Darwin Returns to the Galapagos: Genetic and Morphological Analyses Confirm the Presence of Tramea darwini at the Archipelago (Odonata, Libellulidae)

Simple Summary Flying insects are able to colonize oceanic islands by both active and passive dispersal. Ten species of dragonflies are found in the Galapagos archipelago, located at 900 km from mainland South America. Shortly after the publication of Darwin’s “The Origin of Species”, one of the dragonflies from these islands was named after him as Tramea darwini. However, subsequent studies considered it to belong to another continental species of the same genus known as Tramea cophysa. Here, we studied a series of specimens of Tramea collected in 2018 from the Islands of San Cristobal, Isabela, and Santa Cruz, with the aim of determining their specific identity, through a combination of molecular and morphological analyses. Our results indicate that the Galapagos specimens examined belong to Tramea calverti, another continental species, and not to T. cophysa as previously thought. Following the principle of priority in taxonomic nomenclature, Tramea calverti, which was described in 1910 by Muttkowski, should hereafter be considered a synonym of Tramea darwini, which was described in 1889 by Kirby; hence, the species named after Darwin is to be considered valid, inhabiting both the Galapagos islands and continental America. Abstract The status of the Tramea species present in the Galapagos Islands (Odonata, Libellulidae) has been the subject of a long-standing debate among odonatologists. Here, we use molecular and morphological data to analyze a series of specimens from this genus collected in 2018 from the Islands of San Cristobal, Isabela, and Santa Cruz, with the aim of determining their relationship with Tramea calverti Muttkowski and with their currently considered senior synonym T. cophysa Hagen. We combined sequencing of mitochondrial and nuclear DNA with morphological examination of several specimens of Tramea, including representatives of continental T. cophysa and T. calverti. Our molecular analyses place the Tramea from Galapagos in the same clade as T. calverti, with T. cophysa as a closely related species. The morphological analyses found only one consistent difference between T. cophysa and T. calverti: the presence of an accessory lobe in the male vesica spermalis of T. cophysa that is absent in T. calverti and in the Tramea from Galapagos. In agreement with our genetic results, the overall morphological differences documented by us indicate that the Galapagos material examined is conspecific with T. calverti. In light of this, and following the principle of priority in taxonomic nomenclature, Tramea calverti Muttkowski, 1910 should hereafter be considered a junior synonym of Tramea darwini Kirby, 1889.


Introduction
Islands are laboratories of evolution in action, mainly due to their isolation and small populations, which drive rapid changes in the biota [1]. For these reasons, A total of 19 specimens of T. darwini were collected during a field trip to the Galapagos Islands of Isabela, Santa Cruz, and San Cristobal in August 2018 by AC-R and MOL-C (see Supplementary Table S1 for details on collection locations). Adult individuals were captured with a hand net and placed in 80% ethanol until DNA extraction. Legs from dried specimens of several Tramea species (T. cophysa, T. calverti, Tramea abdominalis (Rambur, 1842), Tramea basilaris (Palisot de Beauvois, 1817), Tramea binotata (Rambur, 1842), Tramea carolina (Linnaeus, 1763), Tramea lacerata Hagen, 1861 and Tramea virginia (Rambur, 1842)) belonging to RWG's personal collection were also used for DNA extraction (see Supplementary  Table S1).
Total genomic DNA was extracted from individual legs using the GeneJet DNA extraction kit (ThermoFisher Scientific, Waltham, MA, USA), following the manufacturer's protocol. To reconstruct the phylogenetic relationships between the Tramea species, we amplified fragments of the mitochondrial 16S and Cytochrome Oxidase I (COI) genes, together with the nuclear Internal Transcribed Spacer (ITS), using previously published primers [23][24][25] (see Supplementary Table S2). PCR reactions were carried out using the DreamTaq Green PCR Master Mix (ThermoFisher Scientific, Waltham, MA, USA). Prior to sequencing, unincorporated primers and dNTPs were removed using Shrimp Alkaline Phosphatase and Exonuclease I (New England Biolabs, Ipswich, MA, USA). Cleaned PCR products were sequenced in both directions using BigDye v.3.1 chemistry (Applied Biosystems, Foster City, CA, USA) in an ABI 3730xl DNA Analyzer (Applied Biosystems, Foster City, CA, USA), by the Macrogen Spain sequencing services.

Genetic Analyses
After sequencing, chromatograms were visually inspected, trimmed and automatically assembled using Geneious v. 9.1.7 (https://www.geneious.com). For some of the ITS sequences, we obtained superimposed traces, characteristic of sequences containing heterozygous insertions/deletions (indels). Allelic sequences were reconstructed using Indelligent v.1.2 [26]. We run BLAST searches for all DNA sequences at the NCBI website (https://blast.ncbi.nlm.nih.gov/Blast.cgi), to ensure that they were not derived from contaminations.
Sequences were aligned using MAFFT [27,28], as implemented in Geneious v 9.1.7. Genetic differentiation between species (p-distance) was estimated for each dataset in MEGA X [29], using the pairwise deletion option, which removes all ambiguous positions for each sequence pair. Mitochondrial DNA (mtDNA) alignments were concatenated for phylogenetic analyses. Phylogenetic relationships among Tramea species were reconstructed using maximum likelihood (ML) and Bayesian inference (BI) approaches. To increase the robustness of the analyses, previously published sequences from several Tramea species downloaded from GenBank (https://www.ncbi.nlm.nih.gov/genbank/) were added to our datasets. Pantala flavescens was selected as the outgroup for the phylogenetic analyses (see Supplementary Table S1).
ML analyses were carried out using RAxML 7.2.8 [30,31] as implemented in Geneious v 9.1.7., using the rapid bootstrapping and search for best scoring ML tree option, under the GTR + I + G model. Support for the nodes was estimated by running 1000 bootstrap replicates. For BI analyses, we used MrBayes 3.2.6 [32,33], as implemented in Geneious v 9.1.7. BI searches were run for 1.1 million generations, with default priors and with the GTR + I + G substitution model. Resulting phylogenetic trees were edited with TreeGraph 2 [34].
To further confirm the species delimitation within our datasets, we used the single locus distance-based delimitation methods implemented by Automated Barcode Gap Detection (ABGD) [35]. Analyses were run at the ABGD web server (https://bioinfo.mnhn. fr/abi/public/abgd/abgdweb.html). Fasta files including the aligned ingroup sequences (i.e., the sequences from the Tramea species) for each locus (ITS, COI, and 16S) were used as input files for the analyses, which were carried out with the default options.

Morphological Analyses
We examined ten individuals of Tramea from the Galapagos, to determine if any morphological characters would corroborate the placement of these populations under T. calverti or T. cophysa: two individuals from Isabela island (1 ♂, 1 ♀), six individuals from San Cristobal island (3 ♂♂, 3 ♀♀) and two from Santa Cruz island (1 ♂, 1 ♀). Our material from Galapagos was compared to 22 individuals of T. cophysa (15 ♂♂, 7 ♀♀) from southeastern Brazil and northern Argentina, and 46 individuals of T. calverti (27 ♂♂, 19 ♀♀) ranging from Arizona (USA) south to northern Argentina. We also examined specimens of both species taken together at Ilha de Marambaia in Rio de Janeiro State, Brazil in order to further determine whether the morphology of the Tramea from Galapagos matched either T. cophysa or T. calverti.
Specimens of T. cophysa and T. calverti belonging to RWG's collection were killed by injecting them with acetone, in order to preserve color patterns and to prevent lateral pressure distortions. Afterwards, they were placed in envelopes and steeped in acetone for 24 h, to promote drying. The Tramea specimens collected at the Galapagos by MOL-C and AC-R, which were preserved in 80% ethanol, were injected with acetone and steeped in acetone for 24 h, prior to morphological analyses.
Specimens were examined under a Zeiss Discovery V20 Stereo Microscope at magnifications ranging from 7.5× to 150×. Entire specimens and wings were scanned at 1200 DPI using an Epson Perfection V600 Photo Scanner. Heads and abdomens were photographed using a Leica MC 170 HD digital camera attached to the microscope at varying magnifications and stacked using Helicon Focus software. Vesicas were removed from each male, soaked in hot water, and cleaned with a brush so that the various lobes and details of this structure could be observed. Using a pair of watch forceps, resulting preparations were temporarily pinned with 0.10 minutens and fastened to small lumps of dentist's wax submerged in 95% ethanol and illustrated via a camera lucida.
The following morphological characters were examined: Head-coloration of vertex, postfrons and labrum in males and females; wings-coloration of wing membrane and extent of Hw basal spot in males and females; vesica spermalis; abdomen-coloration of ventral tergites; and S8 in males and females (abbreviations: Hw = hind wing; S = abdominal segment).

Genetic Analyses
The BLAST searches identified all our obtained sequences as similar to other odonate sequences available in the NCBI database. The final datasets (excluding the outgroup, P. flavescens) consisted of 44 COI and 16S sequences and 37 ITS sequences. The 16S dataset was 505 bp long, with 39 variable sites and 32 parsimony informative sites; the COI dataset was 451 bp long, with 98 variable sites and 93 parsimony informative sites. The ITS dataset was 894 bp long (including gaps), with 188 variable sites and 85 parsimony informative sites. All the sequences generated in this study have been deposited in the GenBank sequence database (https://www.ncbi.nlm.nih.gov/genbank), under accession numbers MW246873-MW246955 (see Supplementary Table S1).
Results of phylogenetic analyses were congruent for both datasets and also for the BI and ML analyses, placing T. darwini in the same clade as the continental T. calverti, while T. cophysa appears as a sister/closely related species. These relationships were supported by high posterior probability and bootstrap values (Figure 1). In agreement with the results of the phylogenetic analyses, the interspecific distances for both the mitochondrial and nuclear datasets group T. darwini with T. calverti. The average distances between T. darwini and T. calverti were 0% for 16S, 0.4% for COI and 0.3% for ITS; while the average distances between T. darwini and T. cophysa were 1.3% for 16S, 3.3% for COI, and 6.7% for ITS (see Supplementary Tables S3 and S4).
The ABGD species delimitation analyses identified nine groups (i.e., phylogenetic species) for the mtDNA loci and eight groups for the nDNA locus (see Supplementary Information Data S1). The groups comprised by (i) T. cophysa and (ii) T. darwini and T. calverti were both recovered as different species in all cases ( Figure 1).

Morphological Analyses
We found only one morphological difference between T. cophysa and T. calverti. The vesica spermalis of T. cophysa in lateral view possesses an accessory lateral lobe ( Figure  2c,d), which is absent in T. calverti (Figure 2a). The vesica in specimens of Tramea from the Galapagos also lacks the accessory lateral lobe (Figure 2b). We found no consistent differences in the morphology of the male hamules and cerci nor in female morphologies. In agreement with the results of the phylogenetic analyses, the interspecific distances for both the mitochondrial and nuclear datasets group T. darwini with T. calverti. The average distances between T. darwini and T. calverti were 0% for 16S, 0.4% for COI and 0.3% for ITS; while the average distances between T. darwini and T. cophysa were 1.3% for 16S, 3.3% for COI, and 6.7% for ITS (see Supplementary Tables S3 and S4).
The ABGD species delimitation analyses identified nine groups (i.e., phylogenetic species) for the mtDNA loci and eight groups for the nDNA locus (see Supplementary Information Data S1). The groups comprised by (i) T. cophysa and (ii) T. darwini and T. calverti were both recovered as different species in all cases (Figure 1).

Morphological Analyses
We found only one morphological difference between T. cophysa and T. calverti. The vesica spermalis of T. cophysa in lateral view possesses an accessory lateral lobe (Figure 2c,d), which is absent in T. calverti (Figure 2a). The vesica in specimens of Tramea from the Galapagos also lacks the accessory lateral lobe (Figure 2b). We found no consistent differences in the morphology of the male hamules and cerci nor in female morphologies. The wing membrane in T. cophysa is hyaline in both sexes, and the Hw basal spot is dark brown and can be variable in extent (see Figures 4a-d and 5a,b). The Hw spot in females is often reduced occupying the basal half or less of the length of the Hw base (Figures 4b,d and 5b). In mainland populations of T. calverti the wings are slightly infused with amber (Figures 4e-h,m,n and 5d,e) and the Hw spot is of a slightly lighter brown and is more extensive in both sexes. The size and extent of the Hw patch in mainland populations tends to be uniform exhibiting less variability than in T. cophysa (Figures 4e-h,m,n  and 5d,e). The Galapagos populations exhibit characters of both species, with the Hw patch being variable as in T. cophysa and the wing membrane having less of an infusion of amber coloration compared to mainland populations (Figures 4i-l and 5c,d). The vertex, postfrons, and labrum consistently differed in coloration, only in the males, as follows: the dorsal surface of the vertex in T. cophysa is dark metallic violet matching the coloration of the postfrons (Figure 3a,b). The dorsal surface of the vertex in T. calverti is pale brown (may be obscured in postmortem preservation) and differs from the dark metallic violet luster of the postfrons (Figure 3g,h,j,). Additionally, the metallic violet luster in T. cophysa extends anteriorly covering the entire postfrons. In T. calverti, the dark metallic violet luster is confined to about the basal half of the postfrons. The labrum is mostly dark brown in males of T. cophysa (Figure 3b) but mostly pale brown in males of T. calverti (Figure 3h,j). Males of Tramea from the Galapagos consistently matched the coloration of T. calverti (Figure 3d,f). We detected no consistent difference in female head coloration between the two species (Figure 3c,e,i,k).
The wing membrane in T. cophysa is hyaline in both sexes, and the Hw basal spot is dark brown and can be variable in extent (see Figures 4a-d and 5a,b). The Hw spot in females is often reduced occupying the basal half or less of the length of the Hw base (Figure 4b,d and Figure 5b). In mainland populations of T. calverti the wings are slightly infused with amber (Figure 4e-h,m,n and Figure 5d,e) and the Hw spot is of a slightly lighter brown and is more extensive in both sexes. The size and extent of the Hw patch in mainland populations tends to be uniform exhibiting less variability than in T. cophysa (Figure 4e-h,m,n and Figure 5d,e). The Galapagos populations exhibit characters of both species, with the Hw patch being variable as in T. cophysa and the wing membrane having less of an infusion of amber coloration compared to mainland populations (Figures 4i-l and 5c,d).     Coloration of the ventral abdominal tergites generally differs between T. cophysa and T. calverti. The majority of the ventral tergites in both sexes of T. cophysa are dark brown and of the same color as the lateral carinae in most of the abdominal segments (Figure 6a,b). The ventral tergites of some females (Figure 6c) may possess a lighter color differing from the dark lateral carinae but in those females, a light dusting of pruinosity is usually present on the more anterior segments (Figure 6c). The ventral tergites in T. calverti are always pale brown, with dark markings confined to the vicinity of the lateral carinae (Figure 6d-g). No pruinosity was observed in any of the T. calverti examined. Abdominal S8 is entirely dark brown in males of T. cophysa (Figure 7a,b), and the same is often observed in the females of this species ( Figure 7c); but, in some females S8 shows a pale coloration confined to the lower half of the segment (Figure 7d). Both sexes of T. calverti consistently have an inverted pale brown semicircular spot at the base of S8 (Figure 7e,f), which is red or brown in live specimens. Specimens of Tramea from the Galapagos (Figure 7g,h) were consistent in coloration and pattern with T. calverti.

Discussion
The results of our genetic analyses show that the samples of T. darwini collected by us at three islands in the Galapagos archipelago (San Cristobal, Isabela, and Santa Cruz), all belong to a clade that also includes the mainland species currently known as T. calverti, with T. cophysa as a closely related species. The phylogenetic relationships are concordant between nuclear and mitochondrial DNA and supported by high bootstrap and posterior probability values in both cases. Furthermore, the genetic distances between the Tramea from Galapagos and T. calverti are also lower than the interspecific distances between the Galapagos individuals and T. cophysa (see Results and Supplementary Tables S3  and S4). If we consider the 2% threshold commonly used as a limit between different species [36], we can conclude that the Tramea from Galapagos belong to T. calverti and not to T. cophysa. ABGD species delimitation methods provided further support for the placement of T. darwini and T. calverti within the same group, separated from T. cophysa.
In agreement with the molecular results, our morphological analyses indicate also a closer relationship between the material collected in the Galapagos islands and mainland T. calverti. This includes also the single morphological character noted above that consistently differentiates both T. calverti and T. cophysa: the accessory lobe that occurs in the male sperm vesica of T. cophysa (Figure 2c,d) but is absent in both T. calverti (Figure 2a) and the Tramea from Galapagos (Figure 2b).
In their exhaustive work on the T. cophysa complex, De Marmels and Rácenis [14] treat T. darwini Kirby, 1889, as a junior synonym of the older name, T. cophysa Hagen, 1867; following Calvert [16] (p. 303) who stated: "Tramea darwini based on a female with a much reduced basal wing marking has been shown by Mr. Currie [1901: 386] to vary greatly in this respect. His material is before me and I cannot separate it from some of the examples from Ecuador, while the Haytian female captured by W. Cabot has the basal brown spot of the hind wings reaching no farther backward than 1 mm. beyond the apex of the membranule." De Marmels and Rácenis had access to only two females of T. darwini (Galapagos Is.: Albemarle, 6.VIII.1955; and Charles), a fact that may have accounted for their placement of the Galapagos specimens under T. cophysa.
The original description of T. cophysa states: "Die drei letzten Ringe oben und die vorhergehenden längs der Bauchkante schwarz; unten vom dritten Ringe an schwarz [Body brown, the last three abdominal segments at the top and those entirely along the edge of the abdomen black; at the bottom black from the third abdominal segment.]", a characteristic not present in T. calverti.
A clue to the true identity of the Galapagos material was partially rectified by Peck [13] (pp. 313 and 316) as follows: "[T. cophysa] may have been confused with T. calverti, and its literature records may refer to T. cophysa. De Marmels and Rácenis (1982) clarify the characters and distributions of T. calverti and T. cophysa, and list only T. cophysa from the Galapagos. The key in De Marmels and Rácenis (1982) should be consulted. Dunkle (1989, and pers. comm.) has seen material of T. calverti but not T. cophysa from the Galapagos. My material contains only specimens of T. calverti. I have examined Galapagos specimens in USNM and CAS collections and found specimens of T. calverti which had been labelled as T. cophysa. If T. cophysa was actually once present and is now absent in the Galapagos it represents a case of natural extirpation of island populations". Peck's key to Tramea (couplet number 6) also differentiates clearly between both species: Our morphological analyses have shown that the Tramea collected in the Galapagos tally with those of T. calverti and not T. cophysa; although some differences in wing pattern and coloration between T. calverti and the Galapagos specimens do exist. The female lectotype of T. darwini (Figure 4i) possess almost no Hw basal wing spot, which likely led Kirby to describe the species as new. Variability in Hw basal wing spot pattern was also illustrated by Asahina [21] and encompasses a greater variability compared to our samples.
Regarding the male vesica spermalis, Gloger [22] stated that he was unable to find any differences in this structure between T. darwini and T. cophysa ("Ich habe die Penis aller mir zur Verfügung stehenden Exemplare des Kontinents (nur sogen. Form c, mit folgender Herkunft: Argentinien: San Isidro, Prov. Buenos Aires; Playadito, Prov. Corrientes; Mascasín, Prov. La Rioja; Bolivien: Roboré) mit denen der Galapagos-Ausbeute verglichen, ohne Differenzen zu finden. [I have compared the penis of all forms of the continent available to me (only so-called form c, with the following origin: Argentina: San Isidro, Prov. Buenos Aires; Playadito, Prov. Corrientes; Mascasín, Prov. La Rioja; Bolivia: Roboré) with those of the Galapagos yield without finding any differences.]"). At first glance, we could not find any differences in the vesica between T. darwini and T. cophysa, until we prepared the structure as described in the Materials and Methods section. Some of the various lobes present in the vesica are difficult to see, and they may be obscured in dried specimens as the ones examined by Gloger [22].
We suggest the following couplet changes in the key provided by De Marmels and Rácenis [14] for separation of T. cophysa from other members of the T. cophysa group, as follows: Given our results, we consider that the Tramea populations present in Galapagos are conspecific with Tramea calverti. Following the principle of priority, this species should be referred to by its older name, T. darwini [15]; from which T. calverti [37] becomes a junior subjective synonym.

•
Our genetic analyses show that the Tramea species from Galapagos belongs to a clade that comprises also the continental species T. calverti, and T. cophysa appears as a closely related species.

•
Concordant with the results of the genetic analyses, the morphology of the Tramea collected in Galapagos is closer to T. calverti than to T. cophysa.

•
Only one morphological character has been found which consistently discriminates between both species: an accessory lobe in the male sperm vesicle, which appears in T. cophysa, but is absent in both T. calverti and the Tramea from Galapagos. • Given these results, and following the taxonomic principle of priority, the Tramea species currently found in Galapagos should be referred to by its older name, T. darwini.

Supplementary Materials:
The following are available online at https://www.mdpi.com/2075-4 450/12/1/21/s1, Table S1: List of Tramea species included in the present study; Table S2: Primer combinations used to amplify mitochondrial (COI and 16S) and nuclear (ITS) DNA of the Tramea species included in this study; Table S3: Estimates of divergence over sequence pairs between the Tramea species included in this study for the mtDNA datasets; Table S4: Estimates of divergence over sequence pairs between the Tramea species included in this study for the nDNA dataset; Data S1: Results of species delimitation analyses.  Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: DNA sequences obtained from this study are available in the GenBank database, and data are available in the Supplementary Information File. All the Tramea specimens examined and/or used for genetic analyses are deposited in the collections of Adolfo Cordero-Rivera and Rosser W. Garrison.