Locations of origin of the specimens used in the morphometric analyses (acronyms as in Tables I and II). 

Context in source publication

Context 1

... endemic to peninsular Italy. The real taxonomic status of A. majellensis is however not clear (Greuter et al. 1989; Conti et al. 2005). According to Bianchini (1982), in fact, they are distinct taxa , even though « . . . in the areas of contact are frequent the presence of individuals with intermediate aspect, perhaps hybrids»; on the contrary, according to other authors (Pinto da Silva 1972) A. majellensis must be included in A. canescens because of their poor morphological separation. It is noteworthy that Boissier, who first described A. majellensis (Boissier 1848), in a subsequent work questions the independence of the two taxa : in Flora Orientalis (Boissier 1879) he wrote in a note to A. majellensis description « . . . A. canescens e Dalmatia et Montenegro ab A. majellensis non specifice differt» ( A. canescens growing in Dalmatia and Montenegro does not differ specifically from A. majellensis ). Moreover, in Flora d’Italia (Bianchini 1982) two subspecies are recognised under each of the two species, which are broadly sympatric and scarcely differentiated. In particular, subsp. gracilis of A. canescens is based on Armeria gracilis of Tenore (1831), who himself proposed it as a doubtful species, only partially distinguishable by characters related to hairiness. On the other hand, subsp. ausonia of A. majellensis was very synthetically described by Bianchini (1982), on the basis of a single herbarium specimen by Levier, conserved in the herbarium FI (Pignatti 1977) and coming from Mount Cocurello, a toponym now referable to Mount Coccorello, in the Velino Group, Central Apennine. According to Conti et al. (2005) the subspecies gracilis (Ten.) Bianchini is to be considered part of the variability of Armeria nebrodensis (Guss.) Boiss., endemic to Sicily, and therefore to be excluded from the peninsular Italy. In that work (Conti et al. 2005) the two subspecies of A. majellensis are considered, in the same manner as the species, as doubtful taxa . The study of the protologues (Host 1827; Boissier 1848) and the type materials (Scassellati et al. 2011) do not contribute to clarify the real status of A. majellensis and its relationship with A. canescens , since there are no evident features allowing to distinguish clearly the two species. For all these reasons we chose to perform a basic morphometric study using materials coming from extensive field collections throughout the Italian distribution area of the taxa , some herbarium specimens and samples coming from the loci classici of the species (Dalmatian Mounts for A. canescens and Majella Mounts for A. majellensis ). Such strategy is partially based on the finding that the characters measured are largely coincident with those used in artificial hybridisation study, in which they proved to be under multigenic control and allowed even distinguishing different hybrid generations (Nieto Feliner 1997). We chose to analyse the variability at a population level, as done both in Armeria (Nieto Feliner et al. 2001; Fuertes Aguilar et al. 2011) and in other genera (e.g. Brus et al. 2011; Iberite et al. 2011; Shiran et al. 2011). Between 2009 and 2010, we collected plants from 13 populations (materials now kept in the herbarium URT, Table I), occurring mostly not only on limestone but also on volcanic [Vivaro (VIV)], sandy [Laga (LAG)] and granite [Sila (SIL)] substrata. The elevation range of our field collections varies between 555 m a.s.l. (VIV) and 2259 m a.s.l. [Gran Sasso (GS)]. The habitat types in which sampled plants were growing are pastures and meadows, with discontinuous and continuous coverage, in the sub- mountain, mountain and sub-alpine vegetation belts. We also used 18 herbarium specimens from the herbaria CLU and RO and personal collections (Table II). Of those specimens, 11 are samples of closely related Italian species [i.e. Armeria denticulata (Bertol.) DC., Armeria saviana Selvi, Armeria macropoda Boiss., Armeria brutia Brullo et al. and Armeria aspromontana Brullo et al.] and 7 are specimens of Armeria from Mount Velino, locus classicus of A. majellensis subsp. ausonia , described by Bianchini (1982). The locations of origin of the specimens used in the morphometric analyses are shown in Figure 1. With the aim of exploring the phenetic affinities among this group of morphologically and related Italian plants, we used a morphometric multivariate approach. Twenty-one morphological quantitative characters (Table III) were measured in 148 dried specimens. One hundred and thirty correspond to the 13 studied populations (10 specimens per population) and 18 to the herbarium specimens on loan. The characters describing the vegetative parts were measured with a digital calliper (precision 0.01 mm), except for the length of mature scape for which we used measuring tape. For characters describing the inflorescence and the flower, due to their smaller size, we used a different method, in order to maximise the accuracy. First we acquired digital images of the calyx and the different parts of the inflorescence (i.e. outer, median and inner involucral bracts, spikelet bracts) with an Olympus ColorView II Camera. Then we measured all the characters using Cell ˆB, an image-acquisition and archiving software for various microscopy appli- cations. Only for the herbarium specimens we measured all the characters using a stereoscopic microscope with a calibrated viewfinder. The values in the so-obtained matrix (148 specimens £ 21 characters) are averages of three measurements ...