Introduction

The Lucilia group was first recognized by Merxmüller et al. (1977) as a subgroup of the tribe Inuleae, subtribe Gnaphaliinae. It was defined by its dorsally pubescent style branches and consisted of 12 genera, Belloa, Berroa, Chevreulia, Facelis, Leucopholis, Lucilia, probably Gnaphaliothamnus and Raouliopsis, and the dioecious or subdioecious genera Chionolaena, Luciliopsis, Mniodes and Oligandra (Table 1). A few years later, Robinson (1985) considered Luciliopsis to be a synonym of Facelis, placing two of its three species in the new genus Cuatrecasasiella. Anderberg and Freire (1990a) later showed that the type of the genus Luciliopsis belongs in Chaetanthera of the Mutisieae. Two genera were added to this group, Jalcophila (Dillon and Sagastegui-Alva 1986) and Novenia (Freire 1986a). Oligandra was stated later by Freire (1989) to be a synonym of Lucilia, whereas Leucopholis was a synonym of Chionolaena (Anderberg 1991; Freire 1993). Based on a morphological cladistic analysis, Anderberg (1991) established five subtribes for the tribe Gnaphalieae: Angianthinae Benth., Relhaniinae Less., Cassiniinae Anderb., Gnaphaliinae, and Loricariinae Anderb. That study indicated that the Lucilia group, as earlier understood, was a highly polyphyletic assemblage and also excluded Novenia from the Gnaphalieae. The remaining genera were placed in three clades, one composed of Gnaphaliothamnus and Chionolaena (subtribe Cassiniinae), the second including Loricaria, Mniodes and Raouliopsis (subtribe Loricariinae) and the third one, Lucilia group s. str. (subtribe Gnaphaliinae), was composed of Belloa, Berroa, Chevreulia, Cuatrecasasiella, Facelis, Lucilia and Jalcophila. Subsequently, Anderberg and Freire (1991) added two further taxa to the Lucilia group s. str., viz. Gamochaetopsis and Luciliocline. Both were segregated from Lucilia which was shown to be non-monophyletic. Anderberg and Freire (1991) recognized a Lucilia group s. str., comprising nine genera (Table 1) with the following traits: very slender pappus bristles basally connate, dorsally pubescent style branches, purple tipped florets and polychromous phyllaries. Later, Dillon (2003), based on a previous cladistic analysis of morphological characters (Dillon 2000) revealed the need for a new circumscription for the group, suggesting a close potential relationship between Chionolaena, Gamochaeta, Micropsis, Stuckertiella and the other members of the Lucilia group (Table 1).

Table 1 Comparison of the genera composition of Lucilia group and its genera (in bold face taxa with different placement) through the treatments
Full size table

Most of the genera included in the Lucilia group have been monographed since they were described, i.e., Berroa, Facelis, Micropsis and Stuckertiella (Beauverd 1913a, b, c, d); Loricaria and Mniodes (Cuatrecasas 1954a, b); Belloa (Cabrera 1958); Lucilia (Freire 1986b); Jalcophila (Dillon and Sagástegui-Alva 1986); Chionolaena (Freire 1993; Nesom 2001; Loeuille et al. 2011); Luciliocline (Anderberg and Freire 1991); Gamochaeta (Nesom 1990a; Freire and Iharlegui 1997), but most of their species have been placed in two or more different genera (Table 1). The delimitation of the genus Lucilia is probably the most controversial, and the limits between Lucilia, Belloa and Luciliocline have been discussed in recent years. As an example, Rémy (1847) created Belloa on the basis of Lucilia chilensis. Cabrera (1958) expanded the monotypic genus Belloa to 11 species. Dillon and Sagástegui-Alva (1985) described three new species of Belloa and transferred another four to this genus (one from Mniodes and three from Lucilia), raising the number of Belloa species to 18 (Dillon and Sagástegui-Alva 1991) or 16 (Dillon and Sagástegui-Alva 1992). Freire (1986b, 1987) merged Belloa with Lucilia. Anderberg and Freire (1991) maintained Belloa as an independent genus for the species with prostrate stems and proposed the genus Luciliocline for the species with erect or ascending stems. Finally, Dillon (2003) recognized Belloa as a monotypic genus and transferred the remaining species of Belloa to Luciliocline. The evident instability in the classification of the Lucilia group reflects the general scarcity of morphological characters traditionally considered relevant for the classification of the group and their high level of homoplasy. This affects mainly not only the generic boundaries, but also the circumscription of the Lucilia group, and this urged for an investigation based on molecular data to shed light on the classification.

In this context, a phylogenetic analysis including molecular and morphological information could be useful to elucidate the taxonomic arrangement of the taxa mentioned above. Both chloroplast and nuclear DNA markers have been used previously to address the phylogenetic relationships in the tribe Gnaphalieae, mainly focused on African and Australian taxa (Bergh and Linder 2009; Bayer et al. 2000, 2002; Galbany-Casals et al. 2010; Smissen et al. 2011), but molecular information on South American groups remains unknown or poorly studied (Ward et al. 2009).

The aims of this work were, using two plastid (rpl32-trnL and trnL-F) and two nuclear DNA regions (ITS and ETS), to test the various hypotheses of relationships within the Lucilia group, using a broad sampling of the Gnaphalieae.

Materials and methods

Taxon sampling

A special effort was made to cover the morphological and geographical variation of Lucilia group and allied genera in South America. The sampling for plastid and nuclear sequencing includes material from Argentina, Bolivia, Brazil, Chile, Colombia and Ecuador. The analyses comprised a total of 56 species, 21 belonging to the Lucilia group sensu Anderberg and Freire (1991), representing 51 % of the group. At least one species each of the 9 genera included in this group were used in the analyses (Appendix 1 in ESM). In order to assess the placement of the Lucilia group members within the tribe, 35 species belonging to closely related genera were used as outgroups (Appendix 1 in ESM). Trees were rooted with Relhania as outgroup.

Using DNA obtained from plants collected in wild populations and dried in silica gel, or from herbarium material, a total of 117 sequences (representing 36 species) were generated for this study, while 64 sequences (for 20 species from the outgroups) were obtained from GenBank (see Appendix 1 in ESM). A complete list of the specimens used in this study, the locality where each specimen was collected, the herbarium voucher information and GenBank accession numbers are shown in the Appendix 1 in ESM.

Morphological characters

For the morphological matrix, 21 ingroup taxa and 12 outgroup taxa (Achyrocline ramosissima, A. tomentosa, Antennaria dioica, Chionolaena arbuscula, C. campestris, Gamochaeta serpyllifolia, Loricaria colombiana, Micropsis dasycarpa, Mniodes andina, Pseudognaphalium badium, P. lacteum, Stuckertiella capitata), were analyzed using 35 morphological characters taken from a previous work (Anderberg and Freire 1991) and adding the following character, i.e., leaves not densely imbrıcate (0)/densely imbricate (1) (Appendix 2 in ESM). Characters were taken from direct examination of herbarium specimens (Appendix 1 in ESM), and information was completed from literature (Anderberg and Freire 1991).

Molecular methods

Total genomic DNA from silica-dried material was extracted using a modified CTAB protocol from Doyle and Dickson (1987), while exactions from herbarium material were performed using the DNeasy Plant Mini Kit (QIAGEN Inc., Hilden, Germany). Genomic DNA was used as a template to amplify four regions: the chloroplast fragment containing the trnL (UAA) intron and the intergenic spacer between the trnL (UAA) 3′exon and the trnF (GAA) gene (trnL-F region) using primers C and F from Taberlet et al. (1991); the rpl32-trnL intergenic spacer (chloroplast) using primers rpl32F and trnL(UAG) from Shaw et al. (2007); the nuclear region ETS using the reverse primer 18S-ETS (Baldwin and Markos 1998) and one of the following forward primers: AST1, ETS1 or ETS2 (Bayer et al. 2002); and the nuclear marker ITS using the primers ITS4 and ITS5 from White et al. (1990).

Polymerase chain reactions (PCR) were carried out on 25 μl volumes, each reaction containing between 50 and 100 ng of DNA, 1.5 U of Taq polymerase (Invitrogen Life Technologies, São Paulo, Brazil), 1× PCR buffer, 5 mM MgCl2, 0.2 mM of each primer and 0.025 mM dNTP each. In species for which these protocols were unsuccessful, 0.4 % bovine serum albumin was included as additive to increase the yield of PCR reactions. The annealing temperatures ranged between 48 and 52 °C for the chloroplast markers and 56–60 °C for the nuclear markers. PCR products were electrophoresed on a 1 % TBE agarose gel stained with SYBR safe DNA gel stain (Invitrogen, Brazil). Macrogen Inc. (Korea) performed the cleaning of the PCR products using the Montage PCR purification kit from Millipore following the manufacturer’s protocol. Sequencing reactions were also performed by Macrogen Inc.

Sequence analyses

Sequence data were edited and contigs assembled using the program Chromas Pro ver. 1.34 (Technelysium Pty, Ltd, Tewantin, Australia), and the matrix was edited using the software BioEdit (Hall 1999). Sequences were aligned using the program MAFFT (http://mafft.cbrc.jp/alignment/server/; Katoh et al. 2002), using the default parameters and an automatic alignment strategy, with subsequent visual inspection and manual revision. Data matrices are deposited on TreeBase (accession number 16182).

Phylogenetic analyses

Three maximum parsimony analyses were conducted, the first including only the nuclear markers, the second including the complete molecular data set and the third combining the molecular information with the morphological data in a combined analysis (Nixon and Carpenter 1996). Searches were carried out using the software TNT (Goloboff et al. 2008), with the characters equally weighted and considering gaps as missing data. The heuristic searches were performed as follows: 1,000 series of random addition sequences (RAS), swapping the trees with tree bisection-reconnection (TBR), plus an additional rearrangement of all the most parsimonious trees found. Branch support was evaluated using Jackknifing (JK; Farris et al. 1996), which was calculated by performing 5,000 pseudoreplicates, each consisting of 10 RAS.

Bayesian inference of phylogeny with posterior probabilities (PP) was conducted with MrBayes ver. 3.2.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003) for the three combined analyses mentioned above. The models of molecular evolution were selected using the Akaike Information Criterion (AIC) as implemented in jModelTest ver. 2.1.4 (Table 2; Guindon and Gascuel 2003; Darriba et al. 2012). The chosen models were GTR + G for rpl32-trnL, trnL-F, ETS and GTR + I + G for ITS. The standard discrete model was used for the morphological data (Lewis 2001). Searches used default settings for all parameters. Four simultaneous independent runs initiated from random start trees were run for 10 million generations, sampling from the posterior distribution of trees every 1,000 generations (for a total of 10,000 samples). Several strategies were employed to confirm that chains had achieved stationarity and the appropriate burn-in for each analysis was determined: (1) we plotted overall—lnL, (2) we examined the standard deviation of split frequencies, (3) we examined the potential scale reduction factor (PSRF) and (4) we checked graphically the convergence of the MCMC by monitoring the cumulative posterior split probabilities and among-run variability of split frequencies using the online tool AWTY (Wilgenbusch et al. 2004). A total of 25 % of the samples were discarded as burn-in.

Table 2 Main characteristics and results for the different regions and datasets analyzed
Full size table

Bayesian posterior probabilities (PP) values were described as high (0.95–1.00), moderate (0.90–0.94) and low (0.50–0.89), while parsimony jackknife (JK) values were described as high, (90–100 %), moderate (70–89 %) and low (≤69 %). Jackknife values of less than 50 %, though extremely low, were also considered since in some cases they can be showing just lack of information instead of inconsistencies in the phylogenetic signal (Goloboff et al. 2003). Nodes with Bayesian posterior probabilities of less than 0.5 were considered as unsupported and the values are not reported.

Results

In total we included in the analyses 54 ETS sequences, of which 35 are new; 45 ITS sequences, of which 28 are new; 39 rpl32-trnL, intergenic spacer sequences, of which 28 are new; and 43 trnL intron intergeneric spacer sequences, from which 26 are new (Appendix 1 in ESM). The main characteristics of the datasets, along with the corresponding tree statistics are summarized in Table 2.

Both Parsimony and Bayesian inference analyses showed highly congruent topologies for molecular data or combination of molecular data and morphological characters. All the analyses showed short branches for the Lucilia group, mainly in the nodes (Figs. 1, 2, 3).

Fig. 1
figure 1

Consensus phylogram obtained from the Bayesian analysis of ITS + ETS sequences (nuclear markers). Bayesian posterior probabilities (PP) and Jackknife values from the parsimony analyses are shown associated with the branches. An asterisk indicates a branch missing from the parsimony analysis. Dotted branches were shortened to have a clearer topology

Full size image
Fig. 2
figure 2

Consensus phylogram obtained from the Bayesian analysis of the complete molecular dataset (ITS + ETS + rpl32-trnL intergenic spacer + trnL intron + trnL-F intergenic spacer sequences). Bayesian posterior probabilities (PP) and Jackknife values from the parsimony analyses are shown associated with the branches. An asterisk indicates maximum support for both analyses (i.e., >0.99 and 98 %). Dotted branches were shortened to have a clearer topology

Full size image
Fig. 3
figure 3

Consensus phylogram obtained from the bayesian combined analysis (molecular data set + morphology). Bayesian posterior probabilities (PP) and Jackknife values from the parsimony analyses are shown associated with the branches. An asterisk indicates a branch missing from the parsimony analysis. Resolution of clade C1 in the parsimony analysis. Dotted branches were shortened to have a clearer topology

Full size image

The analysis of the plastid markers resulted in a highly polytomized tree, with almost no support and no resolution below the node that groups the FLAG clade (tree not shown), so they are not described here.

Nuclear markers (ITS+ETS)

The nuclear markers combined showed that the Lucilia group as previously circumscribed is not monophyletic, due to the inclusion of Antennaria, Gamochaeta, Stuckertiella, Micropsis, Loricaria, Chionolaena and Mniodes (Fig. 1) in a high supported clade for the Bayesian analysis (0.99 PP, 3 % JK). Most of the basal nodes (from the outgroups) were lost in the parsimony analysis, having a tree with more polytomies as a result of that analysis (Fig. 1). However, inside the Lucilia group s.l. both analyses showed almost the same topology (with the exception of the placement of the Antennaria clade, and the inner structure of the Belloa + Luciliochline  +  Mniodes clade) (Fig. 1). Since the clades of the Lucilia group s.l. are also present in the other analyses, they will be discussed in detail below, in the combined analysis.

Complete molecular data set (trnL-F + rpl32-trnL + ITS + ETS)

The analysis of the four molecular markers (Fig. 2) recovered the same topology provided by the nuclear analysis, with two exceptions: Micropsis is grouped together with two species of Facelis (vs. with three species of Lucilia in the nuclear analysis; Fig.1) and Luciliocline subspicata is placed in a basal polytomy of the Belloa pp + the remaining two species of Luciliocline+Mniodes clade (Fig. 2) (vs. grouped with the remaining two species of Luciliocline in nuclear DNA analysis; Fig. 1). The major difference between the parsimony and the Bayesian analyses is that the first fails to recover the Lucilia group s.l. as monophyletic (Fig. 2), but instead places the major clades that conform it as a polytomy (not shown). Since the polytomy does not contradict the topology obtained in the Bayesian analysis, we consider the results of this analysis as valid.

Combined analysis (molecular data + morphological characters)

Considering both the Parsimony and the Bayesian combined analyses, the Bayesian inference gave the highest resolution. Therefore, we adopt the hypothesis obtained from Bayesian analysis for description and discussion of the results. This topology is shown with the addition of PP/JK values (Fig. 3).

From the topology inferred by Bayesian analysis, and following the Relhania and Oedera clades, four clades were recovered. One, clade was constituted by six genera, Vellereophyton, Cassinia, Acanthocladium, Edmondia, Stuartina, and Ewartia (clade 1, 0.84 PP); a second one (clade 2, 1.00 PP, 100 % JK), was constituted by the genera Pseudognaphalium (P. badium. P. lacteum), Helichrysum (H. stoechas), Anaphalis (A. nepalensis, A. margaritacea), and Achyrocline (A. tomentosa, A. ramosissima); the third one (clade 3, 0.91 PP) was constituted by Plecostachys serpyllifolia and Gnaphalium uliginosum, and the forth clade (clade 4, 0.96 PP, 70/JK) comprised the rest of the genera included in the analyses.

One major clade can be identified within clade 4, which is missing in the parsimony analysis (Fig. 3): clade C (1.00 PP) subdivided into clade C.1 (1.00 PP, 100 % JK) including Antennaria chilensis and A. dioica; clade C.2 (1.00 PP, 99 % JK) including Chevreulia acuminata and Cuatrecasasiella argentina; clade C.3 (1.00 PP, 92 % JK) with two subclades, C.3a (0.98 PP; 94 % JK) including Jalcophila boliviensis and J. ecuadoriensis; and subclade C.3b (1.00 PP, 95 % JK) including Loricaria colombiana, Belloa plicatifolia, B. schultzii, B. longifolia, B. piptolepis, B. kunthiana, Mniodes andina, Luciliocline subspicata, L. santanica, and L. lopezmirandae. Finally, clade C.4 (1.00 PP, 87 % JK) which could be subdivided into subclades, subclade C.4a (0.75 PP, 78 % JK) including Belloa chilensis, Chionolaena arbuscula, C. campestris; subclade C.4b (1.00 PP, 90 % JK) including Lucilia linearifolia, Berroa gnaphalioides, Lucilia lycopodioides, L. nitens, L. acutifolia, Facelis retusa, F. plumosa, and Micropsis dasycarpa; and subclade C.4c (1.00 PP, 94 % JK) including Gamochaeta longipedicellata, G. americana, Stuckertiella capitata, Gamochaetopsis alpina, and Gamochaeta serpyllifolia.

Two minor clades can be also identified within clade 4 (Fig. 3): clade A (1.00 PP, 100 % JK) and clade B (0.83 PP) composed of the genera Leontopodium, and Logfia  +  Filago, respectively. These two clades were recovered as sister clades of clade C.

Discussion

The present analyses of Gnaphalieae are consistent with previous phylogenies (Bergh and Linder 2009; Ward et al. 2009; Galbany-Casals et al. 2010; Smissen et al. 2011) showing a relatively small southern African clade, i.e., ‘Relhania and Oedera clades’ at the base and then a ‘crown radiation’ including African and non-African taxa. Within the ‘crown radiation’ two of the four clades recovered are also equivalent to those found in previous phylogenies (Bergh and Linder 2009; Ward et al. 2009; Galbany-Casals et al. 2010; Smissen et al. 2011), i.e., clade 2 formed by Helichrysum, Anaphalis and Pseudognaphalium, called HAP clade (for the first letters of the genera, according to Smissen et al. 2011) which showed the maximum statistical support (Figs. 2, 3) and the highly supported clade 4, formed by Filago, Logfia, Leontopodium, Antennaria, Gamochaeta, called FLAG clade (for the first letters of the large genera, according to Galbany-Casals et al. 2010) (Figs. 2, 3), principally distributed in Eurasia and America.

Clade 3 comprises Gnaphalium uliginosum and Plecostachys serpyllifolia (with moderate support in the combined analyses; Fig. 3), both belonging to subtribe Gnaphaliinae (Anderberg 1991). Unlike Bergh and Linder’s (2009) phylogeny, Plecostachys serpyllifolia, from South Africa, was not grouped within the FLAG clade.

Clade 1 comprises six taxa (with low support in the combined analyses; Fig. 3): Vellereophyton dealbatum (subtribe Gnaphaliinae-Gnaphalium group; Anderberg 1991), from South Africa, as sister species to two subclades: the first one was composed of Edmondia sesamoides (subtribe Gnaphaliinae-Syncarpha group; Anderberg 1991) from South Africa, and two Australian species, Ewartia catipes (subtribe Casssiniinae-Anaphalis group; Anderberg 1991), and Stuartina muelleri (subtribe Gnaphaliinae -Gnaphalium group; Anderberg 1991); and a second subclade was composed of two other Australian species, Acanthocladium dockeri (subtribe Cassiniinae; Anderberg 1991) and Cassinia longifolia (subtribe Casssiniinae-Cassinia group; Anderberg 1991).

From these four clades, only clade 2 (HAP clade) is recovered in the phylogeny of Gnaphaliinae by Anderberg (1991) The remaining three clades show, as was previously noted (Bayer et al. 2000, 2002; Ward et al. 2009; Galbany-Casals 2010; Smissen et al. 2011), with the necessary caution due to the limited sampling of our study, that DNA data do not support some of the relationships among groups of genera and subtribes identified by Anderberg (1991), for example in clade 4 (FLAG clade), Leontopodium and Logfia + Filago are related to Antennaria (instead of Leontopodium and Filago placed in the subtribe Gnaphaliinae, and Antennaria placed in the subtribe Cassiniinae), or in clade 1, Acanthocladium and Cassinia are related to Edmondia and Stuartina (instead of Acanthocladium and Cassinia placed in subtribe Cassiniinae, and Edmondia together with Stuartina in subtribe Gnaphaliinae, but in different groups, i.e., Syncarpha group and Gnaphalium group, respectively).

In general, South American taxa of Gnaphalieae have not been sampled in previous molecular phylogenetic analyses. Only Ward et al. (2009) included Jalcophila and Mniodes in their phylogeny of Gnaphalieae, in which Jalcophila (J. peruviana), Mniodes (M. andina) and Gamochaeta (G. pensylvanica) are all part of the ‘crown radiation’.

Generic composition of the Lucilia group s. lat.

The analysis of the complete molecular data set, and molecular data combined with morphological characters showed that the Lucilia group, as previously circumscribed, is not monophyletic. The combined analysis showed that the genera of the Lucilia group sensu Anderberg and Freire are grouped with seven other genera (clade C, Fig. 3): Gamochaeta is distributed in America with some species known as cosmopolitan weeds, Stuckertiella, Micropsis, Loricaria and Mniodes from South America, Chionolaena found in Central and South America, and Antennaria mainly distributed in Northern Hemisphere with only three species known from South America. Four of these genera were previously placed by Anderberg (1991) in subtribe Loricariinae (Loricaria and Mniodes) and subtribe Casssiniinae (Antennaria and Chionolaena). The relationship between these last two taxa of the tribe Cassiniinae and other members of the Lucilia group was previously suggested by Dillon (2003).

Anderberg and Freire (1991) diagnosed the South American Lucilia group by its very slender pappus bristles which are often connate in groups, dorsally pubescent style branches and polychromous involucral bracts. Only one morphological synapomorphy, i.e., slender pappus bristles basally connate or coherent at the base by patent cilia, was found for the ancestral node that defines the present arrangement of Lucilia group s. lat. Although, only two of c. 50 species of Antennaria were sampled in this study, and it possesses pappus bristles basally with patent cilia (vs. connate in the remaining taxa of Lucilia group), this genus was tentatively included here in this group.

Within these mainly American members, called Lucilia group s. lat., two main groups were recovered with our sampling. The first one (clade C2 + C3, Fig. 3), comprises Chevreulia, Cuatrecasasiella, Jalcophila, Loricaria, Belloa pro parte, Mniodes and Luciliocline, principally characterized by having caespitose, subcaespitose or cushion forming habit (rarely shrubs), with leaves arranged in rosettes or densely imbricate, and achenial trichomes globose or clavate (rarely elongated) (Table 3; Figs. 4, 6g). The second one (clade C4, Fig. 3) comprised Belloa chilensis, Chionolaena, Lucilia, Berroa, Facelis, Micropsis, Gamochaeta, Stuckertiella and Gamochaetopsis, principally characterized by having ascending or erect stems (rarely prostrate), with leaves commonly scattered, and achenial trichomes elongated (Table 3; Figs. 5, 6i). They also conform, each one, two geographical well-defined groups. In fact, the genera of clade C4 occur mainly at low altitude areas in southern Brazil, Uruguay, eastern Paraguay and northern Argentina, but except Berroa and Micropsis, they also include species from some part of the Andes. Conversely, most genera of the clade C2+C3 are endemic to the Andes; only Chevreulia includes species from low altitudes. A minor third group (clade C1, Fig. 3) was also recovered with our sampling which comprises the genus Antennaria, from artic to temperate regions of the Northern hemisphere with only three species in the Andes of South America, and principally characterized by having dioecious and often apomictic habit, with ascending or erect stems, leaves alternate, often rosulate and achenial trichomes usually clavate.The genus Mniodes was also considered by Merxmüller et al. (1977) to be a member of the Lucilia group. However, Anderberg (1991) placed Mniodes with Loricaria (Fig. 4b) in the subtribe Loricariinae because of its dioecy habit. Our results show that both genera are nested within clade C of the Lucilia group (Figs. 1, 2, 3).

Table 3 Morphological characters of the genera of Lucilia group s. lat. (key characters in bold face)
Full size table
Fig. 4
figure 4

Morphological diversity in the Lucilia group s. lat.-clade C2 and C3: a Jalcophila boliviensis (photo S. Beck), b Loricaria colombiana (photo J.C. Ospina), c Mniodes piptolepis (photo F. O. Zuloaga), d Mniodes kunthiana (photo E. Urtubey), e Mniodes schultzii (photo R.I. Meneses), f Mniodes subspicata (photo E. Urtubey), g Mniodes santanica (photo S. Beck), h Cuatrecasasiella argentina (photo S. Beck) and i Chevreulia sarmentosa (photo M. Bonifacino)

Full size image
Fig. 5
figure 5

Morphological diversity in the Lucilia group s. lat.-clade C4: a Berroa gnaphalioides (photo M. Bonifacino), b Lucilia acutifolia (photo M. Bonifacino), c Facelis retusa (photo M. Bonifacino), d Micropsis dasycarpa (photo E. Urtubey), e Belloa chilensis (photo M. Baeza), f Chionolaena costaricensis (photo M. Bonifacino), g Gamochaetopsis alpina (photo N. Bayón), h Gamochaeta longipedicellata (photo S. Beck) and i Stuckertiella capitata (photo S. Beck)

Full size image
Fig. 6
figure 6

Micrographs of some members of the Lucilia group s. lat.: a–e style branches of disc florets, a Cuatrecasasiella argentina (Urtubey et al. 474-SI), b Gamochaeta neuquensis, (Ezcurra 120-SI), c Gamochaetopsis alpina (Cabrera 5902-LP), d Lucilia nitens, (Urtubey 390-SI), e Mniodes lopezmirandae (Urtubey and Freire 433-SI); f achene with elongated hairs (Belloa chilensis, Ezcurra 2262-BCRU), g achene with globose hairs (Mniodes lopezmirandae, Urtubey and Freire 433-SI), h clavate achenial hairs (Gamochaetopsis alpina, Cabrera 5902-LP), i elongated achenial hairs (Belloa chilensis, Ezcurra 2262-BCRU) and j globose achenial hairs (Gamochaeta grazielae, Deble and Deble 4501-SI). Scale bars a–d, g, i 50 µm, e 100 µm, f 200 µm, h, j 20 µm

Full size image

The close relationship between Facelis and Micropsis (Fig. 5c, d) had been previously recognized by Dillon (2003). Micropsis, which is a genus of five species restricted to South America, was defined by paleaceous receptacles where paleae enclose the florets, and consequently associated with the genera of the Filago group by Anderberg (1991). As such, the strong support in the combined analysis (Fig. 3) for the clade of Micropsis + Facelis confirms the inclusion of Micropsis in the Lucilia group.

The genus Chionolaena (Fig. 5f) was considered by Merxmüller et al. (1977) to be a member of the Lucilia group. Later, Anderberg (1991) placed Chionolaena (including Leucopholis), in the Anaphalis group (subtribe Cassiniinae), principally diagnosed by having disc florets functionally male. Subsequently, Dillon (2003) suggested a close potential relationship between Chionolaena and the other members of the Lucilia group. Our results are consistent with those of Merxmüller et al. and Dillon’s placement of genus Chionolaena within the Lucilia group (Fig. 3).

Anderberg and Freire (1991) considered Gamochaeta to be the sister taxon of the Lucilia group. Gamochaeta (Fig. 5h) was not considered by these authors as belonging to the Lucilia group, since it possesses truncate style branches apically pilose (vs. style branches dorsally pilose, Fig. 6a–e). Dillon (2003) suggested a relationship between Gamochaeta and other members of the Lucilia group, which was confirmed by our results, which show that Gamochaeta and its sister genus Stuckertiella (Fig. 5i) are nested with Gamochaetopsis (Fig. 5g) within clade C (Fig. 3).

In our analyses including molecular markers and a combination of molecular and morphological data, some main clades correspond to those recognized in previous morphological phylogenies. As such, for discussion of systematic relationships of Facelis, Chevreulia, Cuatrecasasiella and Jalcophila we refer to Anderberg and Freire (1991) with which our phylogenetic inferences are congruent. Gnaphaliothamnus, which was considered as a synonym of Chionolaena (Nesom 2001; Loeuille et al. 2011) or as a distinct genus (Nesom 1990b, c; Anderberg 1991; Freire 1993), and Raouliopsis, have not been included in our analyses, so their phylogenetic positions remain untested.

Placement of the Lucilia group s. lat.

The position of Lucilia group s. lat., as defined in the present study, is the same in all trees, with Leontopodium, and Filago within the clade 4, i.e., the FLAG clade (Figs. 1, 2, 3). Smissen et al. (2011) hypothesized 2n = 14 as the base number for the tribe Gnaphalieae and a 2n = 28 ancestor for the FLAG clade (derived from 2n = 14 by polyploidy), showing multiple shifts from 2n = 14 to 2n = 28 in the tribe. In this sense, the counts of 2n = 28, 56, 62, 63, 70, 80, 84, 85, 100 in Antennaria, Chouksanova et al. 1968; Juel 1900; Jörgensen et al. 1958; Löve and Löve 1982a, b; Packer and McPherson 1974; Urbanska-Worytkiewicz 1967; n = 14 in Chionolaena lavandulifolia (=Gnaphalium lavandulifolium), De Jong and Longpre 1963; n = 14 in Gamochaeta spiciformis, Moore 1981; n = ca. 14 in Loricaria thuyoides Turner et al. 1967; 2n = 28 in Lucilia acutifolia and L. nitens, Freire 1986c; n = 12 in Luciliocline subspicata (=Belloa punae), Fernández Casas and Fernández Piqueras 1981; and n = 11–12 in Stuckertiella capitata, Spooner et al. 1995, are consistent with Smissen et al. hypothesis and provide good evidence that the Lucilia group is part of the FLAG clade.

Taxonomic implications

Monophyly of Belloa and its relationships with Luciliocline and Mniodes

Our results suggest that Belloa sensu Cabrera (1958) and sensu Anderberg and Freire (1991) is not monophyletic. Instead, these results are consistent with Dillon’s study (2003) that considered Belloa as a montotypic genus (B. chilensis). In fact, all the analyses show that the type species of Belloa (B. chilensis) of the Andes in the southern part of the continent is nested in the strongly supported clade C4 (Figs. 1, 2, 3), whereas the remaining sampled species of Belloa (excluding type) distributed in the Andes from Venezuela to north western Argentina were grouped with Jalcophila, Loricaria, Mniodes and Luciliocline in an also strongly supported clade C3 (Figs. 2, 3). The monotypic genus Belloa is here diagnosed by having prostrate stems, caespitose, closely leafy, solitary capitula surrounded by the terminal leaves, tubular corollas and clavate to elongate achenial trichomes (Figs. 5e, 6f, h, i).

Dillon (2003) transferred the remaining species of Belloa (except B. kunthiana) sensu Anderberg and Freire (1991) to Luciliocline and diagnosed this genus by its globose achenial trichomes (Fig. 6g). In this sense, Belloa kunthiana because of its elongated achenial trichomes was considered by this author as a member of the genus Lucilia [Lucilia kunthiana (DC.) Zardini]. However, in contrast to Dillon (2000, 2003), our results here showed Belloa kunthiana placed in the clade together with Belloa pro parte (excluding type) (Figs. 1, 2, 3), and morphologically B. kunthiana is more closely related to the other species of Belloa than to Lucilia because of its prostrate stems with rosulate leaves and leaf hairs with apical, flagellate cell slightly swollen at the base (vs. ascending or erect stems distantly leafy and leaf hairs with apical, flagellate cell bulbose at the base in Lucilia). This suggests that the resemblance of Belloa kunthiana to species of Lucilia regarding its elongated achenial trichomes is due to parallel evolution, which has led to non-monophyletic groups.

All the analyses showed the species of Luciliocline sensu Dillon (but including Belloa kunthiana) as a strong supported clade with Mniodes nested within it and Loricaria as its sister clade (Figs. 1, 2, 3). Mniodes and Loricaria (Fig. 4b–g) have little morphological similarity (i.e., leaves densely imbricate), but both were placed in the subtribe Loricariinae by Anderberg (1991) because of their dioecious habit. Instead, Mniodes is vegetatively similar to the highly reduced Belloa species (Fig. 4d, e), and strong support was recovered for the clade Mniodes andina + Belloa kunthiana in the analysis of the four molecular data sets (Fig. 2) and low support in the Bayesian combined analyses (Fig. 3). The phylogeny of the Lucilia group suggests at least three independent acquisitions of dioecy in Jalcophila clade, i.e., Loricaria, Mniodes, Cuatrecasasiella and other three independent acquisitions of subdioecy (disc-florets functionally male) in the Belloa clade, i.e., Chionolaena, Lucilia (Freire 1989) and Stuckertiella (Table 3). Since sex separation appears as an evolutionary convergence and may be the consequence of selective pressures particular to Andean environments, the homogamous capitula cannot be a character diagnosing Mniodes.

Although in our combined Bayesian analysis the species of Luciliocline sensu Dillon appears as two subclades, i.e., the subclade Belloa pro parte with the genus Mniodes closely related to Belloa kunthiana, and the subclade Luciliocline, they are moderately and weakly supported, respectively (Fig. 3). In addition, as more specimens were examined, traditional morphological features that characterized each subclade including prostrate stems forming mats, being closely leafy, and solitary capitula (subclade Belloa pro parte + Mniodes), and ascending stems, distantly leafy with capitula arranged in glomerules (subclade Luciliocline) showed overlapping, i.e., ascending stems with capitula in glomerules are present in Belloa piptolepis.

Mniodes is the oldest generic epithet attached to this clade; consequently, here we delimit subclade Belloa pro parte + Mniodes + subclade Luciliocline as Mniodes (A. Gray) Benth. (Figs. 3, 4c–g). Although there is no unequivocal synapomorphy for Mniodes, based on our own observations and literature, the species of this genus are relatively small plants ranging from pulvinate-caespitose with stems tightly compacted and densely leafy to subcaespitose with ascending stems densely leafy at the base, with capitula sessile, globose achenial trichomes only elongated in B. kunthiana and cytologically with a basic haploid number of n = 12.

Placement of Jalcophila boliviensis

Jalcophila includes four species; two of them (J. boliviensis and J. ecuadoriensis) were sampled in this study. The first one, J. boliviensis (Fig. 4a), was recognized as a member of Jalcophila (Anderberg 1991; Anderberg and Freire 1991) but transferred to Gamochaeta by Dillon (2003). Gamochaeta and Jalcophila boliviensis are distantly related to each other in our results (Figs. 1, 2, 3). This treatment is congruent with the observations of the morphological features and previous treatments (Anderberg and Freire 1990b, 1991). Style branches of Jalcophila boliviensis are dorsally papillose unlike those apically pilose in Gamochaeta. Involucral bracts of Jalcophila boliviensis are subequal, whereas in Gamochaeta the outermost bracts are shorter than the inner bracts. The capitula of Jalcophila boliviensis are pedunculate at maturity unlike those usually sessile in Gamochaeta. Jalcophila boliviensis shares these characters, i.e., dorsally papillose style branches, subequal involucral bracts and pedunculate capitula, with J. ecuadoriensis, and both species also have 4-lobed perfect florets. In addition, Jalcophila boliviensis and J. ecuadoriensis were sister taxa in all the analyses, implying that the genus might be monophyletic. The type of Jalcophila was not included in our present analyses so the relationship between J. boliviensis and the generic type remains to be tested.

Monophyly of Lucilia and its relationships with Berroa, Facelis and Micropsis

Lucilia, Berroa, Facelis and Micropsis formed a well-supported clade within the clade C4 (Fig. 3). Facelis characterized by its plumose bristles pappus, scabrid above, and Micropsis characterized by it pappus absent or composed of a short crown, also formed a well-supported clade, and the first was monophyletic. The monotypic genus Berroa, diagnosed by its achenes with elongated 8–12 twisted apically hairs, was placed in a basal politomy inside clade C4. Conversely, all the analyses showed the species of Lucilia sensu Anderberg and Freire, traditionally diagnosed by their elongated achenial trichomes, scabrid pappus and leaf hairs with apical, flagellate cell bulbose at the base, as non-monophyletic. In fact, although four of its eight species sampled in this study have been recognized as distinct and belonging to Lucilia in previous treatments (Freire 1986b; Anderberg 1991, Anderberg and Freire 1991; Dillon 2000, 2003), our data show Lucilia linearifolia in a basal politomy and the remaining three species of Lucilia (L. lycopodioides, L. nitens, L. acutifolia) in a well-supported clade sister of Facelis + Micropsis. Lucilia linearifolia occurs in southern Brazil and northwestern Argentina and presents discolorous leaves. The remaining sampled species exhibit similar distribution (except L. nitens that is found in southern Brazil, Uruguay and northeastern Argentina) and present concolorous leaves. Future studies employing more variable DNA regions, with more extensive sampling, will be valuable before a recircumscription of Lucilia can be proposed.

Monophyly of Gamochaeta and its relationships with Stuckertiella and Gamochaetopsis

Our results suggest that Gamochaeta, as currently defined, is not monophyletic. All the analyses showed the clade Gamochaeta + Stuckertiella + Gamochaetopsis as highly supported, as well as the clade Gamochaeta serpyllifolia + Gamochaetopsis (Figs. 1, 2, 3). Stuckertiella is a genus of two species confined to South America (Peru, Bolivia, Argentina), and it was characterized by having 4-merous florets. Gamochaetopsis is a monotypic genus restricted to austral Chile and adjacent Argentina, differentiated by its achenes with short clavate hairs. Gamochaeta includes c. 60 species (three were sampled in this study) in the New World with some species adventives in the Old World (Nesom 1990a). The close relationship between Gamochaeta and Stuckertiella had been previously inferred in a morphological phylogenetic analysis (Dillon 2000) and diagnosed by truncate apically pilose style branches and achenes with globose myxogenic twin hairs. Gamochaetopsis was recognized as an isolated distinct genus (Anderberg and Freire 1991; Dillon 2000, 2003). The phylogeny of Gamochaeta is one of our ongoing projects, so before synonymizing Gamochaetopsis and Stuckertiella with Gamochaeta, we await the outcome of that work.

Taxonomic treatment of the Lucilia group s. lat.

Morphological characters for each genus are shown in Table 3. This treatment does not include the genera Gamochaeta, Gamochaetopsis and Stuckertiella which will be treated in a forthcoming paper.

Belloa J. Rémy in Gay, Fl. Chil. 3(3): 336. 1847. Type: B. chilensis (Hook. and Arn.) J. Rémy

=Lucilia Cass. sect. Lucilioides DC., Prodr. 7: 46. 1838.

One species, S. America (southern Andes): Belloa chilensis (Hook. and Arn.) J. Rémy, in Gay, Fl. Chil. 3(3): 336, t. 38. 1847 [=Lucilia chilensis Hook. and Arn., Chevreulia nivea Phil., Gnaphalium andicola Kuntze, Lucilia araucana Phil., L. chilensis Poepp. and Endl., L. frigida Reiche, L. nivea (Phil.) Cabrera]

Berroa Beauverd, Bull. Soc. Bot. Genève 5: 210. 1913. Type: B. gnaphalioides (Less.) Beauverd

One species, S. America (S Brazil, Uruguay, C and NE Argentina): Berroa gnaphalioides (Less.) Beauverd, Bull. Soc. Bot. Genève 5: 211. 1913 (=Lucilia argentea Hook. and Arn., L. gnaphalioides Less.)

Chevreulia Cassini, Bull. Soc. Philom. Paris, 1817: 69. 1817. Type: C. sarmentosa (Pers.) S.F. Blake

Six species, S. America (S Brazil, Bolivia, Paraguay, Uruguay, Chile, N and C Argentina, Falkland Islands): Chevreulia acuminata Less., Linnaea 5: 261.1830 (=C. elegans Rusby; C. diemi Cabrera, Notas Mus. La Plata, Bot. 7: 114. 1942; C. filiformis Hook. and Arn., C. longipes Wedd.); C. lycopodioides (d’Urv.) DC., Prodr. 7: 45. 1838 (=Gnaphalium lycopodioides d’Urv.); C. pusilla DC., Prodr. 7: 45. 1838 (=C. lanceolata J. Rémy); C. revoluta A. A. Schneid. and R. Trevis., Syst. Bot. 36 (3): 782. 2011; C. sarmentosa (Pers.) S.F. Blake, Proc. Biol. Soc. Washington 38: 85. 1925 [=Tussilago sarmentosa Pers., Chevreulia stolonifera Cass., nom. illeg., Chevreulia thouarsii J. Rémy, nom. illeg., Gnaphalium calycinum Poir., Xeranthemum caespitosum (cespitosum) Thouars].

Chionolaena DC., Prodr. 5: 397. 1836. Type: C. arbuscula DC.

=Leucopholis Gardner, London J. Bot. 2: 10. 1843. Type: L. phylicoides Gadner

=Parachionolaena M.O.Dillon and Sagást. Arnaldoa 1(2): 42. 1992. Type: P. columbiana (S. F. Blake) Dillon and Sagást.

=Pseudoligandra M.O. Dillon, Taxon 39(1): 127. 1990. Type: P. chrysocoma (Wedd.) Dillon and Sagást.

Twenty-seven species, C. and S. America (C Mexico, Costa Rica, Guatemala, N Colombia, S Brazil): Chionolaena adpressifolia Loeuille, Kew Bull. 66: 263. 2011; C. aecidiocephala (Grierson) Anderb. and S.E. Freire, Notes Roy. Bot. Gard. Edinburgh 46: 40. 1989 [=Anaphalis aecidiocephala Grierson, Gnaphalium aecidiocephalum (Grierson) L.O. Williams]; C. arbuscula DC., Prodr. 5: 397. 1836; C. campestris Deble, Kew Bull. 66: 264. 2011; C. canastrensis J.N. Nakaj., Kew Bull. 66: 266. 2011; C. capitata (Baker) S.E. Freire [=Achyrocline capitata Baker, Leucopholis capitata (Baker) Cufod.]; C. chrysocoma (Wedd.) S.E. Freire, Ann. Missouri Bot. Gard. 80: 415. 1993 [=Oligandra chrysochoma Wedd., Pseudoligandra chrysochoma (Wedd.) M.O. Dillon and Sagást.]; C. columbiana S.F. Blake, J. Wash. Acad. Sci. 25: 312. 1935 [=Parachionolaena columbiana (S.F. Blake) M.O. Dillon and Sagást.]; C. concinna (A.Gray) Anderb. and S.E. Freire, Notes Roy. Bot. Gard. Edinburgh 46: 40. 1989 [=Gnaphalium concinnum A. Gray, Anaphalis concinna (A.Gray) Grierson]; C. costaricensis (G.L. Nesom) G.L. Nesom, Sida 19: 850. 2001 (=Gnaphaliothamnus costaricensis G.L. Nesom); C. cryptocephala (G.L. Nesom) G.L. Nesom (=Gnaphaliothamnus cryptocephalus G.L. Nesom); C. durangensis (G.L. Nesom) G.L. Nesom, Sida 19: 850 (2001); C. eleagnoides Klatt, Leopoldina 23: 88. 1887 (=Gnaphalium eleagnoides (Klatt) S.F. Blake); C. isabellae Baker, in Martius, Fl. Bras. 6(3): 130. 1882 (=C. glaziovii Baker, C. innovans Wawra); C. jeffreyi H. Rob., Phytologia 55: 121. 1984; C. juniperina Loeuille, Kew Bull. 66(2): 268. 2011; C. latifolia (Benth.)Baker, in Martius, Fl. Bras. 6(3): 132. 1882 [=Leucopholis latifolia Benth., Chionolaena breweri Steyerm. and Maguirre, C. glomerata Baker, Lucilia breweri (Steyerm. and Maguire) V.M. Badillo]; C. lavandulifolia (Kunth) Benth. and Hook.f. ex B.D. Jacks., Index Kew. 1: 516. 1893 [=Elichrysum lavandulifolium Kunth, Chionolaena lavandulaceum (DC.) Hemsley, Elichrysum lavandulifolium (Wild.) D. Don, Gnaphalium lavandulaceum DC., G. lavandulifolium Willd., G. lavandulifolium (Kunth) S.F. Blake]; C. lychnophorioides Sch. Bip., Jahresber. Pollichia 20-21: 391. 1863; C. macdonaldii (G.L. Nesom) G.L. Nesom, Sida 19(4): 850. 2001 (=Gnaphaliothamnus macdonaldii G.L. Nesom); C. mexicana S.E. Freire, Ann. Missouri Bot. Gard. 80: 427. 1993; C. phylicoides (Gardner) Baker, in Martius, Fl. Bras. 6(3): 131. 1882 (=Leucopholis phylicoides Gardner); C. salicifolia (Bertol.) G.L. Nesom, Sida 19(4): 850. 2001 (=Helichrysum salicifolium Bertol., Gnaphaliothamnus salicifolius (Bertol.) G.L. Nesom); C. sartorii Klatt, Leopoldina Heft 23: 89. 1887 [=Gnaphalium sartori (Klatt) F.J. Espinosa-García]; C. seemannii (Sch.Bip.) S.E. Freire, Ann. Missouri Bot. Gard. 80: 432. 1993 (=Gnaphalium seemannii Sch.Bip.); C. stolonata (S.F. Blake) Pruski, Phytoneuron 2012-1: 4 (=Gnaphalium stolonatum S.F. Blake); C. wittigiana Baker, in Martius, Fl. Bras. 6(3): 129. 1882.

Cuatrecasasiella H. Rob., Fl. Neotrop. Monogr. 2(Suppl.): 14. 1985. Type: C. isernii (Cuatrec.) H. Rob.

Two species, S. America (Andes from Ecuador to N Argentina): Cuatrecasasiella argentina (Cabrera) H. Rob., Fl. Neotrop. Monogr. 2(Suppl.): 15. 1985 (=Luciliopsis argentina Cabrera); C. isernii (Cuatrec.) H. Rob., Fl. Neotrop. Monogr. 2(Suppl.): 15. 1985 (=Luciliopsis isernii Cuatrec.).

Facelis Cass., Bull. Sci. Soc. Philom. Paris (1819): 94. 1819. Type: Facelis retusa (Lam.) Sch. Bip.

Three species, S. America (S Brazil, Peru, Bolivia, Paraguay, Chile, Uruguay, Argentina): Facelis lasiocarpa (Griseb.) Cabrera, Physis (Buenos Aires) 10: 280. 1931 (=Filago lasiocarpa Griseb., Facelis capillaris Rusby, F. schultziana Beauverd, Lucilia erecta Benoist); F. plumosa (Wedd.) Sch. Bip., Linnaea 34: 532. 1866 (=Lucilia plumosa Wedd., Facelis weddelliana Beauverd, nom. nov. pro Lucilia plumosa Wedd.); F. retusa (Lam.) Sch. Bip., Linnaea 34(5): 532.1865 [=Gnaphalium retusum Lam., Facelis apiculata Cass., nom. illeg. pro Gnaphalium retusum Lam., F. retusa (Lam.) Sch. Bip. var. andicola (Nees) Beauverd, F. retusa (Lam.) Sch. Bip. var. candelabrum Beauverd, F. retusa (Lam.) Sch. Bip. var. patula Beauverd, F. retusa (Lam.) Sch. Bip. fo. congesta Beauverd, F. retusa (Lam.) Sch. Bip. fo. gigantea Beauverd, F. retusa (Lam.) Sch. Bip. fo. laxa Beauverd, F. retusa (Lam.) Sch. Bip. fo. nana Beauverd, F. retusa (Lam.) Sch. Bip. fo. planifolia Beauverd, Helichrysum retusum Spreng., Leptalea apiculata (Cass.) D. Don ex Hook. and Arn., Pteropogon chilense Fisch. and Meyer, nom. nud., P. andicola Nees].

Jalcophila M.O. Dillon and Sagást., Brittonia 38(2): 162. 1986. Type species: J. peruviana M.O. Dillon and Sagást.

Four species, S. America (Andes of Colombia, Ecuador, Peru, and Bolivia): Jalcophila boliviensis Anderb. and S.E. Freire, Brittonia 42: 139. 1990 (=Gamochaeta boliviensis (Anderb. and S.E. Freire) M.O. Dillon and Sagást.); J. colombiana S. Díaz and Veléz-Nauer, Revista Acad. Colomb. Ci. Exact. 19(72): 25. 1994; J. ecuadoriensis M.O. Dillon and Sagást., Brittonia 38(2): 165. 1986; J. peruviana M.O. Dillon and Sagást., Brittonia 38(2): 163. 1986.

Loricaria Wedd., Chlor. And. 1: 165, t. 27 A, B, C. 1856. Lectotype (designated by Cuatrecasas, 1954: 152): L. thuyoides (Lam.) Sch. Bip.

= Tafalla D.Don, Edinb. N. Phil. Journ.: 273. 1831, non Ruiz and Pav. 1794. Type: no stated.

Twenty two species, S. America (Andes from Colombia to Bolivia): Loricaria anceps Sch.Bip., Bonplandia 8: 259. 1860; L. antisanensis Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 157. 1954; L. azuayensis Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 158. 1954; L. cinerea D.J.N. Hind, Kew Bull. 59(4): 541. 2005; L. colombiana Cuatrec., Trab. Mus. Nac. Ci. Nat., Ser. Bot. 29: 33, Fig. 12. 1935; L. complanata (Sch. Bip.) Wedd., Chlor. And. 1: 167. 1856 (=Baccharis complanata Sch.Bip.); L. ferruuginea (Ruiz and Pav.) Wedd., Chlor. And. 1: 166. 1856 (=Molina ferruginea Ruiz and Pav.); L. graveolens (Sch.Bip.) Wedd., Clor. And. 1: 167. 1856 (=Baccharis graveolens Sch.Bip.); L. ilinisssae (Benth.) Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 162. 1954 (=Baccharis illissae Benth.); L. lagunillensis Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 162. 1954; L. leptothamna (Mattf.) Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 163. 1954 (=Tafalla leptothamna Mattf.); L. lucida Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 163. 1954; L. lycopodinea Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 164. 1954; L. macbridei Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 164. 1954; L. ollgaardii M.O. Dillon and Sagást., Phytologia 59: 228. 1986; L. pauciflora Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 165. 1954; L. puracensis Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 166. 1954; L. reticulosa Kuntze, Revis. Gen. Pl. 1: 352. 1891; L. scolopendra Kuntze, Revis. Gen. Pl. 1: 352. 1891; L. thuyoides (Lam.) Sch.Bip., Bonplandia 8: 260. 1860 [=Conyza thujoides Lam., Baccharis thujoides (Lam.) Pers., Loricaria (Tafalla) microphylla Hieron., L. stenophylla Wedd., L. stuebelii Hieron., L. thyoides (Lam.) Sch.Bip., L. thujoides (Lam.) Kuntze, Molina incana Ruiz and Pav., Tafalla thyoides (Lam.) D.Don; synonyms according to Hind 2011]; L. thyrsoidea (Cuatrec.) M.O. Dillon and Sagást., Phytologia 59: 230. 1986; L. unduaviensis Cuatrec., Feddes Repert. Spec. Nov. Regni Veg. 56: 170. 1954. (revision needed).

Lucilia Cass., Bull. Sci. Soc. Philom. Paris: 32. 1817. Type: Serratula acutifolia Poir. = L. acutifolia (Poir.) Cass.

=Oligandra Less., Syn. Gen. Comp.: 123. 1832. Type: O. lycopodioides Less.

=Pachyrhynchus DC., Prodr. 6: 255. 1838. Type: Pachyrhynchus xeranthemoides DC.

=Hymenopholis Gardner, London J. Bot. 7: 88.1848. Type: Hymenopholis imbricata Gardner

Eight species, S. America (S Brazil, Paraguay, Uruguay, N and C Argentina, Andes of Bolivia, Argentina and Chile): Lucilia acutifolia (Poir.) Cass., Dict. Scienc. Nat., ed. 2. 27: 264. 1823 (=Serratula acutifolia Cass., L. jamesonii Baker, L. lundii Baker, L. microphylla Cass., Pachyrhynchus xeranthemoides DC.); L. eriophora J. Rémy, in Gay, Fl. Chil. 3: 335. 1848; L. ferruginea Baker in Martius, Fl. Bras. 6 (3): 114. 1882; L. linearifolia Baker, in Martius, Fl. Bras. 6 (3): 114. 1882; L. lycopodioides (Less.) S.E. Freire, Taxon 38: 298. 1989 (=Oligandra lycopodioides Less., Hymenopholis imbricata Gardner, L. flagelliformis Wedd., L. glomerata Baker); L. nitens Less., Linnaea 5: 363. 1830; L. recurva Wedd.; Chlor. And. 1: 156. 1856; L. tomentosa Wedd., Chlor. And. 1: 157. 1856 (=L. squarrosa Baker).

Doubtful species: Lucilia saxatilis V.M. Badillo, Ernstia 50: 9, 1988. According to the original diagnosis which is accompanied by an illustration, is likely refererable to genus Gamochaeta by its small capitula arranged in glomerules, forming short spikes (vs. capitula arranged in leafy racemes in Lucilia), but no details of style branches are found nor description neither illustration. Consequently, until the type of Badillo’s name is located or materials can be provided, the identity of L. saxatilis is uncertain.

Micropsis DC., Prodr. 5: 459.1836. Type: M. nana DC.

Five species, S. America (S Brazil, Paraguay, Uruguay, NE Argentina, and C Chile): Micropsis australis Cabrera, Notas Mus. La Plata 3, Bot. 20: 147. 1938; M. dasycarpa (Griseb.) Beauverd, Bull. Soc. Bot. Genève, sér. 2, 5: 224. 1913 (=Filago dasycarpa Griseb.); M. nana DC., Prodr. 5: 460. 1836; M. ostenii Beauverd, Bull. Soc. Bot. Genève, sér. 2, 5: 226. 1913 (=M. bonariensis Manganaro); M. spathulata (Pers.) Cabrera, Notas Mus. La Plata 9, Bot. 46: 254. 1944 (=Gnaphalium involucratum Lam., nom. illeg. hom., non Foster, G. bracteatum Willd., nom. illeg. hom., non Lamarck, Evax spathulata Pers., Micropsis involucrata Cabrera).

Mniodes (A.Gray) Benth., Gen. Pl. 2(1): 301. 1873. Antennaria Gaertn. sect. Mniodes A. Gray, Proc. Acad. Arts. 5: 138. 1861. Lectotype (designated by Cuatrecasas, 1954: 3): M. andina (A. Gray) Cuatrec.

=Merope Wedd., Chlor. And. 1: 160. 1856, pro parte, nom. illeg. [excl. M. erythractis Wedd. = Gamochaeta erythractis (Wedd.) Cabrera], non Merope M.J.Roemer 1846 (Rutaceae). Lectotype (designated by Anderberg and Freire, 1991: 189): M. piptolepis Wedd.

=Belloa pro parte, non J. Rémy, syn. nov.

=Luciliocline Anderb. and S.E. Freire, Bot. J. Linn. Soc. 106: 187. 1991. Type: L. lopezmirandae (Cabrera) Anderb. and Freire, syn. nov.

Seventeen species, S. America (Andes from Venezuela to NW Argentina): Mniodes andina (A. Gray) A.Gray ex Cuatrec., Fol. Biol. Andina 1: 3. 1954 (=Antennaria andina A. Gray); M. aretioides (Sch. Bip.) Cuatrec., Fol. Biol. Andina 1: 3. 1954 [=Baccharis aretioides Sch. Bip., Antennaria aretioides (Sch. Bip.) A. Gray, Merope aretioides (Sch. Bip.) Wedd.]; M. coarctata Cuatrec., Fol. Biol. Andina 1: 4. 1954; Mniodes burkartii (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Gnaphalium burkartii Cabrera, Notas Mus. La Plata, Bot. 13, Bot. no. 56: 10, Fig. 2. 1948 [=Belloa burkartii (Cabrera) Cabrera, Lucilia burkartii (Cabrera) S.E. Freire, Luciliocline burkartii (Cabrera) Anderb. and S.E. Freire]; Mniodes catamarcensis (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Belloa catamarcensis Cabrera, Bol. Soc. Argent. Bot. 7: 81, Fig. 1. 1958 [=Lucilia catamarcensis (Cabrera) S.E. Freire, Luciliocline catamarcensis (Cabrera) Anderb. and S.E. Freire]; Mniodes kunthiana (DC.) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Conyza kunthiana DC., Prodr. 5: 379. 1836, nov. nom. [=Conyza pusilla Kunth 1820 non C. pusilla Houtt. 1782, Belloa kunthiana (DC.) Anderb. and S.E. Freire, Lucilia conoidea Wedd., L. kunthiana (DC.) Zardini, Lucilia pusilla Hieron. 1900 nov. nom., L. venezualensis Steyerm., L. violacea Wedd., Merope kunthiana (DC.) Wedd.]; Mniodes lehmannii (Hieron.) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Lucilia lehmannii Hieron, Bot. Jahrb. 19(5): 51. 1894 [=Belloa lehmannii (Hieron.) Anderb. and S.E. Freire]; Mniodes longifolia (Cuatrec. and Aristeg.) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Lucilia longifolia Cuatrec. and Aristeg., Fl. Venez. 10(1): 367. 1964 [=Belloa longifolia (Cuatrec. and Aristeg.) M.O. Dillon and Sagást., Luciliocline longifolia (Cuatrec. and Aristeg.) M.O. Dillon and Sagást.]; Mniodes lopezmirandae (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Belloa lopezmirandae Cabrera, Bol. Soc. Argent. Bot. 7: 83, Fig. 2. 1958 [=Lucilia lopezmirandae (Cabrera) S.E. Freire, Luciliocline lopezmirandae (Cabrera) Anderb. and S.E. Freire, Belloa turneri Sagást. and M.O. Dillon, Luciliocline turneri (Sagást. and M.O. Dillon) M.O. Dillon and Sagást.]; Mniodes pickeringii (A.Gray) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Lucilia pickeringii A. Gray, Proc. Amer. Acad. Arts 5: 138. 1862 [=Belloa pickeringii (A.Gray) Sagást. and M.O. Dillon, Luciliocline pickeringii (A.Gray) M.O. Dillon and Sagást.]; Mniodes piptolepis (Wedd.) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Lucilia piptolepis Wedd., Chlor. And. 1: t. 26, f. B. 1856 [=Belloa piptolepis (Wedd.) Cabrera, Gnaphalium piptolepis (Wedd.) Griseb., Luciliocline piptolepis (Wedd.) M.O. Dillon and Sagást.]; Mniodes plicatifolia (Sagást. and M.O. Dillon) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Belloa plicatifolia Sagást. and M.O. Dillon, Phytlogia 58: 384, f.3F-K. 1985 [=Lucilia plicatifolia (Sagást. and M.O. Dillon) S.E. Freire, Luciliocline plicatifolia (Sagást. and M.O. Dillon) M.O. Dillon and Sagást.]; M. pulvinulata Cuatrec., Fol. Biol. Andina 1: 5. 1954 (=M. ferreyrae Cuatrec.); Mniodes radians (Benth.) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Gnaphalium radians Benth., Pl. Hartw.: 207, tab. 35b. 1845 [=Belloa radians (Benth.) Sagást. and M.O. Dillon, Lucilia radians (Benth.) Cuatrec., Luciliocline radians (Benth.) M.O. Dillon and Sagást.]; Mniodes santanica (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Gnaphalium santanicum Cabrera, Notas Mus. La Plata, Bot. 13, Bot. no. 56: 12, Fig. 3. 1948 [=Belloa santanica (Cabrera) Cabrera, Lucilia santanica (Cabrera) S.E. Freire, Luciliocline santanica (Cabrera) Anderb. and S.E. Freire, Belloa spathulifolia Sagást. and M.O. Dillon, Luciliocline spathulifolia (Sagást. and M.O. Dillon) M.O. Dillon and Sagást.]; Mniodes schultzii (Wedd.) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov, Basionym: Merope schultzii Wedd., Chlor. And. 1: 163. 1856 [=Belloa cerrateae (Ferreyra) M.O. Dillon and Sagást., B. schultzii (Wedd.) Cabrera, Lucilia schultzii (Wedd.) A.Gray, Luciliocline schultzii (Wedd.) M.O. Dillon and Sagást., Mniodes cerrateae Ferreyra]; Mniodes subspicata (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey, comb. nov. Basionym: Belloa subspicata Wedd., Chlor. And. 1: 159. 1856 [=Lucilia subspicata (Wedd.) Hieron., Luciliocline subspicata (Wedd.) Anderb.].

Note: Mniodes tunariensis (Kuntze) Hieron. ex Weberb. = Novenia acaulis (Benth. and Hook. f. ex B.D. Jacks.) S.E. Freire and F.H. Hellw.

Concluding remarks

Our study, using plastid (rpl32-trnL, trnL-F) and nuclear (ITS and ETS) DNA loci, together with morphological characters, and using a broad sampling of the tribe, suggested that the South American Lucilia group sensu Anderberg and Freire (1991) is paraphyletic and that the genera Antennaria, Chionolaena, Gamochaeta, Loricaria, Micropsis, Mniodes, and Stuckertiella should be included for the establishment of the monophyly in the group. This mainly American Lucilia group s. lat. is diagnosed by having slender pappus bristles that are basally connate or coherent at the base by patent cilia. The phylogeny found here corroborates previous generic circumscription (Anderberg, 1991) of most of its genera, i.e., Berroa, Chevreulia, Cuatrecasasiella, Facelis, and Micropsis. Belloa is recovered as a monotypic genus in correspondence with Dillon’s study (2003). Luciliocline and the remaining species of Belloa sensu Anderberg and Freire (1991) are placed in the genus Mniodes. Thirteen new combinations are proposed: Mniodes burkartii (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes catamarcensis (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes kunthiana (DC.) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes lopezmirandae (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes lehmannii (Hieron.) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes longifolia (Cuatrec. and Aristeg.) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes pickeringii (A.Gray) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes piptolepis (Wedd.) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes plicatifolia (Sagást. and M.O. Dillon) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes radians (Benth.) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes santanica (Cabrera) S.E. Freire, Chemisquy, Anderb. and Urtubey; Mniodes schultzii (Wedd.) S.E. Freire, Chemisquy, Anderb. and Urtubey, and Mniodes subspicata (Wedd.) S.E. Freire, Chemisquy, Anderb. and Urtubey. Lucilia sensu Anderberg and Freire (1991) is non-monophyletic; future studies employing more variable DNA regions with extensive sampling will be valuable before a recircumscription of Lucilia can be proposed. The monotypic genera Stuckertiella and Gamochaetopsis are closely nested within the large genus Gamochaeta; however, greater taxonomic sampling is required before to make taxonomic changes for these genera.