate is independent of the rate of evolution of morphological size, shape, and absolute morphological specialization in a large clade of birds. Am. Nat. 193, E78–E91 (2019).
Harvey, M. G. et al. The evolution of a tropical biodiversity hotspot. Science 370, 1343–1348 (2020).
Emerson, B. C. & Kolm, N. Species diversity can drive speciation. Nature 434, 1015–1017 (2005).
Schluter, D. Evidence for ecological speciation and its alternative. Science 323, 737–741 (2009).
Wiens, J. J. Speciation and ecology revisited: phylogenetic niche conservatism and the origin of species. Evolution 58, 193–197 (2004).
Chown, S. L. & Gaston, K. J. Areas, cradles and museums: the latitudinal gradient in species richness. Trends Ecol. Evol. 15, 5 (2000).
Barraclough, T. G. & Nee, S. Phylogenetics and speciation. Trends Ecol. Evol. 16, 391–399 (2001).
Albert, J. S., Tagliacollo, V. A. & Dagosta, F. Diversification of neotropical freshwater fishes. Annu. Rev. Ecol. Evol. Syst. 51, 27–53 (2020).
Melo, B. F. et al. Accelerated diversification explains the exceptional species richness of tropical characoid fishes. Syst. Biol. 71, 78–92 (2022).
Cassemiro, F. A. S. et al. Landscape dynamics and diversification of the megadiverse South American freshwater fish fauna. Proc. Natl Acad. Sci. USA 120, e2211974120 (2023).
Burress, E. D. & Tan, M. Ecological opportunity alters the timing and shape of adaptive radiation. Evolution 71, 2650–2660 (2017).
Helmstetter, A. J. et al. Viviparity stimulates diversification in an order of fish. Nat. Commun. 7, 11271 (2016).
Hoorn, C. et al. Amazonia through time: andean uplift, climate change, landscape evolution, and biodiversity. Science 330, 927–931 (2010).
Albert, J. S., Val, P. & Hoorn, C. The changing course of the Amazon River in the Neogene: center stage for Neotropical diversification. Neotrop. Ichthyol. 16, e180033 (2018).
Guerrero-Jiménez, C. J. et al. Pattern of genetic differentiation of an incipient speciation process: the case of the high Andean killifish Orestias. PLoS ONE 12, e0170380 (2017).
Cardoso, Y. P., Jardim de Queiroz, L., Bahechar, I. A., Posadas, P. E. & Montoya-Burgos, J. I. Multilocus phylogeny and historical biogeography of Hypostomus shed light on the processes of fish diversification in La Plata Basin. Sci. Rep. 11, 5073 (2021).
López-Fernández, H. & Albert, J. S. Paleogene Radiations. in Historical Biogeography of Neotropical Freshwater Fishes (ed Albert, J.) 0 (University of California Press). https://doi.org/10.1525/california/9780520268685.003.0006 2011.
Sibert, E. C. & Norris, R. D. New Age of Fishes initiated by the Cretaceous−Paleogene mass extinction. Proc. Natl Acad. Sci. USA 112, 8537–8542 (2015).
Ramírez-Barahona, S. & Sauquet, H. & Magallón, S. The delayed and geographically heterogeneous diversification of flowering plant families. Nat. Ecol. Evol. 1–7 (2020).
Gould, S. J. & Eldredge, N. Punctuated equilibria: the tempo and mode of evolution reconsidered. Paleobiology 3, 115–151 (1977).
Pennell, M. W., Harmon, L. J. & Uyeda, J. C. Is there room for punctuated equilibrium in macroevolution? Trends Ecol. Evol. 29, 23–32 (2014).
Martin, C. H. & Richards, E. J. The paradox behind the pattern of rapid adaptive radiation: how can the speciation process sustain itself through an early burst? Annu. Rev. Ecol. Evol. Syst. 50, 569–593 (2019).
Folk, R. A. et al. Rates of niche and phenotype evolution lag behind diversification in a temperate radiation. Proc. Natl Acad. Sci. USA 116, 10874–10882 (2019).
Lee, M. S. Y., Sanders, K. L., King, B. & Palci, A. Diversification rates and phenotypic evolution in venomous snakes (Elapidae). R. Soc. Open Sci. 3, 150277 (2016).
Holland, B. S. & Hadfield, M. G. Origin and diversification of the endemic Hawaiian tree snails (Achatinellidae: Achatinellinae) based on molecular evidence. Mol. Phylogenet. Evol. 32, 588–600 (2004).
Wellenreuther, M. & Sánchez-Guillén, R. A. Nonadaptive radiation in damselflies. Evol. Appl. 9, 103–118 (2016).
Hipsley, C. A., Miles, D. B. & Müller, J. Morphological disparity opposes latitudinal diversity gradient in lacertid lizards. Biol. Lett. 10, 20140101 (2014).
Kozak, K. H., Weisrock, D. W. & Larson, A. Rapid lineage accumulation in a non-adaptive radiation: phylogenetic analysis of diversification rates in eastern North American woodland salamanders (Plethodontidae: Plethodon). Proc. R. Soc. B Biol. Sci. 273, 539–546 (2006).
Uy, J. A. C., Moyle, R. G. & Filardi, C. E. Plumage and song differences mediate species recognition between incipient flycatcher species of the solomon islands. Evolution 63, 153–164 (2009).
Rundell, R. J. & Price, T. D. Adaptive radiation, nonadaptive radiation, ecological speciation and nonecological speciation. Trends Ecol. Evol. 24, 394–399 (2009).
Rahbek, C. et al. Building mountain biodiversity: geological and evolutionary processes. Science 365, 1114–1119 (2019).
Butlin, R. et al. What do we need to know about speciation? Trends Ecol. Evol. 27, 27–39 (2012).
Steinbauer, M. J. et al. Topography-driven isolation, speciation and a global increase of endemism with elevation. Glob. Ecol. Biogeogr. 25, 1097–1107 (2016).
Golestani, A., Gras, R. & Cristescu, M. Speciation with gene flow in a heterogeneous virtual world: can physical obstacles accelerate speciation? Proc. R. Soc. B Biol. Sci. 279, 3055–3064 (2012).
Allen, A. P., Gillooly, J. F., Savage, V. M. & Brown, J. H. Kinetic effects of temperature on rates of genetic divergence and speciation. Proc. Natl Acad. Sci. USA 103, 9130–9135 (2006).
Morlon, H. Diversity hotspots: coldspots of speciation? Science 370, 1268–1269 (2020).
Lehner, B. & Grill, G. Global river hydrography and network routing: baseline data and new approaches to study the world’s large river systems. Hydrol. Process. 27, 2171–2186 (2013).
Rabosky, D. L. Automatic detection of key innovations, rate shifts, and diversity-dependence on phylogenetic trees. PLoS ONE 9, e89543 (2014).
Jetz, W., Thomas, G. H., Joy, J. B., Hartmann, K. & Mooers, A. O. The global diversity of birds in space and time. Nature 491, 444–448 (2012).
Vasconcelos, T., O’Meara, B. C. & Beaulieu, J. M. A flexible method for estimating tip diversification rates across a range of speciation and extinction scenarios. Evolution 76, 1420–1433 (2022).
Siqueira, A. C., Morais, R. A., Bellwood, D. R. & Cowman, P. F. Trophic innovations fuel reef fish diversification. Nat. Commun. 11, 2669 (2020).
Title, P. O. & Rabosky, D. L. Tip rates, phylogenies and diversification: what are we estimating, and how good are the estimates? Methods Ecol. Evol. 10, 821–834 (2019).
Plummer, M., Best, N., Cowles, K. & Vines, K. CODA: convergence diagnosis and output analysis for MCMC. R. N. 6, 7–11 (2006).
Rabosky, D. L. et al. BAMMtools: an R package for the analysis of evolutionary dynamics on phylogenetic trees. Methods Ecol. Evol. 5, 701–707 (2014).
Rabosky, D. L. Extinction rates should not be estimated from molecular phylogenies. Evolution 64, 1816–1824 (2010).
Mitchell, J. S., Etienne, R. S. & Rabosky, D. L. Inferring diversification rate variation from phylogenies with fossils. Syst. Biol. 68, 1–18 (2019).
Rabosky, D. L. et al. An inverse latitudinal gradient in speciation rate for marine fishes. Nature 559, 392 (2018).
Reznick, D. N., Furness, A. I., Meredith, R. W. & Springer, M. S. The origin and biogeographic diversification of fishes in the family Poeciliidae. PLoS ONE 12, e0172546 (2017).
Brosse, S. et al. FISHMORPH: a global database on morphological traits of freshwater fishes. Glob. Ecol. Biogeogr. 30, 2330–2336 (2021).
Claverie, T. & Wainwright, P. C. A morphospace for reef fishes: elongation is the dominant axis of body shape evolution. PLoS ONE 9, e112732 (2014).
Toussaint, A. et al. Non-native species led to marked shifts in functional diversity of the world freshwater fish faunas. Ecol. Lett. 21, 1649–1659 (2018).
Caves, E. M., Sutton, T. T. & Johnsen, S. Visual acuity in ray-finned fishes correlates with eye size and habitat. J. Exp. Biol. 220, 1586–1596 (2017).
Corral-López, A., Garate-Olaizola, M., Buechel, S. D., Kolm, N. & Kotrschal, A. On the role of body size, brain size, and eye size in visual acuity. Behav. Ecol. Sociobiol. 71, 179 (2017).
Luiz, O. J. et al. Does a bigger mouth make you fatter? Linking intraspecific gape variability to body condition of a tropical predatory fish. Oecologia 191, 579–585 (2019).
Kopf, R. K., Yen, J. D. L., Nimmo, D. G., Brosse, S. & Villéger, S. Global patterns and predictors of trophic position, body size and jaw size in fishes. Glob. Ecol. Biogeogr. 30, 414–428 (2021).
Knouft, J. H. & Page, L. M. The evolution of body size in extant groups of North American freshwater fishes: speciation, size distributions, and cope’s rule. Am. Nat. 161, 413–421 (2003).
Albert, J. S. & Johnson, D. M. Diversity and evolution of body size in fishes. Evol. Biol. 39, 324–340 (2012).
Benson, R. B. J., Godoy, P., Bronzati, M., Butler, R. J. & Gearty, W. Reconstructed evolutionary patterns for crocodile-line archosaurs demonstrate impact of failure to log-transform body size data. Commun. Biol. 5, 1–4 (2022).
Paterno, G. B., Penone, C. & Werner, G. D. A. sensiPhy: an r-package for sensitivity analysis in phylogenetic comparative methods. Methods Ecol. Evol. 9, 1461–1467 (2018).
Goolsby, E. W., Bruggeman, J. & Ané, C. Rphylopars: fast multivariate phylogenetic comparative methods for missing data and within-species variation. Methods Ecol. Evol. 8, 22–27 (2017).
Penone, C. et al. Imputation of missing data in life-history trait datasets: which approach performs the best? Methods Ecol. Evol. 5, 961–970 (2014).
Venditti, C., Meade, A. & Pagel, M. Multiple routes to mammalian diversity. Nature 479, 393–396 (2011).
Craig, J. M., Kumar, S. & Hedges, S. B. Limitations of phylogenomic data can drive inferred speciation rate shifts. Mol. Biol. Evol. 39, msac038 (2022).
Froese, R. & Pauly, D. FishBase: World Wide Web electronic publication. www.fishbase.org (2022).
Dengler, J. Which function describes the species–area relationship best? A review and empirical evaluation. J. Biogeogr. 36, 728–744 (2009).
Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G. & Jarvis, A. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25, 1965–1978 (2005).
Caissie, D. The thermal regime of rivers: a review. Freshw. Biol. 51, 1389–1406 (2006).
Linke, S. et al. Global hydro-environmental sub-basin and river reach characteristics at high spatial resolution. Sci. Data 6, 283 (2019).
Karger, D. N. et al. Climatologies at high resolution for the earth’s land surface areas. Sci. Data 4, 170122 (2017).
Morlon, H. et al. RPANDA: an R package for macroevolutionary analyses on phylogenetic trees. Methods Ecol. Evol. 7, 589–597 (2016).
Zachos, J. C., Dickens, G. R. & Zeebe, R. E. An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature 451, 279–283 (2008).
Dormann, C. F. et al. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography 36, 27–46 (2013).
Groemping, U. Relative importance for linear regression in R: the package relaimpo. J. Stat. Softw. 17, 1–27 (2007).
Diniz‐Filho, J. A. F., Bini, L. M. & Hawkins, B. A. Spatial autocorrelation and red herrings in geographical ecology. Glob. Ecol. Biogeogr. 12, 53–64 (2003).
Acknowledgements
This study was supported in part by the Coordenação de Aperfeiçoamento de Pessoa de Nível Superior—Brasil (CAPES)—Finance Code 001; the Swiss Federal Institute for Forest, Snow and Landscape (WSL); and the National Institutes for Science and Technology (INCT) in Ecology, Evolution and Biodiversity Conservation, supported by MCTIC/CNPq (proc. 465610/2014-5) and FAPEG (proc. 201810267000023). F.O.C. was supported by a doctorate and a ‘sandwich’ fellowship from CAPES. C.H.G. and E.B. acknowledge funding support from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant 787638) and the Swiss National Science Foundation (grant 173342). M.T.P.C. and C.H.G. are financially supported by the Swiss National Science Foundation (SNSF, no. 315230_197753). We also thank the Graham group for the insightful discussions on the paper.
Author information
Authors and Affiliations
Contributions
F.O.C., C.S.D. and C.G. designed the study. F.O.C. collected and analyzed the data with the crucial assistance of M.T.P.C., E.B., F.A.S.C. and J.A. All authors (F.O.C., C.S.D., M.T.P.C., F.A.S.C., E.B., J.S.A., R.O.W., C.H.G.) actively participated in the interpretation and discussion of the results, providing valuable insights throughout the drafting process.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Cerezer, F.O., Dambros, C.S., Coelho, M.T.P. et al. Accelerated body size evolution in upland environments is correlated with recent speciation in South American freshwater fishes. Nat Commun 14, 6070 (2023). https://doi.org/10.1038/s41467-023-41812-7
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-023-41812-7
- Springer Nature Limited