Lysymachio foeminae-Atractylidetum cancellatae Di Pietro, Fortini, Misano et Terzi ass. nov. hoc loco (Table 1)

Lysymachio foeminae-Atractylidetum cancellatae typicum

Name - Lysymachio foeminae-Atractylidetum cancellatae typicum Di Pietro, Fortini, Misano et Terzi subass nov. hoc loco (Holotypus rel. 5, Table 1).

Characteristic taxa - Atractylis cancellata, Helianthemum salicifolium, Filago eriocephala, Lysimachia foemina.

Dominant taxa - Atractylis cancellata.

High frequency taxa - Atractylis cancellata, Helianthemum salicifolium, Filago eriocephala, Lysimachia foemina, Hippocrepis biflora.

Synecology - Therophytic communities colonizing the Pliocene clayey substrates of the Lucanian badlands and developed on extremely restricted areas (1–2 m2 or even less). The Lysymachio-Atractylideto typicum (Fig. 6) acts as a poor-in-species community (average number of species per relevé = 12) which characterizes the relatively stable parts of the badland’s south-facing slopes. In particular this community is located above and below the steeper part of the badland characterized by accelerated erosion and occupied by the Camphorosmo-Lygeetum sparti Brullo et al. 1990. In this context the Lysymachio foeminae-Atractylidetum cancellatae gives rise to two separated therophytic transiction fringes between the Camphorosmo-Lygeetum and the communities developed on its upper and lower sides. In most cases the communities developed at the upper side are those dominated by Sulla coronaria and Scorpiurus muricatus occurring at the top of each single badland unit. The communities developed at the Camphorosmo-Lygeetum lower side are generally dominated by Parapholis incurva and Sulla capitata these species finding their optimal conditions at the badland footslopes where the run-off waters tend to deposit the finest clays.

Life forms and chorology - Therophytes dominates in all the life forms spectra followed by the hemicryptophytes. Geophytes are completely absent and chamaephytes are very scarce. In the chorological spectrum steno-Mediterraneans dominate followed by Euro -Mediterraneans. The other chorotypes exhibit percentages always lower than 4% (Fig. 7).

Syndynamic - The Lysimachio-Atractylidetum typicum can be considered as a fragment of the micro-permanent vegetation mosaic forming the “badland system” micro-geosygmetum. A reduction of the erosive action of the flooding waters leads to the rapid colonization of the space occupied by the Lysymachio foeminae-Atractylidetum cancellatae by the communities dominated by Sulla coronaria (upper part of badland) or Sulla capitata (lower part of badland) and subsequently by the final scrub stage of the Helictotricho convoluti-Pistacietum lentisci Di Pietro et Misano 2010.

Lysymachio foeminae-Atractylidetum cancellatae ononidetosum reclinatae

Name - Lysymachyo foeminae-Atractylidetum cancellatae ononidetosum reclinatae Di Pietro, Fortini, Misano et Terzi subass nov. (Holotypus: rel. 12, Table 1).

Differential taxa of subassociation - Ononis reclinata, Bupleurum baldense, Coronilla scorpioides.

Dominant taxa - Atractylis cancellata.

High frequency taxa - Atractylis cancellata, Helianthemum salicifolium, Filago eriocephala, Lysimachia foemina, Brachypodium distachyon, Hypochoeris achyrophorus, Euphorbia exigua, Coronilla scorpioides, Poa infirma, Crupina vulgaris, Helianthemum jonium, Fumana ericifolia.

Synecology - On the limestone substrates of the Murgia delle Gravine (in this case in the Gravina of Matera), this community features small pockets of soil interposed between the steppe-like grasslands and the micro-garrigues and the Mediterranean maquis communities. The most natural and long-lasting aspect of this subassociation develops where a very thin layer of soil covers the flat and more or less crackless limestone rocky surfaces. However, there are also other aspects related to the more or less disturbed areas placed on the edge of the paths created by grazing cattle trampling. Compared to the Lysimachio-Atractylidetum typicum occurring within the clayey Badlands area, the subass. ononidetosum reclinatae shows a significantly higher floristic richness (average number of species per relevé = 24). This is due to the wider pool of available annual species coming from the surrounding habitats (dry grasslands and micro-garrigues on limestone) and to the fact that the calcareous lithosols of the Murgia area are less oligotrophic than the sub-halophytic clays of the badlands.

Life forms and chorology - Therophytes dominates in all the life forms spectra followed by chamaephytes and hemycryptophytic with similar percentages. In the chorological spectrum steno-Mediterraneans taxa dominate followed by Euro-Mediterraneans. The other chorotypes exhibit percentages always lower than 5% (Fig. 7).

Syndynamic - The progressive vegetation succession consists of a first step dominated by the Stipa austroitalica steppe-like grasslands and a second one characterized by Mediterranean maquis scrubs (Helictotricho-Pistacietum lentisci, Juniperetum phoeniceae-macrocarpae Di Pietro et Misano 2010, Cyclamino repandi-Paliuretum spinae-christi Biondi et al. 1999).

Syntaxonomy - The Lisymachio foemine-Atractylidetum cancellatae is an association provisionally described (sub Anagallido foeminae-Atractylidetum cancellatae) for the vegetation mosaic of the Lucanian badlands (Fascetti et al. 2001; Di Pietro et al. 2003) whose validation has never been formalized. The new samples from the Murgia territories of Apulian and Basilicata regions allowed to enlarge the distribution area and the ecological spectrum of this association that we consider as divided into two subassociations: typicum and ononidetosum reclinatae. In this paper we validate the Lisymachio foemine-Atractylidetum cancellatae taking into account the chronology of the studies that have concerned this community over time. Accordingly, we have here established as subassociation “typicum” the one comprising the relevés performed on the salty-clayey substrates of the Lucanian badlands, which is the place where the association was originally (invalidly) described. Instead, the subassociation Lisymachio foemine-Atractylidetum cancellatae ononidetosum reclinatae is restricted to the limestone substrates of the Murgia plateau of the Matera administrative territory.

The Lisymachio foemine-Atractylidetum cancellatae is strictly limited to southern Italy where A. cancellata range extends. At present, we have no knowledge about other associations in which A. cancellata acts as dominant species. The only reference for the Italian peninsula is the association Atractylido-Neatostometum apuli described for western Sicily (Brullo et al. 1994) which exhibit some similarities due to a shared Brachypodietalia distachyi floristic background. However, Atractylido-Neatostometum apuli exhibits a higher sub-halophytic and xerophilous character, probably related to the rocky coastal habitats in which this association was described, and a floristic composition quite different from that of Lysimachio-Atractylidetum cancellatae (see A-N synoptic column in Table 1). In addition to the lack of Neatostoma apulum, there are other high-frequency and characteristic species of Atractylido-Neatosometum apuli that are missing in Lysimachio foeminae-Atractylidetum cancellatae (e.g., Anthemis arvensis, Anisantha fasciculata, Catapodium marinum, Kundmannia sicula, Moraea sisyrinchium, Sulla spinosissima) which do not allow to classify our relevés under this community.

The classification of Lysimachio foeminae-Atractylidetum cancellatae at syntaxonomic ranks higher than association is rather complex. The ephemeral communities that develop on limestone substrates or more generally on the neutral-alkaline ones that do not have an extremely limiting factor in the mineral component of the soil tend to be dominated by a group of wide-distribution steno-Mediterranean species occurring throughout the whole Mediterranean coastal and subcoastal areas. These communities are often distinguished each-other by differences in the abundance-dominance values of the species more than by the occurrence of unshared diagnostic characteristic species. As regards the classification of the Lysimachio-Atractylydetum cancellatae in the higher-rank syntaxa, two different options are available: i) to adopt a conservative choice considering few large-scale syntaxa well-distinguishable from a floristic and chorological point of view, ii) to opt for a more divisive choice considering a syntaxonomic framework based on a higher number of alliances having a more or less similar floristic composition. De Foucault (1999) followed this second option distributing the European therophytic dry grasslands into many alliances and orders. In our specific case, we included Lysimachio-Atractylidetum into the Stipo-Trachynietea distachyae class, whose verbal diagnosis (“Mediterranean calciphilous annual and ephemeroid swards and grasslands”, Mucina et al. 2016) is well suited for the community in issue. As regards the rank of order, Rivas-Martínez (1978) classified Atractylis cancellata among the characteristic species of the order Trachynietalia distachyae. In contrast, Brullo et al. (2001) classified A. cancellata as characteristic species of the order Stipo-Bupleuretalia semicompositi which they considered as xerothermic variant of the Trachynietalia distachyae. Biondi et al. (2014) confirmed the syntaxonomic framework of Brullo et al. (2001) except for the rank of class where the Stipo-Trachynietea (sub: Stipo-Brachypodietea distachyae (Br.-Bl. in Br.-Bl., Emberger & Molinier 1947) Brullo 1985) was considered a later synonym of Helianthemetea guttati (sub: Tuberarietea guttatae (Br.-Bl. in Br.-Bl., Roussine & Négre 1952) Rivas Goday & Rivas-Mart. 1963). On the other hand, Mucina et al. (2016), who considered the Trachynietalia distachyae (synonym of Brachypodietalia distachyi) as a strictly W-Mediterranean order, advanced the hypothesis of classifying the Stipo-Bupleuretalia semicompositi in the Saginetea maritimae this order being typified by the Plantagini-Catapodion marini, an alliance comprising the sub-halophilous therophitic communities of the coastal areas. Accordingly, Mucina et al. (2016) proposed the new order Ptilostemono stellati-Vulpietalia ciliatae for “the Central and Eastern Mediterranean therophytic swards on shallow sandy and loamy soils over limestone and gypsum substrates”. In this new provisional order the authors included the following alliances: Onobrychido-Ptilostemonion stellati, Vulpio ciliatae-Crepidion neglectae, Xeranthemion annui, Vulpion ligusticae. It is interesting to note that, basing on different reasons, the Prodrome of Italian Vegetation (Biondi et al. 2014) and the Eurovegchecklist (Mucina et al. 2016) considered the xerothermic therophytic communities of the central Mediterranean as belonging to an order other than Trachynietalia distachyae (Brachypodietalia distachyi). Mucina et al. (2016) based their hypothesis on biogeographical reasons (Brachypodietalia distachyi was classified as a strictly W-Mediterranean order) whereas Biondi et al. (2014) based their hypothesis on bioclimatic reasons (Trachynietalia distachyae would not be occurring in dry Thermo and Infra-Mediterranean bioclimates where it would be substituted by Stipo-Bupleuretalia semicompositi). In our opinion, the main problem is that Stipo-Bupleuretalia and Brachypodietalia distachyi exhibit a clear floristic-ecological-biogeographical overlapping. It is not by chance that De Foucault (1999) considered the alliance Trachynion distachyae (proposed under the new name Sideritido-Brachypodion distachyi) as partly belonging to the Stipo-Bupleuretalia semicompositi and partly to the Trachynietalia distachyae. The syn-chorological question is therefore to establish whether Brachypodietalia distachyi (= Trachynietalia distachyae syntax. syn.) could be a proper reference for the therophytic communities of southern Italy, considering that it was already and abundantly used for similar communities in central Italy (Filesi et al. 1996; Biondi et al. 1997; Di Pietro et al. 2002; Fanelli et al. 2010). Instead, the ecological question is to establish whether Brachypodietalia distachyi is sufficiently "xerophilous" to include the highly xerothermic communities of southern Italy. To answer these questions, we can do nothing but base ourselves on the original diagnosis of the syntaxa in question and on their groups of characteristic species. The order Trachynietalia distachyae was originally defined as including all the therophytic communities occurring throughout the Mediterranean basin on neutral-basic substrates (Rivas-Martínez 1978). Among the characteristic species of this order are included the majority of the annual species we found in the Lysimachio-Atractylidetum cancellatae, such as Atractylis cancellata, Brachypodium distachyon, Catapodium rigidum, Crupina crupinastrum, Euphorbia exigua, Linum strictum, Stachys romana (etc.). The original diagnosis of the alliance Trachynion distachyae (sensu Rivas-Martínez 1978) perfectly overlapped the diagnosis of the order, as regards the chorological aspects, where instead it differed from it in ecological terms. In fact, the ecological range of Thero-Brachypodion was originally defined as including only the therophytic communities distributed within the potential vegetation areas for the Quercetalia ilicis and the thermophilous fringe of Quercetalia pubescenti-petraeae. Instead, the therophytic communities developed within the potential vegetation areas for the Pistacio-Rhamnetalia alaterni were referred to the Stipion retortae. Among the characteristic species of the Thero-Brachypodion there are taxa (Bupleurum baldense, Campanula erinus, Ononis reclinata) which occur in our samples, together with taxa (Euphorbia sulcata, Dianthus nudiflorus, Hornungia petraea and Neatostema apulum) which occur in Italy but not in our samples, and other taxa (the majority) which exhibited a W-Mediterranean distribution non including Italy (e.g., Arabis parvula, Astragalus stella, Brachyapium dichotum, etc.). On the other hand, Stipo-Bupleuretalia semicompositi is typified by a sub-halophilous coastal alliance (Sedo-Ctenopsion gypsophilae), but includes also many "non halophilous" associations belonging to the thermo-Mediterranean bioclimate. In the Aspromonte massif (Brullo et al. 2001), Atractylis cancellata was reported as a high-frequency species in the communities of the alliance Onobrychido-Ptilostemonion stellati. This alliance, which included the neutral and basiphilous therophytic communities of the Infra- and Thermo-Mediterranean bioclimates was proposed as a geographical vicariant of the Stipion retortae in the central Mediterranean. However only Ptilostemon stellatus, out of the seven characteristic species designated in the diagnosis (Brullo et al. 2001), exhibits a distribution restricted to C-Mediterranean (southern Italy and W-Balkans) whereas the others (Onobrychis caput-galli, Crucianella angustifolia, Hippocrepis ciliata, Filago eriocephala, etc.) exhibited a wide Mediterranean distribution.

The associations belonging to the Onobrychido-Ptilostemonion stellati exhibit a clear prevalence of xerophytic therophytes (e.g., Stipellula capensis, Trifolium scabrum, Evax pygmaea, Ononis reclinata, Plantago lagopus) which are also common in the most of the Brachypodietalia distachyi communities. It is therefore confirmed that Brachypodietalia distachyi and Stipo-Bupleruetalia semicompositi are widely overlapping and the presumed vicariant distribution areas of these two syntaxa (i.e., Brachypodietalia distachyi in the western Mediterranean and Stipo-Bupleuretalia semicompositi in the central-eastern Mediterranean) are not consistent in floristic terms. Atractylis cancellata being originally considered a characteristic species of the Brachypodietalia distachyi and the latter order having nomenclatural priority over Stipo-Bupleuretalia semicompositi we have opted to use Brachypodietalia distachyi as a reference for Lysimachio-Atractylidetum cancellatae and to consider Stipo-Bupleuretalia semicompositi as a pro-parte syntaxonomic synonym. As regards the choice of the alliance we have decided to assign a surplus of diagnostic value (both in ecological and biogeographic terms) to the guide-species of the association. This procedure is not a novelty in terms of syntaxonomic classification. For example, the order Pinetalia halepensis is essentially based on the dominance of Pinus halepensis in communities almost completely composed of Pistacio-Rhamnetalia and Quercetalia ilicis characteristic species (see Pesaresi et al. 2017). As regards grasslands the most emblematic case is probably the decisive role played by the circumboreal Carex myosuroides (= Kobresia myosuroides) in the decision of using the class Carici-Kobresietea instead of Elyno-Seslerietea for the classification of the rush-sward grasslands of the central Apennines although the latter were almost completely composed of SE-European orophytes (Biondi et al. 2000; Blasi et al. 2003; Chytry et al. 2015). For this reason, the large dominance of Atractylis cancellata in the Lysimachio-Atractylidetum does not allow, in our opinion, to classify this association neither in the Trachynion distachyae nor in other alliances considered by Mucina et al. (2016) as possibly occurring in Italy, such as Vulpio ciliatae-Crepidion neglectae and Vulpion ligusticae. In fact, A. cancellata is a South-Mediterranean/Irano-Turanian species typical of habitats which are too dry and warm to be related to the original diagnosis of the Trachynion distachyae (Izco 1974; Rivas-Martinez 1976; Lapraz 1982; Izco et al. 1986). It is not by chance that Guinochet (1978) proposed the alliance Atractylido cancellatae-Stipion retortae Guinochet 1978 for the dry sub-steppic territories of north-Africa. For this reason, we classified the Lysimachio-Atractylidetum in the Onobrychido-Ptilostemonion stellati and proposed the displacement of this alliance from Stipo-Bupleretalia semicompositi to the Brachypodietalia distachyi. The Onobrychido-Ptilostemonion stellati was originally described as an alliance not specifically linked to a particular type of substrate. For this reason, it appears as well suited to host the Lysimachio-Atractylidetum cancellatae, this association being found on different bedrock types, such as Pliocene clays, Quaternary sands, and limestones. In chorological terms the Onobrychido-Ptilostemonion stellati acts as a central Mediterranean vicariant of the W-Mediterranean Stipion retortae and the north-eastern-African Atractylido-Stipion retortae Guinochet 1978. However, adopting a less divisive syntaxonomic interpretation, these three alliances could be brought together within a single central and western Mediterranean alliance whose nomenclatural priority would go to the Stipion retortae. In fact, precisely to the Stipion retortae made reference Guarino and Pasta (2017) in their classification of the therophytic ephemeral swards of the alkaline loamy and clayey substrates of western Sicily.

Campanulo erini-Micromerietum microphylli ass nov. (Table 2)

Name - Campanulo erini-Micromerietum microphylli Di Pietro, Misano, Fortini et Terzi ass nov. (Holotypus rel. 6, Table 2).

Characteristic taxa - Micromeria microphylla, Campanula erinus, Festuca danthonii subsp. danthonii, Festuca myuros subsp. myuros, Silene conica.

Dominant taxa - Micromeria microphylla, Medicago minima, Stipellula capensis, Trifolium scabrum, Poa bulbosa, Helianthemum salicifolium, Ononis reclinata.

High frequency taxa - Micromeria microphylla, Campanula erinus, Ononis reclinata, Silene conica, Medicago minima, Erodium cicutarium, Trifolium scabrum, Stipellula capensis.

Synecology - Small-size communities (1 m2 or less) developed on the extremely superficial substrates of the limestone outcrops of the Murgia hills where they are found on flattish sites characterized by extremely shallow soils (1-2 cm deep) (Fig. 8). The community exhibits the co-dominance of Micromeria microphylla and various annual species. Micromeria microphylla in addition of being the syntaxon name-giving species is the only high-frequency chamaephyte occurring in the community (the other one, Ajuga iva, occurs in two relevés only). Where soils are slightly deeper, the Campanulo-Micromerietum microphyllae tends to be colonized by taller annual grasses, such in particular Stipellula capensis (sub-nitrophilous variant with Stipellula capensis, Table 2). The preferential location of the community in environments not completely free from anthropic disturbance (especially human and animal trampling) justifies the occurrence of some sub-nitrophilous annual species, such as Erodium cicutarium, Plantago lagopus, Rostraria cristata, Sonchus tenerrimus.

Life forms and chorology - Therophytes dominate in all the life forms spectra. These are followed by the chamaephytes, whose high values in the cover spectrum are almost exclusively due to the contribution of Micromeria microphylla. In the chorological spectrum Euro-Mediterranean species dominates followed by the steno-Mediterranean ones. Surprisingly high are the percentages of the Eurasiatic component (Fig. 9).

Syndynamic - The evolution of the soils leads the Campanulo erini-Micromerietum microphylli to be substituted by therophytic communities of greater size, as those dominated by Stipellula capensis, and subsequently by Mediterranean garrigues with Corydothymus capitatus and Helianthemum jonium. The final stage of the dynamic succession is a Mediterranean maquis community dominated by Pistacia lentiscus and Phillyrea latifolia.

Syntaxonomy - Despite the apparent floristic simplicity, the Campanulo erini-Micromerietum microphylli is nevertheless problematic as regards its classification at the highest ranks of syntaxonomy. The physiognomic dominance of Micromeria microphylla, although not overwhelming, would suggest a possible classification in the Cisto incani-Ericion multiflorae, in particular in the micro-garrigue sub-alliance Thymo capitati-Helianthemenion jonii Di Pietro and Misano 2010 which was described precisely in the western Murgia Plateau (Di Pietro and Misano 2010). On the other hand, the preferential location on natural calcareous pavement, and the occurrence in the community of Petrorhagia saxifraga, accompanied by therophytes such as Arenaria leptoclados, Campanula erinus, Herniaria glabra, Silene conica, Sabulina tenuifolia and Valerianella muricata suggest to classify the Campanulo erini-Micromerietum microphylli in the Sedo-Scleranthetea. A third hypothesis, would be that of classifying this community in the Trachynietea distachyae taking into account the absolute dominance of the steno-Mediterranean therophytic component (both as number of species and cover percentage). A fourth and last option (unlikely in truth) would be to consider this community in the less synanthropic fringe of Chenopodietea, considering the partial anthropogenic character of the environments in which this community was found. The phytosociological literature is rather scarce as regards communities similar to the Campanulo-Micromerietum. In Sicily Micromeria microphylla is reported as a guide species of the association Putorio calabricae-Micromerietum microphyllae nom. inval. (Brullo and Marcenò 1979; Brullo et al. 2004) which was subsequently validated with the name Micromerio microphyllae-Putorietum calabricae Brullo et Marcenò ex Terzi, Jasprica et Caković 2017 (Terzi et al. 2017). This is a chasmophytic association composed of two subassociations (typicum and chiliadenetosum bocconei) which was classified in the Dianthion rupicolae, an alliance endemic to southern Italy in the class Asplenietea trichomanis (Terzi et al. 2017, 2018). The chasmophytic behavior of Micromeria microphylla was also reported by Lucchese and Pignatti (2013) who recorded the sporadic occurrence of this species within the ancient Roman walls of Ostia Antica (Rome, Italy) and hypothesized its arrival to be related with the transport of marble blocks from distant locations. The authors named the community with dominant Micromeria microphylla occurring on the walls of Ostia Antica simply as "Micromeria microphylla community" and classified it in the Parietarietea judaicae. However, none connection with Asplenietea trichomanis or Parietarietea judaicae, either ecological or floristic, was identifiable for the Campanulo erini-Micromerietum microphylli of the Murgia hills. A classification in the Ononido-Rosmarinetea does not seem feasible too. Although the Ononido-Rosmarinetea communities are abundant in the study area, especially on rocky or pebbly substrates, these are normally characterized by a clear physiognomic dominance of nano-phanerophytes and chamaephyes (Di Pietro and Misano 2010). Looking exclusively at the lithological and geomorphological characteristics of the sampled sites (fissured and flat limestone pavements), the reference to the Sedo-Scleranthetea would seem the most appropriate. However, observing the specific composition of the community, the most typical species of Sedo-Scleranthetea were absent, in particular those belonging to the genus Sedum and therophytes, such as Alyssum alyssoides, Erophyla verna, Hornungia petraea, Saxifraga trydactylites, Thlaspi perfoliatum (etc.). The dominance of steno-Mediterranean therophytes suggested us to classify the Campanulo erini-Micromerietum microphylli in the Stipo-Trachynietea distachyae (Brachypodietalia distachyae, Trachynion distachyae). In fact, although, Micromeria microphylla is the species displaying the highest specific cover index considering the entire phytosociological table, its cover values in several samples are comparable to those of some therophytes, such as Campanula erinus, Medicago minima or Stipellula capensis. Moreover, the extremely small size of Micromeria microphylla make this species physiognomically and functionally comparable to a small therophyte. However, the doubt in the choice between Trachynietea distachyae and Sedo-Sclerantherea is legitimate and recurrent. In particular, coenological overlaps and classification difficulties between Sedo-Scleranthetea and Trachynietea distachyae are well-known especially as regards those communities in which the dominant guide-species are therophytes (i.e., Chaenorhino-Saxifragetum trydactylites, Acino-Trifolietum scabri etc.). In our case, the situation is even more complex and would seems to be unbalanced towards the Sedo-Scleranthetea since the community has a micro-chamaephyte as guide-species. However, the aforementioned reasons seem consistent with our choice to opt for the Stipo-Trachynietea. Fanelli (2007), in order to discriminate between the two classes emphasized on the important diagnostic role played by the bryophytic component in the Sedo-Scleranthetea communities, whereas it would be significantly lower in the Stipo-Trachynietea ones. In the case of Campanulo-Micromerietum microphylli the role of the bryophytic component was found to be very low (substancially limited to a low cover percentage of Grimmia pulvinata) and this would further support our choice to classify the community in the Stipo-Trachynietea.

STIPO AUSTROITALICAE-HYPARRHENIETUM HIRTAE Biondi et Guerra 2008 (Table 3)

The communities dominated by Hyparrhenia hirta characterize the lower parts of the Murgia of Gravina and replace the grasslands dominated by Stipa austroitalica and Scorzonera villosa subsp. columnae of the summit areas of the Murgia plateau. These communities exhibits intermediate floristic features between three classes (Festuco-Brometea, Lygeo-Stipetea and Festuco hystricis-Ononidetea striatae) and testify the coenological transition between the Apennine-Balkan biogeographical province (Euro-Siberian Region) and the Adriatic province (Mediterranean Region). This transition is clearly observable moving from the summit areas of the Murgia plateau, still affected by cold winds from north-eastern Europe, to the lowlands facing the Taranto Gulf where the influence of the warm winds of south-western origin occurs. Hyparrhenia hirta is absolutely the dominant species of these xerothermic steppe-like communities especially when developed on partially leached soils rich in fine debris. Stipa austroitalica is constantly occurring in the dominant tall grasses layer, whereas Scorzonera villosa subsp. columnae do the same in the undergrowth. In addition, the Hyparrhenia hirta communities host some species typical of the Hippocrepido-Stipion austroitalicae dry-grasslands (Festuco-Brometea) such as Hippocrepis glauca, Anthyllis vulneraria subsp. rubriflora, Convolvulus elegantissimus. The steno-Mediterranean and Euro-Mediterranean component however remains dominant and would justify the classification of this community in the Lygeo-Stipetea, Hyparrhenietalia hirtae and Hyparrhenion hirtae as already proposed in Biondi and Guerra (2008). As regards the rank of order we object to a purely nomenclatural question, namely we have followed Mucina et al. (2016) which consider Cymbopogono-Brachypodietalia ramosi to be a prior syntaxonomic synonym of Hyparrhenietalia hirtae Rivas-Mart. 1978 (nom. inval.). Instead, the choice of the alliance is a more complex question. Hyparrhenion hirtae and Cymbopogono-Brachypodion ramosi are both accepted in the Eurovegchecklist as geographically vicariant alliances, the first occurring in the western Mediterranean and the southern Regions of the central Mediterranean whereas the second being restricted to the eastern Mediterranean. In our opinion, the classification of the Stipo austroitalicae-Hyparrhenietum should be based on comparing the floristic-coenological characteristics of the sampled communities with the original diagnoses of the two aforementioned alliances. Accordingly, we noticed that our Hyparrhenia hirta grasslands are characterized by the simultaneous occurrence of two taxa (Stipa austroitalica and Scorzonera villosa subsp. columnae) belonging to the Stipa pennata s.l. and Scorzonera villosa s.l. collective taxa respectively. This is a typical feature of the eastern Adriatic Mediterranean steppe-like grasslands too, where the two taxa in issue are Scorzonera villosa subsp. villosa and Stipa eriocaulis. Moreover, in the Stipo austroitalicae-Hyparrhenietum hirtae occur East-European or amphi-Adriatic species (Dasypyrum villosum, Charybdis pancration, Hippocrepis glauca, Melica transsylvanica, Thapsia garganica, Asyneuma limonifolium, Centaurea deusta, Ornitoghalum gussonei) which are species not included in the Hyparrhenion hirtae range. For this reason, we classified the Stipo austroitalicae-Hyparrhenietum in the Cymbopogono-Brachypodion ramosi. However, considering that the above-mentioned differential species of Stipo-Hyparrhenietum are all belonging to classes other than Lygeo-Stipetea, a different framework could be hypothesized. In practice it would be a question of considering a single Mediterranean alliance for both “eastern” and “western” Hyparrhenia hirta communities and delegating possible biogeographic differentiation to the lower syntaxonomic ranks, such as suballiances or associations. This consideration would be also supported by the fact that the species which were originally considered as characteristic species of both Cymbopogono-Brachypodion ramosi and Cymbopogono-Brachypodietalia ramosi (Allium sphaerocephalon, Briza maxima, Crucianella latifolia, Heteropogon contortus, Hyparrhenia hirta, Medicago minima, Lagurus ovatus, Linum strictum and Lotus edulis), display a wide Mediterranean distribution and do not suggest Cymbopogono-Brachypodion ramosi as an alliance restricted solely to the central-eastern Mediterranean. The situation does not change even adding the species considered by Horvatić (1973; 1975) as transgressive or ingressive in Cymbopogono-Brachypodietalia ramosi (e.g., Allium subhirsutum, Bituminaria bituminosa, Carex divulsa, Clinopodium nepeta, Filago germanica, Hedypnois rhagadioloides, Hymenocarpos circinnatus, Lotus ornithopodioides, Oloptum miliaceum, Scorpiurus subvillosus, Trifolium angustifolium, Trifolium scabrum). In fact, more than the list of characteristic species of Cymbopogono-Brachypodion, it was the great biogeographic selectivity of the characteristic species of the Hyparrhenion hirtae (see Diaz-Garretas and Asensi 1999) that discouraged Hyparrhenion hirtae to be used as syntaxonomic reference for our relevés. In addition to species occurring throughout the whole Mediterranean basin (e.g. Andropogon distachyos, Hyparrhenia hirta, Heteropogon contortus, Bothriochloa ischaemum, Convolvulus althaeoides), the Hyparrhenion hirtae exhibits characteristic species which are strictly western Mediterranean (Daucus crinitus and D. setifolius) or species not occurring in the Adriatic area (Aristida adscensionis, Carex depressa, Tricholaena teneriffae), or not occurring in Italy (Eragrostis papposa) or restricted to the coastal areas of northern Africa (Dichanthium annulatum, Cenchrus ciliaris, Stipagrostis sahelica, Tetrapogon villosus). Accordingly, the classification of Stipo austroitalicae-Hyparrhenietum hirtae in Cymbopogono-Brachypodion ramosi seemed the most logical choice. The phytosociological literature of the Italian Peninsula does not currently report any reference to Cymbopogono-Brachypodion ramosi. This alliance is not taken into consideration in Biondi et al. (2014) not even as a possible synonym of Hyparrhenion hirtae. In fact, references to this alliance are all coming from the coastal sectors of the Dinarids or from Albania (Hoda and Mersinllari 2000). However, it is possible that the Thymo-Hyparrhenietum hirtae described in Di Pietro et al. (2002) for the coastal sectors of southern Lazio, originally classified in the Hyparrhenion hirtae, could be moved in the Cymbopogono-Brachypodion by virtue of its successional and/or spatial link with eastern European potential forest vegetation types (Blasi and Di Pietro 1998).

MEDICAGO DISCIFORMIS-BRACHYPODIETUM DISTACHYI Biondi et Guerra 2008 (Table 4, rels 1–4)

This community shows a clear dominance of Brachypodium distachyum and characterizes small calcareous outcrops or preferential walking sites for animals within the perennial communities dominated by Scorzonera villosa subsp. columnae, Stipa austroitalica and Hyparrhenia hirta. Due to the few relevès carried out in this paper, it does not seem appropriate to promote a too deep syntaxonomic discussion. However, as already highlighted in the previous paragraph, the syntaxonomic classification of the Mediterranean therophytic communities in Italy still remains a heated point of debate (see Biondi et al. 1997; Di Pietro et al. 2002; Fanelli et al. 2010; Di Pietro et al. 2017). In this case we have decided to make reference to the Medicago disciformis-Brachypodietum distachyi, which was already described for the study area (Biondi and Guerra 2008). This association is probably the only therophytic association of the Trachynietea distachyae described in Italy displaying a clear dominance of Brachypodium distachyum. Precisely the dominance of Brachypodium distachyon suggested us to include this association in the Trachynion distachyae. Instead, Biondi and Guerra (2008) proposed to classify this association in the new alliance Hypochaeridion achyrophori Biondi & Guerra 2008, (eastern geo-vicariant of Trachynion distachyae) having the following characteristic species: Hypochoeris achyrophorus, Ononis reclinata, Lotus ornithopodioides, Coronilla scorpioides. In particular, the therophitic communities of the Murgia hills were included in the suballiance Ononidenion ornithopodioidis Biondi & Guerra 2008 whose differential species were selected in Ononis ornithopodioides, Trifolium lucanicum, Bromus fasciculatus, Odontites lutea, Ammoides pusilla, Romulea bulbocodium and Convolvulus elegantissimus. In our opinion, the whole characteristic component of the Hypochaeridion achyrophori does not identify any geographical peculiarity since all the characteristic species display a wide Mediterranean distribution. Accordingly, we don't think there is any reason to distinguish floristically and biogeographically Hypochaeridion achyrophori from the Trachynion distachyae. According to Mucina et al. (2016) the Hypochaeridion achyrophori, as well as being invalid, is to be considered a syntaxonomic synonym of Vulpio ciliatae-Crepidion neglectae Poldini 1989. Where small depressions are formed on the limestone crust with greater soil coverage, especially in areas where the passage of grazing animals is more frequent, annual species with slightly more anthropogenic character, such as Anisantha madritensis and Lagurus ovatus take over as dominant (Table 4, rels 5–6). In syntaxonomic terms, these communities can be interpreted as the transition between Brachypodietalia distachyi and Brometalia rubenti-tectorum (Chenopodietea).

PETROSEDUM OCHROLEUCUM SUBSP. MEDITERRANEUM community (Table 5)

In environmental situations of undulating rocky outcrops characterized by micro-depressions in the calcareous plateau profile, communities with a clear dominance of Petrosedum ochroleucum subsp. mediterraneum accompanied by a few other chamaephytes and a rich therophytic component are found. The degree of vegetation cover is quite high as there are very few parts of the sampling plots not covered by a thin layer of soil at least. From a syntaxonomic point of view, the coenological role of Petrosedum ochroleucum was already known in the study area, this taxon being one of the guide species of the association Sedo ochroleuci-Saturejetum cuneifoliae already described for the small-size terraces bordering the Gravina gorges (Di Pietro and Misano 2010). However, this reference cannot be used for our relevés the habitat type of the Sedo-Saturejetum cuneifoliae being quite different and Satureja cuneifolia being completely absent in our relevés. In the Italian peninsula, other communities dominated by species belonging to the genus Petrosedum have already been described, such as the Sedetum rupestris-sexangularis Di Pietro et al. 2006 of the Prenestini mountains in the Lazio region pre-Apennines (Di Pietro et al. 2006) and the Linario purpureae-Petrosedetum rupestris Ciaschetti et al. 2020 of the hilly and lower-montane belts of the Abruzzo Apennines (Ciaschetti et al. 2020). However, both these communities have different guide species (e.g., Petrosedum rupestris) and exhibit quite different floristic compositions compared with those of the Apulian communities. A suitable syntaxonomic reference already published not being available and our relevés being too few to propose a new association, we preferred to preliminary name this community simply as Petrosedum ochroleucum subsp. mediterraneum comm. In terms of higher synatxonomical ranks, the most plausible references are Alysso-Sedion, Alysso-Sedetalia and Sedo-Scleranthetea.