Research Article |
Corresponding author: Blanca E. Hernández-Baños ( behb@ciencias.unam.mx ) Academic editor: Knud Jønsson
© 2020 Melisa Vázquez-López, Juan J. Morrone, Sandra M. Ramírez-Barrera, Anuar López-López, Sahid M. Robles-Bello, Blanca E. Hernández-Baños.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Vázquez-López M, Morrone JJ, Ramírez-Barrera SM, López-López A, Robles-Bello SM, Hernández-Baños BE (2020) Multilocus, phenotypic, behavioral, and ecological niche analyses provide evidence for two species within Euphonia affinis (Aves, Fringillidae). ZooKeys 952: 129-157. https://doi.org/10.3897/zookeys.952.51785
|
The integration of genetic, morphological, behavioral, and ecological information in the analysis of species boundaries has increased, allowing integrative systematics that better reflect the evolutionary history of biological groups. In this context, the goal of this study was to recognize independent evolutionary lineages within Euphonia affinis at the genetic, morphological, and ecological levels. Three subspecies have been described: E. affinis godmani, distributed in the Pacific slope from southern Sonora to Guerrero; E. affinis affinis, from Oaxaca, Chiapas and the Yucatan Peninsula to Costa Rica; and E. affinis olmecorum from Tamaulipas and San Luis Potosi east to northern Chiapas (not recognized by some authors). A multilocus analysis was performed using mitochondrial and nuclear genes. These analyses suggest two genetic lineages: E. godmani and E. affinis, which diverged between 1.34 and 4.3 My, a period in which the ice ages and global cooling fragmented the tropical forests throughout the Neotropics. To analyze morphometric variations, six morphometric measurements were taken, and the Wilcoxon Test was applied to look for sexual dimorphism and differences between the lineages. Behavioral information was included, by performing vocalization analysis which showed significant differences in the temporal characteristics of calls. Finally, Ecological Niche Models were estimated with MaxEnt, and then compared using the method of Broennimann. These analyses showed that the lineage distributed in western Mexico (E. godmani) has a more restricted niche than the eastern lineage (E. affinis) and thus we rejected the hypotheses of niche equivalence and similarity. Based on the combined evidence from genetic, morphological, behavioral, and ecological data, it is concluded that E. affinis (with E. olmecorum as its synonym) and E. godmani represent two independent evolutionary lineages.
Euphonia affinis, Euphonia godmani, independent evolutionary lineages
The integration of genetic, morphological, ecological, and behavioral data in systematic studies provides information on the evolutionary history of species and their populations, allowing a better assessment of species limits (
Species limits on birds have been studied using different approaches, including the use of morphological characters (
DNA sequences have been useful to complement morphological and geographical information. Phylogenetics, molecular clocks, diversification rates, genetic populations and coalescence analyses have documented that geological complexity, heterogeneity of the environment, and climatic oscillations may have influenced patterns of genetic diversity, demography and divergence within species (
Euphonia affinis is a member of the family Fringillidae, subfamily Euphoniinae (
Geographic distribution and morphotypes of Euphonia affinis, sampling, phylogeny, and haplotype networks. A geographic distribution of E. affinis: in blue E. a. godmani, in yellow E. a. affinis, and in red E. a. olmecorum (Geographic distribution modified from NatureServe shapefile in ArcGIS, ArcMAP 10.2.2; Esri, Redlands, CA, USA). Tissue sampling locations are indicated by circles in the map. Plumage morphotypes of E. a. godmani (female and male) with white undertail coverts, and E. a. affinis (female and male) with yellow undertail coverts. The previously proposed subspecies E. a. olmecorum (not shown) is similar to E. a. affinis, but paler plumage in females and a purple-blue back in males have been reported. B haplotype networks obtained for the mitochondrial gene ND2 and the nuclear genes ODC, MUSK, GAPDH intron 11, and BRM intron 15. Samples from the western distribution, assigned as E. a. godmani, are shown in blue and from the eastern distribution, assigned as E. a. affinis are indicated in yellow, E. a. olmecorum in red. C bayesian Inference concatenated phylogeny of E. a. godmani (west) and E. a. affinis-E. a. olmecorum (eastern Mexico, Central America).
In the present work, we applied integrative taxonomy to identify the independent evolutionary lineages within Euphonia affinis using four types of characters, multilocus genetic data, morphometric data, behavioral, and environmental niches. Based on the allopatric distribution of subspecies E. a. affinis (Eastern Mexico and Central America) and E. a. godmani (West of Mexico), as well as in the distinctive character of subcaudal feathers, we expect to recognize at least two independent evolutionary lineages that can be proposed to elevate at the species level.
Our goals were to: 1) obtain a phylogenetic hypothesis for Euphonia affinis subspecies using multilocus genetic data. 2) Associate the genetic variation and divergence times with historical geographic processes and barriers. 3) Describe the pattern of morphometric, behavioral, and environmental variation in Euphonia affinis, and associate it with genetic variation and phylogenetic relationships. Hence, our hypothesis is that multiple independent evolutionary lineages exist within the Euphonia affinis complex, and our objective is to define them with the integration of multilocus genetic data, morphometric, behavioral, and environmental data. Furthermore, we discuss a potential promotion of those lineages to species status.
For the ingroup we used 19 tissues from Euphonia affinis affinis and four from E. affinis godmani; for the outgroup we obtained one tissue sample from Chlorophonia occipitalis, two from Euphonia chlorotica, one from E. luteicapilla and one from Haemorhous mexicanus (Suppl. material
Genomic DNA was isolated using the Qiagen DNeasyTM kit (Qiagen Inc., Valencia, CA, USA) following the manufacturer’s protocol. We amplified five molecular markers: one mitochondrial ND2 (NADH Dehydrogenase Subunit 2, Sorenson et al. 1999) and four nuclear genes ODC (Ornithine Decarboxylase,
We constructed the alignment of each gene using the CLUSTAL IW (
To resolve heterozygotes in nuclear sequences, we used a Bayesian approach in PHASE v 2.1 (
Divergence times were estimated from the multilocus dataset with three genes, the mitochondrial gene ND2, and the two nuclear genes with sequences for all samples and outgroups (ODC and GAPDH) using Beast v1.8 (
Six morphometric measurements of 355 specimens (233 males and 122 females) were taken from the following collections (see Suppl. material
The normality was tested with the Shapiro-Wilk test of normality in R (
We obtained 19 recordings of Euphonia affinis calls from the Xeno-Canto (XC; http://www.xeno-canto.org) open access database. We used only call recordings in which the subspecies was identified and in which Euphonia was identified as the foreground species. We visualized and measured spectrograms of these recordings using the Raven Pro 1.6 software (Cornell University, Ithaca, NY). We visually inspected the spectrograms, and from each recording we selected one call section that did not overlap with any background vocalizations or other sounds. In recordings where more than one call variant occurred (for example, variants with differing number of notes), we selected one of the most frequent type. The most common call type for this species consists of a short series (2 to 4 notes) of whistled notes with decreasing pitch. Since recording conditions were not standardized, we only took frequency and duration measurements, which are not heavily affected by distance. We measured low and high frequencies (LowFreq and HiFreq), change in frequency (DeltaF), duration of call (DeltaT), number of notes (Notes) and emission rate (Speed; number of notes divided by duration). All measured variables were rescaled by log transforming them.
Unpaired Two-Samples Wilcoxon Test were carried out on individual variables to test for differences between the two groups. We also performed a principal component analysis (PCA) to explore the relation between the two groups in multivariate space. All the scripts and input data are in https://github.com/almamelisa/Euphonia-affinis-complex.
The georeferenced records were obtained from the specimens used in the morphometric and genetic analyses, 102 for E. affinis affinis and 29 for E. affinis godmani. To define the M area (accessibility area; sensu Soberón & Peterson, 2005) for each evolutionary lineage herein identified, we plotted the record points onto the biogeographic provinces of the Neotropical region (
For the first explorative analysis, we used the 19 bioclimate layers from WorldClim and assessed which variables were the most important for the model, according to the Jackknife test calculated in MaxEnt (
To evaluate the models we calculated the partial ROC (Receiver Operating Characteristic) in the web tool Niche Tool Box (https://shiny.conabio.gob.mx:3838/nichetoolb2/), the parameters were 0.05 proportions of omission, 50 random points percentage and 500 iterations. Also, in MaxEnt, we made four models projections one in the M area of each lineage and the remaining three were made to obtain the paleodistribution; we projected the ENM in the last maximum glacial period (~ 22,000 years ago) considering two general circular models: MIROC-ESM (Hasumim and Emori 2004) and CCSM (
The multilocus dataset analyses revealed a well-supported monophyly for the Euphonia affinis complex and recovered two main phylogroups: one included the samples from western Mexico (E. affinis godmani) and the other comprised samples from eastern Mexico and Central America (E. affinis affinis and E. affinis olmecorum) (Fig.
The haplotype network obtained with ND2 sequences showed two geographically structured haplogroups: a western group and an eastern-CA group (Fig.
Diversity indices, nucleotide content, evolution model, variation sites, and alignment base pairs of mtDNA and nDNA.
Gene | H-A | Hd | Pi | D | NC | MEM | PIS | MS | Alignment BP | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Hd | Σ | SD | Pi | σ | SD | %T | %C | %A | %G | 1049 | ||||||
ND2 | 12 | 0.95 | 7E-04 | 3E-02 | 0.019 | 7E-05 | 8E-03 | -0.01* | 26 | 32.5 | 31.6 | 10.4 | TVM+G | 358 | 598 | (997–1041) |
ODC | 7 | 0.56 | 6E-03 | 8E-02 | 0.004 | 9E-07 | 9E-04 | -0.67* | 36.6 | 16.9 | 27.3 | 19.1 | HKY | 10 | 535 | 556 |
MUSK | 11 | 0.86 | 1E-03 | 4E-02 | 0.004 | 3E-07 | 5E-04 | -0.70* | 32.6 | 16.8 | 30.5 | 20.3 | TPM1uf | 8 | 476 | 500 |
GAPDH | 3 | 0.17 | 5E-03 | 7E-02 | 0.0006 | 1E-07 | 7E-02 | -1.13* | 25.4 | 19.9 | 21 | 33.7 | HKY | 1 | 262 | 280 |
BRM | 4 | 0.5 | 5E-03 | 8E-02 | 0.002 | 2E-06 | 3E-04 | -0.30* | 34.6 | 12.9 | 34.6 | 17.9 | HKY | 2 | 288 | 304 |
ND2 | ODC | MUSK | GAPDH | BRM | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 1 | 2 | 3 | 4 | 1 | 2 | 3 | 4 | 1 | 2 | 3 | 4 | 1 | 2 | 3 | 4 | |
1 | ||||||||||||||||||||
2 | 0.11** | 0.01** | 0.01** | 0.00** | 0.01** | |||||||||||||||
3 | 0.14* | 0.17* | 0.01** | 0.01** | 0.02** | 0.02** | 0.05** | 0.05** | 0.16* | 0.06* | ||||||||||
4 | 0.15* | 0.16* | 0.09* | 0.01** | 0.01** | 0.01* | – | – | – | 0.05** | 0.05* | 0.00** | 0.07* | 0.07* | 0.07* | |||||
5 | 0.14* | 0.15* | 0.09* | 0.03* | 0.01** | 0.01** | 0.01* | 0.01** | – | – | – | – | 0.121* | 0.121* | 0.115* | 0.115* | – | – | – | – |
Euphonia affinis godmani and E. affinis affinis split 2.6 Mya (1.5–4.0 Mya, 95% HPD), during the Late Pliocene-Early Pleistocene (Fig.
Ultrametric phylogenetic tree obtained by BEAST using ND2, ODC, and GAPDH concatenated matrix. The rhombus node represents the calibration point 17.1104 My with a 95% HPD of (14.7743, 19.6278) (see methods), dark gray circle node represents the E. affinis origin and light gray circle node represents the break between E. a. godmani and E. a. affinis. Above the branch the diversification dates (My) and in brackets the 95% HPD. Below branch the number indicated the posterior probability. The green area corresponds to the period when lowland dry forests had a greater expansion in Western Mexico.
A total of 355 specimens was analyzed, of which 180 were males and 97 females of E. affinis affinis, and 53 males and 25 females of E. affinis godmani. Morphometric sexual dimorphism was found in three variables: TLE (Tail Length), WC (Wing Chord), and BD (Bill Depth) (Fig.
PCA analyses for both sexes showed that the first two principal components explained a large proportion of the variance (70.37%, Table
Morphometric analyses results. A) Females boxplots and PCA for WC, TLE, and BD morphometric characters. B) Males boxplot and PCA for WC, TLE, and BD morphometric characters. C) Boxplot and PCA for TL, BL, and BW. WC, TLE, and BD characters were analyzed by separated sex, because the analyses indicated sexual dimorphism (see results and Table
Median, Unpaired Two-Samples Wilcoxon Test to evaluate differences between lineages and sexes, p-value < 0.05 (in bold).
Between lineages | ||||||
Both sexes | PC1 | PC2 | PC3 | |||
affinis | godmani | p-value | 37.1% | 33.2% | 29.6% | |
TL | 12.93(11.1–14.6) | 13.12 (11.6–13.9) | 3.10E-02 | 0.74 | -0.0047 | -0.67 |
BL | 6.62(4.7–7.9) | 6.52(6.0–6.9) | 1.36E-04 | 0.48 | 0.76 | 0.42 |
BW | 5.92(1.9–6.9) | 6.05(4.5–6.6) | 1.42E-03 | 0.57 | -0.63 | 0.51 |
Females | PC1 | PC2 | PC3 | |||
affinis | godmani | p-value | 45.96% | 34.74% | 19.28% | |
WC | 52.16 (46.6–55.97) | 53.83 (49.3–55.2) | 1.15E-04 | 0.77 | 0.4 | -0.48 |
TLE | 28 (23.8–37.6) | 27.66 (22.3–32.0) | 0.227 | -0.24 | 0.93 | 0.26 |
BD | 4.11 (3.4–5.0) | 4.32 (3.3–4.9) | 5.68E-03 | 0.84 | -0.09 | 0.52 |
Males | PC1 | PC2 | PC3 | |||
afifnis | godmani | p-value | 47.7% | 35.5% | 16.8% | |
WC | 53(47.7–60.1) | 54.61(51.4–57.0) | 7.64E-08 | 0.7 | 0.56 | -0.41 |
TLE | 28.66(21.0–34.0) | 27.33 (22.4–32.5) | 3.27E-03 | -0.4 | 0.85 | 0.31 |
BD | 4.19(3.5–5.1) | 4.68(3.7–6.6) | 1.43E-13 | 0.87 | -0.06 | 0.48 |
Sexual dimorphism | ||||||
E. a. affinis | E. a. godmani | |||||
females | males | p-value | Females | males | p-value | |
TL | 13.017(11.1–14.6) | 12.908(11.3–14.1) | 0.11 | 13.213(12.3–13.9) | 11.6–13.9) | 0.43 |
BL | 6.607(4.7–7.6) | 6.633(5.0–7.9) | 0.09 | 6.537(6.0–6.9) | 6.523(6.0–6.9) | 0.60 |
BW | 5.947(4.5–6.9) | 5.914(1.9–6.6) | 0.99 | 6.043(4.5–6.4) | 6.06(4.5–6.6) | 0.81 |
WC | 52.167(46.6–55.9) | 53(47.7–60.1) | 3.19E-04 | 53.833(49.3–55.2) | 54.613(51.4–57.0) | 6.22E-03 |
TLE | 28(23.8–37.6) | 28.667(21.0–34.0) | 1.48E-02 | 27.667(22.3–32–0) | 27.333(22.4–32.5) | 0.78 |
BD | 4.117(3.4–5.0) | 4.1985(3.5–5.1) | 0.08 | 4.327(3.3–4.9) | 4.683(3.7–6.0) | 4.40E-03 |
None of the frequency variables measured differed significantly between godmani and affinis groups (N =19). On the other hand, we found that the emission rate of E. a. affinis is much lower than that of E. a. godmani, on average 2.9 notes/s versus 5.09 notes/s. These differences are statistically significant (W = 0, p < 0.001). The first two Principal Components together explain 78.75% of variance. The first PC separates both groups unambiguously (Fig.
Our models obtained a high mean value for AUC (Area Under the Curve) ratio values and statistically significant, 1.68 for E. affinis affinis and 1.81 for E. affinis godmani (***P < 0.05), this indicates a good fit of ENM’s. According to the Jackknife test and contribution variables obtained by MaxEnt, the most important variable for E. affinis affinis model was BIO 15 (precipitation seasonality), and the variable BIO8 (mean temperature of wettest quarter) for E. affinis godmani. We present the ENM predictions in four levels in Fig.
The third part is the projection of the models in Last Glacial Maximum conditions (LGM 21–18,000 years ago), for E. affinis affinis showing a reduction in their environmental suitability along the present distribution, with predictions in areas like the Yucatan Peninsula and the western coast of Mexico with a gap at the western coast of the Tehuantepec Isthmus, unlike Present predictions where Central America has only small patches with predictions for E. affinis affinis (Fig.
The results of ecological overlap for the environmental PCA exhibit a large niche of E. a. affinis, while E. a. godmani exhibits an ecological niche compaction. A total variance of 83.67% is explained for the three principal components, with 41.04% for PC1, 29.886% for PC2 and 12.733% for PC3 (Fig.
Ecological niche modelling and its projection in the geographic areas for E. a. affinis (yellow) and E. a. godmani (blue). In all four panels (a-d), the contribution values of each environmental variable of ENM’s is illustrated in the left and the projection of the Ecological niche conditions in the geographic distribution area is shown in the maps. a Ecological Niche projected in the current geographic distribution area of E. affinis and E. a. godmani. b ENM’s projected into the geography for each lineage. c ENM of E. a. affinis and E. a. godmani projected in the Last Maximum Glacial ecological conditions. d ENM of E. a. affinis and E. a. godmani projected in the Last Inter Glacial ecological conditions.
Equivalence and similarity tests in environmental space for E. a. affinis and E. a. godmani. A PCA of Ecological niche for of E. affinis lineages and the variables contribution to the analyses. The gray gradient indicates the density of the occurrences of the lineages, and the dashed and solid line indicates the 50% and 100% of the environmental background B graphical results of the equivalency tests comparing the two lineages. For both tests (equivalence and similarity) we only presented values for the D metrics. For all graphs the D observed values of the overlap niche analyses are present with the black diamond. The p value is showing in each graphic, all of them not significant for these analyses C graphical results of the similarity test comparing the two lineages in both directions (E. a. affinis vs. E. a. godmani and vice versa), ns = Not significant, p > 0.05.
Tanagra affinis Lesson 1842, Rev. Zool. 5: 175.
Tanagra affinis affinis; Miller et al. 1957, Cooper Ornithol. Soc. Pac. Coast Avifauna 33: 298.
Euphonia affinis affinis; Dickerman, 1981, Occ. Pap. Mus. Zool. Louisiana State. Univ. 59: 3.
Euphonia affinis olmecorum Dickerman, 1981, Occ. Pap. Mus. Zool. Louisiana State. Univ. 59: 4, syn. nov.
Males. Yellow forehead, back black with bluish to violet glow, black throat, yellow from chest to belly, yellow subcaudal coverts feathers (
Females. Forehead olive-yellow and gray, olive-green back. The throat is olive-yellow, with a yellow belly, subcaudal coverts feathers also in yellow (
Through Gulf slope of Mexico from Nuevo Leon, S Tamaulipas and E San Luis Potosí to N Chiapas, Yucatan Peninsula, E of Guatemala, Belize to N Honduras; in the Pacific slope from W Oaxaca, Mexico to NW Costa Rica (
Euphonia godmani Brewster, 1889, Auk 6: 90.
Tanagra affinis godmani; Miller et al. 1957, Cooper Ornithol. Soc. Pac. Coast Avifauna 33: 298.
Euphonia affinis godmani; Dickerman, 1981, Occ. Pap. Mus. Zool. Louisiana State. Univ. 59: 1.
Male. Very similar to E. affinis with white undertail coverts feathers (
Geographical distribution: Along to Pacific slope of Mexico from SE Sonora S to C Guerrero (
We provide molecular, morphological, behavioral, and environmental niche evidence supporting the existence of two evolutionary lineages within the Euphonia affinis complex (E. godmani and E. affinis).
We found three lines of evidence in the molecular data to support the taxonomic split of the E. affinis complex at species level. The first is the reciprocal monophyly between E. affinis and E. godmani found in the multilocus analysis using mitochondrial and nuclear genes, where the samples of E. affinis olmecorum are included into E. affinis. These results agree with the proposals by Ridgway and Friedman (1901) and the taxonomic proposal of
Our analysis revealed significant differences between E. a. godmani and E. a. affinis in six characters among lineages. Bill Depth and Wing Chord are bigger for E. a. godmani, while E. a. affinis has bigger dimensions on Tail Length. Even though the rest of characters have significant differences, in the PCA plots the Tail Length, Bill Length and Bill Width characters do not show dispersion between both lineages, so we can assign Wing Chord, Tail Lenth and Bill Depth as diagnostic characters for males, and Wing Chord and Bill Depth as diagnostic characters for females. These results are similar to Phaethornis mexicanus morphometric patterns, a species also distributed along the Atlantic and Pacific Slope, where the Pacific lineage also shows bigger dimensions vs. the Atlantic lineage (Arbeláez-Cortés and Navarro-Sigüenza, 2013).
Our results show that there are significant differences in the temporal characteristics of calls between E. godmani and E. affinis while we found that there is little divergence in spectral structure or frequency measurements. E. godmani emits call notes at a significantly faster rate than E. affinis. Many bird species are highly sensitive to temporal cues in recognizing conspecific vocalizations (Dooling and Prior, 2016), which suggests that while call structure and frequency in this complex has been conserved, variation in tempo could be an important cue in conspecific recognition.
Euphonia affinis and E. godmani represent two different lineages with no significant conservatism in their ecological niches (west vs. east). The env-PCA, also, showed a larger ENM for E affinis, respect to E. godmani, also the western lineage has a limited ability to predict its ENM in the geographic area of E. affinis. The western coast of Mexico is characterized by a highly contrasting dry season vs. a wet season over the year, this characteristic is unique with respect to the eastern tropical area, so E. godmani has become restricted to these conditions. These results are similar to other taxa with sister lineages distributed along the Pacific and Atlantic slopes in Mesoamerican (Hernández-Canchola and León-Paniagua, 2017). It is interesting that E. godmani shows a reduction in ecological niche, while E. affinis presents a broader ecological niche. That may suggest a scenario where E. godmani was able to invade the western area of Mexico, and, in the absence of ecological competition from other Euphonias, it adapted and specialized to the floristic resources, as well as to the temperature and precipitation conditions of the area. While E. affinis conserved a broader ecological niche, as reflected in its geographical distribution, allowed it to explore more regions and resources, even in the presence of different species of Euphoniinae.
Lineage divergence between E. godmani (western Mexico) and E. affinis (eastern Mexico and Central America) occurred ~ 2.6 Mya (1.5–4.0 Mya HPD 95%), a range between the Pliocene and Pleistocene epochs. During the Pliocene, the Sierra Madre Occidental and the Transmexican Volcanic Belt finished emerging, which made the Pacific Slope drier than the Atlantic slope, due to the hillside effect (Graham and Dilcher, 1995). Additionally, the drier conditions were favored by meteorological phenomena that made the Pacific coast warmer than the Atlantic coast in the northern hemisphere (
In addition to the consequences of the orographic changes of the Pliocene, during the Late Pliocene, global and continuous cooling periods were frequent, and during the Pleistocene the climatic oscillations were defined by glacial and interglacial periods (
We incorporated different kinds of information to help us identify lineages within the Euphonia affinis species complex and understand the speciation process (
We thank the Museo de Zoología Alfonso L. Herrera (MZFC,
Tables S1, S2, S3. Sampling, genbank sequences and sequences of primers
Data type: species data
Raw morphometric data and collection information
Data type: morphological data