Concerted genetic, morphological and ecological diversification in ...

Molecular Ecology (2011) 20, 1936–1951

doi: 10.1111/j.1365-294X.2011.05065.x

Concerted genetic, morphological and ecological diversification in Nacella limpets in the Magellanic Province ´ LEZ-WEVAR,* T. NAKANO,† J. I. CAN ˜ E T E ‡ and E . P O U L I N * C. A. GONZA *Instituto de Ecologı´a y Biodiversidad, Departamento de Ciencias Ecolo´gicas, Facultad de Ciencias, Universidad de Chile, Las ˜ un˜oa, Santiago, Chile, †Department of Geology and Paleontology, National Museum of Nature and Science, Palmeras # 3425, N Tokyo, Japan, ‡Departamento de Recursos Naturales, Universidad de Magallanes, Punta Arenas, Chile

Abstract Common inhabitants of Antarctic and Subantarctic rocky shores, the limpet genus Nacella, includes 15 nominal species distributed in different provinces of the Southern Ocean. The Magellanic Province represents the area with the highest diversity of the genus. Phylogenetic reconstructions showed an absence of reciprocal monophyly and high levels of genetic identity among nominal species in this Province and therefore imply a recent diversification in southern South America. Because most of these taxa coexist along their distribution range with clear differences in their habitat preferences, Nacella is a suitable model to explore diversification mechanisms in an area highly affected by recurrent Pleistocene continental ice cap advances and retreats. Here, we present genetic and morphological comparisons among sympatric Magellanic nominal species of Nacella. We amplified a fragment of the COI gene for 208 individuals belonging to seven sympatric nominal species and performed geometric morphometric analyses of their shells. We detected a complete congruence between genetic and morphological results, leading us to suggest four groups of Nacella among seven analysed nominal species. Congruently, each of these groups was related to different habitat preferences such as bathymetric range and substrate type. A plausible explanation for these results includes an ecologically based allopatric speciation process in Nacella. Major climatic changes during the Plio-Pleistocene glacial cycles may have enhanced differentiation processes. Finally, our results indicate that the systematics of the group requires a deep revision to re-evaluate the taxonomy of Nacella and to further understand the Pleistocene legacy of the glacial cycles in the southern tip of South America. Keywords: adaptive radiation, ecological speciation, Nacella, Patagonia, Patellogastropoda, Southern Ocean Received 5 November 2010; revision received 27 January 2011; accepted 9 February 2011

Introduction True limpets of the order Patellogastropoda are the basal group in the evolution of Gastropoda, as revealed by morphological and molecular studies (Ponder & Lindberg 1997; Lindberg 1998; Harasewych & McArthur 2000; McArthur & Harasewych 2003; Nakano & Ozawa

Correspondence: C. A. Gonza´lez-Wevar, Fax: 056 02 2727363; E-mail: [email protected]

2004, 2007). Typically, taxonomic studies in the group have been based on shell morphology and external characters, but the high degree of variability and homoplasy detected in form and coloration has led to taxonomic confusion (Ridgway et al., 1998; Sasaki 1999; Espoz et al. 2004; Nakano & Spencer 2007; Lindberg 2008). Recent molecular studies have greatly improved the systematics and taxonomy of Patellogastropoda at the levels of family, genus and species. Seven families, Lotiidae, Acmaeidae, Pectinodontidae, Patellidae, Lepetidae, Eoacmaeidae and Nacellidae, are currently  2011 Blackwell Publishing Ltd

C O N C E R T E D E V O L U T I O N I N M A G E L L A N I C L I M P E T S 1937 recognized in the order using different molecular markers (Nakano & Ozawa 2007). Nacellidae includes two genera, Cellana and Nacella, with clearly disjoint distributions (Powell 1973; Lindberg 1998, 2008). Cellana has more than 37 species and subspecies mainly distributed in tropical to warmtemperate waters of the Indo-Pacific regions (Powell 1973; Lindberg 1998). Its distribution range includes southern South Africa, the east coast of Africa, Egypt, the Arabian Sea and expands eastward as far north as Japan. Recently, one species of Cellana was reported in the Atlantic coast of Africa, but it appears to have been introduced from Indo-Pacific regions (Nakano & Espinosa 2010). Cellana is also distributed in oceanic islands of the Indian and Pacific Oceans such as the Bonin Islands, Guam, the Hawaiian islands, Society Island and the Juan Fernandez archipelago in the East Pacific. Finally, its southernmost distribution reaches Australia and New Zealand, as far south as subantarctic islands such as Auckland, Snares, Bounty, Antipodes, Chatham, and Campbell. In the Campbell islands, Cellana strigilis strigilis co-exists with its sister genus Nacella, specifically N. terroris (Powell 1973). The genus Nacella comprises 15 nominal species mainly distributed in four of the described biogeographical provinces of the Southern Ocean (Griffiths et al. 2009): Antarctica, the Magellanic Province and subantarctic islands of the Antipodean and Kerguelenian provinces (Powell 1973). One South American species, Nacella clypeater, expands its distribution north up to Arica (Peruvian) along the Humboldt Current System (Valdovinos & Ru¨th 2005). In the Magellanic Province, Nacella represents one of the dominant groups of marine benthic macroinvertebrates, especially on rocky boulders and rocky shores all along the Magellan Strait and Patagonia (Guzma´n 1978; Rı´os et al. 2003; Bazterrica et al. 2007). According to Powell (1973), and considering that more than 50% of the species have been described from this Province, it has been considered as the centre of origin and diversification of the genus, from whence it expanded east-

ward through the west wind drift. However, this hypothesis has been rejected by recent phylogenetic analyses that showed that the Magellanic species of Nacella are the most derived clade in the genus (Gonza´lezWevar et al. 2010a). Based on morphological studies, at least seven species of Nacella have been described for the Magellan Province, but the taxonomy of the genus in this region is still unclear (Table 1). Powell (1973) recognized five valid taxonomic units: N. deaurata, N. flammea, N. fuegiensis, N. magellanica and N. mytilina, one ecomorph (N. deaurata form delicatissima) and two subspecies (N. magellanica venosa and N. magellanica chiloensis). Recently, Valdovinos & Ru¨th (2005) in a revision of Nacella from southern South America concluded that all the morphological species described in the Magellanic Province, with the exception of N. fuegiensis (considered as synonym of N. magellanica by the authors), are true taxonomic units with diagnostic differences in shell morphology (thickness and coloration), mantle tentacles, foot coloration and radular tooth morphology. The Magellanic species of Nacella also display ecological, bathymetrical and latitudinal distribution differences (Morriconi & Calvo 1993; Morriconi 1999; Rı´os et al. 2003; Bazterrica et al. 2007). The species N. magellanica, N. deaurata, N. delicatissima and N. fuegiensis are commonly found in intertidal rocky environments, while N. flammea and N. mytilina are subtidal (Fig. 1a). Most of the species live attached to rocky substrates where they graze on microalgae, diatoms and bacterial films. Only N. mytilina lives exclusively attached to kelp such as Macrocystis pyrifera (Valdovinos & Ru¨th 2005) and Lessonia flavicans (C.A. Gonza´lez-Wevar, personal observation). In terms of geographical distribution, N. delicatissima, N. magellanica and N. deaurata have a wide distribution, from Chiloe´ Island in the Pacific to the Buenos Aires in the Atlantic, including the Magellan Strait, Tierra del Fuego, Southern Patagonia and the Falkland Islands (Powell 1973; Valdovinos & Ru¨th 2005). Other species such as N. flammea and N. fuegiensis exhibit

Table 1 Taxonomy of Nacella in the Magellanic Province according to four different studies. Filled squares (grey) indicate how the particular study supports synonymies between and among taxa Nominal species

Powell (1973)

Valdovinos & Ru¨th (2005)

de Aranzamendi et al. (2009)

This study

N. N. N. N. N. N. N. N.

N. fuegiensis N. deaurata

N. N. N. N. N. N. N. N.

n.i. N. deaurata Ecotype of N. magellanica or N. deaurata N. magellanica n.i. n.i. N. mytilina n.i.

N. deaurata

fuegiensis deaurata delicatissima magellanica chiloensis venosa mytilina flammea

N. magellanica

N. mytilina N. flammea

n.i., not included in the analyses.  2011 Blackwell Publishing Ltd

magellanica deaurata delicatissima magellanica chiloensis venosa mytilina flammea

n.i. N. magellanica

N. mytilina N. flammea

´ LEZ-WEVAR ET AL. 1938 C . A . G O N Z A (a)

Nacella magellanica 2m 1m

Nacella deaurata

–1 m

Nacella mytilina

Nacella fuegiensis

–5 m

–10 m

Nacella flammea –25 m

(b)

–80º –30º

–75º

–70º

–65º

–60º

–55º –80º –30º

–75º

–70º

–65º

–60º

–55º –80º –30º

–75º

–70º

–65º

–60º

(iii)

(ii)

(i)

–55º –30º

–35º

–35º

–35º

–35º

–40º

–40º

–40º

–40º

–45º

–45º

–45º

–45º

–50º

–50º

–50º

–50º

–55º

–55º –80º

–75º

–70º

–65º

–60º

–55º

–80º

–55º –75º

–70º

–65º

–60º

–55º

–80º

–55º –75º

–70º

–65º

–60º

–55º

Fig. 1 (a) Bathymetrical distributions of the nominal species of Nacella in Punta Santa Ana, Magellan Strait. (b) Latitudinal distribution of the analysed nominal species of Nacella along the Magellan Province based on the descriptions of Powell (1973) and Valdovinos & Ru¨th (2005). (i) N. magellanica, N. deaurata; (ii) N. flammea, N. fuegiensis and N. mytilina; (iii) N. chiloensis and N. venosa.

a narrower distribution, from Ayse´n (4532¢LS; 7204¢LW) to the Magellan Strait (Valdovinos & Ru¨th 2005). N. chiloensis and N. venosa are restricted to Chiloe Island (Powell 1973). Finally, N. mytilina has the widest distribution; it is found in the Magellan Strait, Cape Horn, southern Patagonia, the Falkland Islands and also Kerguelen Island in the Kerguelenian Province (Powell 1973; Fig. 1b). Considering that this species is the only kelp-dweller of the genus, its wide distribution could be attributed to long-distance dispersal by rafting (Donald et al. 2005; Thiel & Gutow 2005; Waters 2007). Recent analyses based on mitochondrial DNA agree with previous morphological and molecular studies,

supporting the monophyly of both Nacella and Cellana and their sister relationship (Gonza´lez-Wevar et al. 2010a). High levels of genetic divergence among Nacella lineages belonging to different Southern Ocean provinces support the existence of transoceanic discontinuities in the genus along its distribution. These genetic discontinuities in the genus have also been detected with nuclear data sets (Actine, 28S rDNA; Gonza´lezWevar 2010). For instance, Magellanic and Kerguelenian subantarctic groups of the genus show high levels of genetic divergence, supporting the existence of physiological or dispersal constraints for Nacella larvae and adults to long-distance dispersion. This result  2011 Blackwell Publishing Ltd

C O N C E R T E D E V O L U T I O N I N M A G E L L A N I C L I M P E T S 1939 contrasts with the high levels of genetic affinity detected in different groups of marine taxa such as Macrocystis (Coyer et al. 2001), Mytilus (Ge´rard et al. 2008), Durvillaea (Fraser et al. 2009, 2010, 2011), Sterechinus (Dı´az et al. 2011) and trochid gastropods (Donald et al. 2005) between distant provinces in the Southern Ocean. Also, high levels of genetic divergence were detected between Nacella species from Antarctica and the Magellanic Province, in contrast to the Antarctic– Magellanic connection described considering species lists in several groups of marine benthic invertebrates (Brandt et al. 1999; Crame 1999; Arntz 2005; Linse et al. 2006; Griffiths et al. 2009). In fact, the closest relative of the Magellanic species of the genus is not the Antarctic limpet N. concinna, but N. clypeater from the Peruvian Province (Gonza´lez-Wevar et al. 2010a). Divergence time estimations using a relaxed Bayesian method suggest that the separation of Nacellidae (Nacella–Cellana) took place after the middle miocene climatic transition (MMCT) 14 Ma (Gonza´lez-Wevar et al. 2010a). This period was characterized by drastic climatic and oceanographic changes in the Southern Ocean (Flower & Kennett 1994; Zachos et al. 2001; Mackensen 2004; Lewis et al. 2008; Verducci et al. 2009). Gonza´lez-Wevar et al. (2010a) proposed that the diversification of Nacella took place at least in two major rounds, long after the separation of the Provinces that they currently inhabit along the Southern Ocean. The first round, at the end of the Miocene (between 9.0 and 5.0 Ma) originated the main lineages of the genus distributed in the Antarctic and Kerguelenian provinces, as well as the separation of Nacella lineages in South America. The second corresponded to a recent Pleistocene diversification of the genus (2.0 to 0.4 Ma) in the Magellanic Province, where different morphological species showed high levels of genetic identity and an absence of reciprocal monophyly (de Aranzamendi et al. 2009; Gonza´lez-Wevar et al. 2010a). However, previous studies in the genus analysed a limited number of Magellanic nominal species of Nacella and a low number of specimens. Based on these preliminary results, the extremely short branches and the lack of reciprocal monophyly among Magellanic species of Nacella could be explained by two different hypotheses. First, the previously described Nacella species in this Province correspond to a single morphologically variable species with multiple habitat affinities. Considering that these Magellanic morphospecies occur in sympatry and exhibit ecological differences they could even be interpreted as ecotypes. Alternatively, the high number of Nacella species in this Province could reflect a very recent diversification process in this region followed by rapid morphological and ecological differentiation as proposed by Gonza´lezWevar et al. (2010a).  2011 Blackwell Publishing Ltd

To evaluate these hypotheses, we measured the degree of genetic differentiation among seven taxa (N. chiloensis, N. deaurata, N. flammea, N. fuegiensis, N. magellanica, N. mytilina and N. venosa). Based on the mtDNA (cytochrome oxidase subunit I, COI), diversity pattern in sympatric Nacella nominal species, we propose to verify whether the nominal taxa of this genus correspond to a single panmictic genetic unit in the Magellanic Province with a broad distribution and a wide range of habitat preferences or whether they correspond to discrete genetic units. Predictions for these hypotheses are as follows: (i) If Nacella nominal species correspond to different ecotypes of a single phenotypically variable species, we expect to find a lack of genetic differentiation among the analysed units found in sympatry. According to this, the existence of different morphotypes, described as different species, would be a consequence of morphological changes associated with the different habitats where settlement and growth occur. In this case, nominal species described in this region would correspond to morphotypes or ecotypes, and the underlying mechanism would be phenotypic plasticity. Such a situation has been described in the Antarctic limpet N. concinna, where two very different morphotypes (a shallow water form and a deeper water one) previously described as different species belong to the same genetic pool (Beaumont & Wei 1991; Hoffman et al. 2010; Gonza´lez-Wevar et al. 2011; but see de Aranzamendi et al. 2008). (ii) Alternatively, if sympatric morphotypes associated with nominal species correspond to different reproductive units, we expect to detect significant genetic differentiation among them, even if they originated recently on a mutational time scale. In this scenario, related species should share the most common haplotypes but exhibit differences in their frequencies because of genetic drift. In this case, the different genetic units found in sympatry could correspond to true biological species, and the underlying mechanism for morphological variation would be adaptive processes associated with niche differentiation. We also performed geometric morphometric comparisons in the analysed specimens to determine the degree of differentiation among them to compare these results with the genetic information.

Materials and methods Sampling and specimen identification Magellanic species of Nacella were identified based on shell morphology and diagnostic external characters following Powell (1973) and Valdovinos & Ru¨th (2005). We included in the analyses Nacella nominal species that clearly constitute different units such as N. deaurata,

´ LEZ-WEVAR ET AL. 1940 C . A . G O N Z A N. flammea, N. magellanica and N. mytilina. We also incorporated N. chiloensis, N. fuegiensis and N. venosa, whose taxonomic status is still unclear (Valdovinos & Ru¨th 2005). To avoid genetic variation because of geographical distribution, we obtained samples of N. deaurata, N. flammea, N. fuegiensis, N. magellanica and N. mytilina from intertidal and subtidal zones of the same locality, Punta Santa Ana (5337¢58¢¢LS; 7054¢50¢¢LW), Magellan Strait. We also collected samples of N. chiloensis and N. venosa from Pelluco, Puerto Montt (4128¢51¢¢LS; 7254¢11¢¢LW; Fig. 1) because of its northern distribution in the Magellanic Province. In spite of the distribution ranges specified by Powell (1973), none of the other Nacella morphospecies were found in this area.

DNA Extraction, PCR amplification and alignment Animals were fixed in ethanol (95%), and total DNA was extracted from mantle tissue using the salting-out protocol (Aljanabi & Martinez 1997). Shells of all the specimens were kept for further geometric morphometric analyses. A partial fragment of the mtDNA gene cytochrome C oxidase subunit I (COI) was amplified using universal primers described by Folmer et al. (1994). PCR amplifications were performed following Gonza´lez-Wevar et al. (2010a); amplicons were purified using QIAquick Gel Extraction Kit (QIAGEN) and sequenced in both directions. All haplotype sequences were deposited in GenBank under the Accession Numbers HQ997162–HQ997364. Sequences were edited with Proseq 2.91 (Filatov 2002) and aligned with Clustal W (Thompson et al. 1992). Using MEGA 5.0 (Kumar et al. 2008), COI sequences were translated into amino acids to check for sequencing errors and ⁄ or the presence of pseudogenes.

Genetic comparisons among Magellanic species of Nacella We determined the levels of polymorphism in the Magellanic morphotypes of Nacella using standard molecular indices such as the haplotype number (k), haplotypic diversity (H), the number of segregating sites (S), average pairwise sequence differences (P) and nucleotide diversity (p) with DnaSP 5.00.07 (Librado & Rozas 2009). For comparison purposes, we also reconstructed the distribution of pairwise differences among Nacella taxa. We estimated the levels of genetic differentiation between the analysed units through mean pairwise differences among groups (NST) and through their haplotype frequencies (GST) in Arlequin v.3.11 (Excoffier et al. 2005). We performed permutation tests (20 000

random iterations) of haplotype identities to confirm the presence of differences among the analysed units. The comparisons of these indices will let us determine whether the described morphospecies of Nacella from Punta Santa Ana and Pelluco are distinct genetic units. We also performed AMOVA analyses in Arlequin to determine the partition that maximizes the differences among groups. We reconstructed genealogical relationships among Nacella units using median-joining haplotype networks with Network 4.5.1.0 (Ro¨hl 2002). This method allows the reconstruction of phylogenies based on intra- and interspecific molecular data such as mitochondrial DNA variation. Haplotype network analyses were carried out individually for each morphospecies, and we also constructed an overall network with all analysed individuals.

Geometric morphometrics analyses Shell-shape variation among Magellanic Nacella morphospecies from Punta Santa Ana and Pelluco was measured using outline analyses based on elliptic Fourier analyses. Outlines were drawn from digital photographs using a two-dimensional projection of the lateral shape of the shells of the same specimens used in the genetic comparisons. Only adult specimens (>4 cm) were included in the analyses. Elliptic Fourier transformations were made using SHAPE software (Iwata & Ukai 2002). Elliptic Fourier descriptors (EFDs) can be used to depict any kind of form and have been effectively applied to the estimation of different shapes in plants and animals (Iwata & Ukai 2002). The method is based on separate Fourier decompositions of the incremental changes of the x and y coordinates as a function of the cumulative length along the outline (Renaud & Michaux 2007). With the module ChainCoder, we extracted the contours of the objects and the relevant information was stored as chain codes. We obtained the normalized EDSs from the chain-coded contour with the module Chc2Nef, and the coefficients of the EFDs were calculated by discrete Fourier transformation following Kuhl & Giardina (1982). These coefficients were normalized, based on the ellipse of the first harmonic, to be invariant with respect to size, rotation and starting point. With the module PrinComp, we performed the principal components analyses on the variance–covariance matrix of the EFD coefficients. Principal components compile the information of the variation contained in the coefficients (Rohlf & Archie 1984) and were estimated using PAST v.1.77 (Hammer et al. 2001). Multivariate analysis of variance (MANOVA) was performed with PAST to evaluate the importance of between-group differentiation relative to within-group  2011 Blackwell Publishing Ltd

C O N C E R T E D E V O L U T I O N I N M A G E L L A N I C L I M P E T S 1941 variation. We performed Hotelling paired comparisons tests to determine the significance of the morphological differences. Finally, to estimate the percentage of correctly re-assigned individuals in each of the analysed nominal species, we used the information contained in the principal components of shell morphology in a multivariate discriminant analyses with Statistica V.7.0.61.0 (StatSoft Inc 2004).

Results Molecular analyses The COI sequence data set of Magellanic Nacella nominal species included 658 bp amplified from 208 individuals. We included 163 individuals belonging to N. deaurata (n = 25), N. fuegiensis (n = 28), N. magellanica (n = 25), N. flammea (n = 31) and N. mytilina (n = 54) from Punta Santa Ana, Magellan Strait, and 45 specimens of N. chiloensis (n = 23) and N. venosa (n = 22) from Pelluco, Puerto Montt (Table 2). As expected with coding regions, no indels or stop codons were detected in the complete data set, and we detected one substitution, from leucine (L) to phenylalanine (F), both nonpolar and neutral. This change corresponded to a transition from C to T in the third position of codon 85. Sequences were A–T rich (60.5%) compared to the C–G content (39.5%), as observed in other molec-

ular studies in Nacella (de Aranzamendi et al. 2009; Gonza´lez-Wevar et al. 2010a; Gonza´lez-Wevar et al. 2010b). Results obtained from the comparisons between Magellanic nominal species of Nacella exhibited high levels of genetic similarity. Only 50 characters were variable (7.4%), and 26 were parsimony informative (3.8%). Genetic diversity indices such as the number of polymorphic sites, haplotype and nucleotide diversities were variable among morphospecies (Table 2). Haplotype diversity varied among the analysed units with N. mytilina exhibiting the lowest value (0.392), while N. flammea showed the highest one (0.892; Table 2). The number of haplotypes and polymorphic sites also varied among species with N. mytilina showing the lowest values even when we analysed the double number of specimens in this nominal species compared to the other ones. Finally, the average number of nucleotide differences (P) and nucleotidic diversity (p) were low and variable between morphospecies (Table 2). GST and NST paired comparisons between the different morphological units from Punta Santa Ana showed significant differences (P < 0.001, Table 3), except for the comparison between N. deaurata and N. fuegiensis. In samples from Pelluco, no significant genetic differences were found between N. venosa and N. chiloensis. However, both nominal species showed significant genetic differences with all other units from Punta

Table 2 Number of individuals per nominal species and their respective diversity indices based on mtDNA (COI) sequences Morphospecies

n

k

H

S

P

p

n MG

N. N. N. N. N. N. N.

23 25 31 28 25 54 22

12 10 15 7 12 8 11

0.814 0.780 0.892 0.611 0.879 0.392 0.827

11 12 17 10 17 8 11

1.336 2.980 2.963 2.794 2.374 0.559 1.459

0.0019 0.0047 0.0044 0.0041 0.0035 0.0008 0.0021

22 25 30 25 22 30 20

chiloensis deaurata flammea fuegiensis magellanica mytilina venosa

n, analysed specimens; k, haplotype number; H, haplotypic diversity; S, polymorphic sites; P, average nucleotide differences; p, nucleotidic diversity; n MG, number of individuals included in the geometric morphometric analyses.

Table 3 GST (below the diagonal) and NST (above the diagonal) pairwise comparisons among the analysed morphospecies of Nacella. 20 000 iterations. Statistically significant differences are marked in bold Morphospecies 1. 2. 3. 4. 5. 6. 7.

N. N. N. N. N. N. N.

chiloensis deaurata flammea fuegiensis magellanica mytilina venosa

1

2

3

4

5

6

7

— 0.171 0.118 0.290 0.011 0.438 0.035

0.205 — 0.145 0.003 0.124 0.449 0.166

0.320 0.173 — 0.231 0.089 0.378 0.113

0.281 0.022 0.176 — 0.239 0.517 0.285

0.013 0.104 0.228 0.174 — 0.417 0.086

0.778 0.681 0.577 0.694 0.715 — 0.434

0.032 0.203 0.315 0.278 0.016 0.773 —

 2011 Blackwell Publishing Ltd

´ LEZ-WEVAR ET AL. 1942 C . A . G O N Z A AMOVA analyses. A total of fifteen haplotypes were detected in Nacella flammea with a dominant haplotype (H32; 39%) and several haplotypes in low frequency (Fig. 2c). As observed in Table 3, N. mytilina had the lowest levels of genetic diversity with 78% of the individuals sharing the same haplotype (H24; Fig. 2d). Dominant haplotypes in N. flammea (H32; Fig. 2c) and N. mytilina (H24; Fig. 2d) were only present in these nominal species. Nacella nominal species can be assigned to three different patterns of pairwise difference distribution, further supporting the high levels of genetic affinities among these units (Fig. 2). First, an L-shaped distribution characterized N. mytilina (Fig. 2d), as expected in a star-like genealogy. Second, a unimodal pattern of distribution characterized N. flammea (Fig. 2c), N. magellanica (Fig. 2e), N. chiloensis (Fig. 2f) and N. venosa (Fig. 2g). Finally, a bimodal distribution distinguished the nominal species N. deaurata (Fig. 2a) and N. fuegiensis (Fig. 2b). The overall median-joining network resulted in a complex genealogy representing the high levels of genetic similarity among them (Fig. 3). We recovered 53 haplotypes; at least three of them (H1, H2 and H24) at high frequency (14.8%, 17.2% and 20.3%, respectively). H1 was present in four morphospecies (N. deaurata, N. fuegiensis, N. magellanica and N. flammea), while H2

Santa Ana, with the exception of their putative species, N. magellanica (Powell 1973). AMOVA analyses suggest that the partition best explaining the variance among groups included the following relationships: group (i) N. chiloensis—N. magellanica—N. venosa; group (ii) N. deaurata—N. fuegiensis; group (iii) N. flammea; and group (iv) N. mytilina, corroborating GST and NST paired comparisons. In this partition, variation among groups explained more than 30% of the genetic variance, while variation among morphospecies within groups accounted for