Taxonomy and conservation of Pancratium maritimum (Amaryllidaceae) and relatives in the Central Mediterranean

Abstract

Pancratium maritimum L. (Amaryllidaceae) is a geophyte occurring in the Mediterranean region, from the Black Sea to part of the Atlantic coast. This plant is receiving much attention from the international scientific community due to its value as a bioindicator, the potential industrial value of its chemical compounds, and its use as a commercial ornamental plant. Plant morphometry and sequences of three plastid DNA regions (rbcL, matK, trnH-psbA) were used to assess the phenotypic and genetic variability of this taxon and its closest congeneric species (in particular Pancratium linosae, from the volcanic island of Linosa) in the Central Mediterranean (Sicily, Tunisia and surrounding islands). Pancratium maritimum and P. linosae cannot be distinguished based on morphological and genetic data and should belong to the same taxon. Our results also highlight a diversified gene pool in P. maritimum that is worth preserving. The lectotypes of the names Halmira stellaris, Pancratium angustifolium and Pancratium foetidum are here designated.

Keywords:

• sea daffodil
• morphometry
• genetic variability
• DNA barcoding
• molecular markers

Introduction

Pancratium maritimum L. is one of the best known and most typical species of the Mediterranean littoral. In recent years, the level of attention given to this species has increased,because of its value as a bioindicator, the potential industrial value of its chemical compounds, and its use as a commercial ornamental plant (Abbassy et al. 1998; Berkov, Evstatieva, and Popov 2004; Bogdanova et al. 2009; El-Hadidy et al. 2011; Georgiev et al. 2011; Sanaa et al. 2013; Youssef and Frahm 1998). Alginic acid and derivatives were successfully extracted from the bulbs of this species (Sanaa et al. 2013). Several alkaloids with pharmacological activity were also found in P. maritimum. The most interesting of these was galanthamine, an acetylcholinesterase inhibitor (Berkov, Evstatieva, and Popov 2004). This species has also been thoroughly investigated from the biological (Eisikowitch and Galil 1971; Medrano, Guitián, and Guitián 2000), chemical (Ronsted et al. 2012), micro-morphological (Perrone et al. 2015), and genetic (Sanaa and Fadhel 2010; Sanaa et al. 2014; De Castro et al. 2012, 2014; De Felice et al. 2013; Di Maio and De Castro 2013) points of view.

Recently, many typical Mediterranean ecosystems have been severely altered by the increasing pressure of human activities and the consequences of global changes (de Montmollin and Strahm 2005; Sciandrello, Guglielmo, and Spampinato 2014). Among threatened habitats, sandy shores are one of the most endangered. Their severe decline is mainly the result of the human activities, tourism and construction development, which have modified the vegetation composition and the shorelines by alteration of the flow of water and sand deposition (Guarino et al. 2008). Moreover, the proliferation of invasive exotic species, such as Carpobrotus acinaciformis (L.) L. Bolus, has reduced the habitat available to native, less competitive species (Celesti-Grapow et al. 2010).

In the central Mediterranean, four species of Pancratium occur: Pancratium maritimum L., Pancratium linosae Soldano & F. Conti, Pancratium foetidum Pomel, and Pancratium illyricum L. (Conti et al. 2005; Euro+Med 2006). Pancratium maritimum and P. linosae are coastal plants, whereas P. foetidum and P. illyricum are rupicolous-heliophilous plants that also grow in the hinterland, up to 1350 m above sea level (Maire and Quézel 1959; Pignatti 1982). Pancratium maritimum, in contrast to most of the other species of Pancratium, has a wide distribution extending all over the Mediterranean, up to the Black Sea and to part of the Atlantic coast (Euro+Med 2006). In contrast, P. linosae is known only from the Pelagian Islands (De Castro et al. 2012). We therefore aimed to investigate the taxonomic independence of P. linosae within the area of distribution of P. maritimum.

In the context of development of a molecular database of the threatened flora of Sicily and surrounding areas (Giovino et al. 2015), the potential of the plastid markers, rbcL, matK and trnH-psbA, was assessed for Pancratium.

Material and methods

Pancratium was studied in the central Mediterranean, i.e. Sicily and Tunisia along a latitudinal transect from 35°N to 38°N. Specimens were collected in the field and voucher specimens were deposited in the Herbarium Mediterraneum Panormitanum (PAL) (Table 1). Field identification was based on morphological characters in the mature stage. Individual plants were cultivated in the Botanical Garden of Palermo to exclude environmental variability, for subsequent comparative morphological analyses. For each population, character traits were measured for 10 individuals (five from Zembra due to the small size of this population) with 10 replicate measurements from each individual. Measurements were taken using an electronic calliper. The quantitative characters considered were: #1 Leaf length (cm); #2 Leaf width (cm); #3 Flowering stem length (cm); #4 No. of flowers per stem; #5 External length of tepals (cm); #6 Paracorolla total length (cm); #7 Paracorolla teeth length (cm). The means of the measures used for statistics are presented in the Supplemental data (File 1). According to Boyd (2002) and Peruzzi et al. (2015) these characters were subjected to a Principal Component Analysis, with the individuals a priori assigned to the eight populations (Figure 1). Each character was also subjected to univariate analysis (analysis of variance or Kruskal–Wallis test with corrections for multiple comparisons, Pearson correlation coefficients, Tukey’s honest significant difference test and Bonferroni, respectively), using PAST version 3.06 (Hammer, Harper, and Ryan 2001; Hammer 2015). Pearson correlation coefficients (r) among the seven characters measured are presented in the Supplemental data (File 2). A discriminant analysis for the eight a priori groups recognized was performed. The scatter plot of specimens along the first two canonical axes is shown (Figure 2). The range of each continuous numerical character was represented using box-and-whisker plots (Figure 3).

Table 1. Synoptic table of the populations studied reporting the field identification, location, habitat characteristics and voucher details.

Taxon	Location	Environment	Coordinates	Date	Collector	Voucher ID	Herbarium	Revision	Code in statistical analysis
Pancratium linosae Soldano & F.Conti	Lampedusa, Pelagie Islands	Carbonate dunes	35°29'55" N 12°35'55" E	08/06/2012	G. Domina	P4.C	PAL101658	P. maritimum L.	Lampedusa
Pancratium linosae Soldano & F.Conti	Linosa, Pelagie Islands	Volcanic dunes	35°51'42" N 12°51'18" E	11/06/2012	G. Domina	P6.C	PAL96715	P. maritimum L.	Linosa1
Pancratium linosae Soldano & F.Conti	Linosa, Pelagie Islands	Volcanic dunes	35°51'20" N 12°52'52" E	07/08/2013	G. Domina	P11u	PAL101657	P. maritimum L.	Linosa2
Pancratium maritimum L.	Cannatello, Sicily	Carbonate dunes	37°14'21" N 13°37'01" E	12/06/2012	G. Domina	P5.C	PAL101659	P. maritimum L.	Cannatello
Pancratium maritimum L.	Marsala, Sicily	Carbonate dunes	37°44'502 N 12°28'12" E	27/11/2013	G. Domina	P16u	PAL101655	P. maritimum L.	Marsala
Pancratium maritimum L.	Balestrate, Sicily	Carbonate dunes	38°02'47" N 12°59'47" E	27/11/2013	G. Domina	P17u	PAL101654	P. maritimum L.	Balestrate
Pancratium maritimum L.	Chebba, Tunisia	Carbonate dunes	35°13'32" N 11°09'27" E	25/04/2013	G. Domina	P12u	PAL101653	P. maritimum L.	Chebba
Pancratium maritimum L.	Zembra, Tunisia	Carbonate dunes	37°07'05" N 10°48'41" E	13/05/2013	G. Domina	P13u	PAL101652	P. maritimum L.	Zembra

Figure 1. Principal components analysis based on the seven considered morphological characters, with groups corresponding to the eight populations. PC1: Eigenvalue 55.81, % variance 81.50; PC2: Eigenvalue 8.17, % variance 11.93. Yellow: Chebba; magenta: Zembra; red: Lampedusa; light green: Linosa 1; dark green: Linosa 2; brown: Cannatello; dark blue: Marsala, light blue: Balestrate.

Figure 2. Discriminant analysis based on the seven considered morphological characters, with groups corresponding to the eight populations. Axis 1: Eigenvalue 0.40153, % variance 38.6; Axis 2 Eigenvalue 0.32117, % variance 30.88. Yellow: Chebba; magenta: Zembra; red: Lampedusa; light green: Linosa 1; dark green: Linosa 2; brown: Cannatello; dark blue: Marsala, light blue: Balestrate.

Figure 3. Plots of the six continuous numeric characters. For each sample, the 25–75% quartiles are drawn using a box. The median is shown with a horizontal line inside the box. The whiskers are drawn from the top of the box up to the largest data point less than 1.5 times the box height from the box, and similarly below the box. Outliers values are shown as stars. No significant differences among populations are highlighted. T(white): Overall; 1(yellow): Chebba; 2(magenta): Zembra; 3(red): Lampedusa; 4(light green): Linosa 1; 5(dark green): Linosa 2; 6(brown): Cannatello; 7(dark blue): Marsala, 8(light blue): Balestrate.

The barcoding approach was adopted in support of the morphological investigation. Multiple individuals for each taxon were used for molecular analysis. Plant material for DNA extraction consisted of young lyophilized leaves. Genomic DNA extraction was based on CTAB protocol for plant tissue (Doyle and Doyle 1987).

The three plastid barcoding regions rbcL, matK, trnH-psbA, were assessed by adopting polymerase chain reaction primers and conditions suggested by the Consortium for the Barcode of Life (CBOL) (Fazekas et al. 2012; Dunning and Savolainen 2010) (Table 2). When making the choice of markers we considered the relevance of the compromise between the discrimination level supported by a marker and amplification and sequencing success (Chase et al. 2005, Hollingsworth et al. 2009). The choice of trnH-psbA, as an additional marker, appeared logical to discriminate morphologically close samples.

Table 2. Barcoding polymerase chain reaction primers adopted in this study.

matK	Dunning and Savolainen, 2010
390F	CGATCTATTCATTCAATATTTC
3F_KIM	CGTACAGTACTTTTGTGTTTACGAG
rbcL	Fazekas et al., 2012
rbcLa_F	ATGTCACCACAAACAGAGACTAAAGC
rbcLa_R	GTAAAATCAAGTCCACCRCG
trnH-psbA	Fazekas et al., 2012
trnH_F	CGCGCATGGTGGATTCACAATCC
psbA_R	GTTATGCATGAACGTAATGCTC

Polymerase chain reaction amplifications were performed with the Gene^®Amp PCR System 9700 (Applied Biosystems, Foster City, CA, USA). Products were purified and bi-directionally sequenced (Amersham Biosciences DYEnamic™ ET Terminator Cycle Sequencing Kits), according to the Sanger protocol for AB3730XL DNA Analyzer (Applied Biosystems). The resulting electropherograms were screened for errors and assembled into contigs using Sequencer software 4.10 (Gene Codes Corporation, Ann Arbor, MI, USA). The sequence alignments were carried out by MUSCLE and phylogenetic Neighbour-Joining. A tree was generated for molecular identification, based on a Kimura two-parameter model, using Mega 6 software (Kimura 1980; Saitou and Nei 1987; Tamura et al. 2013). The comparison included all new sequences generated, a subset of the most closely related sequences, and the significant BLAST results, downloaded from GenBank database (Table 3). Sampling, morphological and molecular data generated in this investigation were submitted to the BOLD database under the dedicated project code FMED (www.boldsystems.org).

Table 3. Reference data of the newly generated sequences and the subset of the related sequences for each taxon considered in this study.

Taxa	Location	Reference study	Voucher ID	rbcL molecular ID	matK molecular ID	trnH-psbA molecular ID
Pancratium maritimum L.	Lampedusa, Pelagie Islands	in this study	P4.C	FMED032-14_rbcLa	FMED032-14_matK	FMED032-14_trnH-psbA
Pancratium maritimum L.	Linosa, Pelagie Islands	in this study	P6.C	FMED024-12_rbcLa	FMED024-12_matK	FMED024-12_trnH-psbA
Pancratium maritimum L.	Linosa, Pelagie Islands	in this study	P11u	FMED034-14_rbcLa	FMED034-14_matK	FMED034-14_trnH-psbA
Pancratium maritimum L.	Cannatello, Sicily	in this study	P5.C	FMED033-14_rbcLa	FMED033-14_matK	FMED033-14_trnH-psbA
Pancratium maritimum L.	Marsala, Sicily	in this study	P16u	FMED037-14_rbcLa	FMED037-14_matK	FMED037-14_trnH-psbA
Pancratium maritimum L.	Balestrate, Sicily	in this study	P17u	FMED038-14_rbcLa	FMED038-14_matK	FMED038-14_trnH-psbA
Pancratium maritimum L.	Chebba, Tunisia	in this study	P12u	FMED035-14_rbcLa	FMED035-14_matK	FMED035-14_trnH-psbA
Pancratium maritimum L.	Zembra, Tunisia	in this study	P13u	FMED036-14_rbcLa	FMED036-14_matK	FMED036-14_trnH-psbA
Pancratium linosae Soldano & F.Conti	Linosa, Pelagie Islands	De Castro et al., 2012; Unpublished*	CAT:Brullo PM4	FN594932	FN594918*	n.a.
Pancratium maritimum L.	Linosa, Pelagie Islands	De Castro et al., 2012	CAT:Brullo PM4	FN594932	n.a.	n.a.
Pancratium maritimum L.	Lampione, Pelagie Islands	De Castro et al., 2012	PM-LM1, NAP	FN594931	n.a.	n.a.
Pancratium maritimum L.	Lampedusa, Pelagie Islands	De Castro et al., 2012	PM-BR1, NAP	FN594930	n.a.	n.a.
Pancratium maritimum L.	Lampione, Pelagie Islands	De Castro et al., 2012; Unpublished*	CAT:Brullo PM1	FN594931	FN594917*	n.a.
Pancratium maritimum L.	Lampedusa, Pelagie Islands	De Castro et al., 2012; Unpublished*	CAT:Brullo PM3	FN594930	FN594916*	n.a.
Pancratium maritimum L.	Foce Simeto, Sicily	De Castro et al., 2012	PM-FS13, NAP	FN594927	FN594913	n.a.
Pancratium maritimum L.	Gallipoli, Italy	De Castro et al., 2012	PM-O18, NAP	FN594928	FN594914	n.a.
Pancratium maritimum L.	Ischitella, Italy	De Castro et al., 2012	PMO1, NAP	FN594929	FN594915	n.a.
Pancratium maritimum L.	Barcelona, Spain	De Castro et al., 2012	PM-BP1, NAP	HE565507	n.a.	n.a.
Pancratium maritimum L.	Portugal	Schaefer et al., 2011	BM 2008/364	HM850226	HM850507	n.a.
Pancratium maritimum L.	Spain	Gage et al., 2011	32786	n.a.	HM011039	n.a.
Pancratium maritimum L.	Casablanca, Morocco	De Castro et al., 2012	PM-Q9, NAP	HE565506	n.a.	n.a.
Pancratium maritimum L.	Darnah, Libya	De Castro et al., 2012	PM-L10, NAP	HE565508	n.a.	n.a.
Pancratium maritimum L.	Baltim, Egypt	De Castro et al., 2012	L. 20245, K	HE565501	n.a.	n.a.
Pancratium maritimum L.	Elafonissi, Crete	De Castro et al., 2012	PM-CR1, NAP	FN594926	n.a.	n.a.
Pancratium maritimum L.	Botanical Garden of theHebrew University, Israel	De Castro et al., 2012; Unpublished*	JBG:Leschner 2006	FN594933	FN594919*	n.a.
Pancratium illyricum	Tunisia	Gage et al., 2011	32785	n.a.	HM011038	n.a.
Pancratium illyricum	Sardinia	Ronsted et al., 2012	Ronsted 370 (C)	n.a.	JX464566	n.a.
Pancratium sickenbergeri	Negev desert, Israel	De Castro et al., 2012; Unpublished*	CAT:Brullo PM5	FN594935	FN594920*	n.a.
Pancratium foetidum	Botanical Garden of Catania, Sicily	Unpublished	CAT:Brullo PM2	FN594923	FN594909	n.a.
Pancratium canariense	Cyprus	Chen et al., 2013	Chase 17733	JX903160	JX903570	n.a.
Pancratium tenuifolium	South Africa	Unpublished	De Castro 71907	n.t.	FN594921	n.a.
Pancratium arabicum	El Borollos, Egypt	De Castro et al., 2012	De Castro, NAP	HE603957	n.a.	n.a.
Sternbergia lutea	Trieste, Italy	Bruni et al., 2012	MIB:ZPL:04510	HE963697	HE967010	HE966837
Lycoris aurea		Unpublished	Yuan050920	n.a.	n.a.	GQ923938
Hippeastrum puniceum		Unpublished	PDBK2012-0631	n.a.	KC704508	KC704245
Narcissus tazetta var. chinensis		Unpublished	PDBK2012-0014	KC704786	KC704517	KC704254

n.a. : not available sequence.

Text in bold type indicates sequences with uncertain taxa attribution.

Results

The principal components analysis (Figure 1) and the discriminant analysis (Figure 2) indicated considerable overlap across the populations in this study. The cases correctly classified by discriminant analysis according to the groups assigned a priori were 36%. Univariate analysis of all the characters considered did not indicate any significant differences across populations. Morphological characters (Figure 3) show continuous variation and do not fall into distinct population groupings.

All the barcoding fragments were successfully amplified and sequenced from the samples investigated. rbcL did not show any variation. Low rbcL polymorphism has been reported in numerous plant lineages (e.g. Chase et al. 2007; de Vere et al. 2012; Costion et al. 2011) and this marker was therefore discarded from subsequent analyses. Conversely, matK and trnH-psbA showed significant levels of genetic variation.

The phylogenetic tree, both of matK and trnH-psbA, showed concordant results, indicating no genetic divergence between the reference sequences of P. maritimum, recovered from GenBank, and those of the majority of the samples collected in this study (Figures 4 and 5). However, two Sicilian populations (taxa ID: P16u, P17u), originating from Balestrate and Marsala (north Sicily), were different on both markers.

Figure 4. Phylogenetic neighbour-joining tree generated from the most significant matK gene sequences listed in Table 3.

Figure 5. Phylogenetic neighbour-joining tree of the most significant trnH-psbA IGS sequences listed in Table 3.

No divergence was observed when comparing samples from plants in the in vivo collection.

Discussion

There was no statistically significant variation in morphological character traits of leaves, stems and flowers between the cultivated plants from the eight different locations, originally recognized as P. maritimum and P. linosae. This indicates that these plants belong to a single taxon. Considerations in Di Silvestro, Ardenghi, and Parolo (2010), De Castro et al. (2012) and Perrone et al. (2015) on analysis of plants in wild populations, lead in the same direction, although the significant differences detected in the field provide evidence for high within-species phenotypic variability.

In spite of the overall low level of genetic variability observed among the sequences generated in this study, matK and trnH-psbA revealed a level of genetic variation that is high enough to support some considerations and help in clarifying some doubtful sequences associated with this group in GenBank. First, the matK sequence of P. maritimum from Marsala and Balestrate, matched precisely the reference sequence ascribed to P. linosae (ID: FN594918, voucher CAT: Brullo PM4). This result still argues for considering P. maritimum and P. linosae as phenotypic variants of the same taxon. Second, we also highlight the incorrect GenBank taxa assignment for the matK sequence referred to P. illyricum from Tunisia (Table 3, see ID: HM011038, voucher 32785). The phylogenetic tree showed this voucher specimen referable to P. maritimum (Figure 4). Moreover P. illyricum is known to be endemic to Corsica and Sardinia, so the sampling location is rather unlikely to refer to this species.

The trnH-psbA intergenic spacer was used for the first time on Pancratium taxa. Quite surprisingly, it showed low molecular variation; even though it is known to be highly variable in complex groups (Chase et al. 2007; Chen et al. 2010; Hollingsworth, Graham, and Little 2011; De Mattia et al. 2011). Nevertheless, this fragment confirmed the variability of the two Sicilian samples (P16u, P17u), supported also by matK.

The evaluation of additional regions could represent a possibility for deeper investigations on the genetic variability of such taxa, in order to increase the resolution of molecular discrimination (Di Maio and De Castro 2013). Nonetheless, other molecular studies on the P. maritimum group, using further markers (i.e. trnL-trnF, trnL, ITS), also demonstrated a low level of genetic variation (Gage et al. 2011; De Castro et al. 2012). A population genetic study could help to assess gene flow between different locations, and identify differentiated gene pools (see below).

The following classification of the species of Pancratium occurring in the central Mediterranean is therefore proposed:

Pancratium maritimum L., Sp. Pl.: 291. 1753

= P. linosae Soldano & F. Conti, Annot. Checkl. Italian Vasc. Fl. 20. 2005. Type: homotypic with P. angustifolium Lojac.

= P. angustifolium Lojac., Fl. Sic. 3: 82. 1909., non M. Roem., Syn. Mon., 4 : 178. 1847. Typus: Lectotype here designated: Pancratium …. n. spec. Linosa si coltiva e si attenda che fiorisca [manu Tineo] / Pancratium angustifolium Lojac. [manu Lojacono] PAL 63913!

Typus: Lectotype: “Narcissus maritimus” in Morison, Pl. Hist. Univ., 2: 365, s. 4, t. 10, f. 28, 1680 designated by Wijnands in Jarvis & al. (ed.), Regnum Veg. 127 : 73 (1993).

Pancratium foetidum Pomel, Nouv. Mat. Fl. Atl. 253. 1874

= Pancratium collinum Coss. & Durieu, Ann. Sci. Nat., Bot., IV, 1: 228. 1854 [nom. nud.]

Typus: Lectotype here designated: Pancratium foetidum, Oran [manu Pome] / Pancratium foetidum Pomel, (type !), Oran, Pomel [manu Maire] MPU 005173!

Pancratium illyricum L., Sp. Pl.: 291. 1753

= Zouchia illyrica (L.) Raf. Fl. Tellur. 4: 22. 1838

= Almyra stellaris (Salisb.) Salisb., Trans. Hort. Soc. London 1: 336. 1812 [nom. illeg.]

= Halmira stellaris Parlatore, Nuov. Gen. Sp. Monocot.: 28. 1854 Typus: Lectotype here designated: Halmira stellaris Parl. Fl. It., Pancratium illyricum Linn., da Moris in Marzo1847, in apricis collinis: 8bri, 9bri Sardinia, FI fototeca3010B!

= Pancratium stellare Salisb., Trans. Linn. Soc. London 2: 74. 1794 [nom. illeg.]

Typus: Lectotype designated by Peruzzi et al. 2013: 828: Herb. Van Royen No. 897.326–441 (L).

Conservation

The morphological and genetic variability observed highlights the importance of preserving as many sites as possible in order to maintain a large gene pool. In fact, all genotypes resulting from local adaptation to a particular habitat are independent genetic units deserving conservation. Sandy shores of the Mediterranean are homogeneous habitats; they host species with broad distributions and are subject to continuous variations (Sànchez Doreste, Domina, and Caujapé-Castells 2005; Jimenez-Alfaro et al. 2015). In this case, as supposed by Picó and Van Groenendael (2007) and by Geraci, Rimondo, and Troia (2009), it is reasonable to believe that fragmentation of the original population and the consequent geographical isolation of the sub-populations have promoted the genetic variation observed. The patchy genotypic distribution can be explained in terms of the seed dispersal capability of the plants and in terms of the importance of Mediterranean sea currents in dispersal processes, allowing the establishment of new colonies that are distant from the parent population (Balestri and Cinelli 2004; Westberg and Kadereit 2009; Raimondo et al. 2012; De Castro et al. 2014 Sanaa et al. 2015). Pancratium maritimum has a relatively very high dispersal capacity, giving rise to probable panmixy throughout the whole central Mediterranean as observed by Kadereit et al. (2005) in other coastal species such as Cakile maritima Scop., Crithmum maritimum L. and Halimione portulacoides (L.) Aellen. The continuous decline of P. maritimum reduces the size of populations and their internal variability. It is therefore desirable to preserve P. maritimum populations in situ with a “dynamic” strategy that includes management of ecosystems, maintenance of gene flow, reintroduction of plants, environmental education, etc. Pancratium maritimum, taking into account its broad distribution, its distinctive appearance, and its iconic status, should be adopted as symbol plant of the sandy shore habitat in conservation initiatives.

Previous experiences of replant–transplant processes showed that overall performance of plants tends to be higher in populations in their native sites in comparison with populations replanted in non-native sites (Joshi et al. 2001; Bischoff et al. 2006). From this, it follows that it is important to know the distribution of genotypes in the field. This is necessary to be able to select the most suitable material for reintroduction in sites that have become too distant from the existing natural populations. This distance is related to the average dispersal distance of each species.

Ex situ conservation should be based on the collection of as large a number of seeds as possible from different populations in mainland and island habitats in order to conserve as much variability as possible. The use of this plant as an ornamental, from plants grown from seeds, and not bulbs dug from the wild, as generally happens, would increase the reserves of genetic material available without reducing the number of mature individuals in natural populations.

Conclusions

This work confirms the validity of the integrated approach linking molecular and morphological analysis, promoted by the CBOL, as the best solution for assigning a real meaning to the levels of nucleotide divergence observed and for developing an accurate universal database for molecular identification (Rougerie et al. 2009, Dunn 2003).

Pancratium maritimum is a highly variable species, from both the genotypic and phenotypic points of view, and this has allowed it to colonize sandy shores of different geologies and in some cases, rocky cliffs. This implies a concerted and intense effort to collect, conserve and reintroduce propagules from as many individuals and populations as possible to ensure the conservation of this species.

The distinction between P. maritimum and P. linosae is not supported by consistent morphological or molecular data. The samples of P. linosae from its locus classicus represent just one example of the broad intraspecific variation that can be found along the coast of the Mediterranean. The particular morphology observed on Linosa appears to be related to the nature of the extremely arid volcanic sands in which it grows. The same probably applies to the individuals of P. maritimum growing in rocky fissures on the islet of Lampione (CAT: Brullo PM1 in Figure 4 and Table 3), which show very similar morphological characters (Lo Cascio and Pasta 2012). In this case too, speciation by genetic drift does not need to be invoked, and character traits observed within this population are well within the range of the broad morphological and genetic variation observed in P. maritimum.

Notes on contributors

Antonio Giovino, PhD is a researcher at the Agricultural Research Council (CRA). The main focus of his research is on molecular studies and conservation of native higher plants. During the past years he studied the conservation biology of Mediterranean plants. Contribution: Molecular analysis and Morphometry.

Gianniantonio Domina, PhD is a researcher at the University of Palermo of Applied Botany, General Secretary of OPTIMA (the Organization for the Phyto-Taxonomic Investigation of the Mediterranean Area). The main focus of his research is about plant taxonomy and nomenclature. Contribution: Field work and Morphometry.

Giuseppe Bazan, PhD is Associate Professor at the University of Palermo of Applied Botany. The main focus of his research is about geographical information systems applied to botany and to plant conservation. Contribution: Morphometry

Patrizia Campisi, PhD is a researcher at the University of Palermo of Applied Botany. The main focus of her research is on conservation of higher plants and mosses, and on taxonomy of mosses. Contribution: Morphometry

Silvia Scibetta, PhD is an associated researcher at the Agricultural Research Council (CRA). The main focus of her research is on molecular analysis and taxonomy of higher plants and fungi. Contribution: Molecular analysis and Morphometry.

Supplemental data

Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/12538078.2015.1089416.

Electronic supplementary file 1. Morphological characters (mean) used for the statistical analysis.

Electronic supplementary file 2. Pearson correlation coefficients (r) among the seven characters measured. All correlations were significant with p > 0.0001 (Student’s t-test) with the lone exception of External tepals length.

Acknowledgements

Our grateful thanks go to Dr Chira Nepi, curator of FI, for the digital loan of the exsiccata of Halmira stellaris and to Dr Sandro Lanfranco (University of Malta) for his careful critical reading of the manuscript.

Additional information

Funding

The field trips in Chebba and Zembra (Tunisia) were performed thanks to the contribution of APAL (Coastal protection and planning agency of Tunisia) and of the PIM (Small Mediterranean Island) Initiative coordinated by the Conservatoire du littoral (France).